Geofluids: origin, migration and evolution of fluids in sedimentary basins

Geofluids: Origin, Migration and Evolution of Fluids in Sedimentary Basins

Geological Society Special Publications Series Editor A . J .

FLEET

GEOLOGICAL SOCIETY SPECIAL PUBLICATION NO. 78

Geofluids: Origin, Migration and Evolution of Fluids in Sedimentary Basins EDITED

BY

JOHN PARNELL School of Geosciences The Queen's University of Belfast, UK

1994 Published by The Geological Society London

THE GEOLOGICAL SOCIETY

The Society was founded in 1807 as the Geological Society of London and is the oldest geological society in the world. It received its Royal Charter in 1825 for the purpose of 'investigating the mineral structure of the Earth'. The Society is Britain's national society for geology with a Membership of 7500 (1993). It has countrywide coverage and approximately 1000 members reside overseas. The Society is responsible for all aspects of the geological sciences including professional matters. The Society has its own publishing house which produces the Society's international journals, books and maps, and which acts as the European distributor for publications of the American Association of Petroleum Geologists and the Geological Society of America. Fellowship is open to those holding a recognized honours degree in geology or cognate subject and who have at least two years relevant postgraduate experience, or who have not less than six years relevant experience in geology or a cognate subject. A Fellow who has not less than five years relevant postgraduate experience in the practice of geology may apply for validation and subject to approval, may be able to use the designatory letters C. Geol (Chartered Geologist). Further information about the Society is available from the Membership Manager, The Geological Society, Burlington House, Piccadilly, London WlV 0JU, UK. Published by The Geological Society from: The Geological Society Publishing House Unit 7 Brassmill Enterprise Centre Brassmill Lane Bath BAI 3JN UK (Orders: Tel. 0225 445046 Fax 0225 442836) First published 1994 © The Geological Society 1994. All rights reserved. No reproduction, copy or transmission of this publication may be made without written permission. No paragraph of this publication may be reproduced, copied or transmitted save with the provisions of the Copyright Licensing Agency, 90 Tottenham Court Road, London WlP 9HE, UK. Users registered with Copyright Clearance Center, 27 Congress Street, Salem, MA 01970, USA: the item-fee code for this publication is 0305-8719/94 $7.00. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library ISBN 1-897799-05-5 Typeset by Type Study, Scarborough Printed by Alden Press, Oxford, UK

Distributors USA AAPG Bookstore PO Box 979 Tulsa Oklahoma 74101-0979 USA (Orders: Tel. (918)584-2555 Fax (918)584-0469) Australia Australian Mineral Foundation 63 Conyngham Street Glenside South Australia 5065 Australia (Orders: Tel. (08)379-0444 Fax (08)379-4634) India Affiliated East-West Press PVT Ltd G-l/16 Ansari Road New Delhi 110 002 India (Orders: Tel. (1 !)327-9113 Fax (11)326-0538) Japan Kanda Book Trading Company Tanikawa Bldg 3-2 Kanda Surugadai Chiyoda-Ku Tokyo 101 Japan (Orders: Tel. (03) 3255-3497 Fax (03) 3255-3495)

Contents

Preface

vii

FYFE, W.S. The water inventory of earth: fluids and tectonics Large-scale fluid flow

VAN BALEN, R. & CLOETINGH, S. Tectonic control of the sedimentary record and stress-induced fluid flow: constraints from basin modelling

9

DEMING, D. Fluid flow and heat transport in the upper continental crust

27

JEssoP, A.M. & MAJOROWlCZ,J.A. Fluid flow and heat transfer in sedimentary basins

43

PmLLIpS, G.N., WILLIAMS, P.J. & DE JoNfi, G. The nature of metamorphic fluids and significance for metal exploration

55

Deformation and fluid flow

SmSON, R.H. Crustal stress, faulting and fluid flow

69

MUIR WOOD, R. Earthquakes, strain-cycling and the mobilization of fluids

85

KNIPE, R.J. & MCCAIG, A.M. Microstructural and microchemical consequences of fluid flow in deforming rocks

99

STEPHENSON,E.L., MALTMAN,A.J. & KNIPE,R.J. Fluid flow in actively deforming sediments:

ll3

'dynamic permeability' in accretionary prisms Fluid flow and reservoir evolution

BJORLYKKE,K. Fluid-flow processes and diagenesis in sedimentary basins

127

RINGROSE, P.S. t~ CORBET/',P.W.M. Controls on two-phase fluid flow in heterogeneous

141

sandstones Fluid chemistry; metal-organic interactions

HANOR, J.S. Origin of saline fluids in sedimentary basins

151

GIORDANO,T.H. & KHARAKA,Y.K. Organic ligand distribution and speciation in sedimentary basin brines, diagenetic fluids and related ore solutions

175

FILBY, R.H. Origin and nature of trace element species in crude oils, bitumens and kerogens: implications for correlation and other geochemical studies

203

NICHOLSON, K. Fluid chemistry and hydrological regimes in geothermal systems: a possible link between gold-depositing and hydrocarbon-bearing aqueous systems

221

Fluid evolution: migration and precipitation of hydrocarbons and metals

MANN, U. An integrated approach to the study of primary petroleum migration

233

SIMONEIT,B.R.T. Organic matter alteration and fluid migration in hydrothermal systems

261

PARNELL, J. Hydrocarbons and other fluids: paragenesis, interactions and exploration potential inferred from petrographic studies

275

vi

CONTENTS

FOWLER, A.D. The role of geopressure zones in the formation of hydrothermal Pb-Zn Mississippi Valley type mineralization in sedimentary basins

293

METCALFE, R., ROCHELLE, C.A., SAVAGE,D. • HIGGO, J.W. Fluid-rock interactions during continental red bed diagenesis: implications for theoretical models of mineralization in sedimentary basins

301

Tracers of fluid evolution

DUDDY, I.R., GREEN, P.F., BRAY, R.J. & HEGARTY,K.A. Recognition of the thermal effects of fluid flow in sedimentary basins

325

BALLENTINE,C.J. & O'NIONS,R.K. The use of natural He, Ne and Ar isotopes to study

347

hydrocarbon-related fluid provenance, migration and mass balance in sedimentary basins

Geofluids: introduction JOHN PARNELL

School of Geosciences, Queen's University of Belfast, Belfast B T71NN, UK

Geological fluids are a central theme linking the petrography and chemistry of all rock types, deformation processes on the microscopic to the continental scale, and the concentration of economic resources. The fundamental importance of fluid migration and evolution to rock composition and structure is reflected in a growing interest in fluid processes, including a series of successful conferences on water-rock interaction (Kharaka & Maest 1992). The papers in this volume are intended to give a state-of-theart review of the whole spectrum of geofluids research. In an introductory account, Fyfe summarizes the water inventory of the planet Earth, and emphasizes the importance of quantifying water fluxes at all levels within the crust, and in particular the fluxes resulting from subduction and continental collision. The relationship between fluid migration and heat flow helps to explain why fluid fluxes are of fundamental importance to the formation of mineral resources.

Large-scale fluid flow In the past decade several models have been proposed for large-scale fluid flow at continental margins and across continental interiors. Van Balen & Cloetingh describe the constraints imposed by basin modelling upon theories for tectonic control of the sedimentary record and stress-induced fluid flow. They use a dynamic numerical model to investigate the effect of short-term variations in the level of intraplate stresses on fluid flow and sedimentation patterns. Increases in stress strongly influence the hydrodynamic regime during the post-rift phase of basins by causing an increase in meteoric water influx and compactional flow. Deming shows that the magnitude of convective heat transport through the upper continental crust is exponentially dependent upon fluid velocity and depth of fluid circulation. The major mechanisms for fluid flow in the upper crust are topography, fluid released by sediment compaction, phase changes or metamorphism, and free convection. The most effective mechanism for transporting heat is topographically-driven flow,

which may be responsible for some ore-forming processes in the North American mid-continent. Jessop & Majorowicz emphasize that heat transport through fluid flow is as effective as conduction, leading to wide contrasts in both lateral and vertical heat flow in some basins. This applies not just to basins above sea level with an obvious topographic driving force, but also to some sub-sea basins. Phillips et al. review the nature of metamorphic fluids, and their role in ore formation. Metamorphic fluids in many gold provinces are dominated by water, carbon dioxide and hydrogen sulphide, with low salinity, reflecting an abundance of mica, carbonate and sulphide in the rocks. A less common metamorphic fluid involving evaporite-bearing sequences is saline, with the potential to transport both gold and base metals.

Deformation and fluid flow Several contributions examine the interrelationships between deformation and fluid flow, in which fluids help to enable deformation, and faulting helps to facilitate fluid migration. Sibson discusses crustal stress, faulting and fluid flow. Deviatoric stress exerts both static and dynamic effects on rock permeability and fluid flow, modulating flow systems in the Earth's crust. Textural evidence from hydrothermal veins suggests that fluid flow in fault-related fracture systems generally occurs episodically, and that stress cycling effects may be widespread. Muir Wood shows how empirical observations of hydrological changes following major earthquakes allow the prediction of subsurface fluid flow during active tectonism. The type of change is dependent on the style of fault displacement: normal faults displace large volumes of fluid from the crust, while reverse faults draw fluids into the crust. Knipe & McCaig review the interactions between deformation and fluid flow. They show that the various deformation mechanisms possible in rocks have different effects on fluid flow which depend upon the associated volume changes. Microstructural analysis of deformed rocks provides information on fluid flow pathways, fluid chemistry and the amount of fluid involved. Stephenson et al.

viii

PREFACE

discuss the importance of inter-related fluid flow and deformation during the evolution of accretionary prisms. With the aid of experimental data, they show that the permeability measured in an actively deforming material contains a dynamic component in addition to the classical notion of capacity to transmit fluid.

Fluid flow and reservoir evolution Inevitably some of the most detailed studies of fluid flow in the past decade have been those related to the evolution of hydrocarbon reservoirs. Bj0rlykke relates the flow of fluids through basins to transport of heat and dissolved ions, and consequent diagenetic reactions. The greatest potential for transporting mass and creating secondary porosity is through meteoric water flow, as the flow rate may be very substantially greater than typical compaction-driven flow. The flow behaviour of immiscible fluids in permeable sandstones is assessed by Ringrose & Corbett, who show that capillary forces result in significant amounts of both trapping and bypassing of the non-wetting phase. In a typical water-wet oil/water system the amount of trapped oil varies between 38% and 65% depending upon the patterns of rock heterogeneity.

Fluid chemistry; metal-organic interactions Research on the chemistry of groundwaters and hydrothermal fluids in sedimentary basins has highlighted the role of organic species in complexing with organic species. Hanor reviews the origins of saline brines in sedimentary basins. Thermodynamic buffering by silicatecarbonate-(halide) mineral assemblages is a first-order control on subsurface fluid compositions. Where fluid composition is rockbuffered its ultimate origin may be obscured by its most recent history; however some nonbuffered components, such as chlorine and bromine, can be useful in providing information on the original end-member fluid compositions. Giordano & Kharaka discuss the diagenetic processes involving dissolved acids. The dissolved acids are important as a control on pH and buffer capacity, as organic ligands to form aqueous complexes with metals and other inorganic species, as reducing agents controlling the Eh of fluids, and through breakdown as a source of carbon dioxide and hydrocarbon species. Filby describes the origin and nature of trace element species in crude oils, bitumens and

kerogens. Nickel and vanadium metalloporphyrins are formed during sedimentation/early diagenesis of oil source rocks, and the relative abundances of the metals are related to depositional environment. Complexes of other trace elements in crude oils may be primary, including products of mineral-kerogen reactions, or secondary, from interactions between oils with mineral matter or formation waters during migration, maturation or biodegradation. Nicholson describes compositions of geothermal fluids which range from gold-depositing dilute waters to saline, oilfield brines, and proposes that techniques used to study active geothermal systems may be applicable to both gold exploration and hydrocarbon reservoir modelling.

Fluid evolution: migration and precipitation of hydrocarbons and metals Fluids have a fundamental role as the agents of migration and concentration of hydrocarbons and metals. Mann presents an approach to the study of primary petroleum migration, integrating sedimentological, petrophysical, organic geochemical and numerical modelling methods. Primary migration probably proceeds through diffusion into pore/fracture systems where a petroleum bulk phase develops, possibly with aqueous solutions. Simoneit describes the alteration of sedimentary organic matter to petroleum hydrocarbons by reductive reactions in modern hydrothermal systems. This alteration occurs under high pressure, over a wide temperature range, and in a very brief geological time. Petroleum generation, expulsion and migration occurs as a single continuous process during hydrothermal activity. Parnell shows how paragenetic relationships between hydrocarbons and inorganic minerals provide information on the relative timing of hydrocarbon migration and the migration of other fluids. Co-migration of hydrocarbons and aqueous fluids is evinced by the occurrence of large quantities of authigenic silicate minerals within some hydrocarbon residues. Fowler describes the fluid pathways and driving mechanisms for lead-zinc mineralizing brines, and considers the possible role of overpressuring in their formation. Fluid flow must be fast to advect the heat needed in mineralizing fluids from deep within a basin. In shale-dominated basins, overpressured zones immediately below platform carbonate rocks can be a proximal source of heat and metals. The shales act as thermal barriers, and high fluid pressure ruptures the overlying rocks to provide vertical pathways for hot mineralizing brines

PREFACE into carbonate host rocks. Metcalfe et al. describe the chemistry of fluid-rock interactions during continental red bed diagenesis. Models for fluid evolution in this environment can be useful in understanding how the relatively high content of heavy metals in ferromagnesian/ aluminosilicate detritus can be concentrated into red bed-hosted ore deposits.

Tracers of fluid evolution Several highly specialized analytical techniques have evolved which help to trace the pathways and consequences of fluid flow. Duddy et al. explain how apatite fission track analysis can be used to determine palaeotemperature histories. T e m p e r a t u r e ~ l e p t h profiles can be used to distinguish the effects of flow of hot fluid through a basin from heating due to the simple conduction of basal heat flow. Ballentine & O'Nions show how the relative abundance of the rare gases helium, neon and argon in crustal, mantle and atmosphere-derived components of fluids can be distinguished according to their isotopic distribution. This data provides information on the physical processes experienced in the fluid and, combined with mass balance calculations, can be used to constrain fluid provenance and transport.

ix

The Geofluids '93 conference, from which this volume developed, was an initiative in collaboration with Steve Lawrence (Quad Consulting) and Chris Cornford (IGI Ltd), and included a special session on Deformation and Fluid Flow organized by Rob Knipe (University of Leeds). The organizers were ably supported by a committee including Sally Cornford, Bert Kennedy, Richard Bray, Graham Harman, Janet James, Les Oldham, Des Horscroft and Dave Naylor, and also the editorial committee (editor, Aiastair Ruffeii and Norman Moles) and the staff of Quad Consulting and IGI Ltd. The conference was supported by The Geological Society of London, The Institution of Mining and Metallurgy and The Institute of Petroleum. Support from these bodies reflects the wide significance of fluids research and the potential for interchange of approaches between the hydrocarbon and minerals industries. This volume should help to foster a greater appreciation of how geofluids research can contribute to the understanding of the evolution of sedimentary basins and their resource potential.

Reference KHARAKA, Y.F. & MAEST, A.S. 1992. Water-Rock Interaction (2 vols). Balkema, Rotterdam.

The water inventory of the Earth: fluids and tectonics W.S. FYFE

Department of Earth Sciences, University of Western Ontario, London, Ontario, Canada N6A 5B7 Abstract: As the human population continues to expand, to approach ten billion, there will be an increasing demand upon all Earth's resources. Of particular importance will be resources related to energy, soil and water, all of which involve geofluids. Problems associated with waste disposal also involve systems for protecting near-surface water quality. There is an urgent need to quantify the fluid inventory and fluid dynamics of the near surface, while understanding volatiles and their deep recycling is fundamental to our understanding of the major dynamic processes of the Earth.

Over the past century, the growth of human population (now almost 100 million per year), supported by diverse and complex technologies, and increasing expectations of quality of life, have placed vast new demands on Earth resources, which come mainly from the atmosphere, hydrosphere, soil and the top few kilometres of the solid crust. At this time, we are concerned with the need for careful management of our fossil carbon fuel resources from which we derive most of our energy and our usable water resources, which, in many nations, are approaching limits on the local scale. There is little doubt that oil-gas (and coal) will eventually be replaced by other energy sources, such as solar, geothermal and biomass sources. However a limiting resource for human occupation of Earth may well be water. There is an urgent need for the best possible inventories of all such resources, and a growing need to integrate the knowledge from all involved with the study of geofluids. It was, perhaps, with the development of observations from space that we acquired a new sense of the limits and fragility of our wet planet. Studies of crust and hydrosphere history have also revealed that, unlike our nearest planetary neighbours, for about 4 billion years of recorded history, the Earth has always had a massive hydrosphere which has never boiled or totally frozen (Broecker 1985). Earth has some remarkable environmental buffer systems, which we still do not adequately understand. We also know that micro-organisms have been present for essentially all of our recorded history, and new studies (Schopf 1992) suggest that photosynthetic organisms may have been present much earlier than has been suggested by other workers. The outer surface and the top, porouspermeable layers of Earth are wet, and the

recent search for micro-organisms at depth (stimulated by consideration of deep biocorrosion of nuclear waste systems), indicates that rocks may well host life at all places where temperatures do not exceed 100°C (Pederson 1993). Studies of surface heat-flow patterns (for example, the spectacular new results from the German deep drilling, KTB) clearly show that, in the outer layers of Earth, fluids transport a substantial amount of heat out of the crust. The work of Straus & Schubert (1977) showed that the adiabatic gradient for fluid convection in a porous medium is almost always exceed in crustal situations. Studies of fluid convection through the oceanic crust, oceanic heat flow studies and the famous hot and cold discharge systems have shown that the sea floor crust is water-cooled, with something like half the thermal energy removed by fluid convection (Fyfe & Lonsdale 1981; Boul6gne & Pflumio 1992). Another growing world problem, which has focused our attention on deep and near surface fluids and their motions, is that of disposal of wastes of all types (urban, chemical, agricultural, n u c l e a r . . . ) , and the scale of the problems which will be associated with a human population of 10 billion is vast. In most cases, the available technologies are not adequate.

The inventory In general, we have only a moderate knowledge of the global inventory of water-rich fluids. Most accessible, the oceans and ground water are the most massive (Berner & Berner 1987). Of much smaller mass is water in the ice caps (which we must leave alone!) and surface waters in rivers and lakes. There is a massive quantity of water

From PARNELL,J. (ed.), 1994, Geofluids: Origin, Migrationand Evolution of Fluidsin Sedimentary Basins, Geological Society Special Publication No. 78, 1-7.

2

W.S. FYFE

(similar to the ocean mass), contained in the hydrous mineral phases of the crust. While the inventory of volatiles fixed in the crust (H20, CO2, N compounds, halogen compounds) are often not well-quantified, we know that change in the P - T conditions of their environment may lead to their evolution or fixation, processes which often lead to chemical transport related to formation of mineral resources. The rock mechanics of the crust would be very different on a planet without water (Fyfe et al. 1978). However, below the accessible crust, the inventory of volatiles in the deep Earth, mantle and core, is very uncertain. Thompson (1992) has recently reviewed the situation of water in the mantle. There are many ways in which water and other volatiles may be held in the mantle. Thus Fyfe (1970) discusses the substitution of SiO44- by ( 0 H ) 4 4- at high pressures and possible tetrahedral CO44-, Pawley et al. (1993) discuss hydrogen in stishovite and Schrauder & Navon (1993) report solid carbon dioxide in diamond. But, while there is evidence that the core must contain some elements of low atomic number, the candidates (O, H, C, etc.), are less certain. Anderson (1993) has recently reviewed the problem of helium isotopes in deep crustal systems. He proposed that recycling from the surface from solar inputs may be more important than deep primordial sources (see also Fyfe 1987).

The global water cycle Most discussions of the global water cycle are restricted to consideration of atmospherebiosphere-hydrosphere interactions within the top few kilometres of the planet. Certainly, in terms of usable water for humans, these systems dominate our fluid resources. The global state of water resources has recently been reviewed by Postal (1992). But, if we broaden our interests to all geochemical fluxes related to fluid motions and hydrosphere evolution, and particularly the flux of bio-essential elements (Mn, Fe, Co, Cu, Ni, Zn, etc.) into the oceans, it is necessary to consider much deeper interactions (see Gillet 1993). If we wish to quantify the total evolution of the hydrosphere and all volatile systems, over geologic time, the deep systems must be considered. All phenomena which involve mantle convection, and the magmatic-tectonicmetamorphic expressions of such convection (plate tectonics) result in volatile transport.

Rising magma convection cells We are increasingly aware that the nature of the surface of our planet is largely controlled by

heat-mass transfer in the mantle, even down to the core-mantle boundary, and by erosional processes involving surface atmospherehydrosphere-biosphere interactions. The major sites of rising mantle convection cells are the great ocean ridge systems, where new basaltic crust is formed. It is also recognized that the same processes, but certainly with different spacing and higher intensity, operated in the ancient crust (see Kroner & Lager 1992; Fyfe 1974). The water cooling processes which transfer almost half the energy from the ridge systems were first modeled by Lister (1977) and Davis & Lister (1977), whose general ideas have been well confirmed by a host of later direct observations. When magmas cool, they contract, and zones of high porosity and permeability must be common. Boul6gne & Pflumio (1992) report a bulk porosity of 15-20% in ocean crust, and a mean hydrothermal flux from the ridges of about 40 km 3 a -~ and, for the overall ocean floor flux, 148kin 3 a -~. Given that the ocean mass is 1.4 x 1021kg, this implies recycling of the entire ocean mass through the sea floor systems in about 10 million years; a massive exchange process. The discharged fluids (hot or warm) are enriched in many species (CO2, CH4, HzS, H2, SiO2, Li, Mn, Fe, Cu, Zn, etc.), and this process makes a major contribution to the geochemistry of ocean water, sediments, and to the nutrient fluxes for the marine biomass. In addition, many of the great sulphide ore deposits with Cu, Zn (Ag, Au) are related to such processes over geological time. A process of great significance to the global water inventory is the hydration of the seafloor crust that must accompany the cooling process. Altered seafloor basalts are highly hydrated, the limiting case being the transformation of peridotites to serpentinites (see MacDonald & Fyfe 1985). While the ridge processes are intense, slow convective cooling occurs over the entire ocean floor regions, and the early ideas of Hess (Fowler 1990) on massive serpentinization appear confirmed. The total mass of water contained in the ocean crust is similar to that in the ocean and, in addition, large quantities of CO2, ammonium and halogen compounds must be present, but this inventory is not well known (Fyfe & Lonsdale 1981). The hydrothermal processes associated with hot spot phenomena, and with the great continental flood basalts are less well studied. Of the present Earth's surface, almost 7% has been influenced by recent hot spot events (Fyfe 1992a). When the mass of some flood basalt events is considered, it even seems possible that

WATER INVENTORY OF THE EARTH cooling processes could perturb atmospheric oxygen (Fyfe 1990).

Subduction and fluids: subduction-induced mantle convection There is very little ocean floor ophiolitic crust older than 200 million years. Ocean crust is subducted at a rate almost equal to its rate of formation, but the materials subducted are very different from those which form the ridges. The original basaltic crust and gabbro-peridotite basement has been changed to a spilitic crust (with H20, CO2, S, U, etc.) and the gabbroperidotite basement has been partially converted to amphibolite-serpentinite before subduction (Fyfe 1992b). In the last few years, we have also come to recognize that pelagic sediments may be subducted on a scale of cubic kilometres per year. This concept, originally proposed by Gilluly (1971), once strongly opposed by those involved with isotope systematics, has now become respectable because of the direct observation of trenches and the structure of lithosphere near trenches. Thus, Hilde & Uyeda (1983) and Uyeda (1983) and the more recent Kaiko Project (Lallemand et al. 1986; Le Pichon 1986) have clearly shown that, when the lithosphere bends, it cracks in the upper part and forms horst and graben structures which fill with sediments. If there is not enough sediment to fill the structure, tectonic erosion of the overplate occurs: Japan is being tectonically eroded and underthrusted. Such studies clearly show that initially light materials which are tectonically trapped may move towards the mantle (Lallemand & Malavieille 1992). It is of note that there is little evidence for sediment subduction in the case of the Northern Cascadia subduction zone (Davis & Hyndman 1989). Consideration of volatiles shows that, for species like H20 and CO2, the recycling of the major reservoirs occurs with time constants of the order of a billion years, at the present rate of subduction. The Earth has a hydrosphere, so that return flow must be moderately efficient. But the present processes, and their scales, should warn us that any purely steady-state model of ocean volumes and other volatile reservoirs may be inadequate. Thermodynamics tells us that heating bodies degass, while cooling bodies adsorb gases. Thus if a cooling Earth continues to convect, the ocean volume might be expected to diminish. Is this what has happened on Mars (Carr 1987)? Processes involved in the return flow of volatiles are complex. At the initial stages of

3

thrusting, pore fluids are literally squeezed out and pass up the thrust structures, reducing friction on the thrusts (Anderson 1981). Exotic fauna, originally characteristic of ridge vents, have now been found in deep trenches, but the fluids are cold, and would not transport large quantities of silica or metals. Recently, the Canadian Lithoprobe Project (Yorath et al. 1985) has been studying the structure of the subduction of the Juan de Fuca plate beneath Vancouver Island. There are complex fault structures above the subducting slab. Seismic studies have revealed the complexity of the structures beneath the continental edge. The very young and active faults provide evidence for fluid flow transporting hydrocarbons, Mn-Fe-Ag, and related species. Recently, ODP Leg 110 Scientific Party (1987) reported gases including methane being produced, presumably by the reprocessing of subducted organic debris. Lewis et al. (1988) have described the thermal influences of slab dewatering. Once the preliminary compression stage has passed, metamorphic processes will dominate, eventually leading to the formation of eclogites from basalts; kyanite and garnet-bearing rocks from pelagic sediments, and even kyanitecoesite-pyrope rocks (Chopin 1984). At this stage, fluids may hydrofracture their way to the surface along faults. Some fluid may be carried to very great depths in minerals like phlogopite, a natural product of the metamorphism of K-bearing spilite or pelagic sediments in an ultramafic mantle environment. When deep degassing occurs, with hotter mantle above, a new chain of events must occur. Water injected into this hotter mantle (water which will essentially be a soup of SiOz-alkalis and trace metals) will soften the mantle, and lead to convection and plume formation. As plumes of contaminated mantle rise, they will melt to produce the contaminated basalts we call andesites. Thus, mantle convective motions are induced by the rising fluids. The small amount of fluids are amplified into a much larger heat flow process, which eventually forms the mountain chains of the volcanic areas of Andean type. In a general way, every gram of fluid introduced will probably lead to something like 100-1000 times the mass of volcanic rock. The injected fluids have led to a process which drains energy and mass out of the overlying mantle wedge. Studies of heat flow and electrical conductivity across subduction zones show the scale of this energy transfer process, catalysed by subduction. Almost one third of the continental crust of

4

W.S. FYFE

the Americas has been influenced by recent subduction events. But we should not forget that it is the volatile loading in the sea floor environment that has led to this process. It would not occur on a dry planet. Once mantle plume processes start above a subduction zone, an array of fluid-mass-energy transfer processes occur (see Rice 1985; Fyfe 1987): (i) basaltic andesite rises, extrudes, intrudes and underplates continental crust; (ii) the crust melts, producing granitic plutons and acid volcanics; (iii) the basal crust undergoes progressive metamorphism; (iv) magmas mix, and complex hybrids are produced near the Moho region; (v) ultra-high-temperature gases are injected into the base of the crust from the andesite magmas and assimilated dense crustal components, which founder in the underplate magmas; (vi) high-level plutons and volcanics are watercooled by deep groundwaters in the high heat flow near-surface environments; (vii) high topography is created, and deep ground water circulation through fractured intrusives and porous volcanics must follow. The total fluid fluxes which must result from the entire array are impressive. We are not at a stage to quantify such fluxes, but we can make some order of magnitude calculations. For example, if we consider the western Americas, where the plutonic-volcanic terrains extend for about 20000km with a width of 500km, and assume that a 5 km thickness of crust has been 'granitized' or melted, the total acid igneous mass is about 5 x 107km 3. Given a 108 year cycle, pluton production is 0.5 km3a -~. Andesite production is about 2km3a -~ (Thorpe 1982). Thus, about 2.5 km 3 of magma can be watercooled per year. The fluid fluxes will thus be about 20% of that of ocean ridges. In this case, there is no question that the fluxes of fluids, just as the volcanic eruptions, will not be steady state, as plutons rise and Tamboras erupt. There will be large spatial variation of fluid flux, fluctuations which may influence the global temperature for years, or even cause mass extinctions (Officer et al. 1987; Grove 1988). It is the knowledge of these events, their intensity and frequency distribution which are needed for the IGBP. Recently, we have become interested in the problem of dewatering of slabs, and the possibility of fluidized bed injection associated with such processes. Le Pichon et al. (oral comm.

1990) showed recent observations of the large (30km 3) warm mud volcanoes being extruded from the Barbados accretionary complex. Barriga et al. (1992) have described a number of tectonic regions where such processes may occur, and A. Ribeiro has named the process 'eduction', where fluidized beds are injected from mantle depths. One of the common consequences of thrusting is the rapid burial of fluid-rich rocks, including water-saturated sediments and metamorphic rocks with variable water contents. Physico-chemical processes that range from compaction to prograde metamorphism will tend to produce less hydrated rocks plus water. Hydraulic fracturing is a common mechanism when impermeable rocks cap fluidrich zones, and fast and concentrated fluid flow may generate vein mineral deposits (Fyfe & Kerrich 1985). Tectonically-induced fluid generation, at all lithospheric levels, can lead to rock fluidization and injection. These processes are well known by sedimentologists and neotectonics specialists, as they are responsible for intrusive sediments and seismically-induced mud and sand volcanoes. This process is suggested to explain some of the blueschisteclogite-serpentine associations of California and even, perhaps, the coesite rocks of the Italian Alps (see Barriga et al. 1992; Maekawa et al. 1993).

Continental collisions and associated strikeslip faults At the present time, most subduction processes involve the underthrusting of oceanic crust in situations near continental margins. But, periodically, a continent is moved into the zone of subduction. Molnar & Gray (1979) considered the problem of the possible subduction of such a continental edge, and concluded that it was indeed possible in terms of density relations. Over the past 50 Ma major collision events have occurred, and are still proceeding at a rate of 5 cm a ~ in the Himalayas. While slightly different models of detail occur, there is no question that India is currently being thrust under Asia (Allegre et al. 1984; Barazangi & Ni 1982). In this region, a section of crust, 1500 x 3000 kin, has been doubled in thickness. Present seismic results provide evidence for a jagged Moho at 60-80km depth with 10km steps. The region appears to be highly electrically conductive (Pham et al. 1986), with active zones of conductive fluids or melts. The area of thickened crust is similar in area to about 60% of continental Australia, and a section of crust of

WATER INVENTORY OF THE EARTH Australian size has been reworked in the process. In Fyfe (1986) some aspects of the problem of fluid fluxes when crust on this scale is heated and compressed due to the thrusting and shortening processes is considered. Essentially, the underthrust rocks will be dehydrated, lose CO2 and other volatiles, and eventually melt. The young plutons of Himalayas show such geochemical features (e.g. extreme initial 87Sr/86Sr etc.) as would be expected. The quantities of metamorphic water which may be expelled up fracture zones and thrust planes may be similar to the mass of the present day ice caps. If thick carbonate sequences are involved, CO2 release could perturb the atmosphere and its greenhouse effect. A particularly interesting situation can occur if large salt basins are involved in which even ocean salinity could be influenced (we tend to forget the frequency and scale of large continental salt deposits). In a collision event of this scale, a vast perturbation in the volatile element flux, from H20 to salt, hydrocarbons and even elements such as Hg and As, etc., is likely. It is also likely that the present rise of the 87Sr/S6Sr ratio of the oceans is related to enhanced weathering of evolved continental crust at high elevations (Burke et al. 1982). It is clear that a collisional event of this magnitude will change the global environment, by altering global wind and ocean current patterns. But there may be even more subtle and fluctuating influences on global ocean chemistry. Such processes are not quantitatively understood, but the record of change must exist in ocean sediments. On a smaller scale, similar phenomena may occur on regional strike-slip faults, when these form plate boundaries. Thus, the present models of structure on the great Alpine Fault of New Zealand shows major regions of over-ride with thick crust (Allis 1981). The strike-slip process requires fluids for lubrication, and the override processes will produce the necessary fluids. As Oliver (1986) has suggested, when major thrusts occur, there must be a huge 'squeegie' effect ahead of all overthickening tectonics. As the load moves forward (e.g. Tibet over India), a vast front of fluid expulsion must advance before the thrusts. Fluids would be typical pore fluids, zeolite facies fluids and salt and hydrocarbons, if appropriate rocks are present. Thus, there could be fluid pulses of highly different geochemistry as a thrust develops. For a thrust front of Himalayan scale (3000 km), moving at 10cma x, the fluid expulsion rate could attain 0.5 km 3 a ~. If this fluid is rich in hydrocarbons or salt, local influences could be dramatic. It is

5

also unlikely that the thrust motions would be steady state. It is interesting to note that the crustal granites formed during these thickening processes may also lead to the formation of large deep aquifers. Thus, the rising plutons and dykes described by Le Fort et al. (1987) and Thakur (1987) may act as excellent aquifers once they cool, crystallize, and contract and thus direct fluid flow. We have described these phenomena in Saudi Arabia, where acid dykes coming off plutons are almost totally converted to epidote by flooding with descending meteoric fluids (Marzouki et al. 1979). Plutons will cause local fracturing as is well shown by Drummond et al. (1988) in their description of granodiorites being converted to tonalites by salt-water flooding.

The impact of high elevations on fluid flows Subduction and collision processes produce extensive belts of high elevation on our planet. These are normally associated with vast fault and thrust structures, and discontinuities associated with magma activity, plutonism and dyke emplacement. Deep gravity driven flow in such systems is now becoming well documented (e.g. Nesbitt & Muehlenbachs 1989; Rye & Bradbury 1988). The scale of this giant hydrogeological process, which may float frontal thrusts, can obviously be significant (Fyfe 1986). High elevations also drive rapid erosion and ocean sedimentation. As Milliman & Meade (1983) estimate, present continental erosion moves about 1.7 x 1016g a-x. Given the mass of continental crust of 1.6 x 1025g, erosion could reprocess all crust in a billion years. A metre of rain per year could dissolve the Himalayas to sea level in 100 million years! Very thick (5-10km) accumulation of sediment are common near many of the great delta systems of the world (Burk & Drake 1974). For example, such sediment piles are common around the entire margin of Brazil, the Amazon delta regions, and the great Bengal Fan system. For example, what happens when ocean floor crust, with a thick deep serpentine layer, is loaded with over 10 km of sediment? At thermal equilibration, the base of the now 20 km section could reach 600°C. The peridotites will deserpentinize, and hot, highly reduced fluids will rise, transporting a range of metals. Faulting in such regions is ubiquitous. The chemistry of systems where reduced hot fluids interact with organic-rich cover sediments along fault systems must be highly anomalous. Hot seeps along such sediment sections are common (Grassle 1985).

6

W.S. FYFE

There is a great need for submersible observations in these systems. There is an obvious final question: does such loading and metamorphism of the oceanic basement provide the guide for future sites of subduction?

Concluding statement As understanding environmental change, and providing human resources, become great challenges for modern Earth science, the necessity to quantify the total global water inventory, and the inventory of other volatiles, is essential. W e must better understand the surface biD-nutrient cycles, and the impact of water use on climate changes. For example, we need models of climate impact when all rivers are d a m m e d and evaporation and evapotranspiration replaces runoff. Water, water quality, and life are inextricably connected. A n intriguing question is whether the mass of the oceans has been constant over geological time? If water is being subducted in large quantities, and if the planet is cooling, will the mantle slowly absorb this volatile? Or, as Frank (1990) has suggested (a not popular view), do we receive constant additions of water via small comets. It is time we visited Mars, and studied a planet where Gaia has failed. I would like to dedicate this brief review to Geoffrey Brown of the Open University, who lost his life while in Columbia on a volcano watch. Geoff, a wonderful student and colleague, worked with me in Manchester, to do some of the classic work on the melting of dry metamorphic rocks.

References ALLEGRE, J.C., COURTILLOT, V. 8£ 33 others. 1984. Structure and evolution of the Himalayan-Tibet orogenic belt. Nature, 307, 17-22. ALLIS, R.G. 1981. Continental Underthrusting beneath the Southern Alps of New Zealand. Geology, 9,303-307. ANDERSON, D.L. 1993. Helium -3 from the mantle: primordial signal or cosmic dust? Science, 261, 170-176. ANDERSON, R.N. 1981. Surprises from Glomar Challenger. Nature, 293,261-262. BARAZANGI,M. & NI, J. 1982. Velocities and propagation of Pn and Sn beneath the Himalayan arc and Tibetan plateau: Possible evidence of underthrusting of India continental lithosphere beneath Tibet. Geology, 10, 179-185. BARRIGA, F.J., FYFE, W.S., LANDEFELD, L.A., MUNHA, J. & RraEmo, A. 1992. Mantle education: tectonic fluidization at depth. Earth Science Reviews, 32, 123-130. BERNER, E.K. & BERNER, R.A. 1987. The Global Water Cycle. Prentice Hall inc. New Jersey.

BOULOGNE,J. & PFLUMIO,C. 1992. Global chemical fluxes and fluid flow through the seafloor. In K.J. Hsu and J. THIEDE, (eds) Use and misuse of the seafloor. John Wiley and Sons, New York, 305-318. BROECKER, W.S. 1985. How to Build a Habitable Planet, Eldigio Press, New York. BURr, C.A. & DRAKE, C.L. 1974. The geology of continental margins. Springer-Verlag. BURKE,W.H., DENISON,R.E., HETHERINGTON,F.A., KOEPNICK,R.B., NELSON,H.F. & OTa, J.B. 1982. Variation of sea water 87Sr/86Srthroughout Phaerozoic time. Geology, 10, 516-519. CARR,M.H. 1987. Water on Mars. Nature, 326, 30-35. CHOPIN, C. 1984. Coesite and pure pyrope in highgrade blueschists of the Western Alps: a first record and some consequences. Contributions to Mineralogy and Petrology, 86, 107-118. DAVIS, E.E. & HYNDMAN,R.D. 1989. Accretion and recent deformation of sediments along the Northern Cascadia subduction zone. Geological Society of America Bulletin, 101, 1465-1480. DAVIS, E.E. & LISTER, C.R.B. 1977. Heat flow measurements over the Juan de Fuca Ridge: evidence for widespread hydrothermal circulation in a highly heat transportive crust. Journal of Geophysical Research, 82, 4845--4860. DRUMMOND,M.S., RAGLAND,P.C. & WESLOWSKI,D. 1986. An example of trondhjemite genesis by means of alkali metasomatism: Rockford granite, Alabama Appalachians. Contributions to Mineralogy and Petrology, 93, 98-113. FOWLER,C.M.R. 1990. The Solid Earth. Cambridge University Press, Cambridge. FRANK,L.A. 1990. The big splash. Avon Books, New York. FYFE, W.S. 1970. Lattice energies, phase transformations and volatiles in the Earth's mantle. Physics Earth and Planetary Interiors, 3, 196-200. 1974. Archaean tectonics. Nature, 249, 338-399. 1986. Fluids in deep continental crust. In BARAZANGI, M. & BROWN,L. (eds) Reflection Seismology: The Continental Crust. American Geophysics Union, Geodynamics series, 14, 33-39. 1987. The fluid inventory of the crust and its influence on crustal dynamics. In FRITz, P. & FRAPE, S.K. (eds) Saline waters and gases in crystalline rocks. Geological Association of Canada, Special Paper, 33, 1-3. 1990. Geosphere forcing: plate tectonics and the biosphere. Palaeogeography, PalaeoclimatoIogy, Palaeoecology, 89, 185-181. 1992a. Magma underplating of continental crust. Journal of Volcanology and Geothermal Research, 50, 33-40. 1992b. Geosphere Interactions On A Convecting Planet: Mixing and Separation. In HUTZIN~ER,O. (ed.) The Handbook of Environmental Chemistry, 1, part F. Springer Verlag Berlin Heidelberg, 1-26. & KERRICR, R. 1985. Fluids and Thrusting. Chemical Geology, 49,353-362. & LONSDALE,P. 1981. Ocean floor hydrothermal

WATER INVENTORY OF THE EARTH activity, in EMILIAN, C. (ed.), The Oceanic Lithosphere, John Wiley and Sons, 589-638. , PRICE, N.J. & THOMPSON, A.B. 1978. Fluids in the Earth's Crust. Elsevier, Amsterdam. GILLET, P. 1993. L'eau du manteau terrestre. La Recherche, 255,676-685. GILLULY, J. 1971. Plate tectonics and magmatic evolution. Geological Society of America Bulletin, 82, 2387-2396. GRASSLE, J.F. 1985. Hydrothermal vent animals: Distribution and biology. Science, 229, 85-400. GROVE, J.M. 1988. The Little Ice Age. Methuen, London. HILDE, T.W.C. • UYEDA,S. 1983. Convergence and subduction. Tectonophysics, 99, 85-400. KRONER,A. & LAYER,P.W. 1992. Crust formation and plate motion in the early Archean. Nature, 256, 1405-1411. LALLEMAND, S. t~ MALAVIEILLE,J. 1992. L'erosion profondes des continents. La Recherche, 23, 1388--1397. LALLEMANT, S., LALLEMAND,S., JOVIET, L. & HUCHON,P., 1986. Kaiko: l'exploration des losses du Japan. La Recherche, 17, 1344-1357. LE FORT, P., CUNEY, M., DENIEL, C., FRANCELANORD, C., SHEPPARD,S.M.F., UPRETI, B.N. & VIDAL, P., 1987. Crustal generation of the Himalayan leucogranites. Tectonophysics, 134, 39-57. LE PICHON,X. 1986. Kaiko, voyage aus extremites de la met. Editions Odile Jacob, Paris. LEWIS, B.T.R., BENTOWSKI, W.H., DAVIS, E.E., HYNDMAN,R.D., SOUTHER,J.G. & WRmHT, J.A. 1988. Subduction of the Juan de Fuca plate: thermal consequences. Journal of Geophysical Research, 93, 15,207-225. LISTER, C.R.B. 1977. Qualitative models of spreading center processes, including hydrothermal penetration. Tectonophysics, 37,203-218. MACDONALD, A.H. & FYFE, W.S. 1985. Rates of serpentization in seafloor environments. Tectonophysics, 116,123-132. MAEKAWA, H., SHOZUI, M., ISHII, T., FRYER, P. & PEARCH, J.A. 1993. Blueschist metamorphism in an active subduction zone. Nature, 364,520-523. MARZOUKI, F., FYFE, W.S. & KERRICH, R. 1979. Epidotization of diorites at A1 Hadah, Saudi Arabia: fluid influx into a cooling pluton. Contri-

butions to Mineralogy and Petrology, 68,281-284. MILLIMAN, J.D. & MEADE, R.H. 1983. World-wide delivery of river sediments to the oceans. Journal of Geology, 91, 1-21. MOLNAR, P. & GRAY, D. 1988. Subduction of continental lithosphere: some constraints and uncertainties. Geology, 7, 58-63. NESBITr, B.E. & MUEHLENBACHS, K. 1989. Origins and movement of fluids during deformation and

7

metamorphism in the Canadian Cordillera. Science, 245,733-736. ODP LEG I IO SOEN~FIC PARTY, 1987. Expulsion of fluids from depth along a subduction-zone decollement horizon. Nature, 326,785-788. OFFICER, C.B., HALLAM,A., DRAKE,C.L. & DEVINE, J.D. 1987. Late cretaceous and parysmal Cretaceous/Tertiary extinctions. Nature, 326, 143149. OLIVER, J. 1986. Fluids expelled tectonically from orogenic belts: their role in hydrocarbon migration and other geologic phenomena. Geology, 14, 99-104. PAWLEY, A.R., MCMILLAN, P.F. & HOLLOWAY,J.R. 1993. Hydrogen in stishovite, with implications for mantle water content. Science, 261, 10241026. PEDERSEN,K. 1993. The deep subterranean biosphere. Earth Science Reviews, 34,243-260. PHAM, V.N., BOYER,D., THEOME,P., YAN, X.C., LI, L. & JIN, G.V. 1986. Partial melting zones in southern Tibet from magnetotelluric results. Nature, 319,310-312. POSTAL, S. 1992. Last oasis-facing water scarcity. W.W. Norton and Co. New York. RICE, A. 1985. The mechanism of the Mt. St. Helens eruption and speculations regarding Soret effects in planetary dynamics. Geophysical Surveys, 7, 303-384. RYE, D.M & BRADBURY,H.J. 1988. Fluid flow in the crust: an example from a pyrenean thrust ramp. American Journal of Science, 288,197-235. ScnovF, J.W. 1992. Microfossils of the Early Archean Apex chart: new evidence of the antiquity of life. Science, 260,640-646. SCHRAUDER, M. & NAVON, O. 1993. Solid carbon dioxide in a natural diamond. Nature, 365, 42--44. STRAUS, J.M. & SCHUBERT,G. 1977. Thermal convection of water in a porous medium: effect of temperature- and pressure-dependent thermodynamic and transport properties. Journal of Geophysical Research, 82, 325-333. THAKUR, V.C. 1987. Plate tectonic interpretation of the Western Himalayas. Tectonophysics, 134, 91-102. THOMPSON, A.B. 1992. Water in the Earth's upper mantle. Nature, 358,295-300. THORPE, R.S. 1982. Andesites. John Wiley and Sons. UYEDA, S. 1983. Comparative subductology. Episodes, 1983, 19-24. YORATH, C.J., GREEN, A.G., CLOSES,R.M., SUTHERLAND GROWN, A., BRANDON, M.T., KANASEWlCH, E.R., HYNDMAN, R.D. & SPENCER, C. 1985. Lithoprobe southern Vancouver Island: seismic reflection sees through Wrangellia to the Juan de Fuca plate. Geology, 13,759-762.

Tectonic control of the sedimentary record and stress-induced fluid flow: constraints from basin modelling R. V A N B A L E N & S. C L O E T I N G H Tectonics~Structural Geology Group, Vrije Universiteit, De Boelelaan 1085, 1081 H V A m s t e r d a m , The Netherlands

Abstract: Many basins show deviations from the simple subsidence pattern predicted by the

stretching model for basin evolution. Incorporating finite strength of the lithosphere during rifting and stresses acting on the lithosphere during the post-rift phase of basin evolution can successfully explain these deviations. Finite strength of the lithosphere during rifting causes flexurally supported rift-shoulders. The erosion of these rift-shoulders induces both additional uplift at the basin margin and sediment loading-induced subsidence in the basin centre, causing tilting of syn-rift sediments. A 'break-up' unconformity results from the progressive diminishing of this tilting towards the end of the rifting period, separating the syn- and post-rift sediments. As the amount of erosion is primarily controlled by climatic conditions, basins having the same tectonic history but, for example, located at different lattitudes will not have developed the same pattern of stratigraphic fill. The same effect is expected if another mechanism causes the uplift of the basin margin, as, for example, a change in the level of intraplate stress. Plate reorganizations are the main cause for changes in the level of intraplate stress. We demonstrate that the in-plane stress variations affect the fluid flow regime in rifted basins, with possible implications for the diagenesis of sediments, primary migration of hydrocarbons, faulting and localization of economic resources. Examples include the North Sea and Pannonian Basin. An increase in the level of compressive stress causes flank uplift and basin centre subsidence, inducing a contemporaneous increase of meteroic water influx and compaction-driven fluid overpressures. An increase in the level of tensile inplane stress induces the opposite effects.

Over the last few years considerable progress has been made in quantifying the effects of forces originating from plate boundaries on differential uplift and subsidence and the stratigraphic record in rifted sedimentary basins (Cloetingh et al. 1985, 1989; CIoetingh & Kooi 1992). In fact, in many cases it appears to be very difficult to discriminate between eustatic and stress induced on- and offlap patterns in the stratigraphy (Kooi & Cloetingh 1992; Reemst et al. 1994). The influence of intraplate stresses is not restricted to rifted basins; stresses also affect the stratigraphy of foreland basins in a way comparable to the effect of eustatic sealevel changes (e.g. Peper et al. 1992, 1994; Posamantier & Allen 1993). The effect of inplane stress variations on the compaction-driven fluid flow system in rifted basins can be considerable, as shown by numerical modelling by Van Balen & Cloetingh (1993). Fluid overpressures and flow velocities are altered by as much as 30%. This can account for important phenomena as hydrofracturing and episodic dewatering with associated thermal, diagenetic and migration processes. Observations of present-day stress fields by

earthquake focal mechanism studies, in situ stress measurements and break-out analyses (Zoback 1992; Miiller et al. 1992) shows that a large number of basins formed in an extensional regime, such as the North Sea (Kooi & Cloetingh 1989) and the Pannonian Basin (Horvfith 1993), are in a situation of present-day compression. The most popular model for extensional basin formation, the McKenzie stretching model, has been extensively applied to the North Sea and Pannonian Basin (Sclater & Christie 1980; Sclater et al. 1980; Royden & D6venyi 1988). The main reason for this has been the observed thinning of crustal and sub-crustal lithosphere under these basins, compatible with the predictions of the stretching model. The stretching model also predicts a decay of subsidence during the post-rift phase of extensional basin evolution. This is, however, not consistent with recent observations of anomalous subsidence during the Plio-Quaternary times in the North Sea Central Graben and Pannonian Basin, showing the need for a new generation of basin formation models incorporating constraints on lithosphere stress and rheology (Cloetingh et al. 1993).

From PARNELL,J. (ed.), 1994, Geofluids: Origin, Migrationand Evolution of Fluids in Sedimentary Basins, Geological Society Special Publication No. 78, 9-26.

10

R. VAN BALEN & S. CLOETINGH -20"

30'

S.-.l'q~',m.~Ic fP,oc'b.~ /

Su-br~kouts

m,m~,OIF O w O ~ Ovom* I n g n

E~lr~rd S~lm ~ ~mm~na s~m ~ ~Ike ~ ~ Rqmw unwo~ sblm ~

Da~a

0 t28 252 378 mmlm~:=lil=.m~

N

i.N~m" c:~r.N, iif~Emlm¢

~i~

Fig. 1. The stress map for Europe compiled by the World Stress Map team (modified after Miiller et al. 1992). The map shows the directions of the maximum horizontal stress. Clearly shown are the consistent NW-SE orientation in the North Sea Basin area, and the interference of the NW-SE and NE-SW (the Dinaride trend) directions in the Pannonian Basin area. The stress field for Europe (see Fig. 1) shows two main directions for the maximum horizontal stress: a N W - S E orientation in northwestern Europe and a N E - S W orientation in southeastern Europe. The first stems from the ridge push in the Atlantic ocean, the second from the

subduction processes in the Aegean area (MOiler et al. 1992). Recent investigations have revealed the basinwide existence of a giant normal fault network in the Palaeogene successions of the North Sea Basin (Cartwright 1993, pers. comm.), which is related to over-

BASIN MODELLING CONSTRAINTS pressuring. The timing of the normal faulting coincides with a period of acceleration of Pliocene-Recent subsidence in the North Sea Basin (Cloetingh et al. 1990). This acceleration of subsidence can be explained by an increasing compressive intra-plate stress (Kooi & Cloetingh 1989). The model of Van Balen & Cloetingh (1993) can, therefore, explain the overpressuring which caused the activation of the normal faults in the Palaeogene successions. The Pannonian Basin is a large intra-montane basin developed inside the Alpine chain. The basin originated from a combination of tectonic escape (Ratschbacher et al. 1991) and subduction related processes (Horvfith 1993). The main rifting period for this basin was in the MidMiocene (Badenian). Horvfith & Cloetingh (pers. comm.) have proposed a model in which an increase in the compressive stress field explains the observed anomalous Late Pliocene - Quaternary subsidence in the Pannonian Basin. In the present paper we discuss the effect of this change in magnitude of the compressive stress on the compaction driven fluid flow regime in the two deepest sub-basins in the Pannonian Basin. A full presentation of the results of a forward modelling study of the stratigraphy and fluid flow in the Pannonian Basin will be presented elsewhere (Van Balen et al. 1994).

Intraplate stresses and anomalous

11

2" c o m p r e s s i o n ~

tension

^W:

rift-shoulde~

I

_1

-- --0.5

I

rifted _ .

~ '

--0 -0.5 --3

E v

hinterland

E qJ

-

0

-

3

"o

lithospheric necking during rifting:

,4:,',i:i:i:!:!:i,'i:i'i:i

-0

necking depth"',',',,,','"',"'"""";'"","

E v e-

"0

- 30 :;sub-crust?.:'" isostas~-'-:-~-'.-T-:-~.-:-: -'.

surface expression after isostatic rebound: rifted

rift-sh0ulder'~- - 3 ,-,

subsidence of rifted basins

The stretching model, mathematically formulated by McKenzie (1978), can account for the long term post-rift subsidence (thermal sagging), but does not incorporate rift-basin topography (and, therefore, the distribution of syn-rift deposits) and observed short term deviations of the subsidence pattern in the post-rift phase. Anomalous Pliocene-Quaternary subsidence, observed in a large number of basins in the Northern Atlantic/Mediterranean region (Cloetingh et al. 1990; Cloetingh & Kooi 1992; Zijerveld et al. 1992), is consistent with the model proposed by Cloetingh et al. (1985, 1989). In this model in-plane forces, originating from processes operating at plate boundaries (slabpull, ridge push), modulate the bending of the lithosphere beneath extensional basins (see also Cloetingh & Kooi 1992). The surface expression of this process is differential uplift and subsidence operating on a long wavelength (up to several hundreds of kilometres). The vertical loads acting on the lithosphere determine which parts of the basin are going to be uplifted and which will subside. For models invoking a

3"o

(b) Fig. 2. (a) The effect of changes in the lithospheric stress regime on the subsidence and uplift pattern in a rifted basin. An increase in the level of compressive stress induces flank uplift and basin center subsidence, whereas an increase in the level of tensile stress induces the opposite differential movements. (b) Due to finite strength during rifting the lithosphere necks around its strongest part. Depending on the depth of necking, the isostatic rebound forces can be directed upwards, causing flexurally supported rift-shoulders. shallow depth of lithospheric necking, in the case of compressive forces, the largest downward-pointing lithospheric loads (located under the deepest part of the basin) coincide with areas of enhanced subsidence, while uplifted areas are located where the loads are the least (at the basin flanks). In the case of tensile forces, crustal movements are oppositely directed (see Fig. 2a)

12

R. VAN BALEN & S. CLOETINGH

Such crustal movements are associated with onlap and offlap patterns, bounding system tracts (e.g. Van Wagoner et al. 1988), and can therefore equally well explain relative sea-level changes inferred from these patterns as, for example, glacio-eustatic variations (Cloetingh 1991; Posamentier & Allen 1993) and sediment supply (Schlager 1993). Regional intraplate stress level changes can occur on time scales varying between a few tens to a few million years (Philip 1987). N u m e r i c a l model Basin formation model

Our numerical model for basin formation and evolution is based on the stretching principle, originally proposed by McKenzie (1978). In this model, the lithosphere is kinematically thinned during the rifting time, causing a closer spacing of the geotherms (the geothermal gradient is increased). Relaxation of the disturbed temperature field causes cooling of the rocks, which therefore contract, giving rise to subsidence. The model employs different extension factors for the crustal and mantle part of the lithosphere (Royden & Keen 1980). An important modification of the pure-shear stretching model was incorporated by Kooi et al. (1992) in forward stratigraphic modelling, based on inferences by Braun & Beaumont (1989) and Weissel & Karner (1989). In this model the lithosphere has finite strength during rifting (see Fig. 2b). The actual depth of necking is controlled by bulk rheological properties of the lithosphere. The average response to extension in a multi-layer temperature-dependent rheology with strong layers in the upper crust and upper mantle (Carter & Tsenn 1987; Cloetingh & Banda 1992), therefore, can be cast in terms of a level of necking at lower crustal levels, which itself acts as a detachment level in which major faults sole out (Kooi et al. 1992; Van der Beek et al. 1994). Dynamical analyses of rifted basin formation (Braun & Beaumont 1989; Bassi et al. 1993; Dunbar & Sawyer 1989) show that the lithosphere necks around the strongest part of the crust (i.e. a mid-crustal level), even when vertically orientated crustal or mantle weakness zones are taken into account. Kinematic modelling of rift-shoulder evolution constrained by fission-track data suggests necking depths between 10 and 15 km for the Saudi Arabian Red Sea margin and the Transantarctic Mountains (Van der Beek et al. 1994). Modelling of Neogene western Mediterranean extensional basins formed in collisional regimes including

the Gulf de Lions (Kooi et al. 1992) and the Valencia Trough (Janssen et al. 1993) also yield a best fit for necking depths at mid-crustal levels. In contrast, modelling of the Tyrrhenian Sea Basin points to a deep level of necking at a lower crustal level, around 25km depth (Spadini & Cloetingh pers. comm.). McKenzie's original model assumes that local isostasy during rifting creates a topographic 'hole'. Here we adopt a model in which this feature results from lithospheric necking (see Fig. 2b). Models not invoking finite-strength of the lithosphere require larger amounts of lithospheric thinning, leading to an overestimation for paleo-heatflow (Kooi et al. 1992). This has major implications for the determination of the oil-window as the amount of heat available for hydrocarbon maturation is overestimated (Kooi & Cloetingh 1989). Another effect of introducing necking and flexural isostasy is that, depending on the depth of necking, the vertical loads acting on the lithosphere after stretching can be directed upwards. If the depth of necking is 'sufficiently deep', the basin is overdeep and the lithosphere is in an upward state of flexure, instead of downward. Riflshoulders are supported flexurally by this mechanism (see Fig. 2b). Rift-shoulder erosion

During their uplift, the rift-shoulders will be eroded, causing additional isostatic uplift of the basin margin. At the same time, the erosion products are deposited in the basin, inducing isostatic subsidence (see Fig. 3). Because riftshoulder topography is created by an isostatic uplift caused by lithospheric thinning, this process is particularly important during the syn-rift phase of basin evolution. decreasing

decreasing

K

trans

erosion

.... ============================================================

Fig. 3. Syn-rift erosion of rift-shoulders causes an

additional isostatic uplift at the basin margin due to unloading. At the same time, the increased sediment loading enhances the subsidence in the nearby basin.

BASIN MODELLING CONSTRAINTS In order to quantify the effect of rift-shoulder erosion on basin evolution, sedimentation and erosion are treated as dynamical processes which are modelled numerically as a diffusion problem. The governing equation is (Kenyon & Turcotte 1985; Syvitski et al. 1988; Flemings & Jordan 1989; Peper etal. 1994): Oh 32h Ot - Kt .... Ox---7

(1)

This equation expresses the law of conservation of mass, whose justification of which depends on the transport medium (see Angevine et al. 1990; Kenyon & Turcotte 1985). The transportation coefficient, Kt..... depends on the transport medium and the sediment type. Values of Kt.... can be derived theoretically, although in some cases this is very difficult. In our modelling we have applied empirical values from the literature (e.g. Flemings & Jordan 1989). In general Kt .... increases basinward until the shelf, from where it decreases basinward. The diffusion equation is solved numerically using a one dimensional implicit finite-difference scheme. Two simulation runs are presented in Fig. 4a and b, which illustrate the difference in basin development predicted by models incorporating different amounts of riftshoulder erosion. The same parameters were used for the simulations, the only difference being the presence or absence of erosion of the riftshoulders. During the syn-rift phase, sediments are eroded from the riftshoulders, whereas for the post-rift phase of basin evolution an external source for the sediment is assumed. The stratigraphy predicted by a model ignoring the effects of erosion is shown in Fig. 4a. The stratigraphy is characterized by a continuous onlap of sediments onto the rift-shoulder, with a very minor horizontal distance effected by the shoreline migration. The inclusion of erosion (see Fig. 4b) has two main effects compared with no erosion. The syn-rift sediments, which are derived from the eroded rift-shoulder, experience major syn-rift tilting. This tilting is caused by both erosion-produced additional uplift at the basin margin and enhanced sediment loading in the basin (see Fig. 3). As the amount of tilting diminishes progressively towards the end of the rifting period, a 'break-up' unconformity separates the syn- and post-rift sediments. As shown by Fig. 4b, because of erosion of the riftshoulder, a platform area has developed at the basin margin, causing the post-rift sediments to onlap over a larger horizontal distance. The modelling shows the importance of including

13

riftshoulder erosion in forward basin analyses. As shown by a comparison of Fig. 4a and Fig. 4b, models using the same set of stretching factors can lead to quite different predictions for the basin fill patterns. The 'amount' of rift-shoulder erosion (the value of Kt.... at the riftshoulder area) is dependent amongst others on climatic conditions. If the climate is humid, erosion will be very efficient, whereas in arid climatic conditions erosion is absent. Therefore, the differences in stratigraphy shown in Fig. 4a and 4b could be indicators for respectively dry and wet climatic conditions during rifting. Eustatic sea-level changes (Haq et al. 1987), in-plane stress variations (Cloetingh et al. 1985) and sediment supply (Schlager 1993) are known to influence the development of stratigraphic sequences. However, our results show that climate-controlled erosion of rift-shoulders affects the syn-rift basin fill pattern through its coupling with a tectonic process, i.e. tilting due to isostatic uplift in the eroding area and subsidence in the nearby depocentre. R e s p o n s e o f the basins to vertical a n d h o r i z o n t a l loads

The response of the lithosphere to vertical and horizontal loads, adopting a thin elastic plate analogy for the mechanically strong part of the lithosphere is given by: 02 OX2

O2W O2...~W ( D(x)--~x 2 ) + F 0X 2 + (Pa-Pfin)gw (X) = q(x)

D-

(2)

Ec T~3 12 (1 - vz)

(for notation see Table 1). The effective elastic thickness (Te) of the lithosphere is taken to be equal to the depth of the 450°C isotherm (Kooi et al. 1992), and, therefore, varies in space and time during basin evolution. Temperature calculations are performed on a finite-difference mesh, which is deformed due to the thinning and extension process. The mesh also deforms due to the thermal contraction, which removes the need to make the Boussinesq approximation. In the case of the Pannonian Basin we found that models invoking this approximation underestimate thermally induced subsidence (up to 800m). The finite difference calculations are fully implicit in the vertical direction and explicit in the horizontal direction. The boundary conditions for the temperature calculations are zero heatflow

14

R. VAN BALEN & S. CLOETINGH I

T

i

r

T

--7

on,ap

°

! •

i

E'-" I1) "EI C~l

0

1O0

200 distance (km)

300

0

1O0

200 distance (kin)

300

A

E'-"

v" v

Q.

Fig. 4. (a) Dynamic sedimentation modelling result showing the stratigraphy in the absence of rift-shoulder erosion, representing dry climatic conditions. Each stratigraphic interval represents 2 Ma. The stratigraphy shows a minor onlap to the rift-shoulder. (b) Result of the dynamical sedimentation modelling incorporating rift-shoulder erosion, representing wet climatic conditions. The onlap of sediments to the basin margin is considerable. The syn-rift deposits show tilting due to the couple of additional uplift at the margin and subsidence in the basin. A 'break-up' unconformity results from the ceasure of the tilting. across the side boundaries of the domain, 0 ° C at the top and a user specified (usually 1350 ° C) temperature at the depth of the base of the pre-stretch lithosphere. Sedimentation rates are determined using water depth profiles; i.e. the bathymetry as a function of time is given as an input parameter. The modelling approach taken allows for differ-

ent sediments to be deposited, which obey different hydraulic characteristics and compaction trends.

Fluidflow modelling The module which calculates fluid overpressures and sediment porosities has been extensively

BASIN MODELLING CONSTRAINTS Table 1.

Notation

D Ec

flexural rigidity (N m) Youngs modulus of lithospheric material (Pa) horizontal force (N) total fluid pressure (Pa) lithostatic or overburden pressure (Pa) effective pressure (Pa) effective elastic thickness (m) displacement due to compaction, relative to a fixed point (m) parameters for empirical laws relating porosity to effective pressure and permeability to porosity acceleration of gravity (m/s 2) permeability (m 2) horizontal and vertical permeability (m 2) vertical load due to sediments and seawater (Pa) time (s) Poisson's ratio vertical deflection (m) fluid overpressure, pressure above hydrostatic (Pa) fluid compressibility (Pa-~) porosity, porosity at sedimentation density of asthenosphere (kg/m 3) density of crust (kg/m 3) density of basin fill (kg/m 3) density of grains (kg/m 3) fluid density (kg/m 3) fluid viscosity (Pa s) divergence, gradient

F Pn

PHxH Peff Te 1I...1 a. b, c g k khor, kvert

q t v w qb 13 qb, ~b0 Pa Pc pfill Ps Pn tx VT, V

described in Van Balen & Cloetingh (1993). In this approach porosity is represented by a sediment dependent exponential function of the effective pressure (Shi & Wang 1986):

+z = +oe-be°"z z

(3)

z

eo.= f, (1-+Op~dz+f (+z)o.gdz- P~ The permeability is a function of porosity. Anisotropy is accounted for, as shown in equations 4 (see Bethke 1985). khor = 10 -"+b+,

kver = c. khor

(4)

In contrast to Van Balen & Cloetingh (1993), each finite-element can represent a mixture of sediment types. The overall porosity for the finite-element is given by the volumetric mean porosity. The overall permeabilities in the direction parallel and perpendicular to the strata are determined by considering the analogous problem of electrical resistors wired parallel or in

15

series (permeabilities are 'inverse' resistors, see equations 5). Khortot = aKhorsl + (1 --a) Khors2 1/gvertot =

(5)

a/gversl + (1-a)/Kvcrs2

a = volume percentage of sediment sl. Assuming incompressible grains, fluid overpressure is determined (see Bethke 1985; Van Balen & Cloetingh 1993) by:

+13 0(,~ + 0°Vzm)= VT{_~[V(O)]) at 1 0+ ( 1 - + ) Ot

(6)

The differential equations are solved using Galerkin finite elements. The resulting equations are mutually dependent. This non-linearity is solved using a Picard iteration. The three modules (basin subsidence, sedimentation and fluid flow) are each executed every time step. Because the fluid flow calculations require the smallest step size, this module contains an automatic time step-size adjuster. A typical time step in the simulation is 1000 years.

lntraplate stress variations and compactiondriven fluid flow Van Balen & Cloetingh (1993) have proposed a model for the effect of changing inplane stresses on the fluid flow regime in rifted basins, assuming a very shallow level of necking (see Fig. 5). In this model, an increase in the level of compressive intraplate stress induces basin flank uplift and basin center subsidence. This leads to enhanced sediment exposure at the basin margin and an increase in the gravity potential. Therefore, under humid conditions an increase of meteoric water influx will occur. At the same time, sedimentation rates in a basinward position have amplified (due to an increase of sediment accumulation space and hinterland weathering), inducing an increase in compaction-driven fluid velocities and overpressures. The opposite processes are expected during an increase in the level of tensile intraplate stress, or a drop in the magnitude of compressive stress. Numerical simulations support this hypothesis. The stress-induced perturbations in fluid velocities and overpressures were found to be up to the order of 30%. Because intraplate stress events have a finite duration, stresses are ultimately relaxed, leading to differential motions opposite to those during the increase in the level of stress. Therefore, the intraplate stress-induced perturbation of the

16

R. VAN BALEN & S. CLOETINGH

~

~ w S ° ° ~ ~ + , ,, (m) ~ -soo ....

t

--.,.

I . . . . . . . . . . . . .

L I I I

COMPRESSION

t I

<----

~ meteoric water

~,~,~,~,~,~,~,~&~,~,,

•

I I

I--

"-'-~-~-~-~-~-~'~-. "'";~;~;~;~;~;~.,..

4-0

~2 "0

• ".'.,,,,,.-.-,',',',,,',',',',-,-,-,-,-,

+ 500

,-,-,

I I I

~,"

TENSION

tsssssis~

~3

> 4-0

]

~:~'~'~{~'~',',

I

BASEMENT

4-z

,I

================================================= 4-3 "O

Fig. 5. The effect of changing magnitudes of in-plane stresses on the fluid flow regime in sedimentary basins. (a) An increase in the level of compressive intraplate stress enhances sedimentation rates in the basin centre and, therefore, increases the compaction-driven fluid overpressures. At the same time, the meteoric water infiltration increases at the uplifted basin margin. The upper panel depicts the stress induced deflection. (b) An increase in the magnitude of tensile inplane stress results in the opposite effects on the fluid flow regime. The compaction-inducedfluid overpressures are reduced and the meteoric water incursion is hampered due to the subsidence at the basin margin.

fluid flow regime takes the shape of a pulse. The implications of this mechanism include the faulting and fracturing of strata, the alteration of heat flow, diagenesis and localization of economic resources such as hydrocarbons and metals. Kooi & Cloetingh (1989) have shown that the

late stage-acceleration of subsidence in the Central North Sea can be explained by the observed N W - S E - o r i e n t a t e d maximum horizontal stress (Miiller et al. 1992). Seismic investigation of the Palaeogene sediments in the North Sea Basin show evidence for overpressuring. A three-dimensional system of hexagonal

BASIN MODELLING CONSTRAINTS organized normal faults has developed during or shortly after deposition of these sediments (Cartwright 1993, pers. comm.), during the build-up of the compressive intraplate stress field. This is consistent with the model proposed by Van Balen & Cloetingh (1993). The effect of a drop in eustatic sea level at Mid-Oligocene of the Gulf of Mexico Basin was investigated by Bethke et al. (1988). Their modelling predicts an overall drop in fluid overpressure, while at the same time the amount of meteoric water incursion increases. This can be explained by a basinward shift of the highest sedimentation rate, causing the rate of compaction above the zone of high overpressuring to decrease, and thereby to decrease the overpressuring (the zone of overpressure tends to migrate basinward). The increase in sediment exposure at the basin margin and the increase of the gravity potential of the groundwater table (due to the drop of sea level) enhances rainwater infiltration into the basin. Our model predicts a different contemporaneous link between overpressuring and meteoric water infiltration caused by a relative change in sea level (see above). As the Mid-Oligocene is characterized by plate-reorganizations, these could have caused the same effects as the Late Neogene reorganizations (Cloetingh et al. 1990), i.e. rapid uplift and subsidence. Therefore, the MidOligocene eustatic sea level drop could be substantially enhanced by a tectonic component. A compressive stress-induced basin margin uplift has caused the documented increase in meteoric water infiltration in the Gulf of Mexico Basin (Harrison & Tempel 1993). This was probably coeval with an increase in fluid overpressures. Diagenesis of sediments is the key indicator of the paleo-fiuid regimes in sedimentary basins (Horbury & Robinson 1993). It documents important changes in the fluid regime induced by, amongst others, faulting, regional heating and meteoric water incursion. Widespread diagenetic changes are concentrated during periods of greatest water flow and thus are effectively episodic, related to changes in the hydraulic evolution of the basin (Harrison & Tempel 1993). There are several indications for cementation events in the North Sea Basin. Gluyas et al. (1993) have, for example, documented basinwide quartz and illite cementation phases taking place within a restricted time. The quartz cementation seems to coincide with basinwide rapid subsidence and heating, contemporaneous with a similar event at the Norwegian Margin (Jensen & Dor6 1993). As shown by Reemst et al. (1994), the increase of subsidence in the

17

offshore part of the Norwegian Margin can be explained by a Pliocene-Quaternary increase in the level of compressive in-plane stress, which has major implications for the hydrocarbon generation and migration. The increased heatflow might be explained by advection of heat caused by an increase in the ~amount of compaction-driven fluid flow. Inplane stresses accumulating in the flexed lithosphere are ultimately relaxed by faulting. These faults can be both permeable and impermeable, depending on the deformation mechanisms, the fluid pressure history, the spatial and temporal pattern of dilation and the diagenesis induced by the fluids using the fault zone (Knipe 1993). The relaxation itself causes the opposite differential movements as during the build-up of stress, causing the stress-induced changes in the fluid flow regime to be pulsewise (Van Balen & Cloetingh 1993).

Application to the Pannonian Basin O r i g i n o f the b a s i n

The Pannonian Basin is a large intra-montane basin located inside the Alpine chain. It is bordered in the east by the Alps, in the north and east by the Carpathians and in the south by the Dinarides (see Fig. 6). The basin originated from a combination of tectonic escape from the eastern Alps (Ratschbacher et al. 1991) and subduction-related processes along the Carpathian front. The latter include subduction rollback (Stegena et al. 1975; Horv~th & Berckhemer 1982), steepening of the subducting slab due to the eastward mantle flow (Doglioni 1990) and mantle diapirism (Stegena et al. 1975; Becker 1993). The basement consists of Palaeogene retroarc basin deposits and Palaeozoic to Cretaceous rocks, stacked on top of each other as imbricate nappes during the Cretaceous Alpine collision (Csontos et al. 1992; Tari et al. 1992). The Neogene (Mid-Miocene) extension is characterized by the development of rift- and pull-apart basins and the formation of metamorphic core complexes (Rumpler & Horvfith 1988; Tari et al. 1992). A detailed discussion of the origin of the basin is given by Horwith (1993). During the whole post-rift period the basin was a lake, with water depths reaching more than 1000 m (Horvfith et al. 1988; K~izm6r 1990). The basin was filled by a large delta system which moved from the Carpathian domain towards the inner parts of the basin (Great Hungarian Plain).

18 /

R. VAN BALEN & S. CLOETINGH ~(,

_~ ~,

116

118

I~A~H~I~H nt~'~filca thrta= tnH th,~ P ~ n n n n i ~ n

t20 l~,n,~ira

122

124

~26

Foredeep

•, Alpines Ikaline ; surface uaternary

km

Fig. 6. The Pannonian Basin and its surroundings. A-A' denotes the modelled cross-section. The contour lines represent the depth to the pre-Neogene basement, after Horvfith (1993).

A n o m a l o u s late-stage subsidence and inplane stress during the Late Pliocene Quaternary Abundant evidence is available for anomalous subsidence in the Pannonian Basin during the Late Pliocene and Quaternary (Horvfith & Cloetingh pers. comm.; Van Balen et ai. 1994). In general, continuing and increased subsidence is recorded in the basin center, while decreased subsidence and uplift are recorded at the flanks. The subsidence curves shown in Royden et al. (1983) and Demetrescu & Polonic (1989) show evidence of anomalous subsidence. Late-stage acceleration of subsidence and uplift can be observed in several wells. Recent movements in the Romanian part of the basin, as evidenced by geodetic measurements, are uplift at the basin margins and subsidence in the basin centre (L~z~rescu et al. 1983). Studies of the Danube river terraces and travertine horizons show clear evidence for a 200-300 m uplift of the Transdanubian Central Range (R6nai 1974; Horvfith & Cloetingh pers. comm.). Furthermore, Quaternary erosion has removed great parts of Late Pliocene sediments (R6nai 1985) with

exceptions of parts which were covered by contemporaneous basaltic lava flows (Horvfith & Cloetingh pers. comm.). The top of the Pliocene mega-sequence is located 600-650m below the Pliocene palaeo-base level in the deep sub-basins and 300-400m above it in the mountainous areas, giving rise to about 1000 m difference within the Pannonian Basin (excluding the Carpathian area) (R6nai 1974). Quaternary sediments are very thin in the mountainous areas (50-70 m) and are shown to onlap on to the Pliocene sequences by R6nai (1985).

Fluid)tow and hydrocarbon migration in the Pannonian Basin The main source rocks for hydrocarbons in the Pannonian Basin are Miocene and Early Pliocene marls (Dank 1988; Sarkovid et al. 1991). The Miocene marl deposits are generally overpressured in the Pannonian Basin (Clayton et al. 1990). Modelling of maturation and overpressuring (Horwith et al. 1987; Szalay 1988) showed that the hydrocarbon generation started between 9 and 6 Ma ago. Overpressures started to

BASIN MODELLING CONSTRAINTS develop at the same time. The high amounts of overpressure and a possible change of the overall stress field (Horvfith et al. 1987) resulted in local tensile hydrofracturing of basement. These fractures are sometimes filled with hydrocarbons. The most common hydrocarbon traps are compactional anticlines over basement highs (Dank 1988). Along our modelling profile three major hydrocarbon occurrences are located at the margins of the deep sub-basins, in compactional anticlines (Dank 1988). Recently, in the Serbian part of the Pannonian Basin, fields have been discovered in non-structural traps (pinch-outs, flexures, etc.) (Sarkovid et al. 1991). Hydrostatic fluid pressures are present down to a depth of about 1800 m, below which the fluid overpressures start to develop (Clayton et al. 1990; Szalay 1982, 1988). The amount of overpressuring as measured at the flanks of the B6k6s Sub-basin can be up to 15 MPa, while in the Mako Trough the maximum measured value is around 45 MPa. In general, the overpressure is nearly hydrostatic over the basement highs. M o d e l l i n g strategy

The modelled profile through the Pannonian Basin has a roughly SW-NE orientation and transects the Sava, Drava, Mak6 and B6k6s sub-basins (see Fig. 6). In order to make a forward model for the compaction-driven fluid flow system, a good stratigraphic model is required. The thickness of the crust and subcrustal lithosphere before the Neogene extension for the Pannonian Basin basement are assumed to be comparable to the current thicknesses in the Eastern Alps (F. Horv~th pers. comm. 1993), giving a crustal thickness of 42.5 km and a sub-crustal lithosphere thickness of 80 km before rifting. The current sub-crustal lithosphere thickness below the Great Hungarian Plain is about 40 km or less. A Neogene basement depth map (Horv~ith 1985), a lithospheric thickness map (HorvS.th 1993), palaeogeographic maps of the main sedimentary sequences (Juh~.sz 1991) and a cross-section, going through the B6k6s subbasin, analysed by using seismic sequence stratigraphic principles (Vakarcs et al. 1994) were used to constrain the stratigraphic forward modelling. Based on literature (e.g. B6rczi & Phillips 1985; Horv~ith et al. 1988; Mattick et al. 1988; Juh~isz 1991; KS.zmer 1990) the stratigraphy is subdivided into 7 chrono- and lithostratigraphic units (see Table 2). Ages of the different units are constrained by the analyses of Vakarcs et al. (pers. comm.).

19

Table 2. Stratigraphy of the Pannonian Basin Unit

Age (Ma)

Facies

Sand fraction

Q AP DP DS PD DB B

2.4-- 0.0 3.9- 2.4 5.5- 2.4 6.3- 5.5 8.2- 6.3 14.8- 8.2 16.5-14.8

Alluvialplain Alluvialplain Delta plain Delta slope Pro delta Deep basin Basal sst

0.2 0.2 0.8 0.2 0.2 0.0 1.0

For modelling purposes the lithology has been simplified to a mixture of a 'marl' and a 'sandstone' type of sediment. The porosity versus effective pressure trends for the two lithologies are based on porosity versus depth data given by Szalay (1982) (see Fig. 7). The 'deep' data for marls were not used, because these porosities are abnormally high due to fluid overpressure and, presumably, secondary porosity (Szalay pers. comm. 1993). The data were fitted to exponential curves of the type given by equation 2. The relationship between porosity and permeability is reconstructed from permeability-depth curves and the porosity- depth curves for the two lithologies given by Szalay (1982). These data were fitted to curves obeying equation 4.

depth versus porosity

0.0 1.0 ~'2.0 ~3.0 4.0

+ = sandstone

"

0.0

0.1

0.2 porosity

Fig. 7. Porosity versus depth curves for marls and sandstones in the Pannonian Basin, after Szalay (1982).

20

R. VAN BALEN & S. CLOETINGH Stratigraphy in the Mako en Bekes sub-basins

oi t O,

~o O O t.r) O ¢.O

300.0

400.0 distance (krn)

Fig. 8. Stratigraphic cross-section through the Pannonian Basin.

Results Stratigraphic modelling. The stratigraphic crosssection based on the data sources mentioned earlier is shown in Fig. 8. Clearly recognizable are the truncations (or offlaps) at the basin margins. The goal for the stratigraphic forward modelling was to reproduce the same stratigraphic cross-section using all the available constraints, in order to constrain the lithospheric necking depth during rifting. The necking depth is constrained by the current earthquake depths, which are most frequent at depths between 4 and 10 km, and by the low positive bouguer gravity anomaly measured above the Great Hungarian Plain (Bielik 1988), which indicates the downward state of flexure of the lithosphere. The forward modelling demonstrates that a shallow necking depth of 7 km gives the best fit to the observed stratigraphy. Deeper necking depths lead to unacceptable palaeo-waterdepths for the syn-rift sediments, whereas shallower necking depths require too high crustal extension factors. Both effects are due to the large amount of subcrustal extension. The result of the forward modelling is shown in Fig. 9. Also shown in the same figure are the best estimates, obtained from forward modelling, for the lake level variation (the Pannonian Basin was not connected to the world oceans during the post-rift period) and palaeo-water depth for the deepest part of the Mako Trough (the position of the Hod-I well). For the lake level variation we have taken into account the observation made

by Tari et al. (1992) an d Vakarcs et al. (1994) that the lake level varied with the eustatic sealevel changes, though not necessarily with the same amplitude. The most dramatic lake level change occurs around the Messinian (6.3Ma). Lake level dropped and subsequently rose 100-200 m. The palaeo-water depths are less constrained; especially the deep basin marls (DB) were deposited at unknown waterdepths ( > 1000 m). The deltaic sediments are better constrained, due to the geometry of delta systems. We need a palaeo-water depth maximum of 1300 m in order to make the model fit. Although huge, it seems plausible. The lake-level variations we have applied are the same as the eustatic sealevel changes given by Haq et al. (1987) with the exception of the Messinian where we included a 150 m sea level drop and subsequent rise in the Late Pliocene to Quaternary. During this time interval there is, according to Haq et al. (1987) an eustatic sea-level rise. We found that during this time interval there cannot have been a lake-level rise in the Pannonian Basin, because this would result in onlapping of sediments on the basement highs, which is not observed. We have therefore kept the lake level fixed since the Late Pliocene. The increase in the level of compressive intraplate stress started at 3.9 Ma, reaching a value of 300MPa at 2.4Ma and subsequently dropping to a present-day value of 200 MPa. As shown by an inspection of Fig. 9, the overall stratigraphy and basin shape predicted by the model compares well with the observed

BASIN MODELLING CONSTRAINTS

21

E¢~ ¢-

"0

tD

0

1O0

200 distance (km)

Fig. 9. The stratigraphy as derived by our forward modelling. The insets show the adopted lake level curve and the palaeo-water depth curve for the deepest part of the Mako trough (the position of the Hod-I well).

large scale pattern in the basin geometry and basin fill. Observed differences in the fine structure of the predicted and observed stratigraphy are probably caused by simplifications in the model assumptions and partly due to uncertainties in the available data. The latter include unknown lake-level changes, previous intraplate stress events, and the assumption that lithostratigraphic units are equivalent to chronostratigraphic units. Although this is approximately true for the older units, this might be wrong for the AP unit, explaining why in contrast to observation, in the model the AP unit onlaps on the basin margin.

Table 3. Numerical values adopted in fluid flow

Compaction-driven fluid flow modelling in the Pannonian Basin. Using our forward strati-

36MPa, due to the increased sedimentation rates above the two sub-basins (2.4Ma). The prediction for the current amount of overpressure is depicted in Fig. 10c. The maximum overpressure reaches 41 MPa, which is in excellent agreement with observations in the Hod-I well (Szalay 1982, 1988). The overpressures are maximal in the DB unit, as it is the least permeable. The overpressures decrease in the deeper B unit. This has implications for the hydrocarbon migration, as discussed below. The B6k6s Basin has developed a slightly less overpressure than the Mako Trough.

graphic model we have made an analysis of the compaction-driven fluid flow system in the Pannonian Basin. The composition of the stratigraphic units is shown in Table 2. The empirical parameters for porosity and permeability laws are depicted in Table 3. The results for three different time slices are shown in Fig. 10. The fluid overpressures prior to the compressive stress event are shown in Fig. 10a. The maximum overpressure is approximately 21MPa, reflecting the dramatic increase in sedimentation rates during the deposition of the DS unit due to the arrival of the Pannonian delta in the deep sub-basin (5.5 Ma). The effect of the increase of compressive stress on the fluid overpressures is shown in Fig. 10b. The maximum overpressure has dramatically increased to

calculations Sediment type

Parameter

Value

Sandstone

4) logkhor

0.475e -45 I°-~Peff -- 18 + 8+ 1.62 2.7gcm -3 0.65e -s's2 t°-SPeff -21 + 8+

khor/kv~rt Marl

p~ + Iogkhor khor/kvert P~

11.7 2.7g cm-3

Discussion and conclusions Models incorporating the finite strength of the lithosphere during rifting and changes in the stress regime during the post-rift phase of

22

R. VAN BALEN & S. CLOETINGH

(a)

A

EOJ

t~)

0

100

200 distance (km)

(b) o

,,e t-i1) "0

1O0

200

distance (kin) extensional basins can successfully explain a number of important features recognized in the basin fill, not compatible with models invoking local isostasy for the basin formation and equating post-rift subsidence to thermal subsidence. Better understanding of the rheology of the lithosphere and its interplay with erosion processes and stress changes enhances the quality of predictions generated by forward models for basin stratigraphy and fluid flow. In particular, the modelling has demonstrated that an increase in the level of compressive stress or a decrease in the amount of tensile stress can lead to substantial flank uplift and basin centre

subsidence. Therefore, more meteoric water will infiltrate from the margins into the basin, due to increased sediment exposure and gravity potential. At the same time, sedimentation rates in a basinward position increase, leading to an increase in compaction-driven fluid overpressuring and flow velocities. An increase of tensile stress or a decrease of compressive stress have the opposite effects. As intraplate stresses can change sign and magnitude on short time scales, the fluid regime also can change very rapidly due to this mechanism. This process triggers faulting and fracturing, alters heatflow and diagenesis (Horbury & Robinson 1993), enhances primary

BASIN MODELLING CONSTRAINTS

23

(c)

ECq ,¢ ¢Q~ "o

0

100

200

distance (krn) Fig. 10. The results from the compaction-driven fluid flow modelling. Thin lines represent the stratigraphic layering, thick lines are isobars (1, 10, 20, 30 and 40 MPa). (a) Fluid overpressures at 5.5 Ma, showing the effect of the increased sedimentation rates due to the arrival of the Pannonian delta. The maximum overpressure is 21 MPa. (b) The overpressuring regime at 2.4 Ma, demonstrating the impact of the stress-induced subsidence on the fluid overpressures, which have increased to a maximum value of 36 MPa. (c) The present overpressures as predicted by our modelling. The maximum value of 41 MPa is in excellent agreement with observations. migration of hydrocarbons and transportation of metal-rich brines in a pulse-like way. The forward modelling of fluid flow in the Pannonian Basin has shown that an increase in the level of compressive stress during the late stage evolution of this extensional basin has altered the compaction-driven fluid overpressuring substantially. Fluid overpressures have almost doubled due to the effect of the increase of compressive stress. This has a number of implications for the diagenesis of sediments and primary migration of hydrocarbons. The hydrocarbon generation started between 9 and 6 Ma (Horv~th etal. 1987; Szalay 1988), implying that during the stress-induced change of the fluid regime in the Pannonian Basin the hydrocarbons were mature and actively expelled. The dramatic increase of overpressure has probably enhanced the primary migration considerably. Our modelling also shows that the overpressures decrease with depth in the deepest part of the basin, enhancing primary migration into this unit from the overlying sediments, enhancing the prospectivity of the deepest unit. This research was funded by the IBS (Integrated Basin Studies) project, part of the JOULE I1 research programme funded by the Commission of European Communities (contract nr. JOU2-CT 92-0110). F. Horvgtth and L. Lenkey made significant contributions

to the modelling of the Pannonian Basin. M. G61ke kindly provided us with the European stress map (Fig. 1). IBS contribution No. 7 Publication No. 8 of the Netherlands Research School of Sedimentary Geology.

References ANGEVINE, C.L., HELLER, P. & PAOLA, C. 1990.

Quantitative sedimentary basin evolution. American Association of Petroleum Geologists, Continuing Education Course Note Series, 32. BASSI, G., KEEN, C.E. & POTTER,P. 1993. Contrasting styles of rifting: models and examples from the eastern Canadian margin. Tectonics, 12,639-655. BECKER, A. 1993. Contemporary state of stress and neotectonic deformation in the CarpathianPannonian region. Terra Nova, 5,375-388. B~RCZl, I. & PHILLIPS, R.L. 1985. Processes and depositional environments within Neogene deltaic-lacustrine sediments, Pannonian Basin, southeast Hungary. Geophysical Transactions, 31, 55-74. BETHKE, C.M. 1985. A numerical model of compaction-driven groundwater flow and heat transfer and its application to the paleohydrology of intracratonic sedimentary basins. Journal of Geophysical Research, 90, 6817-6828. --,

HARRISON, W.J., UPSON, C. • ALTANER, S.P.

1988. Supercomputer analyses of sedimentary basins. Science, 239,261-267.

24

R. VAN BALEN & S. CLOETINGH

DEMETRESCU,C. 8z POLONIC,G. 1989. The evolution of the Pannonian Depression (Romanian sector) as derived from subsidence and heat flow data. Tectonophysics, 164,287-299. DOGLIONI, C. 1991. A proposal for the kinematic modelling of W-dipping subductions - possible applications to the Tyrrhenian-Appenine system. Terra Nova, 3,423--434. DUNBAR, J.A. & SAWYER,D.S. 1989. HOW preexisting weaknesses control the style of rifting. Journal of Geophysical Research, 94, 7278-7292. FLEMINGS, P.B. & JORDAN, T. 1989. A synthetic stratigraphic model of foreland basin development. Journal of Geophysical Research, 94, 3851-3866. GLUYAS, J.G., GRANT, S.M. & ROBINSON,A.G. 1993. Geochemical evidence for a temporal control on sandstone cementation. In: HORBURY, A.D. & ROBINSON, A.G. (eds), Diagenesis and Basin Development, AAPG Studies in Geology Series, 36, 23-33. HARRISON, W.J. & TEMPLE, R.N. 1993. Diagenetic pathways in sedimentary basins. In: HORBURY, A.D. & ROBINSON, A.G. (eds) Diagenesis and Basin Development, AAPG Studies in Geology Series, 36, 69-86. HAQ, B.U., HARDENBOL,J. & VAIL, P. 1987. Chronology of fluctuating sea levels since the Triassic. Science, 235, 1156-1167. HORBURY, A.D. & ROBINSON, A.G. (eds) 1993. Diagenesis and Basin Development, AAPG Studies in Geology Series, 36. HORV~,TH, F. 1985. Thickness of Neogene-Quaternary basin fill (map). Hungarian Geological Institute, Budapest, Hungary. 1993. Towards a kinematic model for the Pannonian Basin. Tectonophysics, 226,333-357. & BERCKHEMER,H. 1982. Mediterranean backarc basins. In: BERCKHEMER, H & HSU, K. (eds) Alpine-Mediterranean geodynamics. Geodynamic Series 7, American Geophysical Union and Geological Society of America, Boulder Colorado, 141-173. - - , DOVENYI,P., PAPP, S., SZALAY,A. & HARGITAY, M. 1987. Tectonic control on migration and trapping of hydrocarbons in the Pannonian Basin. In: DOLIGEZ, B. (ed.) Migration of hydrocarbons in sedimentary basins, Editions Technip, Paris, 667-681. --, SZALY, A. & ROYDEN, L. 1988. Subsidence, thermal, and maturation history of the Great Hungarian Plain. In: ROYDEN, L.H. & HORV~,TH,F. (eds) The Pannonian Basin: A Study in Basin Evolution, American Association of Petroleum Geologists, Memoirs, 45,355-372. JANSSEN, M.E., TORNI~,M., CLOETINGH,S. & BANDA, E. 1993. Pliocene uplift of the eastern Iberian margin: inferences from quantitative modelling of 208,221-241. the Valencia Trough. Earth and Planetary Science DANK, V. 1988. Petroleum geology of the Pannonian Basin, Hungary: an overview. In: ROYDEN, Letters, 119,585-598. JENSEN, R.P. & DOR~, A.G. 1993. A recent NorL.H. & HORV~TH, F. (eds) The Pannonian wegian Shelf heating event - fact or fantasy? In: Basin: A Study in Basin Evolution. American DORI~, A.G. et al. (eds) Basin Modelling: AdAssociation of Petroleum Geologists Memoirs, vances and Applications. Norwegian Petroleum 45,319-331.

BIELIK, M. 1988. Analyses of the stripped gravity map of the Pannonian Basin. Geologica Carpathica, 39, 99-108. BRAUN, J. & BEAUMONT, C. 1989. A physical explanation of the relationship between flank uplifts and the breakup unconformity at rifted continental margins. Geology, 17,760-764. CARTER,N.L. & TSENN, M.C. 1987. Flow properties of continental lithosphere. Tectonophysics, 136, 27-63. CARTWRIGHT, J.A. 1993. Giant mud cracks in the Paleogene of the North Sea: evidence for basin scale hydrofracturing of overpressured shale sequences, extended abstract. In: PARNELL, J., RUFEELL, A.H. & MOLES, N.R. (eds) Geofluids '93, Torquay. 4-7 May 1993, 172-174. CLAYTON, J.L., SPENCER,C.W., KONCS, I. & SZALAY, A. 1990. Origin and migration of hydrocarbon gases and carbon dioxide, B6k6s Basin, southeastern Hungary. Organic Geochemistry, 15, 233-247. CLOETINGH,S. 1991. Tectonics and sea level changes a controversy? In: MULLER, D., WEISSEL, H. & MCKENZIE, D. (eds) Controversies in modern geology: a survey of recent developments in sedimentation and tectonics. Academic Press, London, 249--277. -& BANDA, E. 1992. Mechanical structure of Europe's lithosphere. In: BLUNDELL, D., MULLER, S & FREEMAN, R. (eds) A continent revealed: The European Geotraverse. Cambridge University Press, 80-91. & KooI, H. 1992. Intraplate stresses and dynamical aspects of rift basins. Tectonophysics, 215,167-185. --, GRADSTEIN, F., KOOI, H., GRANT, A. & KAMINSKI,M. 1990. Plate reorganization: a cause for rapid Neogene subsidence and sedimentation around the North Atlantic? Journal of the Geological Society, London, 147,495-506. , KooI, H. & GROENEWOUD,W. 1989. Intraplate stress and sedimentary basin evolution. In: PRICE, R.A. (ed.) Origin and Evolution of Sedimentary basins and these energy and mineral resources. American Geophysical Union, Geophysical Monographs, 48, 1-16. , McOuEEN, H. & LAMBECK, K. 1985. On a tectonic mechanism for regional sea level variations. Earth and Planetary Science Letters, 75, 31-61. - - , SASSI, W. & HOVATH,F. (eds) 1993. The origin of sedimentary basins: inferences from quantitative modelling and basin analyses. Tectonophysics, 226, 1-504. CSONTOS, L., NAGYMAROSY, A., HORVATH, F. & KovAc, M. 1992. Tertiary evolution of the Intra-Carpathian area: a model. Tectonophysics,

BASIN MODELLING CONSTRAINTS Society (NPF), Special Publications, 3, Elsevier, Amsterdam, 85-106. JUH~,SZ, G. 1991. Lithostratigraphical and sedimentological framework of the Pannonian (s.1.) sedimentary sequence in the Hungarian Plain (Alf61d), Eastern Hungary. Acta Geologica Hungarica, 34, 53-72. KAZM~R, M. 1990. Birth, life and death of the Pannonian Lake. Palaeogeography, Palaeoclimatology, Palaeoecology, 79,171-188. KENYON, P.M. & TURCOTrE, D.L. 1985. Morphology of a delta prograding by bulk sediment transport. Geological Society of America Bulletin, 96, 14571465. KNIPE, R.J. 1993. The influence of fault zone processes and diagenesis on fluid flow. In: HORBURY,A.D. & ROBINSON, A.G. (eds) Diagenesis and Basin Development, AAPG Studies in Geology Series, 36,135-151. Kool, H. & CLOETINGH, S. 1989. Intraplate stresses and the tectono-stratigraphic evolution of the central North Sea. In: TANKARD, A.J. & BALKWILL, H. (eds) Extensional tectonics and stratigraphy of the North Atlantic margins. American Association of Petroleum Geologists Memoir, 46,541-558. &~ 1992. Lithospheric necking and regional isostasy at extensional basins 2. Stress induced vertical motions and relative sea level changes. Journal of Geophysical Research, 97, 1757317591. , -& BURRUS, J. 1992. Lithospheric necking and regional isostasy at extensional basins 1. Subsidence and gravity modelling with an application to the Gulf of Lions margin (SE France). Journal of Geophysical Research, 97, 1755317571. L~,Z.~RESCU, V., CORNEA, I., R.g.DULESCU, F. & POPESCU, M. 1983. Moho surface and recent crustal movements in Romania: geodynamic connections. Proceedings of the Institute of Geology and Geophysics in Bucarest, 63, 83-91. MCKENZIE, D.P. 1970. Some remarks on the development of sedimentary basins. Earth and Planetary Science Letters, 40, 25-31. MATHCK, R.E., PHILIPS, R.L. & RUMPLER, J. 1988. Seismic stratigraphy and depositional framework of sedimentary rocks in the Pannonian Basin, southeastern Hungary. In: ROYDEN, L.H. & HORV~.TH, F. (eds) The Pannonian Basin: A Study in Basin Evolution. American Association of Petroleum Geologists Memoirs, 45, 117-146. MOLLER, B., ZOBACK, M.L., FvcrIS, K., MASTIN, L., GREGERSEN, S., PAVONI, N., STEPHANSON,O. & LJUNGGREN, C. 1992. Regional patterns of tectonic stress in Europe. Journal of Geophysical Research, 97, 11783-11803. PEPER, T., BEEKMAN, F. & CLOETINGH, S. 1992. Consequences of thrusting and intraplate stress for foreland basin development and peripheral areas. Geophysical Journal International, l l l , 104-126. , VAN BALEN, R. & CLOETINCH, S. 1994. Implications of orogeny, intraplate stress variations -

-

25

and eustatic sealevel change for foreland basin stratigraphy - inferences from numerical modelling. In: DOROBEK,S. & Ross, G. (eds) Stratigraphy and foreland basins. SEPM Special volume, in press. PHILIP, H. 1987. Plio-Quaternary evolution of the stress field in Mediterranean zones of subduction and collision. Annales Geophysicae, 5b, 301-320. POSAMENTIER,H.W. & ALLEN, G.P. 1993. Variability of the sequence stratigraphic model: effects of local basin factors. Sedimentary Geology, 86, 91-109. RATSCHBACHER,L., FRISCH,W., LINZER,H. & MERLE, O. 1991. Lateral extrusion in the Eastern Alps, part 2: structural analyses. Tectonics, 10, 257271. REEMST, P., CLOETINGH, S. & FANAVOLL, S. 1994. Tectono-stratigraphic modelling of Cenozoic uplift and erosion in the SW Barents Sea. Marine and Petroleum Geology, in press. R6NAI, A. 1974. Size of Quaternary movements in Hungary's area. Acta Geologica Academiae Scientiarum Hungaricae, 18, 39--44. 1985. Limnic and terrestial sedimentation and the N/Q boundary in the Pannonian Basin. In: KRETZOI, M. & P~csl, M. (eds) Problems of the Neogene and Quaternary. Akad6miai Kiad6, Budapest, 000-4)00. ROVDEN, L.H. & D6V~NVI, P. 1988. Variations in extensional styles at depth across the Pannonian Basin. In: ROYDEN, L.H. & HORVATH, F. (eds)

The Pannonian Basin: A Study in Basin Evolution. American Association of Petroleum Geologists Memoirs, 45,235-255. & KEEN, C.E. 1980. Rifting processes and thermal evolution of the continental margin of eastern Canada determined from subsidence curves. Earth and Planetary Science Letters, 51, 343-361. --, HORVATH, F. & NAGYMAROSY,A. 1983. Evolution of the Pannonian Basin system 2: subsidence and thermal history. Tectonics, 2, 91-137. RUMPLER, J. & HORV.~TH, F. 1988. Some representative seismic reflection lines from the Pannonian Basin and their structural interpretation. In: ROYDEN, L.H. & HORV~,TU,F. (eds) The Pannonian Basin: A Study in Basin Evolution. American Association of Petroleum Geologists Memoirs, 45,153-171. SARKOVI(~, M., STANKOVI(~,S. & MILOSLAVLJEVR~,S. 1991. Petroleum geology of the southeastern part of the Pannonian Basin. Terra Nova, 3, 550-554. SCHLAGER, W. 1993. Accommodation and supply - a dual control on stratigraphic sequences. Sedimentary Geology, 86,111-136. SCLATER, J.G. & CHRISTIE, P.A.F. 1980. Continental stretching: an explanation of the .post-midCretaceous subsidence of the central North-Sea Basin. Journal of Geophysical Research, 85, 3711-3739. , ROYDEN, L., HORV~TH, F., BURCHFIEL, B.C., SEMKEN,S. & STEGENA,L. 1980. The formation of the Intra-Carpathian basins as determined from

26

R. VAN BALEN & S. CLOETINGH

subsidence data. Earth and Planetary Science Letters, 51,139-162. Sin, Y. & WANG, C. 1986. Pore pressure generation in sedimentary basins: overloading versus aquathermal. Journal of Geophysical Research, 91, 21532162. STEGENA, L., G~CZY, B. & HORVXTH, F. 1975. Late Cenozoic evolution of the Pannonian Basin. Tectonophysics, 26, 71-90. SWITSKI, J.P.M., SMrrH, J.N., CALABRESE,E.A. & BOUDREAU, B.P. 1988. Basin sedimentation and the growth of prograding deltas. Jounral of Geophysical Research, 93, 6895-6908. SZALAY,A. 1982. A rekonstrukci6s szemldletii fOldtani

kutatds lehetOs~gei a szdnhidrogdn-perspektivdk elOrejelz~sdben [Possibilities of the reconstruction of basin evolution in the prediction of hydrocarbon prospects]. PhD thesis, Hungarian Academy of Sciences, Budapest. 1988. Maturation and migration of hydrocarbons, SE Pannonian Basin. In: ROYDEN, L.H. & HORV~,TH, F. (eds) The Pannonian Basin: A Study in Basin Evolution. American Association of Petroleum Geologists Memoirs, 45,347-354. TARI, G., HORV.~TH,F. d(z RUMPLER,J. 1992. Styles of extension in the Pannonian Basin. Tectonophysics, 208,203-219. VAN BALEN,R. ¢~zCLOETINGH,S. 1993. Stress-induced fluid flow in rifted basins. In: HORBURY, A.D. & ROBINSON, A.G. (eds) Diagenesis and Basin Development. AAPG Studies in Geology Series, 36, 87-98.

~,

LENKEY,L., CLOETINGH,S. • HORVXTH,F. 1994. Two-dimensional modelling of stratigraphy and fluid flow in the Pannonian Basin. Marine and Petroleum Geology, in press. VAN DER BEEK, P., CLOETINGH,S. & ANDRIESSEN,P. 1994. Mechanisms of extensional basin formation and vertical motions at rift flanks: Constraints from tectonic modelling and fission-track thermochronology. Earth and Planetary Science Letters, 121, 417-433. VAN WAGONER, J., POSAMENTIER, H.W., MITCHUM, R.M., VAIL, P.R., SARG, J.F., LOUTIT, T.S. & HARDENBOL, J. 1988. An overview of fundamentals of sequence stratigraphy and key definitions. Sea-level changes - an integrated approach, The Society of Economic Paleontologists and Mineralogists Special Publications 42, 40-45. WEISSEL, J.K. & KARNER, G.D. 1989. Flexural uplift of rift flanks due to mechanical unloading of the lithosphere during extension. Journal of Geophysical Research, 94, 13919-13950. ZOBACK,M.L. 1992. First and second order patterns of stress in the lithosphere: The World Stress Map Project. Journal of Geophysical Research, 97, I 1703-11728. ZIJERVELD, L., STEPHENSON, R.A., CLOETINGH, S., DUIN, E. & VAN DEN BERG, M. 1992. Subsidence analysis and modelling of the Roer Valley graben (Southeastern Netherlands). Tectonophysics,

208,159-171.

Fluid flow and heat transport in the upper continental crust DAVID

DEMING

School o f Geology & Geophysics, University o f Oklahoma, Norman, O K 73019-0628, USA

Abstract: The upper 10-15 km of the continental crust is saturated with aqueous brines and is sufficiently permeable to allow the circulation of these fluids. The most important driving forces for fluid flow in the continental crust are topography gradients, sediment compaction and diagenesis, and buoyancy forces. Topographically-driven flow is generally the most efficient mechanism for mass and heat transport, and therefore has recently been favoured in theories of the origin of Mississippi Valley-type lead-zinc deposits. The magnitude of heat transport by fluid movement in the crust has an exponential dependence upon the depth of circulation. In areas of the continental crust with appreciable topography gradients (c.0.01) and permeabilities greater than c. 10-17m2, the background thermal state will likely be appreciably perturbed by groundwater movement. Fluid circulation in the upper crust is both continuous and pervasive due to topography and fluid-density gradients. The old conception of the continental crust as an unchanging body of solid rock should be replaced by a paradigm that recognizes the continental crust as a two-component system; a solid framework which continuously evolves through thermal, chemical, and mechanical interaction with crustal fluids.

Importance of fluid and heat flow With the advent of plate tectonics, the concept of a static Earth was replaced by a new picture of the Earth's crust as a domain caught up in a process of continual evolutionary change. An increasing knowledge of the importance of fluids in the Earth's crust is now invalidating the old concept of the crust as being composed of unchanging bodies of solid rock. As geological fluids move through the crust they enter into chemical reactions with solid components of the crust, mobilize and transport both heat and mass, alter the geologic and geophysical properties of crustal rocks, and create economic concentrations of important ore minerals. Fluids play an important role in nearly all geological processes and are ubiquitous in the continental crust to depths of at least 10 to 15km (see Table 1 and Nesbitt & Muehlenbachs 1989, 1991; Zoback et al. 1988; N R C 1990; Oliver 1986, 1992). In the new paradigm, the continental crust is viewed as a two-component system: a solid framework that evolves through interactions with geological fluids. A convenient division may be made between hydrological regimes in the upper and lower continental crust. In the upper continental crust, the permeability of sedimentary and crystalline rocks may be sufficiently high to allow continuous flow systems that may conceivably persist for tens of millions of years. Below depths of

Table 1. Possibleindicators offluids in the continental crust (after Nur & Walder, 1990, p. 114) Indicator

Depth range

Water table Deep wells Reservoir induced seismicity Crustal low velocity zones Crustal electrical conductivity zones Oxygen isotopes Metamorphism Crack healing and sealing Hydrothermal ore deposits Crustal seismic attenuation zones Low stress on faults Silicic volcanism Fluid inclusions

0-2 km 0-12 km 0-12 km 7-12 km 10-12 km 0-12 km >20 km ? >5 km 7-12 km 0-10 km near surface ?

approximately 10-20km, however, the continental crust is thought to be essentially impermeable. Open fractures probably cannot exist below the brittle-ductile transition, and unfractured crystalline rocks are otherwise as impermeable as window glass. Nevertheless, there is considerable geochemical evidence from examination of exhumed terrains that large amounts of fluid have circulated through the lower crust (Walther 1990). One way in which the inferences of large fluid fluxes can be reconciled with the rheological requirements of low permeability is episodic hydrofracturing (Nur & Walder 1990).

From PARNELL,J. (ed.), 1994, Geofluids: Origin, Migrationand Evolution of Fluids in Sedimentary Basins, Geological Society Special Publication No. 78, 27-42.

27

28

D. DEMING S A L I N I T Y (%o) 0 0

200 I

I

400 I

if

• - ~-..

I

500

i

• m

•

m

E a:: I000

v

.m

•m

a

•

•w

•

•

w'lwm

1500-

3

w

i,

2000

.

'

' '

.

.

.

.

.

.

'

II0

'

....

I

. . . . . . . .

I00

I000

Total Dissolved Solids (mgll)

4 Fig. 1. Maximum observed salinities of pore fluids from selected sedimentary basins (after Hanor 1979; Ranganathan & Hanor 1987). Theoretical calculations based upon the physical properties of aqueous solutions (Norton & Knapp 1977; Norton 1984, 1990; Nur & Walder 1990) have shown the feasibility of this mechanism, and there is field evidence that cyclic flow mechanisms probably exist near magmatic intrusions in the upper crust (Titley 1990). Nature of fluids in the crust Fluids found in the continental crust are primarily aqueous brines; most data concerning these brines come from sedimentary basins. The salinity of aqueous brines from sedimentary basins is commonly found to increase with increasing depth (Fig. 1) to a maximum of more than 40% solids by weight (Dickey 1969; Hanor 1979). Stable isotope studies indicate that most of the formation fluids found in sedimentary basins are not connate water trapped at the time of sediment deposition, but of meteoric origin (Clayton et al. 1966; Kharaka 1986). The presence of juvenile fluids in the crust is relatively rare (White 1974; Truesdell & Hulston 1980). In comparison to sedimentary environments there is less known about the nature of fluids in crystalline rocks, no doubt due to the scarcity of boreholes and the limited availability of fluid samples. One of the best data sets that is available comes from a study of crustal fluids in the Canadian Shield. In the late 1970s a number of relatively shallow (c. 1600m, maximum depth) boreholes were drilled in crystalline rocks of the Canadian Shield for the purpose of studying possible sites for the disposal of nuclear-fuel waste products. A study of fluid samples from

Fig. 2. Salinities of pore fluids from crystalline rocks in the Canadian Shield (after Frape & Fritz 1987). these boreholes showed that the dominant fluid

was a highly saline Ca-Na-C1 brine. Total dissolved solids content was found to increase exponentially with depth, from fresh water near-surface (c.100m) to c.150g1-1 at c.1500m depth (Fig. 2). Had the boreholes penetrated depths exceeding c. 1600 m, salinity presumably would have continued to increase. However, the rate of increase could not have been maintained as extrapolation of the observed trend would have exceeded saturation values near 2000m depth. Analyses showed the isotopic signature of brines recovered from the Canadian boreholes to be very much different from fluids from sedimentary or geothermal environments, evidently reflecting a long history of intense fluid-rock interactions (Frape & Fritz 1987).

Mathematical description of fluid flow Hydraulic head

Fluid movement in porous media occurs in response to head gradients and buoyancy forces. Hydraulic head h is mechanical energy per unit mass of fluid, equivalent to the height above an arbitrary datum to which fluid will rise in a tube, otherwise known as the potentiometric surface (Hubbert 1940). It is convenient to divide head into an elevation and pressure term, h = z + P/pg

(1)

where h is hydraulic head, z is elevation above an arbitrary datum, P is fluid pressure, P is fluid density, and g is the acceleration due to gravity (Freeze & Cherry 1979). A persistent fallacy, first corrected by Hubbert (1940), is that fluids flow from regions of high pressure to low pressure. As can be seen by inspection of

FLUID FLOW AND HEAT TRANSPORT equation (1), head gradient is proportional to pressure gradient only for the special case where elevation is constant. Thus fluids do not necessarily flow in the direction of decreasing pressure. For example, water in a drinking glass does not spontaneously flow from the bottom to top of the glass. Equation of flow

Neglecting inertial forces, a continuum (volumeaveraged) description of the laminar flow of an inhomogeneous fluid through an anisotropic porous medium is (Bear 1972) q = - ( E / I x ) [VP+pgVz]

(2)

where q is the Darcy velocity or specific discharge (volumetric flow rate per unit time divided by cross-sectional_ area perpendicular to flow direction), E is the permeability tensor, Ix is fluid dynamic-viscosity, VP is the gradient of fluid pressure, z is elevation, and g, the acceleration due to gravity, is assumed to act in the z-direction. Under the assumptions stated above, equation (2) always holds true, regardless if flow is steady-state or transient. The Darcy velocity q is distinguished from linear velocity, v, which is the average speed with which individual water molecules move through a porous medium. In a porous medium of fractional porosity qb, linear velocity v is related to Darcy velocity q by v=q/+.

(3)

By defining an equivalent fresh-water hydraulic head h, h = z + P/pog

(4)

equation (2) can be rewritten as q=-(~pog/ix)[vh+P-P°Vz] Po

(5)

where Po is a reference fluid density. Thus fluid flow results not only from a head gradient Vh, but also fluid-density gradients which give rise to buoyancy forces. It follows that in the general case of a fluid of variable density, head is not a potential field as q is not exactly proportional to Vh (Bear 1972). However, if an assumption of constant density fluid is made, equation (5) reduces to the familiar form of Darcy's Law (Darcy 1856; Hubbert 1969) q = - (~ 9og/tx)Vh

(6)

q = - KVh

(7)

or

where K, the hydraulic conductivity tensor, is a

29

function of both the porous medium (rock) and the fluid moving through it. Hydraulic conductivity

Some of the factors that determine hydraulic conductivity may vary over orders of magnitude in different geological settings; others tend to be relatively constant throughout the crust. Fluid density, p, is determined by temperature, salinity, and pressure. Of these three factors, temperature and salinity are usually more important. For example, the density of pure water decreases by c.8% as temperature increases from 25 to 150°C at constant pressure. Assuming that this increase of temperature corresponds to a depth increase of c.5km (25°Ckm-1), the corresponding increase of pressure under hydrostatic conditions is c.50 MPa. An increase in pressure of 50 MPa at 25°C results in c.2% increase in water density. Thus the decrease in density due to thermal effects is about four times larger than the density increase due to pressure effects. In sedimentary basins, the decrease in fluid density due to increased temperature is usually more than offset by increases in salinity that result in fluid density increasing with increasing depth. However, in the absence of appreciable salinity gradients, fluid columns in the crust are potentially unstable and may be susceptible to convective overturn. The acceleration due to gravity, g, is essentially constant over the Earth's surface, varying from c.9.83 ms -2 at the poles, to c.9.78 ms -z at the equator for elevations at sea level and below. Dynamic fluid viscosity, Ix, is strongly temperature-dependent. An empirical formula for the dynamic viscosity of pure water is Ix = 2.414 x 10-5 x 10 (248"37/(T+133"15)) (8) where tx has units of k g m - l s -1 and T is temperature in degrees Celsius (Touloukian et al. 1975). Dynamic viscosity decreases by approximately an order of magnitude as temperature increases from 0 to 150° C. All other factors being equal, the reduction of viscosity with increasing temperature tends to promote deep flow in the crust (Smith & Chapman 1983). The effect of salinity is to increase the dynamic viscosity, but the magnitude of the effect is generally smaller than the temperature dependence (see Phillips et al. 1981). Generally the most significant factor that determines hydraulic conductivity is permeability, k. The permeability of Earth materials varies over more than 11 orders of magnitude, from values as high as 10-9 m 2 for

30

D. DEMING

karst limestone, to values lower than 10-2°m 2 for some shales and unfractured crystalline rocks. Presumably the average permeability of lithologies such as salt may be even lower than 10-2°m 2. One of the primary tenets of hydrogeology is that there are no completely impermeable rocks or geological materials (Bredehoeft et al. 1982). A possible exception is ice-rich permafrost, in that any water attempting to move through such a layer simply freezes. Rock permeability is highly variable. For example, permeability measurements on cores from sedimentary basins that sample geological horizons with uniform lithologies commonly vary over five orders of magnitude. Rock permeability also changes with scale (Brace 1980, 1984; Garven 1986; Ciauser 1992). Estimates of permeability made from pumping tests in wells commonly yield values one to two orders of magnitude greater than those derived from measurements on core samples. Core measurements may fail to sample fracture networks, the presence of which can dramatically increase the effective permeability. At the present time, it is not well understood if permeability also increases from the well scale (101-103m) to the regional scale (105-106m). Much of what we know about the change of permeability with scale comes from the studies of sedimentary basins in the central and western US by Bredehoeft and his colleagues at the US Geological Survey. Bredehoeft et al. (1983) found that for their numerical models of regional groundwater flow to yield acceptable matches to hydraulic head data, it was necessary to specify the permeability of the Cretaceous shale confining layer in South Dakota as one to three orders of magnitude higher than values derived from in situ well tests and laboratory measurements on cores. Similarly, Belitz & Bredehoeft (1990) found the regional scale permeability of the shale confining layer in the Bighorn Basin in the western US to be 1000 to 10000 times higher than would be inferred from laboratory measurements. The dramatic increase from the laboratory to regional scale was attributed to extensive high-angle thrust faulting. In contrast, Belitz & Bredehoeft (1988, 1990) estimated that the permeability of the shale confining layer in the Denver Basin of the western US was in the range 10-19-10-2°m 2, and exhibited no significant increase from the laboratory to regional scales. From these three studies, it would appear that scale dependence of confining layer permeabilities is entirely dependent upon fracturing and faulting, and thus the regional geology and tectonic setting. Clauser (1992) reviewed data from crystalline

rocks and noted that permeability increases approximately three orders of magnitude from the core (10-2-10 -1 m) to the well scale (101102m), but found no corresponding increase from the well to the regional scale (103-105 m). Deming (1993) studied regional-scale (c.300 km) groundwater flow and concomitant heat transport in the North Slope Basin, Alaska. Simulations of coupled heat and fluid flow in an idealized model showed that the general trend of near-surface conductive heat flow estimated from borehole data could be reproduced by numerical models incorporating permeability data from core measurements. Thus, it would appear in the case of the North Slope Basin that there is no apparent increase of permeability from the core to the regional scale. However, the direct comparison of core data with largescale indirect estimates is complicated by considerations such as possible bias in selection of intervals for core measurements and the choice of an appropriate algorithm for averaging core data (e.g., arithmetic, geometric, or harmonic mean?). These considerations made it difficult to draw conclusive inferences in this instance. Diffusion equation

Through the implicit assumption of constantdensity fluid, hydraulic head can be approximated as a hydraulic potential (or pseudopotential) for the purpose of obtaining insight into geological problems for which an exact answer is unobtainable. In this case, there are no buoyancy forces and the transient movement of fluids in the crust is described by the diffusion equation. In complex and heterogeneous geological environments, the assumption of constant-density fluid is nearly always violated, but the diffusion equation nevertheless is a useful approximation that can be used to obtain insight into the physics of fluid flow in the crust. For a porous medium that is both homogeneous and isotropic, the diffusion equation is Oh _ g V2h Ot Ss

(9)

where h is hydraulic head, t is time, K is hydraulic conductivity, V2h is the curvature of the head field, and Ss is a material property called specific storage. Specific storage refers to the physical ability of a porous medium to store fluid, and is largely determined by the compressibility of the porous medium itself and the compressibility of the fluid which saturates it S~= pg(c~ + CB)

(10)

FLUID FLOW AND HEAT TRANSPORT where P is fluid density, g is the acceleration due to gravity, + is porosity, c~ is porous medium compressibility, and B is fluid compressibility. Freeze & Cherry (1979) give the compressibility of water as 4.4 x 10-mPa -~, and cite a range of compressibilities for consolidated rocks that range from 10 -9 to 10-~IPa -1. Ge & Garven (1992) reviewed porous medium compressibilities for consolidated rocks and found most shale compressibilities to be of the order of 10 -9 Pa -~ , with other lithologies (sandstone, limestone, granite, quartzite, etc.) commonly of the order of 10-m-10 -~1Pa -1. Neuzil (1986) pointed out that most compressibility data reflect measurements in the laboratory at human time scales, and therefore may underestimate in situ compressibilities. Estimates of in situ compressibilities for sedimentary rocks derived from porosity-depth relationships commonly are one to three orders of magnitude higher than laboratory data. However, porosity reduction in the subsurface may not solely reflect mechanical compaction, as porosity may also be reduced by chemical processes such as mineralization. At the present time, it appears that our best estimate is that most rocks fall in the range 10 - u -< et <- 10 -8 Pa -~, with shales occupying the upper end of this range. For these compressibilities and a nominal porosity 4, of 10%, & ranges from c.10-4 to c. x 10-Tm -~. Characteristic l e n g t h - t i m e A useful rule-of-thumb derived from analogy with heat conduction theory is that the characteristic time -r required for a transient in the scalar pseudopotential h to diffuse a distance y is given by (Lachenbruch & Sass 1977). y2 "r-4D (11) where the hydraulic diffusivity D is K D = S~

(12)

For a characteristic sedimentary rock permeability of 10-1sin 2, specific storage in the range 10-5-10 -6 m -1 , and dynamic fluid viscosity of 5 × 10-4m s kg -I (pure water at 50 ° C), D is in the range 0.002 to 0.02m2s -1. Thus, for example, the time needed for a transient to diffuse through a regional flow system 100 km long is in the neighborhood of 4000-40 000 years. In contrast to fluid flow, the primary direction of conductive heat transport in the Earth is vertical and the thermal diffusivity (c.10 -6 m2s -1) is much lower than the hydraulic

31

diffusivity. Thus the characteristic time for transient thermal perturbations to diffuse is usually much greater than that required for hydraulic systems. For example, in the case of a sedimentary basin 10km thick, the time required for a significant degree of thermal equilibration through heat conduction is about 8 × 105 years.

Mathematical description of heat transport Fluid movement in the crust can be an effective agent for the redistribution of heat. The transport of heat by conduction and advection in a porous medium is described by the diffusionadvection equation 0 -~[T(OC)pr] = V . [XVT] -[qoC"

VT]+A*

(13)

where T is temperature, (pC)pr is the product of the bulk (porous rock) density and heat capacity (both fluid and matrix), p is fluid density,X is the porous rock thermal-conductivity tensor, q is Darcy velocity, C is fluid heat capacity, and A* is internal (usually radioactive) heat production per unit volume. Vertical fluid movement is usually required to perturb the thermal regime in the Earth's crust; little or no heat is transported by the horizontal movement of groundwater because isotherms are almost always parallel to the ground surface. If it is assumed that heat flow is vertical only, there are no heat sources or sinks, fluid Darcy velocity does not vary with respect to space, and physical properties such as thermal conductivity, density, and heat capacity are invariant with respect to both space and time, equation (13) can be simplified to OT . 02T OT 9Cpr 0-'7 = hz Oz----T - qzpC-~z

(14)

where OZT/Oz z is the curvature of the temperature field and qz is Darcy velocity in the z-direction. Lachenbruch & Sass (1977) showed that for steady-state (OT/Ot = 0) heat transport across a layer of thickness Az OtoJObot = e Az/=

(15)

where Otop is the conductive heat flow at the top of a layer of thickness Az, Obot is the conductive heat flow at the bottom of the layer, and s = h/pCqz

(16)

where h is the thermal conductivity of a saturated porous medium, P and C are the density and heat capacity of a fluid moving with Darcy velocity qz, and qz is negative for

32

D. DEMING Darcy Velocity (m/yr) 10 .3

101

=

E "-" .

10 2

P,

-

] 'J~'"l

10 .2

,

I Ill,,,

I

+~+

10 0

10 -1

,

,

, ill,,

I

'7'

'1/.

/

101

'" ' " "

9'

_

/

0_

o "6 r--

(3.

10 3

a q02

10 4

I

I

I IIII1[

I

I I Illlt

3. Ratio of conductive heat flow at top of layer (0]) through which groundwater is circulating, to conductive heat flow at bottom of layer (02) as function of Darcy velocity and depth of circulation (after Deming 1994).

Fig.

downward flow. In general, downward movement of groundwater depresses conductive heat flow and the geothermal gradient, while upward movement of groundwater increases conductive heat flow and the geothermal gradient. According to equation (15), the magnitude of a thermal anomaly due to subsurface fluid circulation increases exponentially as the depth and velocity of circulation increase linearly. Assuming typical values of ~ . = 2 . 5 W m -] K -], p = 1000kg m-3, and C = 4200J kg -t K -], the near-surface conductive heat-flow (and geothermal gradient) is reduced by 41% for downward fluid movement through 1 km at a Darcy velocity of 1 cm per year (Fig. 3). If the extent of fluid movement reaches a depth of 5 kin, then conductive heat flow at the surface is reduced by 93% for the same Darcy velocity of 1 cm per year. As a consequence of the exponential dependence upon depth, the background thermal regime in the Earth's crust can be appreciably perturbed by fluid moving at relatively low velocities (10-3-10-2m a-l). Significant disturbances are possible even in sedimentary basins composed of thick sequences of aquitards (permeabilities ~10-14-10-17m2); the presence of factors such as high-permeability aquifers, extensive fracture networks, and conspicuous signs of underground flow (e.g., artesian wells) is not a prerequisite. By applying boundary conditions of fixed temperature at the top and bottom of a layer of

thickness Az, equation (14) can also be solved for T ( z ) . Assuming steady-state conditions (OT/Ot = 0), Bredehoeft & Papadopolous (1965) showed that T ( z ) = To +

ATe(Bz/Az)e8 - 1

1

(17)

where To is the temperature at the top of a layer Az thick, AT is the total temperature change across the layer, z is depth, and B = CpqzAz/~.

(18)

By differentiating equation (17) twice with respect to z, I obtain oZT Oz 2

AT

Bz e (Rz/Az]

(19)

e R - l Az2

thus curvature in a temperature field introduced by fluid movement depends exponentially upon depth. As a consequence, in regions of downward flow it is possible to obtain nearly linear temperature profiles that appear conductive if the depth of the temperature log is less than the total depth of circulation. Under these circumstances it may be possible to obtain temperature logs which appear to be largely linear and conductive leading to the incorrect conclusion that there is no significant fluid flow. Domenico & Palciauskas (1973) considered regional groundwater flow and heat transport in sedimentary basins and derived analytical solutions for head, temperature, and near-surface,

FLUID FLOW AND HEAT TRANSPORT conductive heat flow. They considered steadystate flow in an idealized, two-dimensional flow system with homogeneous and isotropic physical properties, and showed that a relative gauge of the magnitude of convective to conductive heat transport (Peclet number) under these circumstances is given by the dimensionless group Pe =

AhKAzgC 2Lhpr

(20)

where Ah is the total head drop across a basin of length L and depth Az, K is hydraulic conductivity, P is fluid density, C is fluid specific heat, and ~pr is the thermal conductivity of a porous rock (matrix and fluid). By taking the average head gradient to be Ah/L, equation (20) can be written as P~-

qAzpC 2hpr

(21)

where q is the Darcy velocity, and the familiar dimensionless group from equations (15) and (18) appears again. For nominal values of k = 1 0 - 1 5 m z, p=lOOOkgm -3, g = 9 . 8 m s -2, C = 4200 J kg -1K -1, ~kpr "= 2 . 5 W m -1K -1 , Ix = 5 × 10-4kg m -1 s -I, Ah L -1 = 0.01, Az = 5000m, Pe = 1, and it is apparent that thermal regimes in sedimentary basins with average permeabilities of the order of 10-~Sm ~ will be perturbed by groundwater flow. If the average permeability is an order of magnitude lower (10 -I6 m2), Pe ~ 0.1 and the magnitude of advective heat transport will still be significant. Only at average permeabilities of the order of 10-17mZ and lower will fluid flow fail to be an important mechanism for heat transport in sedimentary basins.

Fluid-flow mechanisms and examples The major physical and thermal driving forces for fluid flow in the upper continental crust may be conveniently divided into three groups: (1) topography or gravity, (2) fluid released by sediment compaction, phase changes, or metamorphism, and (3) buoyancy forces resulting from density gradients due to temperature or salinity gradients. It is possible to invoke many other driving forces that do not fit into the above categories. These might include, for example, tectonic compression (Ge & Garven 1992), seismic pumping (Sibson this volume) and thermal expansion or aquathermal pressuring (Domenico & Palciauskas 1979). All of the preceding, however, generally pale in comparison to the major physical and thermal driving forces listed above, with the caveat that it may be possible to find specific examples of temporally

33

or spatially restricted situations in which a driving force dominates that would otherwise be insignificant in most other settings.

Topography The topography of the water table usually mimics the topography of the ground surface. Unlike the solid Earth, however, fluids readily flow downhill, from regions of high potential energy (head) to regions of low potential energy. If precipitation and infiltration in regions of high elevation are sufficient to recharge the water table, a continuous supply of groundwater is available to maintain flow. The time constant (equation 11) which describes the diffusion of transients in regional flow systems is usually orders of magnitude lower than the characteristic time scale over which geological processes which create topographic gradients (e.g., mountain building) operate, and quasi steady-state flow systems result. It is scarcely too strong a statement that topographically driven fluid-flow occurs anywhere where there is topographic relief; it is virtually pervasive throughout the upper continental crust. A characteristic situation is a foreland basin (Fig. 4). In the mountain range and its foothills recharges takes place and fluid moves downward and out into the neighbouring foreland basin in the direction of decreasing hydraulic head. The moving fluid seeks the path of least resistance, or highest hydraulic conductivity. Often, this may be the lowest stratigraphic unit in a sedimentary basin. Through a coincidence of stratigraphy, Cambrian rocks are often permeable sandstone aquifers. The moving fluid then follows a largely horizontal path through the basin, with eventual discharge probably taking place near the distal edge, depending upon the geometry and hydraulic conductivity structure of the basin. Total transport distances of several hundred to a thousand kilometres or more are conceivable. For example, isotopic and trace-element analyses of saline groundwater discharging from natural springs and artesian wells in central Missouri (Banner et al. 1989) show that these fluids most likely originated as meteoric recharge in the Rocky Mountains, c. 1000km to the west (Fig. 5). The existence of regional, topographically-driven groundwater flowsystems has also been documented for several sedimentary basins in western North America by Bredehoeft and his co-workers at the US Geological Survey through measurements of hydraulic head and permeability. These include the Kennedy Basin in South Dakota (Bredehoeft et al. 1983), the Denver Basin in

34

D. DEMING 32-

1"

E

-

.

.

.

.

.

.

.

.

.

.

.

~£££=

.

.

.

.

.

.

.

.

.

.

.

" :?--_~-

..

0-

v

Z

o_

°1 --

.J uJ "3--

I l t l l l

500 km

-4 -5

* t t t t t t t

Fig. 4. Schematic illustration of topographically-driven groundwater flow in a sedimentary basin (after Garven & Freeze 1984a).

RECHARGE

co,o

KANSAS

oo

MISSOURI

~ont

High e ~

Plains aquff

-

- Great Plains _ _ _ -_~-_~ confining system-.7------ S.~ s~lS~'e-~.~'1~"~--- - - G' e~" ~,\~,~S~. ___ _ - -

\

~',~"

;'-"",;.,':".".""

'

;

~ ,,.,. ,,-':" ., ,.-,,~.." ; ....

- . ,. ,.

•

~

0

:

,-

aquifer

system 2 km

FLOW DIRECTION

W Fig. 5. Hypothetical groundwater flow system in Central Great Plains area of the United States extends c. 1000km (after Banner et al. 1989). Colorado (Belitz & Bredehoeft 1988), and the Big Horn Basin in Wyoming (Bredehoeft e t al. 1992). Geophysical field studies in the same and similar settings have found thermal anomalies consistent with heat transport by subsurface fluid movement. These include the western Canadian sedimentary basin (Majorowicz & Jessop 1981; Hitchon 1984; Majorowicz 1989), the Kennedy, Denver and Williston Basins of the Great Plains Province of the central United States (Gosnold 1985, 1990), the Uinta Basin in Utah (Chapman e t al. 1984; Willett & Chapman

1987), the Rheingraben in Germany (Clauser & Villinger 1990; Person & Garven 1992); the Great Artesian Basin in Australia (Cull & Conley 1983), and the North Slope Basin in Alaska (Deming e t al. 1992). The North Slope Basin in Alaska is a well-documented case of thermal anomalies in a sedimentary basin that are apparently due to subsurface fluid flow driven by regional topographic gradients. From 1977 to 1984, the US Geological Survey made repeated temperature measurements in the upper sections ( > 9 0 0 m

FLUID FLOW AND HEAT TRANSPORT

35

100

(a) Heat FIow (mW/m2)

80

.r FCK

KP'

60 40 20 0

- - - - ~ ' EKU --'-,,..,-~t~----____r~--_.~.~ ¢:r,.u

o 3----

~ 1 _

_

-~ ~

SBE

KOL

INI

SON ETK NINFCK NKPATI

JWD DRP

- J ~ l ~ l / . . ~ ..,

¢-

"~ -

..

/ - ,~.- ' - - 2 ° ° ~ -

,,-" _,,, :..." .....:-'I"/.~',;;

8r (O) Temperature(°C) "'"" 91--

.

0 ....

~0km

"""....:'<

............

Fig. 6. Estimates of (a) near-surface heat flow and (b) subsurface temperature in Northern Slope Basin, Alaska. Heat flow estimates are based on relatively shallow (average c. 600 m) equilibrium temperature logs (+ 0.1 ° C) and thermal conductivity measurements. Isotherms represent interpolation of equilibrium temperatures (about + 3-5 ° C) estimated from extrapolation of series of bottom-hole temperatures (after Deming et al. 1992). depth) of 21 boreholes drilled for petroleum exploration in the North Slope Basin (Lachenbruch et al. 1987, 1988). The repetition of these temperature measurements over a period of several years enabled extrapolation to a thermal state that was near equilibrium ( _ 0.1 ° C), and allowed relatively accurate (c. _ 20%) estimates of the near-surface conductive heat flow to be made by combining temperature logs with several hundred thermal conductivity measurements (Deming et al. 1992). Bottom-hole temperature data measured at the bottom of these same wells during drilling and geophysical logging were also collected and used to estimate temperature and thermal gradients at greater depths (Blanchard & Tailleur 1982; Deming et al. 1992). From analysis of this data set (Deming et al. 1992) it became clear that heat flow varies systematically throughout the North Slope Basin in a manner consistent with the probable existence of a regional scale (c. 330 km) groundwater flow system which transports heat by advection from regions of high elevation in the Brooks Range and its foothills to lower elevations on the Arctic coastal plain (Fig. 6). Estimated, near-surface heat flow varies from a low of 27 mW m -2 in the foothills of the Brooks Range, to a high of 90 mWm -2 near the Arctic

coastline to the north. Similarly, at a depth of 3km, the maximum difference in equilibrium formation-temperatures estimated from series of bottom-hole temperatures is 60 ° C, about 20 times the estimated level of noise in these data. Deming et al. (1992) considered a number of hypotheses to explain the geothermal pattern observed in the North Slope Basin. These included lateral changes in radioactive heat generation, thermal transients related to tectonic factors such as rifting or sedimentation, migration of bodies of water over the land surface, conductive heat refraction, and a change in heat flux at the base of the lithosphere due to some unknown cause. However, although each of these mechanisms could, and probably does, contribute to scatter in the data, Deming et al. (1992) concluded that none of them could reasonably explain the observed systematic changes in heat flow and temperature by themselves. In contrast, the hydrological hypothesis is simple and elegant, relying upon a physical mechanism (regional groundwater flow) whose existence has been documented for several sedimentary basins from measurements of hydraulic head and permeability (Bredehoeft et al. 1982). Flow velocities in topographically driven

36

D. DEMING

groundwater flow systems are directly proportional to head gradient and rock permeability, as required by Darcy's Law. Deming et al. (1992) inferred average Darcy velocities in the North Slope Basin, Alaska, as being of the order of 10 cm per year by linking near-surface heat loss and gain to mass flux through the basin. Average Darcy velocities in the Great Artesian Basin in Australia are also c. 10 cm per year, as inferred from dating of radioactive chlorine isotopes (Cathles 1990). The relatively high fluid velocities associated with topographicallydriven flow systems implies that they are efficient transport mechanisms for both heat and mass, and thus likely to play important roles in the concentration of economic resources in the continental crust. Garven & Freeze 1984a, b, Garven 1985, Bethke 1986, Leach & Rowan 1986, Bethke & Marshak 1990, Sverjensky & Garven 1992 and Garven et a1.1993, have invoked topographically-driven fluid flow as a mechanism for the formation of Mississippi Valley-type (MVT) lead zinc deposits. These ore deposits typically form near the margins of sedimentary basins at relatively shallow depths (c. 500-1500 m) (Jackson & Beales 1967; Sverjensky 1986; Anderson & McQueen 1988; Ohle 1991). Studies of fluid inclusions from MVT deposits constrain the temperature of the ore-forming fluids to be in the range c. 100-150 ° C, and the salinity to be c. 20% by weight (Roedder 1977; Leach 1979; Gregg & Shelton 1989; Shelton et al. 1992). The flow velocities needed to reproduce the relatively high temperatures at which MVT deposits apparently formed at shallow depths are generally in the range 1-10 m a -1 (Bethke & Marshak 1990), which is c. 100-1000 times greater than flow velocities observed in sedimentary basins today. Fluids moving updip at Darcy velocities slower than c. 1-10m a -1 simply cool off by conductive heat loss to the surface before they reach the sites of ore deposition. However, the apparent requirement of high flow velocities makes it more difficult to satisfy the salinity constraint. As pointed out by Deming & Nunn (1991), the conundrum is that relatively high flow velocities are needed to satisfy the thermal constraints, but the higher the flow velocities the more quickly the supply of solute is exhausted, and thus the more difficult it is to satisfy the salinity constraint. The difficult problem of insufficient heat has led to the invocation of novel mechanisms to supply the missing energy. Garven et al. (1993) suggested that transient thermal pulses associated with the early phases of topographically driven flow systems could explain the upper

range of fluid temperatures. However, the thermal peaks present in Garven et al.'s (1993) models are mathematical artifacts that reflect an instantaneous emplacement of a topographic driving force. In nature, topography builds gradually over periods of time that are long compared with the characteristic times of basin hydraulic systems, and such transients do not exist (see equation 11 and related discussion). Deming (1992) speculated that perhaps orogenic activity in the uplifted area in which topographic flow systems originate could enhance the fracture permeability of the upper continental crust sufficiently that free convection might be permissible, with resulting large, transient additions of heat into the regional flow system. However, this idea is admittedly speculative. At the present time, not all questions relating to the formation of MVT ore deposits have been satisfactorily answered; each new hypothesis apparently leads to a new set of questions and problems. Understanding the genesis of these ores poses a fascinating problem whose complete resolution will ultimately entail an understanding of both the physical and chemical mechanisms involved. S e d i m e n t compaction Porosity in sedimentary basins decreases with increasing depth. Thus the pore fluids originally present in near-surface rocks have evidently been expelled as these rocks became buried at increasing depth through time. Approximate estimates of average flow rates can be obtained from a simple 'wastebasket' model of fluid expulsion. An imaginary column or wastebasket with cross-sectional area A is constructed from the ground surface to the bottom of a basin. Sediment enters the wastebasket at the top with a velocity R, where R is the sedimentation rate. The total mass accumulation rate is RA, and the total rate at which pore fluid enters the basin is RA+o, where +o is the porosity at zero depth. A maximum estimate of the average fluid velocity due to sediment compaction can be obtained by assuming that all of the pore fluid which enters the basin is expelled and setting +o = 0.50. The average Darcy velocity, q is then q=O.5RA/A -=0.5R

(22)

Thus it is easily shown that the average fluid flow-rate due to compaction in sedimentary basins is proportional to the sedimentation rate. Intracratonic basins, such as the Williston, Michigan, and Illinois basins on the North

FLUID FLOW AND HEAT TRANSPORT American Craton, have average sedimentation rates c. 10m Ma -~, and maximum depths c. 5 km. The maximum average Darcy velocity of expelled pore fluids is thus c. 5 m Ma -~ = 1.58 x 10-~3m s -~, and, according to equation (15), the effect of vertical fluid flow through 5 km would be to increase or decrease the average thermal gradient by c. 0.1%. For higher sedimentation rates c. 100-1000m Ma -~, such as may occur in foreland basins, the maximum average Darcy velocity is c. 500m Ma -~ = 1.58 x 10 -~1 m s -1. For a vertical flow through 10km of sediments, the thermal gradient would be increased or decreased by a factor of 1.3. These are the maximum possible rates of fluid flow and maximum possible depths of circulation. The decrease of porosity with depth, for example, implies that the quantity of pore fluid available for deep circulation would likely be very limited. However, the magnitude of heat transport depends exponentially upon depth of circulation (equation 15). Thus compaction in sedimentary basins is a relatively inefficient mechanism for heat transport. The quantity of fluid involved is very limited, and the rate of expulsion is usually too slow to have significant effects on the thermal regime unless flow is temporally or spatially focused. In a detailed mathematical analysis of compaction-driven flow in intracratonic basins, Bethke (1985) showed that the diffusion equation (equation 9) could be modified by introducing forcing terms related to sediment compaction, Oh at

i

1 {K~72h + - 1 3___~ +Bpg 1-60t OT + +or a--7 - +BpgVzm)

(23)

where o~ is the thermal expansion coefficient for water (c. 5 x 10 -4 K-l), T is temperature, B is the isothermal compressibility of water ( - 4.3 x 10 -~° Pa-1), p is fluid density, g is the acceleration due to gravity (c. 9.8m s-2), Vzm is the settling grain (matrix, not pore fluid) velocity with respect to a point of fixed elevation, and I have simplified Bethke's (1985) original and more generalized formulation which does not assume homogeneous and isotropic permeability. Equation (23) is similar to equation (9) with the exception of three forcing terms on the right side. The first term, 1 -1+

a¢~ Ot' repre-

sents pore collapse, the physical squeezing of fluid out of rock pores due to sediment comaT paction. The second, +¢x --~, is aquathermal

37

pressuring due to fluid expansion upon heating, and the third, ~BpgVzm, represents the loss of potential that occurs as pore fluid moves downward as a result of burial and subsidence. The relative importance of these terms can be compared through an order-of-magnitude calculation. Assuming a sedimentation rate of 50m Ma -~, a geothermal gradient of 25 ° C km -1, an exponential porosity-depth relationship, + = 0.40 x exp[-depth/3 km], ¢x~ 5 x 10 -4 K -1, and a packet of sediment that starts at the surface and is buried to a depth of 1000m in 20Ma, the first term c. 0.01 Ma -~, the second c. 0.0002 Ma -~, and the third c. 0.00007 Ma -1. Thus, for an average sedimentation rate of 50m Ma -~ pore collapse is approximately two orders of magnitude more important for driving fluid flow than aquathermal pressuring or losses of potential energy (head) due to burial. Permeability anisotropy in sedimentary basins has important consequences for compactiondriven fluid flow. Sedimentary basins tend to be organized into flat-lying layers of contrasting permeability. Flow across layers is controlled by the lowest permeability in the sedimentary sequence, while flow parallel to bedding planes is controlled by the highest permeability. Thus permeability on the basin scale tends to be highly anisotropic. Quantitative estimates of basinscale permeability anisotropy are rare because it is not possible to directly measure permeability on a scale of 104-106m. Deming (1993) used idealized computer simulations to estimate the effect of groundwater flow in the North Slope Basin, Alaska, on the near-surface conductive heat flow and found that most model simulations which successfully reproduced the heat-flow data had basin-scale permeability anisotropies in the range kx/kz ~ 100-1000. An important consequence of permeability anisotropy is that compaction-driven flow in sedimentary basins is inherently two- or three-dimensional. Onedimensional compaction simulations which assume vertical flow (e.g. equation 22) may be used to obtain order-of-magnitude estimates of average mass flux rates, but actual flow is at least two-dimensional. Pore fluids expelled from compacting shales are likely to travel vertically until they encounter an aquifer, and then migrate horizontally to the edge of a basin (Bethke 1985). For example, Bredehoeft et al. (1988) interpreted estimates of fluid pressure from the South Caspian basin to indicate that pore fluids from compacting shales drain into adjacent sands and travel laterally hundreds of kilometers through the sands to points of eventual discharge. Due to the very high anisotropies which are possible in sedimentary

38

D. DEMING 2 .~

basins (kx/kz~- 100-1000), one-dimensional simulations may be inappropriate even if the lateral extent of the basin is orders of magnitude greater than the vertical.

4

E

Convection due to buoyancy forces Free convection, or convection due to buoyancy forces, occurs as a result of density gradients that result largely from thermal expansion or salinity gradients. For the simple case of a one-dimensional, homogeneous porous medium confined between two impermeable boundaries, free convection due to thermal expansion occurs if the Rayleigh number (R) exceeds a critical value (Turcotte & Schubert 1982).

R= °tgpZCkAz2~l

(24)

~X where ot is the fluid coefficient of thermal expansion, g is the acceleration due to gravity, p is fluid density, C is fluid heat capacity, k is permeability of the porous medium, Az is the height or thickness of the convecting cell, -y is the thermal gradient, ix is the fluid dynamicviscosity, and X is the thermal conductivity of the porous medium. If I take the critical Rayleigh number (Rcr) to be 27.1 (Turcotte & Schubert 1982), a = 1 . 0 × 1 0 -3 K -1, g = 9 . 8 m s - : , p = l l 0 0 k g m -3 (for a brine), C = 4200 kg m -3, ix = 5 × 10 -4 Pa s -1 (pure water at 50 ° C), and X = 2.5W m -1 K -~, the permeability needed to initiate free convection can be calculated as a function of thickness (h) and thermal gradient (~,) (Fig. 7). For an upper crustal thickness of h = 10km and ~/= 25 ° C km -1, a permeability of about 3 x 10 -16 m 2 is required. Compilations of data by Brace (1984) and Clauser (1992) show that average permeabilities of crystalline rocks in the continental crust on the kilometer scale are of the order 10-15m2; thus convection is apparently feasible over a representative range of thickness and thermal gradients. However, this analysis does not consider the increase of fluid density with depth that occurs as salinity increases. Data from sedimentary basins show that salinity increases with increasing depth, although the rate of increase tends to decrease as depth increases (Hanor 1979). Similarly, salinity data from relatively shallow boreholes (c. 1600m, maximum depth) drilled in crystalline rocks of the Canadian Shield found that total dissolved solids content increased exponentially with increasing depth to an average value c. 150mg 1-1 at 1500m depth (Frape & Fritz 1987). This salinity gradient, however,

Go w

o_ "1-12 14 10

/ 20

30

40

THERMAL GRADIENT (°C/km) Fig. 7. Log~0permeability (m 2) needed for free convection to occur in a porous medium as a function of thermal gradient and cell thickness (after Deming 1992). must decrease at greater depths, because in situ fluids would soon become saturated. For example, extrapolating the rate of increase observed from 0-1500 m to a depth of 2000 m yields a total dissolved solids c. 400mg 1-1. In the presence of a flat salinity gradient it is conceivable that the geothermal gradient may be sufficiently high so as to lead to a density inversion and convective overturn if the average crustal permeability is high enough. If convective overturns occur on a crustal scale, they must be temporally rare. Lachenbruch & Sass (1977) showed that the sudden onset of convection in a formerly quiescent layer dramatically cools the area through which fluid is circulating and also conductively mines heat from below the convection cell. The cooling continues until insufficient heat is available to drive the convective process, or the background flux of heat from the Earth is sufficient to maintain a quasi steady-state circulation system (as it does in geothermal areas). Once a convection cell shuts down, it may take tens of millions of years for the lithosphere to regain lost heat as the thermal recharge must occur conductively. The inherently catastrophic nature of convective episodes has led some authors to speculate that free convection in the crust may be responsible for unexplained phenomena that apparently require large amounts or sudden releases of heat. Deming (1992) suggested that episodic releases of heat from free convection triggered by orogeny may have played a role in the formation of Mississippi Valley-type leadzinc ores in the North American midcontinent. Nunn (1992) speculated that what appears to be

FLUID FLOW AND HEAT TRANSPORT episodic thermal subsidence recorded in the stratigraphy of intracratonic sedimentary basins (e.g. the Michigan Basin) could be due to catastrophic releases of heat from recurrent episodes of free convection. Rayleigh-type analyses (see above) assume convection between two plates which are exactly parallel to each other and perpendicular to the gravity vector. However, in many geologic situations these assumptions are violated (Criss & Hofmeister 1991). In porous media that are tilted with respect to the gravity vector by more than about c. 5°, convection always occurs, although the magnitude of heat transport may be of the same order as that due to conduction. Criss & Hofmeister (1991) suggest that crustal intrusions and subduction zones are examples of geological situations where convective circulation may be common although its presence would not be indicated by a Rayleigh-type analysis. Bethke (1989) pointed out that the full form of Darcy's Law (equation 2) implies that vertical and lateral flow must take place anytime a non-zero lateral fluid-density gradient exists. It is thus evident that density-driven flow is probably both pervasive and perpetual throughout the upper continental crust, although its magnitude may be small in comparison to topographically-driven flow. The cumulative effect of low rates of mass transport over geological time may, as Bethke (1989) suggested, go a long way towards explaining diagenetic alteration in basin sediments. However, fluid motion probably does not occur if a fluid-density gradient falls below a critical value; there must be a lower limit to the physical applicability of Darcy's Law. That is, there must be a point at which the inertial and frictional forces which are usually neglected are large enough to prevent any fluid movement. Exactly where this point occurs is unknown at the present time, and more work in this area is needed. Conclusion

The upper 10-15 km of the continental crust is ubiquitously saturated with aqueous brines and is sufficiently permeable so as to allow for the circulation of these fluids. Fluid circulation in the upper crust is both continuous and pervasive due to topography and fluid-density gradients. The old conception of the continental crust as an unchanging body of solid rock should be replaced by a paradigm that recognizes the continental crust as a two-component system; a solid framework which continuously evolves

39

through thermal, chemical, and mechanical interaction with crustal fluids.

References

ANDERSON,G.M. & MACQUEEN,R.W. 1988. Mississippi Valley-type Lead-Zinc Deposits. In: ROBERTS, R.G. & SHEAHAN,P.A. (eds) Ore Deposit Models. Geoscience, Canada, Reprint Series, 3, 79-90. BANNER, J.L., WASSERBURG,G.J., DoasoN, P.F., CARPENTER,A.B. & MOORE,C.H. 1989. Isotopic and trace element constraints on the origin and evolution of saline groundwaters from central Missouri. Geochimica et Cosmochimica Acta, 53, 383-398. BEAR, J. 1972. Dynamics of fluids in porous media. Dover (edition 1988), New York, 764. BEL1TZ,K. & BREDEHOEFT,J.D. 1988. Hydrodynamics of Denver Basin: explanation of subnormal fluid pressures. American Association of Petroleum Geologists Bulletin, 72, 1334-1359. & 1990. Role of confining layers in controlling large-scale regional ground-water flOW. In: NEUMAN,S.P. & NERETNIEKS,I. (eds) Hydrogeology of low-permeability environmentshydrogeology, selected papers, 2. Verlag-Heinz, Hannover, Germany, 7-17. BETHKE, C.M. 1985. A numerical model of compaction-driven groundwater flow and heat transfer and its application to the paleohydrology of intracratonic basins. Journal of Geophysical Research, 90,6817--6828. 1986. Hydrologic constraints on the genesis of the Upper Mississippi valley district from Illinois Basin brines. Economic Geology, 81,233-249. 1989. Modeling subsurface flow in sedimentary basins. Geologische Rundschau, 78, 129-154. & MARSHAK, S. 1990. Brine migrations across North America - the plate tectonics of groundwater. Annual Review of Earth and Planetary Sciences, 18,287-315. BLANCHARD,D.C. & TAILLEUR,I.L. 1982. Temperatures and interval geothermal-gradient determinations from wells in National Petroleum Reserve in Alaska. US Geological Survey Open File Report, 82-391. BRACE, W.F. 1980. Permeability of crystalline and argillaceous rocks. International Journal of Rock Mechanics, Mineral Science and Geomechanical Abstracts, 17,241-251. - 1984. Permeability of crystalline rocks: New in situ measurements. Journal of Geophysical Research, 89, 4327-4330. BREDEHOEFT,J.D. & PAPADOPULOS,I.S 1965. Rates of vertical groundwater movement estimated from the Earth's thermal profile. Water Resources Research, l, 325-328. , BACK, W. & HANSHAW, B.B. 1982. Regional ground-water flow concepts in the United States: historical perspective. In: NARASIMHAN,T.N. (ed.) Recent trends in hydrogeology. Geological Society of America Special Papers, 189,297-316.

40 --,

D. DEMING BELITZ, K. & SHARP-HANSEN, S. 1992. The hydrodynamics of the Big Horn Basin: a study of the role of faults. American Association of Petroleum Geologists Bulletin, 76, 530-546. , DJEVANSHIR,R.D. & BELITZ,K.R. 1988. Lateral fluid flow in a compacting sand-shale sequence: South Caspian Basin. American Association of Petroleum Geologists Bulletin, 72,416-424. , NEUZIL, C.E. 8,~ MILLY, P.C.D. 1983. Regional

Flow in the Dakota Aquifer: a study of the role of confining layers. US Geological Survey WaterSupply Paper 2237. CATHLES, L.M. 1990. Scales and effects of fluid flow in the upper crust. Science, 248,323-329. CHAPMAN,D.S., KEHO, T.H., BAUER, M.S. & PICARD, M.D. 1984. Heat flow in the Uinta basin determined from bottom hole temperature (BHT) data. Geophysics, 49,453-466. CLAUSER, C. 1992. Permeability of crystalline rocks.

American Geophysical Union LOS Transactions, 73,233 & 237-238. & VILLINGER, H. 1990. Analysis of conductive and convective heat transfer in a sedimentary basin, demonstrated for the Rheingraben. Geophysical Journal International, 100, 393-414. CLAYTON, R.N., FRIEDMAN,I., GRAF, D.L., MAYEDA, T.K., MEENTS, W.F. & SHIMP, N.F. 1966. The origin of saline formation waters. I. Isotopic composition. Journal of Geophysical Research, 71, 3869-3882. CRlss, R.E. & HOFMEISTER,A.M. 1991. Application of fluid dynamics principles in tilted permeable media to terrestrial hydrothermal systems. Geophysical Research Letters, 18, 199-202. CULL, J.P. & CONLEY,D. 1983. Geothermal gradients and heat flow in Australian sedimentary basins.

--

BMR Journal of A ustralian Geology and Geophysics, 8,329-337. DARCY, H. 1856. Les Fontaines Publiques de la Ville de Dijon. Dalmont, Paris. DEMING, D. 1992. Catastrophic release of heat and fluid flow in the continental crust. Geology, 20, 83-86. 1993. Regional Permeability Estimates from Investigations of Coupled Heat and Groundwater Flow, North Slope of Alaska. Journal of Geophysical Research, 98, 16,271-16,286. 1994. Overburden rock, temperature and heat flow - essential elements of the petroleum system. In: MAGOON, L.B. & DOW, W.G. (eds) The Petroleum System - From Source to Trap. American Association of Petroleum Geologists, in press. & NUNN, J.A. 1991. Numerical simulations of brine migration by topographically driven recharge. Journal of Geophysical Research, 2485-2499. - - , SASS,J.H., LACHENBRUCH,A.H. & DERITO,R.F. 1992. Heat flow and subsurface temperature as evidence for basin-scale groundwater flow. Geological Society of America Bulletin, 104, 528-542. DICKEY, P.A. 1969. Increasing concentration of subsurface brines with depth. Chemical Geology, 4, 361-370.

-

-

-

-

9 6 ,

DOMENICO, P.A. & PALCIAUSKAS,V.V. 1973. Theoretical analysis of forced convective heat transfer in regional ground-water flow. Geological Society of America Bulletin, 84, 3803-3814. & -1979. Thermal expansion of fluids and fracture initiation in compacting sediments. Geological Society of America Bulletin, 90, 953-979. FRAPE, S.K. & FRITZ, P. 1987. Geochemical trends for groundwaters from the Canadian Shield. In: FRITZ, P. & FRAPE, S.K. (eds) Saline water and gases in crystalline rocks. Geological Association of Canada, Special Papers, 33, 19-38. FREEZE, R.A. & CHERRY, J.A. 1979. Groundwater. Prentice-Hall, Englewood Cliffs, New Jersey, 604. GARVEN, G. 1985. The role of regional fluid flow in the genesis of the Pine Point Deposit, Western Canada Sedimentary Basin. Economic Geology, 8O, 307-324. 1986. The role of regional fluid flow in the genesis of the Pine Point Deposit, Western Canadian Basin - A reply. Economic Geology, 81, 10151020. & FREEZE, R.A. 1984a. Theoretical analysis of the role of groundwater flow in the genesis of stratabound ore deposits, 1, Mathematical and numerical model. American Journal of Science, 284, 1085-1124. -& 1984b. Theoretical analysis of the role of groundwater flow in the genesis of stratabound ore deposits, 2, Quantitative results. American Journal of Science, 284, 1125-1174. --, GE, S., PERSON, M.A. & SVERJENSKY, D.A. 1993. Genesis of stratabound ore deposits in the midcontinent basins of North America. I. The role of regional groundwater flow. American Journal of Science, 293,497-568. GE, S. & GARVEN, G. 1992. Hydromechanical modeling of tectonically driven groundwater flow with application to the Arkoma foreland basin. Journal of Geophysical Research, 97, 9119-9144. GOSNOLD, W.D. 1985. Heat flow and ground water flow in the Great Plains of the United States. Journal of Geodynamics, 4, 247-264. 1990. Heat flow in the Great Plains of the United States. Journal of Geophysical Research, 95, 353-374. GREGG, J.M. & SHELTON, K.L. 1989. Minor- and trace-element distributions in the Bonneterre Dolomite (Cambrian), southeast Missouri: Evidence for possible multiple-basin fluid sources and pathways during lead-zinc mineralization. Geological Society of America Bulletin, 101, 221-230. HANOR, J.S. 1979. Sedimentary genesis of hydrotherreal fluids. In: BARNES, H.L. (ed.) Geochemistry of Hydrothermal Ore Deposits. John Wiley, New York, 137-168. HITCHON, B. 1984. Geothermal gradients, hydrodynamics, and hydrocarbon occurrences, Alberta, Canada. American Association of Petroleum Geologists Bulletin, 68,713-743. HUBBERT, M.K. 1940. The theory of ground-water motion. Journal of Geology, 48,785-944.

-

-

-

-

-

-

FLUID FLOW AND HEAT TRANSPORT 1969. The theory of ground-water motion and related papers. Hafner, New York. JACI~SON, S.A. & BEALES, F.W. 1967. An aspect of sedimentary basin evolution, the concentration of Mississippi Valley-type ores during the late stages of diagenesis. Canadian Petroleum Geology Bulletin, 15,393-433. KHAaAKA, Y.K. 1986. Origin and evolution of water and solutes in sedimentary basins. In: HITCHON, B., BACHU,S. • SAUVEPLANE,M. (eds) Proceedings, Third Canadian~American conference on hydrogeology, hydrogeology of sedimentary basins: application to exploration and exploitation. National Water Well Association, Dublin, Ohio, 173-195. LACHENBRUCH,A.H. & SASS, J.H. 1977. Heat flow in the United States and the thermal regime of the crust. In: HEACOCK,J.G. (ed.) The Earth's crust, its nature and physical properties. American Geophysical Union, Geophysical Monographs, 20,626-675. , , LAWVER, L . A . , BREWER, M.C., MARSHALL, B . V . ,

MUNROE, R . J . ,

KENNEELY,

J.P., JR., GALANIS,S.P., JR. & MOSES, T.H., JR. 1987. Temperature and depth of permafrost on the Alaskan arctic slope. In: TAILLEUR, I. & WEIMER, P. (eds) Alaskan North Slope Geology, 2. Bakersfield, California, Pacific section, Society of Economic Paleontologists and Mineralogists, 545-558. ,

,

,

~

,

1988. Temperature and deptla of permafrost on the Alaskan arctic slope. In: Gavc, G. (ed.) Geology and exploration of the National Petroleum Reserve in Alaska, 1974 to 1982. US Geological Survey Professional Papers, 1399, 645-656. LEACH, D.L. 1979. Temperature and salinity of the fluids responsible for minor occurrences of sphalerite in the Ozark region of Missouri. Economic Geology, 74,931-937. & ROWAN, E.L. 1986. Genetic link between Ouachita foldbelt tectonism and the Mississippi Valley-type lead-zinc deposits of the Ozarks. Geology, 14,931-935. MAJOROWICZ, J.A. 1989. The controversy over the significance of the hydrodynamic effect on heat flow in the Prairies Basin. In: BECK, A.E., GARVEN, G. & STEGENA, L. (eds) Hydrological Regimes and Their Subsurface Thermal Effects. American Geophysical Union, Geophysical Monographs, 47,101-105. & JESSOP, A.M. 1981. Regional heat flow patterns in the western Canadian sedimentary basin. Tectonophysics, 74,209-238. NESBITT, B.E. & MUEHLENBACHS,K. 1989. Origins and movement of fluids during deformation and metamorphism in the Canadian Cordillera. Science, 245,733-736. & -1991. Stable isotopic constraints on the nature of the syntectonic fluid regime of the Canadian Cordillera. Geophysical Research Letters, 18,963-966. NEUZIE, C.E. 1986. Groundwater flow in low-

-

41

permeability environments. Water Resources Research, 22, 1163-1195. NORa'ON, D. 1984. A theory of hydrothermal systems. Annual Reviews of Earth and Planetary Sciences, 12, 155-177. 1990. Pore fluid pressure near magma chambers. In: The role of fluids in crustal processes. National Academy Press, Washington, DC, 42-49. & KNAPP, R. 1977. Transport phenomena in hydrothermal systems: Cooling plutons. American Journal of Science, 277,937-981. NUNN, J.A. 1992. Free thermal convection beneath the Michigan basin: thermal and subsidence effects (abstract). LOS, 73/43 Supplement, 547. NRC (NATIONAL RESEARCH COUNCIL) 1990. The role of fluids in crustal processes. National Academy Press, Washington, DC, 170. NUR, A. & WALDER, J. 1990. Time-dependent hydraulics of the Earth's crust. In: The role of fluids in crustal processes. National Academy Press, Washington, DC, 111-127. OHLE, E.L. 1991. Lead and zinc deposits. In: GLUSKOTER, H.J., RICE, D.D. & TAYLOR, R.B. (eds) Economic Geology. Geological Society of America, The Geology of North America, P-2, 43-62. OLrVER, J. 1986. Fluids expelled tectonically from orogenic belts: their role in hydrocarbon migration and other geologic phenomena. Geology, 14, 99-102. 1992. The spots and stains of plate tectonics. Earth Science Reviews, 32, 77-106. PERSON, M. & GARVEN, G. 1992. Hydrologic constraints on petroleum generation within continental rift basins: theory and application to the Rhine Graben. American Association of Petroleum Geologists Bulletin, 76,468-488. PHILLIPS, S.L., IGBENE, A., FAIR, J.A., OZBEK, H. & TAVANA, M. 1981. A technical handbook for geothermal energy utilization. Lawrence Berkeley Laboratory Report LBL- 12810, 46. RANGANATHAN, V. d(z HANOR, J.S. 1987. A numerical model for the formation of saline waters due to diffusion of dissolved NaC1 in subsiding sedimentary basins with evaporites. Journal of Hydrology, 92, 97-102. ROEDDER, E. 1977. Fluid inclusion studies of ore deposits in the Viburnum trend, southeast Missouri. Economic Geology, 72,474-479. SHELTON, K.L., BAUER, R.M. & GREGG, J.M. 1992. Fluid inclusion studies of regionally extensive epigenetic dolomites, Bonneterre Dolomite (Cambrian), southeast Missouri: Evidence of multiple fluids during dolomitization and leadzinc mineralization. Geological Society of America Bulletin, 104,675-683. SIBSON, R.H. 1994. Crustal stress, faulting, and fluid flow. This volume. SMITH, L. & CHAPMAN, D.S. 1983. On the thermal effects of groundwater flow 1. Regional scale systems. Journal of Geophysical Research, 88, 593-608. SVERJENSKY, D.A. 1986. Genesis of Mississippi -

-

-

-

42

D. DEMING

Valley-type lead-zinc deposits. Annual Reviews of Earth & Planetary Science, 14, 177-199. & GARVEN, G. 1992. Tracing great fluid migrations. Nature, 356, 481-482. TITLEV, S.R. 1990. Evolution and style of fracture permeability in intrusion-centered hydrothermal systems. In: The role of fluids in crustal processes. National Academy Press, Washington, DC, 50-63. TOULOUKIAN,Y.S., SAXENA,S.C. & HESTERMANS,P. 1975. Viscosity, Thermophysical properties of matter. TPRC Data Series, Plenum, New York, 11. TRUESDELL, A.H. & HULSTON, J.R. 1980. Isotopic evidence in environments of geothermal systems. In: FRIZ, P. & FONTES, J. CH. (eds) Handbook of environmental isotope geochemistry. Elsevier, New York, l, 179-226. TURCOTTE, D.L. & SCHUBERT,G. 1982. Geodynamics applications of continuum physics to geological problems. John Wiley, New York.

-

-

WALTHER,J.V. 1990. Fluid dynamics during progressive regional metamorphism. In: The role of fluids in crustal processes. National Academy Press, Washington, DC, 64-71. WHrrE, D.E. 1974. Diverse origins of hydrothermal ore fluids. Economic Geology, 69,954-973. WILLETT,S.D. & CHAPMAN,D.S. 1987. Temperatures, fluid flow, and the thermal history of the Uinta Basin. In: DOLIGEZ,B. (ed.) Migration of Hydrocarbons in Sedimentary Basins. Editions Technip, Paris, 522-533. ZOBACK,M.D., ELDERS,W.A., VAN SCHMUS,W.R. & YOUNKER, L. 1988. The role of continental scientific drilling in modern earth sciences, scientific rationale and plan for the 1990's: a report to the interagency coordinating group for the continental scientific drilling program from the workshop on continental scientific drilling convened at Stanford University. August, 1988 (available from Mark Zoback, Stanford University).

Fluid flow and heat transfer in sedimentary basins A L A N M. J E S S O P 1 & J A C E K A . M A J O R O W I C Z

2

i 333 Silver Ridge Crescent NW, Calgary, Alberta, T3B 3T6 Canada 2 105 Carlson Close, Edmonton, Alberta, T6R 2J8 Canada

Abstract: One of the major assumptions underlying the science of geothermics over the last 100 years has been that heat is transported through the Earth's crust only by conduction, except in the immediate vicinity of active volcanoes. In old shields this assumption may be justifiable, but in sedimentary basins the ability of water to move through permeable aquifers provides a means of heat transport that may under certain conditions be as effective as conduction. The thermal history of a basin depends on its tectonic origin and the circumstances of its development. The dissipation of excess heat associated with basin formation and the transfer of the continuous heat supply from the basement depend on the thermal properties of the strata and their water content. In some basins widespread contrasts of heat flow are observed, both laterally and vertically, suggesting forced convection by water migration. In basins where part of the surface is above sea-level a hydrological drive may be inferred from the topographic surface. In sub-sea basins this drive is not available in the same form. Nevertheless, patterns of heat-flow variation have been observed for which some form of water flow seems to be the most probable explanation. A complete quantitative description of the present thermal state has not yet been developed for any basin. Some two-dimensional numerical models have been developed, but, since flow directions may vary between aquifers, three-dimensionalmodels must be the ultimate objective. There is scope to extend such models to the geological past by using palaeotemperature indicators such as vitrinite reflectance and fission-track measurements. So far we have been able to establish only an understanding of the principles and processes that control the thermal state of basins, so that detailed description of small parts or two-dimensional models may be developed.

U p to about 1975 the escape of heat from the interior of the Earth was often considered to be controlled entirely by conduction, except for those special circumstances where temperatures are high enough to convert the rock to a mobile fluid. Since the observation of hot water vents on the ocean floor, and the detailed exploration of geothermal resources, the assumption of purely conductive heat flow has been shown to be inadequate and it is generally acknowledged that water flow is capable of transporting significant quantities of heat within the Earth's crust. At the other end of the scale, distortion of heat flow has been observed even in areas of old and stable shields (Drury et al. 1984; Lewis & Beck 1977). Sedimentary basins, in which water migration on a large scale had been recognized by the hydrogeologists for many years, are now seen as areas of special geothermal concern, because of the very large quantities of water that are contained in the sedimentary rocks and its potential for movement. Water is a material of high thermal capacity and mobility, capable of severely distorting conductive geothermal regimes. Moving water is also capable of

transporting hydrocarbons and collecting them into pools, and its study is thus relevant to hydrocarbon exploration (Toth 1980).

The geothermal time-scale Thermal processes occur in a span of time that depends on the duration of changes in the external influences, the thickness of the basin or crust in question and the thermal properties within the crust. A thermal time constant of changes controlled by conduction may be derived from the dimensionless time parameter that is common in all analysis of transient conductive problems. This parameter is defined by:

T

=

st/a 2

where s is thermal diffusivity, t is time, and a is the size of the zone or body, which in this context is the depth. This parameter usually appears as an exponent, and if we make it equal to unity, we obtain an exponential time constant given by:

t=

a2]s

From PARNELL,J. (ed.), 1994, Geofluids: Origin, Migrationand Evolution of Fluids in Sedimentary Basins, Geological Society Special Publication No. 78, 43-54.

43

44

A.M. JESSOP & J.A. MAJOROWICZ

which may be generally regarded as the time taken for a transient thermal event to progress to about 63% of completion. Thermal diffusivity may be assumed to be of the order of 1 mm 2 s -1, and it will not usually be less than 0.5mm 2 s -1 or greater than 2 m m 2 s -1. A basin of modest depth of 2 km has a time constant of about 100ka, but if we wish to consider the whole crust, of thickness 30 km, the time constant is 30 Ma. This time constant is relevant to changes by conduction only and is not strictly applicable to systems involving convective heat transfer. This is an inexact indication only, and is no substitute for a complete analysis of any particular situation. Because of the long time required for complete equilibrium, few processes in sedimentary basins are in a steady state. Change is a continual process of deposition, erosion, and diagenesis, and all change is followed by thermal adjustment. However, these thermal response times are short in comparison with geological time, and it may be assumed that in older basins the thermal events associated with initial rifting or subsidence have dissipated. Understanding of the thermal state of a basin involves a time sequence of conditions, not just the present time. In many basins this implies a history of 600 million years or more of continuous change.

The thermal regime In a purely conductive crust heat transfer may be assumed to be vertically uniform, and geothermal gradient varies inversely with the thermal conductivity of the rocks. Provided lateral heat flow is negligible, temperature at any depth below the surface may be calculated as a function of depth z as:

12= Vo + f

(Q/K) dz

where Vo is surface temperature, Q is equilibrium conductive heat flow, which is uniform at all depths, and K is conductivity, which is a function of depth. Where movement of water occurs in fractures or permeable formations such vertically uniform heat flow may not be assumed. Water in a porous rock has an effect on the conductivity, since the conductivity of water is lower than that of almost all rocks. However, the presence of water does not itself affect heat flow, only the movement of water, and even then only under certain circumstances. The equation of heat transfer in a continuous medium may be written as

. 0212 0212 021/' (pC)m'~'/= gm {,"-~-X2+ 7y2 -F "~Z2) Ov Ov Ov -- (pC) f (wx--~-x--t-Wy-~y--~-Wz-~z) + a

(1)

where v is temperature, P is density, C is specific heat, K is conductivity, w is water flow rate, A is rate of generation of heat, subscript m refers to properties of the combined rock and fluid medium, subscript f refers to the fluid only, and dimensional subscripts define the direction of flow. It is here assumed that conductivity is uniform and isotropic, but this is frequently untrue, particularly in shales. The usual source of heat within the medium is radioactivity, but chemical change may also contribute heat and A may be positive or negative. The second brackets contain terms that describe the effect of flow, and each is a multiple of the flow rate and temperature gradient. If one of these quantities is zero in any direction the term vanishes, and since horizontal temperature gradient may usually be taken to be zero, purely horizontal water flow has no effect on the temperature field. Turbulent flow through a horizontal aquifer has the effect of a slight increase in the vertical thermal conductivity, and hence a slight effect on the thermal regime. Such flow may also act as a supply mechanism for zones where there is a vertical component of flow, and a horizontal flow may be the unseen link between two areas of major anomalies of opposite sense. In and around an inclined aquifer the horizontal gradient may not be assumed to be zero and there is a resultant lateral transfer of heat. A more comprehensive summary of the thermal regime has been given by Toth (1980).

Basins and their setting In terms of the Earth as a whole, the sedimentary basins are no more than a thin skin, providing a thermal boundary condition by which heat may be transported laterally. Sedimentary basins exceed 10km in depth only at their deepest, compared with the total Earth radius of about 6400km. Since a large basin, such as the Western Canada Basin, is about 600km wide, the ratio of width to maximum depth may be at least 60 and the ratio of width to average depth may be as much as 200. Each of the many basins in the world has developed from one of a number of possible starting mechanisms, each of which has its own thermal character. Some of the important mechanisms will be described below. However, the reason for the beginning of many basins is

HEAT TRANSFER IN SEDIMENTARY BASINS poorly understood, and there are many individual characteristics that prevent a simple categorisation of basins. The sedimentary strata of each basin are deposited on a subsiding basement that itself has a thermal state progressing with time. A lower boundary condition is imposed by the heat supply from the mantle. Within accumulating sediment the temperature rises towards an equilibrium level in the geothermal gradient. Conditions in the sedimentary strata control the shape of the temperature profile. In general, periods of deposition mean that the temperature of any stratigraphic unit will rise and periods of emergence and erosion mean that temperatures will fall. However, the thermal state of the strata reacts slowly to changes in heat supply or temperature at the base or the surface, and the time needed for equilibrium to be attained may be very long. The dominant mode of heat transfer is normally conduction, but there are other mechanisms that play a significant part at certain stages of the life of a basin, sometimes to the extent of overwhelming the normal conductive process. Volcanic intrusion and salt diapirism can provide very strong local heat transfer over limited periods of time. The presence of highly conductive salt columns may influence the conductive field for as long as they remain in place. Chemical changes of diagenesis and dehydration of clay minerals may provide sources or sinks of heat in the formations in which they occur, normally in young basins. As basins mature the thermal processes tend to become generally less active, but the movement of water, driven mainly by topographic contrast does not depend on age. This can occur wherever and whenever there are continuous aquifer formations through which the water can travel. Further, the direction and flow rate may change, reacting to the shape of the surface as it is changed by erosion or tectonic forces. We can thus divide the consideration of the thermal state of basins into two parts: a thermal stimulus and input from the starting mechanism; and the thermal state inside the basin, controlled by the boundary heat flow, the physical properties and processes within the basin. Heat flow from a stable cratonic basement probably is normally within the range 25 to 55 mW m -2, but in young basins associated with island arcs, continental collisions or rifts the basement heat flow is much higher, and may even be supplemented by volcanic action. Unfortunately, we are generally unable to make observations of thermal state below the veneer of sediments. Except for some measurements of radioactive

45

components, for example by Burwash and Cumming (1976), we have little data. Distortion of heat flow in the impermeable basement by effects within sediments is probably small, but long-term variation of the effective boundary temperature of the basement may be more severe than in exposed shield areas because of the possibility of major changes in groundwater flow patterns. The extent to which heat flow is distorted by hydrodynamic flow in any individual basin depends on several factors, including topographic contrast across the surface of the basin, driving forces for the water, continuity of aquifer formations, supply of water from the surface, strata, or the basement and presence of strong heat sources. Of these the first, topographic contrast across the surface of the basin, is probably the most important. All basins contain some porous and permeable strata, but water moves only if there is sufficient hydraulic potential to drive it. Since the geometry of basins is related, at least partially, to their origin, it may be expected that some types of basin are more prone to hydrodynamic redistribution of heat than others. However, no strict correspondence may be assumed, and each basin must be considered individually. A few of the more common types of basin will be considered.

Intracratonic basins Intracratonic basins surrounded by basement material and having no strong topographic contrast from one side to the other, such as the Paris Basin or the Michigan Basin, may be expected to be relatively free of hydrodynamic redistribution of heat. These two examples are of regular and symmetrical shape, and there is no apparent source of external drive for water in the strata. Housse & Maget (1976) have produced maps of principal reservoirs within the Paris Basin, with depth to the upper surface, thickness, temperature, water chemistry and transmissivity for each. Figure 1 shows the temperature of the Dogger Formation and the depth to its top. The generally parallel contours of temperature and depth indicate a uniform temperature gradient, with little distortion due to lateral heat transport. Similar maps of the other reservoirs also show reasonably uniform gradient, indicating a generally uniform heat flow. The Heat Flow Map of Europe (Cermak & Hurtig 1979) shows moderately high heat flow of 80 mW m -2 on the north and south flanks of the basin, with a shallow trough below 70 mW m -2 running east to west across the central part. This pattern does

46

A.M. JESSOP & J.A. MAJOROWICZ

OO

2OE

\'~

:

4°E

I

50 k

--m

6°E

l

I'

~./- -, - ~ ,

O. Amiens

'~; i)

-..

F:

! i

5 " 0 ° ~

\..

fo<,

">0o \ ; '.

x

BELGIUM

__

'

\

I I

~Paris

i

(

t ,

/

~ J I'~'--

""

/ i 4J

/

1

~Orleans

/ \

,,,

'~

,,,

\'..

)

,..-5oo,--'/

K - - ~ - ' - - - - 4 0 °~

/

z ,

/,

-- A°

\

Temperature (°C) FRANCE

~ 3 0 o ~

Depth Contours --500m--Fig. 1. Temperature of the Dogger Formation of the Paris Basin, shown by continuous line contours, with the depth to the top of the formation shown by dashed contours. Redrawn from Housse & Maget (1976). not correspond to any major topographicallycontrolled recharge or discharge zones, and the variation in heat flow may be substantially derived from the basement. A similar situation has been reported by Speece et al. (1985) for the Michigan Basin. Temperature profiles have been calculated for each well, on the assumption of uniform conductivity throughout each individual formation and a uniform basement heat flow over the whole area of the basin. Comparison of theoretical temperature profiles with observations where available suggests that effects of heat transfer mechanisms other than conduction are minor and produce no more than small local perturbations. Residuals between observed and calculated bottom-hole temperatures show a definite regional pattern, which may be interpreted as a relatively low heat flow in the areas

under which the basement is composed of mafic Keweenawan lavas, as revealed by gravity mapping. The Williston Basin shows definite thermal effects that are believed to be caused by water flow (Majorowicz et al. 1986b). This intracratonic basin has been overlain by the sediments of a foreland basin and it has become partially dependent on these new surroundings. Water flow and hence heat flow within the Williston basin are now influenced strongly controlled by the hydrological and thermal regimes of a much larger sedimentary body.

Foreland basins A foreland basin has a zone of compression and uplift on one side, which provides the source of the material with which the basin is filled and

4?

HEAT TRANSFER IN SEDIMENTARY BASINS 120°W I

60 ~

117 ° I

113' I

110 ~ 1

106 ° I

102°W 60 °

58 ~

MAN. Ft.

58 °

Vermilion

B.C. ALTA. 56 ~

Grande

! ;~\°,.,

Pr i

SASK.

.

O

54:

56 °

°Fort MacMurray

Peace River

54 ° OCold

Lake

'~.Lloydminster

H i n t o n:~:i:!i:i:

Nipawin O o Prince Albert

North Battleford

52 ~

52 °

~XQ [mW/m2] ~

o

~

>20

50 °

50 ~

Lethbridge

<-20

48°N

48~N

I 120°W

I 117

I

I

113

110

I 106 ~

I 102°W

Fig. 2. Observed difference in average heat flow above and below the Palaeozoic-Mesozoic boundary beneath the Canadian Prairies. AQ is the heat flow above the boundary minus the heat flow below. Redrawn from Majorowicz et al. (1986b).

continues to provide a source of surface water at a high elevation. It is very probable that this water penetrates fractures and that the area of high elevation constitutes a strong recharge zone. Low land on the far side of the basin provides the discharge zone. Examples that have been studied include the Western Canadian Basin, the Aquitaine Basin, the Swiss Molasse Basin and the Uinta Basin. Majorowicz et al. (1985, 1986a, b) have described the thermal regime of the sediments below the Canadian prairies. They have shown that the heat flow in the basin is not vertically uniform and not controlled entirely by conduction. A significant difference is observed between the heat flows, calculated by averages of squares of 30 km by 30 km, above and below the Palaeozoic-Mesozoic boundary. This boundary was chosen as a convenient dividing line for practical purposes, and it is not implied that the change in heat flow is exactly coincident with that boundary, or that the change is abrupt. Figure 2 shows the differences observed in the Canadian provinces from British Columbia to

Manitoba. The variation shows a clear trend of increasing heat flow with depth in the deeper, southwestern parts of the basin and decreasing heat flow with depth in the shallower northeastern parts. The line of vertically uniform heat flow follows a sinuous course from northeast to southwest. Since the heat flow shown here is derived from statistical averages over 30km squares, this general picture is a gross simplification of the detailed situation, and detailed measurement at any single site may not always agree exactly. In Manitoba a strong anomaly of high heat flow is observed above the Palaeozoic surface and a negative anomaly appears below. This is the classic signature of a strong upward water flow, and in this area, on the flank of the Williston Basin, water flow has an upward component provided that it is flowing toward the edge of the basin in a northeasterly direction. This anomaly is open to the northeast, since there are no data in the shallowest sediments. The contrast in heat flow is as great as 40 to 90roW m -2 below and above the boundary;

48

A.M. JESSOP & J.A. MAJOROWICZ

more than a doubling of the heat supply from the basement. The dip of the basement in this area is about 0.0063 or about 1 in 160. In an approximate analysis, it is assumed that the system is one of steady state, the differential with respect to time in Equation 1 is set to zero, horizontal gradients are neglected, and the equation becomes:

10

30 I

20 i

TEMP

°C

2p

40 I

4O

i

i

6O i

TEMP. GRADIENT mK/m 209

~ •

~

KMR

~.

KUC 60C

K

021;

KLC

012

m - ~ = (~C) , wz ~z

KM 80(i

which is equivalent to: g = goexp(-(pC)fw~z/Km)

-.

tj- ] ~

JV

__.

=

(2)

where g is geothermal gradient and subscript 0 refers to the elevation z = 0. The total depth of the sediments is a little over 1000m. If it is assumed that the change in heat flow takes place in 50% of this depth and that conductivity is 2 W mK -~, the above heat-flow contrast implies geothermal gradients of 45 and 20mK m -1, and substitution into Equation 2 gives a vertical water flow rate of 0.8 nm s-1 o r 24 mm per year. If we assume that this flow is the vertical component of flow parallel to the strata, the implied total flow rate is 120nm s -1 or 4 m per year. This is a high flow rate, but it is not impossible in formations becoming thinner in the direction of flow. To the west, at Regina, a well drilled as a water production well but never used for its intended purpose has provided an unusual opportunity for long-term observation of temperature in an undisturbed setting (Jessop & Vigrass 1989). The temperature and temperature gradient profiles from this well are shown in Fig. 3. The heat flow profile, derived from these temperatures and associated conductivity measurements, shows two distinct sections. In the Upper Clastic Unit, of Jurassic and Cretaceous age, heat flow decreases more or less uniformly from 75 m W m -2 at 300 m to 51 mW m -2 at about 950 m. This has been interpreted as indicating an upward component of water flow in this zone of 0.27 nm s -1 or 8 mm per year, using methods of calculation by Bredehoeft & Papadopulos (1965) and Mansure & Reiter (1979). Because of the presence of some sands, particularly the Manville Formation (710-800m), the continuity of this upward component is not assured, but the trend in heat flow seems to cross these aquifers. This analysis tells us nothing about a lateral component of flow, but a model study of the thermal response of a brineinjection well near Regina (Jessop 1987) provides an estimate of lateral flow in the Manville Formation of 76nm s -~ or 2.4m per year. The

I

DD i~iii~iiill

~)

~

"1

\

I mK/m

I

20

I

°C

I ~o

Shale

~

Granite

Sandstone

~

Limestone

~o ~

~

Dolomite

Evaporite Anhydrite

Fig. 3. Temperatures in the well at Regina, Saskatchewan, with calculated tempel .ure gradient and the simplified lithological section. Changes in gradient are mostly related to contrasts in conductivity. Gradient is calculated over intervals of 25 m, but calculation over intervals of 10 m is shown by the dashed line near depths of 1100 m. The very low gradient at 2050m to 2180m is caused by flow of water in the open hole between aquifers of different hydraulic potentials.

lateral flow in the shales cannot be expected to be more than 10% of this value, which leaves it more than an order of magnitude greater than the inferred vertical component. The hypothesis that the observed distortions of the conductive thermal regime are caused by hydrological flow is not universally approved. It has been disputed by Bachu (1988) on the basis of hydrological properties of the aquifers and by Bachu & Burwash (1991) on the basis of distribution of heat flow from the Precambrian basement. Majorowicz & Jessop (1993) have

HEAT TRANSFER IN SEDIMENTARY BASINS pointed out that uncertainties in measurement and modelling of thermal conductivity may account for part of the large vertical variations in apparent heat flow, but no interpretation has been offered that is able to account for the variations in their entirety. Chapman et al. (1984) have described a suite of heat-flow estimates in the Uinta Basin, derived from bottom hole temperatures and assessments of conductivity from measured data of samples from each formation. The heat flow shows a definite trend of increase from 40 to 65 mW m -2 from north to south. This has been interpreted as the result of water flow from the Uinta Mountains to the north of the basin, flowing southward down steeply-dipping formations and escaping upwards towards the south flank of the basin. Unfortunately, the data from this basin are concentrated in the northern part and the whole thermal picture is not available. The basin to the north of the Caucasus shows both high and low areas of heat flow, the low showing some similarity with the foothills of Alberta, and the high similar in appearance to the high of northeastern British Columbia. Sukharev et al. (1964) have explained the contrasts in heat flow and temperature at specified depths in terms of conductivity contrasts, although flow of hot water is mentioned briefly. In the light of present thought on the influence of water flow, it seems probable that conductivity is not the only controlling factor and that hydrodynamic control is much more important than indicated. The Aquitaine Basin shows heat flow increasing towards the shallower parts, furthest from the hydraulic drive of the Pyrenees (Gable 1979). Without more detailed evidence this cannot be separated with confidence from the high heat flow over the west side of the Central Massif. Scbegg (1992) has presented data of vitrinite reflectance from the Swiss Molasse Basin. The level of coalification is too high for reasonable assumptions of burial and paleogeothermal gradients. Since the present thermal regime is believed to be influenced by ground-water flow from the elevated Alps, it is hypothesized that maturity of organic matter has been influenced over geological time by a similar effect. Rybach (1984) has shown that in the Swiss Molasse Basin the present geothermal gradient gradually decreases from 35 mK m -1 in the north to less than 25 mK m -1 in the south, at the border of the Alps, a situation similar to that of the Alberta foreland basin in Canada. Darcy velocities of 5 mm a -1 lead to a calculated isotherm distribution in the subsurface that matches the present temperature field.

49

Rift basins and passive margins Favre & Stampfli (1992) have described a sequence of rifting that has three phases. In the first phase crustal thinning occurs over a wide area and multiple lateral graben systems develop. In the second phase extension is concentrated in a central rift zone and the lateral grabens evolve into rift-rim basins. In the third phase sea-floor spreading commences and new oceanic lithosphere forms between the separated continental blocks. The first phase is normally accompanied by uplift over a broad area as a result of thermal buoyancy. An example is the East African Rift. In the second phase extension is concentrated in a central rift zone, subsidence of the margins occurs and sedimentation takes place. An example of this is the Red Sea. In the third phase all remaining excess heat in the margins dissipates and subsidence continues due to both cooling and sediment loading. The separated edges of the two continental blocks recede from each other and are termed 'passive margins'. This is shown in the Atlantic Ocean and the coast of Africa and North America. The rifting process may stop at any point. Examples of rifts that stopped developing are the North Sea, the Rhine Graben, and Baffin Bay. The North Sea is a failed rift: where the action of rifting stopped before sea-floor spreading could commence. Andrews-Speed et al. (1984) have analysed thermal data from the wells of the British sector of the North Sea. They have chosen to present maps of heat flow above and below a level of 1000m below sea-level, as shown in Fig. 4, rather than in terms of any stratigraphic boundary, but the effect is very similar to the results of Maj orowicz et al. (1986a, b) on the Canadian prairies. In the upper 1000 m, heat flow is high in the shallower part of the basin to the southeast, approaching the English coast; whereas below 1000m the heat flow is higher beneath the central part of the basin, with low heat flow adjacent to the coast. The shallow part of the North Sea Basin to the southeast shows a strong similarity to the situation in the Williston Basin, the shallow edge of the Western Canada Basin, and the central part of the basin is similar to the western edge of the Western Canada Basin. However, there is no mountain range to provide hydraulic drive. If the heat-flow anomalies are caused by water migration, some other driving force must be in operation. The Heat Flow Map of Europe (Cermak & Hurtig 1979) shows a high heat flow anomaly off the coast of Lincolnshire and Norfolk, corresponding to the high heat flow

A.M. JESSOP & J.A. MAJOROWICZ

50

::i?ii?, .::..:

\

,° °°

North

North Sea

Sea

"

Z':':

. . . . .

°o

•

.--

.--

-..-,

.......-...-_,

.-::.;;.~

.:ii?i?i?i?i?ii??ii?iii:::::

:: -Z',' . . .

HEAT FLOW .

.

.

-,';.~

°,

".'.,-:.7." ,'

.

BELOW

lO00m

ABOVE

lO00rn

Fig. 4. Estimates of heat flow above and below the 1000 m level in the North Sea Basin. Redrawn from Andrews-Speed et al. (1984). above 1000 m, closely resembling the Manitoba anomaly. The east coast of Canada is a passive margin, resulting from Atlantic rifting. Reiter & Jessop (1985) have used industrial data from a number of wells on the Nova Scotia and Labrador shelves to produce estimates of heat flow. In Fig. 5 the data are shown as indicators of upward or downward water flow. An upwardly increasing heat flow corresponds to an upward water flow, and vice versa. It is assumed here that the heat flow estimate of each section has an error of no more than 10%, and so variations are separated in Fig. 5 into those greater than and less than 10%. The driving mechanism for the water flow has not been explained. A geopressure zone shows a strongly enhanced heat flow by this method of estimation, because of the reduction in effective conductivity in the over-saturated strata, a reduction that is not acknowledged by the use of estimates of conductivity by rocktype. Thus a geopressure zone gives the same thermal signature as a strong downward water flow. Fortunately, the geopressure system tends to give a stronger contrast in heat flow, with the

contrast appearing at the same depth within neighbouring wells, so that there is a reasonable chance of recognising this feature. Figure 5 shows that the indications of upward water flow are grouped generally in the area of the Abenaki sub-basin and extending to the south-west, where the geopressure zones of Sable Island are found. Wells showing downward flow are mainly to the north of this area. As in the North Sea, the reasons for water flow are not known. Correia et al. (1990) have examined in detail the Jeanne d'Arc Basin, a sub-basin of the Labrador Shelf. They have noted a tendency for apparent heat flow to increase with depth in the northern part and to decrease with depth in the southern part of the basin, and they have interpreted this as indicating a southward and upward flow of water extruded from the deep northern part. The number of data and the weak correlations mean that the interpretation is speculative. Deming et al. (1992) have observed a distribution of heat flow on the North Slope of Alaska that cannot be explained by conduction alone. They have accepted that the hypothesis of

H E A T TRANSFER IN SEDIMENTARY BASINS 64*

62 ° I

GULF

Sr

5B °

60 ('

I

I

OF N C E

LA WRE

FOUNDLAN

c

SYDNEY BASIN

ORPHEUS GR,4BEN

~ --~

45 °

H/aH/

_._-I0

~ s

s~8 8 0

.;

yl

.... .--~

/7" ....

i

-

/

~

. . . .

//'/~

--

--43 °

\ \ \ \ \

Base of Slope

~i_\41

o I

I

I

I

I

Fig. 5. Estimated heat flow in wells on the Nova Scotia Shelf, interpreted in terms of indicated vertical direction of water flow. Triangular symbols indicate wells where variations in heat flow imply water flow, pointing up or down to show direction. Solid triangles indicate strong heat flow contrasts and open triangles indicate contrasts that are within the limits of error of the estimate of heat flow. Open circles denote wells showing little or no contrast. X shows wells where geopressure zones are indicated by heat flow estimates. Redrawn from Reiter & Jessop (1985).

52

A.M. JESSOP & J.A. MAJOROWICZ

hydrothermal distortion is the only mechanism that can produce the observed effect. They comment that 'Not only do the thermal data seem to require the existence of ground-water flow, but they also provide substantial insight into the nature of the flow system'. This marginal basin has the Brooks Range, a high and rugged terrain, as the water source and the elevated hydraulic potential to drive the sea-ward flow. Deming (1993) has confirmed the interpretation by means of numerical models on a profile across the basin, and has shown quantitatively that the observed heat flow can be explained by hydrological distortion of the thermal regime.

Anomalous basins The Hungarian Basin has been extensively drilled, both for hydrocarbons and for geothermal water. Strong circulation systems, driven from the surrounding mountains, show high temperatures in the basin and artesian flow in many parts (Toth 1980). Many geothermal wells in the southeastern part of Hungary supply water at temperatures from 70 to 100° C, from the Upper Pannonian beds at depths to 2500 m, fed by circulating groundwater. The groundwater system, powerful as it is, is believed to be incapable of accounting for the general high temperature and excess heat flow found throughout the basin (Horvath et al. 1979). The concept of a shallow asthenosphere with partially molten material, domed as a result of the tectonic processes that have produced the basin and the mountain ranges of southern Europe, has been proposed to account for the anomalous thermal regime. The thermal state of the Hungarian Basin is thus substantially the result of an unusually high heat flow from the basement, a situation that is probably applicable to many young basins.

Contrasting styles of analysis The examples reviewed show a variety of ways of using the available thermal and hydrological data; first, to describe the thermal state of the basin in question and, second, to progress from that to an understanding of the reasons behind that state. Some of these examples are based on large data-sets, for example the papers by Majorowicz et al. (1985, 1986a, b) have been based on over 63 000 data of temperature from over 36000 wells in the provinces of Alberta, Saskatchewan and Manitoba. Even so, the data distribution is uneven, both horizontally and vertically, and statistical methods are essential

to derive a regional picture. Other examples are based on one well only (Jessop & Vigrass 1989). Closely spaced temperature, conductivity and lithological data in one vertical line is capable of providing estimates of vertical water velocity. Smith & Chapman (1983) chose to investigate the generic thermal response in short twodimensional profiles of permeable rocks. They used models of profiles 40km long and 5km deep, with uniform heat supply from below, with various configurations of layered and anisotropic permeability and relying on the assumption of mirror-image repetition at each end. Various systems of water table were used. Results were expressed in terms of heat flow immediately below the water table. It was shown that water flow can have a major effect on the temperature and heat flow within the permeable block. Deming (1993) has confirmed a previous qualitative interpretation (Deming et al. 1992) by a quantitative numerical analysis over a two-dimensional profile. Since water flow in a deep basin with multiple aquifers may not be everywhere in the same direction, such twodimensional models have limitations. However, they are a valuable tool in relatively simple circumstances. Garven & Vigrass (1985) have constructed models representing sections of the north flank of the WiUiston Basin. This type of modelling inevitably demands a compromise between the numerical magnitude of the model and the requirements for adequate representation of the properties concerned. They also have shown that convective disturbance of the temperature field occurs, with some indication of the importance of contrasts in permeability and the role of impermeable layers such as evaporites. The numerical techniques to simulate parts of a sedimentary basin, with all the necessary thermal and mechanical properties, are available. Schegg (1992) and Majorowicz & Jessop (1981) have used coalification data to derive indications of the palaeogeothermal regime. These are valuable supporting data, and they refer to a thermal history. Other indicators of thermal history, such as fission-track measurements of palaeotemperature, can provide further insights. The diverse methods used by these writers to approach the problem of combining thermal and hydrological observations show that this aspect of the understanding of mechanisms within basins is still under development. Much remains to be done to acquire all the relevant data and to provide a quantitative interpretation that can be accepted with confidence by all.

HEAT TRANSFER IN SEDIMENTARY BASINS

Conclusions

53

& BURWASH,R.A 1991. Regional-scale analysis of the geothermal regime in the Western Canada The science of geothermics has changed over the Sedimentary Basin. Geothermics, 20, 387-407. last 20 years to include a major association with BURWASH, R.A. & CUMMING, G.L. 1976. Uranium hydrogeology. The two are now inextricably and thorium in the Precambrian basement of western Canada. 1. Abundance and distribution. linked. The old concept of conductive heat flow, Canadian Journal of Earth Sciences, 13,284-293. which used to be taken as the norm, may now be CERMAK, V. ¢~ HURTIG, E. 1979. Heat Flow Map of accepted as valid only where it can be demonEurope. In: CERMAK, V. & RYBACH, L. (eds) strated that hydrodynamic effects are negligible. Terrestrial heat flow in Europe. Springer-Verlag. There is now sufficient d o c u m e n t e d evidence of CHAPMAN,D.S., KEHO,T.H., BAUER,M.S. 8z PICARD, geothermal distortion that appears to be caused M.D. 1984. Heat flow in the Uinta Basin by water m o v e m e n t to establish within reason determined from bottom hole temperature that such an effect is real in sedimentary basins. (BHT) data. Geophysics, 49,453-466. The observations agree with the model analyses CORREIA, A., JONES, F.W. & FRICKER, A. 1990. Terrestrial heat-flow density estimates for the in indicating that water can and does move Jeanne d'Arc Basin, offshore eastern Canada. significant amounts of heat within the permeable Geophysics, 55, 1625-1633. strata of basins. The full details of the present thermal regime DEMING,D. 1993. Estimates of permeability on a basin scale from observations of terrestrial heat flow in of any one basin have not been presented. A full a sedimentary basin, North Slope of Alaska. In: description and explanation of hydrothermal PARNELL, J., RUEEELL, A.H. & MOLES, N.R. effects depends on a more comprehensive data (eds) Geofluids '93 Torquay, UK, 4-7 May 1993, set than is generally available. Some factors, 88-91. such as permeability, n e e d improved measure- - - , SASS, J.H., LACHENBRUCH, A.H. & Rrro, R.F.DE. 1992. Heat flow and surface temperament techniques, but much remains to be done ture as evidence for basin-scale ground-water in linking the implications of water flow to be flow, North Slope of Alaska. Geological Society derived from data now available of the hydraulic of America Bulletin, 104,528-542. potential field, the temperature field, and DRURY, M.J., JESSOP, A.M. & LEWIS, T.J. 1984. The chemical content of formation waters. In many detection of groundwater flow by precise tembasins this means three-dimensional analysis, perature measurements in boreholes. Geothersince in any one vertical column water in mics, 13,163-174. different aquifers may be moving in significantly FAVRE, P. & STAMPFLI, G.M. 1992. From rifting to different directions. passive margin: the examples of the Red Sea, Central Atlantic and Alpine Tethys. TectonoThere is considerable scope in this field of physics, 215, 69-97. research for linking the present to the past. The GABLE, R. 1979. Draft of a geothermal flux map of thermal regime reacts slowly to outside changes, France. In: CERMAK, V. • RYBACH, L. (eds) but compared to times in the geological past, the Terrestrial heat flow in Europe. Springer-Verlag, development of the present regime is relatively 206-217. rapid. Indicators of past temperature, such as GARVEN,G. ,~ VIGRASS,L.W. 1985. Modelling of deep organic maturity and fission-track annealing groundwater flow in Saskatchewan. Earth Physics have a valuable part to play in combination with Branch, Open File 85-17. more conventional geothermal techniques of HORVATH, F., BODRI, L. & OTTLIK, P. 1979. Geothermics of Hungary and the tectonophysics of basin analysis. the Pannonian Basin "Red Spot". In: CERMAK,V. & RYBACH, L. (eds) Terrestrial heat flow in Europe. Springer-Verlag, 206-217. HOUSSE, B. & MAGET, P. 1976. Potentiel GeoReferences thermique du Bassin Parisien. Bureau de Recherche Geologique et Miniere & Elf Aquitaine. ANDREWS-SPEED, C.P., OXBURGH, E.R. & COOPER, B.A. 1984. Temperatures and depth-dependent JEssoP, A.M. 1987. Observation of lateral water flow in an aquifer by thermal logging. Geothermics, 16, heat flow in western North Sea. Bulletin of the 117-126. American Association of Petroleum Geologists, -t~ VIGRASS, L.W. 1989. Geothermal measure68, 1764-1781. ments in a deep well at Regina, Saskatchewan. BACHU, S. 1988. Analysis of heat transfer processes Journal of Volcanology and Geothermal Reand geothermal pattern in the Alberta Basin, search, 37,151-186. Canada. Journal of Geophysical Research, 93, LEWIS, T.J. & BECK, A.E. 1977. Analysis of heat flow 7767-7781. data - detailed observations of many holes in a BREDEHOEFT,J.D. & PAPADOPULOS,1.S. 1965. Rates of small area. Tectonophysies, 41,41-59. vertical ground water movement estimated from MAJOROWlCZ,J.A. & JESSOP,A.M. 1981. Present heat the Earth's thermal profile. Water Resources flow and a preliminary paleogeothermal history of Research, l, 325-328. --

54

A.M. JESSOP & J.A. MAJOROWICZ

the Central Prairies Basin, Canada. Geothermics, 10, 81-93. - & ~ 1993. Relation between basement heat flow and the thermal state of the sedimentary succession in the Alberta Plains. Bulletin of Canadian Petroleum Geology, 41,358-368. , JONES, F.W. & JESSOP, A.M. 1986b. Geothermics of the Williston Basin in Canada in relation to hydrodynamics and hydrocarbon occurrences. Geophysics, 51,767-779. , - - , LAM, H.L. & JESSOP, A.M. 1985. Regional variations of heat flow with depth in Alberta, Canada. Geophysical Journal of the Royal Astronomical Society, 81,479-487. - &~ 1986a. Terrestrial heat flow and geothermal gradients in relation to hydrodynamics in the Alberta Basin, Canada. Journal of Geodynamics, 4, 265-283. MANSURE, A.J. & REITER, M.A. 1979. A vertical groundwater movement correction for heat flow. Journal of Geophysical Research, 84, 3490-3496. REITER, M.A. & JESSOP, A.M. 1985. Estimates of terrestrial heat flow in offshore eastern Canada. Canadian Journal of Earth Sciences, 22, 15031517. RVBACn, L. 1984. Geothermal conditions of the Swiss Molasse Basin and implications for hydrocarbon

potential. Revue de l'Institut Franqais du Petrole, 39, 143-146. SCHECG, R. 1992. Thermal maturity of the Swiss Molasse Basin: indications for paleogeothermal anomalies? Eclogae Geologicae Helvetiae, 8 5 , 745-764. SMITH, L. & CHAPMAN,D.S. 1983. On the thermal effects of groundwater flow. 1. Regional scale systems. Journal of Geophysical Research, 88, 593-608. SPEECE, M.A., BOWEN, T.D., FOLCIK, J.L. & POLLACK,H.N. 1985. Analysis of temperatures in sedimentary basins: the Michigan Basin. Geophysics, 5 0 , 1318-1334. SUKHAREV,G.M., TARANUKHA,Y.K. & VLASOV, S.P. 1964. Geothermal features of Caucasian oil and gas deposits. International Geological Review, 6, 1541-1556. TOTH, J. 1980. Cross-formational gravity-flow of groundwater: a mechanism of the transport and accumulation of petroleum (the generalized hydraulic theory of petroleum migration). In: ROBERTS, W.H. & CORDELL,R.J. (eds) Problems of petroleum migration, American Association of Petroleum Geologists Studies in Geology, 10, 121-167.

The nature of metamorphic fluids and significance for metal exploration G. NEIL PHILLIPS, PATRICK

J. W I L L I A M S

& GEOFFREY

DE JONG

National Key Centre in Economic Geology, James Cook University, 4811, Townsville, Australia Abstract: Fluids are active at all levels in the crust during metamorphism. Devolatilization of many rock packages leads to the evolution of H20, CO2 and/or S as a fluid phase that is commonly of low salinity. The composition of the fluid phase released during devolatilization is controlled by the dominant mineral assemblage in the source region, and the evolution of such fluid is most dramatic where major changes in modal mineralogy occur; e.g. across metamorphic facies boundaries. Evidence for this loss of volatiles during metamorphism is provided by metamorphic assemblages that typically contain less volatile components as grade increases. The metamorphism of mafic rock successions can yield a metamorphic fluid rich in H20, CO2 and S, and close in composition to fluids inferred in the formation of many gold-only deposits. The close affinity between gold and these fluids arises from the stability of the Au-S complexing. Where evaporites are present in the succession, the fluids released during devolatilization can be quite saline. A further potential source of saline fluids in these environments is from crystallization of certain magmas. Saline fluids have a composition favourable for transporting gold and base metals, which is likely to explain the occurrence of gold as a co-product in many deposits where saline fluids are inferred. The contrast between saline and low salinity fluids during metamorphism is not only reflected in ore metals, but also in the associated alteration and element mobility. Only a few elements are mobile in the low salinity fluids, and the alteration mineralogy is strongly influenced by the nature of the host rocks. Saline fluids readily transport many elements, and a similar alteration assemblage can be superimposed on many different precursors.

Overview of crustal fluids during metamorphism Metamorphism is commonly synchronous with the evolution of various fluid types at different levels of the crust. For example, at deeper levels (i.e. brittle-ductile to ductile regimes; Sibson 1991), metamorphic fluids and magmatic fluids can both play significant roles and may even share the same fluid pathways at broadly similar times. As these two fluid types move to higher crustal levels, they can potentially mix with seawater, connate waters and/or meteoric waters. The elevated thermal activity during metamorphism relates to the greater activity of these different fluid types at the various crustal levels. Because of this variety of fluids during metamorphism, we make the distinction here between metamorphic fluids (sensu stricto), and fluids that are simply active during metamorphism. Metamorphic fluids (sensu stricto) are those released by devolatilization of mineral assemblages. In most lithologies, this means dehydration, decarbonation and/or desulphidation to produce H 2 0 + CO2 + H2S bearing

fluids that are generally of low salinity. Increased temperature is the main control on devolatilization reactions, mainly because of the higher entropy of the reaction products (because there is a fluid phase released). If one accepts that different fluid types can share channelways, especially during thermal peaks, some of our approaches to separating fluid types on the basis of isotopic composition are in need of revision. One example is ore deposit studies that use isotopes to determine the source of a particular component in the fluid (e.g. O, H, C, S, Pb, Sr). It is vital to determine whether such a component is an accident of mixing, or a fundamental and essential part of the ore-forming process. This problem exists in the mesothermal environment but is more acute in the uppermost crust (see Fig. 1). It is commonly better to focus on the overall process controlling the ore-forming fluid than to postulate a source on the basis of individual components (although the latter approach has produced very useful results in simple systems). A broader issue is the conceptual separation of diagenetic fluids from metamorphic fluids.

From PARNELL,J. (ed.), 1994, Geofluids: Origin, Migrationand Evolution of Fluidsin Sedimentary Basins, Geological Society Special Publication No. 78, 55-68.

55

56

G.N. PHILLIPS E T A L . I

METEORIC

SEA W A T E R

~!!!~::::!!i!iiii!~

÷ ÷ + ÷ ÷ ÷ ÷ ÷ ÷ ÷ ÷ ÷ ÷ ÷ ÷

:~ ~:~:

:/

:~

+ ÷ + ÷ ÷ ÷ + ÷ ÷ t ÷÷÷÷÷÷ + ÷ ÷ + + + + + ÷ + + + ÷ ÷ ÷ ÷ ÷ ÷ ÷ + + ÷ ÷ + ÷ ÷÷+÷÷++÷÷ ÷÷ ÷÷÷÷÷÷++÷÷÷, ÷+++÷÷+++++. +++÷÷+++++÷+ ÷+++÷++÷÷+++ ÷++÷÷++÷÷+÷÷ ++++÷++÷÷+++

:

: :

:

,

Fig. 1. Cross-sections of the upper few kilometres of the crust showing the differing fluid types characteristic of sedimentary, magmatic and metamorphic domains. Fluid mixing is possible at all depths, but potentially very complex at shallow depths.

This is often inferred to be around 200 ° C mainly because researchers working on problems of diagenesis see this temperature as an upper limit to many of the processes which they concentrate on. At the same time, metamorphic petrologists are taking a greater interest in lower temperature regimes where they see metamorphism (formation of new minerals) occurring around 100°C (Frey 1987), and involving devolatilization reactions to produce metamorphic fluids. Probably more important than these definitions of diagenesis and metamorphism is the need to recognize a significant overlap between the two processes. Of major significance for studies in sedimentary basins is the reduction of porosity. While porosity is significant, diagenetic fluids are likely to dominate. Once porosity becomes negligible, there is minimal capacity to retain connate (diagenetic) fluids, and from then on metamorphic fluids become volumetrically more important. Again, this is a conceptual endmember description of a more complex natural situation. Large-scale fluid migration in the crust has been recognized for several years. Meteoric fluids can enter sedimentary basins and move

hundreds of kilometres as their composition evolves. Similarly, the migration through rock masses of brines in oil fields and sea water at mid-ocean ridges has been well documented. More recently, petrologists have recognized the importance of fluid migration during metamorphism (Fyfe et al. 1978; Etheridge et al. 1983; Yardley 1986; Wood & Walther 1986; Ferry 1988a, b, 1992; Craw 1988; Phillips 1988; Symmes & Ferry 1991; Baumgartner & Ferry 1991; Ferry & Dipple 1991; Valley et al. 1990) and considerable progress has been made on mechanisms of migration (Etheridge et al. 1984; Robert & Brown 1986a; Cox et al. 1987; Thompson 1987; Sibson et al. 1988; Oliver et al. 1990). Significant debate has centred around whether fluids present during metamorphism merely escape in single-pass flow or circulate in large-scale hydrothermal systems. Although there have been theoretical arguments against the latter (e.g. Wood & Walther 1986; Yardley 1986), observation of rocks in several metamorphic terrains has provided indications of convective fluxes that effected significant heat and mass transfer (e.g. Ferry 1986, 1988a, b,

METAMORPHIC FLUIDS AND METALS

a)

b)

57

ore-forming potential of all fluid types (meteoric, seawater, diagenetic, metamorphic, magmatic) has been realized, and the role of many intrusions need not be as the supplier of metals, but might be the source of fluid, the source of heat, or simply a response to the same heat event that caused the fluid generation. In some cases, the emplacement of an intrusion can generate significantly more metamorphic fluid than the magmatic fluid evolved by the pluton itself. The link between fluids and ore deposits in this review reflects the focus of research in the National Key Centre In Economic Geology on the overlap between metamorphism, ore formation, field geology and fluids. This overlap is seen as an area of geoscience that has been somewhat neglected in the past. In particular, we emphasize the contrasting fluids found in metamorphic terrains that contain gold-only, and base metal-rich, deposits.

The metamorphic devolatilization process Fig. 2. Schematic mesoscopic representation of the sites of fluids. (a) The diagenetic stage with high porosity and permeability with excess pore fluid (connate water). (b) The metamorphic stage with low porosity. The pore fluid is in boundary and tensile cracks and the volatiles are locked in the mineral structure. See also Etheridge et al. 1983, 1984. 1992; Oliver & Wall 1987; Rumble et al. 1982; Yardley et al. 1991). These studies suggest that fluid flow occurred during or after peak metamorphism in large hydrothermal systems whose architecture was governed by stratigraphy, regional-scale structures and local variations of stress-induced permeability. Hanson (1992) questioned the methodology of earlier theoretical models, and predicted that fluid circulation is favoured at or after the metamorphic peak when dehydration wanes and permeability is high. Economic geologists have long-recognized the importance of fluids (Lindgren 1913; Barnes 1979), but ore-fluid studies have often developed in isolation from other fluid-related disciplines (such as petroleum geology and structural geology) and taken independent research paths. Considerable merging of these paths is now occurring (e.g. the Geofluids '93 Conference). This isolation of ore-fluid studies from other branches of geology is seen in the role of magmatic fluids in ore-formation. Granites (outcropping, indicated by geophysics, or simply postulated) have been linked to almost all ore deposit types. Over the last few decades, the

Thermodynamic modelling of the devolatilization process has been restricted to relatively simple theoretical systems, rather than natural systems. The results of this modelling are not always intuitive, nor predicted by examination of assemblages in rocks. An often overlooked aspect of devolatilization is the important control that host rock composition has on the stability of volatilebearing minerals. For example, a pure calcite rock may be stable well into the granulite facies, whereas calcite in a shale may partake in complex decarbonation reactions even in the greenschist facies. Thus, CO2 is released from marly shales and basalts well before it is released from pure limestones, even though breakdown of the limestone will ultimately yield more CO2 than the other rock types. Similar principles affect sulphides and hydrous minerals and are well illustrated for massive sulphide deposits and their alteration haloes. The pelite system (quartz-micas) would appear to present a relatively simple case of an H20-rich fluid being evolved upon breakdown of muscovite and biotite. This is true for mica-rich rocks, except for the relatively common occurrence of minor sulphides, carbonates or carbon. These last three components are predicted to be involved in early devolatilization reactions. One confirmation of this prediction was made by Ferry (1984) who considered the biotite isograd as a decarbonation front in low grade pelites. Importantly, Ferry (1984) showed that each sample underwent its own particular

58

G.N. PHILLIPS E T A L .

devolatilization reaction depending upon its specific bulk rock composition. Furthermore, the specific reactions involved some desulphidation and dehydration/hydration along with the decarbonation. As such, laboratory experiments involving dehydration (or decarbonation) of a mineral assemblage are only very general approximations for some of the actual reactions occurring in nature. The ultramafic system has been modelled by Will et al. (1990a) who suggest that the resulting fluids are H20-CO2 rich with XCO2 controlled by the mineral assemblages and varying from Xco2 of less than 0.1 to greater than 0.5. Pure carbonate sequences are well treated in the extensive work on CaO-MgO-SiO2-CO2 systematics. In these systems, the resulting fluid is CO2-rich and the temperatures of reaction are predictable from experiments (Greenwood 1975; Slaughter et al. 1975; Walther & Orville 1982). However, the natural situation is usually more complex and less is known of carbonate assemblages with hydrous minerals, and especially sulphides. The metamorphism of sulphide-bearing sequences (or assemblages) is well treated by economic geologists for massive sulphide ores where sulphide introduction pre-dates metamorphism. Mineral stabilities are well constrained by experiments, but the ease of sulphide assemblage resetting upon cooling demands caution when these experimental results are being applied to natural assemblages. Mixed sulphide - silicate (with or without carbonate) assemblages are characteristic of alteration haloes around many ore deposits. These have generally escaped detailed investigation, although there have been a few important studies where sulphide deposits are inferred to have undergone subsequent medium to high grade regional metamorphism (Froese 1969, 1976; Nesbitt & Kelly 1980; Nesbitt 1982, 1986a, b). In an unmineralized situation, Ferry (1981) inferred that the shift from pyrite to pyrrhotite in graphitic schists represented desulphidation that was primarily driven by fluid infiltration rather than mere temperature rise. As a site of diverse lithologies, alteration zones could be exploited more to investigate the relationship between fluids and assemblages during metamorphism. Although there is some difficulty in determining the nature of the early-formed metamorphic fluid in complex rocks and the answers may not be intuitive (e.g. CO2-rich fluids from marly pelites, aqueous fluids from shaly carbonates), it should still be remembered that the volume of this early-formed fluid could be very small, especially if the minor phases are exhausted

early. The overall metamorphic fluid (early plus late) will still reflect the balance of volatilebearing minerals in the original rock type. In nature, mixing between early-formed fluids from high in the pile and later-formed fluids deeper in the pile will tend to cause mixing, especially higher in the pile. One of the closest models to a complex natural system is the Ca-Fe-Mg-Si-A1-H20-CO2 model of mafic and greywacke compositions (Will et al. 1990b). The thermodynamics successfully represent the greenschist to amphiboiite facies transition in the 400-500° C range for rocks with chlorite - calcite - quartz producing amphibole - plagioclase. The most interesting aspects of the modelling are the predicted Xco2 of the fluid in the range 0.1~).4 depending on pressure, and the relatively small temperature interval over which the rocks are predicted to release most of their volatiles (i.e. perhaps 10-20 ° C). The resulting prediction is that of an H20-C02 fluid of low salinity produced at c. 500°C, and with minor HzS. Several researchers have noticed the similarity of this predicted fluid to the fluid compositions that had been recorded in many gold deposits (Smith et al. 1984; Ho et al. 1985; Robert & Kelly 1987; Powell et al. 1991).

Low salinity fluids in gold provinces A majority of the World's gold production has come from 'gold-only' deposits in which base metals are uneconomic and/or negligible. Such types include the Archaean greenstone gold provinces (25000 tonnes Au), Archaean Witwatersrand gold (45 000 tonnes Au), slate belt or turbidite-hosted gold (5000 tonnes Au), and parts of the epithermal gold provinces (e.g. Carlin deposits). Exceptions to this group are deposits that produce gold and base metals such as volcanic-hosted massive sulphides, Cu-Au deposits in ironstones and Cu-Au porphyries: these deposits with associated base metals appear to have different fluid characteristics (Phillips & Powell 1992). The fluids responsible for these gold and gold-base metal deposits have been relatively well studied during the 1980s. Numerous studies of the Archaean greenstone deposits have revealed low salinity, H20-CO2 fluids of elevated temperature (250-400 ° C); similar fluids are known in the Witwatersrand goldfields (Phillips et al. 1990), and in slate belt gold deposits in Victoria and Alaska (Goldfarb et al. 1991). Some gold-bearing fluids in epithermal provinces have these same characteristics,

METAMORPHIC FLUIDS AND METALS particularly those from deposits dominated by gold rather than base metals or silver. An interesting aspect of the gold-bearing fluids is its nearly constant composition despite a wide range of host rocks. This is easily demonstrated in the Kalgoorlie district of Western Australia where similar fluids are recorded from deposits in a dolerite sill (Golden Mile), a metabasalt, a felsic porphyry stock and an ultramafic conglomerate (Ho et al. 1985): the Alleghany District of California demonstrates this same relationship (Bohlke 1989). Thus it appears that the characteristic fluid composition is a parameter inherited prior to entering the depositional site. The pathways for fluid migration have been well studied at a number of major gold deposits where they are either the actual ore material, or spatially related to ore. At the site of deposition, fluid channelways included shear zones, tensile cracks, breccias and contacts, and shear zones subsidiary to large systems are particularly important (Fig. 3). The mode of fluid movement prior to entry into the depositionai site is less certain, but devolatalization is inferred to occur on an individual grain scale before fluids amalgamate and move in discrete channelways. The geometry of the channelways is extremely important in focusing fluids to form major deposits; for example, major shear zones that flatten with depth provide a means of trapping deeper fluids over a wide area and focussing them into and along the shear zone. A feature of the alteration associated with these auriferous fluids is the minimal migration of many major, transition and high field strength elements in all but the most intensely altered zones surrounding the deposits (Phillips & Groves 1983; Robert & Brown 1986b). This means that alteration mineral assemblages are strongly dictated by the composition of original rock types, as demonstrated by variations in the carbonate mineralogy around gold deposits (Kerrich & Fyfe 1981; Phillips & Brown 1987; Bohlke 1989). The host rocks play an important part in the fluid - wallrock interactions linked to gold deposition, especially Fe-rich lithologies in desulphidation reactions, and carbon-bearing rocks in redox reactions (Phillips & Groves 1983; Bohlke 1989). A likely origin for these fluids is metamorphic devolatilization based on their timing and overall composition. In a careful review of fluid origins in the Archaean greenstone gold deposits, Perring et al. (1987) concluded that a metamorphic source for the fluid was most likely although they identified two possible problems relating to detailed fluid timing relative to peak

59

metamorphism and the genesis of amphibolite facies gold deposits. More recently, these problems have been addressed by Powell et al. (1991) who demonstrated that variable timing of devolatilization relative to peak metamorphism was not only explained (by minor erosion), but expected. The other concern of Perring et al. (1987) relating to gold deposits in amphibolite facies is less serious, because rising geotherms during and after gold emplacement would naturally lead to some gold deposits being overprinted by higher grade metamorphism. Other gold deposits in amphibolite facies domains have formed after peak metamorphism. It should also be noted that other researchers see a more complex involvement of magmatic, mantle, meteoric and metamorphic fluids in the gold-bearing systems, but this may not be unexpected in the scenario of the potential fluid mixing postulated earlier (see also Fig. 1): a major challenge is how to decipher these different fluid components, and to decide which are essential for gold mineralization. Even though this very distinctive fluid (low salinity, H20-CO2, reduced sulphur) occurs in many deposits, the fluid source has not been without controversy. Studies focusing on the stable isotope composition of the fluids (H and O isotopes) in these mesothermal environments have identified a meteoric component, and from this, have postulated a meteoric source for the gold-bearing fluid. Other workers have suggested that the meteoric component represents a late influx down the same structures that channelled an earlier auriferous metamorphic fluid into the deposit (Kyser & Kerrich 1991; Goldfarb et ai. 1991). The quartz vein material is thus preserving a different stage of fluid migration than that being preserved in the surrounding wallrock alteration halo. Other studies using Pb and Sr isotopes have identified specific sources for the Pb and Sr (e.g. granite source, Mueller et al. 1991). Unfortunately, the relationship between the Pb and Sr, and Au is far from clear, and it does not immediately follow that the gold or even the fluid that carried the gold came from the same source as the Pb and Sr. One of the major limitations arises from the possibility that the metamorphic fluid may have carried little or no Pb or Sr because of its low salinity. The difficulties in isolating a specific goldbearing fluid arise from the multiple usage of channelways even as deep as the mesothermal environment, and the limitations of stable isotope studies, particularly those that overlook the basic character of the fluids themselves. The difficulties become more complex higher in the

60

G.N. PHILLIPS E T A L .

iiiiiii,i ,il

COUNTRY ROCKI

Fig. 3. Diagramatic pathway for metamorphic fluids from grain-scale devolatilization to major (?listric) shear zones. The fluids can potentially enter second and third order structures and form major gold deposits (partly modelled on Kalgoorlie gold deposit, Phillips 1986; Sigma gold deposit, Robert & Brown 1986a, b; Etheridge et al. 1983). The shear zone (top diagram) has been enlarged for clarity. Scales are only order-of-magnitude.

METAMORPHIC FLUIDS AND METALS

61

LEGEND

°~N~IS,OUN

Elolse Cu (Au'r)

ISA BLOCK

1

I c.-..,...-.

• Mmonan ~ laseMetals+ Au) l ~

Granites

Metesediments -predominantly

altered

Metesediments -l~redomlnantly unaltered

L)

Majo¢Fault Faul!

c c Y - Ck~ncurry c

,tun/Zn

~

EASTERNFOLD BELT'r

LadyLoa'eaa

~

i

MDF -

Mount Dora Fault

•

Deposit

km

,co-~?-y: : :.: :

tl ;y.,n.:,Cu~,. . . . . . . . Mount Isa • Pb-Zn-Ag/Cu

.,+oo.

~,

/!~ii!~!~!:!:!:i:i:i:i:!:!:i:i:i:!~ ~.. j

~7;~,,~.. 0

.o.,,noton Pb-Ag-Zn

O Osborne

20km

Cu-Au

Fig. 4. Map showing the Cloncurry area within the Mt Isa Block of northern Australia, and the Eastern Fold Belt and associated ore deposits south of Cloncurry. Some deposits are potentially very large; others are interesting prospects. The Williams Granite is the large pluton between Kuridala and Pegmont. The Doherty Formation covers the area on the east of the Williams Granite marked as 'metasediments - predominantly altered'. The Maronan Supergroup is to the east of the Doherty Formation and mostly comprises 'metasediments - predominantly unaltered'. crust where mixing is greatly facilitated (e.g. the epithermal environment).

Regional metasomatism around Cloncurry, Northern Australia Low-salinity metamorphic fluids have been widely documented in mineralized and unmineralized settings over the last decade. In contrast, the saline fluids referred to below are less well documented, and their origin is less clear. However, their presence during metamorphism can be demonstrated by timing relationships involving structural features and alteration assemblages. The saline fluids are described here because they illustrate that not all fluids during metamorphism need be of low-salinity character. We believe that it is no accident that mineralization types vary between the low salinity and high salinity examples chosen.

The Cloncurry district (Fig. 4) forms the eastern part of the Mount Isa Inlier, which is one of several large Proterozoic blocks in central Australia. The metamorphic rocks range in grade from upper greenschist to upper amphibolite facies and are intruded by batholithic, late to post-tectonic granitoids. Multiple deformation included a regionally extensive, upright D2 deformation synchronous with peak metamorphism, and a predominantly retrograde D3 deformation. The district has major petrological and economic significance. It is an outstanding example of regional-scale metasomatism with altered rocks exposed over hundreds of square kilometres, and it has an exceptional endowment of significant base metal and gold deposits. An important factor in the development of these features is likely to have been the presence of metasedimentary sequences with large components of evaporites and carbonates that acted

G.N. PHILLIPS ET AL.

62

:.:.:.:.:.:.:.:..

.'.:.:.:.:.~-~'.:.:.:.:.:...~

))))))))):.:.... ):'Z'.e,, ,.e. ")): •. :......:..:::::::... .....:...,,,A .. :.....

. . . . +

......................:.:.:.:.:.::: ~ ::::::::::::::::::::::::::::::::::

~+

+

+

+

+

+

+

+

+

:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.: :....: :..::..::......:.

b6

%

!iiii!iiiiiiiiiiiiiiiiiii

21°30 '

"DINGO

I'

.~t ~

LEGEND +

[

I Alluvium

+

2 km Granites

+ +

Calc-silicate

rocks +

Albite

a c t l n o l i t a rocks, brecclas

Metasediments

*

Fault

t T

Metamorphic layering

(dip/s trlk e)

MARAMUNGEE

+~ +

+ +

D2 fold axis C~

High strain zone

140°50'

+

Fig. 5. District map around Maramungee Creek enlarged from Fig. 4, showing the details of the large scale metasomatism. The Williams Granite is shown on the western margin. The Doherty Formation includes the calc-silicate rocks and breccias" the Maronan Supergroup includes the 'metasediments' marked to the east.

as a source of saline fluids during a protracted series of tectonothermal events. Major batholithic granites were emplaced during the latter stages of this evolution and are also likely to have contributed fluid components.

Nature and scale o f metasomatism Widespread scapolite alteration was described within a large tract of rocks west of Cioncurry by

Edwards & Baker (1959). Studies of an area close to the Mary Kathleen deposit (Fig. 4) by Oliver & Wall (1987) have shown that the rocks there, belonging to a widespread metamorphosed carbonate-evaporite-clastic-metadolerite association known as the Corella Formation, underwent interaction with distinctively saline fluids under conditions close to those of peak metamorphism. This produced scapolite, aibite and large calcite pods (Oliver &

METAMORPHIC FLUIDS AND METALS Wall 1987; Oliver et al. 1990). Calculated fluid:rock ratios were highly variable and controlled by strain variations induced during the generation of the dominant N-S-oriented structures formed in association with peak metamorphism. However it was estimated that the total fluid flux was large compared to the mass of rock, such that the fluid must have included a major externally-derived component (Oliver et al. 1990). This is supported by recent stable isotope work which indicates that the infiltrating fluid was not initially in stable isotope equilibrium with the Corella Formation (N. Oliver, pers. comm. 1993). Extensive syn-metamorphic alteration features are also developed south of Cloncurry (Jaques et al. 1982; Laing 1991; Williams & Phillips 1992; Phillips & Williams 1993) where large volumes of rock were affected by a zoned hydrothermal system. The main, but by no means sole, locus for this activity is a stratigraphically problematical package of rocks that has been referred to as the Doherty Formation (e.g. Blake 1982). Current work suggests that much of the Doherty Formation consists of strongly altered products of essentially the same precursor rocks as constitute the Corella Formation further west (Williams & Phillips 1992). The alteration system is closely related in space to a major regional fault (the Cloncurry Fault, Fig. 4) which has undergone a series of movements under both ductile and brittle conditions. This fault juxtaposed carbonate-evaporitebearing sequences in the west against predominantly clastic sequences to the east ('Maronan Supergroup' of Beardsmore et al. 1988, Fig. 4) early in the deformation history of the region (Loosveld 1989). The alteration system south of Cloncurry is well-exposed close to the trace of the Cloncurry Fault, east of the Williams Batholith (Figs 4 and 5). This is particularly true of a number of east-draining watercourse sections that provide good three-dimensional outcrops and transect the strike of the dominant structures. In the area of Fig. 5 alone the intensely metasomatized zone is over 2km by 15km, and this is only part of a much more extensive altered zone (Fig. 4). The eastern part of the Cloncurry area is composed predominantly of migmatitic paragneisses that constitute the structurally-deepest exposed portion of the Maronan Supergroup. This succession confirms mid to upper amphiboiite facies metamorphism (sillimanite-K feldspar) but is extensively retrogressed to muscovite-bearing gneisses and schists along a thick shear zone adjacent to their contact with the Doherty Formation. This contact is folded

63

with the result that belts of muscovite schist also occur further west. The Doherty Formation contains large amounts of metasomatic breccia together with less altered metadolerite and banded feldspar-rich metasedimentary rocks that generally lack mica but commonly contain calc-silicate minerals. The m e t a s o m a t i z e d z o n e in detail: M a r a m u n g e e Creek section.

The Maramungee Creek section in the eastern Selwyn Range area exposes the contact between the Maronan Supergroup in the east (here expressed as retrograde muscovite schists), and the Doherty Formation to the west (Fig. 6). Through much of this creek section it is possible to see the variability of alteration types that make up the regional-scale metasomatized zone. The effects of post-D3 phases of regional metasomatism are comparatively weak in the Maramungee Creek area, which consequently provides an opportunity to study features developed after the syn-peak metamorphic D2 event. The dominant characteristic of the metasomatism is alkali exchange controlled by structural position and host rock composition. There is a pattern of alteration zone geometry controlled by pre-existing D2 structures, even though the alteration itself is retrograde. In the east, the tectonic contact between the Maronan Supergroup and the Doherty Formation is characterized by metasomatic breccias (quartz + sodic plagioclase + actinolite + magnetite) which overprint muscovite schists. The breccias are themselves heterogeneously deformed, reflecting the evolution of anomalous high strain at this structural position. The main metasomatized zone is several hundred metres thick and comprises breccias and sodic- and calc-silicate rocks with several distinct mineral assemblages (Fig. 6). A late to post D2 age for the metasomatism is inferred from the undeformed nature of the albite-scapolite, and the veins and replacement that destroy $2 in the schist. To the west, granites, microcline-bearing rocks, and tremolite rocks dominate. Shear zones up to hundreds of metres thick form the locus of sodium metasomatism. Their central portions are overprinted by brecciatextured rocks whereas adjacent rocks are cut by complex sheeted vein systems containing the same minerals as found in the breccias. Alteration assemblage

A striking feature of the alteration in the Cloncurry area is the dominance of actinolite -

64

G.N. PHILLIPS

ET AL.

GEOLOGY OF THE MARAMUNGEE CREEK AREA ii!iilililililii!ii~" o o :i::iiii::ii!i!iiiiiii!i~!!!a" o

o o o

~ 0 0o

• ~~ • t

o oo o o o o o

~

o o

o o

.

~

~°,%? • ,%.-~., . . v. .v. v v., ~. o o o • ~ .~ 0°0% • :, ~ ,"~ , .v .v .v v. v v v . ."-~ ,"" ~ . .

,

',vvvv

~

~.! ++ ++ 1

~ , i ~ . .~",I)., ', . . . . . . . .

i

~2

.+,

v),/

~++; "~

o o o OoOo o

:::::: ::: : : ~ . :::: ::::: : : : . ,::- : :::: :::: ~ + ::: ::-

o o o o o o o o o o

oo00 00%0

OoO % o oo 0

0°. *.0

I

,,*"

: :?:~:~::~ ":~::~. :~:: ::~. ~:.::~:.::i~,~.~. ::.':.:: ::

70

~

::i

::~:~l

<<-:--

~-

-

.

,~,+ +

+÷+,

cover sequence

retrograde schist

I

~,vvv

" *

----------

"~

4850oomE

LEGEND

:t°

'i~::~,:~

487000 mE

486~omE

488000 mE

METASOMATIC ROCKS ~

foliated granite, granite sheets [ quartz-feldspar rock

~

metadolerite

~

hb-Ca scap-plag-bi

i --~ "pristine"

] kfs-ab-trem/act-mt ab- act-kfs- mt-ilm-tit kfs-Na o o tz sca -bi- la

~_ less altered ti1

~

Fault Shear zone

• I Breccia. "'v's1-so//Slsheared breccia .". S2

most altered

ab-cpx-trem/act-cc-mt-

~ 0

El-2 metres

1000 1

Fig. 6. Detailed map of the Maramungee Creek section showing relationships between metasomatic rock units. Abbreviations: ab, albite; trem, tremolite; act, actinolite; rot, magnetite; hb, hornblende; Ca-scap, calcium-rich scapolite; qz, quartz; plag, plagioclase; bi, biotite; kfs, K-feldspar; ilm, ilmenite; tit, titanite; Na-scap, sodium-rich scapolite; cpx, clinopyroxene; cc, calcite. The Cloncurry Fault is 1 km to the west of this map. Virtually all the units shown comprise the Doherty Formation.

albite-bearing assemblages over hundreds of square kilometres of metasomatized rocks. The feldspar is An0-8 (rarely up to An14), and the assemblage also contains variable amounts of other minerals, particularly titanite, magnetite, tremolite, diopside, microcline, biotite, scapolite and quartz. The actinolite - albite assemblages are apparently stable in virtually all altered rock types including granite, pelite, marble and mafic intrusive rocks. The proportion of actinolite to albite varies substantially over centimetres, and is not directly a function of rock type. The specific characteristics of this alteration to a large part reflect the mobility of many elements during the metasomatism. The alteration assemblages overprint the peak metamorphic mineral assemblage and, in places, overprint retrograde muscovite schists.

Mobility of elements On the basis of the actinolite - albite assemblage in a large number of rock types, and the wide

range of vein minerals including titanite, magnetite and ilmenite, many elements were mobile during alteration. Unlike the low salinity, gold-only system, the Cloncurry assemblages have few coexisting phases, many mobile components, and high inferred variance. The dominant bulk rock compositional change effected by the alteration varies with rock type, with an overall replacement of K by Na. However, many other elements were also redistributed including rare earth elements, Y and Fe, the latter of which is locally concentrated as magnetite in the breccias. The normally immobile element Ti, which was leached from some altered rocks, has been concentrated in some veins (e.g. as ilmenite and titanite).

Type of fluid The nature of the fluids present during regional metamorphism has not been fully characterized, but much can be gleaned from the mineral

METAMORPHIC FLUIDS AND METALS assemblages and some preliminary fluid inclusion evidence. Both Na ( + C I ) - and Ca-rich scapolites occur in the Maramungee Creek section (Fig. 6). The more Ca-rich varieties occur in weakly altered metasedimentary rocks associated with hornblende, phlogopite, quartz and feldspars and are inferred to have been stable during regional metamorphism. Na (+Cl)-rich scapolite occurs as selvages on albite-rich veins adjacent to shear zones and also as an alteration product of metamorphic plagioclase in metadolerite. The latter occurrences demonstrate high C1 activities in the fluid (cf. Mora & Valley 1989) and in view of their field and petrographic relationships provide an indirect indication that fluids of high salinity were involved in the alkali metasomatism (cf. Oliver & Wall 1987). The saline nature of fluids is further indicated by fluid inclusion data from vein quartz occurrences both within altered schist of the eastern contact zone and from shear zones further west. Three phase (liquid-vapour-salt) inclusions in which the daughter salt has the halite form give entrapment temperatures in the range 400650°C corrected for estimated pressures of 200-300MPa. NaCl-equivalent salinities are in the range 28-32% with some freezing determinations beyond halite saturation. To the immediate west of the Williams Batholith, Beardsmore (1993) has identified a high temperature (500 ° C), oxidizing brine (6570wt% salt) which underwent dilution and mixing during cooling. He implicates this fluid in the genesis of the Cu-Au deposit of Mt Dote, and suggests a late D3 timing for the sulphide deposition. Ore deposits

There are two particular reasons for pursuing a closer link between the Cloncurry fluids described above and mineralization. First, the inferred fluid compositions (high temperature, saline, oxidizing) are ideal for the transport of metals such as Au, Cu, Ag, Pb and Zn (Crerar et al. 1985); and second, there is a regional spatial association between the widespread alteration and numerous ore occurrences. The Cloncurry area has abundant metalliferous deposits (Fig. 4) and for a few years in the early part of the twentieth century was the largest source of copper in the British Commonwealth. Mining activity subsequently waned but recent years have seen a resurgence in interest that led initially to the discovery and development of two relatively small mines at Selwyn (Au-Cu) and Tick Hill (Au). In the last few

65

years, this area has emerged as a world class metal province that rivals the well known resources of the Mount Isa district (Derrick 1992). This success has included recent discoveries of medium to large sized deposits including Osborne (Cu-Au-Co), Ernest Henry (Cu-Au) and Cannington (Pb-Ag-Zn). There are also many other significant deposits that are currently subeconomic or awaiting feasibility studies (Pegmont, Osborne - Trough Tank, Maramungee, Maronan, Eloise, Mt Elliott, Mt Dore, Kuridala). For some of these deposits an origin during metamorphism appears well established (e.g. copper-gold deposits), but for others their origin remains contentious and may be interpreted as metamorphic or pre-metamorphic depending upon the emphasis placed on the various evidence.

Discussion The gold provinces and the Cloncurry district provide contrasting examples of fluids active during metamorphism. Major differences exist in the types of alteration assemblages, the mobility of different elements, the control of host rock on alteration assemblages, and ultimately the ease of recognizing different rock types after intense alteration. The gold-bearing fluids are well established as being synchronous with metamorphism, of low salinity, and dominated by H20-CO2 and some HzS. Such a composition is favourable for transporting gold without base metals, as the soft acid (Au ÷) forms a particularly stable complex with reduced sulphur, and the base metals are more stable with harder bases such as C1- (Crerar et al. 1985; Wood et al.1987). The immobility of many elements during this alteration means that the distribution of several major elements remains a function of original rock type, and thus the host rocks exert a major control on alteration mineral assemblages. One important source for the gold fluids is likely to have been devolatilization of chloritecarbonate-quartz-albite assemblages during amphibolite facies metamorphism, commonly involving mafic rocks. The Cloncurry fluids were also active during metamorphism, but their origin is not so clear. A saline fluid composition is inferred from the mineral assemblages and the related fluid inclusions. High chlorinity favours base metal (and gold) transport in solution, and also facilitates the mobility of many other alkali, transition and even rare earth elements (Crerar et al. 1985). The mobility of many elements during alteration means that a common alteration assemblage is

66

G.N. PHILLIPS ET AL.

superimposed upon many host rock types, rather than the host rock dictating the balance of major elements in any sample. Some care must be exercised in determining the origin of these fluids, and the timing, overall composition and viable source regions are at least as important as the stable isotope composition. The isotopic composition of major fluid components can be used to some effect if they are sufficiently diagnostic of a specific source (commonly they are not), but particular care is required in using the isotopic signature of minor components of the fluid. Small degrees of mixing can easily swamp a system with a minor element if two fluids mix that have highly variable concentrations of that component: such a scenario is likely if a small fraction of saline fluid mixes with a low salinity fluid. An important difference between the two examples is the crustal level at which each is being studied. The gold fluids are predominantly known from studies of material at the site of deposition where the fluid has been well focused: the Cloncurry fluids are being studied mostly away from mineralization where they are pervasive rather than well focused.

over a decade. We are grateful to many mining operations in Western Australia, Canada, Brazil, South Africa and Zimbabwe for access to Archaean greenstone gold deposits, to forty companies with Witwatersrand gold mines, to slate belt gold mines in Alaska and Victoria, and to operators of epithermal deposits in Queensland and Nevada for access to these deposits. The research in the Cloncurry area has received strong support from Aberfoyle, Australian Consolidated Minerals, Battle Mountain (Australia), Billiton, MIM, BHP, Placer Pacific, WMC, the Queensland Tertiary Education Foundation, the Key Centre In Economic Geology and Australian Research Council. Support from the W. C. Lacy Scholarship Fund, and an Overseas Postgraduate Research Scholarship (G.D.J.) considerably facilitated the research at Cloncurry. The Commonwealth Science Council and conference organizers are thanked for facilitating travel by G.N.P. to the Geofluids '93 Conference. Numerous colleagues have been generous with their ideas over the years, including S. Cox, D. Groves, K. Lawrie, G. Morrison, R. Myers, P. Pollard, R. Powell, F. Robert, M. Rubenach, R. Spencer, R. Taylor, V. Wall and T. Will. Drafting by M. Myers and critical reviews of the manuscript by M. Rubenach, G. Dong, F. Robert and N. Oliver are gratefully acknowledged.

References Summary The last decade has seen a rapidly growing awareness of the role of fluids during metamorphism. The mere link between heat and metamorphism guarantees that several different fluid types will be active at the one time, and possibly sharing c o m m o n fluid channelways. It can be difficult to separate fluids of different origin after they mix. The volumes of these fluids can be substantial, and their elevated temperature gives them considerable ore-forming potential. Fluid movement at the metamorphic stage is structurally controlled (rather than permeability-porosity driven as in diagenesis) and utilizes shear zones, tensile cracks and grain boundaries. The main components of some metamorphic fluids are H20, CO2 and H2S with low salinity reflecting the abundance of micas, carbonates and sulphides in many sequences (and paucity of halite). Less common metamorphic fluids can be generated from the devolatilization of carbonate and especially evaporitelbearing successions and the resulting products are a saline fluid, widespread alteration, and the potential to transport and deposit base metals and gold in economic quantities. This review represents a culmination of research on ore deposits and fluids in metamorphic terrains for

BARNES, H.L. (ed.) 1979. Geochemistry ofhydrothermal ore deposits. Wiley, New York. BAUMGARTNER,L.P. &FERRY,J.M. 1991. A model for coupled fluid-flow and mixed-volatile mineral reactions with applications to regional metamorphism. Contributions to Mineralogy and Petrology, 106,273-285. BEARDSMORE, T.J. 1993. The role of metamorphic fluids in the petrogenesis of Mount Dore - style breccia - hosted copper - gold deposits in the eastern Mt Isa Inlier, Queensland. Geological Society of Australia, Abstracts, 35, 54-55. --, NEWBERV, S.P. & LAING, W.P. 1988. The Maronan Supergroup: An inferred early volcanosedimentary rift sequence in the Mount Isa Inlier and its implications for ensialic rifting in the Middle Proterozoic of northwest Queensland. Precambrian Research, 40--41,487-507. BELL, T.H. 1983. Thrusting and duplex formation at Mount Isa, Queensland, Australia. Nature, 304, 493-497. BLAKE, D.H. 1982. A review of the Corella Formation, Mount Isa Inlier. Bureau of Mineral Resources Journal of Australian Geology Geophysics, 7, 113-118. BOHLKE, J.K. 1989. Comparison of metasomatic reactions between a common COz-rich vein fluid and diverse wall rocks: intensive variables, mass transfers, and Au mineralization at AIieghany, California. Economic Geology, 84,291-327. Cox, S.F., ETHERIDGE,M.A. & WALL,V.J. 1987. The role of faults in syntectonic mass transport and the localization of metamorphic vein-type ore deposits. Ore Geology Reviews, 2, 65-86.

METAMORPHIC FLUIDS AND METALS CRAW, D. 1988. Shallow-level metamorphic fluids in a high uplift rate metamorphic belt; Alpine Schist, New Zealand. Journal of Metamorphic Geology, 6, 1-16. CRERAR,D., WOOD, S., BRANTLEY,S. • BOCARSLY,A. 1985. Chemical controls on solubility of oreforming minerals in hydrothermal solutions. Canadian Mineralogist, 23,333-352. DERRICK,G.M. 1992. Brothers in arms: the interaction of geology and geophysics in the Mount Isa Inlier. Exploration Geophysics, 23, 117-22. EDWARDS,A.B. & BAKER,G. 1959. Scapolitization in the Cloncurry district of northwestern Queensland. Journal of the Geological Society of Australia, 1, 1-31. ETHERIDGE,M.A., WALL,V.J. & VERNON,R.H. 1983. The role of fluid phase during regional metamorphism and deformation. Journal of Metamorphic Geology, 1,205-226. --, & Cox, S.F. 1984. High fluid pressures during regional metamorphism and deformation: implications for mass transport and deformation mechanism. Journal of Geophysical Research, 4344--4358. FERRY, J.M. 1981. Petrology of graphitic sulfide-rich schists from south-central Maine: an example of desulfidation during prograde metamorphism. American Mineralogist, 66,908-930. 1984. A biotite isograd in South-Central Maine, USA: mineral reactions, fluid transfer, and heat transfer. Journal of Petrology, 25,871-893. 1986. Reaction progress: a monitor of fluid-rock interaction during metamorphic and hydrothermal events. In: WALXHER,J.V. & WOOD,B.J. eds. Fluid Rock Interactions During Metamorphism. Springer-Verlag, New York, 61-88. 1988a. Infiltration-driven metamorphism in northern New England. Journal of Petrology, 29, 1121-1159. 1988b. Contrasting mechanisms of fluid flow through adjacent stratigraphic units during regional metamorphism, south-central Maine, USA. Contributions to Mineralogy and Petrology, 1-12. 1992. Regional metamorphism of the Waits River Formation, Eastern Vermont: Delineation of a new type of giant metamorphic hydrothermal system. Journal of Petrology, 33, 45-94. & DIPPLE, G.M. 1991. Fluid flow, mineral reactions and metasomatism. Geology, 19, 211-214. FREY, M. 1987. Low temperature metamorphism. Blackie, New York. FROESE, E. 1969. Metamorphic rocks from the Coronation mine and surrounding area. Canadian Geological Survey paper, 68-5, 55-77. -1976. Applications of thermodynamics in metamorphic petrology. Canadian Geological Survey paper, 75-43. FYFE, W.S., PRICE, N.J. & THOMPSON, A.B. 1978. Fluids in the Earth's crust. Elsevier, Amsterdam. GOLDFARB,R.J., NEWBERRY,R.J., PICKTHORN,W.J. & GENT, C.A. 1991. Oxygen, hydrogen, and sulfur isotope studies in the Juneau gold belt, Southeastern Alaska: constraints on the origin of hydro,

B 8 9 ,

- -

-

-

-

-

-

-

9 8 ,

67

thermal fluids. Economic Geology, 86, 66-80. GREENWOOD, H.J. 1975. Buffering of pore fluids by metamorphic reactions. American Journal of Science, 273,573-593. HANSON,R.B. 1992. Effects of fluid production on fluid flow during regional and contact metamorphism. Journal of Metamorphic Geology, 10, 87-97. Ho, S.E., GROVES,D.I. & PHiLLiPS,G.N. 1985. Fluid inclusions as indicators of the nature and source of ore fluids and ore depositional conditions for Archaean gold deposits of the Yilgarn Block, Western Australia. Geological Society of South Africa, Transactions, 88,149-158. JAQUES, A.L., BLAKE,D.H. & DONCHAK,P.J.T. 1982. Regional metamorphism in the Selwyn Range area, northwest Queensland. Bureau of Mineral

Resources Journal of Australian Geology and Geophysics, 7,181-196. KERRICH,R. & FYFE, W.S. 1981. The gold-carbonate association: source of CO2 and CO2 fixation reactions in Archaean lode deposits. Chemical Geology, 33,265-294. KYSER, T.K. & KERRICH, R. 1991. Retrograde exchange of hydrogen isotopes between hydrous minerals and water at low temperature. In: TAYLOR, H.P., O'NEIL, J.R. & KAPLAN, I.R. (eds) Stable isotope geochemistry. Geochemical Society, Special Publications, 3,409-422. LAINC, W.P. 1991. Base metal + gold mineralization styles in the Cloncurry Terrane, Mount Isa Inlier. In: LAWRiE, H.C.F. (ed.) Base metal deposits symposium. James Cook University Economic Geology Research Unit Contribution 38, 77-88. LINDGREN, W. 1913. Mineral deposits. New York, McGraw-Hill. LOOSVELD, R.J.H. 1989. The intra-cratonic evolution of the central eastern Mount Isa Inlier, northwest Queensland, Australia. Precambrian Research, 44,243-276. MORA, C.I. & VALLEY, J.W. 1989. Halogen-rich scapolite and biotite: implications for metamorphic fluid-rock interaction. American Mineralogist, 74, 721-737. MUELLER, A.G., DE LAETER, J.R. & GROVES, D.I. 1991. Strontium isotope systematics of hydrotherreal minerals from epigenetic Archean gold deposits in the Yilgarn Block, Western Australia. Economic Geology, 86,780-809. NESBITr, B.E. 1982. Metamorphic sulphide-silicate equilibria in the massive sulphide deposits at Ducktown, Tennessee. Economic Geology, 77, 364-378. 1986a. Oxide-sulphide-silicate equilibria associated with metamorphosed ore deposits. Part I: Theoretical considerations. Economic Geology, 81,831-840. 1986b. Oxide-sulphide-silicate equilibria associated with metamorphosed ore deposits. Part II: Pelitic and felsic volcanic terrains. Economic Geology, 81,841-856. -& KELLY,W.C. 1980. Metamorphic zonation of sulphides, oxides, and graphite in and around the ore bodies at Ducktown, Tennessee. Economic Geology, 75, 1010-1021. -

-

-

-

68

G.N. PHILLIPS ET AL.

OLIVER, N.H. & WALL, V.J. 1987. Metamorphic

plumbing system in Proterozoic caic-silicates, Queensland, Australia. Geology, 15,793-796. --, VALENTA, R.K. & WALL, V.J. 1990. The effect of heterogeneous stress and strain on metamorphic fluid flow, Mary Kathleen, Australia, and a model for large-scale fluid circulation. Journal of Metamorphic Geology, 8, 311-331. PERRING, C.S., GROVES, D.I. & HO, S.E. 1987. Constraints on the source of auriferous fluids for Archaean gold deposits. In: Ho, S.E. & GROVES, D.I. (eds) Recent advances in understanding Precambrian gold deposits. Geology Department and Extension, University of Western Australia Publications, 11,287-306. PHILLIPS, G.N. 1988. Widespread fluid infiltration during metamorphism of the Witwatersrand goldfields: generation of chloritoid and pyrophyllite. Journal of Metamorphic Geology, 6, 311-332. & BROWN,l.J. 1987. Host rock and fluid control on carbonate assemblages in the Golden Mile Dolerite, Kalgoorlie gold deposit, Australia. Canadian Mineralogy, 25,265-273. - - & GROVES,D.I. 1983. The nature of Archaean gold bearing fluids as deduced from gold deposits of Western Australia. Journal of the Geological Society of A ustralia, 30, 25-39. & POWELL,R. 1992 Gold-only provinces and their common features. James Cook University Economics Geology Research Unit, Contribution 43. & WILLIAMS, P.J. 1993. Metamorphism, metasomatism and mineralization in the Cloncurry Belt. Australian Institute of Geoscientists Bulletin, 13, 31-33. - - , LAW, J.D.M. & MYERS, R.E. 1990. The role of fluids in the evolution of the Witwatersrand Basin. South African Journal of Geology, 93, 54-69. POWELL, R., WILL, T.M. & PmLLIPS, G.N. 1991. Metamorphism in Archaean greenstone belts: calculated fluid compositions and implications for gold mineralization. Journal of Metamorphic Geology, 9, 141-150. ROBERT, F. & BROWN, A.C. 1986a. Archean goldbearing quartz veins at the Sigma Mine, Abitibi greenstone belt, Quebec: Part I. Geologic relations and the formation of the vein system. Economic Geology, 81,578-592. & 1986b.Archean gold-bearing quartz veins at the Sigma mine, Abitibi greenstone belt, Quebec: Part II. Vein paragenesis and hydrothermal alteration. Economic Geology, 81,593-616. -& KELLY, W.C. 1987. Ore-forming fluids in Archean gold-bearing quartz veins at the Sigma mine, Abitibi greenstone belt, Quebec, Canada. Economic Geology, 82, 1464-1482. RUMBLE, D., FERRY,J.M., HOERING,T.C. & BOUCOT, A.J. 1982. Fluid flow during metamorphism at the Beaver Brook fossil locality, New Hampshire. American Journal of Science, 282,886-919. SIBSON, R.H. 1991. Fault structure and mechanics in relation to greenstone gold deposits. In: ROBERT, F., SHEAHAN,P.A. & GREEN, S.B. (eds) Greenstone gold and crustal evolution. Geological Association of Canada, 54-60. -

-

-

-

-

-

--.,

ROBERT,F. & POULSEN,K.H. 1988. High-angle reverse faults, fluid pressure cycling and mesothermal gold-quartz deposits. Geology, 16, 551555. SLAUGHTER, J., KERRICH, D.M. & WALL, V.J. 1975. Experimental and thermodynamicstudy of equilibria in the system CaO-MgO-SiOz--H:O-CO2. American Journal of Science, 275,143-162. SMITH, T.J., CLOKE, P.L. & KESLER, S.E. 1984. Geochemistry of fluid inclusions from the Mclntyre-Hollinger gold deposit, Timmins, Ontario, Canada. Economic Geology, 79, 1265-1285. SYMMES, G.H. & FERRY, J.M. 1991. Evidence from mineral assemblages for infiltration of pelitic schists by aqueous fluids during metamorphism. Contributions to Mineralogy and Petrology, 108, 419-438. THOMPSON, A.B. 1987. Some aspects of fluid motion during metamorphism. Journal of the Geological Society, London, 144,309-312. VALLEY,J.W., BOHLEN, S.R., ESSENE, E.J. & LAMB, W. 1990. Metamorphism in the Adirondacks: The role of fluids. Journal of Petrology, 31,555-596. WALL, V.J. & CEPLECHA,J.C. 1976. Deformation and metamorphism in the development of goldquartz mineralization in slate belts. 25th International Geological Congress, Abstracts, 1, 142143. WALTHER, J.V. & ORVILLE, P.M. 1982. Volatile production and transport in regional metamorphism. Contributions to Mineralogy and Petrology, 79,252-257. WILL, T.M., POWELL,R. & HOLLAND,T.J.B. 1990a. A calculated petrogenetic grid for ultramafic rocks in the system CaO-FeO-MgO-A1203SiO2-CO2-H20 at low pressures. Contributions to Mineralogy and Petrology, 105,347-358. -& GUIRAUD, M. 1990b. Calculated greenschist facies mineral equilibria in the system CaO-FeO-MgO-A 1203-SiO2-COe-H20. Contributions to Mineralogy and Petrology, 104, 353-368. WILLIAMS, P.J. & PHILLIPS, G.N. 1992. Cloncurry mapping project 1990: Geology of the Selwyn Range (McKinlay River and Maramungee Creek areas). James Cook University Economic Geology Research Unit Contribution 40. WOOD, B.J. & WALTHER,J.V. 1986. Fluid flow during metamorphism and its implications for fluid-rock ratios. In: WALTHER, J.V. • WOOD, B.J. (eds) Advances in Physical Geochemistry, 5. SpringerVerlag, 89-108. WOOD, S.A., CRERAR, D.A. & BORCSIK, M.P. 1987. Solubility of the assemblage pyrite-pyrrhotitemagnetite-sphalerite-galena-gold-stibnitebismuthinite-argentite-molybdenite in H20NaC1-CO2 solutions from 250°C to 350°C. Economic Geology, 82, 1864-1887. YARDLEY,B.W.D. 1986. Fluid migration and veining in the Connemara Schists, Ireland. In: WALTHER, J.V. & WOOD, B.J. (eds) Advances in Physical Geochemistry, 5. Springer-Verlag, 109-131. - - . , BOrrRELL, S.H. & CLIFF, R.A. 1991. Evidence for a regional-scale fluid loss event during midcrustal metamorphism. Nature, 349, 151-154.

Crustal stress, faulting and fluid flow RICHARD

H. S I B S O N

Department o f Geology, University o f Otago, P O Box 56, Dunedin, N e w Zealand Abstract: Differential stress exerts both static and dynamic effects on rock-mass

permeability, modulating fluid flow in the Earth's crust. Static stress fields impose a permeability anisotropy from stress-controlled features such as faults, extension fractures, and stylolites which, depending on the tectonic regime, may enhance, or counteract existing anisotropic permeability in layered rock sequences. Textural evidence from hydrothermal veins suggests, however, that fluid flow in fault-related fracture systems generally occurs episodically and that dynamic stress cycling effects are widespread. In the vicinity of active faults that undergo intermittent rupturing, permeability and fluid flux may be tied to the earthquake cycle through a range of mechanisms, leading to complex interactions between stress cycling, the creation and destruction of permeability, and fluid flow. Mechanisms for fluid redistribution include: (1) various forms of dilatancy (localized to the fault zone or extending through the surrounding rock mass) related to changes in shear stress and/or mean stress that occur during the fault loading cycle; (2) localized post-seismic redistribution around rupture irregularities, especially dilational jogs and bends which act as suction pumps; and (3) post-seismic discharge of fluids from overpressured portions of the crust through fault-valve action when ruptures breach impermeable barriers. All of these processes may be involved in fluid redistribution around active faults, but they operate to varying extents at different crustal levels, and in different tectonic regimes.

This paper explores the dynamic interplay between faulting and fluid flow in the Earth's crust. In recent years there has been a growing appreciation of the extent and varying character of fluid flow at all crustal levels (Oliver 1986; Cathles 1990; Fyfe 1990; Torgerson 1990, 1991). Sources for crustal fluids include meteoric water derived from the atmosphere through infiltration of the crust, connate and formation waters and hydrocarbon fluids in sedimentary basins, water derived from dehydration reactions during prograde metamorphism, and magmatic fluids derived from deep crustal levels or the mantle (Fyfe et al. 1978; Irwin & Barnes 1980; Gold & Soter 1984; Etheridge et al. 1984). Major driving forces for fluid migration include gradients in hydraulic potential that arise from topography, localized heat sources, sedimentary compaction, metamorphic dewatering, and mantle degassing. Because many of the driving potentials are long-lasting, there has been a tendency to analyse flow systems for constant permeability models (e.g. Norton 1982; Bethke 1986). Hydrothermal vein systems, however, provide abundant evidence for massive focused fluid flow along faults and fractures and their textures frequently record a history of incremental deposition, suggesting that the flow was intermittent (e.g. Hulin 1925; Newhouse 1942; Ramsay 1980; Sibson 1981). Gold-quartz

veins, especially, document the passage of substantial volumes of aqueous fluid through fault-fracture networks (e.g. Kerrich 1986; Cox et al. 1991; Sibson 1992a). As an example, consider the Mother Lode vein system of early Cretaceous age developed within the Melones fault zone in the western Sierra Nevada foothills of California (Knopf 1929). Individual quartz veins are hosted on reverse faults within the fault zone. In places, continuous veins averaging over 1 m in thickness can be traced for kilometres along strike and have been mined to depths in excess of 1 km. Per kilometre of fault strikelength, the volume of fault-hosted quartz is therefore c. 106 m 3. Established reverse displacements on the hosting faults range up to 100 m or so, with ribbon vein textures suggesting hundreds of episodes of hydrothermal deposition. Even assuming a 100% efficient precipitation mechanism, well in excess of 109 m 3 of aqueous fluid (the fluid volume of a supergiant oil-field!) would have to be flushed through the fault per kilometre strike-length to deposit this volume of quartz, given its low solubility (Fyfe et al. 1978). And the vein system extends along strike for at least 200km! A likely inference, therefore, is that fluctuations in stress and fracture permeability associated with episodic fault slip act to modulate primary flow systems. In particular structural settings, these fluctuations in stress and permeability appear to trigger episodes of

From PARNELL,J. (ed.), 1994, Geofluids: Origin, Migration and Evolution of Fluids in Sedimentary Basins,

Geological Society Special Publication No. 78, 69-84.

69

70

R.H. SIBSON

hydrothermal precipitation and play a crucial role in the formation of mineral deposits. No attempt is made to review the mathematical theory of time-dependent fluid redistribution in poroelastic media which is a subject of considerable complexity (e.g. Rice & Cleary 1976; Rudnicki & Hsu 1988). The approach adopted is semi-quantitative at best, and seeks only to provide simple physical models that account for the geological evidence of largescale fluid redistribution around faults in certain tectonic environments. Because of the uncertainties in many of the critical parameters and constitutive relationships, and the heterogeneous character of material properties (strength, permeability, etc.) in the Earth's crust, the analysis is couched in terms of classical rather than modern fracture mechanics. While this approach is less than satisfactory in quantifying the different mechanisms for fluid redistribution, it does identify areas where future detailed modelling may prove worthwhile.

Seismogenic crust Studies of tectonically active terrains have shown that fault displacements in the upper continental crust are largely accomplished by increments of seismic slip along pre-existing structures. The depth extent of seismic activity, representing the zone of unstable frictional sliding, is temperature-dependent. For moderate or higher geothermal gradients away from areas of subduction, this seismogenic regime is restricted to the upper one-third to one-half of deforming continental crust (Sibson 1983). In the context of plate tectonics, therefore, the Earth's crust may be divided into intraplate regions, where the rate of seismic activity is low and the stress field is effectively static over long time periods, and areas of active deformation associated with plate boundaries where stress fields in the vicinity of active faults are coupled to the earthquake stress cycle and are timedependent. Thus, whereas fluid flow in intracratonic regions is unlikely to be affected by comparatively short-term stress and permeability cycling, flow in tectonically active areas may undergo episodic perturbation on time-scales of the order of the repeat times of large earthquakes (10-104 years). Subsidiary perturbations may be induced by second order seismicity.

freely linked through to the Earth's surface. In such circumstances, the pore-fluid factor, hv = Pf/crv = Pd(ogz) ~ 0.4, where Pe and O-v are, respectively, the fluid pressure and vertical stress (overburden pressure) at a depth, z, in the crust, p is the average rock density, and g is the gravitational acceleration. However, there is a great deal of evidence that at depths greater than a few kilometres within deforming crust, fluid pressures commonly rise above hydrostatic towards lithostatic values (hv---~1) (see below, Figs 7 & 9). The condition kv ~ 1 is also widely assumed to prevail in areas undergoing prograde regional metamorphism (Fyfe et al. 1978; Etheridge et al. 1984). Hubbert & Rubey (1959) first called attention to the role of overpressured (suprahydrostatic) fluids in reducing the frictional strength of faults, which can be represented by a criterion of Coulomb form: 7f = C + ~LsO'n' : C q- ~l,s(Orn -- Pf)

(1)

where C is the cohesive or cementation strength of the fault (small for an active structure), ixs is the static coefficient of rock friction, and o-, is the normal stress on the fault. From laboratory experiments, Ixs = 0.75 + 0.15 for a broad range of rock types (Byerlee 1978), and field observations suggest that natural faulting involves comparable friction coefficients (Sibson 1990a). Fault reactivation thus occurs when the shear stress along the fault, a', equals "re, and may be brought about by rising shear stress, decreasing normal stress, or by an increase in fluid pressure. This simple criterion has been shown to be applicable to several cases of induced seismic activity occurring in the top few kilometres of the crust (Nicholson & Wesson 1990) and may well remain valid throughout the seismogenic zone, though fault failure at depth may also be affected by time-dependent processes such as stress corrosion (Das & Scholz 1981). Note, however, that there is now good evidence for seismic rupturing within overpressured portions of the crust such as the Western Taiwan fold/thrust belt and regions adjacent to the San Andreas fault in California (Sibson 1990a). Complex coupling between episodes of fault failure, the creation and destruction of fracture permeability, and fluid redistribution is especially likely in such overpressured regions (see later discussion on fault-valve activity).

Coupled time-dependent permeability Fluid pressure and fault stability Hydrostatic fluid pressures prevail where interconnected pore space and fracture systems are

While inactive faults often act as impermeable seals through the presence of clay-rich gouge and hydrothermal cementation (Smith 1980;

CRUSTAL STRESS, FAULTING & FLUID FLOW Hooper 1991), the intrinsic roughness of natural fault surfaces (Power et al. 1987) has the implication that freshly ruptured faults should become highly permeable, if tortuous, channelways post-failure. Rupture zone permeability is, however, likely to be short-lived. Evidence from geothermal fields suggests that hydrothermal flow along fractures rapidly leads to hydrothermal precipitation and self-sealing (Grindley & Browne 1976; Batzle & Simmons 1977). Experiments show that flow of hot water ( T > 200 ° C) along pressure gradients in granite leads to dramatic reduction in permeability through hydrothermal dissolution and reprecipitation (Morrow et al. 1981). In addition, laboratory experiments suggest that diffusional crack healing in 'wet' quartz is fast at T > 200° C (Smith & Evans 1984; Brantley 1992). Solution-precipitation creep is also particularly effective in clogging porosity and lowering permeability in fine-grained quartz-bearing rocks over the temperature range 200-400°C (McClay 1977; Sprunt & Nur 1977; Cox & Etheridge 1989). Observational and experimental periods over which these various processes are effective in reducing permeability range from hours to months. Thus in the vicinity of active faults, there must be continual competition between the creation and destruction of fracture permeability. While permeability destruction through hydrothermal activity is known to be fast-acting in high-level geothermal systems, it also seems likely to be particularly effective in the bottom half of the continental seismogenic regime at, say, 7-15 km depth. These processes of porosity and permeability destruction (Walder & Nur 1984; Nur & Walder 1990) counter the arguments of Brace (1980, 1984), based on laboratory and in situ measurements of rock permeability, that crystalline rocks are generally too permeable to allow the development of fluid overpressures. They play an important role in some of the recent models that attempt to account for the apparent weakness of major fault zones in terms of fluid overpressuring (Byerlee 1990, 1993; Sibson 1990a; Rice 1992; Sleep & Blanpied 1992). Static stress field effects Stress-controlled features affecting rock permeability include brittle faults, microcracks, extension fractures, and stylolitic solution seams. Their relationship to a stress field with principal compressive stresses 0-1 > 0-2> 0-3 in homogeneous, isotropic rock is as shown in Fig. 1, which illustrates the range of fault/fracture/

71

stylolite mesh systems that may develop in different circumstances. Faults tend to initiate as Coulomb shears containing the 0-2 axis and lying at + 20--30° to 0.1 (Anderson 1951). In 'classical' fracture mechanics, macroscopic extension fractures (mode I cracks in the parlance of modern fracture mechanics) form perpendicular to 0.3 by natural hydraulic fracturing in accordance with the criterion: P f = 0-3 q- Z

(2)

provided (0-1-0.3) < 4T, where T is the long-term tensile strength of the rock (Secor 1965). Microcracks developing by grain impingement also have a preferred orientation perpendicular to 0-3. Impermeable stylolites develop as anticracks in surfaces subperpendicular to o-1 (Fletcher & Pollard 1981). The different components may combine in various fault/fracture/stylolite meshes as shown, but the full range of components need not be present in all cases. For instance, the development of stylolites by pressure solution processes depends critically on factors such as grain size and composition, being far more widespread in fine-grained sedimentary rocks, particularly limestones (Groshong 1988). Faults and fractures have the potential to impart directional permeability to the rock mass (e.g. du Rouchet 1981), but may also become choked with hydrothermal deposits and/or clayrich gouge and alteration products (e.g. Roberts 1991). Moreover, the permeability characteristics of a fault depend to some extent on the character of the host rock. Faults developing in initially high-porosity sedimentary rock may, through comminution and porosity collapse, form deformation bands that are relatively impermeable in comparison with the wallrock (Aydin & Johnson 1978). In contrast, faults in initially low-porosity rocks may enhance permeability locally through cataclastic brecciation. Extension fracturing on both microscopic and macroscopic scales can greatly enhance permeability in the 0-1/0-2plane, the effect becoming pronounced as Pf---~0-3. Comparatively impermeable stylolitic seams are associated with tabular zones of reduced porosity (Groshong 1988; Carrio-Schaffhauser et al. 1990) and, overall, will generally tend to restrict flow perpendicular to the 0-J0-3 plane. Existing anisotropic permeability in layered rock sequences may therefore be either enhanced or reduced, depending on the type of stress field prevailing and the dominant attitude of the layering (Fig. 1). In the case of predominantly horizontal layering, and for the more

72

R.H. SIBSON

Fig. 1. Faults (Coulomb shears), extension fractures (hachured), and stylolites (squiggly lines) forming in a triaxial stress field, illustrating possible combinations of these stress-controlled features which might affect permeability of the rock mass. A Hill fault/fracture mesh is shown in the lower left. Diagram represents a thrust fault stress regime when upright, a normal fault regime when viewed sideways, and a strike-slip regime in plan view. common situation where fracture-enhanced permeability predominates over impermeable stylolitic bands, bedding-parallel permeability will be enhanced in a compressional tectonic regime, while trans-bedding permeability will tend to be enhanced in extensional and strikeslip regimes (cf. du Rouchet 1981). In viewing the diagrams, the enhancement of out-of-plane permeability in the 0-2 direction should also be noted; this is likely to be of special significance in strike-slip regimes. Where the dominant anisotropy is sub-vertical, as in belts of upright folds with well-developed slaty cleavage, transanisotropy permeability will be enhanced by low-angle faults and extension fractures developed through horizontal compression. The situation may become vastly more complicated in regions of complex structure where rotational strains have developed in heterogeneous stress fields.

Fluid expulsion through Hill fault~fracture meshes Hill (1977) showed how the seismological characteristics of earthquake swarms could be accounted for by the passage of hydrothermal or magmatic fluids through a 'honeycomb mesh' of interlinked shear and extension fractures (Figs 2 & 3). Such meshes may develop in extensional and strike-slip settings because individual hydraulic extension fractures can only extend over

limited depth intervals in the Earth's crust (Secor & Pollard 1975). The ability of such fault/fracture meshes to transmit fluids is highly sensitive to the value of the effective least compressive stress, 0-3' = (0-3 - Pf), but becomes large when 0"3' ~ 0. This can be achieved close to the Earth's surface in areas of extensional and strike-slip tectonics, but the meshes may also develop at depth in the crust when fluid pressure approaches the lithostatic load (hv ~ 1). As a consequence of the contrasting mechanical properties of alternating layers, Hill fault/ fracture-meshes are particularly likely to develop in horizontally-layered rock sequences undergoing extension with 0.3 horizontal and 0-1 vertical (Fig. 3). Under conditions of increasing fluid pressure, the brittle failure mode of intact rock is governed by the level of differential stress, (0"1-0"3), in relation to its tensile strength, T (Secor 1965). Provided (o-1- 0"3) < 4T, hydraulic extension fractures form perpendicular to 0"3 in accordance with equation (2), whereas if (0-~ - o-3) > 4T, the rock fails in shear in accordance with the Coulomb criterion (eqn 1). Thus, in a typical sequence of alternating sandstones and shales with T~st> T~h, the shales may fail by development of brittle Coulomb shears (perhaps in conjugate sets), while the sandstone fails by hydraulic extension fracturing to form a fault/fracture mesh. Natural fracture meshes of this kind may develop over a range of scales, but generally tend to be less ordered than

CRUSTAL STRESS, FAULTING & FLUID FLOW

73

! L

shale

~

~

~

~

!

(31 ..iiiiii .....!.

t

shale sandstone

I

'I

COMPRESSIONAL REGIME

I

t

EXTENSIONAL REGIME

Fig. 2. Schematic representation of stress-control of permeability by 'brittle' structures (faults, extension fractures (hachured), and stylolites (wavy lines)) affecting an anisotropic, sandstone-shale sequence in compressional and extensional stress regimes. (Yl

I (Y3

A

EXTENSIONAL CHIMNEY 1"

'FUZZY' NORMAL FAULT

Fig. 3. Hill fault/fracture meshes acting as fluid conduits within an extensional stress regime, either as extensional chimneys or as 'fuzzy' normal faults. Note that the passage of large fluid volumes leads to increased disorganisation and brecciation within the fault/fracture mesh.

those portrayed diagrammatically in Fig. 3. The passage of large fluid volumes through fault/ fracture meshes leads to brecciation with rock fragments initially bounded by combinations of shear and extensional fractures. Continued flow causes rotation and progressive comminution of the fragments in a process akin to fluidisation. Such processes may have contributed to the extensive tabular bodies of high-dilation breccia within the Monterey Shales of California

(Redwine 1981) and similar formations elsewhere. Fault/fracture-meshes of this kind, developed in well-layered sedimentary sequences, tend to form either as subvertical extensional 'chimneys', or as 'fuzzy' normal faults with an overall component of shear across the mesh (Fig. 3). In the Monterey Shales they have apparently acted as major fluid expulsion structures, transporting large volumes of aqueous and hydrocarbon

74

R.H. SIBSON

fluids across the bedding anisotropy, the relict mesh structure being preserved through cementation by silica, carbonate, and bitumen. Such structures likely developed by the breaching of seals to overpressured fluid compartments.

I

i

Stress cycling effects Earthquake ruptures spread over portions of existing fault surfaces at c.3 km s -1 (corresponding to the shear wave velocity in the upper crust), so that in the case of moderate (c.M5) to large (c.M7) ruptures with dimensions of kilometres to tens of kilometres, the period of coseismic rupturing typically lasts for seconds to tens of seconds (Sibson 1989). Accompanying effects relevant to fluid redistribution include: (i) the creation of transient fracture permeability along the primary rupture zone; (ii) the development of subsidiary fracturing at specific structural sites such as fault jogs; (iii) shear stress reduction over the rupture plane and intensification at the rupture tips, with localized enhancement or reduction of mean stress around fault irregularities; and (iv) the abrupt reduction of fluid pressure at dilational sites within the rupture zone.

Several time-scales must therefore be considered in relation to the earthquake stress cycle; long-term accumulation of shear stress through the inter-seismic period lasting tens to perhaps many thousands of years, possible pre-seismic stress changes associated with precursory slip, the rapid co-seismic drop in shear stress during rupturing which at any one place occurs over a period of a few seconds, and a period of post-seismic adjustment (corresponding to the aftershock phase) which may last for days to years depending on the size of the rupture and the physical characteristics (strength, heterogeneity, permeability, etc.) of the rock mass (Fig. 4). Elasticity theory suggests that significant stress cycling around a seismically active fault should be restricted to a response zone whose extent compares broadly with characteristic rupture dimensions, extending laterally for perhaps 10-15 km from ruptures that occupy the full depth of the continental seismogenic zone (Sibson 1989). The highest amplitude cycling occurs close to the causative fault but is likely to be locally exaggerated in the vicinity of jogs and other fault irregularities. However, it should be noted that in recent years there has been increasing recognition of triggering of remote seismicity and other effects, likely to be fluidrelated, that occur far outside the expected response zone derived from linear elasticity theory (e.g. Hill et a/.1993).

A

Ax I

int°rseismicl

t

I

EQ

EQ

EQ

I

I

I

TIME Fig. 4. Variations in mean stress (6-) during the cyclic accumulation and release of shear stress (-r) on thrust and normal faults that are optimally oriented for frictional reactivation with ~ = 0.75. Different phases of the earthquake stress cycle and fluid movement in and out of the response zone are indicated. Dotted lines indicate possible timedependence of stresses post-failure.

Dilatancy related to the fault loading cycle The state of dilatational strain in a rock mass may be affected by variations in the levels of both shear stress and mean stress. In the search for precursory effects to fault failure, attention was focused initially on possible dilatant effects associated with the earthquake cycle of shear stress accumulation and release, the belief being that dilatant strains should develop in the rock mass around a fault as shear stress increases during fault loading (Fig. 4). Fluids would be drawn into the dilatant rock volume pre-failure, only to be expelled post-failure once shear stress was relieved. This dilatancy-diffusion hypothesis for earthquake prediction (Nur 1972; Scholz et al. 1973) was predicated on the development of extensive high-stress (a->100MPa) microcrack dilatancy in the surrounding rock mass. It formed the basis of the original seismic pumping concept developed by Sibson et al. (1975) to account for fluid redistribution in the vicinity of seismically active faults. Twenty years on, there is little evidence that this particular form of dilatancy or the requisite stress levels are widespread in deforming seismogenic crust (Hickman 1991), so that the original seismic pumping mechanism must be considered invalid. However, as Nur (1975) has pointed out, other forms of dilatancy with different stress-dependences may also operate in the vicinity of fault zones, and need to be

75

CRUSTAL STRESS, FAULTING & FLUID FLOW Table 1: Postulated dilatancy mechanisms

Mechanism

Primary dependence

Distribution of dilatancy

Depth within seismic zone

References

High-stress microcrack dilatancy at high -r and high ~-' = (b - Pf) Low-stress microcrack dilatancy from subcritical crack growth at low a" Low-stress microcrack dilatancy at low -r, high Pf, and low ~' 'Sand-pile' dilatancy under low -r and low &'

Aa-

Laterally extensive Laterally extensive

Deep

Nur 1972; Scholz et al. 1973 Crampin et al. 1984

Laterally extensive? Restricted to fault zone

Deep

Laterally extensive? Laterally extensive

Shallow

Fischer & Paterson 1989 Nur 1975; Marone et al. 1991 Nur 1975

Shallow?

Sibson 1991

Laterally extensive? Restricted to fault zone

Shallow-deep

Sibson 1981

Deep

Borraidale 1981; Cox & Etheridge 1989

A-r A-r A~

Existing joint/fracture dilatancy

A,r and/or ATr?

A?rcompactive effects in highly fractured or porous material at low ~r' Hydrofracture dilatancy under low ~rand high Pe Grain-scale particulate flow under low ?r' and high Pf

A?r APf

Apf

evaluated as possible dilatancy pumping mechanisms. The different dilatancy mechanisms may be grouped into those sensitive primarily to varying shear stress, those sensitive to variations in mean stress, and those driven by high fluid pressure levels (Table 1), though some degree of overlap occurs. More than one of the mechanisms may contribute to fluid redistribution in any particular circumstance. Shear stress sensitive dilatancy is likely to predominate in low porosity rocks, whereas high porosity rocks may develop significant dilatant strains from variations in mean stress. It has to be emphasized, however, that no firm constitutive relationships applicable to the real Earth have yet been established for any of the postulated mechanisms. Another key issue, affecting the volume of fluid involved in dilatancy pumping, is the extent of the region experiencing cyclic dilatancy. Are significant dilatant strains restricted to the fault zone itself, to its immediate surrounds, or do they extend over broad areas of the surrounding rock mass? Dilatancy dependent on shear stress variations. Nur (1975) recognized three varieties of dilatancy dependent on shear stress; microcrack dilatancy which is sensitive to small variations in shear stress at high ambient levels of shear stress, sand-pile dilatancy which is stresssensitive at low ambient levels of shear stress, and joint dilatancy in fractured rock masses which maintains an approximately uniform

Shallow-deep

Shallow

stress sensitivity. As previously discussed, it seems unlikely that high-stress microcrack dilatancy is widespread in seismogenic crust. Crampin et al. (1984) have suggested, however, that microcrack dilatancy may also develop by slow sub-critical crack growth under comparatively low shear stress levels and give rise to extensive-dilatancy anisotropy ( E D A ) over broad areas of deforming crust. In addition, Fischer & Paterson (1989) have demonstrated that significant microcrack dilatancy may also develop under low stress levels at high fluid pressures and low effective confining pressures. Sand-pile dilatancy is likely to develop in any granular aggregate material under low levels of shear stress and must undoubtedly operate in the granular cataclastic material contained within brittle fault zones at high crustal levels (e.g. Marone et al. 1990). However, the volume of fluid involved is then likely to be comparatively minor. Extensively fractured rock masses adjacent to fault zones at high crustal levels are also likely to experience joint dilatancy as shear stress varies. Dilatancy dependent on mean stress variations Tectonic shear stress on a fault cannot generally be increased without also changing the normal stress on the fault (altering its frictional strength) and the level of mean stress (g-= [o-1 + o-2 + o-3]/3) (Sibson 1991). This in turn affects the state of dilatational strain, especially within high porosity rock masses. Patterns of fluid flow

76

R.H. SIBSON

during fault loading must therefore be influenced by poroelastic effects arising from changes in 6. as the shear stress on a fault increases. This is illustrated in Fig. 4 for the end-member cases of thrust and normal faults that are optimally oriented for frictional reactivation; the coupled changes in mean stress (A6.) are somewhat greater than the shear stress drop (AT). For the two cases illustrated, a typical earthquake shear stress drop, A,r = 1 MPa, causes a post-failure increase in mean stress, A6. = 1.25 MPa in the vicinity of a normal fault, or a post-failure decrease, A6. = 1.25 MPa, in the case of a thrust fault. This corresponds to a change of _+ 125 m in equivalent hydrostatic head. Full analysis of these poroelastic effects in crust with a heterogeneous distribution of stress and permeability is likely to be complicated (e.g. Rice & Cleary 1976) but, neglecting for the time-being any shear stress related dilatancy and rupture tip effects, it is clear that as shear stress rises during the loading of a thrust fault, mean stress also increases so that fluids should be driven away from the fault and out of the response zone by poroelastic compaction, only to be drawn back in post-failure. In the case of normal faults, the to-and-fro motion should be reversed, with fluids drawn in towards the fault during loading to be expelled post-failure. For strike-slip faulting, the coupling of mean stress to shear stress can lie anywhere between these end-member cases (Sibson 1993). More complicated patterns of fluid redistribution from mean stress changes are to be expected at rupture tips (Muir Wood & King 1993). Fluid redistribution from varying mean stress is likely to be important when the effective mean stress, 6. = (6. - P0 ~ 0, and the ratio (A6./6.') is large. These conditions will be satisfied in the near-surface, where intensely fractured portions of the upper crust respond as a fluid-saturated, blocky aggregate. In such settings, cyclic variations in mean stress should induce to-and fro-motion of fluids, with fluid pressures staying close to hydrostatic. A6. effects may, however, also be strong at depth in the crust when 5- -(5- - Pf) ~ 0 and the rock mass is deforming by particulate flow (see below). The 1983 M7.3 earthquake at Borah Peak, Idaho, which involved rupturing of a rangebounding normal fault, provided a good example of post-seismic discharge coupled to an increase in mean stress postfailure. The earthquake was followed by a large fluid discharge (c.0.3km 3) over a period of several months, including immediate post-seismic fountaining from a series of fissures striking parallel to the rupture trace (Wood et al. 1985). The discharge

seems largely attributable to the increase in horizontal stress and mean stress in the upper few kilometres of the crust consequent on shear stress relief along the normal fault (Sibson 1991; Muir Wood & King 1993).

Dilatancy driven by fluid pressure.Arrays of extension fractures develop in the crust by hydraulic fracturing when the requisite conditions (Pf = o-3 + T, and (o-l - o-3) < 4T) are met (Secor 1965). Because the tensile strength of rocks is generally low (T < 10 MPa), such hydrofracture dilatancy is a low differential stress phenomenon. The necessary conditions may be satisfied by hydrostatic levels of fluid pressure within perhaps a few hundred metres of the earth's surface in extensional and strike-slip regimes, but fluid overpressures with hv ~ 1 are needed at greater depths (Sibson 1981). Provided the rock mass retains a finite tensile strength, fluid pressures in excess of the lithostatic load (~.v-> 1) are required at all depths for hydraulic extension fractures to develop in compressional stress regimes where o.v = o.3. Localization of hydrofracture arrays to the immediate vicinity of fault zones suggests that, in many cases, the fault zones are themselves the principal conduits for the migration of overpressured fluids (Cox et al. 1991). The development of hydrofracture dilatancy in compressional regimes that are strongly fluid overpressured is an accompaniment to extreme fault-valve action (see below). Grain-scale microcrack dilatancy and associated particulate flow (approximately equivalent to sand-pile dilatancy) may also develop in fluid overpressured rock masses where Xv~ 1, and the effective mean stress, 6.'= ( 6 . - P f ) ~ 0 (Borraidale 1981; Cox & Etheridge 1989; Fischer & Paterson 1989; Knipe 1989). Mixed dilatancy effects. In most natural settings it is likely that more than one of these various dilatancy mechanisms may be operating at a given time. For example, consider the coupled variation of both shear and mean stress around thrust and normal faults as illustrated in Fig. 4. In the case of the thrust fault, any dilatancy related to increasing shear stress during fault loading is opposed by the coupled rise in mean stress, whilst any post-failure tendency for crack closure from reduced shear stress is counteracted by the lowered mean stress. In the case of normal faults, however, development of dilatancy during loading is favoured both by the increasing shear stress and by the coupled reduction in mean stress. Post-failure, reduced shear stress and increased mean stress both

CRUSTAL STRESS, FAULTING & FLUID FLOW

77

ru/pture nucleation ~'/

dilational jog

compressional bend

~

"

~

" ~'~ compressional jog

Fig. 5. Seismotectonic carbon of an irregular rupture trace (not to scale), showing areas of enhanced (+ and cross-hatched) and reduced ( - and diagonal hachures) mean stress arising from a rupture propagating from left to right. Note that the response of isolated fault bends depends on the direction of rupture propagation. Diagram represents a map view of a strike-slip fault, or a cross-section through a dip-slip fault.

contribute to crack closure. On these arguments, dilatancy effects throughout the fault loading cycle should be more pronounced in extensional rather than compressional stress regimes. However, another factor to be considered is that at the same depth and fluid pressure level (kv value), the shear stress required for frictional reactivation of a thrust fault is about four times that needed to reactivate a normal fault (Sibson 1974). Effects from shear stress related dilatancy are therefore likely to be more pronounced in the vicinity of thrust faults. Until the details of the stress levels driving faulting and the constitutive laws for the different dilatancy mechanisms are fully resolved, it is impossible to evaluate the contributions of the various mechanisms to fluid redistribution in the crust. Of critical concern, because it affects the volume of fluid redistribution by dilatancy pumping, is the question of whether cyclic dilatational strains are generally localized to the material within fault zones, or whether they extend over broad regions of the surrounding crust. While there is considerable geological evidence for localized dilatancy within shear zones at all crustal levels (Hobbs et al. 1990), documentation of microstructures demonstrating cyclic dilatancy on a regional scale is generally lacking. Thus, the extent to which dilatancy pumping of one form or another may redistribute fluids, and in particular may draw fluids down from the near-surface to levels of the crust undergoing prograde metamorphism, as proposed by McCaig (1988), remains unresolved.

Post-seismic fluid redistribution Rupturing of an irregular fault surface leads to abrupt post-failure changes in mean stress

localised around the structural irregularities (Segall & Pollard 1980; Sibson 1989) (Fig. 5). As a consequence, there is a tendency for fluids to be redistributed from areas of raised mean stress to areas of lowered mean stress (Nur & Booker 1972). Fluids are driven out of compressional jogs and bends where mean stress increases post-failure, while the most intense fluid influx, coupled with aftershock activity, is concentrated in regions of sharply reduced mean stress such as dilational jogs and bends. At these dilational sites, rapid slip transfer during rupture propagation causes abrupt local reductions in fluid pressure below ambient (hydrostatic?) values. In some instances, the induced suctional forces may lead to rupture arrest (Sibson 1985, 1989). Dilational fault jogs and bends thus act essentially as suction pumps and are often characterized by multiply recemented wallrock breccias resulting from repeated hydraulic implosion. Larger dilational structures typically comprise a mesh of extension veins, implosion breccias, and subsidiary shears in the form of a Hill fault/ fracture mesh. Active geothermal systems are commonly localized within dilational sites in extensional and transtensional fault systems. Where transected by faults in the top kilometre of the seismogenic zone (the epithermal environment), pressure reductions from slip transfer may induce episodes of boiling within the hydrothermal system, triggering mineral deposition throughout the phase of aftershock activity (Sibson 1987). The association of epithermal mineralization with the high levels of extensional/transtensional fault systems thus arises from the coincidence of this 'boiling horizon' with the near-surface region where extensive hydrofracturing may occur under hydrostatic fluid pressure conditions (Sibson 1992a).

78

R.H. SIBSON

(TECTONIC'~ -i-~ ~LOADING'J ~

I FAULT I I INSTABILITY I I ~ = C +/~s(an-Pr)

I

• R U P T U, R IxN G

//

(PERMEABILITY) k DESTRUCTION) Self"Se aling "

/

~CREATION OF "~ rHYDROTHERMAL'~ FRACTURE Fluid Discharge PRECIPITATION Pf

Decrease

Fig. 6. Schematic representation of fault-valve activity, illustrating the dependence of frictional fault failure on the cycling of both tectonic shear stress and fluid pressure (after Sibson 1992b). Fault-valve action in o v e r p r e s s u r e d crust

From the time of Hubbert & Rubey's (1959) seminal paper on the mechanics of low-angle thrusting, it has become increasingly clear that overpressured fluids play a critical role in faulting, and may be especially important in the case of faults that are unfavourably oriented for frictional reactivation (Cello & Nut 1988; Sibson 1990a). Recent oil-field studies have demonstrated the widespread occurrence of fluid pressure compartments in sedimentary basins, and analysis of oil migration paths suggests that overpressured compartments become breached from time to time with episodic discharge of overpressured fluids (Hunt 1990; Powley 1990). There is also now good evidence that, in at least some areas, seismic rupturing is occurring within overpressured portions of the crust. Fault-valve action thus occurs wherever ruptures transect suprahydrostatic gradients in fluid pressure and breach impermeable barriers, leading to upwards fluid discharge along the transient permeability of the fault zone and local reversion towards a hydrostatic fluid pressure gradient before hydrothermal self-sealing of the rupture zone occurs, and fluid overpressures rebuild at depth. Under such circumstances, the timing of successive failure episodes is controlled by the cycling of both tectonic shear stress and fluid pressure through the interseismic period (Fig. 6). A key issue, affecting the volume of fluid available for post-failure discharge and the time for fluid replenishment of the fault zone in the interseismic period, is the extent to which fault zones are locally overpressured in relation to the surrounding crust (Fig. 7). On mechanical grounds, a case can be made that valve action is most likely to be prevalent in compressional or transpressional fault systems. Steep reverse faults are particularly likely to be

effective as fault-valves because they are severely misoriented for reactivation, and frictional shear failure can only initiate when slightly supralithostatic fluid pressures are attained (Sibson etal. 1988; Sibson 1990b). Under these conditions, arrays of subhorizontal extension fractures develop by hydraulic fracturing adjacent to the fault (hydrofracture dilatancy) prior to failure, to form a lithostatically pressured reservoir that, together with associated grain-scale dilatancy, may contain a substantial fluid volume (Fig. 8). Fault failure and rupturing through the upper crust then allows fluid from the overpressured reservoir to drain upwards, with fluid pressures dropping rapidly towards hydrostatic values. The discharge may be accompanied by phase separation (if CO2 is present), rapid mineral precipitation and hydrothermal self-sealing of the fault. Fluid pressure then rebuilds towards the critical value needed to trigger the next episode of fault slip, and the cycle repeats. Such faults may give rise to extreme fault-valve action involving the discharge of large fluid volumes, with fluid pressure potentially cycling between pre-failure lithostatic and post-failure hydrostatic levels. Extreme valve action appears responsible for many mesothermal Au-quartz lodes hosted in steep reverse or reverse-oblique faults, such as the Mother Lode vein system described previously (Sibson et al. 1988; Cox et al. 1991; Boullier & Robert 1992). A typical vein assemblage consists of flat-lying extensional veins, inferred to have developed by hydraulic fracturing prior to episodes of fault slip, in mutual cross-cutting relationships with steep fault-veins formed during episodic upwards discharge along the steeply dipping fault zone (Fig. 8). The hosting shear zones are typically of mixed discontinuous-continuous ('brittle-ductile') character and developed under greenschist

CRUSTAL STRESS, FAULTING & FLUID FLOW

FLUID PRESSURES INSIDE FAULT ZONE

I

!1

I

I

FLUID PRESSURES OUTSIDE FAULT ZONE ~'V

1.0

I

I

i

I

1.0 i i

I I

I I

79

I

m

¢¢

:i ".n~ 's

....

,,,///////~;,~,,~

I W

¢)" a

Fig. 7. Hypothetical plots of pore-fluid factor, ~,,, versus depth, illustrating pre-failure and post-failure fluid pressure gradients inside a fault zone undergoing fault-valve action in relation to possible gradients (curves A, B, C, D) in the surrounding rock mass. Curve A represents an end-member case where the pressure distributions inside and outside the fault zone are identical. Curves B, C, and D represent different degrees of localization of fluid overpressure within the fault zone, with D representing a crust that is hydrostatically pressured throughout the seismogenic zone. Conceivably, curves A-D could reflect fluid pressure gradients with increasing distance from a major fault zone.

facies metamorphic conditions at crustal levels corresponding to the base of the seismogenic zone (depths of c . 1 0 ± 5 k m ) with the vein material precipitated from low salinity, mixed H20-CO2 fluids (Robert & Kelly 1987). Localization of the prefailure arrays of hydraulic extension fractures to the vicinity of the fault zones (generally to within tens to hundreds of metres) suggests that the fault zones are locally overpressured with respect to the surrounding crust, but the degree of comparative overpressuring is unclear (see Fig. 7). Valving action leading to the formation of mesothermal vein systems may also occur at depth within strikeslip fault systems in the vicinity of compressional jogs and bends (Fig. 5). This is in direct contrast to epithermal mineralization which typically

develops in extensional/transtensional fault systems (Sibson 1992a) While extreme valve action (involving the episodic discharge of large fluid volumes and the formation of major hydrothermal vein systems) may be comparatively rare and restricted to specific tectonic settings, the frequent presence of minor syn-tectonic veins in fault zones (e.g. Chester et al. 1993), and the growing evidence from fluid inclusion studies for fluid pressure cycling (e.g. Parry & Bruhn 1990; Mullis 1990), suggest that basic fault-valve activity is widespread in the earth's crust. Even minor valving action involving small fluid volumes is likely to have a considerable effect on the nucleation and recurrence of earthquakes (Sibson 1992b). The depositional environment of mesothermal vein

80

R.H. SIBSON

~'~. ,~ ZO N E I +~'~(~1 ~'~,rupture nucleatiosinte,~!,,l,IIl,III

~3

Pfl I

"h'yd~s6tic TIME

Fig. 8. Schematic diagram of extreme fault-valve action and associated fluid-pressure cycling on a

high-angle reverse fault. Mesothermal gold-quartz lodes form in the region of large rupture nucleation and intense fluid pressure cycling near the base of the seismogenic zone.

systems also suggests that the region towards the base of the seismogenic zone may in general represent a time-dependent interface between hydrostatically and lithostatically pressured portions of the crust, as represented schematically in Fig. 9. General fault zone model The absence of a localized heat flow anomaly centred on the trace of the San Andreas strike-slip fault in California limits the timeaveraged frictional shear resistance in the seismogenic zone to < 2 0 M P a (Lachenbruch & McGarr 1990). Additional evidence for a very weak fault zone comes from the accumulation of data suggesting that the San Andreas fault is extremely unfavourably oriented for frictional reactivation within the regional stress field, lying at 65-85 ° to the greatest compressive stress (Mount & Suppe 1987). Present knowledge of the frictional characteristics of rock materials suggests that, despite the problem of containment, fluid overpressures are the most probable weakening mechanism that could account for this low frictional strength. As a consequence, a

range of models accounting for the generation and maintenance of fluid overpressures in transcrustal fault zones have recently been proposed. A key issue in the different models is whether the high fluid pressures are derived from the fault zone acting as a migratory conduit for overpressured fluids (Sibson 1990a; Stark & Stark 1991; Rice 1992), or whether the fluid overpressures are continually regenerated from essentially the same fluid volume during cyclical loading (Byerlee 1990, 1993; Sleep & Blanpied 1992). The development of extensive hydrothermal veining in fault zones, especially the gold--quartz mineralization precipitated at structural levels corresponding to the lower half of the seismogenic zone, provides geological evidence that fault zones, in at least some circumstances, are acting as migratory conduits for the passage of rather substantial fluid volumes (Kerrich 1986). If fluid overpressures and associated faultvalve action leading to fluid pressure cycling are as widespread as evidence is beginning to suggest, then fluid pressure gradients within transcrustal fault zones must be regarded as time-dependent, affecting the shear resistance profiles derived from rheological modelling. Figure 9 is an attempt to illustrate the effects on rheology and fault-strength for a transcrustal fault zone that are likely to arise from fluid pressure cycling associated with fault-valve action. Both the depth and amplitude of the peak shear resistance become time-dependent. In addition, the integrated strength of the fault zone is at a minimum pre-failure and reaches its maximum value at the end of the post-failure discharge phase, before self-sealing occurs and fluid pressure starts to reaccumulate. Some of the implications of such a model for rupture nucleation and recurrence have been explored by Sibson (1992b) Discussion While static stress fields may induce directional permeability in the rock mass, textural evidence from hydrothermal veins hosted in fault/fracture systems suggests that fluid flux through such features is commonly episodic. Moreover, it is difficult to account for the great fluid mobility in the vicinity of fault zones in terms of static stress fields and constant permeability. It appears that in many circumstances fault and fracture permeability must continually be renewed for them to remain as effective channelways. Stress cycling thus appears as the great of fluid flow in deforming upper crust, affecting flow systems driven by long-term hydraulic

modulator

CRUSTAL STRESS, FAULTING & FLUID FLOW T, °C m

FAULT ROCK METAMORPHISM

0 m

FAULT STRENGTH

0

100

200

I

I

300 MPa

FLUID PRESSURE

0

100 ~,~. I

I

200

300

400

I

I

,

~ \.~~..~~

Hydrothermal Alteration -- 100

!:i:!:i:i:i:!q:i:i:i:!:!:! Zeolite

81

.

.

.

z, km

500MPa I

0

-

hydrostatic regime

.

s -

Facies

--200 !:i:i:i:i:i:!:i:i:i:i:!:i:!

\X ~

PrehnitePumpellyite Facies

_.o --

400

iiiiiiii!iiiiiiiill iiiiiiii Greenschist Facies

_,oo ii ililiiiiiiiiiiiiilil

x

~

~

stfailure

\

\

.

\ \ prefailure

suprahydrostatic regime

lO

-

,.. .~'v < o 9• ~ ~u.,,< !

~f

-!

pos

~fithostatic regime

20-

( ~'v > 0.9 ) Amphibolite -- 600

Facies

\

25-

\ Fig. 9. General fault zone model incorporating fluid pressure cycling, assuming a linear geotherm of 25° C km-1. The extent of the different fluid pressure regimes and their relationship to metamorphic environment is conjectural. potentials such as topography, intrusive heat sources, metamorphic dewatering, etc. A considerable range of mechanisms tied to the earthquake stress cycle may contribute to fluid redistribution around fault zones. 'Suction pump' effects at dilational jogs and bends have characteristic structural associations, as have the vein assemblages that result from 'extreme fault-valve action'. However, with the exception of hydrofracture dilatancy linked to extreme valve action, the relative importance for largescale fluid redistribution of the various pumping mechanisms involving pre-failure dilatancy and post-failure fluid expulsion remains uncertain. From consideration of hydrothermal veining at the scale of individual outcrops, it may be difficult or impossible to discriminate between the different mechanisms for fluid redistribution. In evaluating the possibilities, critical account must be taken of the structural site, the mode of faulting at the time of hydrothermal deposition, the level of exposure, and the evidence for a predominantly hydrostatic or suprahydrostatic fluid pressure regime. Similar difficulties arise when attempting to attribute post-seismic discharge at the ground surface to a particular fluid redistribution mechanism (e.g. Sibson 1981; Rojstaczer & Wolf 1992; Muir Wood & King 1993). Systematic monitoring of discharge chemistry over extended time periods

may help to distinguish varying depths of origin for the discharging fluids, and allow discrimination between the different mechanisms. It is, however', becoming increasingly evident that fluid redistribution tied to the earthquake stress cycle has application to the development of fault-hosted mineralization and perhaps also to the migration of hydrocarbons in certain tectonic settings (e.g. Burley et al. 1989). Gold-quartz veins, especially, record the episodic passage of substantial volumes of aqueous fluid through fault zones and allied fracture networks (Cox et al. 1991; Boullier & Robert 1992). Regions near the top and the bottom of the seismogenic zone are favoured sites for the development of fault-hosted mineralization, with fluctuations in stress and permeability acting as triggers for episodes of hydrothermal precipitation at different stages of the fault loading cycle (Sibson 1992a). Rapid slip transfer across upwelling hydrothermal systems hosted within dilational jogs and bends in extensional and transtensional fault systems leads to abrupt pressure reductions, triggering episodes of boiling within the epithermal environment and mineral deposition throughout the phase of aftershock activity. Disseminated epithermal mineralization may also develop from the to-and-fro passage of fluids driven by cyclic loading within intensely fractured portions of

82

R.H. SIBSON

the uppermost crust. At deeper levels in compressional/transpressional fault systems, intense valving action towards the base of the seismogenic zone gives rise to high-amplitude fluid pressure cycling and the formation of mesotherreal gold-quartz lodes. There is, therefore, mounting evidence for mechanical involvement of fluids with all stages of the earthquake cycle. In the lower half of the seismogenic zone, the competition between creation and destruction of fracture permeability plays a critical role in the accumulation of fluid overpressures, affecting earthquake nucleation and recurrence, while transient fluid pressure reductions at dilational jogs and bends contribute to rupture arrest and aftershock activity (Sibson 1992b). The cyclical character of fluid pressure gradients arising from fault-valve action on seismically active faults has the implication that rheological models of fault zones and crustal shear strength profiles must also be considered time-dependent, with integrated shear strength at a minimum pre-failure, but attaining a maximum value postfailure at the end of the discharge phase. Many of the ideas expressed here have arisen through interaction and conversations over many years with N. Brown, J. Moore, H. Poulsen, N. Price, and F. Robert, while S. Cox, I. Main, J.-R. Grasso and an anonymous reviewer provided constructive criticism of the manuscript. I thank the organizers of Geofluids '93 for the financial support which made my participation in the conference possible.

References ANDERSON,E.M. 1951. The Dynamics of Faulting and Dyke Formation with Application to Britain. 2nd edn, Oliver & Boyd, Edinburgh. AVDIN, A. & JOHNSON, A.M. 1978. Development of faults as zones of deformation bands and as slip surfaces in sandstones. Pure and Applied Geophysics, 116,932-942. BATZLE, M.L. & SIMMONS, G. 1977. Geothermal systems: rocks, fluids, fractures. American Geophysical Union Monograph, 20,233-242. BETHKE, C.M. 1986. Hydrologic constraints on the genesis of the Upper Mississippi Valley mineral district from Illinois basin brines. Economic Geology, 81,233-249. BORRAIDALE, G.J. 1981. Particulate flow and the generation of cleavage. Tectonophysics, 72,306321. BOULLIER, A-M. & ROBERT, F. 1992. Paleoseismic events recorded in Archean gold-quartz vein networks, Val d'Or, Abitibi, Quebec. Journal of Structural Geology, 14, 161-179. BRACE, W.F. 1980. Permeability of crystalline and argillaceous rocks. International Journal of Rock Mechanics and Mining Science, 17,241-251.

1984. Permeability of crystalline rocks: new in situ measurements. Journal of Geophysical Research, 89, 4327-4330. BRANTLEV, S. 1992. The effect of fluid chemistry on quartz microcrack lifetimes. Earth and Planetary Science Letters, 113,145-156. BURLEY,S.D., MULLIS,J. & MATTER,A. 1989. Timing diagenesis in the Tartan Reservoir (UK North Sea): constraints from combined cathodoluminescence microscopy and fluid inclusion studies. Marine and Petroleum Geology, 6, 98-120. BYERLEE, J.D. 1978. Friction of rocks. Pure and Applied Geophysics, 116,615--626. 1990. Friction, overpressure and fault-normal compression. Geophysical Research Letters, 17, 2109-2112. -1993. Model for episodic flow of high-pressure water in fault zones before earthquakes. Geology, 21,303-306. CARRIO-SCHAFFHAUSER, E., RAYNAUD, S., LATIERE, H.J. & MAZEROLLE,F. 1990. Propagation and localisation of stylolites in limestones. In: KNIPE, R.J. & RUTTER,E.H. (eds) Deformation Mechanisms, Rheology and Tectonics. Geological Society, London, Special Publications, 54, 193-199. CATHLES,L.M. 1990. Scale and effects of fluid flow in the upper crust. Science, 248,323-329. CELLO,G. & NUR, A. 1988. Emplacement of foreland thrust systems. Tectonics, 7,261-271. CHESTER, F.M., EVANS, J.P. & BIEGEL, R.L. 1993. Internal structure and weakening mechanisms of the San Andreas Fault. Journal of Geophysical Research, 98,771-786. Cox, S.F. & Etheridge, M.A. 1989. Coupled grainscale dilatancy and mass transfer during deformation at high fluid pressures: examples from Mount Lyell, Tasmania. Journal of Structural Geology, 11,147-162. ~, WALL,V.J., ETHERIDGE,M.A. & POTTER,T.F. 1991. Deformational and metamorphic processes in the formation of mesothermal vein-hosted gold deposits - examples from the Lachlan Fold Belt in central Victoria, Australia. Ore Geology Reviews, 6,391-423. CRAMPIN, S., EVANS, R. & ATKINSON, B.K. 1984. Earthquake prediction: a new physical basis. Geophysical Journal of the Royal Astronomical Society, 76, 147-156. DAS, S. & SCHOLZ, C.H. 1981. Theory of timedependent rupture in the earth. Journal of Geophysical Research, 86, 6038-6051. Du ROUCHET, J. 1981. Stress fields, a key to oil migration. American Association of Petroleum Geologists Bulletin, 65, 74-85. ETHERIDGE,M.A., WALL,V.J., Cox, S.F. & VERNON, R.H. 1984. High fluid pressures during regional metamorphism and deformation:implicationsfor mass transport and deformation mechanisms. Journal of Geophysical Research, 89, 4344-4358. FISCHER, G.J. & PATERSON, M.S. 1989. Dilatancy during rock deformation at high temperatures and pressures. Journal of Geophysical Research, 94, 17,607-17,617. FLETCHER, R.C. & POLLARD, D.D. 1981. Anticrack

CRUSTAL STRESS, FAULTING & FLUID FLOW model for pressure solution surfaces. Geology, 9, 419-424. FYFE,W.S. 1990. Plate tectonics and the hydrosphere. In: NESBITT, B.E. (ed.) Fluids in Tectonically Active Regimes of the Continental Crust. Mineralogical Association of Canada Short Course Handbooks, 18,299-312. , PRICE, N.J. & THOMPSON, A.B. 1978. Fluids in the Earth's Crust, Elsevier, Amsterdam. GOLD, T. & SOTER,S. 1984. Fluid ascent through the solid lithosphere and its relation to earthquakes. Pure and Applied Geophysics, 122,492-529. GRINDLEY,G.W. & BROWNE,P.R.L. 1976. Structural and hydrological factors controlling the permeability of some hot-water geothermal fields. In: Proceedings of the 2nd United Nations Symposium on Development and Use of Geothermal Resources. New York, 377-386. GROSHONG,R.H. 1988. Low-temperature deformation mechanisms and their interpretation. Geological Society of America Bulletin, 100, 1329-1360. HICKMAN,S.H. 1991. Stress in the lithosphere and the strength of active faults. Review of Geophysics Supplement, 29,759-775. HILL, D.P. 1977. A model for earthquake swarms. Journal of Geophysical Research, 82, 1347-1352. - - , REASENBERG,P.A. & 29 OTHERS1993. Seismicity remotely triggered by the magnitude 7.3 Landers earthquake. Science, 260, 1617-1623. HoBas, B.E., MULHAUS, H-B. & ORD, A. 1990. Instability, softening, and localization of deformation. In: KNIPE, R.J. & RUTrER, E.H. (eds) Deformation Mechanisms, Rheology and Tectonics, Geological Society, London, Special Publications, 54, 143-166. HOOPER, E.C.D. 1991. Fluid migration along growth faults in compacting sediments. Journal of Petroleum Geology, 14, 161-180. HUBBERT, M.K. & RUBEY, W.W. 1959. Role of fluid pressure in the mechanics of overthrust faulting. Geological Society of America Bulletin, 70, 115205. HULIN, C.D. 1925. Structural control of ore deposits. Economic Geology, 24, 15-49. HUNT, J.M. 1990. Generation and migration of petroleum from abnormally pressured fluid compartments. American Association of Petroleum Geologists Bulletin, 74, 1-12. IRWIN, W.P. & BARNES,I. 1980. Tectonic relations of carbon dioxide discharges and earthquakes. Journal of Geophysical Research, 8 5 , 3115-3121. KERRICH, R. 1986. Fluid infiltration into fault zones: chemical, isotopic, and mechanical effects. Pure and Applied Geophysics, 124,226-268. KNIPE, R.J. 1989. Deformation mechanisms - recognition from natural tectonites. Journal of Structural Geology, 11,127-146. KNOPF, A. 1929. The Mother Lode System of California. US Geological Survey Professional Paper, 157. LACHENBRUCH,A.H. & MCGARR, A. 1990. In: WALLACE, R.E. (ed.) The San Andreas Fault System, California. Stress and Heat Flow. US Geological Survey Professional Paper, 1515, 261-277.

83

MCCAIG, A.M. 1988. Deep fluid circulation in fault zones. Geology, 16, 867-870. McCLAY, K. 1977. Pressure solution and Coble creep in rocks. Journal of the Geological Society, London, 134, 57-70. MARONE, C., RALEIGH, C.B. & SCHOLZ, C.H. 1990. Frictional behavior and constitutive modeling of simulated fault gouge. Journal of Geophysical Research, 95, 7007-7025. MORROW,C., LOCKNER,D., MOORE,D. & BYERLEE,J. 1981. Permeability of granite in a temperature gradient. Journal of Geophysical Research, 86, 3002-3008. MUIR WOOD, R. & KING, G.C.P. 1993. Hydrological signatures of earthquake strain. Journal of Geophysical Research, 98, 22035-22068. MOUNT, V.S. & SUPPE,J. 1987. State of stress near the San Andreas fault: implications for wrench tectonics. Geology, 15, 1143-1146. MULLIS, J. 1988. Rapid subsidence and upthrusting in the Northern Appenines, deduced by fluid inclusion studies in quartz crystals from Porretta Terme. Schweizerische Mineralogische und Petrographische Mitteilungen, 68,157-170. NEWHOUSE, W.H. 1942. Ore Deposits as Related to Structural Features. Princeton University Press, Princeton, New Jersey. NICHOLSON, C. & WESSON, R.L. 1990. Earthquake hazard associated with deep well injection - a report to the U.S. Environmental Protection Agency. US Geological Survey Bulletin, 1951. NORTON, D.L. 1982. Fluid and heat transport phenomena typical of copper-bearing pluton environments, south-eastern Arizona. In: TrrLEY, S.R. (ed.) Advances in Geology of the Porphyry Copper Deposits, Southwestern North America. University of Arizona Press, Tucson Arizona, 59-72. NUR, A. 1972. Dilatancy, pore fluids, and premonitory variations of t/tp travel times. Bulletin of the Seismological Society of America, 62, 1217-1222. - 1975. A note on the constitutive law for dilatancy. Pure and Applied Geophysics, 133,197-206. - & BOOKER, J.R. 1972. Aftershocks caused by pore-fluid flow? Science, 175,885-887. & WALDER,J. 1990. Time-dependent hydraulics of the Earth's crust. In: The Role of Fluids in Crustal Processes. National Academy of Sciences, Washington, DC, 113-127. OLIVER, J. 1986. Fluids expelled tectonically from orogenic belts: their role in hydrocarbon migration and other geologic phenomena. Geology, 14, 99-102. PARRY, W.T. & BRUHN, R.L. 1990. Fluid pressure transients on seismogenic normal faults. Tectonophysics, 179,335-344. POWER, W.L., TULLIS, T.E., BROWN, S., BOITNOTT, G.N. & SCHOLZ,C.H. 1987. Roughness of natural fault surfaces. Geophysical Research Letters, 14, 29-32. POWLEY, D.E. 1990. Pressures and hydrogeology in petroleum basins. Earth Science Reviews, 29, 215-226. RAMSAY,J.G. 1980. The crack-seal mechanism of rock deformation. Nature, 284,135-139.

84

R.H. SIBSON REDWINE, L. 1981. Hypothesis combining dilation, natural hydraulic fracturing, and dolomitization to explain petroleum reservoirs in Monterey Shale, Santa Maria area, California. In: GARRISON, R.E., DOUGLAS, R.G., PISOA~O, K.E., ISAACS, C.M. & INGLE, J.C. (eds) The Monterey

Formation and Related Siliceous Rocks of California. Special Publications of the Society of Economic Paleontologists and Mineralogists, 221-248. RICE, J.R. 1992. Fault stress states, pore pressure distributions, and the weakness of the San Andreas fault. In: EVANS, B. & WONG, T-F. (eds)

Fault Mechanics and Transport Properties of Rocks. Academic Press, New York, 475-503. ~& CLEARY,M.P. 1976. Some basic stress diffusion solutions for fluid-saturated elastic porous media with compressible constituents. Reviews of Geophysics and Space Physics, 14,227-241. ROBERT, F. & KELLY,W.C. 1987. Ore-forming fluids in Archean gold-bearing quartz veins at the Sigma Mine, Abitibi greenstone belt, Quebec, Canada. Economic Geology, 82, 56--74. ROBERTS, G. 1991. Structural controls on fluid migration through the Rencurel thrust zone, Vercors, French Sub-Alpine Chains. In: ENGLAND, W.T. & FLEET, A.J. (eds) Petroleum Migration. Geological Society, London, Special Publications, 59,245-262. ROJSTACZER, S. & WOLF, S. 1992. Permeability changes associated with large earthquakes: an example from Loma Prieta, California. Geology, 2 0 , 211-214. RUDNICKI, J . W . & H s u , T-C. 1988. Pore pressure changes induced by slip on permeable and impermeable faults. Journal of Geophysical Research, 93, 3275-3285. SCHOLZ, C.H., SYKES,L.R. & AGGARWAL,Y.P. 1973. Earthquake prediction: a physical basis. Science, 181,803-810. SECOR, D.T. 1965. Role of fluid pressure in jointing. American Journal of Science, 263,633-646. -& POLLARD, D.D. 1975. On the stability of open hydraulic fractures in the earth's crust. Geophysical Research Letters, 2,510-513. SEGALL, P. & POLLARD, D.D. 1980. Mechanics of discontinuous faults. Journal of Geophysical Research, 85, 4337-4350. SIBSON, R.H. 1974. Frictional constraints on thrust, wrench and normal faults. Nature, 249, 542544. -1981. Fluid flow accompanying faulting: field evidence and models. In: SIMPSON, D.W. & RICHARDS, P.G. (eds) Earthquake Prediction: an International Review. American Geophysical Union, Maurice Ewing Series, 4, 593-603. 1983. Continental fault structure and the shallow earthquake source. Journal of the Geological Society, London, 140,741-767. - 1985. Stopping of earthquake ruptures at dilational fault jogs. Nature, 316,248-251. 1987. Earthquake rupturing as a hydrothermal mineralising agent. Geology, 15,701-704.

1989. Earthquake faulting as a structural process.

Journal of Structural Geology, 11, 1-14. 1990a. Rupture nucleation on unfavorably oriented faults. Bulletin of the Seismological Society of America, 80, 1580-1604. 1990b. Conditions for fault-valve behaviour. In: KNIPE, R.J. & RUTTER, E.H. (eds) Deformation Mechanisms, Rheology and Tectonics. Geological Society, London, Special Publications, 54, 15-28. 1991. Loading of faults to failure. Bulletin of the Seismological Society of America, 81, 2493-2497. 1992a. Earthquake faulting, induced fluid flow, and fault-hosted gold-quartz mineralization. In: BARTHOLOMEW, M.J., HYNDMAN, D.W., MOGK, D.W. & MASON, R. (eds) Basement Tectonics 8:

Characterization and Comparison of Ancient and Mesozoic Continental Margins. Proceedings of the 8th International Conference on Basement Tectonics, Butte, Montana, Kluwer, Dordrecht, 603-614. 1992b. Implications of fault-valve behaviour for rupture nucleation and recurrence. Tectonophysics, 21 l, 283-293. 1993. Load-strengthening versus load-weakening faulting. Journal of Structural Geology, 15, 123-128. ~, MOORE, J.McM. & RANKIN, A.H. 1975. Seismic pumping - a hydrothermal fluid transport mechanism. Journal of the Geological Society, London, 131,653-659. ~, ROBERT, F. & POULSEN, K.H. 1988. High-angle reverse faults, fluid pressure cycling and mesothermal gold-quartz deposits. Geology, 16, 551555. SLEEP, N.H. & BLANPtED, M.L. 1992. Creep, compaction, and the weak rheology of major faults. Nature, 359,687-692. SMITH, D.A. 1980. Sealing and non-sealing faults in Louisiana Gulf Coast Salt Basin. American Association of Petroleum Geologists Bulletin, 64, 145-172. SMITH, D.L.& EVANS, B. 1984. Diffusional crack healing in quartz. Journal of Geophysical Research, 89, 4125-4135. SPRUNT, E.S. & NUR, A. 1977. Destruction of porosity through pressure solution. Geophysics, 42, 726741. STARK, C.P. & STARK,J.A. 1991. Seismic fluids and percolation theory. Journal of Geophysical Research, 96, 8417-8426. TORGERSON, T. 1990. Crustal-scale fluid transport: magnitude and mechanisms. LOS, Transactions of the American Geophysical Union, 71, l, 4, 13. 1991. Crustal fluid flow: continuous or episodic.

LOS, Transactions of the American Geophysical Union, 72, 18--19. WALDER,J. & NUR, A. 1984. Porosity reduction and crustal pore pressure development. Journal of Geophysical Research, 89, 11,539-11,548. WOOD, S.H., WURTS, C., BALLENGER, N., SHALEEN, M. & TOTORICA,D. 1985. The Borah Peak, Idaho, earthquake of October 28, 1983: hydrologic effects. Earthquake Spectra, 2,125-150.

Earthquakes, strain-cycling and the mobilization of fluids R. M U I R W O O D

E Q E , Newbridge House, Clapton, Gloucestershire GL54 2 L G , UK Abstract: Employing empirical observations of the hydrological changes that follow major

earthquakes, it becomes possible to predict subsurface fluid flows that accompany tectonic activity. Coseismic hydrological changes are found to be critically dependent on the style of fault displacement. Normal faults involve post-seismic compressional elastic rebound and displace large volumes of fluid from the crust: rainfall equivalent discharges have been found to exceed 100 mm close to the fault and remain above 10 mm at distances greater than 50 km; the total volume of water released in two M7 normal fault earthquakes in western USA was 0.3-0.5 km 3. In contrast, reverse fault earthquakes involve extensional elastic rebound that increases crustal porosity, drawing fluids into the crust. The magnitude and distribution of the water-discharge for normal fault earthquakes has been compared with deformation models calibrated from seismic and geodetic information, and found to correlate with the crustal volume strain down to a depth of up to 5 km. These results suggest that fluid-filled cracks are ubiquitous throughout the brittle continental crust, and that these cracks open and close through the earthquake cycle. By repeatedly drawing fluids into microcracks distributed throughout a very large volume of rock, strain-cycling can achieve very significant chemical fractionation. Different classes of tectonics impose particular configurations of fractures and according to the origin of fluid recharge, lead to a range of different types and styles of mineralization. Seismic strain-cycling can also produce primary and secondary oil migration within a single causative mechanism.

A study has been undertaken to collect, and where possible quantify, hydrological changes accompanying numerous (more than 200) earthquakes in many different parts of the world (Muir Wood & King 1993). In order to resolve such changes it is necessary to focus on areas where there is hydrogeological communication from the surface to depth. Such communication is most obvious where there are high permeability deep-sourced conduits, releasing thermal springs, but has also been demonstrated more widely from certain types of reservoirinduced seismicity in which earthquakes are triggered by a raised fluid pressure at the surface (Simpson et al. 1988). (Focusing on areas in which there is no barrier between surface water and fracture flow deep in the upper crust inevitably disqualifies the collection of observations from most sedimentary basins, and other platform areas overlain by low-permeability sediments.) Within this study it was found that once coseismic hydrological effects could be isolated from an inevitable groundwater 'noise', a simple relationship could be demonstrated between the style of fault displacement and its hydrological 'signature' (Muir Wood & King 1993). In most igneous and metamorphic rock, as well as well-lithified sediments, mobile water is held

and transported in fractures. The response of these fracture systems to the coseismic strain field of different classes of earthquakes indicates that porosity is responding to the prevailing stress-state, as is implied by numerous experimental studies (Walsh 1965; Batzle et al. 1980). In a region undergoing extensional faulting, continuing strain distributed through the crust causes appropriately oriented high-angle fractures to dilate, thereby increasing crustal porosity. Pore-pressures are sustained by slow infilling of the dilated fractures with fluid (Fig. 1). At the time of a major normal fault rupture strain, formerly distributed through the crust, becomes concentrated on the fault. As a result the region of the crust that 'surrounds' the fault (ie. the footwall and hanging wall) undergoes elastic rebound in compression. In contrast, in a region undergoing compressional tectonic deformation, in the interseismic period negative strain (volume decrease) in the crust closes high-angle fractures and reduces crustal porosity. At the time of the fault rupture, as strain is transferred into fault displacement the surrounding crust undergoes elastic rebound in extension. Hence in the region around a normal fault rupture the coseismic decrease in crustal porosity should lead to the postseismic expulsion of

From PARNELL,J. (ed.), 1994, Geofluids: Origin, Migrationand Evolution of Fluidsin Sedimentary Basins, Geological Society Special Publication No. 78, 85-98.

85

86

R. MUIR WOOD

NORMAL FAULTING !

ti,, i.......

I

b)

t

Interseismic extension

Coseismic compressional elastic rebound

REVERSE FAULTING

d)

c)

Interseismic compression

Coseismic extensional elastic rebound

Fig. 1. For extensional faulting the interseismic period (a) is associated with crack opening and increase of effective porosity. At the time of the earthquake (b) cracks close and water is expelled. For compressional faulting the interseismic period (c) is associated with crack closure and the expulsion of water. At the time of the earthquake (d) cracks open and water is drawn in.

water. In the region surrounding a compressional fault rupture expanded porosity reduces hydraulic pressures leading to water being drawn into the crust. In both normal and reverse fault earthquakes the impact on the height of the water-table (and consequent well-levels and superficially sourced spring flows) will be dependent on how effectively this water-table is connected to the fracture-flow system at depth. River flows provide the most important resource for quantifying these post-seismic changes; by sampling changes in discharge, unrelated to precipitation, across catchments typically >100 km 2 in area, it becomes possible to average the crust on a scale whose lateral dimension is of the same order as the hydraulically conductive thickness of the crust. River discharge data can then be normalised for the area of the drainage basin to achieve a 'rainfall equivalent' discharge,

either in the form of velocity (for a daily average) or cumulative linear 'thickness'. N o r m a l v. r e v e r s e f a u l t s

For the purposes of illustration the form of these hydrological changes can be seen from the contrast between two major, surface-rupturing, high-angle dip-slip earthquakes: the 1959 M7.3 (normal) Hebgen Lake (Montana) earthquake and the 1896 M7.2 (reverse) Rikuu (North Honshu, Japan) earthquake. Many other comparable examples of normal and reverse fault earthquake hydrological effects, as well as strike-slip, oblique slip and subsurface faultrupturing events can be found discussed and mapped in Muir Wood & King (1993). (a) Three river flow profiles (with entirely separate catchments) are shown in Fig. 2 from

MOBILIZATION OF FLUIDS IN EARTHQUAKES

18 Madison river 16

m3per

rain

14 12 10 September 19.59

August

: ~ier~

river

August

30

O c l ~ Days

r~n

September 1959

l

October Days

Gailatin river

1 o I,.............................. August

, ............................. September 1959

,.................... October

1

EVays

Fig. 2. River flow data for three rivers in the vicinity of the 17 August 1959 Hebgen Lake, Montana earthquake, in the days around the time of the earthquake. Average monthly flows plotted as thin line.

87

discharges, are shown in Fig. 3. By summing the individual catchments the cumulative volume of water discharged in the region around the earthquake is equivalent to c. 0.5 km 3. (b) The 31 August 1896 Rikuu (N. Honshu) earthquake involved reverse fault surfacerupture over a distance of 36kin, with a maximum uplift of 3.5 m. This was accompanied by an antithetic reverse fault rupture located some 15 km to the east and in the hanging wall of the main fault. Hence all the near-surface faulting appears to have been reverse, implying that all near-surface elastic rebound was extensional. No changes in river flows were reported following the earthquake, but hot springs supplying bath-houses at Oshuku, Tsunagi and Osawa dried up after the earthquake, while there was a significant reduction in flow at the thermal springs at Namari and Yuda (Yamasaki 1900 and see Fig. 4). All of these lie in the hanging wall of the main fault, to the east and within 20 km of the principal antithetic fault. All the hot springs eventually re-appeared, indicating that the water-flow was temporarily recharging the additional porosity in the crust that followed the earthquake. In contrast, close to the northern end of the main fault, between the main fault and its antithetic fault, a permanent new hot spring formed at Sengan-Toge following the earthquake.

Coseismic strain models the region of southeast Montana for the period before and after the 18 August 1959 Hebgen Lake earthquake (data from USGS Water Supply Reports and Stermitz 1964). For both sets of plots average flows (observed over a period of 15 years surrounding the year following the earthquake) are shown for comparison. From the daily plots it can be seen that in all three rivers a peak in flow arrived within four days following the earthquake. There was no precipitation around this time to explain such a surge and the increase in flow that follows the earthquake was sustained relative to the trend of the monthly averaged flow curve for more than 200 days. From expected flow curves for these rivers (average flows calibrated from rivers within the same region but outside the influence of the earthquake) it is possible to assess the overall volume of the water release in each catchment. This appears to show an almost linear decay following the initial peak (Muir Wood & King 1993). Typical decay times to half peak flow are 100-150 days. These cumulative flows, normalised for the individual drainage basins around the fault into rainfall equivalent

In order to examine in more detail the expected magnitude and extent of hydrological effects, strain models of coseismic deformation have been generated using a boundary element program in which a dislocation element is introduced into an elastic medium (for details of the model see King & Ellis 1990). Figure 5 is an illustration of the dilational strain changes that accompany a normal fault displacement, both in cross-section (Fig. 5a) and plan-view at two depths in the crust (Fig 5b and c). The predominant strain changes are compressional. Strain changes of reverse fault earthquakes are largely equivalent although of opposite sign to those of normal fault earthquakes. Models for simple dip-slip earthquakes always reveal 'strain-shadows' of opposite sign beyond the ends of the fault (see Fig. 5b and c). There is some evidence from both normal and reverse fault earthquakes for hydrological effects reflecting such opposite signed strain effects in these locations (Muir Wood & King 1993). The creation of the new hot-spring at the end of the Rikuu earthquake fault corresponds with a

88

R. MUIR WOOD

cumulative 'rainfallequivalent'discharge < l mm

~

~

1-20mm

~

fault scarp ~ ' - River

20-40mm

~

[

I >40mm

catchment

Fig. 3. Cumulative excess flow following the Hebgen Lake earthquake normalised for the area of the catchments into excess rainfall equivalent (ram).

strain-change anticipated to cause an expulsion of water.

Near-fault strain The above account concerns regional, or 'farfield', coseismic strain. However, close to the fault, irregularities in fault geometry, and variations in displacement, begin to dominate the strain-field. The most notable effects accompany jogs where displacement passes from one fault plane to a parallel one (see Sibson 1986). As jogs can have a dimension of metres to kilometres, and for major earthquakes involve the transfer of several metres of displacement, self-evidently they can be subject to very significant levels of strain; potentially higher

than 10 -2. This strain can be either compressional or dilational. Within the jog the coseismic stress changes do not simply accompany elastic rebound but instead are loaded directly by adjacent fault rupture. Along some strike-slip and oblique-slip fault ruptures, where jogs communicate with the surface, these effects may cause fountain outbursts of water (as in the 28 October 1983 earthquake at Borah Peak, Idaho; Waag & Lane 1985) or a collapse of the water-table (as along parts of the Fairview Valley, Nevada fault in the 16 December 1954 earthquake; Zones 1957).

The properties of crustal drainage In the case of the Hebgen Lake earthquake the

MOBILIZATION OF FLUIDS IN EARTHQUAKES

89

se°g o ~'~ k

T

gi"

~-JOshuku ~ ,

39°30 N ~~~Yuda ~ .

~'-X~sa w ' ~ -~_- ~ '

40km',.. 141 E O

m,

1

New hot spring Decrease in flow of hot spring

O Hot spring ceased flowing

river J fault-scarp

Fig.4.

Impact on hot-springs of the 31 August 1896 Rikuu (North Honshu) M7.2 reverse fault earthquake. Barbed lines are surface-fault reverse ruptures; barbs on upthrown side of the fault.

observed discharge for a traverse perpendicular to the strike of the principal fault rupture has been compared (see Muir Wood & King 1993) with predicted discharge for a two dimensional coseismic strain model, calibrated from known parameters of fault displacement, dip and depth for the earthquake. The closest fit is with the 5 km depth prediction both in general shape and in amplitude (Fig. 6). This comparison suggests that the water that emerges at the surface is associated with fracture systems extending to considerable depth. From the fit with the model it would appear that either all the volumetric strain down to a depth of 5 km is accommodated by crack closure, or else fracture connectivity extends to greater depths and only part of the strain results in a reduction in porosity. The simplest way to describe the form of the observed flows is to assume that strainlocalization and drainage are accomplished in

the same series of planar uniform cracks, open at the surface and closed at some depth, subject to a pressure change over a depth range (Muir Wood & King 1993). The crack width is found to be the parameter most strongly controlling the decay of the flow v. time profile while the dead depth (depth to the top of the section of the crack undergoing a sudden strain) and the crack width together determine the rise time. Most of the observations are fit by models with crack widths of about 0.03 mm and effective dead depths of about 2 km. For such characteristic crack apertures, extending to a depth of 5 kin, cumulative rainfall equivalent discharges of 20-40mm imply crack separations less than 4-8 m.

Implicationsof the empiricalobservations These observations offer (effectively for the first time) the potential to explore the macroscale

90

R. M U I R W O O D -60.00

a)

,o® ~

-60 O0 -60.00

-30.00

0.0

-30.(~)

0.0

30.(X)

60.(X)

~iiiiiiiiiiiiiiiiiiiii!i~:::i :~....

..:.~:~:;:~:;~:,~:,:,:~:,:~ii~,iiiiiiliiiiiiiiiiiil ~

30.00

-60,00

6000

[Voi-iirne

Strains at ~ = - : - _ ~ the Surface

3000

30.00

0.0

60.00

[Strains ~ ~ [at 5km ~ ~ l ] ~

- -

-30.00

....-~it!iiiiiil

-30.o0

O0

0.0

30.00

30.00

60,00

60.00

b)

.....

c)

Fig. 5. Cross-section (a) and plan-views at the surface (b) and at a depth of 5 km (c) of volumetric strains modelled around a normal fault. The scale units are kilometres; slip for all the models is I m. Strain levels are shaded from a background grey to lighter shades (negative strain) and to darker shades (positive strain): strain-steps of 2 x 10 -5.

SO0

450 400

1959 Hebgen Lake earthquake

Total 'rainfall equivalent' discharge: mm

A

350

I

~ 10 km

predicted from

3OO 250

2oo~ 150 100 50 0

,

-50

-80

-60

, distance from faui! rupture:, km -40

-20

0

20

,

,

1

40

60

80

Fig. 6. Observed (1959-1960) and predicted cumulative rainfall-equivalent discharges for a cross-section perpendicular to the strike of the Hebgen Fault. A constant 6 m displacement has been modelled on the fault dipping at 45 ° and extending to 15 km, consistent with the 17 August 1959 earthquake. Solid lines are predicted from the areal strain summed to depths of 2, 5 and 10 km.

MOBILIZATION OF FLUIDS IN EARTHQUAKES

91

Mohr-Coulomb failure criterion

a)

Fig. 7. Stress-cyclingin extensional (1) and compressional (2) tectonic regimes: (a) pre-seismic, (b) post-seismic.

properties of crustal drainage. The fluvial discharge information ultimately constrains the minimum depth of the percolation threshold (the depth below which all fluids are trapped under lithostatic pressure), the aperture of the characteristic fractures that dominate crustal drainage, and the minimum separation of such fractures. The comparability of these parameters for earthquakes in different parts of the world suggests that they are a fundamental property of the upper continental crust undergoing extension. Hence these parameters are likely to be relevant to the conditions that prevailed when episodes of rifting were occurring early in the life of many sedimentary basins. From empirical hydrological data it is not so easy to quantify the same key parameters of crustal drainage and porosity changes associated with a region involved in active compressional tectonic deformation. As deeply circulating water generally makes up a small contribution to surface discharge, a reduction in this component is unquantifiable in river flows, although notable in thermal springs. A higher ambient horizontal stress will reduce average fracture apertures, closing a greater portion of the walls of fractures, sealing fluid flow, and thereby causing a greater spacing between open flowing fractures. As the stress cycling that accompanies reverse faulting never reduces horizontal stresses below the highest levels encountered (post-seismically) in a compressional environment (see Fig. 7), there is inevitably a lower proportion of strain accommodated in the opening and closing of fractures. This means that porosities, the change in porosity in response to a strain increment and vertical permeabilities will all be lower than in an extensional tectonic environment.

S u m m a r y of principal findings These empirical findings have a number of important implications for the understanding of tectonically induced fluid flows in the upper crust. In particular they provide a synoptic, holistic and actualistic view of the sudden changes in hydrogeology that accompany earthquakes. In the past, theories of coseismic changes in hydrogeology have almost all been based on observations collected at individual exhumed outcrops that record multiple episodes of fluid movement accompanying numerous earthquakes millions of years in the past. (i) The character of fluid mobilization is primarily determined by the tectonic style of fault rupture. Normal fault earthquakes are predominant in expelling fluids. In contrast, reverse fault earthquakes cause a sudden increase in crustal porosity, although where a fault is blind, compressional deformation in an overlying anticline may also create lesser magnitudes of fluid discharge. This contrasts with much of the literature on seismic pumping that has assumed that reverse fault earthquakes are the most significant cause of fluid expulsion (see for example Sibson 1990). Part of the confusion in the literature can be traced back to the widely quoted discharge of water that followed the oblique-slip (reverse-sinistral) Kern County, California earthquake of 1952 (Briggs & Troxell 1955). However, when mapped this post-seismic discharge is found to be concentrated at the ends of the fault and in the anticlinal hanging wall, while close to the fault-rupture wells went dry. The pattern of discharge very closely mimics the oblique-slip strain-field modelled for this style of fault displacement (Muir Wood & King 1993).

92

R. MUIR WOOD

(ii) While 'seismic pumping' is an important component of these changes, as was highlighted by the seminal paper of Sibson et al. (1975), the 'seismohydraulic' alterations in hydrogeology through the seismic strain-cycle also affect permeabilities and porosities, as well as the orientations and apertures of fractures involved in fluid flow. Most of these changes are transient but in the vicinity of the fault can also be permanent (Muir Wood 1993). (iii) The volume of displaced fluid represents a significant proportion of total coseismic strain, at least down to the depth at which open fractures are connected. This volume is large; equivalent to a giant oilfield in each Magnitude 7 normal fault earthquake. (iv) The consistency that can be found between models of coseismic strain and the pattern of water release gives confidence in being able to predict the quantity and distribution of fluid displacement for a particular size, depth and style of fault-rupture. (v) Following a major normal fault earthquake the fluid is displaced from a wide region, with a radius typically greater than the length of the fault. For a rift-bounding fault this could cover the full width of the rift. This again challenges a widely held view to be found in the literature, that fluid flow following earthquakes is largely or completely concentrated on the fault that underwent rupture (see again Sibson 1990). Faults are most likely to concentrate flow in poorly-lithified low permeability sediments (see Behrmann 1991; Knipe et al. 1991), or perhaps at great depths in the crust. However in the top few kilometres of an upper crust dominated by fracture flow, if an earthquake was described solely in terms of its hydrological effects at the surface the causative fault itself would probably not be resolved. Hydrologically the earthquake reflects the volumetric strain-field, rather than the volume-conserving process of fault rupture. (vi) These hydrological effects demonstrate the fundamental role of strain-cycling in the crust and the intimate relationship between volumetric strain and the displacement of fluids. The close match between predicted strains and the volume of displaced water is only possible if the upper crust is filled with interconnected microcracks, that represent the dynamic porosity of the rock. (vii) The close connection between strain and fracture porosity proves that a larger, although far slower, change in crustal porosity will accompany a transformation in the tectonic regime. This porosity change will be most marked in the most extreme transition in horizontal stresses between an environment of

active extension to one of active compressional deformation, or vice versa.

Tectonic controls on fracture dilation

Each tectonic environment is accompanied by a specific style of strain cycling, manifest by the opening and closing of appropriately oriented fracture systems. As a result, individual tectonic styles of deformation are likely to generate particular patterns of fracture mineralization, at different stages of the strain cycle (Fig. 8). (a) In an extensional tectonic environment high-angle fractures undergo dilation as a result of continued horizontal strain. Following fault rupture these high-angle cracks partially close, expelling fluid from the rock. The cycle then passes once again into interseismic dilation. As these fractures are located in a stress-regime of reduced ambient horizontal stresses, they are likely to be highly strain-sensitive and hence to undergo relatively large changes in aperture. (b) In a region of compressional deformation any high-angle cracks become progressively closed in the interseismic period, squeezing fluids into near-horizontal cracks (orthogonal to the minimum compressional stress direction) where, if the fluid cannot escape, pressures will be lithostatic. At the time of an earthquake extensional elastic rebound will dilate highangle fractures, releasing fluid pressures in the low angle fracture set. High-angle fractures will continue to contract through the ensuing seismic cycle. High ambient horizontal stresses mean that changes in the aperture of high angle fractures will be significantly smaller than those encountered in extensional tectonic regimes. (c) Strike-slip tectonic environments tend to be dominated by lengthy fault systems, in which individual episodes of fault rupture do not extend along the whole length of the fault. In the vicinity of such faults high-angle fractures, will be subject to alternating episodes of compressional and dilational strain according to the relative disposition of the individual episodes of fault rupture. (d) Jogs can be either compressional (antidilational) or dilational in nature, and can accompany any class of faulting (Sibson 1986). Volume strain accompanying sudden displacement across a compressional jog can raise pore pressures within fractures sufficient to cause spontaneous hydrofracture, creating new, but transient, high-aperture fracture paths through the rock. In contrast, sudden displacement across dilational jogs will create large voids and can cause such a significant reduction in pore

MOBILIZATION OF FLUIDS IN EARTHQUAKES

Inter-seismic

Post-seismic

~

t -'l~ a)

93

pattern of high-angle and low-angle fractures (Fig. 8b) encountered in some gold deposits (see Boullier & Robert 1992) has been formed in a single compressional tectonic environment as also were the similarly oriented (but far shallower) bitumen vein deposits encountered in the Argentinian Andes (Carey & Parnell 1993). Implosion fault breccias, developed in dilational jogs on reverse faults (Fig. 8d), are another important source of gold mineralization (see for example Liu 1990).

Seismohydrauliccontrols of mineralization

iI iI

b)

c)

Fig.8.

Views in section of idealized fracture dilation accompanying seismic strain-cycling in a range of tectonic environments: (a) extensional, (b) compressional, (c) dilational jog, (d) compressional jog. fluid pressures that the rock implodes, generating some form of megabreccia (Sibson 1986). It is apparent that observations of fracture systems in many exposed rocks are describing one or other style of strain-cycling associated with active tectonics at some time in the past. Crack-seal vein infillings are found in many different tectonic environments (Ramsay 1980; Parry et al. 1988) although the relationship of mineralisation to the seismic cycle will vary from one style of nearby faulting to another. One can also predict that the magnitude of individual strain-cycles recorded by crack-seal veins will be determined by the depth of burial of the rock and the prevailing tectonic regime. Some tectonic environments can generate two orientations of fractures at different parts of the strain-cycle. For example the characteristic

Hydrothermal mineralization involves processes whereby some typically low-solubility ionic species, distributed throughout the rock at low concentrations, becomes enriched and concentrated in or around conduits for fluid flow (Sharp & Kyle 1988). Strain-cycling provides an important agent for concentrating mineralization as it repeatedly allows a small volume of water to come into contact with a large volume of rock. Strain-cycling also re-opens fractures, creating new voids in the rock that can form the site of further precipitation. The fundamental differences in strain cycling between extensional and compressional tectonic environments, as well as the most extreme strain effects in jogs, are all likely to have a profound impact on the conditions and styles of mineralization. Of course active tectonics implies the creation of new topography, and the disruption of thermal equilibrium through uplift and subsidence as well as the generation of high conductivity fracture systems, all of which will have a profound influence on fluid flows. However none of these macro-scale influences on fluid flows affects the micro-scale in the manner of strain-cycling. While the principal role of strain-cycling is in repeated fluid infiltration, post-seismically it may also promote significant fluid movements, which may be of particular consequence where thermally and gravitationally driven fluxes are weak or of opposite direction. In the environment of normal faulting, the sudden collapse in porosity that follows fault rupture causes fluid to be expelled from microcracks into arterial cracks connected to the surface. The upward displacement of fluids leads to a reduction of temperature reducing the solubility of most ionic species and thus leading to post-seismic precipitation. In a compressional tectonic environment the rock becomes infilled with fluid following an earthquake and slowly decreases in porosity between earthquakes. As a generality fluid is most likely to rise in the crust,

94

R. MUIR WOOD

and hence undergo a temperature reduction leading to precipitation, towards the end of the interseismic period. A most important determinant of the chemistry and nature of mineralization in all tectonic environments will be the source of the fluid that moves into the rock during the course of the strain cycle. If this fluid comes from deeper levels it will be saturated with dissolved ionic species and will precipitate within the rock; if it arrives from shallower levels then it will have the greater capability to dissolve and remove ionic species out of the rock. Much will depend on the particular hydrogeological context, although this in turn may be dominated by the tectonic location. For example in the subsiding hanging wall of a normal fault, where thick sequences of sediments accumulate, fluids rich in dissolved ions are produced from sediment compaction. The post-seismic reduction in fracture porosity could be the trigger for the discharge of overpressured fluids, as in low permeability sedimentary sequences on continental margins (Hunt 1990). However in the uplifted footwail, recharge is more likely to be dominated by meteoric surface recharge, low in dissolved ions and hence able to leach the rock in the course of strain cycling (Bruhn et al. 1990; Reynolds & Lister 1987). Repeated flushing of water over the 1000-10000 seismic cycles that make up the typical tectonic activity of a major fault, will dramatically alter rock chemistry, as has been noted from granites adjacent to strike-slip faults (Stierman 1984) and from deep shear-zones (Wayne & McCaig 1992). In a compressional tectonic environment, compaction of footwall sediments overridden by the fault may provide a source of fluids. Fluid is squeezed out between earthquakes, producing such phenomena as mud-volcanoes. Compressional activity and resulting uplift may also drive fluids laterally into adjacent foreland basins (Bethke & Marsak 1990). The sudden opening of high-angle fractures in normal and strike-slip faults overlying accretionary wedges can also allow overpressured fluids to escape upwards through hydrofracture in the immediate post-seismic period (Behrmann 1991). However the phenomenon of seismic valving of reverse faults (Sibson et al. 1988), whereby the fault moves in response to an increase in fluid pressures, and then acts as a conduit to fluid flow post-seismically, appears to be a process that only exists deep in the crust, as there is no empirical evidence for water coming out of reverse faults at the surface. In an environment of dip-slip faulting the principal post-seismic hydraulic gradient is near

vertical, approximately parallel to the fault. However across a strike-slip fault, in which volumes of rock subject to compression and dilation are located on opposing sides of the fault, pronounced fluid pressure gradients can develop to encourage horizontal flow through the fault.

Petroleum migration As noted above, in order to be able to map the hydrological consequences of an earthquake it is necessary to study regions in which there is continuity between water stored in fractured crustal rocks, and surface-discharging aquifers. However over a significant proportion of the continents, and in particular the continental shelves, there are extensive sedimentary cover sequences of very low permeability, which prevent, at least in the short-term, any connection between basement fluid pressures and near-surface aquifers. In such environments, if pore-pressure changes cannot be relieved through flow or recharge from the surface, then fluid flow resulting from interseismic and coseismic strain changes will be lateral, either within the fractured crust, fractured low permeability sediments or within some overlying aquifer or aquifers typically near the base of the sedimentary cover. The impact of this strain cycling will be determined by the composition of the available fluid, and the degree to which the system is interconnected with the surface. In the presence of suitable organic material the fluid will inevitably include hydrocarbons. Earthquakes have been observed to mobilize hydrocarbons as in the 12 May 1802 earthquake in north Italy, after which petroleum was procured from fissures near Bardi (Mallet & Mallet 1854). Clearly, if petroleum can be released at the surface post-seismically, it is also being remobilized subsurface. Petroleum generation demands the accumulation of organic-rich sediments and burial to sufficient depths to generate hydrocarbons. These hydrocarbons then have to be liberated from the source rock and migrate to some reservoir formation in which they become trapped and concentrated. The process of primary liberation from the rock, and secondary expulsion to some external reservoir has been much debated (England et al. 1987). The principal problem in achieving primary migration arises from the difficulty of hydrocarbons escaping from fine-grained sediments against very high inter-granular capillary pressures. Secondary migration requires that a

MOBILIZATION OF FLUIDS IN EARTHQUAKES

a)

low permeability source rock

b)

I . . ~ m

Fig. 9 (a) During positive (dilational) strain, fluids are drawn into widening microfractures accomplishing primary migration; (b) during negative (compressional) strain, hydrocarbon fluids are pumped from microfractures into high-permeability channels to accomplish secondary migration.

sufficient concentration of hydrocarbons has accumulated, to allow the fluid to become mobilized as a column. It is not difficult to see how seismic strain cycling provides a potentially important mechanism for achieving both primary and secondary migration as the result of a single process (Fig. 9). Perhaps the simplest way in which to avoid the problems of capillary resistance to primary migration is to generate new microfracture void space in the rocks in which hydrocarbons can accumulate. In well-lithified sediments such microfractures will form as a response to strain-cycling in response to active tectonics. Strain-cycling allows microcracks to open and close repeatedly, drawing hydrocarbons out of the rock matrix, and then pumping them to a larger fracture system or high permeability sedimentary units. What is the evidence for such processes? Microfractures containing either bitumen, or a diagenetic mineral-fill such as calcite, are widespread in source rocks (Duppenbecker et al. 1991). Lindgreen (1987) described diagenetic

95

mineral-bearing microfractures from within the Kimmeridge Clay of the Central Graben of the North Sea with apertures of up to 0.1 mm, oriented both parallel to the bedding and at high angles, in which primary oil migration was accomplished. The orientation of such microfractures should reveal the particular tectonic regime in which they were generated. The role of tectonics in some secondary migration in the North Sea was also demonstrated by Leonard (1984) who found that a faulted connection between source and reservoir gave a fourfold increase in the chance of exploration success in chalk reservoirs in the southern Norwegian sector of the North Sea. However, once liberated from the source-rock, strain-cycling can pump hydrocarbons through pre-existing fracture systems or any suitable high permeability sedimentary formation, not just along a neighbouring active fault. In parts of the North Sea there is also evidence for Tertiary hydrothermal activity, likely to have been tectonically driven (Wayte 1993). Pulsed episodes of fluid flow, driven by strain-cycling, are likely to have the most significant impact on the migration of brines and hydrocarbons, when tectonic activity is resumed fairly late in the evolution of a sedimentary basin. In the Tertiary history of the North Sea basin there was a prominent episode of midPalaeocene extension and late Eocene-early Oligocene compressional wrench faulting (Muir Wood 1989) concurrent with an important phase of petroleum migration (Barnard & Bastow 1991). For the latter phase of compressional (or transpressional) deformation in the rift areas perhaps 100 strain-cycles are implied by the cumulative displacements of these mid-Tertiary structures, with amplitudes typically of about 200 m (see for example Pegrum & Ljones 1984).

Conclusions

Seismic strain-cycling is akin to the function of the lungs in breathing. In order to accommodate distributed strain, interconnected fracture systems must permeate throughout the rock-mass and thereby repeatedly allow a small volume of fluid to be brought into contact with a large area of rock. It is the area of the vessels involved in the infiltration and expulsion that gives the process such enormous potency. The crudest model of 'characteristic' fractures applied to the hydrological consequences of the Hebgen Lake earthquake implies a total area of the vessels from which fluid is expelled of around 10 million km 2. Repeated fluxing of fluids through the

96

R. MUIR WOOD

crack systems that permeate the brittle rockmass has the power to fractionate and concentrate many low-solubility ionic or molecular species distributed in the upper crust. Following normal fault displacement, fluids tend to migrate away from the vicinity of the fault. For the same size earthquake occurring in a sedimentary basin the volumes of displaced fluid can be predicted to be comparable to those found emerging at the surface where there is no impermeable cover (as at Hebgen Lake, involving an estimated 0.5 km 3 of fluid migrating laterally from a region 10000 km 2 in area). The flux on the margins of such an area, entirely covered by a low permeability overburden, would be more than 1000 m 3 of fluid for every metre length of perimeter; in a 10m thick aquifer with a 10% dynamic porosity, fluid would move 1 km. Following a reverse fault, displacement fluid flow will be drawn towards the fault. On the assumption that the strain 'efficiency' (porosity change divided by volumetric strain) of a reverse fault earthquake is at least 20% that of a normal fault, an M7 earthquake would draw in perhaps 100000000 m 3 of fluid. Where hydraulic conductivities in sediments lying on the crystalline basement are significantly higher than in the bedrock itself this will encourage large scale flow with the lighter fluids becoming concentrated in the sedimentary section above the reverse fault. Blind, non-outcropping reverse faults beneath thick sedimentary basins are also some of the most efficient at creating anticlinal hydrocarbon reservoir geometries. The flows created by the large-scale and sudden changes in pore-pressures accompanying seismic strain-cycling may oppose the prevailing fluid flow regimes created by gravitational or thermally-driven fluid flows. Tectonic activity could therefore drive fluids into reservoirs that might otherwise be discounted as potential drilling targets. Employing coseismic and interseismic strain models within basin analysis it should be possible to predict the magnitude and extent of these repeated episodes of fluid flow. Coseismic and interseismic strain changes could also help to explain the development of some zones of over- and under-pressurization within confined fluid reservoirs, where other explanations associated with consolidation secondary porosity reduction or temperature changes are found to be insufficient (Domenico & Palciauskas 1988). In a medium where porosity is dominated by fractures (or connected with such a medium) sealed reservoirs have the potential to act as strain-barometers recording

changes in dynamic porosity through an alteration in fluid pressure (Bodvarsson 1970). The impact on pore-fluid pressures will be determined by the dynamic porosity relative to the strain-sensitivity of the fractured medium. As with Earth-tides the most sensitive response is found in a low dynamic-porosity medium in which the fluid is distributed in strain-sensitive fracture systems. Some confined aquifers are known to produce centimetric changes in welllevels from 10 -s strain Earth-tides (Bredehoft 1967). For a confined low dynamic-porosity medium a reduction in volume of 10-4-10 -5, consequent on seismic strain-cycling, has the potential to raise pore pressures by several bars. Where the rock experiences higher levels of strain in response to the imposition of a new tectonic regime, then an even more marked change in pore-fluid pressures is possible. For example, extreme under-pressurisation can be expected in a confined, low-porosity fractured reservoir where a compressional tectonic regime passes to one of extension. Such a process could partly explain the under-pressurization of the oil-reservoirs at Kimmeridge, Dorset, England (Selley & Stoneley 1987) where mid-Tertiary compressional deformation has passed into Plio-Quaternary extension. Strain models in this paper were produced by G. King at IPG, Strasbourg, whose collaboration is gratefully acknowledged.

References BARNARD, P. C. & BASTOW, M. A. 1991. Petroleum generation, migration, entrapment and mixing in the central and Northern North Sea. In: ENGLAND,W. A. & FLEET,A. J. (eds) Petroleum Migration. Geological Society, London, Special Publications 59,167-190. BATZLE,M. L., SIMMONS,G. & SIEGFRIED,R. W. 1980. Microcrack closure in rocks under stress: direct observation. Journal of Geophysical Research, 85, 7072-7090. BEHRMANN,J. H. 1991. Conditions for hydrofracture and the fluid permeability of accretionary wedges. Earth and Planetary Science Letters, 107, 550558. BETHKE, C. M. & MARSHAK,S. 1990. Brine migration across North America - the plate tectonics of groundwater. Annual Review Earth and Planetary Science Letters, 18,287-315. BOt)VARSSON, G. 1970. Confined fluids as strain meters, Journal of Geophysical Research, 75, 2711-2718. BOUILLIER, A-M. & ROBERT,F. 1992. Palaeoseismic events recorded in Archaean gold-quartz vein networks, Val d'Or, Abitibi, Quebec, Canada. Journal of Structural Geology, 14,161-179.

MOBILIZATION OF FLUIDS IN E A R T H Q U A K E S BREDEHOFT, J.D. 1965. Response of well-aquifer systems to earth tides. Journal of Geophysical Research, 72, 3075-3087. BRIGGS, R. C. & TROXELL, H. C. 1955. Effect of the Arvin-Tehachapi earthquake on spring and stream flow. In: Earthquakes in Kern County California, during 1952. Bulletin of the California Division of Mines, 171, 81-98. BRUHNN, R. L., YONKEE,W. A. & PARRY,W. T. 1990. Structural and fluid-chemical properties of seismogenic normal faults. Tectonophysics, 175, 139-157. CAREY, P. F. & PARNELL,J. 1993. Solid bitumen veins related to overpressuring and thrusting in the Neuquen Basin, Argentina. In: PARNELL, J., RUFFELL, A. H. & MOLES, N. R. (eds) Geofluids '93, Torquay, May 4-7, 1993, 181-185. DOMENICO, P. A. & PALCIAUSKAS,V. V. 1988. The generation and dissipation of abnormal fluid pressures in active depositional environments. In: BACK, W., ROSENSHEIN, J. S. & SEABER, P. R. (eds) Hydrogeology. Geological Society of America, The Geology of North America, 0-2, 435-445. DUPPENBECKER, S. J., DOHMEN, L. & WELTE, D. H. 1991. Numerical modelling of petroleum expulsion in two areas of the Lower Saxony Basin, Northern Germany. In: ENGLAND, W. A. & FLEET, A. J. (eds) Petroleum Migration. Geological Society, London, Special Publications, 59, 47-64. ENGLAND, W. A., MACKENZIE,A. S., MANN, D. M. & QUIGLEY, T. M. 1987. The movement and entrapment of petroleum in the subsurface. Journal of the Geological Society (London), 144, 327-347. HUNT, J. M. 1990. Generation and migration of petroleum from abnormally pressured fluid compartments. Bulletin of the American Association of Petroleum Geologists, 74, 1-12. KING, G. C. P. & ELLIS, M. 1990. The origin of large local uplift in extensional regions. Nature, 348, 689-692. KNIPE, R. J., AGAR, S. M. & PRIOR, D. J. 1991. The microstructural evolution of fluid flow paths in semi-lithified sediments from subduction complexes. Philosophical Transactions of the Royal Society of London, A355, 261-273. LEONARD, R. 1984. Generation and migration of hydrocarbons on the Southern Norwegian Shelf (Abstract). Bulletin of the American Association of Petroleum Geologists, 68,796. LINDGREEN,H. 1987. Molecular sieving and primary migration in Upper Jurassic and Cambrian claystone source rocks. In BROOKS,J. & GLENNIE, K. (eds) Petroleum Geology of North West Europe. Graham & Trotman, London, 357-364. LIu, E. 1990. The metalloctectonics in Eastern Shandong gold metallogenetic province, China. Terra Nova, 2,257-263. MALLET, R. & MALLET,J. W. 1854. Third reporton the facts of earthquake phenomena. Reports of the British Association for the Advancement of Science.

97

MUIR WOOD, R. 1989. Fifty million years of 'passive margin' deformation in North West Europe. In: GREGERSEN, S. & BASHAM, P. W. (eds) Earth-

quakes at North-Atlantic Passive Margins: Neotectonics and Postglacial Rebound. Kluwer,

-

Dordrecht, 7-36. 1993. Neohydrotectonics. Zeitschrift fur Geomorphologie, 94,275-284. & KING, G. C. P. 1993. Hydrological signatures of earthquake strain, Journal of Geophysical Research, 98, 22035-22068. PARRY, W. T., WILSON, P. N. & BRUHN, R. L. 1988. Pore-fluid chemistry and chemical reactions on the Wasatch normal fault, Utah. Geochimica et Cosmochimica Acta, 52, 2053-2063. PEGRUM, R. M. & LJONES, Z. E. 1984. 15/9 Gamma Gas Field, offshore Norway, new trap type for North Sea Basin with regional structural implications. Bulletin of the American Association of Petroleum Geologists, 68, 874-902. RAMSAY,J. G. 1980. The crack-seal mechanism of rock deformation. Nature, 284, 135-139. REYNOLDS, S. J. & LISTER, G. S. 1987. Structural aspects of fluid-rock interactions in detachment zones. Geology, 15,362-366. SELLEY,R. C. & STONELEY,R. 1987. Petroleum habitat in South Dorset. In: BROOKS, J. & GLENNIE, K. (eds) Petroleum Geology of North West Europe, Graham & Trotman, London, 139-148. SHARP, J.M. JR & KYLE, J.R. 1988. The role of ground-water processes in the formation of ore deposits. In BACK, W., ROSENSHEIN, J.S. & SEABER, P.R. (eds) Hydrogeology. Geological Society of America, The Geology of North America, 0-2,461-483. SmsoN, R. H. 1986. Brecciation processes in fault zones: inferences from earthquake rupturing. Pure and Applied Geophysics, 124, 159-173. 1990. Conditions for fault-valve behaviour. In: KNIPE, R. J. & RUTTER, E. H. (eds) Deformation Mechanisms, Rheology and Tectonics, Geological Society, London, Special Publications, 54, 15-28. ~, ROBERT, F. & POULSEN, K. H. 1988. High-angle reverse faults, fluid pressure cycling and mesothermal gold-quartz deposits. Geology, 16, 551555. SIMPSON, D. W., LEITH, W. S. & SCHOLZ,C. H. 1988. Two types of reservoir induced seismicity. Bulletin of the Seismological Society of America, 78, 2025-2040. STERMITZ, F. 1964. Effects of the Hebgen Lake earthquake on surface water. United States Geological Survey Professional Paper, 435-L, 139150. STIERMAN, D. J. 1984. Geophysical and geological evidence for fracturing, water circulation and chemical alteration in granitic rocks adjacent to major strike-slip faults. Journal of Geophysical Research, 89, 5849-5857. WAAG, C. J. & LANE, T. G. 1985. The Borah Peak, Idaho earthquake of October 28, 1983 - structural control of groundwater eruptions and sediment boil formation in the Chilly Buttes area. Earthquake Spectra, 2, 151-168. -

98

R. MUIR WOOD

WALSH, J. B. 1965. The effect of cracks on the compressibility of rocks. Journal of Geophysical Research, 70,381-389. WAYNE, D. M. • MCCAIG, A. M. 1992. Fluid flow during thrusting: a Sr isotope study. In: KHARAKA, Y. & MAEST, A. S. (eds) Water-Rock Interaction. Balkema, Rotterdam, 1559-12664. WAYTE,G. J. 1993. Evidence for the cross-formational movement of hydrothermal fluid within overpressured cells: - an example from the Triassic of the U.K. Central North Sea. In: PARNELL, J.,

RUFFELL, A. H. & MOLES, N. R. (eds) Geofluids '93, Torquay, May 4-7, 1993, 329-332. YAMASA~a, N. 1990. Das grosse japanische erdbebden im nordlichen Honshu am 31 August 1896. Petermanns Mitteilungen, 46,249-255. ZONES, C. P. 1957. Changes in hydrologic conditions in the Dixie Valley and Fairview Valley areas, Nevada, after the earthquake of December 16, 1954. Bulletin of the Seismological Society of America, 47,387-396.

Microstructural and microchemical consequences of fluid flow in deforming rocks R.J. KNIPE & A.M. McCAIG

Department o f Earth Sciences, The University, Leeds LS2 9JT, U K Abstract: The different deformation mechanisms possible in rocks impact differently on

fluid flow processes because of the range of induced volume changes associated with different deformation histories. Microstructural analysis of deformed rocks can provide crucial information for the identification of fluid flow pathways, determination of the physico-chemical properties of the fluid, quantification of the amount of fluid involved and an assessment of the variation in the open/closed nature of the fluid flow system. A critical factor in the efficiency of fluid flow in deformed rocks is the competition between the processes which maintain the connectivity of the high permeability pathways and those which close such pathways. The range of deformation processes which are involved in this competition will be different depending on the tectonic settings, the deformation conditions and the rock types involved. A brief review of the processes which interact to control fluid flow during deformation in sedimentary basins, crystalline basement under low to moderate grade metamorphism and during prograde metamorphism at moderate to high grades is given.

Identification of the pathways available for fluid flow in rocks which experience deformation, together with a detailed understanding of the interactions between fluid flow and deformation mechanisms, are important to the quantification of a number of geological processes. Recent research has highlighted that both deformation and fluid flow are likely to be localized and to occur during transient events (Sibson 1980, 1990; Carter et al. 1990; Cox et al. 1991; Sleep & Blanpied 1992; Byerlee 1993; Knipe 1993). Such complex behaviour provides an additional complication to the quantification of fluid flow processes and restricts modelling of fluid flow in rocks. The main objectives of the study of palaeofluid flow events in rocks are: • identification of the critical flow pathways along which the majority of fluid flow occurred; • determination of the physico-chemical properties of the fluid (e.g. pressure, temperature, composition and the variation of these properties with time); • characterization of the physical properties of the flow path (porosity, permeability, and time-variation of these in relation to deformation); • quantification of the amount of fluid involved (time-integrated flux) and the fluid flow rate

(flux);

•

determination of the scale characteristics of the flow event (dimensions of the volume

involved) and variations in the open/closed nature of the fluid flow system; • assessment of the relationships between fluid-induced reactions, cementation, alteration patterns and the above. There are three main types of evidence which can contribute to the characterization of palaeofluid flow events through rocks. (1) Direct observation of fluid compositions and densities from fluid inclusions (Roedder 1984; Parry & Bruhn 1986; Banks etal. 1991). (2) Physical evidence from the presence of features which require fluid flow, for example veins, cementation, stylolites and fluid inclusion trails (Ramsay 1980; Lespinasse & Cathelineau 1990). (3) Chemical evidence for flow, including vein and cement compositions, metasomatic changes in rock composition, zoning patterns of minerals and cryptic veining in recrystallised rocks (e.g. Beach 1976; Etheridge et al. 1984; Kerrich 1986; McCaig & Knipe 1990; McCaig et al. 1990). Physical indicators help to identify fluid flow pathways and are essential for assessing the mechanisms of flow, while chemical evidence is required to demonstrate that significant flow has occurred. Assessment of fluid flow characteristics and the interactions between deformation and flow can be aided by the analysis of microstructural features generated during fluid

From PARNELL,J. (ed.), 1994, Geofluids: Origin, Migration and Evolution of Fluids in Sedimentary Basins,

Geological Society Special Publication No. 78, 99-111.

99

Scanning proton microprobe (SPM)

Laser ICPMS

Ion microprobe (SIMS)

Laser microprobe stable isotope analysis

Electron microprobe

Transmission electron microscope (TEM, STEM) Cathodoluminescence (CL)

Secondary electron SEM

Backscatter scanning electron microscopy (BSEM)

Optical microscopy

Technique

Rapid in situ trace element determinations at ppm level. Small spot size (1 txm) and little beam spreading in sample. Spot analyses, linescans and element maps. Non-destructive with little sample damage.

Contrast depends on trace element variations, and/or defect structure. Gives good information in monomineralic rocks (e.g. Quartz veins in quartz). Best contrast from hot cathode devices (e.g. SEM) Major element compositional variations on a 5 ixm scale. Vein compositions, zoning patterns. Good backscatter SEM on modern instruments. In situ analysis of O, C, and S isotopes in minerals. Sensitivity good, especially for carbonates. Rapidly becoming a routine technique. Isotopic zoning has been related to fluid flow. In situ trace element and isotopic measurements in minerals. Good sensitivity in favourable circumstances. Spatial resolution 2-10 i~m. In situ trace element determinations on minerals. Rapid and cheap. Detection limits c. 0.5 ppm.

First stage of most studies. Excellent mineral phase contrast. Chemical variations in carbonates enhanced by staining. Porosity identified with coloured resin. Ultrathin sections required in fine grained rocks. Information on microstructures from birefringence contrast. Universal stage gives quantitative information on crystal fabrics, fracture orientations etc. Good mineral phase and compositional contrast. Compositional variations in isotropic minerals. Information on intragranular microstructure in orientation contrast mode, linked to lattice orientation through electron channelling. Good analytical, element mapping and image analysis facilities on modern instruments. Photographic collages allow comparison of fine detail over significant areas. High resolution of surfaces. Good for details of porosity structure and occlusion in rocks affected by diagenesis and hydrothermal processes. Fracture surfaces can be imaged, and etching can reveal microstructural details. Resolution < 1 Ixm. Lattice defects imaged. Compositional variations detectable on sub p~m scale.

Uses

Table 1. Techniques f o r microstructural and microchemical analys&

Many relevant references, e.g. Wintsch & Knipe (1983); McCaig (1984); McCaig & Knipe (1990); Selverstone et al. (1992). Sharp (1992); Elsenheimer & Valley (1992); Chamberlain & Conrad (199l); Dickson et al. (1991). Giletti & Shimizu (1989); Lyon & Turner (1992). Detection limit c. 0.1% for most elements.

Fraser (1990); Wogelius et al. (1992); Jamtveit et al. (1993).

Jackson et al. (1992).

McLaren (1991) for a general review; Knipe (1982); Hippler & Knipe (1990); Philippot & Van Roemund (1992). Burley etal. (1989); Yardley & Lloyd (1989); Lloyd & Knipe (1992).

Specimen preparation difficult. Only small areas can be imaged at once, so harder to relate to other techniques. No luminescence in many situations (e.g. Fe-rich minerals). Origin of contrast sometimes not clear.

Minimum spatial resolution 100 p~m with CO2 laser. Better with Nd-YAG laser, but some minerals cannot be analysed. Many complicating factors make analyses non-routine, and difficult for certain isotopes and matrices. Minimum pit size 20--40 I~m. Lack of homogeneous mineral standards can make calibration difficult. Expensive. Data reduction still time consuming. Fine grained samples require special preparation because of beam penetration (up to 40 p~m).

e.g. papers in Marshall (1987); Knipe (1992, 1993); Evans (1990).

Many relevant references, e.g. Ramsay (1980); McCalaig (1984); Knipe (1982); Brodie & Rutter (1985); papers in Marshall (1987); Cox & Etheridge (1989); Lespinasse & Cathelineau (1990); Evans (1990). Lloyd (1987) for review of techniques. Papers in Marsh all (1987); Prior (1988); McCaig & Knipe (1990); Agar (1990); Knipe etal. (1991), and this paper for examples of application to fluid flow.

References relevant to fluid flow

Compositional information mainly from crystal form. Difficult to relate directly to optical microscopy.

Specimen damage currently limits minerals suitable for electron channelling. Resolution limited to c. 1 ~m. Will not detect veining by same phase. (e.g. quartz in quartz).

Limited scale of observation. Composition information limited. Isotropic materials cannot be imaged under crossed polars.

Limitations

MICROSTRUCTURAL CONSEQUENCES OF FLUID FLOW flow and deformation (e.g. McCaig & Knipe 1990; Knipe et al. 1991; Knipe 1993). A variety of microstructural and microchemical techniques are now available which can be applied to the study of fluid flow during deformation (Table 1) and the success of this approach is enhanced by the ability of modern research microscopes to combine different imaging signals from the same features. New techniques such as the scanning proton microprobe and laser fluorination allow small scale isotopic and trace element variations to be detected, and are likely to be increasingly applied to fluid-rock interaction problems. This short paper reviews the different deformation mechanisms possible in rocks and highlights some of the progress made in understanding the different ways in which deformation influences fluid flow. The objective is to provide a brief overview of recent research and to direct the reader at some relevant literature.

Influence of deformation mechanisms on fluid flow In this paper we are not concerned with the large scale driving forces for fluid flow but with the identification and quantification of fluid flow influenced by deformation. The general direction of flow will always be down the hydraulic gradient, whether this results from the buoyancy associated with thermal convection, from overpressures due to sedimentary or tectonic loading, or from some form of stress-related dilatancy. Deformation can influence fluid flow in two ways. (1) Strain-related changes (increases and decreases) in the host rock porosity or pore geometry will create changes in pore pressures and the pattern of fluid pressure gradients. (2) Permeability can be enhanced or restricted by the deformation (e.g. enhanced by the opening of fractures or grain boundary voids, and restricted by pressure solution, compaction and cementation). These influences are not independent. For example, the opening of an array of fractures may cause a local reduction in the fluid pressure, generating a steep pressure gradient into the dilatation site and enhancing fluid flow potential. However, the pressure drop may also lead to precipitation of cements which will reduce the permeability (cf. Etheridge et al. 1984). In the following sections we outline the consequences of the operation of different deformation mechanisms on permeability and fluid flow and give

101

examples on how microstructural analyses can be used to study these processes. Three sections are presented below, each one corresponding to one major mode of deformation.

Fracture, cataclastic flow, and independent particulate flow Fracture, cataclastic flow and independent particulate flow are the dominant deformation mechanisms in rocks and sediments in the higher, cooler levels of the crust (Groshong 1988; Mitra 1988; Knipe 1989; Lloyd & Knipe 1992). There is little doubt that fractures dominate the permeability of large parts of the upper crust, particularly in crystalline rocks at low pressure and temperature. Excellent reviews of the fracturing processes in rocks have been given by Atkinson (1987) and Scholz (1990), and a recent collection of papers in Evans & Wong (1992) reviews some of the transport properties of fault zones. The fracture strength of materials increases with increasing effective confining pressure ( 0 - c f f = 0"3 - - PH20)while the flow stress needed to activate plastic deformation decreases with increasing temperature. Hence fracture processes generally give way to plastic flow with increasing depth (Sibson 1986; Knipe 1989). However, it is now apparent that there is a considerable depth range over which fracture and ductile flow processes are both involved in deformation events (Knipe 1989, 1990; Lloyd & Knipe 1992). This may be associated with the alternate operation of deformation mechanisms during the repetitive deformation events of fault growth. In addition, brittle fracture, which is associated with rapid fracture propagation after low strains, is only one of a large variety of fracture processes. Slower crack propagation, below the critical stress needed for brittle fracture, can occur either with the aid of a concentration of plastic deformation near the crack tip or by stress corrosion processes by fluids at crack tips (see review in Scholz 1990). In general, the deviatoric stress required for fracture propagation is reduced as the pore fluid pressure is increased, since this reduces the effective pressure. Where the pore fluid pressure exceeds the least principal stress by an amount equal to the tensile strength of the material, hydraulic fracture will occur. Although the controls on these mechanisms are known the recognition of the different fracture propagation processes from microstructural evidence remains an important goal for future natural fracture mechanics.

102

R.J. KNIPE & A.M. McCAIG

Fluid flow through a fractured rock requires both the fracture array to have a suitable geometry and some threshold conditions to be exceeded in order for the array to be linked into a connected fluid migration pathway (Guegen et al. 1987; Main et al. 1990; Odling 1992; Stark & Stark 1991). Fracture/fault populations normally appear to be fractal with respect to the number of features of different size or displacement magnitude (Cowie & Scholz 1992; Gillespie et al. 1992). However, analysis of fluid flow through a fracture array also requires knowledge of the individual geometrical characteristics of the linked pathways, as not all the fractures/faults present need form part of the linked network at the time of flow, and not all the deformation features in the finite array need have formed during the same deformation event. For these reasons, establishing the history of fracture opening and identifying the types of fluids associated with fracture arrays is an important component of the analysis of fluid flow. Microstructural studies can provide information for such analysis. Most fractures which may have contributed to fluid flow in rocks are now represented by either healed micro-fractures, identified by planar secondary fluid inclusion arrays, or by veins (Lespinasse & Cathelineau 1990; Sibson et al. 1988; Boullier & Robert 1992). Except in rare cases of replacement, veins represent a finite strain event which will have increased the local porosity of the rock. The effect on the permeability and fluid flow in the rock depends on the amount, extent and duration that the fractures are open and interconnected. Repeated fracturing, dilatation and sealing by cementation can generate large and extensive veins, but may only be associated with small individual dilatation events (see Knipe & White 1979; Ramsay 1980). Other veins contain large crystals which have grown into a fluid filled cavity of considerable size (1 cm or more). The growth and maintenance of such cavities is an interesting mechanical problem (i. e. what size of open fracture can be maintained by the inherent rock strength under the imposed stress conditions?) although the development of such cavities is favoured by high fluid pressures which hold open the fractures. The presence of veins does not necessarily require either high fluid pressure or large scale fluid flow. For example the quartz vein arrays described by Knipe & White (1979) are associated with pressure solution seams which could have provided the quartz precipitated in the veins. Although some fluid flow must have taken place there is no need to invoke large scale flow

as fluid-enhanced diffusion processes could have dominated over advection. The demonstration of large scale flow through a vein array requires geochemical evidence. For example, Wayne & McCaig (1992) described isotopic data from carbonate veins within Cretaceous carbonates and mylonites from beneath the Gavarnie Thrust in the central Pyrenees. Radiogenic fluid stored in Triassic redbeds forming the basement to the thrust system was introduced into mylonites along the thrust plane, causing a general increase in the 87Sr/86Sr ratios in the deformed carbonates away from seawater-derived values. Within both undeformed and mylonitized carbonates, veins are almost invariably enriched in 8VSrcompared with the adjacent wall rocks, demonstrating that the fluid which precipitated the vein material must have equilibrated with more SVSrenriched rocks further back along the flow path. The minimum distance for the fluid transport in this case is 1.5 km, which is much larger than the average diffusion distance. Flow must have been unidirectional, with fluid moving always from more altered to less altered mylonites, and outwards into less altered wall rocks (Wayne & McCaig 1992). Abundant deformed, recrystallized veins within the mylonites suggest that the majority of fluid flow occurred during fracture events, but that these alternated with periods of quasi-plastic flow. An example of how fractures can affect fluid movement in a mylonite is illustrated in Fig. 1, which shows part of a plagioclase porphyroclast from a mylonite from Andorra in the centraleastern Pyrenees. The mylonite was generated primarily by plastic deformation associated with dynamic recrystallization at a temperature of approximately 400 ° C. The plagioclase porphyroclast (Anlv) contains more than 100 veins of more calcic plagioclase (up to An35). Plagioclase of this composition is stable in adjacent more mafic layers in the mylonite zone (see McCaig & Knipe 1990). The veins demonstrate that fluid moved across the layering for distances of at least a few millimetres. Cross-cutting relationships demonstrate that several generations of fractures are present (Fig. lb), and some of the larger veins clearly formed from several crackseal fracture events. The vein network must represent the product of between 10 and 150 discrete fracturing events, depending on how many separate veins formed during each event. In the dynamically recrystallized matrix of the same rock, trials of isolated Ca-enriched plagioclase grains have been interpreted as the remnants of similar veins crossing the matrix (McCaig & Knipe 1990). This demonstrates two

30 ~m I

I

Fig. 1. (a) Backscattered SEM montage of a plagioclase porphyroclast (An16) from a mylonite in the central Pyrenees (cf. McCaig & Knipe 1990). The porphyroclast contains multiple veins of more Ca rich plagioclase (up to An35). Veins are not present in the surrounding matrix, which consists of fine recrystallized quartz, plagioclase and muscovite. (b) Detail of an area in the lower left of (a), showing three generations of cross-cutting veins of increasingly Na-rich composition. The final generation is a marginal zone or pressure shadow to the porphyroclast.

104

R.J. KNIPE & A.M. McCAIG

3 !~!} 2;:

~

1

Fig. 2. Backscattered SEM images from a fractured dolomitic layer within mylonites beneath the Gavarnie Thrust, central Pyrenees (McCaig et al. 1993). (a) Ankerite and calcite vein in dolomite (dark phase at top edge). Note euhedral ankerite crystals built up by multiple crack-seal events (arrows), and relict screens of dolomite within ankerite. (b) Dolomite veined by Fe-dolomite and calcite, with cavity fill texture at lower left. This is evidence for storage of fluid within fracture-controlled porosity within the dolomite layer. Darkest phase is quartz. important aspects of fluid flow and deformation in mylonites. Firstly, the deformation involves an alternation of fracture and plastic flow, and secondly, fracturing of the stronger phases can create local fluid reservoirs. Knipe et al. (1991) also highlighted the importance of the creation of transient micro-fluid reservoirs for fluid flow in deforming aggregates. In the case of a mylonite the dilatation and veining associated with the hard minerals will be controlled by; (a) the amount and distribution of the hard phase, and, (b) the ability of the matrix to absorb strain created by the fracturing of the hard phases. Higher concentrations of hard phases and lower temperatures will enhance the potential for

larger, more extensive and interconnected fracture arrays for fluid flow in the myionite. On a mesoscopic scale, more competent layers within a deformation zone can also provide a reservoir for storing fluid in fractures. Figure 2 illustrates intense veining and cavity-fill textures in a dolomitic layer within calcite mylonites along the Gavarnie Thrust (McCaig et al. 1993). Complex fluid movement patterns are indicated by a wide variety of vein compositions from calcite to ankerite. Cataclastic flow involving repeated fracturing, grain size reduction and frictional grain boundary sliding is characteristic of deformation in high level fault zones (Blenkinsop & Rutter 1986; Knipe 1989; Sleep & Blanpied 1992). The movement of grains during deformation is accommodated by fracturing triggered by stress concentrations at grain contacts, and by dilatation. Independent particulate flow dominates the deformation of poorly consolidated sediments and fault gouges at low effective confining pressures, where strain is achieved by the sliding of grains past one another (Borradaile 1981). In both these cases, most deformation is likely to be associated with short periods of high strain rate. These transient pulses of deformation will induce periods of dilatation during which both the permeability and the porosity will be enhanced; i.e. the system is more open and fluid flow rates can be high. These periods will be separated by periods of low strain-rate when the system may be effectively closed. Although the general effects and consequences of these alternations in fluid communication are known, the time periods when the system is open, the decay rate of the permeability after the events and the volume affected by deformation events of different magnitude are largely unknown (see discussion in Knipe 1993). It is also unclear what controls the relative importance of fluid flow associated with the linking of fracture/fault zone dilatation over extensive areas or, fluid flow related to the slow migration of a strain (dilatation) wave which effectively transports packets of fluid with it (see Knipe et al. 1991, Fig. 4 for discussion). An attempt to assess the permeability changes which occur during deformation (i.e. quantification of the dynamic permeability) of poorly consolidated sediments has been made by a series of recent experiments (Stephenson et al. this vol.). The paper by Stephenson et al. (this volume) demonstrates the complexity of permeability variations during deformation and again highlights the ability of a fault zone to alternately store and expel fluids. This dual behaviour is also likely on a larger scale where

MICROSTRUCTURAL CONSEQUENCES OF FLUID FLOW segments of fault zones such as dilatational jogs may provide transient storage volumes for fluids (see Sibson 1990, 1993).

Diffusive mass transfer Strain associated with the removal of material from grain boundaries and interfaces experiencing high stresses gives rise to a range of mechanisms grouped together as diffusive mass transfer (Spiers & Schjutjens 1990). The processes involved can be separated into; (a) dissolution processes at the source, (b) migration of material along some pathway, and (c) the precipitation of material at sinks. The migration pathway may be a grain boundary, a thin film of fluid at grain boundaries or a pore fluid. The complexity of these processes is reflected in the range of flow laws which have been constructed for different combinations of source, migration and sink processes (see Knipe 1989 for discussion). Diffusive mass transfer (DMT) is generally associated with a loss of porosity and permeability due either to the removal of material by dissolution or by cementation in pore space. These processes are important to the creation of fault seals or barriers which restrict fluid flow (Knipe 1992, 1993). Because the distribution of such barriers and the timing of their development are important to fluid flow, an understanding of the processes and mechanisms involved in seal evolution is also important. Knipe (1993) has emphasized the role of diffusive mass transfer as a late stage process, which can be enhanced in fault zones where grain size reduction by cataclasis has already occurred. That is, the last stages of movement in a fault zone act to close the high permeability window created by dilatation associated with faulting. The cements which arise from precipitation in sealed fault zones do hold evidence (mineralogy, zoning and isotopic composition) which can aid the identification of the source of fluids involved in any fluid flow (e.g. Burley et al. 1989; Knipe 1993). The time periods when fluid flow is enhanced after large (>M6) earthquakes can be assessed from stream discharge and are usually measured in months (Muir Wood & King 1993).

Dislocation creep Dislocation creep is an intragranular deformation mechanism which relies on the movement of crystal defects (dislocations) to create crystal plasticity. The strain rate is largely dependent on the climb of dislocations which is diffusion controlled. The mechanism dominates the deformation of the middle to lower crust and

105

is a crucial component in the generation of fine grained mylonitic fault zones (Sibson 1986; Knipe 1989). Dislocation creep is not normally associated with enhanced fluid flow because any 'voids' or fractures present at the start of the deformation will lead to stress concentrations which should lead to local straining and closure of such features. However, two aspects of dislocation creep behaviour can be important to the potential for fluid flow. (a) Fluid pressures in 'voids' or fractures during dislocation creep may be close to the confining pressure so that fractures under certain conditions may propagate relatively easily (see below). (b) Under conditions where dislocation creep does not/cannot produce all the strain rate required by the deformation conditions or if there is significant grain boundary sliding, then additional strain accommodation processes may create dilatation and so enhance fluid flow. This dilatation may take the form of 'void' generation on extensional grain boundaries, where strain compatibility cannot be maintained between grains by dislocation creep. Such behaviour may be associated with deformation along a decreasing temperature and/or confining pressure path or related to a transient period of high strain-rate (see Knipe 1990) and is likely to be characteristic of deformation near the transition from quasiplastic to elastico-frictional deformation. The result is an opening of grain boundaries and the ability of fluids to penetrate the deforming aggregate. There is a gradation in the amount of porosity which can be created by this process and the effect on fluid flow will depend upon the deviation from the conditions where 'steady state' deformation by dislocation creep is possible or where strain accommodation by diffusive mass transfer prevents the creation of extended and linked fluid migration pathways. At one end of the spectrum is the behaviour where deformation conditions induce only a small deviation away from the situation where 'void' formation and grain boundary dilatation is suppressed. In this situation small, isolated but migrating grain boundary 'voids' and tubes are created which link only occasionally. The movement velocity, distribution and linking probability of the fluid inclusions will control any fluid flow. It is likely that fluid movement is via the slow migration of micro-packages of fluid which are only linked with other inclusions for short time periods. Larger deviations in the deformation conditions away from those which suppress 'void' development will lead to the creation of more extensive

106

R.J. KNIPE & A.M. McCAIG

arrays of inclusions which are linked for longer periods and therefore are more efficient fluid migration pathways. Under these conditions fluid flow may be associated with pervasive flow through the deforming aggregate although not all the boundaries need to be open at any one time. That is, while fluid-rock interactions may affect all grains, the active fluid migration pathway through the aggregate is not constant but changes as previously active routes are closed or cemented up. This dynamic creation and destruction of fluid migration pathways, where only a small percentage of the grain boundaries act as linked pathways, is important as it allows the aggregate to maintain its strength. This dynamic situation will have a specific range of conditions (strain-rates, stress levels, grain boundary migration rates and temperatures) for specific materials and will give way at lower temperatures and higher strain rates to fracturing. The microstructures which are characteristic of the migration of fluids along selective grain boundaries include the following.

(a) Reaction products from fluid-rock interactions along the selective pathway. McCaig & Knipe (1990) gave an example of this from a mylonite in the Pyrenees where the generation of new plagioclase tracked the movement of a fluid away from a possible source. The particular value of such information is that the direction of flow can be identified from the pattern of selective alteration.

Fig. 3. Backscattered SEM orientation contrast image of dynamically recrystallized calcite from a deformed vein within mylonites beneath the Gavarnie Thrust, central Pyrenees (McCaig et al. 1993). The main source of contrast is differences in lattice orientation within a monomineralic calcite aggregate (Lloyd 1987), but note also the truncated zoning patterns within recrystallized calcite grains. The similar sequence of zones everywhere in the vein indicates pervasive introduction of chemical components (probably Fe) into the grain boundaries during grain growth or grain boundary migration. Note also the microporosity along grain boundaries. Statistical analysis (L-Y. Gong, unpublished data, 1993) shows that the pores are concentrated on grain triple points, and on low stress grain boundaries in the simple shear regime (shear was dextral in the view shown, with the shear plane parallel to the top edge of the photograph).

(b) Precipitation of new phases at microdilatation sites along fluid migration pathways. Extension across grain-boundaries created by the strain incompatibilities between grains may allow the precipitation of new cements in equilibrium with the invading fluid. Examples of this include the development of fine phyllosilicates along commonly orientated grain boundaries in mylonites. McCaig (1987) described the precipitation of chlorite-quartz intergrowths in dilatant shear bands within a phyllonite from the Pyrenees. Figure 3 presents another example of pervasive fluid flow through a mylonite. The deformed calcite vein in the example shown is from mylonites associated with the Gavarnie Thrust in the Central Pyrenees (McCaig et al. 1993). The dynamically recrystallized aggregate present is typical of deformation by dislocation creep accompanied by dynamic recrystallization. The crystallographic orientation contrast images (see Lloyd 1987) also reveal the presence of small scale zoning patterns, which probably arise from minor variations in Fe content (these zoning patterns were not identifiable by spot

micro-chemical analysis in the microprobe). The zoning patterns are frequently truncated by adjacent grain boundaries, probably because of subsequent grain boundary migration. In the monomineralic aggregate these microchemical variations can only be produced by the introduction of new chemical components from an external source. Because the zoning patterns are similar all over the vein the microstructures suggest either growth on selected extensional grain boundaries throughout the aggregate during deformation, or pervasive introduction of components into migrating grain boundaries during dynamic recrystallization. The latter might occur by grain boundary diffusion away from transiently open fluid channelways. The concentration of grain boundary 'voids' on extensional grain boundaries (L-Y. Gong, unpublished data) supports this suggestion. The deformation mechanism path for the mylonite probably involved a transition from dislocation creep, which generated the fine grained aggregate, to one where diffusive mass transfer and

MICROSTRUCTURAL CONSEQUENCES OF FLUID FLOW grain boundary dilatation became more important.

Discussion The above review has outlined the range of fluid flow patterns which may be generated in rocks undergoing deformation by different mechanisms. It is clear that deformation can add high permeability flow paths to the pervasive or bulk flow possible through the undeformed rock porosity. The additional flow pathways associated with deformation are however transient and their effect on fluid flow depends on the permeability evolution of the fabrics in the tectonites, the distribution of strain associated with deformation and the ability of the deformation to link domains, layers or compartments with different fluid pressures. Knipe et al. (1991) recognized three aspects of deformation in fault zones which are important to the future assessment of fluid flow. (a) The deformation mechanism paths and the transient changes in the physical properties of the fault zone associated with displacement. (b) The deformation fabric stability; i.e. the dilatation patterns induced during deformation associated with the generation of different fabrics, the distribution of deformation within fault zones and the duration of the deformation activity. (c) The competition between cyclic or transient deformation events and continuous deformation. Enhanced fluid flow may be associated with; (i) pervasive, continuous creep, (ii) transient deformation events, or, (iii) flow along faults between or after deformation events when, despite the low strain rates, enhanced flow is still possible through the deformation fabrics present. A critical factor in the efficiency of fluid flow in deformed rocks is the competition between the processes which maintain the connectivity of the high permeability pathways and those which close such pathways. The range of deformation processes which are involved in this competition will be different depending on the tectonic setting, the deformation conditions and the rock types involved. A brief review of the processes which interact to control fluid flow during deformation in sedimentary basins, crystalline basement under low to moderate grade metamorphism and during prograde metamorphism at moderate to high grades is given below. It should be emphasized that this division and separation is artificial and interactions between these regimes are possible particularly where exchange of fluids occurs via a large crustal scale fault zone as discussed by Sibson (1986, 1992).

107

Fluid flow and deformation in sedimentary basins While subsidence and sedimentation continue, material in a sedimentary basin will follow a path of increasing pressure and temperature. Fluid pressure increases to values above hydrostatic pressure during burial by a combination of compaction, pressure solution and cementation, and the release of metamorphic fluids and hydrocarbons. The process may be accelerated by deformation if this leads to over-compaction. The extent of fluid pressure increase depends on the permeability of the route to the surface, which depends on the distribution of sealing lithologies and the distribution and sealing properties of faults. This dual role of faults, to act as fluid pathways and as flow barriers in sediments, together with the complexities of the timing of fault array evolution and fluid pressure evolution in different sediments, leads to fluid flow in sedimentary basins being controlled by the creation and destruction of complex 3D migration pathways (Knipe 1993). The influence of evolving fault plane geometries, displacement patterns, tip zone processes and fault rock evolution in controlling the juxtapositions and windows for fluid communication and diagenetic changes are discussed in Knipe (1993). Fluids in the deeper parts of sedimentary basins are frequently hypersaline brines with high concentrations of Na, Ca and C1 (Hanor 1987). Such fluids have a high potential for transporting chemical components (cf. Banks et al. 1991 ; McManus & Hanor 1993), and this may in turn enhance deformation involving diffusive mass-transfer or stress corrosion.

Deformation of crystalline basement rocks at low to moderate metamorphic grades At low temperatures deformation in these rocks is likely to be overwhelmingly dominated by fracture and cataclasis. Significant compaction will not be occurring and any metamorphic reactions will be retrograde, tending to reduce fluid pressures and permeability through the absorption of water and the precipitation of hydrous phases. Many crustal earthquakes nucleate in rocks of this type, since they generally form the strongest part of the lithosphere. Various studies have shown that fluids in rocks of this type tend to be hypersaline brines, both in the near surface (Frape & Fritz 1987), and in shear zones exhibiting cyclic fracture-creep behaviour at temperatures up to 400 ° C (Yardley et al. 1993). In some cases these fluids are

108

R.J. KNIPE & A.M. McCAIG

demonstrably formation waters derived ultimately from the Earth's surface (McCaig 1988; McCaig et al, 1990; Yardley et al. 1993). Crystalline rocks are probably strong enough to support a fracture porosity for considerable periods even if fluid pressures are below lithostatic. Fluids may move down into such rocks from overpressured sediments during deformation or alternatively overthrusting onto sediments will induce upward migration into the crystalline rocks (Kerrich 1986). In some cases, structural geometries do not appear to allow this, and mechanisms involving large scale seismic pumping have been proposed (McCaig 1988). Complex flow patterns are likely to be associated with the tip zones of sub-surface, isolated faulting events, where an asymmetrical distribution of extension and compression adjacent to the fault plane may induce a range of flow patterns between the hanging wall and the footwall (Knipe 1993). Fluid flow is also likely to be different around normal and reverse faults, since mean stress variations may favour movement of fluids into normal faults and away from thrust faults during loading associated with earthquakes (Sibson 1992). It is important to bear in mind however that geochemical evidence for fluid flow often comes from relatively minor faults and shear zones where metasomatic changes can easily be documented (e.g. Beach 1976; McCaig et al. 1990). Stress cycling and fluid flow in the vicinity of these minor structures is likely to be controlled by the earthquake cycle on regionally important seismogenic faults, and they should not be treated in isolation (McCaig 1988). At high temperature (>c.350°C) enhanced plastic deformation will lead to increased fluid pressures, which enhances the possibility of fracturing. It is an interesting paradox that ductile shear zones which contain fluid-filled inclusions or micro-fractures undergoing deformation by dislocation creep are probably also in a near-critical state for fracture. Transient fracturing will lead to enhanced porosity and permeability as discussed earlier in the paper. Even when deformation in mylonites is dominated by ductile or crystal-plastic deformation, fracturing is likely in the more resistant layers or phases and these may act as important transient fluid reservoirs in mylonites, expelling fluid into the adjacent more ductile layers when these fracture intermittently. D e f o r m a t i o n associated with p r o g r a d e m e t a m o r p h i s m at m o d e r a t e to high grades

In these circumstances, fluid pressures will generally initially be high due to the release of

metamorphic fluid and deformation will occur at low differential stresses by mixtures of crystal plasticity, diffusive mass transfer and fracture (e.g. Cox & Etheridge 1989). Some reactions may enhance deformation by inducing stresses through volume changes, creating new weak phases or by enhancing diffusion in fine-grained reaction products. Such reaction-enhanced ductility (White & Knipe 1979; Brodie & Rutter 1985) may also influence permeability behaviour, e.g. in the generation of cleavages (Knipe 198!; Cox et al. 1991) or during deformation/metamorphism of impure dolomites (Yardley & Lloyd 1989).

Conclusions The aim of this paper has been to highlight some of the progress made in recent years towards understanding the interactions between deformation and fluid flow. These interactions are complex and arise from changes in the pore structure, permeability and fluid pressure gradients in rocks during straining. Deformation generates high permeability pathways (dynamic permeability of Stephenson et al. this volume) which enhance fluid flow while the pathways are open and are associated with fluid pressure gradients. Deformation may also create changes in porosity which survive after deformation has ceased and allow continued focused flow of fluids, before eventual sealing. Important objectives of future research will be to quantify these changes, assess how extensive the pathways are, and to establish the time period that such enhanced migration pathways can contribute to fluid redistribution in rocks.

The authors would like to thank E.H. Rutter, A.J. Barker and J. Parnell for comments on the first draft.

References AGAR, S.M. 1990. Fracture evolution in the upper ocean crust: Evidence from DSDP hole 504B. In: KNII,E, R.J. & RurrER, E.H. (eds) Deformation Mechanisms, Rheology and Tectonics. Geological Society, London, Special Publications, 54, 41-50. ATKINSON, B.K. 1987. Fracture Mechanics of Rock. Academic Press, London. BANKS, D.A., DAVIES, G.R., YARDLEY, B.W.D., McCAm, A.M. & GRANT, N.T. 1991. The chemistry of brines from an Alpine thrust system: An application of fluid inclusion analysis to the study of fluid behaviour in orogenesis. Geochimica et Cosmochimica Acta, 55, 1021-1030.

MICROSTRUCTURAL CONSEQUENCES OF FLUID FLOW BEACH, A. 1976. The interrelations of fluid transport, deformation, geochemistry and heat flow in Early Proterozoic shear zones in the Lewisian Complex. Philosophical Transactions of the Royal Society of London, A280, 569--604. BLENKINSOP,T.G. & RUTrER, E.H. 1986. Cataclastic deformation of quartzite in the Moine Thrust Zone. Journal of Structural Geology, 8, 669-682. BORRADAILE, G.J. 1981. Particulate flow and the generation of cleavage. Tectonophysics, 72, 306321. BOULLIER, A-M, & ROBERT, F. 1992. Palaeoseismic events recorded in Archean gold-quartz vein networks, Val d'Or Abitibi, Quebec. Journal of Structural Geology, 14, 161-179. BURLEY, S.D., MULLIS,J. & MATTER, A. 1989. Timing diagenesis in the Tartan Reservoir (UK North Sea): constraints from combined cathodoluminescence microscopy and fluid inclusion studies. Marine and Petroleum Geology, 6, 98-120. BRODIE, K.H. & RUYrER, E.H. 1985. On the relationship between deformation and metamorphism with special reference to the behaviour of basic rocks. In: THOMPSOr%A.B. & RUBLE, D.C. (eds) Kinetics, Textures, and Deformation. Advances in Geochemistry, 4. Springer-Verlag, New York, 138-179. BVERLEE, J.M. 1993. Model for episodic flow of high pressure water in fault zones before earthquakes. Geology, 21,303--306. CARTER, N.L., KRONENBURG,A.K., Ross, J.V. & WILTSCHKO, D.V. 1990. Control of fluids on deformation in rocks. In: KNIPE, R.J. & RuI"rER, E.H. (eds) Deformation Mechanisms, Rheology and Tectonics. Geological Society, London, Special Publications, 54, 1-13. CHAMBERLAIN,C.P. & CONRAD,M.E. 1991. Oxygen isotope zoning in garnet. Science, 254,403-406. COWIE, P.A. & SCHOLZ, C.H. 1992. Displacementlength scaling relationships for faults: Data synthesis and discussion. Journal of Structural Geology, 14, 1149-1165. Cox, S.F. & ETHERIDGE, M.A. 1989. Coupled grainscale dilatancy and mass-transfer during deformation at high fluid pressures: Examples from Mt. Lyell, Tasmania. Journal of Structural Geology, l l , 147-162. , WALL, V.J., ETHERIDGE,M.A. & POTTER, T.F. 1991. Deformational and metamorphic processes in the formation of mesothermal vein-hosted gold deposits- examples from the Lechlan Fold Belt in Central Victoria, Australia. Ore Geology Reviews, 6,391-423. DICKSON, J.A.D., SMALLEY,P.C. & KIRKLAND,B.L. 1991. Carbon and oxygen isotopes in Pennsylvanian biogenic and abiogenic aragonite (Otero County, New Mexico): a laser microprobe study. Geochimica et Cosmochimica Acta, 55, 26072613. ELSENHEIMER,D. & VALLEY,J.W. 1992. In situ oxygen isotope analysis of feldspar and quartz by Nd:YAG laser microprobe. Chemical Geology, 101, 21-42. ETHERIDGE, M.A., WALL, V.J. & COX, S.F. 1984.

109

High fluid pressures during regional metamorphism and deformation: Implications for masstransport and deformation mechanisms. Journal of Geophysical Research, 89, 4344-4358. EVANS, B. & WON6, T-F. 1992. Fault mechanics and transport properties of rocks. Academic Press pp. 524. EVANS, J.P. 1990. Textures, deformation mechanisms and the role of fluids in the cataclastic deformation of granitic rocks. In: KNIVE, R.J. & RUTTER, E.H. (eds) Deformation Mechanisms, Rheology and Tectonics. Geological Society London, Special Publications, 54, 29--40. FRAPE, S.K. & Fgrrz, P. 1987. Geochemical trends for groundwaters from the Canadian Shield. In: FRITZ, P. & FRAPE, S.K. (eds) Saline water and gases in crystalline rocks. Special Papers of the Geological Association of Canada, 33, 19-38. FRASER, D.G. 1990. Applications of the high resolution scanning proton microprobe in the earth sciences: An overview. Chemical Geology, 83, 27-37. GILLESPIE, P.A., WALSH, J.J. t~ WATrERSON, J. 1992. Limitations of dimension and displacement data from single faults and the consequences for data analysis and interpretation. Journal of Structural Geology, 14, 1157-1172. GILLETI'I, B.J. & SHIMIZU, N. 1989. Use of the ion microprobe to measure natural abundances of oxygen isotopes in minerals. In: SHANKS,W.C. III & CRISS, R.E. (eds) New Frontiers In Stable

Isotope Research: Laser Probes, Ion Probes and Small-Sample Analysis. US Geological Survey Bulletin, 1890, 129-136. GROSnONG, R.H. 1988. Low temperature deformation mechanisms and their interpretation. Geological Society of America Bulletin, 100, 1329-1360. GUEGUEN, Y., DAROT, M. & REUSCHLE, T. 1987. Permeability evolution under stress. In: DOLIGEZ, B. (ed.) Migration of Hydrocarbons in Sedimentary Basins. IFP Exploration Resources Conference, 45,281-285. HANOR, J.S. 1987. Origin and migration of subsurface sedimentary brines. SEPM short course 21. HIPPLER, S.J. • KNIPE, R.J. 1990. The evolution of cataclastic fault rocks from a pre-existing mylonite. In: KNIPE, R.J. & RUIq'ER, E.H. (eds)

Deformation Mechanisms, Rheology and Tectonics. Geological Society, London, Special Publications, 54, 71-79. JACKSON, S.E., LONGERICH, H.P., DUNNING, G. & FRYER, B.J. 1992. The application of LaserAblation Microprobe - Inductively Coupled Plasma - Mass Spectrometry (LAM-ICP-MS) to in situ trace-element determination in minerals. Canadian Mineralogist, 30, 1049-1064. JAMTVEIT,B., WOGELIUS,R.A. & FRASER, D.G. 1993. Zonation patterns of skarn garnets: Records of hydrothermal system evolution. Geology, 21, 113-116. KERRICH, R.J. 1986. Fluid transport in lineaments.

Philosophical Transactions of the Royal Society of London, A317, 219-251.

110

R.J. KNIPE & A.M. McCAIG

KNIPE, R.J. 1981. The interaction of deformation and metamorphism in slates. Tectonophysics, 78, 249-272. 1989. Deformation mechanisms - recognition from natural tectonites. Journal of Structural Geology, 11,127-146. 1990. Microstructural analysis and tectonic evolution in thrust systems: examples from the Assynt region of the Moine Thrust Zone, NW Scotland. In: BARBER, D.J. & MEREDITH, P.G. (eds) Deformation Processes in Minerals, Ceramics and Rocks. The Mineralogical Society Series, 1,228-259. 1992. Faulting processes and fault seal. In: LARSEN, R.M. (ed.) Structural and Tectonic

Modelling and its Application to Petroleum Geology. NPF Special Publications, 1,325-343. -

1993. The influence of fault zone processes on fluid flow and diagenesis. In: HORBURY, E.D. & ROmNSON, A.G. (eds) Diagenesis and Basin Development. American Association of Petroleum Geologists Studies in Geology, 36, 135154. -& WHrrE, S.H. 1979. Deformation in low grade shear zones in the O.R.S. from S.W. Wales. Journal of Structural Geology, 1, 53-66. - - , AGAR, S.M. & PRIOR, D.J. 1991. The microstructural evolution of fluid flow paths in semi-lithified sediments from subduction complexes. Philo-

sophical Transactions of the Royal Society of London, A335, 261-273. LESPINASSE, M. & CATHELINEAU, M. 1990. Fluid percolations in a fault zone. A study of fluid inclusion planes (FIP). Tectonophysics, 184, 173-187. LLOYD, G.E. 1987. Atomic number and crystallographic contrast images using SEM: a review of backscattered electronic techniques. Mineralogical Magazine, 51, 3-19. & KNIPE, R.J. 1992. Deformation mechanisms accommodating faulting of quartzite under upper crustal conditions. Journal of Structural Geology, 14, 127-144. LYON, I. & TURNER, G. 1992. The Isolab ® 54 ion microprobe. Chem&al Geology, 101,197-199. MAIN, I.G., MEREDITH, P.G., SAMMONDS, P.R. & JONES, C. 1990. Influence of fractal flaw distributions on rock deformation in the brittle field. In: KNIPE, R.J. & RUTTER, E.H. (eds) Defor-

-

-

mation Mechanisms, Rheology and Tectonics. Geological Society, London, Special Publications, 54, 71-89. MARSHALL,J.D. (ed.) 1987. Diagenesis of Sedimentary Sequences. Geological Society of London, Special Publications, 36. MCCAIG, A.M. 1984. Fluid-rock interaction in some shear zones from the Pyrenees. Journal of Metamorphic Geology, 2,129-141. 1987. Deformation and fluid-rock interaction in metasomatic dilatant shear bands. Tectonophysics, 135,121-132. 1988. Deep fluid circulation in fault zones. Geology, 16,867-870. & KNIPE, R.J. 1990. Mass transport mechanisms

-

-

in deforming rocks: Recognition using microstructural and microchemical criteria. Geology, 18,824-827. - - , WICKHAM,S.M. & TAYLOR,H.P. JR. 1990. Deep fluid circulation in alpine shear zones, Pyrenees, France: Field and oxygen isotope studies. Contributions to Mineralogy and Petrology, 106, 41-60. - - , GONG, L-Y. & WAYNE,D.M. 1993. Mechanisms of permeability enhancement in carbonate mylonites. In: PARNELL,J., RUFFELL, A.H. & MOLES, N.R. (eds) Geofluids "93 Conference Volume, 159-161. MCLAREN, A.C. 1991. Transmission electron microscopy of minerals and rocks. Cambridge Topics in Mineral Physics and Chemistry 2. Cambridge University Press. MCMANUS, K.M. & HANOR, J.S. 1993. Diagenetic evidence for massive evaporite dissolution, fluid flow, and mass transfer in the Louisiana Gulf Coast. Geology, 21,727-730. MrrRA, S. 1988. Effects of deformation mechanisms on reservoir potential in Central Appalachian overthrust belt. American Association of Petroleum Geologists, 72,536--554. MUIR WOOl), R. & KING, G.C.P. 1993. Hydrological signatures of earthquake strain. Journal of Geophysical Research, 98,000-000. ODLING, N.E. 1992. Permeability of natural and simulated fracture patterns. In: LARSON, R.M. (ed.) Structural and Tectonic Modelling and its Application to Petroleum Geology. NPF Special Publications, 1,365-381. PARRY, W.T. & BRUHN, R.L. 1986. Pore fluids and seismogenic characteristics of fault rock on the Wasatch fault, Utah. Journal of Geophysical Research, 91,730-744. PHILIPPOT, P. & VAN ROERMUND, H.L.M. 1992. Deformation processes in eclogitic rocks. Journal of Structural Geology, 14, 1059-1077. PRIOR, D.J. 1988. Fractures and retrogression in garnets from the Alpine Fault mylonites, New Zealand. Journal of the Geological Society, London, 146, 335-347. RAMSAY,J.G. 1980. The crack-seal mechanism of rock deformation. Nature, 284, 135-139. ROEDDER, E. 1984. Fluid inclusions. Reviews in Mineralogy 12. SCHOLZ, C.H. 1990. The mechanics of earthquakes and faulting. Cambridge University Press. SELVERSTONE,J., FRANZ, G., THOMAS, S. t~ GETTY, S. 1992. Fluid variability in 2 GPa eclogites as an indicator of fluid behaviour during subduction. Contributions to Mineralogy and Petrology, 112, 341-357. SHARP,Z.D. 1992. In situ laser microprobe techniques for stable isotope analysis. Chemical Geology, 101, 3-20. SIBSON, R.H. 1980. Transient discontinuities in ductile shear zones. Journal of Structural Geology, 2, 165-171. 1986. Earthquakes and rock deformation in crustal fault zones. Annual Reviews in the Earth and Planetary Sciences, 14,149-175. 1990. Conditions of fault-valve behaviour. In: -

-

MICROSTRUCTURAL CONSEQUENCES OF FLUID FLOW KNIPE, R.J. & RUTTER, E.H. (eds) Deformation Mechanisms, Rheology and Tectonics. Geological Society, London, Special Publications, 54, 15-28. 1992. Implications of fault-valve behaviour in rupture nucleation and recurrence. Tectonophysics, 211,283-293. 1993. Crustal stress, faulting, and fluid flow. In: PARNELL, J., RUFFELL, A.H. & MOLES, N.R. (eds) Geofluids '93 Conference Volume, 137-140. ~, ROBERT,F. & POULSEN, K.H. 1988. High angle reverse faults, fluid pressure cycling & mesothermal gold-quartz deposits. Geology, 16,551-555. SLEEP, N.H. & BLANPIED, M.L. 1992. Creep, compaction and the weak theology of major fault zones. Nature, 359,687--692. SPIERS, C.J. & SCHUTJENS, P.M.T.M. 1990. Densification of crystalline aggregates by fluid phase diffusional creep. In: BARBER,D. & MEREDITH, P.G. (eds) Deformation of Materials. Mineralogical Society Series, 1,334-352. STARK, C.P. & STARK,J.A. 1990. Seismic fluids & percolation theory. Journal of Geophysical Research, 96, 8417-26. STEPHENSON, E.L., MALTMAN, A.J. & KNIPE, R.J. 1993. Fluid flow and microstructure in deforming sediments. In: PARNELL, J., RUFFELL, A.H. & MOLES, N.R. (eds) Geofluids '93 Conference Volume, 148-150.

111

WAYNE, D.M. & McCAIc, A.M. 1992. Fluid flow during thrusting: A Sr isotope study. In: KHARAKA, Y. & MAEST, A.S. (eds) Water-rock interaction. Balkema, Rotterdam, 1559-1664. WINTSCH, R.P. & KNIPE,R.J. 1983. Growth of a zoned plagioclase porphyroblast in a mylonite. Geology, 11,360-363. WHITE, S.H. & KNIPE, R.J. 1979. Transformational and reaction enhanced ductility. Journal of the Geological Society, London, 135, 513-516. WOGELIUS, R.A., FRASER, D.G., FELTHAM, D.J. & WHITEMAN, M.I. 1992. Trace element zoning in dolomite: Proton microprobe data and thermodynamic constraints on fluid composition. Geochimica et Cosmochimica Acta, 56,319-334. YARDLEY, B.W.D. & LLOYD, G.E. 1989. An application of cathodoluminescencemicroscopy to the study of textures and reactions in high-grade marbles from Connemara, Ireland. Geological Magazine, 126,333-337. ~, BANKS, D.A. & MUNZ, I.A. 1993. Fluid penetration into crystalline crust: Evidence from the halogen chemistry of inclusion fluids. In: PARNELL, J., RUEEELL, A.H. & MOLES, N.R. (eds) Geofluids '93 Conference Volume, 350-353.

Fluid flow in actively deforming sediments: 'dynamic permeability' in accretionary prisms E.L. STEPHENSON,

A.J. MALTMAN

& R.J. KNIPE 1

Institute of Earth Studies, University of Wales, Aberystwyth, Wales SY23 3DB, UK 1 Department of Earth Sciences, University of Leeds, Leeds LS2 9JT, UK Abstract: Recent investigations into active accretionary prisms have emphasized both the importance of the deformation/fluid-flow interplay in governing fundamental aspects of prisms and the need for quantitative data on the processes. Even the basic parameter of permeability is little known in this and other geological settings where the sediments are undergoing active deformation. New experimental data are presented which show that permeability during accumulating strain, here called the dynamicpermeability, is not a static value but is highly variable. Moreover, microstructural analysis and precise determination of permeant volume reveal that this dynamic permeability is not solely the varying capacity of the deforming medium to transmit fluid, a quantity here called the Darcyanpermeability, but includes a contribution to the amount of permeant from the medium itself. In the laboratory experiments this additional contribution, called the dynamic component, arises from pore-volume fluctuations associated with microstructural changes, but in nature there may be further mechanical and chemical effects. Applying conventional methods of permeability determination to an actively deforming sediment will necessarily include this dynamic component in the measurement, a consideration relevant to a variety of geological and engineering situations.

The interaction between deformation and fluid flow has recently become an important topic for geologists investigating accretionary prisms. The interest has arisen because at convergent plate margins, where materials tend to be scraped from the subducting plate to build out the prism-shaped mass, accumulating wet sediments are being transformed into virtually dry rocks in an exceptionally dynamic structural environment (e.g. Langseth & Moore 1990). Enormous quantities of fluid have to be lost from the system, and their fate influences a wide range of fundamental plate-margin processes, from magmatism to mineralization, and tectonic behaviour to metamorphism. How do the fluids escape? Are some temporarily trapped? What proportion of the fluid is subducted and what routes does the remainder follow? Such questions are fundamental to an understanding of the architecture and processes of convergent plate margins. Despite this importance, knowledge of how fluids are transported through actively deforming and deformed geological materials is extremely sparse. The fundamental parameter of permeability, for example, has been measured in countless on-land rocks that are assumed to be rigid, static, and isotropic. Any effects of deformation fabrics or ongoing strain have been little explored. Reported below is an

attempt in the laboratory to measure the permeability of sediments during their active deformation, a quantity we refer to as the dynamic permeability. The results not only reveal a substantial difference between the dynamic and the static, at-rest values, but demonstrate that the deforming sediments themselves can contribute to the amount of expelled fluid. This latter observation requires that in such dynamic situations the concept of permeability cannot be treated in the conventional way. The first half of the paper is a brief review of recent progress in investigating fluid flow in actively deforming accretionary prisms. The second half discusses the results of new experiments into permeability variations during the active deformation of sediments.

Recent studies on deformation and fluid flow in accretionary prisms In the last two decades, convergent margins have blossomed from being a poorly investigated, obscurely understood plate-margin type to one with well-established architectural features, and to being a target of major international study. By the mid-1980s, it was known that at most sites of oceanic-plate subduction,

From PARNELL,J. (ed.), 1994, Geofluids: Origin, Migrationand Evolution of Fluidsin SedimentaryBasins, Geological Society Special Publication No. 78, 113-125.

113

114

E.L. STEPHENSON E T AL.

the veneer of sediments on the converging plate and those accumulating in the trench were being intermittently offscraped, in a geometry closely analogous to that well documented from on-land fold-and-thrust belts (Moore & Lundberg 1986). The main features of such prisms of accreted materials are shown in Fig. 1. Where little sediment is available for accretion, such as at sites of intra-oceanic collision, accretion is minimal (e.g. Mariana; Fryer et al. 1992), and some margins appear to show no accretion at all, with all of the oceanic plate material being subducted (e.g. off Guatemala; yon Huene et al. 1985). At some margins, accreted material seems to be removed from beneath the prism by the subducting plate, the so-called subduction erosion (e.g. off NW Japan; Langseth et al. 1981), while at others further material is added to the sole of the prism, a process called underplating (e.g. Makkran; Platt & Leggett 1985). Also, some accretionary prisms are relatively short but immensely thick while others appear thin and areally extensive, leading to variations in prism taper angles. These major differences were reviewed by von Huene (1984). A crucial factor in determining these gross tectonic differences is the behaviour of the boundary between the colliding plates, the detachment zone or decollement, and theories attempting to explain the variations in taperangle have also highlighted the importance of decollement behaviour (e.g. yon Huene & Lee 1983). All the indications are that the fluid behaviour is fundamental (e.g. Westbrook & Smith 1983). Fluid pressures appear to govern, just as in the classic explanations of on-land thrust belts, the friction at the decollement, and hence whether it can slip freely or whether the strain has to be transferred to other features such as propagating thrusts, thus thickening the prism (Davis, et al. 1983). Drilling near the decollement beneath the Northern Barbados Ridge prism during the Deep Sea Drilling Project (DSDP; Moore & Biju-Duval 1984) inadvertently indicated the existence there of near-lithostatic fluid pressures, and a direct measurement of abnormal fluid pressures was achieved during DSDP drilling off Guatemala (yon Huene 1985). Besides this general evolution of enquiry, two particular developments in the late 1980s made fluids the crucial topic of prism research. One emerged from a series of international projects using submersibles and sea-floor imaging to observe the ocean floor (e.g. Carson et al. 1991; Le Pichon et al. 1992 and following articles). The discovery of isolated biological communities and mineral crusts at the outcrops of faults seen on

seismic traces fostered the notion of focused flow, with tectonic features acting as conduits for fluid escape. The other development arose from further drilling of the Barbados prism, now as part of the Ocean Drilling Program (ODP). This drilling revealed clear evidence of channelised flow along major tectonic features. It was indicated by, for example, geochemical anomalies (e.g. enrichment of methane, 180, and 87Sr/S6Sr, and dilution of chloride; Gieskes et al. 1990), thermal effects (Langseth et al. 1990), and syntectonic mineralization (Vrolijk & Sheppard 1991) within the faults and basal decollement. Geophysical studies, particularly involving seismic velocities together with reflection amplitudes and polarities, added further evidence (e.g. Westbrook 1991). Such discoveries revealed exciting new insights into the hydrogeology of prisms, and the neatness of the data from Barbados elevated the deformation-controlled dewatering of this prism to something of a paradigm. Moreover, drilling of prism-like sediments off Peru showed reasonably similar anomalies (e.g. Kastner et al. 1991). However, the concept of tectonic conduits had also raised a host of new questions (Langseth & Moore 1990), and exposed the need for quantitative information. It was expected that the next prism to be targeted by the ODP, the Nankai Prism off SW Japan, would display features similar to Barbados and help provide numerical data on the channelized flow and related effects. In the event, however, not only did in situ measurement continue to prove difficult, but the very evidence for the tectonic focusing of flow was unclear. In fact, in some ways, the structural (Maltman et al. 1992) and other evidence (Taira et al. 1992) was more compatible with a pervasive, intergranular drainage of this prism. Good evidence did emerge that the decollement at Nankai is currently acting as a seal, efficiently separating overpressured sediments on the downgoing plate from the apparently diffusely draining prism. Yet along strike from the drilling site at Nankai, beyond several major strike-slip faults transverse to the prism, observations from a submersible showed the presence of diapirs and bioherms indicative of localized venting (Le Pichon et al. 1992) i.e. prisms can be compartmentalised into different dewatering modes. The relatively small prism drilled during ODP investigations of processes at the Chile triple junction showed, again through thermal and chemical discontinuities (Behrmann et al. 1993), that fluids were moving vigorously in the highly sheared formations, although the processes there are taking place within an unusually high

FLUID FLOW IN DEFORMING SEDIMENTS diffuse dewatering

channelised dewatering ~

.-~~, "' " " .

i'

.

ent . ! -

• d~ut,

. "• ~' /-_ ." ' . ' - . " .• ' . '- . ~ ...

.

'-

• ". ". ". •

/

\

"

~~ I " ,

.

.

i

115

"\/

~ / ~ / / _

.

.

.

.

" . " . -

.

/ /

frontal thrust

.£

/

.

i

,'~\

. x. .

\DOWNGOINGPLATE ~1

.

.

s u b d u c t i o n e r o s on I / /

/~/\..

\"

~/

t/~.-"

~

--" ~

protothrust

....:.:.:.:.:.i:.i.i.:.:.i.:.i::

underD a t i n o ~" "" " p r i s m t a p e .

,, ~ ' \. . .

.

~

...-

.

.

I

~

/ ~ / ~ \ / ~ l t " / ,

/

\l

g t

\

"..t.."

-. , \ ~ " \ .

/

/

~

/

t

/

- / ~/ " / t t i

/

--

.

Fig. 1. Schematic drawing of the toe region of a hypothetical accretionary prism, to illustrate terminology used in the text. In this typical example, the decollement is positioned within the sedimentary pile (stippled) so that the upper sediments are being accreted to the overriding plate while oceanic crust (pecked ornament) and some overlying sediment are underthrust. Stepping up of the decollement within the sedimentary section leads to subduction erosion while stepping down produces underplating. Note that the channelised and diffuse dewatering regimes, although portrayed here for convenience on adjacent hinterland thrusts, may occur at any position within a prism.

thermal gradient. The accretionary prism most recently targeted, at the Cascadia margin, provided a further example of along-strike change in dewatering behaviour, in that sites off Vancouver disclosed diffuse dewatering through pervasively fractured silts, at flow rates no more than a few millimetres per year, whereas sites further south, off Oregon, showed substantial tectonic focusing (Westbrook et al. 1993). Some thrusts carried thermogenic hydrocarbon gases in actively flowing pore fluids, and overpressures (0.2MPa at 94-146m) were successfully documented. Summaries of some relevant aspects of recently drilled prisms are given in Table 1. Some of the instruments deployed at Cascadia have been left in place for future measurement, to evaluate the episodicity of flow. The objective of obtaining in situ measurements points the way to future research in flow-deformation studies in prisms. The planned ODP drilling of the Barbados prism in 1994 will emphasize the long-term deployment of measuring devices. The goal is to quantify the amounts and temporal behaviour of fluid flow in this otherwise well-documented, actively deforming, accretionary prism. It is clear now that different accretionary prisms dewater in different ways. The behaviour may well vary spatially within the prism, and probably fluctuates through time. Quantitative understanding of all these processes is still poor.

Permeability in deforming sediments If overpressures are the key to the structural behaviour of a prism, then permeability must be the fundamental influence on their extent and duration (Moore et al. 1991). It is the permeability that quantifies the capacity of prism

sediments to dissipate fluid pressures in order to maintain an equilibrium gradient, and hence determines the extent of overpressuring. As mentioned above, although the parameter is routinely measured in a host of on-land circumstances, knowledge of permeability in any oceanic sediments is extremely sparse, let alone in examples containing deformation fabrics and those that are undergoing active strain. Taylor & Fisher (1993) reviewed the relatively few permeability data obtained from modern accretionary prisms, through laboratory measurements. They cite values ranging from 10-14 to 10-19 m 2 for cores retrieved from Nankai and, for cores from the Barbados prism (Taylor & Leonard 1990), permeabilities between 10 -15 and 10 -18 m 2. The role that deformation fabrics play in influencing these values is unclear. Evidence of preferred flow along core-scale faults was documented by Knipe (1986a) and along shear-zones by Maltman (1988). Arch & Maltman (1990) advanced an explanation of the latter observations, for zones with marginparallel grain alignments, based on the reduced tortuosity of the flow path. Brown & Moore (1993) have commented that fabric collapse accompanying such alignments should also reduce the porosity and hence reduce the alongzone permeability. Although theory holds this to be true, it does seem that in practice, whether through transient dilation or some other mechanism, shear zones can channel flow at least for some time. Core-scale shear zones can be regarded to some extent as scale models of the large-scale tectonic discontinuities of accretionary prisms (Arch & Maltman 1990), but in some prisms they themselves may play an important role in the bulk dewatering (Byrne et al. 1993).

Anomalies Localized rhodocrosite 28-36 ° , anomalies

87Sr/86Sr Mineralized veins

Thermal gradient

Chile (ODP Leg 141)

Lacks anomalies No veins 110°, linear

50 ° , anomalies

Some localized carbonate, pyrite Highly variable between 30--600 °

Diapirs, bioherms in n/a places Local minima Lacks anomalies Some anomalies at n/a thrusts

Nankai (ODP Leg 131)

Anomalies Sparse

Anomalies Anomalies

n/a

Mud volcanoes, diapirs Anomalies Anomalies

Ocean-floor venting

Cf concentrations 180 values

Peru (ODP Leg 112)

Barbados (ODP Leg 110)

54 ° , linear

51 ° mostly linear

Carbonate veinlets

Gas discharge carbonate deposits Anomalies Pore water anomalies reflect 'spatial heterogeneity of fluid flow'

Oregon (ODP Leg 146)

Linear change 'Little evidence of significant fluid flow that is confined to conduits'

None

Vancouver

Cascadia

FLUID FLOW IN DEFORMING SEDIMENTS Maltman et al. (1993) suggested that the abundance of these kinds of core-scale structures at Nankai may be part of the reason why this prism shows less marked flow along the major tectonic structures (Taira et al. 1993). All the permeability measurements mentioned above were made on laboratory specimens 'at rest', giving a value we refer to as the static permeability. Langseth et al. (1988) pointed out that such determinations may not be an accurate measure of the value within an actively deforming prism, hence the importance of attempting the difficult matter of obtaining reliable in situ measurements. The only known efforts in the laboratory to assess the effect of active strain on flow through prism materials, referred to here as the dynamic permeability, are the preliminary results we gave in Byrne et al. (1993) and the data outlined below. The former results, on cores from the Nankai prism and analogue material, revealed that the dynamic permeability may change by almost 100% at only 25% bulk strain. New data are presented below, in an attempt to refine understanding of this kind of behaviour.

Experimental method We have employed a number of innovative, precise methods, cross-checked with other laboratories, to monitor permeability changes during sediment deformation. Details of the equipment and methods are given elsewhere (Arch, Stephenson & Maltman, in prep). Briefly, the experiments were based on coupling the well-established constant-head method of permeability measurement to a triaxial test arrangement. The two important points to note here are that: (i) the constant-head technique is based on supplying permeant under a constant pressure-head and measuring its ou(flow from a specimen; and (ii) our equipment also allows the amount of inflow to be monitored, even though this would normally be expected to be the same value. Other relevant points are that the specimens are reasonably large, 54 mm diameter cylinders × (typically) 100 mm height, tests were accomplished in a few tens of hours and at room temperature, with readings being automatically taken at ninety second intervals, and the permeant was distilled water. Our work so far has centred on clays and silty clays. The samples have either been retrieved from the active Nankai accretionary prism by the ODP, or generated in the laboratory by consolidating clay/silt slurries.

Experimental results Equilibrium flow through a specimen is known to have been established when any variation of outflow with time is negligible. From Darcy's relation, and knowing the specimen dimensions and the differential head, the permeability of the

117

specimen can be calculated. A typical result of such static permeability tests is shown in Fig. 2a. None of the materials tested so far retains its static permeability value when it undergoes strain. Our results suggest that for a given sediment type the variation depends to a large extent on its state of consolidation. Under- and normally-consolidated materials show a gradual reduction in dynamic permeability with strain (e.g. Fig. 2b) whereas over-consolidated samples of the same composition show an increase (Fig. 2c). Some lightly overconsolidated specimens show, for reasons we have not yet identified, more complex fluctuations superimposed on a progressive dynamic permeability decrease (Fig. 2d). Arch, Stephenson, & Maltman (in prep) suggest that the explanation of these varying behaviours lies in the dilation and contraction of the materials at the grain scale. In summary, more tightly-packed grains, as in an overconsolidated specimen, have to dilate to accommodate the growing strain, thus increasing the permeability either transiently or permanently, whereas poorly consolidated sediments tighten the grain packing with strain, and diminish the permeability. Temporary fluctuations presumably correspond to porosity changes associated with localized reconfigurations of grains, in particular the production of shear zones in conditions of peak stress.

The concept of dynamic permeability The inferences of dilation and tighter packing mentioned above open up a new view of the concept of permeability in active materials. That is, instead of the medium simply being a matrix of grains and pores that passively allows the permeant to filter through, the material itself can add to or subtract from the amount o f fluid. In the classical view, originating with Henry Darcy, permeability is a material constant, representing the capacity of a porous medium to transmit water in response to a hydraulic gradient. There have been discussions in geology on possible physical (e.g. Peach et al. 1987) and chemical (Sample 1990) interactions between the permeant and the permeated matrix, but normally the medium is treated as inert. This is the case in our treatment of static permeability. However, we view dynamic permeability as consisting of two components: the passive matrix element mentioned above, which we propose to call the Darcyan permeability, and a dynamic component, arising from the contribution from the grain matrix itself. Fine semantic matters arise from this; we purposefully avoid referring to the

118

E.L. STEPHENSON E T AL. 1400

5E-16

1200 r~

4E-16

5E-16

~

1000

800 "~

3E-16

4E-16 600

E

E

~- 2E-16

ne 3E-16

~-

400 200

1E-16 0

|

=

=

=

i

5

10

15

20

25

2E-16 0

2

4

2E-15 1E-15

~ >,

1200

b ~

6

8

10

12

14

16

18

20

Strain %

.time - h o u r s

~ BSS~~ ' ' ~

5E-17

] d

350 300

1000

1E-15

¢~ E

800

1E-15

600

8E-16

==

._ "~

4E-17

\

250 200

3E-17 150

~

6E-16

400

=

e~_b_i,ty_ " '-.. . . . . . . . . . .p_~r~_ .. ,_~ . . . . . . . .

4E-16

E 2E-17

200

2E-16 OE+ O0 0

|

i

,

,

5

10

15

20

0 25

30

Strain %

1E-17

]I V

0

lO0

', ;-.-., ,. . . . . ~,. . . . . ; ~.,J

~

;0

;5

2'0

ca .~ m

50 0

25

Strain %

Fig. 2. Summary of various permeability behaviours. (a) Static permeability graph for a sand-kaolinite mixture. (b) Dynamic permeability of an under-consolidated sand-kaolinite sediment. The permeability shows an overall decrease with increasing strain, initially rapid but then at a reduced rate as the specimen consolidates under the applied load. (c) Dynamic permeability of an overconsolidated sand-kaolinite mixture. After initial adjustment (1% strain), the permeability shows a small decrease followed by an increase, presumably as the specimen dilates to accommodate strain. (d) Complex dynamic permeability behaviour, in a lightly overconsolidated ball-clay. Intervals of enhanced permeability are superimposed on an overall gradual decrease.

latter component as a permeability or a flux. Depending on the circumstances, the dynamic component may be positive, negative or zero. Irrespective of this contribution from the dynamic component, in an actively straining sediment the Darcyan permeability itself may well not be of constant magnitude. In a natural active situation, the dynamic component probably arises chiefly through either or both of two effects. The first is diagenesis. For example, cementation (mineral infilling of pore space) will displace some of the pore-fluid. Transformation of organic matter during burial will produce new hydrocarbons, some of which may be in the fluid state (e.g. Berner & Faber 1993). The hydrous content of minerals in the buried sediment may change. Many minerals contain water more or less loosely linked to their molecular structure, and there are numerous geological circumstances where various minerals may take up or lose

water. However, the major process of this kind within accretionary prisms will be clay dehydration. According to Tribble (1990), for example, in the Barbados area 30% of the grain matrix consists of (hydrous) smectite. After burial to temperature conditions around 60° C, the smectite transforms to illite, releasing about a third of its volume as water (Le Pichon et al. 1991). The second effect is mechanical change within the sediment. In materials consolidating or dilating, the bulk of the volume change will be accommodated by porosity adjustment, and because pore-fluids are virtually incompressible (excepting gases), these will be accompanied by fluid intake or expulsion. It is doubtful that in our short, room-temperature experiments involving distilled water, any of the diagenetic effects are being reproduced, but we believe we have detected mechanical contributions to dynamic permeability. The evidence comes from

FLUID FLOW IN DEFORMING SEDIMENTS

;

119

i-

,

Fig. 3. SEM secondary electron images of ball clays at different amounts of bulk strain. (a) Undeformed sample, showing the domainal primary consolidation fabric, here running approximately NNE-SSW, with high-angle clays between creating pockets for fluid storage. (b) Similar material, deformed to 3% strain. The primary fabric, oriented approximately NE-SW in this image, has been intensified. (e) Similar material, deformed to 7% strain. Shear zones, with intensely aligned but undamaged clays, here run NNW-SSE. Localized dilation, to allow rotation of the clay particles, will have caused temporary storage of permeant, later to be released as the clays collapse into alignment.

microstructural analysis and from precise volume-change measurements during the experiments.

Microstructural analysis Some of the test specimens have been examined microscopically in the scanning electron microscope. Summarizing the results, materials consolidated to greater pressures show more intense fabrics, defined by better alignment of the clay flakes and decreased pore space. The effect is well known from nature (e.g. Maltman 1981; Faas & Crockett 1983). Initial shortening of the experimental cylinders continues the consolidation process; Arch (1988) showed that 30% porosity ball clays can achieve up to 12% bulk strain in this way. Continuing strain causes, in line with observations from nature (Maltman 1988; Knipe 1986b) and experiment (Maitman 1987), the sediments to fail by generating arrays

of shear zones. The evidence that these microstructural changes must have contributed to the dynamic permeability is discussed in greater detail below for a selected suite of tests. A series of ball-clay test cylinders was generated in the laboratory and each deformed to different levels of total axial strain. Samples from each were compared using secondary electron imaging of fracture surfaces, oriented roughly perpendicular to the dominant microstructure of the specimen. A control sample, trimmed from a specimen prior to testing, showed a primary fabric resulting from the uniaxial consolidation of the sediment employed in the manufacture of the sample. As illustrated in Fig. 3a, the fabric is domainal, with variations in orientation and intensity of clay particle alignment. Both presumably give rise to local variations in permeability. Also present are regions with particles oriented at a high angle to the main fabric. Being sub-parallel to the

E.L. STEPHENSON ET AL.

120

_

5E-17-

stress

180 -160 -140

4E-17-

-120 E '-JO ¢0 (1)

-100

t~ ,--

3E-17 -80

t~ "13

60 0,.

2E-17-

40

g 20 1E-17 - J 0

2

4

(~

8 10 strain %

1'2

14

1'6

0

18

Fig. 4. Dynamic permeability of a lightly overconsolidated ball clay, showing marked permeability variation as strain increases. Transient increases are superimposed on an overall permeability decrease. Confining pressure = 340 KPa; differential pressure along specimen = 200 KPa; strain rate = 1.02 × 10-4 s-1. The permeability curve is a thirty-point running mean through the outflow data.

direction of consolidation loading, these particles are prompting the formation of pockets in which permeant might be stored and, being in a less mechanically stable orientation, they may be preferred sites for the initiation of shear zones when strain accumulates. The intensification of the primary fabric at low levels of bulk strain is apparent from Fig. 3b. The clay particles show a predominantly face-toface arrangement, trending approximately N E - S W in that image. The porosity of the sample has been reduced significantly as a result of the denser packing of the parallel clay flakes. This loss of pore volume must contribute fluid to the downstream, exiting permeant. Measurement of the dynamic permeability during this period will therefore not be wholly Darcyan, but overestimated by an amount proportional to this porosity reduction. The formation of discrete shear zones at higher levels of strain is illustrated in Fig. 3c. Two such zones are present, showing a sigmoidal swinging of the consolidation fabric into the zones, which are defined by a more intense parallel orientation of the clay plates and a further reduction in porosity. Note that none of the clay particles shows signs of damage. The only conceivable way in which this new arrangement could be achieved is through temporary

dilation while the clay flakes rotate into alignment. During these intervals of dilation, permeant must have been incorporated into the expanded pores, temporarily reducing the throughflow. This stored fluid would then be released in pulses when the realigned clays progressively collapsed to form the shear zone fabric. The existence of such localised, transient reservoirs is in line with the model presented by Knipe et aI. (1991). Therefore, the dynamic permeabilities portrayed in Fig. 2 are not simply registering changes in the capacity of the sediment to transmit fluid, the Darcyan permeability, but must be incorporating varying contributions to the flux from these dynamic components arising from pore-volume changes.

Volume change analysis In the Darcyan view of permeability, any change in the fluid transport capability will be accompanied by a corresponding change in the inflow and outflow of permeant, and these will be identical quantities. Both are therefore not normally considered in the determination of permeability. Our equipment allows the precise monitoring of both the inflow and outflow volumes: should any discrepancy between the two quantities be detected it must represent the

FLUID FLOW IN DEFORMING SEDIMENTS

0.020

12l

stress

180

oo,6 / o

,! !

~ 0.o12.

/

o

-, --

o>

,oo

"~'~'~ ~

o.oos

,,.L. L I I

~,l,t~,JI,

!,~AJ,t.~, I

2

x--

4

6

60

,I

h4 O

00.006004~~~!i'~i'i, )~I~i]~'r '"" ' . )f~. I'' . 0

-80

"0

0.0101

.

.

8 10 strain %

.

20 0 12

14

16

18

Fig. 5. Inflow and outflow volumes of permeant during the dynamic permeability test illustrated in Fig. 4. Each data point through which the curve is drawn represents the fluid entering or leaving the specimen during a ninety-second interval of the test.

activation of a dynamic component in the permeability. We have no reason to suspect in our experiments any change in the density of the sediment grains, so that, assuming such a component in the present experiments is wholly mechanical and not diagenetic, the volume difference must reflect a change in pore space. In other words, differences in volume between the inflow and outflow will give a quantitative assessment of the volume changes inferred from the microstructures. Figures 4 to 7 reveal the dynamic component of permeability and its significance for a ball-clay cylinder deformed to 16% axial strain. Figure 4 shows the dynamic permeability of the deforming clay, calculated in the conventional way from the outflow of permeant from the specimen. Figure 5 shows this outflow as a volume, together with the volume of inflow, which is a lesser value for most of the test. The dynamic permeability of Fig. 4 is therefore not the classical, Darcyan permeability but is incorporating a dynamic component associated with pore-volume reduction within the specimen. The magnitude of the dynamic component is given by the difference between the incoming and outgoing fluxes. During any interval for which the outflow exceeds the inflow the specimen can be taken to have decreased in volume by an amount equal to this difference (given that deformation is accommodated entirely by change in pore fluid volume). Figure 6

shows this dynamic component, expressed as the percentage specimen volume change within each 90s interval of the. entire test. Note that the majority of the data falls in the negative field, indicating that the specimen is undergoing bulk volume loss, at various rates. At the onset of axial loading, where the rate of mechanical dewatering (negative volume change) is at a maximum, the dynamic component reaches its maximum value (at 1% strain). The rate of porosity loss decreases steadily until 6% strain, where it levels off before rapidly increasing again. This latter, sudden increase in drainage rate coincides with the strain softening exhibited in the stress curve. Then, as deformation continues, the rate of volume decrease returns to a value similar to that between 2 and 6% strain, and then gradually decreases until it becomes positive at approximately 14% strain. At this point, the specimen is undergoing bulk dilation and permeant is being stored within the specimen. Here the dynamic component is negative, causing a reduction in the dynamic permeability. The result of removing the dynamic component from the dynamic permeability is given in Fig. 7. This curve is therefore a representation of the variation in Darcyan permeability with strain. Its shape resembles that of the dynamic permeability of Fig. 5, but in inverseforrn. This demonstrates that the conventional methods of determining permeability do not, when applied

122

E.L. S T E P H E N S O N E T A L .

0.002

_ stress

180

o.oo1 o~

0.000

"

-0.001

II

I

/ . . . . . . . i~.11 ~ ~

'~'~'

-160

i! 140 ql' ~ ~ ' ".!lJh --

ii ,

-120

tO

100 -0.002

E = > -0.003 .=

,

, ~1

~i

g

-80 I

'

60

-0.004

"~ =

40

-0.005

20 I

- 0 . 0 0 6 ---~

0

2

.

4

.

6

.

.

8 1'0 strain %

.

1'2

1~4

16

0

18

Fig. 6. Volume changes of the specimen during the test illustrated in Fig. 4. The changes are plotted as an interval bulk volume change, i.e. the percentage change of the initial cylinder volume within a ninety-second interval of the test. Almost all the values are negative, indicating overall volume loss, but at varying rates. Positive slopes (e.g. at 1-6% strain) indicate successively decreasing amounts of volume loss; steeper slopes represent more rapid rates of change. Decreasing rates of volume loss in the negative field may be accounted for by increasing Iocalised dilation at the same time as continuing bulk consolidation. Bulk dilation occurs only temporarily at around 14% strain. to an actively deforming material, measure its fluid transport capacity unless corrections are made for the contributions from the dynamic component. The explanation of the contrast between the Darcyan and the dynamic permeabilities lies in the pore-volume fluctuations that accompany the microstructural changes outlined earlier, and especially the localized effects associated with shear zone formation. For example, at around 1-3.5% strain, the Darcyan permeability is decreasing very slightly, because of the increased packing of the grains in response to the bulk strain. It then begins to increase as a result of dilation associated with the initial development of shear zones, which facilitate increased fluid flow. Even so, the dynamic permeability continues to decrease (Fig. 6), reflecting continuing bulk consolidation of the specimen. The localised, dilating shear zones eventually collapse (7% strain), prompting an increased total outflow, and a peak in the dynamic permeability curve (Fig. 4). The Darcyan component, however, decreases, due to the production of a low-porosity clay fabric within the shear zones, most of which are oriented at a high angle to the axial fluid flow and hence are tending to act as barriers. In ways such as these, the permeated

medium is far from being simply a matrix of static grains and pores that passively allows the movement of fluid, as is normally assumed in permeability work.

Conclusions The observations presented above demonstrate that the dynamic permeability, the permeability measured in an actively deforming material, contains a dynamic component in addition to the classical Darcyan permeability. The amount of fluid exiting from a material is not just a function of the standard Darcyan parameters, but is contributed to by the transmitting medium itself. In our experiments, the dynamic component arises from pore-volume fluctuations associated with microstructural changes during consolidation and deformation, but in nature a host of other mechanical and chemical contributions are possible. The conclusions will apply to other laboratory methods of determining permeability, such as the falling head, flow-pump, and consolidometry techniques, and also to in situ measurements of a natural dynamic situation. The most likely approach in the near future to measuring permeabilities in an accretionary prism will be

FLUID FLOW IN DEFORMING SEDIMENTS 4E-17-

_

123

stress

180 +160 140

3E-17

120 lOO

•~

80

11[''* I~arcyan permeability

0

.

2

.

.

4

. -

6

.

.

.

.

8 10 strain %

.

.

12

.

14

"0

40 r

1E-17 d

u~

20 0

16

18

Fig. 7. The Darcyan permeability of the test illustrated in Fig. 4. Differences in flux due to pore fluid fluctuations (the dynamic component) have been removed from the dynamic permeability using the data shown in Figs 5 and 6; the curve therefore shows the permeability of the material in the conventional sense. The variation in the capacity of the deforming sediment to transmit the permeant is due to microstructural changes within the specimen as a result of deformation.

an indirect derivation from the rate of fluidpressure dissipation and a variety of methods are under development, such as piezometric probes and wireline packers (Langseth et al. 1988). Any data obtained will lead to the dynamic rather than the conventional permeability. The amount of contribution from the dynamic components must ultimately be finite, but it could be significant during certain intervals of fluid flow. Our work is continuing on defining the timings and amounts involved in different circumstances. Bearing in mind that the Darcyan permeability itself varies with strain, the role of dynamic permeability becomes important in any situation where fluid flow through actively deforming sediments is being considered. The ideas have been presented here in the context of accretionary prisms, but they are relevant to numerous other geological settings. The concepts are also relevant to engineering applications, for example the use of clays for waste containment (Arch, Stephenson & Maltman, in prep.), because the linings of many repositories are not static but undergoing small but continuous mechanical movement. In all such dynamic situations, the amount of fluid leaving the system depends not only on the conventional permeability, but a contribution from the system itself.

References

ARCH, J. 1988. An experimental study of deformation microstructures in soft sediments. PhD thesis, University of Wales, Aberystwyth. MALTMAN, A.J. 1990. Anisotropic permeability and tortuosity in deformed wet sediments. Journal of Geophysical Research, 95, 9035-9045. BEHRMAN, J.H., LEWIS, S.D., MUSGRAVE,R. & 25 OTHERS. 1992. Proceedings of the Ocean Drilling Program, Initial Results, 141. - - & - - AND THE ODP LEG 14I SHIPBOARD SCIENTIFIC PARTY. 1993. Tectonic controls on migration of fluids at the South American convergent margin near the Chile triple junction. Geofluids '93 extended abstracts, Torquay, 175177. BERNER, U. & FABER, E. 1993. Light hydrocarbons in sediments of the Nankai accretionary prism (Leg 131, Site 808). Proceedings of the Ocean Drilling Program, Scientific Results, 131, College Station, Texas (Ocean Drilling Program), 185-195. BROWN,K. &:MOORE,J.C. 1993. Comment on Arch & Maltman, Anisotropic permeability and Tortuosity in deformed wet sediments. Journal of Geophysical Research, 98, 17860-17864. BYRNE,T., MALTMAN,A.J., STEPHENSON,E., Son, W. & KNIPE, R. 1993. Deformation structures and fluid flow in the toe region of the Nankai accretionary prism. Proceedings of the Ocean Drilling Program, Scientific Results, 131, 83-101.

124

E.L. STEPHENSON ET AL.

CARSON, B., HOLMES, M.L., UMSTATTD, K., STRASSER, J.C. & JOHNSON, H.P. 1991. Fluid expulsion from the Cascadia accretionary prism: evidence from porosity distribution, direct measurements, and GLORIA imagery. Philosophical Transactions of the Royal Society, A335, 331-340. DAVIS, D.M., SUPPE, J. & DAHLEN, F.A. 1983. The mechanics of fold-and-thrust belts and accretionary prisms. Journal of Geophysical Research, 88, 1153--1172. FAAS, R.W. & CROCKETT, D.S. 1983. Clay fabric development in a deep-sea core: Site 515, Deep Sea Drilling Project Leg 72. Initial Reports DSDP, 72,519-525. FRYER, P., PEARCE, J.A., STOX
Proceedings of the Ocean Drilling Program, Scientific Results, 125, College Station, Texas (Ocean Drilling Program). GIESKES, J.M., BLANC, G., VROLIJK, P., ELDERFIELD, H. & BARNES, R. 1990. Interstitial water chemistry-major constituents. Proceedings of the Ocean Drilling Program, Scientific Results, 110, 155-178. KASTNER, M., ELDERFIELD, H. & MARTIN, J.B. 1991. Fluids in convergent margins: what do we know about their composition, origin, role in diagenesis and importance for oceanic chemical fluxes?

Philosophical Transactions of the Royal Society, A335, 243-259. KNIPE, R.J. 1986a. Faulting mechanisms in slope sediments: examples from Deep Sea Drilling Project cores. In: MOORE, J.C. (ed.) Structural

Fabrics in Deep Sea Drilling Project Cores from Forearcs. Geological Society of America -

-

Memoirs, 166, 45-54. 1986b. Microstructural evolution of vein arrays preserved in Deep Sea Drilling Project cores from the Japan Trench, Leg 57. In: MOORE,J.C. (ed.)

Structural Fabrics in Deep Sea Drilling Project Cores from Forearcs. Geological Society of America Memoirs, 166, 75-87. AGAR,S.M. & PRIOR,D.J. 1991. The microstructural evolution of flow paths in semi-lithified sediments from subduction complexes. Philosophical Transactions of the Royal Society, A335, 261-273. LANGSETH,M.G. & MOORE,J.C. 1990. Introduction to special section on the role of fluids in sediment accretion, deformation, diagenesis, and metamorphism in subduction zones. Journal of Geophysical Research, 95, 8737-8742. --, VON HUENE, R., NASU, N. & OKADA, H. 1981. Subsidence of the Japan Trench forearc region of Northern Honshu. Oceanologica Acta, 4, supplement, 173-179. LANGSETH,M. et al. 1988. The role offluids in sediment

--,

accretion, deformation, diagenesis and metamorphism at subduction zones. A planning document for an international program of cooperation. NATO/NSF. , WESTBROOK, G.K. & HOBART, M.A. 1990. Contrasting geothermal regimes of the Barbados

Ridge complex. Journal of Geophysical Research, 95, 1049-1061. LE PICHON,X., KOBAYASHI,K. ANDTHEKAIKO-NANKAI SCIENTIFIC CREW. 1992. Fluid venting activity within the eastern Nankai Trough accretionary wedge: a summary of the 1989 Kaiko-Nankai results. Earth and Planetary Science Letters, 109, 303-318. MALTMAN,A.J. 1981. Primary bedding-parallel fabrics in structural geology. Journal of the Geological Society, London, 138,475-483. 1987. Shear zones in argillaceous sediments, an experimental study. In: JONES, M.E. & PRESTON, R.M.F. (eds) Deformation of Sediments and Sedimentary Rocks. Geological Society, London, Special Publications, 29, 77-87. - 1988. The importance of shear zones in naturally deformed wet sediments. Tectonophysics, 145, 163--I75. - - , BYRNE, T., KARIG, D.E. & LALLEMANT,S. 1992. Structural geological evidence from ODP Leg 131 regarding fluid flow in the Nankai prism, Japan. Earth and Planetary Science Letters, 109, 463468. - - , KNIPE, R. & PRIOR, D. 1993. Deformation structures at Site 808, Nankai accretionary prisms, Japan. Proceedings of the Ocean Drilling Program, Scientific Results, 131, 123-133. MASCLE, A., MOORE, J.C. AND THE ODP LEG IIO SHIPBOARDSCIENTIFICPARTY.1988. Proceedings of the Ocean Drilling Program, Initial Reports, 110, College Station, Texas (Ocean Drilling Program). MOORE, J . C . & Buu-DUVAL, B. 1984. Tectonic synthesis Deep Sea Drilling Project Leg 78A: Structural evolution of offscraped and underthrust sediment, northern Barbados Ridge complex. Initial Reports of the Deep Sea Drilling Project. US Government Printing Office Washington, DC, 78A, 601-621. - & LUNDBERG, N. 1986. Tectonic overview of Deep Sea Drilling Project transects of forearcs. In: MOORE,J.C. (ed.) Structural Fabrics in Deep Sea Drilling Project Cores from Forearcs. Geological Society of Ameria Memoirs, 166, 1-12. --, BROWN, K.M., HORATH, F., COCHRANE, G., MACKAY, M. & MOORE, G. 1991. Plumbing accretionary prisms: effects of permeability variations. Philosophical Transactions of the Royal Society, A335, 275-288. PEACH, C.J., SPIERS, C.J., TANKINK, A.J. & ZWART, H.J. 1987. Fluid and ionic transport properties of deformed salt rock. Nuclear science and technology series, report EUR 10926. Office for Official Publications of the European Communities, Luxembourg. PLATr, J. & LEGGETT,J. 1985. Large-scale underplating in the Makkran accretionary prism, SW Pakistan. Geology, 13,507-511. SAMPLE, J.C. 1990. The effect of carbonate cementation of underthrust sediments on deformation styles during underplating. Journal of Geophysical Research, 95,9111-9121.

FLUID FLOW IN DEFORMING SEDIMENTS SUESS, E., VON HUENE, R. & THE ODP LEG II2 SHIPBOARDSCIENTIFICPARTY. 1988. Proceedings of the Ocean Drilling Program, Initial Reports, 112, College Station, Texas (Ocean Drilling Program). TAIRA, A., HILL, I. &. 27 OTHERS. 1992. Sediment deformation and hydrogeology of the Nankai Trough accretionary prism: synthesis of shipboard results of ODP Leg 131. Earth and Planetary Science Letters, 109,431-450. , --, & 26 OTHERS. 1991. Proceedings of the Ocean Drilling Program, Initial Results, 131. TAYLOR,E. & FISHER, A. 1993. Sediment permeability at the Nankai accretionary prism, Site 808.

Proceedings of the Ocean Drilling Program, Scientific Results, 131,235-243. & LEONARD,J. 1990. Sediment consolidation and permeability at the Barbados forearc. Proceed-

ings of the Ocean Drilling Program, Scientific Results, 110,289-308. TRIABLE, J.S. 1990. Clay diagenesis in the Barbados accretionary complex: potential impact on hydrology and subduction dynamics. Proceedings of the Ocean Drilling Program, Scientific Results, ll0, 97-110. VON HUENE, R. 1984. Tectonic processes along the front of modern convergent plate margins research of the past decade. Annual Reviews of the Earth and Planetary Sciences, 12,359-381. 1985. Direct measurement of pore fluid pressure, Leg 84, Guatemala and Costa Rica. In: VON

125

HUENE, R. et al. Initial Reports of the Deep Sea Drilling Project. US Government Printing Office, Washington, DC, 84. --, &: LEE, H.J. 1983. The possible significance of pore fluid pressures in subduction zones. In: WATKINS, J.S. • DRAKE, C.L. (eds) Studies in continental marine geology. American Association of Petroleum Geologists Memoirs, 34, 781-789. - - , AUBOIN,J. & 12 OTHERS. 1985. Initial Reports of the Deep Sea Drilling Project. US Government Printing Office, Washington, DC, 84. VROLIJt<, P. & SHEPI'ARD, S.M.F. 1991. Syntectonic carbonate veins from the Barbados accretionary prism (ODP Leg 110): record of palaeohydrology. Sedimentology, 38,671-690. WES~ROOK, G.K. 1991. Geophysical evidence for the role of fluids in accretionary wedge tectonics.

Philosophical Transactions of the Royal Society, A335, 227-242. & SMITH,M.J. 1983. Long decollements and mud volcanoes. Evidence from the Barbados Ridge complex for the role of high pore fluid pressure in the development of an accretionary wedge. Geology, 11,279-283. - - . , CARSON, B., MUSGRAVE, R. AND THE ODP LEG 146 SHIPBOARDSCIENTIFICPARTY. 1993. Fluid flow within a convergent continental margin - results from ODP Leg 146, Cascadia margin. Geofluids '93 extended abstracts, Torquay, 178-180.

Fluid-flow processes and diagenesis in sedimentary basins KNUT BJ~RLYKKE

Department of Geology, Box 1047, University of Oslo, 0316 Oslo 3, Norway Abstract: Flow of fluids in sedimentary basins causes transport of heat and dissolved mass and is therefore potentially important in relation to diagenetic reactions. Rather large fluid fluxes are required, however, for this type of transport to be significant in terms of dissolution and precipitation of minerals. Before diagenetic processes are attributed to pore water flow, semi-quantitative calculation of flow rates and duration of flow should be attempted. Heat flow in sedimentary basins is normally dominated by conduction, except on a local scale. The upwards pore water flux due to compaction is on average smaller than the subsidence rate, and the thermal anomalies caused by compaction are moderate in modern basins. Dissolution and precipitation of minerals which are in equilibrium with the pore water occur when the flow is oblique relative to the isotherms. Due to the low solubility/temperature gradient of the common silicate and carbonate minerals, very large fluxes of pore water are required to transport significant volumes of mass in solution. The greatest potential for transporting solids and creating secondary porosity is during meteoric water flow, because the flow rate then may be several orders of magnitude higher than what is typical for compaction-driven flow. Quartz overgrowth corresponding to 3% of the rock volume requires a total flow of 108cm3 cm -2 if the silica is introduced by vertical compaction-driven flow from a source outside the sandstone. Pore water flow precipitating quartz due to cooling will dissolve carbonate cement at a much higher rate. An external import of silica through the flow of water would be expected to cement up the most permeable pathways, such as fractures and well-sorted permeable sand beds. In the case of carbonate cement the solubility gradient is negative and upwards flowing and cooling pore water would cause dissolution rather than precipitation. Also in the case of carbonate rocks, large-scale mass transfer such as dolomitization is easier to explain as occurring at relatively shallow depths rather than during deeper burial. Ore minerals like galena have a steep solubility/temperature gradient, but concentrated precipitation requires rapid cooling of hot water. Such conditions are most likely to be met when hot fluids are cooled near the surface.

Fluid flow in sedimentary basins has the capacity to transport solids in solution and heat. This may produce thermal anomalies and cause dissolution and precipitation of minerals in reservoir rocks and thus influence diagenetic reactions. At any time, however, the amount of solids in solution in pore water is very small compared to the solid phases. This is particularly true for elements like silica and aluminium, where the concentrations are controlled by the low solubility of silicate minerals. Precipitation of significant volumes of minerals therefore has to be accompanied by dissolution. It is important to determine the typical distances travelled by the ions from the sites of dissolution to the points of mineral precipitation. If pore water is stationary relative to the minerals, transport of solids in solution only occurs by diffusion. The rates of diffusion in sedimentary rocks are functions of the concentration gradients, temperature and the diffusion coefficient of the matrix. The concentration gradients of ions, dissolved in the pore water, is

controlled by the solubility of the mineral species present and the kinetics of dissolution and precipitation. The concentration gradients must be rather low when the minerals are ubiquitous like quartz and illite and frequently also feldspars. The formation water in Jurassic North Sea reservoirs appears not to be far from equilibrium with c o m m o n minerals like quartz, albite, illite and chlorite (Aagaard et al. 1992). High chlorinity may be due to dissolution of evaporite minerals or formed as a residual brine in arid conditions (Egeberg & Aagaard 1989; Egeberg 1992). High salinities will reduce the activity of cations and thereby influence the solubility of silicate and carbonate minerals. Pore waters in sedimentary basins are never stationary relative to minerals, and flow of pore water will always cause some degree of dissolution and precipitation. The low solubilities and solubility/temperature gradients of c o m m o n silicate and carbonate minerals implies, however, that very large fluxes are required to transport enough solids to introduce significant

FromPARNELL,J. (ed.), 1994, Geofluids:Origin, Migrationand Evolutionof Fluidsin SedimentaryBasins, Geological Society Special Publication No. 78, 127-140.

127

128

K. BJORLYKKE

volumes of cement from an outside source into sandstones or limestones. Although meteoric water flow may be significant in terms of influencing the geothermal gradients in sedimentary basins (Chapman 1987), heat transport is mainly controlled by conduction, particularly in the deeper parts of sedimentary basins where flow rates are much lower (Bethke 1985; Demming et al. 1990; Harrison & Summa 1991). To the extent that pore water flow does produce thermal anomalies, it will also influence the temperatures of diagenetic reactions. This paper will focus, however, on the effects of pore water flow on diagenetic reactions. It also presents calculations of the order of magnitude of pore water fluxes required to transport significant volumes of solids in solution and to influence diagenetic reactions. A fuller discussion of fluid flow in sedimentary basins is published by Bethke (1989), BjCrlykke (1993) and is also addressed in other papers in this volume.

Dissolution and precipitation of minerals resulting from fluid flow The solubility of minerals varies with temperature, and the common silicate minerals exhibit increased solubility with increasing temperature. Mineral solubilities are also influenced by the concentration of components other than those present in the minerals, and the salinity (or chlorinity) of the pore water influences the solubility significantly. The calculations of mineral solubilities in pore water are rather complex and will not be discussed here. However, pore water flow, which occurs at an angle to the isotherms, will cause the pore water to be heated or cooled and this will in itself cause dissolution or precipitation of minerals. Assuming equilibrium between the pore water and the minerals, the volume of minerals dissolved or precipitated can be calculated (Wood & Hewett 1984; Bj0rlykke & Egeberg 1993): Vc = FtsinfS(dT/dZ)etT/p

(1)

Vc is the volume of cement precipitated, F is the total flux of pore water (cm3/cm2s), t = time (s), 13 is the angle between the direction of flow and the isotherm, d T / d Z is the geothermal gradient, OL~ris the solubility-temperature function (transfer coefficient) and p is the density of the mineral. Flowing from higher to lower temperatures (upwards), silicate minerals, which usually have a progressive solubility, will precipitate, while carbonate minerals like calcite, which

Depth (z) .

...,.:.,

....

;:::PoreWater Flux(F) :i: cm3/cm 2 ~i?ii

Solubility Gradient of quartz (aT) 1-3 ppm/°C (3 X10-6/°C) ]

Fig. 1. Precipitation of quartz resulting from vertical flow of pore water in equilibrium with quartz. If the pore water is in equilibrium with quartz, the rate of precipitation of quartz is a function of the rate of cooling and the solubility gradient. Assuming that the geothermal gradients are constant and the isotherms horizontal the rate of cooling is a function of the geothermal gradients and the flow rate. The total volume of quartz precipitated by fluid flow through a sandstone can then be calculated according to equation I (Vc = F~(dT/dZ) aT/p). Assuming that the geothermal gradient (dT/dZ) is 30° C km -1, aT = 10-60 C -1 , and the density of quartz (p) 2.7gcm -3, Vc = Ft 3 x 10-9. It follows from this calculation that a total flow (F, = pore water flux integrated over time) of 3 x 108cm 3 cm -2 is required to precipitate 10% quartz cement (Vc = 0.1). This means that a vertical flow corresponding to a water column of 3000 km (30 000 km if the porosity is 10%) must flow through each volume of the reservoir to precipitate 10% quartz cement by volume. have retrograde solubility functions, will dissolve. For quartz, Ot-r ranges from 1 ppm ° C -1 at 75 ° C to 3 ppm ° C -1 at 140 ° C (Wood 1986). The density P of quartz is 2 . 7 g c m -3 and the geothermal gradient is taken to be 30 ° C km -1 (3 x 10 -4° Ccm-1). To precipitate one volume of quartz the component perpendicular to the isotherm must be 3 × 109 cm 3 cm -2, if the pore water is in equilibrium with quartz. For 1% quartz to precipitate (Vc = 0.01) the vertical component of the total flow integrated over time has to be 3.107 cm 3 cm -2. This is comparable to a water column of 30 km (Fig. 1). It is clear that pore water fluxes of this magnitude can not be obtained by compactiondriven flow except by extreme focusing of the flow. The total volume of water contained in a 5 km deep basin is approximately 105 cm 3 cm -2, assuming an average of 20% porosity for the whole sequence. Most of the quartz cementation in North Sea sandstones takes place at burial depths below 2.5-3.0 km (Bj0rlykke et al. 1992; Giles et al. 1992). The volume of pore water available in the basin below this depth is small

FLUID FLOW AND DIAGENESIS

129

.......................................

D i s s o l u t i o n of quartz

~

r. - -_ .- - -. - -_- - - - - _-_-_-_-. _ - _ - _ - ~ _ - _ - _ - _ - _ - _ - _ - _ - .

~

: -

~ M-Oa~iBh-~~ : - : - : = - z _ - -

~.~-~-~-~------\------~-~-~\~\~Z~Z~\~\~\~Z~\~\~\~\~\~\~Z~Z~Z~Z~Z~Z~Z~\~Z~\~Z~Z~Z~ Permeability .---=--=-----------~---------~------------~------------~-~----~------~--~jjIjjjjjjjj~jj~ Profile

Sandstone Be~'-I d t'~ (coarsening

uartz [cite Fig. 2. Convection cells in a sandstone bed will cause quartz cement to be precipitated and calcite to be dissolved where the flow has an upwards component so that the pore water is cooled. The downwards directed flow will cause precipitation of calcite and dissolution of quartz. In the case of Non-Rayleigh convection, the angle of the direction of the flow relative to the isotherm determines the rate of precipitation or dissolution. The rate of quartz precipitation relative to the rate of calcite dissolution depends on the depth (pressure) and on the pH. Calculations show that for pH between 5 and 6 and depths between 2 and 3 km, upwards-flowing (cooling) pore water will dissolve 30-100 times the volume of quartz precipitated (Bj0rlykke & Egeberg 1993). Any remaining calcite in a reservoir sandstone would then be evidence of a rather limited advective supply of silica. The rates of pore water flow by thermal convection are functions of the permeability of the sediment layer (sandstone) and the thickness of the layer. Even thin shale layers will cause the flow to separate and reduce the effective height of the convection cell and thereby the velocity reducing effect on diagenesis. (Bjorlykke et al. 1988)

since the shales have already been compacted and dewatered to a large degree. Assuming an average of 5-10% porosity the volume of pore water is less than 2 × 104 cm 3 cm -2. The pore water therefore has to be focused by a factor of at least l0 s before sufficient fluxes can be obtained to introduce significant quartz cementation in sandstones. This calculation is meant to illustrate the relatively small volumes of pore water available in sedimentary basins at deeper burial. It can be shown that, as an average for a sedimentary basin, the upwards flow of compaction-driven pore water is always lower than the subsidence rate. This means that if the geothermal gradients are constant the pore water will be heated due to subsidence and its solubility will be increased with respect to silicate minerals (Bj0rlykke 1993). Since carbonate minerals such as calcite have retrograde solubilities, cooling of pore water will generally cause dissolution of carbonate rather than precipitation (Wood & Hewett 1984). Locally, however, high gradients in the CO2

~1~

J~

I ~ "~Primary

Lateral flow proportional to the permeability

/ upwards) ~ ;-~ I ~.~:~_~_#_e_-~#_-" ~:}~@~__-_-}}~}~:-_:__--_:~_e_-~ ~e_>-_>-~ ~_~ -~-=~==_==~_-~---=~-~------: ! -_--..Nu~g Fl~w ~ ~ ! ~(precipitationof ~ - _ ~ ~:_~.~E_-_~ ~_2-'quartzdue to ~ Hard Brittle Shale :-'~:_-."cooling) E~=E: ............ -=-: -__.--==]

After some precipitation of quartz cement by advective flOW

==-==~=-==~:~.=-= =~=~=~===~ ==-~~ ~: =_=-__,__::_~_-_~c~==_=_=:=:-:=:====~_==i

Sea Floor :=-------_-_--__----_-_--_-=_-_--=_.2_------; ............................... , ~Shai ~ - _ - - - - - - - - - - 1 L~_-_-..-_-_-_-_-_-_-_-_-~ -_-_-_-_-_--.-_-_-..-_-_-_-_-_-_-~ Precipitation t_=.-------------------.--------------------=========== Flux • Sin (X

ii::iii::i!;}'.;iii:i:iii!iiiii::iii!iii:::ii!iiii;.i .: Isotherm I ~ l ~ - - - - - - - - - - - - - - I- ~------------------':

~

lsotherm

Fig. 3. (a) Illustration of pore water flow from a fracture into adjacent sandstone beds. The rate of quartz cementation will be proportional to the flux of water and the rate of cooling. Most of the quartz cementation will occur in and close to the fracture where the flux is highest. The most permeable layers would receive the highest pore water flux and become cemented up first. (b) When fluids are flowing along a fracture or fault into a sandstone, the rate of cooling is a function of the flow rate and the angle between the direction of flow and the isotherms. Most of the precipitation in the sandstone will occur near the fracture since the angle cxbetween the isotherm and the bed becomes small in the case of relatively horizontal beds.

pressure may produce exceptions to this. At p H 5-6, the solubility gradient for calcite is steeper than for quartz, and as a result, the ratio of carbonate dissolution to quartz cementation in a rising and cooling volume of pore water is high (10-300) (Bj0rlykke & Egeberg 1993). The presence of even small amounts of carbonate cement in sandstones may therefore be used as evidence against introduction of quartz cement (Fig. 2). The source of quartz cement has been the subject of considerable discussions and several authors have suggested that quartz is introduced from an outside source through fluid flow (Gluyas & Coleman 1992; Robinson & Gluyas 1992; Grant & Oxtoby 1992). In addition

130

K. BJORLYKKE

would be reduced (Fig. 3a). This is not what is commonly observed. When water flows into sandstones from faults and fractures the rate of cooling is reduced and most of the precipitation occurs close to the fracture. Particularly in the case of relatively horizontal beds, the angle (a) between the isotherm and the direction of flow will be small, reducing the rate of precipitation for a given flux (Fig. 3b). The porosity distribution in sandstone reservoirs as recorded by core analyses or based on log measurements shows that the porosity is often rather unevenly distributed. Even in sandstones with relatively low porosities, thin beds with high porosities are often observed (Fig. 4). Adjacent beds have experienced the same temperature/pressure history during burial. When they show very different porosities and permeabilities, this is evidence of a compositional and textural control on compaction and quartz cementation. This is consistent with a model where most of the quartz cement is derived internally by pressure solution and other local diagenetic reactions. Mineralogical and textural parameters may also influence rates of precipitation of silica from pore water introFig. 4. Porosity distribution in sandstones buried to duced from an outside source. If the quartz more than 4000 m depth offshore Norway. The log cement had been derived from an outside source derived porosity (CPI log) to the left shows large by pore water flow, the pore water flux would be variations in porosities in the sandstone. The several orders of magnitude higher in the permeabilities (x) measured in the core show that permeable beds than in the low permeability highly permeable layers exist in the middle of much beds. It would then be difficult to explain how more cemented sandstones. The most permeable silica was transported into the low permeability layers have permeabilities of several hundred zones before the permeable zones had lost all millidacus while most of the more well-cemented sandstones have permeabilities around I mD or less. their porosity and permeability. This transport The rate of pore water flow would be about three would have had to be by diffusion and this orders of magnitude higher in the permeable assumes highly supersaturated pore water in the sandstone. Most of the quartz cement introduced main aquifers to provide a high concentration from an external source should therefore be expected gradient. The most permeable beds must then be to precipitate in these layers. characterized by clay coatings on quartz grains or other features inhibiting quartz cementation. Sandstones will continue to be cemented with to the problem of the large fluxes required, this increasing burial and temperature until nearly model would suggest that the rate of quartz precipitation was proportional to the vertical all porosity is lost and the permeability reaches extremely low values. This is easier to explain if component of the flux. The most permeable pathway for fluid flow would then receive most the silica is derived from an internal source. The rate of quartz cementation in a sandstone of the cement which is precipitated from solution due to cooling. This could be beds with will depend to a large extent on the rate of exceptionally high porosity and permeability, or subsidence and heating, and the time factor fractures and faults. These pathways will be probably plays a role (Bloch et al. 1990). In the Gulf Coast basin, Pliocene rocks start to become cemented up before significant cement can be introduced to the less permeable rocks, which quartz cemented at about 4 km depth while this receive a much lower pore water flux. In a occurs at about 3 km in older sedimentary rocks (Harrison & Summa 1991). In the North Sea and graded bed, the high permeability zone would be the Mid-Norway shelf (Haltenbanken) the Plioeither near the top (shallow marine sediments) or near the base (i.e. fluvial channels), and most cene/Pliestocene subsidence may exceed 1 km. A reservoir may then have subsided from 3 to of the quartz cementation would be expected to occur there such that the primary differences 4 k m and much of the most intense quartz

FLUID FLOW AND DIAGENESIS cementation would have occurred over a short period of time. Periods of rapid subsidence will also be characterized by high rates of oil generation and these processes are often linked (Gluyas et al. 1993). This may be mainly due, however, to the control temperature exerts on both processes and does not support the idea that quartz cementation is episodic in nature. There is also considerable evidence that quartz cementation may continue after oil emplacement in a reservoir (Walderhaug 1990; Saigal et al. 1992). At oil saturations above 50-60% the relative permeability of water is close to zero and import of silica through flow of pore water would then be more difficult.

Thermal convection Thermal convection has been suggested by several authors as a mechanism to move solutes through the subsurface (Wood & Hewett 1982, 1984; Davis et al. 1985; Cassan et al. 1981; Haszeldine et al. 1984). Rayleigh convection requires relatively thick sequences (100-300 m) of homogeneous and porous sandstones for the critical Rayleigh number to be exceeded (Bjerlykke et al. 1988). Mathematical modelling of Rayleigh convection has shown that relatively thin shales (0.1 m) or cemented intervals within a sandstone sequence will effectively divide potentially large convection cells into smaller ones, which may then be too small to exceed the critical Rayleigh number (Bj0rlykke et al. 1988). If the critical Rayleigh number is exceeded by as little as 10%, however, pore water flow due to Rayleigh convection will be fast enough to dissolve and precipitate 10% of the quartz within 10 Ma (Palm 1990). This demonstrates that if Rayleigh convection is taking place, the effects on diagenesis would be fast enough to be significant relative to diagenetic processes, but this is probably relatively rare. Non-Rayleigh convection will always take place when the isotherms are not perfectly horizontal. Slight slopes in the isotherms will be caused by dipping beds with different conductivity (Davis et al. 1985). The velocity of non-Rayleigh convection is proportional to the slope of the isotherm but also to the height of the convection cell. Where convection is confined to a few metres-thick sandstone beds separated by shales, the flow velocities will be too slow to be significant in terms of mass transfer during diagenesis (Bjcrlykke et al. 1988). A test would be that all carbonate minerals should be dissolved where significant quartz cementation occurs and that extensive corrosion of quartz and carbonate cementation should occur in other parts of the sequence (Fig. 2).

131

The salinity distribution observed in the North Sea Basin from formation analyses and from resistivity logs suggests that there is an increase in salinity with depth towards underlying evaporites. These salinity gradients show that there is at least at the present day no large-scale convective flow, since this would have homogenised the salinity (Gran et al. 1992).

Flow of water along fractures and permeable faults In a sedimentary basin with alternating sandstones and shales, there is a high permeability contrast between the two lithologies. In basin modelling, tight shales are assumed to have permeabilities which may be as low as 10-1°D, while the permeability of deeply buried sandstones may typically be 1-0.01D (Mudford et al. 1991). Faults and fractures will, therefore, to a very large extent control cross-formational flow. Fractures are continuous void space which may be important conduits for fluid flow. They are normally formed by tension set up during tectonic deformation or due to uplift and unloading. Relative movements along faults may also have a tensional component, so that voids similar to fractures are produced. Fractures are temporary features which after some time will be destroyed by mechanical deformation or become cemented up by precipitated minerals. For a fracture to remain open the rocks around the fracture must have sufficient shear strength to resist the stress imposed by the horizontal stress. Loose sands and soft mudstones therefore have limited potentials for preserving open fractures, since these rocks do not have sufficient strength to resist the horizontal stress. Sand normally does not become significantly quartz-cemented at temperature below about 80°C (2.0-2.5 km burial) and are then loose and poorly indurated when not cemented by carbonate or other types of early cement (Bj0rlykke et al. 1992). Carbonate rocks are more commonly subjected to early cementation and frequently develop a high shear strength capable of preserving fractures at shallow depths. Fractures will cement up by two different mechanisms. (1) Transport by diffusion of solids from surrounding rocks and precipitation of minerals in to the fracture. Minerals in the rocks on both sides of the fracture are under stress from the overburden, and, in some cases, also tectonic stress. Minerals such as quartz and calcite are

132

K. BJORLYKKE Water Boiling Temp (°C) 100 150 200 250 300 '

'

0 -50

Boiling

"x

- 1 oo \~

-15o ~\

- 2 0 0 ®o "10

~

- 2 5 0 -~ - 3 0 0 "-" ~

- 350

\\ ::

-400 -450 -500

Fig. 5. Hot water flowing on fractures will be cooled rapidly near the surface by mixing with ground water and by boiling. At 200° C boiling starts at about 130 m under hydrostatic pressure. This will favour concentrated precipitation of ore minerals like galena. The generated steam may cause severe corrosion of quartz and other minerals. then more soluble than minerals in the fracture, which are only subjected to pore pressure and no stress. There is therefore a thermodynamic drive for minerals to dissolve in the surrounding rock and to precipitate in fractures. The rate of this process, however, is difficult to calculate. (2) Precipitation of minerals by fluid flow through the fracture. The rate of precipitation can, in this case, be calculated, assuming different values for fluid flow rates, the temperature field (geothermal gradients), and mineral solubilities. It is also important to attempt to constrain the fluid flow rates and total fluxes along fractures in sedimentary basins. It is therefore necessary to examine the source of fluids and the mechanisms driving fluid flow.

Precipitation of sedimentary ores Even in very rapidly subsiding sedimentary basins with high sedimentation rates like the Gulf of Mexico, compaction-driven water does not produce very steep thermal anomalies and most of the heat flow is by conduction rather than advection (Harrison & Summa 1991). This is also the case for the North Sea (Hermanrud et al. 1991). This is because the compaction-driven pore water flow rate is on average lower than the subsidence rate and because open permeable fractures are rare in softer, relatively ductile sediments (Bjorlykke 1993). Flow rates on faults and fractures not extending up to the surface or sea floor will be reduced if the flow has to continue through low-permeability shales or

mud to reach the surface. A requirement for abnormally hot fluid to flow on fractures is that the flow rate is high so that most of the heat is not lost by conduction to the surrounding rocks. Thermal convection may produce high pore water fluxes but the rate of cooling and heating would be slow and gradual. Ore minerals and quartz would be precipitated along the limb of the convection cell where cooling occurs. Convective flow would, however, only redistribute what is present in the rocks inside the convection cells. Concentrated precipitation of dissolved minerals requires a change in the chemical environment by reactions with other fluids or minerals. We have shown that mixing of fluid is not likely to be very effective in the case of compactiondriven flow. The solubility, particularly of sulphides, is very temperature dependent and rapid cooling is the most effective way of precipitating minerals like galena and other sulphides (Deloule & Turcotte 1989). At depth rapid cooling of water flowing on fractures is difficult because the heat can not be disseminated away from the fracture fast enough because of the low thermal conductivity of the rocks. Near the surface hot fluids can be cooled more rapidly because the heat loss is higher. Fractures terminating on the sea floor will cause mixing with ocean water and precipitation of sedimentary ores. Sea water also represents a source of sulphur. Intersection with an open karstic system in limestone may also cause rapid cooling and precipitation of Mississippi Valleytype ore deposits (Bj0rlykke 1993). Also, cold

FLUID FLOW AND DIAGENESIS ground water in porous sandstone may cause rapid cooling. Boiling at 200 ° C will occur at 120-130 m depth and add to the heat loss (Fig. 5). Generation of steam may cause very strong corrosion of quartz and other silicate minerals. In the case of fractures which are terminated at some depth and overlain by a several hundred metre thick sequence of soft muddy sediments, the flow rate on the fracture is limited by the permeability of the overlying sediments because the flow has to continue to the surface. This means that the permeability on the faults and fractures in the brittle rocks may not be rate-limiting for the flow rate. The highest water flow rates and the most efficient precipitation of ores are therefore likely to occur near the top of brittle rocks before thick clay-rich sediments have been deposited. Compaction-driven pore water flow in sedimentary basins does not normally produce large thermal anomalies which would cause rapid cooling of fluids. Even in sedimentary basins with high sedimentation rates like the Gulf Basin thermal anomalies are moderate (Harrison & Summa 1991). Thermal anomalies associated with salt domes and overpressured shales are small compared to hot springs in basement rocks where brittle rocks are more likely to cause brecciation and sustain permeabilities on faults and fractures.

133 RAINFALL 100 cm / yr 108cm / m.yr

__

~

--~~--J.

. ~ ~ "

".:.::.~ I A

~UPWARDS FLOW ~ ~ / " 5 km= "~,~1¢ DUE TO COMPACTION % / . / ~

~'+. . . .

"''''+++"

I
I

--I

I ~

+

-~/'++ ***+÷*÷'.+**.+.

NO UPWARDSFLOW

.I "RELATIvE TO / / ~ , / . * + * ÷ * ÷

,L ~

SEA F .

•

+

.

L O O ~ ÷ ÷

÷

•

+

+

TEMP.

#~/X.2~

+

÷

; *~77: * , *, *+ ++ "+ .

÷

.

+

.

.

+

+

÷

PRECIPITATION OF KAOLINITE ~ -~,--\

DISSOLUTION OF FELDSPAR

20-50° × ~ ~ ' x C R L M AY

~}A~EME'N-r" ÷ + ÷ ÷ +

~ "

+.,

r'~ ~ _ ~ J - - / ~ - "

._~.%~_2 R +V n 2.U==.~7-/~)~~

/

2 K++4H4

SiO4

AI 2 Si205 (OH)4

KAI3Si3Olo(OH)2 +2SIO2 +H20 ILLITE F i g . 6. ( a ) T h e f l o w o f m e t e o r i c w a t e r i n t o

Pore water flow at low temperatures: the importance of kinetics The flux of meteoric water through sediments depends on the rainfall, the percentage of infiltration into the ground water, the ground water head, and on the permeability of aquifers their distribution and lateral continuity. It is also clearly seen that the integrated pore water flux flowing through each volume of sediments is inversely proportional to the sedimentation rate (Fig. 6a) Particularly at relatively low temperatures, the kinetics of mineral reactions becomes very important. In the case of quartz the precipitation rate below 80°C becomes negligible, even on geologic time scale, so that opal A and opal CT may exist as meta-stable phases (BjCrlykke & Egeberg 1993). Meteoric water flow in sedimentary basins at shallow depth (<1 km) may be several orders of magnitude higher than for compaction-driven flow (Giles 1987; Harrison & Summa 1991). Although modelling suggests that meteoric water flow may reach 2-3 km into basins like the Gulf of Mexico (Bethke et aI. 1988), the highest flow rates are likely to be found in the shallowest

sedimentary basins depends ultimately on the supply to the ground water by rainfall. Over 1 Ma, the rainfall may be 108cm 3 cm -2 and even if the infiltration is only 10% the supply to the ground water is high. The flux of meteoric water depends on the positions of the aquifers and it will decrease away from the area of recharge. (b) Dissolution of Kfeldspar and precipitation of kaolinite is a reaction which requires a throughflow of pore water to continue so that the reaction products K+ and silica can be removed. If all the K-feldspar is dissolved during meteoric water flushing at shallow depth, illite cannot form at greater depth (3.5-4.5 km) unless there is an external source of potassium. and most proximal parts (Harrison & Summa 1991). Meteoric water flow will dissolve amorphous silica present as fossils or volcanic debris in the sediments, causing the silica concentration to build up since quartz fails to precipitate. Dissolution of feldspar and the formation of kaolinite requires a through-flow of pore water since the reaction is not isochemical: 2KA1Si3Os + 2H + + 9H20 AlzSizOs(OH)4 -I- 2K + + 4H4SIO4 If this reaction is to proceed, potassium and silica must be removed so that the K+/H + ratio

134

K. BJORLYKKE

and the silica concentration remains inside the stability field of kaolinite (Garrels & Christ 1965). Plagioclase and albite will dissolve in a similar way and the reaction then depends on the Na+/H ÷ ratio. Much higher Na+/H ÷ ratios can build up before the pore water passes from the stability field of kaolinite to that of albite (Bj0rlykke & Aagaard 1992) and albite may therefore be selectively dissolved. The remaining K-feldspar and kaolinite may react at higher temperatures (120-140 ° C) during burial diagenesis to form illite (Bjorlykke 1983; Bj0rlykke et al. 1992) (Fig. 6b). If all K-feldspar in a sandstone is leached during meteoric water flushing, kaolinite may remain stable to rather high temperatures because there is no K-feldspar to supply the necessary potassium to form illite. The amount of kaolinite present as a precursor mineral may be the limiting factor controlling illite growth. This again may be related to facies and climate. Sandstones representing distal shelf or turbidites facies may develop little diagenetic kaolinite due to restricted meteoric water flushing (Bjorlykke & Aagaard 1992). The upper Jurassic Fulmar Sandstone (North Sea Basin) represents a turbiditic and distal shelf facies and contains little kaolinite (Saigal et al. 1992). We would therefore predict that very limited amounts of diagenetic illite would form if this sandstone were buried more deeply than its present depth of 3.5 km. The silica released during feldspar dissolution has been considered a source of silica for quartz cementation, and Blatt (1979) calculated the flux of pore water required to introduce silica for quartz cementation through meteoric water flow. He assumed that the pore water contained 32.5 ppm disolved silica, which represents a high degree of super-saturating relative to quartz (6 ppm at 25 ° C). However most of the meteoric water flow will not reach depths where the temperatures are high enough for quartz to precipitate (70-80 ° C) (Bj0rlykke & Egeberg 1992). Silica released through feldspar dissolution must therefore be removed by continued pore water flow. If silica is allowed to build-up in the pore water, smectite rather than kaolinite will become the stable clay mineral (Garrels & Christ 1965; Aagaard & Helgeson 1983). Since dissolution of feldspar and precipitation of kaolinite requires a flux of pore water, the degree of feldspar dissolution and the volume of kaolinite precipitated can to a certain extent be taken as a measure of the integrated pore water flow. The rate of feldspar dissolution also depends on the degree of initial undersaturation

of the pore water and the kinetics of the reaction rate. The initial acidity determines the number of protons available for the leaching process as calculated by Bj0rkum et al. (1990). Dissolution of mica and feldspar does not require acid conditions and the hydrolysis may continue at high pH (Bj0rlykke & Aagaard 1992). The process does, however, consume protons and if there is no other reaction taking place the leaching capacity of meteoric water is limited by the initial acidity in the soil horizon as suggested by Bjcrkum et al. (1990). Probably oxidation of organic matter and carbonate buffering reactions may supply additional protons along the pathway of meteoric water flow, but this is difficult to quantify. The build up of potassium and sodium in meteoric water would be a measure of the degree of dissolution along the flow pathway. We may then assume that one mol of potassium or sodium may have dissolved one mol of feldspar or mica. The volume (V) dissolved by pore water flow of a mineral X can be expressed as follows: V = ff/mxp

where f is the increased concentration of potassium or sodium and p is the density of the mineral, rnx is the atomic weight fraction of potassium or sodium in the mineral and f is the integrated flux over time (cm 3 cm-2). In the case where potassium is dissolved from K-feldspar, mx = 0.14(39/278) and p = 2.9 g c m -3. Ifq = 10 -5 (10 ppm) each volume of water (f = 1 cm 3 cm -2) will dissolve 2.4 × 10 -5 volumes of feldspar. This means that if the potassium concentration in the pore water builds up to a concentration of 10 ppm, it takes 4 x 104 volumes of pore water to dissolve one volume of feldspar. Similar values have been published based on different assumptions (Bj0rkum et al. 1990). Dissolution of 1% feldspar along a 1 km aquifer correspond to the dissolution of 103cm3 cm -2 and this would require an integrated flux of 4 x 107cm3 cm -2. However we do not know how rapid the leaching of mica and the build up of potassium in the pore water occurs. This corresponds to 4 million years of flow at the rate of 10 cm per year, and more intensive leaching requires correspondingly more flow. The pore water flux would in most cases decrease away from the recharge area. The rate of meteoric water flow is limited by the infiltration of rainfall into the ground water and the degree of focusing or dispersal of the flow. These crude calculations show that it requires large volumes of water and long time to leach significant volumes of feldspar and mica. The intensity of leaching is likely to decrease rapidly with depth and the time that

FLUID FLOW AND DIAGENESIS fluvial and shallow water sediments are exposed to meteoric water flushing will be inversely proportional to the sedimentation rate. In the distal shelf and turbidite facies the meteoric water flux would be expected to be lower and this is in accordance with observations from the North Sea Basin (Bj0rlykke & Aagaard 1992). The flow rates of meteoric water are typically several orders of magnitude higher than that of compaction-driven water (Harrison & Summa 1991). It is therefore difficult to understand how sufficient flow is generated to dissolve large volumes of K-feldspar and precipitate kaolinite at depth. It is also difficult to see how the released potassium is removed. The degree of feldspar dissolution observed in sedimentary basins like the North Sea Basin varies with facies. The fluvial and shallow marine sediments have generally been subjected to severe feldspar dissolution. Sandstones representing distal shelf and turbidite facies, on the other hand, show very little evidence of feldspar dissolution and precipitation of diagenetic kaolinite, probably because the flux of meteoric water was lower in these sediments. If dissolution of feldspar and precipitation of kaolinite has occurred at deeper burial as a result of the generation of acids from source rocks, as advocated by Burley et al. (1985) and Burley (1986) the degree of leaching should not be expected to reflect depositional environment and near surface ground water flow. The sandstones most closely associated with the source rocks (Kimmeridge Clay Formation) show generally very little evidence of leaching (Bj~rlykke & Aagaard 1992). Even if very little acid is generated kaolinite can form without removing the potassium if K-feldspar is present. Organic acids generated in the source rock will tend to be neutralised in the source rock itself. During migration the organic acids will diffuse into the water phase because they are highly water soluble (Barth & Bj0rlykke 1993). Dissolution of minerals like feldspar and mica may occur both shortly after deposition and also after tectonic uplift and sub-aerial exposure (Bjorlykke 1983). In the North Sea Basin, the sandstones of the Brent Group were flushed by meteoric water shortly after deposition in the Brent delta and at least partly during tectonic uplift in Late Jurassic times. Some authors consider meteoric water leaching beneath unconformities to be very important (Shanmugam 1988). Bj0rkum et al. (1990) have questioned the significance of the leaching phase associated with the Cimmerian unconformity of the North Sea Basin. They failed to detect leaching of

135

feldspar, and precipitation of kaolinite related to the Upper Jurassic/Cretaceous unconformity in the Gullfaks and Snorre fields is not very significant (Bjorkum et al. 1990, 1993). They interpreted this to be partly because the rate of erosion may have kept pace with the downwards progradation of the dissolution zone. It is also possible that the islands produced during this uplift had a low sub-aerial relief because they consisted of mostly uncemented Jurassic sand. The head of the ground water table would then be low. Meteoric pore water flowing down-dip will gradually approach equilibrium with the mineral phases at a rate dependent on the kinetic reaction rate. The flow rate and the kinetics of feldspar dissolution will strongly influence the distance along which significant mineral dissolution occurs. Feldspar dissolution is ubiquitous in both the marine and the fluvial part of the Brent Formation but, in most cases, not all of the feldspar is dissolved. This is also evidence of a slow reaction rate, since there is no sharp dissolution front. The total amount of leaching is, however, limited by the pore water flux and the composition of the groundwater. As pointed out by BjCrkum et al. (1990) a soil horizon will release carbon dioxide and lower the pH of the ground water flowing into the subsurface. Present day land surfaces are examples of unconformities and the extent of meteoric water flow and dissolution can be studied at them. Hanor & McManus (1988) observed high concentrations of kaolinite down to about 200m below the surface in Cretaceous rocks of the Mississippi coastal plain. In the Alberta Basin, diagenetic kaolinite in isotopic equilibrium with the present ground water several hundred metres below the surface is an indication of recent authigenic growth due to meteoric water flow (Longstaffe 1984). Land areas characterized by relatively permeable aquifers, significant head and low erosion rates have the highest potential for preserving a thick zone of meteoric water leaching below an unconformity. In Triassic aquifers onshore in Britain recent feldspar dissolution has been observed some hundred meters below the land surface (Bath et al. 1987). In summary it seems that meteoric water dissolution in most cases occurs shortly after deposition and is therefore controlled by climate and depositional environments. However meteoric water leaching may also occur after uplift from land surfaces but the depth of such leaching depends on many different factors including rates of ground water flow and erosion. Unfortunately little data is available on the extent and depth of meteoric water dissolution in modern environments.

136

K. BJORLYKKE

Diagenetic effects of fluid flow in carbonate rocks The origin of carbonate cements in limestones has been addressed by many authors (see Bathurst 1976). Is the carbonate cement in limestones derived from local dissolution (also pressure solution) of carbonate minerals or is it derived from a source outside the limestone itself? This discussion is analogous to the discussion about the origin of the quartz cement in sandstones. If calcite cements in limestones are introduced by advective flow over longer distances (>10 m), very high pore water fluxes are required. Bathurst (1976) estimated that each volume of pore water can be expected to precipitate only 10-100 ppm, so that 104-105 volumes of water are required in order to precipitate one volume of cement. However, this depends very much on the pH and the temperature gradient. To supply 50% calcite cement to a 100 m thick limestone, a pore water flux of at least 5 x 107-108 cm 3 cm -2 is required. This is far more than can be obtained by normal compaction-driven flow in sedimentary basins (BjOrlykke 1993). In the case of upwards-focused compaction-driven flow the pore water would be cooled during flow and cause dissolution of calcite rather than precipitation. The fact that 313C values in limestones often show variations that reflect changes in the primary contemporaneous sea water is evidence of a rather closed system (Scholle & Halley 1985). If the calcite cement was introduced from the outside over longer distances, the delicate ~13C pattern from changes in the ocean water composition would have been destroyed. Meteoric water flow is important because it has a composition different from sea water and because of the large pore water fluxes that can be driven by the ground water. In particular, the lower sulphate content in meteoric water is favourable for dolomite formation (Kastner 1984). Ground water is often charged with carbon dioxide from decaying organic matter and therefore acid and carbonates are usually extensively leached in the upper part of the ground water lens. This may develop into a karstic dissolution pattern. However, the pore water will relatively rapidly become saturated with respect to calcite. Aragonite, high Mgcalcite, and other biominerals in fossils will dissolve, since the pore water is under-saturated with respect to these minerals, and low Mgcalcite will precipitate. The rate of meteoric water flow into sedimentary basins is controlled to a large extent by sea level changes, and these influence near-surface carbonate diagenesis

rather markedly (Matthews & Frohlich 1987). As soon as saturation with respect to calcite is reached, meteoric water has a limited net capacity for transporting carbonate in solution, although the solubility of calcite may vary somewhat with the local conditions. Most of the dissolved aragonite and high Mg-calcite must therefore precipitate relatively locally. Because of the small sample size, it is difficult to reach firm conclusions about mass balance based on textural evidence from thin-sections. As the pH and the temperature of the pore water increase along the meteoric flow pathway due to dissolution of feldspar and mica, calcite may precipitate. However, where extensive dissolution occurs in the proximal end of an aquifer system, the dissolved calcite must precipitate in the more distal and deeper parts of the system where the pH and the temperature of the pore water is higher. Precipitation of calcite (Ca ++ + H2CO3 = CaCO3 + 2H +) releases protons which may contribute to the dissolution of feldspar. The most important diagenetic reaction which involves mass transfer on a large scale is dolomitization of calcite. Dolomitization of calcite or aragonite requires that calcite dissolves and dolomite precipitates as the most stable phase: 2CACO3 + Mg ++ = Ca Mg(CO3)2 + Ca ++ Even if dolomite is the most stable phase in a certain type of pore water, this reaction cannot proceed without a continuous supply of Mg ++ and removal of Ca ++. The Dorag model for dolomitization (Badiozamani 1973) assumes mixing of meteoric water and marine connate water, producing a mixture in which dolomite is stable; however, it does not explain the largescale supply of Mg ++ or the removal of Ca ++ which would be required. The concentrations of magnesium in subsurface pore waters are normally rather low (Egeberg & Aagaard 1989), and very high fluxes of pore water are therefore needed to supply the necessary magnesium for dolomitization of large volumes of limestone. Sea water may then provide a source of Mg ++. At low sulphate concentrations, dolomite will precipitate more easily (Kastner 1984). As sulphate is removed from sea water by sulphate reduction, the pore water close to the sedimentwater surface may be in the stability field for dolomite. Atolls and reefs form permeable limestones where sea water may circulate due to tidal forces and thus they form ideal settings for early diagenetic dolomitization (Saller 1984). The fact that dolomitization is often restricted to platform margin facies is also evidence of very early dolomitization influenced by meteoric

FLUID FLOW AND DIAGENESIS water flow (Xun & Fairchild 1987). Sea water may then be both a source of Mg ÷+ and a sink for Ca ++" Dolomitization at greater burial depth is very difficult to understand in terms of magnesium supply. Water released by compaction of shale sequences will normally have a relatively low Mg content, because silicate phases like chlorite will compete for the available Mg ++. Pore water in Cretaceous carbonates in Southern Texas is under-saturated with respect to dolomite at depths greater than 1 km (Land & Prezbindowsky 1981). When there is evidence from fluid inclusions or from stable isotopes that dolomite crystals have precipitated at high temperatures, it does not necessarily follow that dolomitization occurred at these temperatures and burial depths. Recrystallization of fine-grained early diagenetic dolomite without a net supply of magnesium could produce the same geochemical and textural evidence. A useful treatment of the chemical and textural aspects of carbonate dissolution is presented by Hutcheon et al. (1992).

Mixing of pore water Mixing of pore water of different compositions is often inferred as a mechanism for precipitation of minerals. An example of this is the Dorag Model for dolomitization (Badiozamani 1973). In principle, two solutions which are undersaturated with respect to a mineral may become supersaturated with respect to the same mineral after mixing. The best potential for mixing of pore water, however, is at relatively shallow depths where meteoric water and sea water may produce a mixing zone (Taberner & Santisteban 1987). Even in this case, the mixing is often not very effective because the fluids have different densities, and meteoric water may displace more saline pore water in sandstone or limestone aquifers while relatively steep salinity gradients are maintained. In the deeper subsurface, the pore water flow is very low and nearly always laminar. The flow is controlled by potential gradients in the subsurface, which means that fluids are moving in the same direction and the rate of mixing is usually low. Meteoric water may displace underlying more saline water with a sharp boundary between the two types of pore water. Even in sandstones the pores are relatively small (in most cases <0.5 mm) and the flow rate must be several metres per second for the critical Reynold number to be exceeded so that the flow is turbulent. Flow on fractures triggered by hydrothermal or seismic activity

137

may be turbulent, particularly if the fractures are relatively wide (>1 cm).

Conclusions Flow of pore water in sedimentary basins transports heat and solids in solution and therefore may potentially be a significant factor in diagenetic reactions. Mineral solubilities are a function of temperature, and heating or cooling of pore water causes dissolution or precipitation of minerals. However, because of the low solubility/temperature gradients for the common carbonate and silicate minerals, very large pore water fluxes are required to dissolve, transport and precipitate significant quantities of minerals in sedimentary rocks. Higher fluxes may be obtained through focused pore water flow, but this can only explain local cementation or dissolution. Fractures or permeable sandstones, which would receive the highest fluxes, would be cemented up before the pore water could supply significant cement to the less permeable sediments. High pore water flow rates are also required to obtain significant heat transport by advection, and in most cases geothermal gradients are controlled by conduction. The most significant mass transport on a regional scale can be obtained during meteoric water flow, where the pore water fluxes may be several orders of magnitudes higher than for compaction-driven flow on a large scale. The depth of significant meteoric water dissolution is limited both by the flow rate, the kinetic reaction rate and the leaching capacity of water. Non-isochemical reactions like the dissolution of feldspar and the precipitation of kaolinite are direct evidence of pore water flow in sedimentary basins. The ubiquitous occurrence of dissolved feldspar and authigenic kaolinite in sandstones, which represent fluvial and deltaic facies in the North Sea Basin, is evidence of facies-controlled meteoric water leaching. Distal shelf sandstones and turbidites, on the other hand, show little evidence of feldspar dissolution. Compaction-driven pore water flow is generally very low, lower than the sedimentation rate, and its capacity to transport solids in solution is very limited except on a local scale. Extreme pore water fluxes are required in order to introduce carbonate cement or quartz cement in limestones and sandstones from an outside source, and compaction-driven upwardsdirected flow would tend to cause precipitation of cement rather than dissolution of carbonate. The fact that cementation of limestones as well as sandstones continues until very low porosity

138

K. BJ~RLYKKE

values are reached, with increasing burial and after the permeability has b e c o m e very low, also suggests a p r e d o m i n a n t l y internal source of the cements. The highest pore water fluxes in s e d i m e n t a r y basins should be e x p e c t e d to occur along o p e n fractures or faults e x t e n d i n g all the way to the surface. C o n c e n t r a t i o n s of s e d i m e n t a r y ores may form w h e n hot fluids flow at high velocity towards the surface and are rapidly cooled by mixing with sea water or cold groundwater. This research has been supported by VISTA, a research co-operation between the Norwegian Academy of Science and Letters and Det Norske OIjeselskap (Statoil), and by the Norwegian Research Council (NAVF). Comments and suggestions by Per Aagaard and the reviewers P.A. Bj0rkum and C. Zwach have been useful and are greatly appreciated.

References

AAGAARD, P. & HELGESON, H. C. 1983. Activity/ composition relations among silicates and aqueous solutions: II. Chemical and thermodynamic consequences of ideal mixing of atoms of homologicai sites in montmorillonites, illites, and mixed-layer clays. Clays and Clay Minerals, 31, 207-217. --JAHREN, J.S. & EGEBERG, P.K. 1992. North Sea clastic diagenesis and Formation water constraints. In: KnARAKA, K. & MAEST, A.S. (eds) Water rock interaction. Balkema, Rotterdam, 1147-1152.

BADIOZAMANI, K. 1973. The Dorag dolomitization model - Application to the Middle Ordovician of Wisconsin. Journal of Sedimentary Petrology 43, 965-984. BATH, A.H., MILDOWSKI, I.E. & SPIRO, A.E. 1987. Diagenesis of carbonate cements in PermoTriassic sandstones in the Wessec and East Yorkshire-Lincolnshire Basins. U.K. A stable isotope study. In: GOLF, J.C. & WILLIAMS,B.P.J. (eds) Fluid Flow in Sedimentary Basins and Aquifers. Geological Society, London, Special Publications, 34, 127-140. BARTH,T., & BJORLYKKE,K. 1993. Organic acids from source rock maturation: generation potentials, transport mechanisms and relevance for minerals diagenesis. Applied Geochemistry, 8,325-337. BATHURST, R.G.C. 1976. Carbonate sediments and their diagenesis. Developments in Sedimentology, 12, Elsevier, Amsterdam. BETHKE, C.M. 1985. A numerical model of compaction-driven groundwater flow and heat transfer and its application to the paleohydrology of intercratonic sedimentary basins. Journal of Geophysical Research, 90,6817-6828. -1989. Modelling subsurface flow in sedimentary basins, Geologische Rundschau, 78, 129-154. --, HARRISON, W.J., UPSON, C. & ALTANER, S.P.

1988. Supercomputer analysis of sedimentary basins. Nature, 239,261-267. BJORKUM, P.A., KNARUD, R. & BERCAN, M. 1993. How important is the late Cimmerian unconformity in controlling formation of kaolinite in sandstones of the North Sea? Examples from the Snorre Field. In: HORBURY,A. & ROBINSON, A. (eds) Diagenesis and basin Development. American Association of Petroleum Geologists. Studies in Geology, 36,261-269. MJOS, R., WALDERHAUG,O. & HURST, A. 1990. The role of the Cimmerian unconformity for the distribution of kaolinite in the Gullfaks Field, northern North Sea. Sedimentology, 37,396--406. BJORLYKKE, K. 1983. Diagenetic reactions in sandstones. In: PARKER, A. & SELLWOOD,B.W. (eds) Sediment Diagenesis. NATO Advanced Study Institute, Reidel Publishing, Reading, UK, 169213. - 1993. Fluid flow in sedimentary basins. Sedimentary Geology, 86, 137-158. - - & AAGAARD,P. 1992. Clay Minerals in North Sea Sandstones. In: HOUSEKNECHT,D.W. & PITMAN, E.D. (eds) Origin, Diagenesis, and Petrophysics of Clay Minerals in Sandstones, SEPM Special Publications, 47, 65-80. -& EGEBERG, P.K. 1993. Quartz cementation in sedimentary basins. American Association of Petroleum Geologists Bulletin, 77, 1538-1548. - Mo, A. & PALM, E. 1988. Modelling of thermal convection in sedimentary basins and its relevance to diagenetic reactions. Marine and Petroleum Geology, 5,338-351. - - , NEDKVITNE,T., RAMM, M. & SAIGAL, G. 1992. Diagenetic processes in the Brent Group (Middle Jurassic) reservoirs of the North Sea - an overview. In: MORTON, A.C., HASZELDINE,R.S., GILES, M.R. & BROWN, S. (eds) Geology of the Brent Group. Geological Society, London, Special Publications, 61,263-287. BLATT, H. 1979. Diagenetic processes in sandstones. In" SCHOLLE, P.A. & SCHLUGER, P.R. (eds) Aspects of Diagenesis. SEPM, Special Publications, 26,141-157. BLOCH, S., McGOWEN, J.H., DUNCAN, J.R. & BRIZZOLARA,D.W. 1990. Porosity Prediction, Prior to drilling, in sandstones of the Kekiktuk Formation (Mississippian), North Slope Alaska. American Association of Petroleum Geologists Bulletin, 74, 1371-1385. B URLEY,S.D. 1986. The development and destruction of porosity within Upper Jurassic reservoir sandstones of the Piper and Tartan Fields, outer Morey Firth, North Sea. Clay Minerals, 21, 649-694. - - , KANTOROWICZ,J.D. & WAUGH, B. 1985. Clastic diagenesis. In: BRENCHLEY, P.J. & WILLIAMS, B.P.J. (eds) Sedimentology: recent developments and applied aspects. Geological Society, London, Special Publications, 18, 189-226. CASSAN, J.P., GARCIA PALACIOS, C.B., DEE, M., FRITZ, B. & TARDY, Y. 1981. Diagenesis of sandstone reservoirs as shown by petrographical and geochemical analysis of oil bearing

FLUID FLOW AND DIAGENESIS

139

Gulf Coastal plain. Gulf Coast Association of formations in the Gabon Basin. Bulletin Centres Geological Societies, Transactions, 38,495-502. Recherches, Elf-Aquitaine, 5, 113-135. CHAPMAN, R.E. 1987. Fluid flow in sedimentary HAVES,M.J. & BOLES,J.R. 1993. Meteoric recharge in the San Joaquin basin California provided by basins: a geologist's perspective In: GOFF, J.C. & isotope and trace element chemistry of calcite. WILLIAMS,B.P.J. (eds) Fluid Flow in Sedimentary Marine and Petroleum Geology, 10,135-144. Basins and Aquifers. Geological Society, London, Special Publications, 34, 3-18. HARRISON,W.J. & SUMMA,L.L. 1991. Paleohydrology DAVIS, S.H., ROSENBLAT,S., WOOD, J.R. & HEWETT, of the Gulf of Mexico Basin. American Journal of T.A. 1985. Convective fluid flow and diagenetic Science, 291,109-176. patterns in domed sheets. American Journal of HASZELDINE, R.S., SAMSON, I.M. & CORNFORD, C. 1984. Quartz diagenesis and convective fluid Science, 285,207-223. movements, Beatrice Oilfield, UK, Northern DELOULE, E. & TURCO~E, D.L. 1989. The flow of hot North Sea. Clay Minerals, 19,391-402. brines in cracks and the formation of ore deposits. Economic Geology, 84,2217-2225. HERMANRUD, C., EGGEN, S. & LARSEN, R.M. 1991. Investigations of the thermal regime of the Horda DEMMIN6, D., NUNN, J.A. & EVANS, D.G. 1990. Platform by basin modelling: implication for the Thermal effects of compaction-driven ground hydrocarbon potential of the Stord basin, northwater flow from overthrust belts. Journal of ern North Sea. In: SPENCER,A.M. (ed.) GenerGeophysical Research, 95, 6669-6683. ation, accumulation and production of Europe's EGEBERG, P.K. & AAGAARD, P. 1989. Origin and evolution of formation waters from oil fields on hydrocarbon. Special Publications of the European Association of Petroleum Geoscientists, 1, the Norwegian shelf. Applied Geochemistry, 4, 131-142. 65-73. Oxford University Press. EGEBERG, P.K. 1992. Brines in sedimentary environ- HUTCHEON, I., SHEVALIER,M. & ABERCROMBIE,H.J. 1992. pH buffering by metastable mineral-fluid ments. Encyclopedia of Earth Systems Science, 1, 395-407. equilibria and evolution of carbon dioxide fugacity during burial diagenesis. Geochimica et GILES, M.R. 1987. Mass transfer and problems of Cosmochimica Acta. 57, 1017-1027. secondary porosity creation in deeply buried KASTNER, M. 1984. Control of Dolomite Formation. hydrocarbon reservoirs. Marine and Petroleum Geology, 4,188-201. Nature, 311,410--411. LAND, L.S. & PREZmDOWSK1,D.R. 1981. The origin - - , STEVENSON,S., MARTIN,S.V. & CANNON,S.J.C. and evolution of saline formation water. Lower 1992. The reservoir properties and diagenesis of Cretaceous carbonates. South central Texas. the Brent Group: a regional perspective. In: MORTON, A.C., HASZELDINE,R.S., GILES, M.R. U.S.A. Journal of Hydrology, 54, 51-57. LONGSTAFFE,F.J. 1984. The role of meteoric water in & BROWN,S. (eds) Geology of the Brent Group. diagenesis of shallow sandstones: Stable isotope Geological Society, London, Special Publications, 61,289-327. studies of the Milk River Aquifer and Gas Pool, southeastern Alberta. In: MCDONALD, D.A. & GARRELS, R.M. & CHRIST, C.L. 1965. Solutions, SURDAM,R.C. (eds) Clastic Diagenesis. American Minerals and Equilibria. Harper & Row, New York. 450. Association of Petroleum Geologists, Memoirs, GLUYAS, J. & COLEMAN,M. 1992. Material flux and 37, 81-98. MATTHEWS, R.K. & FROHLICH, C. 1987. Forward porosity changes during sediment diagenesis. modelling of bank-margin carbonate diagenesis. Nature, 356, 52-54. ~, ROBINSON, A.G., EMERY, D., GRANT, S.M. & Geology, 15,673-676. OXTOBV, N.H. 1993. The link between pet- MUDEORD, B.S., GRADSTEIN,F.M., KATSUBE,T.J. & BEST, M.E. 1991. Modelling 1D compaction roleum emplacement and sandstone cementation. In: PARKER,J.R. (ed.)Petroleum geology driven flow in sedimentary basins: a comparison of the Scotian Shelf, North Sea and Gulf Coast. of Northwest Europe. Proceedings of the 4th In: ENGLAND, W.A. & FLEET, A.J. (eds) PetConference. Geological Society of London, 1395-1403. roleum Migration. Geological Society, London, GRANT, S.M. & OXTOBY,N.H. 1992. The timing of Special Publications, 59, 65-85. quartz cementation in Mesozoic sandstones from PALM, E. 1990. Rayleigh convection, mass transport, and change in porosity in layers of sandstone. Haltenbanken, offshore mid-Norway: fluid incluJournal of Geophysical Research, 95, 8675-8679. sion evidence. Journal of the Geological Society, ROBINSON, A. & GLUVAS,J. 1992. Duration of quartz London 149,479-482. GRAN, K., BJORLYKKE,K. & AAGAARD,P. 1992. Fluid cementation in sandstones, North Sea and Haltenbanken Basins. Marine and Petroleum Gesalinity and dynamics in the North Sea and ology, 9,324-327. Haltenbanken basins derived from well log data. In: HURST,A., GRIFFITHS,C.M. & WORTHINGTON, SAIGAL,G.C., BJORLYKKE,K. & LARTER,S. 1992. The P.F. (eds) Geological Application of Wireline effects of oil emplacement on diagenetic processes. Examples from the Fulmar Reservoir Logs II. Geological Society, London, Special Sandstones, Central North Sea. American AssociPublications, 66,327-338. ation of Petroleum Geologists Bulletin, 76, 1024HANOR, J.S. &MCMANUS, K.M. 1988. Sediment alteration and clay mineral diagenesis in a 1033. regional ground water flow system, Mississippi SALLER, A.H. 1984. Petrologic and geochemical

140

K. BJORLYKKE

constraints on the origin and subsurface dolomite. Enewetak atoll: An example of dolomitization by normal sea water. Geology, 12,217-220. SCHOLLE, P.A. & HALLEY, R.B. 1985. Burial Diagenesis: out of sight, out of mind! SEPM Special Publication, 36,309-334. SHANMUGAM, G. 1988. Origin, Recognition, and Importance of Erosional Unconformities in Sedimentary Basins. In: KLEINSPEHN,K.L. & PAOLA, C. (eds) New Perspectives in Basin Analysis, Springer Verlag, 83-108. TABERNER, C.T. & SANaqSTEBAN, C. 1987. Mixedwater dolomitization in transgressive beach-ridge system, Eocene Catalanan Basin, NE Spain. In: MARSHALL, J.D. (ed.) Diagenesis of Sedimentary Sequences. Geological Society, London, Special Publications, 36, 123-139. WALDERHAtJG, O. 1990. A fluid inclusion study of quartz cemented sandstones from offshore MidNorway - possible evidence for continued quartz cementation during oil emplacement. Journal of Sedimentary Petrology, 60,203-210.

WOOD, J.R. 1986. Thermal mass transfer in systems containing quartz and calcite. In: GAVVtER, D.L. (ed.) Roles of Organic matter in Sediment Diagenesis. SEPM, Special Publications, 38, 169180. --& HEWE'rr, T.A. 1982. Fluid convection and mass transfer in porous sandstones - a theoretical model. Geochimica et Cosmochimica Acta, 46, 1707-1713. & - 1984. Reservoir diagenesis and convective fluid flow. In: MCDONALD, D.A. & SURDAM, R.C. (eds) Clastic diagenesis. American Association of Petroleum Geologists, Memoirs, 37, 99-111. XUN, Z. &FAIRCHtLD,I. 1987. Mixing zone dolomitiziation of Devonian carbonates, Guangxi, South China. In: MARSHALL, J.D. (ed.) Diagenesis of Sedimentary Sequences. Geological Society, London, Special Publications, 36,157-170.

Controls on two-phase fluid flow in heterogeneous sandstones P.S. R I N G R O S E

& P.W.M. CORBETT

Department o f Petroleum Engineering, Heriot-Watt University, Edinburgh EH14 4AS, UK Abstract: The effective flow behaviour of two immiscible fluids in a permeable medium is a function of fluid/fluid as well as fluid/matrix interactions. These are described by relative permeability functions, which are strongly influenced by the spatial distribution of permeability in sedimentary media, given specified fluid and pore surface properties. For typical subsurface flow rates and patterns of rock heterogeneity in hydrocarbon reservoirs, capillary forces can result in significant amounts of trapping and bypassing of the non-wetting phase. Different types of sedimentary media (e.g. laminated, crossbedded and pervasively faulted strata) result in different, quantifiable, degrees of trapping. Examples of typical water-wet oil/water systems illustrate how the amount of trapped oil can vary between around 40% and 65% depending on the type of heterogeneity. Furthermore, the layered nature of most sedimentary strata means that the relative permeability of two immiscible phases is highly anisotropic. These findings have important implications for reservoir engineering and studies of oil migration.

The flow of liquid hydrocarbons, in both natural oil migration and engineered oil production, generally involves multiphase flow. The main controls on multiphase flow behaviour are significantly different from, and more complex than, the controls on single-phase flow in a permeable medium. We focus here on oil-water (immiscible) flow. We review the main concepts, some experimental observations, and then summarize some of our own work on numerical modelling in heterogeneous rock formations. Three-phase flow (e.g., oil-gas-water) or the flow of two, more-or-less miscible phases (e.g. gas and oil) are beyond the scope of this paper (see Lake 1989 for a more general discussion). The main concepts in oil migration theory are also reviewed elsewhere (e.g. England et al. 1987; England & Fleet 1991). However, this paper does address important questions identified in earlier studies of oil migration, namely:

(a) (b)

'what control does sediment fabric have on two-phase flow?'(Tissot 1987) 'the key parameters (petroleum properties, capillary seals and relative permeability) play a significant role at the regional scale, but are difficult to evaluate.' (Burrus etal. 1991).

Oil-water systems are of major concern to the petroleum industry as the majority of oil fields are produced by some form of water-flooding,

and as oil migration into reservoirs occurs dominantly within an aqueous phase (Schowalter 1979). Furthermore, oil-water systems are analogous to other immiscible flow systems, such as air and water in the under-saturated zone of groundwater aquifers (provided that adjustments are made for the actual interfacial tensions, viscosities and densities of the fluids). Thus, understanding oil-water flow has widespread implications. Models of fluid flow in geological systems require estimates of the permeability of the rock medium concerned. In engineered systems (groundwater or petroleum reservoirs), effective macroscopic values of permeability are estimated from borehole data; either by measurements of core material, log-tool response or pressure transient tests of the surrounding formation (e.g. Haldorsen 1986; Worthington 1991). In natural aquifers, effective permeability may be inferred from boundary conditions, such as records of flow rates at source/sink points (precipitation, spring discharge etc.) and from the geometries of permeable formations (e.g. Deming et al. 1992). These externally-estimated media properties often carry large errors due to the inherent uncertainties about the medium. Parameter estimation becomes even more difficult when assessing two-phase flow processes, because of the additional physical processes involved (e.g., rock-fluid interactions and phase changes), and the additional parameters required (relative

From PARNELL,J. (ed.), 1994, Geofluids: Origin, Migrationand Evolution of Fluids in Sedimentary Basins, Geological Society Special Publication No. 78,141-150.

141

142

P.S. RINGROSE & P.W.M. CORBETT

permeability, capillary pressure, and fluid viscosities). In this paper we take a bottom-up approach to this problem, by considering the following questions: (a) (b)

permeability and wettability. These physical concepts are represented in the two Darcy flow equations required to describe oil-water flow:

what do we know about the principles of two-phase flow in typical rock media? what can we then infer about macroscopic two-phase flow behaviour?

This approach is an important one because of the complexity of multiphase flow, which is not as intuitive as simple, single-phase (Darcy) flow. In multiphase flow, a number of mostly empirical relations are needed to represent the interactions of the fluids with one another and with the porous medium. One fluid may be mobile when another is not; the various fluid phases may adhere to, or be repelled by, mineral surfaces in different ways; and the imposed pressure gradients and gravity forces may act to separate or mix the phases. An understanding of the main concepts of multiphase flow is therefore a prerequisite to any informative discussion of the large scale implications. The multiphase flow effects are sensitive to small scale pore and pore throat characteristics. The small scale effects have to be accounted for in large scale numerical models (reservoir models and basin models) if their relevance is to be quantified or understood. In water-flooding of reservoirs and secondary oil migration, the interaction of rock and fluids at the appropriate rate will determine the flow process. Our work shows that, for reservoir displacements in certain field cases, the capillary effects resulting from lamination can have a significant impact on the field performance. A modelling procedure for capturing the effective flow parameters for representative geological elements at various scales is discussed. This procedure allows the small-scale phenomena to be incorporated in larger scale models, and their effects quantified and understood. Similar techniques can be used for the study of secondary oil migration where the large scale process may also be dominated by small scale phenomena. In this paper, we investigate the importance of capillary pressure in laminated sediments (i.e. a small scale phenomenon). The scale up technique, however, could be extended to examine the effects of other phenomena that might be considered important (e.g. film drainage).

The main concepts Essential to understanding two-phase flow are the concepts of capillary pressure, relative

-k

Uo -

-k

Uw where

IC~°Veo

(1)

k~wV/'w

(21

IXo

~w

u

= velocity (suffices 'o' and 'w' refer to the oil and water phases); k = absolute (or intrinsic) permeability; kr = relative permeability; ix = viscosity; V P = pressure gradient.

The absolute permeability, k, is an intrinsic property of the porous medium, regardless of the fluid properties. In general k is a macroscopic tensor (Pickup et al. 1993), but for the purposes of this discussion it suffices to treat it as a scalar quantity by neglecting the off-diagonal, or crossflow terms. However, the relative permeability, kr, is affected by the fluid properties (viscosity, interfacial tensions, wetting characteristics, etc.) as well as the properties of the porous medium. At any point in the medium the pressure difference between the two phases can be related to the capillary pressure: Pc = P o - Pw

(3)

where Po and Pw are the local oil and water phase pressures. In a specific volume of a porous medium, Pc is the average pressure difference maintained across a collection of fluid-fluid interfaces. For a capillary tube, Pc can be simply related to the interfacial tension, 0., the interface contact angle, a~, and the radius of curvature of the interface, R (in a form of Laplace's equation): Pc -

20- cosa~ R

(5)

In a realistic porous medium the irregular pore geometry and surface properties make this relationship difficult to apply. However, an approximation, using a mean radius of curvature and an inferred interfacial tension, can be useful for understanding the general properties of capillary trapping. For example, approximate estimates of the maximum thickness of a column of oil, Zo, in a reservoir can be estimated using (Berg 1975):

CONTROLS ON TWO-PHASE FLOW

,)

__

Zowhere

Rres g(pw- po)

.,0

ap

"5

(5)

Reap, R~s = pore throat radius in the cap rock and reservoir; p~, po = densities of oil andwater

In some simple, cap-rock/reservoir systems this relationship appears to be validated by independent estimates of pore throat size in the cap and reservoir, the observed oil column height and plausible values of interracial tension inferred from hydrocarbon composition (e.g. Jennings 1987). However, as capillary pressure is strongly influenced by the smaller pores it is not usually clear what measurable pore throat, or set of connected pore throats, is appropriate. A more appropriate function can be drawn from the dimensionless capillary pressure function, introduced by Leverett (1941) and known as the 'j-function':

j(S,w)

=

~

cos 0

143

(6)

In this function, the R term has been replaced by V'-~(b using the hydraulic radius concept and an assumption about the pore size distribution (Lake 1989). The J-function is much more appropriate for relating the measurable parameters Pc, k and +, as borne out by experiment (Amyx et al. 1961) and numerical analysis (Ringrose et al. 1993a). The second important concept for two-phase flow is relative permeability. The relative permeability is defined as the ratio of the effective permeability of one phase to the absolute permeability. Relative permeabilities vary as a function of phase saturation. The typical form of relative permeability curves for a water/oil system is shown in Fig. 1. Note that both phases are only mobile between the two endpoint saturations (irreducible water and residual oil) and that the sum of the two relative permeabilities is less than one. The actual positions of the endpoints, and the shape of the curve between, vary considerably from rock to rock, as discussed below. A third important concept is that of wettability. This is defined as the tendency of one fluid to spread on or adhere to a solid surface in the presence of other immiscible fluids (Craig 1971). In a rock-oil-brine system, it is a measure of the preference that the rock has for either oil or water (Anderson 1986). Under water-wet conditions (which are assumed in the examples below), water contacts the majority of the rock

"B

=e-. '~,

Total permeability

~: o~" =

e-

2

.~ ~

o

E

o

e'~

"8

,~ 0 Wetting fluid saturation Fig. 1. The general form of relative permeability functions for water-oil systems.

surface and tends to occupy the smaller pores. Oil-wet and mixed-wet systems also occur and the type of wetting has a significant influence on both the capillary pressure and the relative permeability (McDougall & Sorbie 1992). In this discussion we assume water wet conditions, but the influence of wettability should be borne in mind when applying the conclusions to different reservoirs.

Some experimental observations Several laboratory studies of secondary oil migration have been conducted in which gravitydriven oil migrates through glass-bead, sand or dolomite packs in vertical or steeply inclined glass columns (Dembicki & Anderson 1989; Catalan et al. 1992). The main findings of these studies can be summarized as follows: (a)

(b) (c)

the basic parameters controlling the minimum oil column height (Equation 5) before migration occurs were confirmed, but required allowance for the effects of phase saturation (Equation 6); oil migration was observed to occur in restricted pathways or conduits; the observed rates of oil migration in experiments (around 10-50m s -~) were much higher than parametric estimates of basin-scale oil migration (around 10 -9 m s -1 , England etal. 1987).

Drawing inferences on geological systems from these experimental results may be inappropriate for two important reasons. Firstly, bead or sand packs have high permeability (usually 1-20

144

P.S. RINGROSE & P.W.M. CORBETT

Darcies) and low heterogeneity (approximately uniform media) which is not characteristic of most sandstones. Secondly, the laboratory experiments carry many artefacts, such as container geometry, grain packing characteristics, edge effects, and inappropriate scales. More recently, core flood experiments (Selle et al. 1993) have investigated secondary migration through real rocks at a more realistic (for the North Sea) dip angle of 7.5 ° . These experiments produced geologically plausible, but still high, migration rates (10-2m s -I) and showed the influence of capillary forces. In high permeability rock, with low threshold capillary pressures (e.g., Bentheimer Sandstone, 2260 mD), they found secondary oil migration to be dominated by buoyancy forces. However, in lower permeability rock (e.g., Berea Sandstone, 66 mD), capillary pressure effects resulted in a significant capillary transition zone at the migration front and distributed migration pathways. The two samples used in this study were of fairly uniform rock and their experiments do not address the issue of rock heterogeneity. They do, nevertheless, show the significant impact of capillary forces. Further experimental work, combined with larger scale, geologically realistic models, are needed to better understand the process of secondary migration through sedimentary rocks.

Numerical models for typical sandstones In this section, we describe effective two-phase flow properties derived using a theoretical framework for the intrinsic permeability structure of sedimentary strata. Probable large-scale effective flow properties for sandstone formations were derived using centimetre-scale probe permeameter data from core and outcropbased models of sedimentary architecture. The models are drawn from reservoir engineering studies, but are then used to consider the implications for basin flow modelling and oil migration studies. Only water-flooding (water displacing oil) is considered, but many of our conclusions could, in principle, be applied to other immiscible fluid systems, such as gaswater. Several decimetre to metre scale models of flow performance in sandstones are illustrated: uniform, layered, and cross-bedded sandstones; a sandstone with small-scale faults, and a model of shallow-marine wavy-bedded interval. Details of the measurements, models and theory are given elsewhere (Hurst & Rosvoll 1989; Corbett et al. 1992; Corbett & Jensen 1993; Ringrose et al. 1993a). Permeability values in the

models vary from 5-1300 mD and are based on calibrated probe permeameter data (air permeabilities). The scale and structure of permeability heterogeneity is based on typical outcrop examples of the sandstone bedforms concerned. Lamina thicknesses are of the order of 2-10 mm and bedforms are 0.1-0.5m high. Porosity variation has not been measured, but has either been assumed constant or related directly to permeability. The effective two-phase flow behaviour of the models has been assessed using the ECLIPSE100 black oil reservoir simulator, which allows the dynamics of flow-heterogeneity interactions to be assessed numerically. This is done by specifying relationships for pore-scale immiscible flow behaviour (based on laboratory measurements, pore-scale network models, and empirical functions). Thus, we assume that the pore-scale flow behaviour (Darcy flow and capillary pressure) is known and related to local permeability, and then use a numerical model to assess the effective larger-scale behaviour for models with defined heterogeneity. The governing equations were given above (Equations 1,2 and 3). The additional empirical relationships used are: Pc

3.0 Se -2/3 Id~ °5

~ o = 0 . 8 5 ( 1 - so) ~

(8)

krw = 0 . 3 ( s o ) 3

(9)

Swc = 0.6 - 0.165 Log k

(10)

Sor = 0.7

(11)

Swc~Se~Sor

(12)

where, Swc = irreducible water saturation, Sot = wetting phase saturation corresponding to the residual oil saturation, Se = effective wetting phase saturation (i.e normalized in the range Swc&r). Note that the permeability dependence of capillary pressure (Equations 6 and 7) and its effect on relative permeability is critical to the proper derivation of capillary-dominated flow properties. Capillary pressures in a heterogeneous rock have the effect of producing saturation variations (correlated to local permeability), which in turn affect the local relative permeabilities. The relative permeability functions (Equations 8 and 9) depict a moderately water wet system. Different assumptions on wettability would alter flow behaviour significantly (McDougall & Sorbie 1992), but similar effects

CONTROLS

130 > O o ~D

.,=

Permeability 120 contrast 110

145

Along-layer flow (capillary-enhanced recovery)

100

...................

,:1 ii

90 0 o

FLOW

100:1 10:1

O

. ,,,,q

ON TWO-PHASE

i:!i

................

2:1

.....

80 70

> 0 o

60

100:1 Cross-layer flow 1000:! :i:i::i:ill ~. (,capillary-hinderedrecovery)

50

•

.001

"

•

•

•

•

•

.01

•

, , I

.

.

.

.

.

.

.I

1

Scaling parameter, x = layer thickness (m) x velocity (m/day) Fig. 2. Oil recovery for laminated systems as a function of scaling parameters in the viscous/capillary scaling group. of rock heterogeneity would still be expected, although manifested in different ways. We have also assumed an empirical scaling for the irreducible water saturation (Equation 10), but have held the residual oil saturation fixed (Equation 11). The later assumption is not strictly valid, but data support for a suitable equation is lacking. Sensitivity analyses have shown that different assumptions for these empirical correlations affect the results numerically but not in terms of the trends observed and general nature of capillary trapping in heterogeneous formations. Using the above relationships, effective relative permeability functions (using the method of Kyte & Berry 1975) have been derived for water-flood in a number of types of sandstone architecture. These functions, which we term 'geopseudos', depict the mobility of oil and water (in a given direction) relative to the average absolute permeability of the rock model. The functions are scale-dependent and incorporate the (unavoidable) effects of numerical dispersion. They should always be derived with reference to the dimensions of large-scale flow model. The functions are also ratedependant. The effects of heterogeneity are reduced at high (viscous-dominated) rates, but enhanced at lower (capillary-dominated) rates. These effects can be understood in terms of a viscous-capillary scaling group (Rapoport 1955):

Viscous/capillary ratio -

Ux 2~cp, o kx(dPffdS)

(13)

where ux

= is the intrinsic, or Darcy, velocity (in the direction x); Ax = system dimension in the x direction; kx = effective permeability in the x direction; IXo = oil viscosity dPddS = capillary pressure gradient with respect to saturation.

The effects of this scaling group are illustrated in Fig. 2, which shows oil recovery for layered models as a function of the scaling parameter, "r, which is the product of rate and layer thickness (the other parameters are held fixed). As the viscous/capillary ratio is reduced the range in recovery increases, depending on the layer orientation and permeability contrast. For this simple layered model, cross-layer flow results in capillary-hindered oil recovery (because oil becomes trapped in isolated high permeability layers), and along layer flow results in capillary enhanced recovery. Both these conditions differ from a uniform rock model, which shows negligible change in recovery as a function of the viscous/capillary scaling group. In natural and engineered systems involving two-phase flow, we expect capillary-dominated behaviour to

146

P.S. R I N G R O S E & P.W.M. C O R B E T T

1.0 Graded bedding 0.8 2 m model 5 cm gridcells 50 to 500 m D layers

o,..q

0.6

O

V.IJJJJJ

~D

n¢

Uniform

Crossbedding

0.4

.,.-i

0.2

0.0 4 0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

Water saturation Fig. 3. Relative oil permeability versus water saturation for horizontal flow in models of 2 m sandstone blocks with graded bedding, cross-bedding, and uniform architecture. predominate. Viscous-dominated behaviour occurs when the medium is uniform, high permeability, or subject to large pressure gradients and flow rates. We have, therefore, evaluated capillarydominated behaviour in a number of typical heterogeneous sandstone models, each with a Darcy velocity of 0.2 m day -1 (2.3 x 10 -5 m s-a), which is typical of waterflooding in oil reservoirs. For brevity, only the oil geopseudos, k~o, are illustrated (as the main effects of heterogeneity are apparent in the non-wetting phase flow behaviour). The pseudos are for x-direction (horizontal) flow, except for the illustration of anisotropy where both horizontal and vertical flows are considered. Our findings are summarized below.

are inter-connected (as in the graded-bed model) then this oil is able to flow efficiently. However, if high permeabilty zones are surrounded or blocked by low permeability zones (as in the cross-bedding model), heterogeneity trapping occurs. The degree of trapping is evident by the saturation-value at which k,o becomes zero (0.55 for the cross-bedding model, 0.7 for the graded-bed model). At water saturations greater than these 'remaining oil endpoints' the medium becomes permeable to water only. After injection of one pore volume (pv) of water, 56% of the original oil becomes trapped in the cross-bedding model compared to only 38% in the graded-bed model. As crossbedding is common in high porosity sandstones, capillary trapping in this architecture is also likely to be common.

Effects of sediment architecture Figure 3 shows kro functions for models of 2 m blocks of sandstone with graded-bed, crossbedding, and uniform architectures. All three models have an arithmetic average permeability of 260mD. The observed differences in oil relative permeability kro, can be understood in terms of capillary/heterogeneity trapping mechanisms. Because of the dependence of capillary pressure on saturation and permeability/ porosity, local redistributions of oil and water occur, with the tendency to trap oil in higher permeability zones. If high permeability zones

Effects of pervasive faulting Figure 4 shows kro functions for parallel lamination (at a smaller scale) compared with the same laminae cut by a pair of small, lowpermeability faults. The faulting causes a large reduction in kro and a significant amount of heterogeneity trapping of oil (65% of original oil in place remains after 1 pv). A plot of saturation distribution in this model (Fig. 5) illustrates the impedance to oil flow that could occur in a pervasively-faulted sandstone horizon. Oil is

CONTROLS ON TWO-PHASE FLOW

1.0

147

Parallel layering

0.8 ,,,,a

0.6 ~D

O

0.4 °,-q

0.2

0.0 0.0

0.1

0.2

0.3 0.4 0.5 Water Saturation

0.6

0.7

0.8

Fig. 4. Relative oil permeability versus water saturation for a 20 cm model of horizontal flow in parallel layering with and without small-scale faulting.

$W O. 1000 0.2500 0.4000 0.5500

to to to to

0.2500 0.4000 0.5500 0.7000

i! ~

•

•

Fig. 5. Water saturation distribution in the fault model (Fig. 4) after injection of one pore volume (pv) of water, for (a) capillary-dominated (0.2 m/day) and (b) viscous dominated (10 m/day) flow rates (oil is represented by lighter tones). retained behind the fractures under capillarydominated conditions (Fig. 5a), but can be displaced by higher pressure gradients (Fig. 5b). Changes in the fault properties (for example, if the fracture permeability was increased) would

alter this behaviour significantly. More work is needed to relate these small-scale effects to larger-scale models of fault architecture, but the potential for tectonic modification of two-phase flow characteristics is clearly evident.

148

P.S. RINGROSE & P.W.M. CORBETT 1.0

°aNn

Wavy bedded ......

0.8 . ,,,,~

e~

I-

, .......

0.6

............ I v3_ _ Lc3

O

. ,,....q

~

0.4

I

5cm gridcells

0.2

0.0

'

0.0

I

0.1

'

,

,

0.2

0.3

0.4

0.5

0.6

0.7

0.8

Water Saturation Fig. 6. Relative oil permeability versus water saturation for horizontal and vertical flow in a model of wavy-bedded facies in the upper Rannoch Formation.

Large-scale anisotropy When assessing large-scale immiscible flow behaviour, the effects of small-scale heterogeneities must be aggregated (Ringrose et al. 1993b). Our approach is to scale up generalised flow models of lamina-scale and bedform-scale architecture, in a number of stages dictated by sequence stratigraphic principles. This procedure was followed to investigate the effects of lamination in a North Sea Middle Jurassic Rannoch Formation reservoir (Corbett et al. 1992). The Rannoch Formation is a shoreface unit dominated by low-angle cross-lamination and wavy bedded intervals (Scott 1992). Figure 6 illustrates one example of scaled-up water-flood behaviour for the wavy-bedded and rippled facies that occurs within the nearshore trough of the upper Rannoch Formation. Permeability contrasts between adjacent laminae of 301300mD are typical within this facies. The nature of the lamination produces significant two-phase flow anisotropy, as well as oil trapping for vertical flow. The effective properties shown are for a 5 m grid block. These grid blocks are combined with those representing the other laminated styles and contrasts in a regular geologically-sensible arrangement at larger scales (Corbett & Jensen 1993). In this way, the effects of small scale phenomena can be carried

through to the large scale reservoir models. In the producing North Sea fields that were studied (Corbett 1993) the capillary effects were shown to be significant and their incorporation led to an improved understanding of the reservoir production process. The sediments studied in this work were shown to be anisotropic at the bedform scale and similar behaviour would be expected if this formation were considered as an oil migration carrier bed.

Discussion The above examples illustrate how the spatial distribution of permeability in rock media is of great importance in determining the effective mobility of two immiscible phases. Although great variation in sandstone architecture does occur, we believe that typical spatial permeability patterns can be established. Evaluation of bedform statistics at outcrop enable the likely amounts of spatial variation of permeability patterns to be determined. Where two-phase flow operates, these geologicallybased estimates of small-scale permeability architecture, supported by available core data, are generally preferable to the (often employed) average/uniform media approximations (see Fig. 3). Even when estimates of permeability variation cannot be made, our calculations for

CONTROLS ON TWO-PHASE FLOW typical sandstone media may help in identifying important transport mechanisms. The above calculations relate to oil reservoir production by water injection (at typical flow velocities of 0.2 m/day). Accordingly, the main concern has been the viscous/capillary scaling group. Natural system oil-water flow rates tend to be much lower (<m/year). Therefore, secondary oil migration processes are likely to be strongly capillary-dominated with important influences of gravity. The intrinsically-derived 'geopseudos' can address some of the problems associated with defining media properties in oil migration studies by extending the study of heterogeneity with respect to the gravity/ capillary ratio. The effects of wettability, and indeed the heterogeneous distribution of pore surfaces with different wettabilities, will also need to be considered. In mature oil provinces, like the North Sea, the availability of good data from producing fields (which has facilitated this study), should be used to improve our understanding of basin flow processes, with a view to developing and locating remaining fields more efficiently. Future oil migration studies should attempt to synthesize these architecture/flowdynamic effects with the geochemical and basin evolution models considered elsewhere. However, our numerical studies of heterogeneous sandstone systems lead us to speculate that many of the inferences on oil migration based on uniform-media approximations (e.g., Hermans et al. 1992) are invalid. We question whether continuous, uniform, high permeability pathways commonly exist in most aquifers. If, on the other hand, heterogeneities of the type described are pervasive, oil migration is likely to be inefficient and slow. This may be an explanation as to why published experimental studies (reviewed above) showed oil migration rates well in excess of likely geological, basin-scale, rates. Conversely, in order to achieve oil migration in heterogeneous sandstones some impetus appears necessary; either large pressure gradients or the opening of connected high-permeability pathways. The most obvious candidates for these pathways are episodically-open fault systems (Sibson 1993), releasing oil trapped by sedimentary and tectonic heterogeneities. Further work on the effective multi-phase flow properties of real (heterogeneous) rock systems is needed. This should focus on more appropriate laboratory study of two-phase flow in rock in order to improve the understanding of the process and on correctly scaled numerical models of reservoir and basin-scale flows to quantify the effects. We have, however, illustrated the importance of introducing a realistic

149

heterogeneous medium into assessments of these systems. Much of this work was done as part of the Reservoir Heterogeneity Project, funded by Amerada Hess, Bow Valley, British Gas, Chevron, Conoco, The Department of Trade and Industry, Deminex, Elf, Exxon, Mobil, and Shell. P.W.M.C. is also grateful to Elf Production Geoscience for financial support. We thank our colleagues G. Pickup, K. Sorbie, J. Jensen and J. Lewis for their insights, discussions and support. Intera Information Technologies Ltd are thanked for the provision of the ECLIPSE finite difference simulator used in this research. Thanks are due to J. Verweij, P. Lemouzey, an anonymous reviewer and the editor for encouraging this work to be presented to a new audience. References

AMvx, J.W., BASS, D.M. JR & WHITING,R.L. 1960. Petroleum Reservoir Engineering. McGraw-Hill, New York. ANDERSON,W.G. 1986. Wettability Literature Survey - Part 1: Rock/Oil/Brine Interactions and the Effects of Core Handelling on Wettability. Journal of Petroleum Technology, October 1986, 1125-1144.

BERG, R.R. 1975. Capillary pressures in stratigraphic traps. American Association of Petroleum Geologists Bulletin, 59,939-956. BURRUS,J., KUHFUSS,A., DOLIGEZ,B. & UNGERER,P. 1991. Are numerical models useful in reconstructing the migration of hydrocarbons? A discussion based on the Northern Viking Graben. In: ENGLAND, W.A. & FLEET, A.J. (eds) Petroleum Migration. Geological Society, London, Special Publications, 59, 89-109. CATALAN,L., XIAOWEN,F., CHATZlS, I. & DULLIEN, F. A.L. 1992. An experimental study of secondary oil migration. American Association of Petroleum Geologists Bulletin, 76,638-650. CORBE'rr,P.W.M. 1993. Reservoir characterisation of a laminated sediment: The Rannoch Formation, Middle Jurassic, North Sea. PhD Thesis, HeriotWatt University, Edinburgh. & JENSEN, J.L. 1993. Application of probe permeametry to the prediction of two-phase flow performance in laminated sandstones (lower Brent Group, North Sea). Marine and Petroleum Geology, 10,335-346. ~, RINGROSE, P.S., JENSEN, J.L. & SORBIE, K.S. 1992. Laminated clastic reservoirs - the interplay of capillary pressure and sedimentary architecture. Society of Petroleum Engineers Annual Technical Conference, Washington, DC, 4-7 October, 1992, Paper SPE 24699. CRAIG, F.F. 1971. The Reservoir Engineering Aspects of Waterflooding. Society of Petroleum Engineers, Monograph Series, 3, Richardson, Texas. DEMBICKI,H., JR & ANDERSON,M.J. 1989. Secondary migration of oil: Experiments supporting efficient movement of separate, buoyant oil phase along

150

P.S. RINGROSE & P.W.M. CORBETT

limited conduits. American Association of Petroleum Geologists Bulletin, 73, 1018-1021. DEMING, D., SASS, J.H., LACnENSRUCH,A.H. & DE RITO, R.F. 1992. Heat flow and subsurface temperature as evidence for basin-scale groundwater flow, North Slope of Alaska. Geology Society of America Bulletin, 104,528-542. ENGLAND, W.A. & FLEET, A.J. 1991. Introduction. In: ENGLAND, W.A. & FLEET, A.J. (eds) Petroleum Migration. Geological Society, London, Special Publications, 59, 89-109. ~, MACKENZIE, A.S., MANN, D.M. & QUIGLEY, T.M. 1987. The movement and entrapment of petroleum fluids in the subsurface. Journal of the Geological Society, London, 144,327-347. HALDORSEN,H.H. 1986. Simulator Parameter Assignment and the Problem of Scale in Reservoir Engineering. In: LAKE, L.W. & CARROLL,H.B. (eds) Reservoir Characterisation, Academic Press Inc., San Diego, 293-340. HERMANS, L., VAN KUYK, A.D., LEHNER, F.K. & FEATHERSTONE, P.S. 1992. Modelling secondary hydrocarbon migration in Haltenbanken. Norway. In: LARSEN, R.M., BREKKE, H., LARSEN, B.T. & TALLERAAS, E. (eds) Structural and Tectonic Modelling and its Application to Petroleum Geology. NPF Special Publications, 1, 305-323. Elsevier, Amsterdam. HURST, A. & ROSVOLL, K.J. 1991. Permeability variations in sandstones and their relationship to sedimentary structures. In: LAKE, L.W., CARROLL, H.B. JR. & WESSON, T.C. (eds) Reservoir Characterisation II. Academic Press, San Diego, California, 166--196. JENNINGS, J.B. 1987. Capillary Pressure Techniques: Application to Exploration and Development Geology. American Association of Petroleum Geologists Bulletin, 71, 1196-1209. KYTE, J.R. & BERRY, D.W. 1975. New Pseudo Functions to Control Numerical Dispersion. Society of Petroleum Engineers Journal, August 1975,269-275. LAKE, L.W. 1989. Enhanced Oil Recovery. Prentice Hall, New Jersey. LEVERETT, M.C. 1941. Capillary behaviour in porous solids. Transactions of the American Institute of Mechanical Engineers, 142, 152-169. McDOUGALL, S.R. & SORBIE, K.S. 1992. Network simulations of flow processes in strongly wetting and mixed-wet porous media. 3rd Conference on the Mathematics of Oil Recovery, 17-19 June,

1992, Delft University Press, Netherlands, 169181. PICKUP, G.E., RINGROSE,P.S. JENSEN,J.L. & SORBIE, K.S. 1994. Permeability tensors for sedimentary structures. Mathematical Geology, 26(2), 227250. RAPOPORT, L.A. 1955. Scaling laws for use in design and operation of water-oil flow models. Transactions of the American Institute of Mechanical Engineers, 204, 143. RINGROSE, P.S., SORBIE, K.S., CORBETT, P.W.M. & JENSEN, J.L. 1993a. Immiscible flow behaviour in laminated and cross-bedded sandstones. Journal of Petroleum Science and Engineering, 9, 103124. - - , FEGHI, F., PIcrtJP, G.E. & JENSEN, J.L. ]993b. Relevant reservoir characterisation: recovery process, geometry and scale. In Situ, 17, 55-82. SCHOWALTER, T.T. 1979. Mechanics of Secondary Hydrocarbon Migration and Entrapment. American Association of Petroleum Geologists Bulletin, 63,723-760. Sco'rr, E.S. 1992. The palaeoenvironments and dynamics of the Rannoch - Etive nearshore and coastal succession, Brent Group, northern North Sea. In: MORTON, A.C., HASZELDINE, R.S., GILES, M.R. & BROWN, S. (eds) Geology of the Brent Group. Geological Society, London, Special Publications, 61,129-147. SELLE, O.M., JENSEN, J.I., SYLTA,~, ANDERSON,T., NYLAND, B. & BROKS,T.M. 1993. Experimental verification of low dip, low rate two-phase (secondary) migration by means of ~/-ray absorbtion (extended abstract) Geofluids '93 Conference, Torquay, England, 4-7 May, 1993, 72-75. SmSON, R.H. 1993. Crustal stress, faulting, and fluid flow (extended abstract) Geofluids '93 Conference, Torquay, England, 4-7May, 1993, 137-140. TISSOT, B. 1987. Migration of hydrocarbons in sedimentary basins: a geological, geochemical and historical perspective. In: DOLIGEZ, B. (ed.) Migration of hydrocarbons in sedimentary basins. Editions Technip, Paris, 1-19. WORTmNGTON,P.F. 1991. Reservoir Characterisation at the Mesoscopic Scale. In: LAKE, L.W., CARROLL,H.B. & WESSON,T.C. (eds) Reservoir Characterisation IL Academic Press, San Diego, California, 123-165.

Origin of saline fluids in sedimentary basins J E F F R E Y S. H A N O R

Department of Geology and Geophysics, Louisiana State University, Baton Rouge, Louisiana 70803-4601, USA Abstract: Subsurface saline waters in sedimentary basins can be divided into three groups based on their anionic composition and salinity: (1) Waters with anions other than Cl dominant. These include Na-HCO3 and Na-acetate waters. Most such waters have salinities of less than 10 000 mg I-1; (2) CI-dominated, halite-undersaturated waters having salinities between 10000 and 250000-300000 mg1-1. These include Na-Cl waters and, at higher salinities, Na-Ca-CI waters; (3) Cl-dominated, halite-saturated waters with salinities typically in excess of 300000 mg I-~. Ca and K become increasingly dominant and Na decreases with increasing salinity. Subaerial evaporation of marine and continental waters and the subsurface dissolution of evaporites both have the potential for producing the range of salinities and dissolved chloride concentrations observed for most subsurface brines, but not their major cation compositions. The broad systematic increase in dissolved Na, K, Mg, Ca, and Sr and decrease in pH and alkalinity with increasing salinity support the hypothesis that the approach toward thermodynamic buffering by silicate-carbonate _+ (halide) mineral assemblages is a first-order control on subsurface fluid compositions, even at temperatures well below 100°C. The chemical potential of chloride or, alternatively, the aqueous concentration of anionic charge, is a master variable which ranks in importance with such other variables as pressure and temperature in driving fluid-rock exchange and controlling bulk fluid compositions. This variable is in turn controlled largely by physical processes of fluid advection and dispersion. Dissolved organic acid anions are associated primarily with low salinity waters, but dissolved metals, such as Cu, Pb, and Zn are preferentially found in brines having salinities in excess of 200 g I-~. The high chloride concentration and low pH of these saline waters may enhance solubilization of metals through chloride complexing.

Approximately 20% by volume of most sedimentary basins consists of pore waters of widely varying temperature, pressure, and chemical composition. The composition of these fluids provides important information on the geochemical, hydrological, thermal, and tectonic evolution of the Earth's crust (Hanor 1988) and insight into a number of important applied problems specifically related to the generation of ore deposits and to hydrocarbon exploration and production. Although discussion of the controls on the composition of pore waters is often presented as a strictly chemical problem, pore water compositions are a direct function of both chemical and physical factors (Hanor 1994). These factors include: (1) the composition of the water physically included in the pore spaces of the sediment at the time of sediment deposition; (2) the net effects of diagenetic exchange of components between the water and (a) the ambient solids which make up the matrix of the sediment and (b) any other fluids, such

as gases or liquid hydrocarbons, which may be present; (3) the net physical transport of material into and out of the sediments by bulk flow and the mixing of waters. All of the above can be accounted for in the following basic equation for the conservation of mass of a dissolved solute or isotopic species:

~C Ot

- "~ Rii + (-v.VC~) + (V.(DVC~))

(1)

where (OCJOt) is the change in the aqueous concentration, Ci, of solute or isotope i with time. Porosity, +, which often appears in such equations, has been assumed here to be constant with time and thus factored out. The first mass balance term on the right, ERij, describes the net effects of diagenetic reaction on the concentration of solute i. R u is the rate of the jth diagenetic reaction which results in the addition or removal of i from solution. The next term, (-v-VCi), where v is the fluid velocity field, defines the net addition or removal of dissolved i

From PARNELL,J. (ed.), 1994, Geofluids: Origin, Migration and Evolution of Fluids in Sedimentary Basins, Geolo~zical Society Special Publication No. 78, 151-174.

151

152

J.S. HANOR

as a function of time as a result of bulk fluid flow through the sediment. The last term on the right hand side of the equation, (V.(DVCi), describes the net rate in change in composition due to diffusional and dispersive mixing. D is a dispersion tensor which describes the magnitude of mixing which occurs in response to a given concentration gradient and fluid velocity. The initial, or connate, composition of the fluid, Ci(0), comes into play when we integrate equation (1) with respect to time. The physical processes of fluid flow and solute transport and dispersion are described in Hanor (1994) and will be mentioned here only in passing. The purpose of this review is to examine the origin of subsurface sedimentary waters from a geochemical perspective, i.e., the initial and present compositions of subsurface fluids and the Rij terms in equation (1). The emphasis will be on deeper, more saline fluids rather than on potable ground waters in shallow meteoric flow regimes. The review will begin with a brief historical overview and then will describe what is known about the composition of aqueous fluids in sedimentary basins. We will then examine hypotheses for the origin of these fluids. The plural, origins, is really more appropriate, because the water molecules and various dissolved solutes in deep waters are generally derived from diverse sources by diverse processes. The discussion will focus on major solutes and on minor and trace components of potential economic interest. This paper builds on earlier reviews on the origin and migration of subsurface brines (Hanor 1979, 1988, 1994), each of which has had a different emphasis. One of the several new contributions in the present paper is the use of a data set, synthesized from a wide range of published formation water analyses, which gives a truly global perspective to the composition of subsurface waters. Other recent perspectives on the origin of subsurface sedimentary fluids can be found in Land (1987, 1992).

Some definitions The term salinity, as used in this paper, is synonymous with total dissolved solids as determined either (1) directly by summing measured dissolved constituents or by weighing solid residues after evaporation, or (2) indirectly from electrical conductivity or spontaneous potential response. A much more precise definition of salinity exists for sea water, but is not required here. Typical concentration units for salinity and major solutes are mg1-1 or gl -I. Conversion to

mass/mass concentration units or molality requires knowledge of fluid density. Many reserve the term brine for pore fluids having salinities in excess of 100000 mg1-1. Carpenter's (1978) classification is the best formal scheme for those who would like to use one. The term brine will be used here loosely, without strict observance of the 100000 mgl -a lower limit. Connate (Latin for 'born-with') fluids are those included in sediment pore spaces at the time of deposition. Connate properties, such as connate isotopic composition or connate Br/C1 ratio, are the physical and chemical properties of an incremental volume of fluid at the time it was deposited as part of package of sediments. Because of chemical alteration and physical migration after deposition, few if any subsurface fluids are truly connate. This author has begun to use elevation (z) rather than depth (D) to represent the relative vertical position of waters because elevation is the variable which appears in numerical problems dealing with vertical fluid, solute and energy transport. Some of the figures thus use elevation rather than depth, where elevation can be taken to be negative depth.

Historical overview Hanor (1983, 1987) has reviewed in detail the history of thought on the origin of subsurface saline waters in sedimentary basins. The following brief discussion is abstracted from these papers. The existence of brine springs was known by prehistoric man, who used salt derived from these waters to preserve food. Written speculation on the origin of subsurface saline waters began nearly 2500 years ago by Anaxagoras (c. 50~428 Bc), who proposed that sea water was produced by evaporation of waters of the earth's interior, leaving a salty residue. Lucretius (c. 99-55 Bc) advocated subterranean cycling of ocean water into the continents and up and back out toward the sea as river water. In the process salt was filtered out. Thought on the origin of brines throughout the Middle Ages was dominated by the belief that sea water found its way into the interior of the earth and was distilled upward through the crust by local sources of heat. The vapor produced by this distillation condensed in cold cavernous spaces and ultimately trickled out onto the earth's surface as fresh water springs. By-products of these processes were salty brines and subsurface deposits of salt. Perrault, however, argued in 1674 on the basis of mass-balance arguments for

ORIGIN OF SALINE FLUIDS the pluvial origin of springs and against the idea that springs result from subterranean cycling and distillation of sea water. Continuing distillation, according to Perrault, should produce a saltfilled earth and an ocean free of dissolved salt. Discovery of rock salt in spatial proximity of subsurface brines, as happened in Cheshire, England in the 1670s, gave rise to the conclusion that salty waters were the product of subsurface dissolution of salt. Hitchcock (1845) summarized the general thought of many geologists of the nineteenth century. 'In many parts of Europe, salt springs are found rising directly from beds of rock salt; so that their origin is c e r t a i n . . . ' ; 'Most American geologists . . . maintain that our salt springs proceed from beds of rock s a l t . . . ' . With improvements in analytical chemistry it became possible to analyse the various dissolved constituents of saline waters. Hunt (1879) noted that in saline waters in lower Palaeozoic carbonates in Canada ' . . . only about one half of the chlorine is combined with sodium; the remainder exists as chlorides of calcium and magnesium, the former predominating, - while sulfates are present only in small amount'. This observation gave rise to an alternative hypothesis for the origin of subsurface brines. Hunt thought that these waters were fossil sea water of Palaeozoic age and that the composition of the oceans then must therefore have been significantly different than the composition of today's sea water. Lane (1908) invoked the Latin term connate (with-born) to describe waters of similar composition. Truly connate fluids of different ages could, in theory, be used to reconstruct the geochemical history of sea water. With the rapid development of the oil and gas industry in the early twentieth century, a large body of information on the composition of formation waters co-produced with hydrocarbons became available. The hypothesis that these oil-field brines were connate waters was soon taken to task. Washburne (1914) doubted that buried sea water would remain as stagnant masses of fluid during periods of deformation and subsequent circulation of meteoric ground water. Richardson (1917) further doubted that an active solvent like water would remain unchanged in chemical composition throughout geological time. The common occurrence of salty waters in gas-producing reservoirs encouraged the brief reappearance of subsurface evaporation as a brine-forming mechanism. Russell (1933), however, in an extensive review of various brineforming processes, showed on the basis of mass balance calculations that evaporation could at

153

best produce only small volumes of saline waters. Russell was apparently the first to suggest that brines might be produced instead by what he termed 'negative osmosis', the passage of water through compacted clay and shale membranes. With the steady improvement of analytical techniques and the widespread availability of analytical instrumentation, the body of information on the solute and isotopic composition of subsurface waters has increased greatly during the last several decades. Improvement in thermodynamic models for complex aqueous solutions and a better understanding of fluid flow in sedimentary basins have provided a rational basis for explaining many aspects of subsurface fluid chemistry. Current thought on the origins of subsurface sedimentary fluids will be reviewed in the remainder of this paper.

Composition of aqueous fluids in sedimentary basins Sources o f data The substantial increase in the body of analytical data available on the composition of subsurface waters in sedimentary basins in the past 20 years reflects in part interest in such waters as indicators of the diagenetic and hydrologic evolution of sedimentary basins. Table 1 lists the sources of analytical data utilized in preparing this review. An attempt was here made to select data sets that not only appear to be analytically reliable, but reflect a range of basinal settings, ages, and salinities. While some of these waters may have a meteoric heritage, they exist predominantly in non-meteoric fluid regimes. The list is by no means exhaustive. Other equally high quality data sets exist but have not been used because of space limitations. The bias towards North American waters simply reflects the greater availability of data from this continent. The analytical data referred to in Table 1 are utilized in this review in different ways. The data base has been used to prepare scatter plots showing the degree of covariance between different properties of these fluids. From this large data set, which comprises nearly 500 analyses, a smaller data set of 50 representative analyses was culled to make plotting of some compositional relations more manageable. The data are used in this paper primarily to establish a global or at least a continent-scale perspective on pore water compositions. The original references should be consulted for

154

J.S. HANOR

Table 1: Sources of data used in scatter plots and graphs Basin

Reference

Alberta Basin, Canada Central Mississippi, USA Illinois Basin, USA Michigan Basin, USA Michigan Basin, USA Northern Gulf Coast, USA Offshore Louisiana, USA Offshore Louisiana, USA Paris Basin, France

Connolly et al. (1990) Kharaka et al. (1987) Stueber & Walter (1991) Wilson & Long (1991) Case (1945) Moldovanyi & Walter (1992) Land et al. (1988)

Pattani Basin, S.E. Asia San Joaquin Basin, USA San Joaquin Basin, USA

Land & Macpherson (1989) Michard & Bastide (1988) Lundegard & Trevena (1990) Fisher & Boles (1990) White (1965)

Data from the following used in some figures as noted: Norwegian Shelf Southwest Louisiana, USA

Egeberg & Aagaard (1989) Hanor (1994)

discussions of the equally important variations in composition which exist on the intrabasinal and intraformational scale. Salinity

The salinities of pore waters in sedimentary rocks vary by approximately five orders of magnitude from a few milligrams per litre in shallow meteoric flow regimes to over 400000 mg1-1 in evaporite-rich basins such as the Michigan Basin, USA. The most saline formation water reported in the literature that this author is aware of is an unusual 643000 mg1-1 CaClz-brine from the Salina formation of the Michigan Basin (Case 1945). Formation water salinity in some sedimentary basins increases with depth, but there are notable exceptions, such as the south Louisiana Gulf Coast, where the most saline waters occur in the upper 3 km of the section (Fig. 1). There is typically a wide range in salinity at any given depth within a given sedimentary basin. As shown by Hanor and colleagues (Hanor et al. 1986; Hanor & Sassen 1990) these spatial variations in salinity provide useful constraints on the sources of dissolved salts and the physical processes which result in their transport and dispersal.

=

S o u t h w e s t Louisiana

-1000

i sea water ~ i TDS

~

-2000-

!

.O (~>

i' ~~ • • • /~ hydropressured : • / fiuids -3000 .... :-.'~.o .;,,t-~.-~ ............ .'. ...........................

E

~,,

°I

m,,v

•

..i.lj

LU

"ti

• .,,,,~. • qm

•

....

.tL....qIIL ...... i.e. .....................................

"

,~ overpressuredfluids

4000 -5000 0

........... 100

• ........ 200 300 TDS, g/L

400

Fig. 1. Total dissolved solids versus elevation (-depth) for waters in a portion of the southwestern Louisiana Gulf Coast (Hanor 1994). While many, if not most, sedimentary basins contain some waters having salinities well in excess of average sea water salinity of 35 g 1-I, not all sedimentary basins contain hypersaline waters. Well known examples include the San Joaquin Basin, California, USA (Fisher & Boles, 1990), and various basins in southeast Asia (Lundegard & Trevena 1990), where salinity in general is that of sea water or less. These tend to be basins devoid of evaporites. A useful concept in the investigation of the chemical evolution of formation waters is that of connate salinity, i.e., the salinity of the water trapped in a given pore space or volume of sediment at the time of sediment deposition. Today, this pore space or sediment may contain water of a far different salinity as the result of some combination of physical processes of displacement and dispersion and chemical processes of dissolution or precipitation. Comparison of connate and present salinities puts important constraints on diagenetic and mass transport models. For example, the clastic sedimentary sequence of the south Louisiana Gulf Coast was deposited in fluvial, deltaic, and normal marine environments. It can thus be concluded that few of these sediments contained fluids significantly more saline than approximately 35 g 1-1, normal sea water salinity, at the time of their deposition. Some of these same sediments today contain fluids with salinities greatly in excess of 100g1-1 as a result of large-scale vertical and lateral transport of brine (Hanor & Sassen 1990). Major anions

Chloride makes up over 95% by mass of the anions in most sedimentary formation waters

ORIGIN OF SALINE FLUIDS

Subsurface Brines

155

Subsurface Brines

Sr

100

8O

~ 60 E

60

E .2

t-

Ca

¢-

40

.£ N 40 0

20

20

<

Mg

O, 0

200000

400000

600000

TDS, mg/L

0

•

,

K

200000

400000

600000

TDS, mg/L

Fig. 2. Relative abundances by mass of major anions and cations in subsurface saline waters. Data from sources cited in text.

having salinities greater than 10 000 mg 1-x (Fig. 2). Less saline waters often have bicarbonate, sulfate, or acetate as the dominant anion. Explaining the origin of saline waters in sedimentary basins is thus in large part the problem explaining the origin of the dissolved chloride. M a j o r cations

In contrast to anionic composition, there is a profound and progressive change in the cationic makeup of sedimentary formation waters with increasing salinity (Fig. 2). Sodium is the dominant cation on a mass basis in low to moderate salinity waters. With increasing salinity, however, the relative proportion of Na decreases and the proportions of K, Mg, and Ca increase. Most notable is the increase in Ca, which typically is the dominant cation by mass in waters whose salinities exceed 300 000 mg 1-1. The reasons for this shift in cationic composition will be discussed later in this paper. Origin of high salinities

Many formation waters are far more saline than the continental or marine connate waters originally incorporated in the sediment at the time of deposition, and the most fundamental problem regarding the composition of these subsurface waters is the origin of their high salinity. Potential processes for generating high salinities are reviewed below. Subaerial evaporation Sea water. The subaerial evaporation of sea water is one of the most important processes

determining the composition of fluids in sedimentary basins. Removal of H20 by progressive evaporation not only produces brines which have the potential for infiltrating down into underlying sedimentary sequences, but results in the precipitation of a succession of Na, Ca, Mg, K and C1 evaporite minerals which may react with ambient pore fluids during burial, profoundly altering subsurface fluid compositions. The general principles governing the evolution in residual brine compositions which result from evaporation of sea water are well laid out in the classic paper by Carpenter (1978). One of the more recent and detailed studies of sea water evaporation is that of McCaffrey et al. (1987), who sampled and analysed brines and salt at a solar salt production facility in the Bahamas. The most saline subaerial brines were concentrated further in the laboratory. McCaffrey et al. report the degree of evaporation of the waters they studied in terms of the ratios (Mg brine/Mg seawater) and (Li brine/Li seawater). We shall refer to these ratios here simply as the evaporation index (EI) for sea water. McCaffrey et al. did not analyse for dissolved bicarbonate but note that calcium carbonate begins to precipitate out at the salt works at an evaporation index of approximately 1.8. All of the other major dissolved constituents progressively increase with increasing evaporation up to an EI of 3.8. At this point, gypsum begins to precipitate, and there is preferential removal of calcium from solution. At an EI of 10.6, halite begins to precipitate, and there is a remarkable depletion in sodium and decrease in the Na/Cl ratio of the residual brines over a very narrow range of C1 values. At evaporation indices

156

J.S. HANOR

Evaporated Sea Water 100

•

Evaporated Sea Water

, ,

Ca

100

80

80

Br

o~

60

g

E c.O 40

40

O 20

20

O,

0

200000

400000

, 600000

TDS, mg/L

0 0

200000

400000

600000

TDS, mg/L

Fig. 3. Relative proportions of major anions and cations in evaporated marine waters. Data from McCaffrey et al. (1987).

between 70 and 80, magnesium sulphate and potassium chloride salts begin to precipitate out. Evaporation of seawater produces two basic types of waters in terms of dominant components (Fig. 3). From a salinity of 35-330 g 1-] , the waters are dominated by Na-CI. At higher salinities, the waters become progressively dominated by Mg-CI-SO4 as a result of preferential removal of Na as halite. Note that while Mg becomes the dominant cation at extreme evaporation, chloride remains the dominant anion throughout the evaporation sequence. Although evaporation of sea water can produce waters having a range of salinities and chloride concentrations which can account for the range of salinities observed for most subsurface brines, evaporation of sea water by itself cannot account for the major anion and cation composition of subsurface brines (Fig. 2). Subaerially-evaporated sea water has sulphate as a major solute over all stages of evaporation. Sulphate in most subsurface brines, in contrast, is minor. The most saline marine brines are dominated by Mg; the most saline subsurface waters are dominated instead by Ca. While it is probable that some subsurface brines have had evaporated marine waters as their ultimate precursors, it is obvious that other processes have been at work to account for their major solute composition. Much of the field and laboratory work which has been done on the evaporation of sea water has taken place in closed system conditions, where there has simply been progressive removal of H20. Sanford & Wood (1991) have recently calculated the changes in fluid composition and the sequence of salts which would

precipitate under open system conditions, where there is continuous fluid loss due to infiltration into underlying sediments. Their work shows that in hydrologically-leaky systems, there may be a significant reduction in the maximum salinities of brines which can be produced by evaporation.

Bromine systematics during evaporation. The behaviour of the minor constituent bromine during subaerial evaporation of sea water provides important constraints on the later interpretation of the origin of halides in subsurface brines. Although bromide and chloride ions are both monovalent anions of similar ionic radii ( B r - = 1.96 .~,, C I - = 1.81 ~ , ) , C1- is strongly preferentially partitioned over Br- into Na, K, and Mg halogen salts during their precipitation, and Br- preferentially remains behind in the aqueous solution. During initial evaporation of sea water, both Br and CI increase in concentration in the residual hypersaline waters, and the Br/C1 ratio of these waters does not vary (Fig. 4). When halite saturation is reached, however, CI is preferentially precipitated out as halite. A small fraction of the Br is incorporated into the halite lattice as Na(C1-Br), but most remains behind in the aqueous solution. As a result, the Br/C1 ratio of the residual brine begins to increase with progressive evaporation. This is well illustrated in a plot of Br versus C1 for evaporated marine waters (Fig. 4). As saturation with respect to the K-Mg-C1 salts is reached, the slope of the Br-CI curve begins to flatten out, because these minerals discriminate against Br somewhat less than halite. The upper practical limit to Br concentrations which can be

ORIGIN OF SALINE FLUIDS

157

Subsurface dissolution of evaporites

6000

EvaporatedSea Water 5000-

K Salts

4000

E 3O00

Mg Salts

m

2000 1000 Sea Water

0

~

T

Halite Saturation

100000 200000 Cl, mg/L

300000

Fig. 4. Variation in dissolved Br and CI during the progressive evaporation of sea water. Data from McCaffrey et al. (1987).

produced by evaporating seawater is approximately 6000 mg 1-1. The practical upper limit for CI is approximately 250000 mg 1-~ . Brines formed by subaerial evaporation of sea water should in theory thus have elevated Br/Cl ratios. Brines formed instead by the dissolution of halite, should have low BrFFDS (Rittenhouse 1967) and Br/C! ratios (Carpenter 1978).

Evaporation of continental waters. Surface brines produced by evaporating continental waters are much more diverse in composition than those produced by evaporation of sea water. This is not surprising given the highly variable starting compositions these waters can have (Berner & Berner 1987). The general controls on the composition of continental lake brines have been discussed by Eugster & Hardie (1978), Eugster & Jones (1979) and references therein. Total dissolved solids in a collection of 27 analyses of saline lakes in western North America (Eugster & Hardie 1978) range from 5510 to 336000mg 1-I. With the notable exception of some highly saline MgSO4 brines, Na is generally the most abundant cation, as it is over the comparable range of salinities in evaporated seawater. One of the more significant differences between evaporated marine and evaporated continental waters, however, is in their anionic signature. Highly saline continental waters can have CI, SO4 or even HCO3 or CO3 as the dominant anion. While some continental brines resemble subsurface brines in overall chemical composition and salinity, there can thus be significant differences, particularly in anionic composition and the relative abundance of dissolved Mg.

Field evidence. An alternative mechanism for producing subsurface brines of high salinity is the dissolution of chloride-bearing evaporites. Well-documented field examples (e.g., Bennett & Hanor 1987) exist in the US Gulf Coast, where the spatial variation in pore water salinity around some salt domes provides clear evidence for the dissolution of these halite-dominated diapirs as the source of brines. Contrary to those who claim that the dissolution of a salt dome influences salinity only in the immediate area of the dome (Macpherson 1992), field mapping of variations in salinity indicates that vertical transport of dissolved salts has occurred over distances of many kilometres vertically and tens of kilometres laterally throughout the surrounding sedimentary sequences (Fig. 5) (Bray & Hanor 1990; Hanor & Sassen 1990). The generally low Br/C1 ratio of the pore fluids in this part of the Gulf Coast is consistent with dissolution of halite rather than subaerial evaporation as the source of CI. Incongruent dissolution of evaporites. Many evaporite sequences contain suites of different evaporite mineral phases. The occurrence of interbedded halite and calcium sulphate minerals, for example, is common. Evaporite sequences formed from residual brines of very high salinity often contain halite and K and Mg salts. Subsurface dissolution of these mixed evaporite assemblages may produce brines which have composition reflecting preferential or incongruent dissolution of the bulk evaporite sediment. Where salt dome dissolution occurs in proximity to meteoric flow regimes, halite is preferentially dissolved and anhydrite is left as a solid residue (Posey & Kyle 1988). Anhydrite may then be hydrated to gypsum, which in turn can be converted to calcite by the bacterial reduction of sulphate and oxidation of hydrocarbons. McManus & Hanor (1988), in contrast, present evidence for the dissolution of both halite and anhydrite simultaneously in deep, saline environments surrounding salt domes. Anhydrite solubility here may be enhanced by the high salinity of the ambient pore fluids. In shallow environments, a brackish Na-C1 fluid is produced; in deeper environments, dissolution may result in a Na-C1 brine with more substantial concentrations of dissolved Ca and SO4. A second example of incongruent dissolution involves the behaviour of Na-K-Mg-C1 mineral assemblages during progressive burial. The aqueous solubility of most halide salts increases

158

J.S. HANOR

NORTH

SOUTH

f!

SALINITY, g/L

£

i i "

o

/

I

I

I \

c~ 10 O2J 0 I

10 mi I

0

I

I

I

17 D I t\i f 2'

e--

Darrow '

salt : ~-'~a1"~20-

.. ~ . . S t . Gabriel D°rne { : ' : : / ' x x x ~,Salt Dome " ~ x ,

~.

i

L4

l

1() km

Fig. 5. Cross-section of a portion of the Louisiana G u l f Coast showing spatial variations in salinity reflecting

subsurface dispersion of dissolved salt away from two salt domes (Bray & Hanor 1990).

with increasing temperature. The solubility of halite, however, increases much less rapidly over the range 25-200 ° C than that of K, Mg, or Ca salts (Phillips et al. 1981). A fluid saturated with respect to a mixture of chloride minerals and subjected to increasing temperature will preferentially dissolve the K, Mg and Ca phases and can actually precipitate halite. The K and Mg phases typically have elevated concentrations of Br, and this type of incongruent dissolution has the capacity to produce Br/CI ratios in excess of those normally associated with subaerial evaporation (Hanor 1988).

M e m b r a n e filtration The need to account for the occurrence of subsurface brines in evaporite-poor sedimentary basins was one of the driving forces behind the investigation in the 1960s to early 1980s of a potential brine-forming process variously known as reverse osmosis, ultrafiltration, or membrane filtration. It was thought problematical, for example, that brines having salinities exceeding 100g 1-1 should exist in the Illinois Basin, USA, which is nearly devoid of evaporites (Graf et al. 1966). The process of formation of brines by membrane filtration involves the hydraulically-driven flow of fluid across semipermeable shale or clay beds (Graf 1982). Neutral water molecules pass through shales more readily than do dissolved ions, which are electrostatically repulsed as a consequence of the presence of electrical double layers around clay mineral grains. In theory, pore fluids on the influent side of a shale membrane will thus become progressively more saline as cross-formational flow and the selective filtration of cations and anions continues. There is no question that compacted shales and clays can behave as semipermeable membranes as has been demonstrated by numerous laboratory experiments (Demir 1988) and by the

typical spontaneous potential log response of brine-saturated clays which reflects retardation of anionic diffusion (Hanor 1988). There is also some evidence that the presence of semipermeable strata can induce osmotic flow of water, that is, the flow of water from fresh to salty beds (e.g. Marine & Fritz 1981). What has been questioned, however, is in what specific basinal settings, if any, is the large-scale production of highly saline waters by membrane filtration important.

A simple field test for membrane filtration. It should be possible to apply the following simple field test in basinal settings where membrane filtration is thought to be an active and significant brine-forming process. As a first-order approximation, membrane filtration can be considered in terms of two competing mass transport processes: (1) hydraulic advection of H20 as a result of a hydraulic force field, and (2) molecular diffusion of H20 arising from a difference in the chemical potential of H20. Other processes inducing the transport of H20 may exist as a result of spatial variations in temperature and electric charge, but will be neglected here. Solute transport must be considered as well. Membrane filtration will produce residual brines only if the net advection and diffusion of solutes is small relative to the advection of H20. The magnitude and direction of the advective flux of water molecules, Jadv, is given by the relation: Jadv = - ( k l ' q ) ( V P - pg)

(2)

where k is intrinsic permeability, r I is fluid viscosity, VP is the fluid pressure gradient, and pg is specific weight. The magnitude and direction of the diffusive flux of water molecules, Jdif, is given by: ./die = -- D~'I-I, H20

(3)

ORIGIN OF SALINE FLUIDS

PREDICTEDSALINITY, g/L(schematic)

HYDRAULICHEAD, km 0

E

1

2

3

4

I

1

I

O

'1

12

MASSIVESANDS

159

o ~ R

lOO I

i

l

ESH-WATER ZONE

I

MEMBRANE EFFLUENT

J

BRINES " ~

OBSERVEDSALINITY,g/L 2OO I

I

0 [

m

too

200 ,

I

HALITE-DERIVED

Q

~_~

NDS_& SHALES

3,1-

4L

--

s-s,vE

(,) -'I/ ~

*-

(b)"

RESIDUAL I MARINEWATERS I

(e)

Fig. 6. Diagram showing vertical variation in isodensity hydraulic head in a typical section of the south Louisiana Gulf Coast (a). Hydraulic head is used here as a proxy for hydraulic force (VP - pg). If membrane filtration were an important brine-forming process here, residual brines should be accumulating within the massive shales of the overpressured zone at the base of the section shown (b). Instead, observed salinities decrease downward (c). (Hanor, unpublished.)

where D is the sediment diffusion coefficient for H20 and IXn2o is the chemical potential of H20. Taking the standard chemical potential of water to be the chemical potential of pure water at any T and P of interest, changes in chemical potential are due only to the presence of solutes. In general, the higher the salinity (TDS), the lower the chemical potential of H20, hence: IXu2o = f(1FFDS)

(4)

Assuming immobility of charged solutes and the absence of other driving forces, membrane filtration can thus occur only where Jadv > -Jdif. Inspection of equations 2, 3 and 4 demonstrates that membrane filtration should be favored in hydraulic settings where ( 7 P - pg) is large and oriented in the same direction as the salinity gradient. Among the largest natural fluid pressure gradients known are those which occur in the transition from hydropressured to overpressured fluid regimes in the south Louisiana Gulf Coast. Because this transition zone usually occurs within a shaly interval, this then should be the ideal setting for the production of brines by membrane filtration. Indeed, the Louisiana Gulf

Coast has often been cited as a type field area where membrane filtration is occurring (Graf 1982). As originally pointed out by this author (Hanor 1984), however, if membrane filtration were the dominant control on pore water salinities in the Louisiana Gulf Coast, then salinities should decrease upward through the deep overpressured sequence (Figs 2 and 6). It is obvious that the observed salinities do not correspond to trends that would be generated by membrane filtration. The saltiest waters are found a b o v e the zone of high hydraulic gradients and massive mudstones. Pore water salinities actually decrease with depth within the pressure transition and the geopressured zone. Although the mudstones exhibit membrane behaviour, as evidenced by their spontaneous potential response, upward flow of fluids through the overpressured zone may be taking place through fracture porosity and faults, rather than the matrix porosity of the sediments. The membranes are thus bypassed. This author is not aware of a single well-documented field example of the production of brines by membrane filtration.

160

J.S. HANOR

Marine aerosols: brines f r o m rain water The Murray Basin, southeast Australia, provides an example of an interesting variant on the role of subaerial evaporation of sea water in the formation of brines. The Murray Basin consists of an approximately 500m thick sequence of Palaeocene to Recent sediments deposited in fluvial to marine setting (Jones et al. 1994). The upper 100 m of the central portion of the basin now contains large volumes of subsurface brines with salinities in excess of 100 g 1-1 which have major solute compositions very much like that of marine and evaporated marine waters. The isotopic composition of many of these fluids, however, is that of evaporated meteoric water. This observation has led recent workers (e.g. Hanor & Evans 1988; Jones et al. 1994) to the conclusion that the solutes in these brines have been derived from the long-term accumulation of marine aerosols. The marine salts are introduced either in dry form or as meteoric precipitation. Because evapotranspiration greatly exceeds annual rainfall there is a net accumulation of solute as brackish surface waters. These waters are further evaporated to brines in lakes in the central part of the basin. The shallow ground water flow regime is not sufficiently dynamic to displace these brines, and they have been infiltrating downward and accumulating above a regional aquitard.

Other mechanisms The occurrence of saline waters in fractured Precambrian crystalline rocks in the Canadian Shield (Frape & Fritz 1987) and Hercynican granites of the Cornwall tin mines (Edmunds et al. 1987) has prompted the suggestion that the elevated salinities in these settings are the result of hydrolysis of Cl-bearing silicates, such as micas and amphiboles. While this mechanism is reasonable for the particular waters in question, it is less likely on mass balance grounds to be the source for the large mass of CI found in most sedimentary basins. Mills & Wells (1919) concluded on the basis of the common association of salty waters in gas-producing reservoirs that the saline waters are the residue formed as a result of the evaporation of pore fluid into gases produced in the subsurface. Russell (1933) showed by simple mass balance arguments that impossibly large volumes of methane would be required to produce significant quantities of brines through evaporation. It may, however, be necessary to invoke a mechanism such as this on a local scale to account for the extraordinary salinity of the

643000mg 1-1 CaCl2-brine from the Salina formation of the Michigan Basin described by Case (1945). The salinity of this water is far in excess of that which can be produced by normal subaerial evaporation or the dissolution of most evaporites. Other mechanisms of primarily historical interest are reviewed by Hanor (1983, 1988).

Reduction in salinity Not all formation waters have elevated salinities relative to their connate precursors. Examples of hyposaline waters in normal marine sediments occur in the US Gulf Coast, particularly in the deep overpressured section (Fig. 2). The apparent reduction of salinity may be due to incursions of meteoric water and/or the in situ production of HaO by the dehydration of mineral phases or organic matter during diagenesis.

Controls on major solute and isotopic composition Major cations and p H Most subsurface waters have major cation compositions which cannot be explained by either: (1) burial or infiltration of subaeriallyevaporated marine or continental waters, or (2) by subsurface dissolution of evaporites. There is, however, a growing body of evidence which supports the hypothesis that major cation compositions in formation waters are extensively buffered or influenced by an approach toward metastable thermodynamic equilibrium with ambient sedimentary mineral phases. Although there is up to an order of magnitude spread in the absolute concentrations of some of the major cations waters at any given chloride concentration, the first order trend for each cation over most of the salinity range is that of increasing concentration with increasing salinity (Fig. 7). An exception is Na, which decreases with increasing salinities beyond approximately 300000 mg 1-1. For moderately saline waters, the monovalent cations Na and K show an approximately 1:1 slope and the divalent cations Mg, Ca, and Sr an approximately 2:1 slope with respect to total dissolved solids on log-log scatter plots. The pH of these waters decreases with increasing salinity from typical values of 7-9 in moderately saline waters to values between 3 and 4 at high salinity, i.e., there is a progressive increase in the activity of H ÷, and the waters become more acidic. Such observations for

ORIGIN OF SALINE FLUIDS 1000000'

Sodium

+ooo ..................+................ , ................ +.....- ° ° 7

10000 1000"

2~

100

'2, .

Magr~esium

1000000"

'l

100000

_..I '~

161

,oooo.........

~

~

iooo

E d,,

[

t

1000

10000

: .........

+

.

.

11 100

1000000

.

.

.

.

.

.

.

i

:

10000'

Y 1000

I~o

I T

E Ioo-

......

"

"

i

,

~

w

..... l....

1

~......

i,1" _,i~ _~ '

10-

/

i .

t

f

O0

1000

10000

100000

,1

',

.

_J

•

-1000000

.

.

.

.

.

I I

i

.

.

.

.

.

.

.

.

-; ..............

100 .................. . . . . . . . . . . . . . . . . . . . . . . . .

O

10 2:1slop i e ~ i - - - - ~

Saturation i I

.1

1000000

100

1000

10000

100000

1000000

TDS, rng/L

1000000

Strontium

100000

+

"

=. 10000"

8

"IQ.

+

+ 100000

................. . ................ :-

~"

,

pH

9

+

i 10000

TDS, mg/L

10

+ ........... ++

Calcium

10000

~

+..-. . . . . . . . .

lOO°.i iiiiiiiiiiii:i i j

1oooooo

!

i

.......

TDS, mg/L

Potassium

100000' ................ +

; ~-Z';,iJ. .....

,~

TDS, mg/L

1000000'

" "

•

. i.~J~.=-

,f ..-y,++

[

100000

+

i .'.'~

.^

Saturation I ,

.1 IO0

~

~

..................~,................!.........~

/ 2 : 1 slope

1

i

~

1oo ..................

:+

10

i

i"

7 ................................

•

!"~

.!.~

...... Ill

..J "~

"

1000'

,.~ O0

6

................

4, ................

i

•

i

E

•.,

100'

.................

~....,.

......

° .4

.......

10'

s.

~

4'

i

3 100

1F '~. '

, 1000

10000

100000

1000000

TDS, mg/L

1 .1 100

1000

10000

100000

1000000

TDS, mg/L

Fig. 7. Scatter plots showing the covarience of log Na, K, pH, Mg, Ca and Sr with log TDS for typical saline waters. Sources of data listed in text. Note the 1:1 slopes for Na and K and the 2:1 slopes for the divalent cations.

waters in the salt d o m e province of the southern Louisiana Gulf Coast led this author (Hanor 1988) to suggest that the first order controls on pore fluid compositions in this region are the

subsurface dissolution of halite and subsequent buffering of pore water compositions by multiphase silicate-carbonate mineral assemblages. Dissolution of salt alone should produce a

162

J.S. HANOR

progressively NaCl-dominated fluid without the observed significant increases in K, Mg, Ca and

106 CalculatedF}uid Compositions i0 s .

St.

Metastable thermodynamic buffering. The idea

that the compositions of subsurface brines are at least partially thermodynamically buffered is implicit in the discussion of DeSitter (1947) and was specifically invoked as far back as Carpenter & Miller's (1969) work on saline water in the Ozark Dome. Thermodynamic buffering as a control on the composition of subsurface waters has been proposed as well for other basinal waters (e.g., Merino 1975; Nesbitt 1985; Hanor 1988; Land et al. 1988; Michard & Bastide 1988; Smith & Ehrenberg 1989; Hutcheon et al. 1993 and references therein). Testing this hypothesis by evaluating the saturation state of pore fluids with respect to common aluminosilicate and carbonate minerals is complicated by the fact that necessary data are often lacking, as is the case with many of the data sets used in this paper. Take, as an example, equilibrium involving albite: NaA1Si3Os + 4H+ = Na + + A13+ + SiO2 ° + 2H20

(5)

Calculating the saturation state of an aqueous fluid with respect to albite requires having reliable values for pH and dissolved sodium, aluminium, and silica, reliable data for the Gibbs free energy of formation of albite and all aqueous species, and an adequate thermodynamic model for the non-ideal behavior of the aqueous phase. Although significant improvements have been made in thermodynamic modeling of aqueous solutions (Helgeson et al. 1981; Pitzer 1987), methods for the treatment of the non-ideal behavior of some key species, such as carbonate and silica, in hypersaline, N a - C a dominated brines are still problematic. In addition, reliable analytical data for dissolved A1 and pH are often lacking. One technique for partially obviating these problems is to invert the process and calculate the bulk solution compositions which should exist in the presence of known or assumed mineral buffers and compare these with observed compositional trends. The results of one such calculation are presented in Fig. 8, which shows the variation in the concentrations of major dissolved species in fluids as a function of dissolved chloride in waters in equilibrium with the hypothetical mineral assemblage, quartz-muscovite-albiteK feldspar-calcite-dolomite at an arbitrary temperature of 100° C, a pressure of 1 bar, and a CO2 fugacity of 10 -3 bars. These constraints, plus the concentration of chloride, are sufficient

I

-r~ 104

.... ~ " -

~"

~

103

[

-~"" I .",

.Na

K

~

101 ¢0

100 10-1 103

J 104 102 Chloride, mg/kg H20

106

Fig. 8. Results of a computer simulation showing the calculated variation in major solute composition as a function of dissolved chloride for fluids buffered by a multiphase (see text) silicate-carbonate assemblage at fixed P, T andfco2 (Hanor 1994).

to fix uniquely the chemical potentials of dissolved species in the nine-component system: Na20-K20-MgO-CaO-A1203-SiOT-CO2HC1-H20 (or, alternatively, NaC1-KC1-MgCI~CaCI2-AICI3-SiO2-CO2-HC1-H20). Allowing CO2 to be mineral-buffered (c.f., Smith & Ehrenberg 1989) requires adding an additional solid phase, for example, chlorite, to the system. Hydrolysis constants for mineral phases and for CO2(g) were taken from Bowers et al. (1984). Conversion between aqueous activities and concentrations was made using a program which utilizes a Pitzer virial equation of state for aqueous solutions (Harvie et al. 1978; Pitzer 1987). While the particular simulation pictured here does not match all of the field data precisely, there are some important general similarities, including the progressive increases in dissolved Na, K, Mg, and Ca, and the decrease in pH and alkalinity with increasing salinity (Figs 7 and 11). Lower salinity waters are Na-dominated, but divalent cations, particularly Ca, become much more important constituents as salinity increases. By changing the buffer system, fco2, and/or T, it is in fact possible to shift the compositional trends of each of the cations and produce model waters dominated by Ca. It is highly unlikely on the basis of the complex and variable petrology of sedimentary rocks and the observed range in formation water compositions at any given chloride value and temperature that a single buffer system is responsible for the range of compositions observed in natural waters. Scatter in the field data sets presented here presumably reflect varying

ORIGIN OF SALINE FLUIDS

163

Quartz saturated, 2b'~C 12"

I

Albite +

-!-

Smectite

+~ z 6

--/--

--

/

I Chlorite

O4

"iv

10-

1

Od

r"

m o

14

I

K-feldspar CD

3 .Kaol

J

6 Kaolinite

mO

Albite

Ilite 2

I

0 0

3

6

9

log a (K+/H +)

2

I 4

I 6

8

log a (Na+/H + )

Fig. 9. Phase diagrams showing mineral stability as a function of the activity ratios Na/H, K/H and Mg/H 2. Fluid composition (black circles) is buffered by an arbitrary phase assemblage. An increase of TDS by an order of magnitude in moderately saline fluids requires approximately an order of magnitude increase in the most soluble cation, Na ÷ (Fig. 7). An increase in the activity of Na ÷ of an order of magnitude requires that the activities of H ÷ and K + also each increase by an order of magnitude and that the activity of Mg2÷ increases by two orders of magnitude. Hence the decrease in pH and the ca. I : 1 and 2:1 slopes on log plots of cations v. TDS on Fig. 7.

P-T conditions, different buffer assemblages, and department from equilibrium. Egeberg & Aagaard (1989) for example, have documented differences in composition between waters from carbonate and sandstone reservoirs in the Norwegian Shelf. Some waters may not be cation-buffered at all, including some saline waters from southwest Louisiana, whose compositions can most simply be explained by congruent dissolution of bulk salt dome evaporites (Hanor 1994). The necessity for having a large number of silicate mineral phases to buffer a multicomponent fluid system may mean that some formation waters in mineralogicallysimple sandstones and carbonates, which are the usual waters sampled, are strongly influenced by reactions occurring in ambient or intercalated mudstones and shales. Covarience in cation composition with salinity. The 1:1 covarience in the log-log plots of the concentrations of monovalent cations with TDS and the 2:1 covarience of divalent cations can be explained as follows. Na is the dominant cation in many subsurface waters simply because the silicate minerals which buffer it are in general the most soluble, not because NaCi may be the ultimate source. Because Na is the dominant cation up to salinities of approximately 330 g 1-1, it follows that an order of magnitude change in

TDS should be complemented by approximately an order of magnitude covarient change in Na. Hence the log-log plot of N a v . TDS has a slope of approximately 1 : 1. Consider next Fig. 9, which shows two generic phase diagrams which relate mineral stability and aqueous fluid composition. Equilibrium with respect to the hypothetical mineral assemblage, quartz, Na-smectite, albite, illite, and chlorite fixes the activity ratios (aNa+/aH+), (aK+/aH+), and (aMg2+/(aH+)2) at any given temperature and pressure. An increase in the activity of Na + of one order of magnitude requires an increase in the activity of H + of one order of magnitude, i.e., the solution becomes more acid. The stoichiometries of the coupled hydrolysis reactions buffering fluid compositions require an increase in the activity of monovalent ions, such as K +, of one order of magnitude, and the activity of divalent cations, such as Mg 2+, of two orders of magnitude. Even though the relations between the concentrations and the activities of the individual cations are strongly non-linear at elevated ionic strengths, broadly similar trends are reflected in the concentration scatter plots, and K + increases by approximately an order of magnitude with increasing salinity, but the divalent cations Mg 2+, Ca 2+, and Sr 2+ increase by two orders of magnitude. Sr may be buffered with respect to Sr-bearing calcites.

164

J.S. HANOR 10000

Southern Arkansas ....j~

1000 NorthSea~ ~ ,~;i S"~~: !'.~..~.]. ..........~.~j . , ~ " ~'~ ,. ,,"o ........' ° a . . ~ Z

100 ~..~i~..4~.~ " ~ 133

10 1'~ .........................

Sea Water Evaporatton Trend

SouthwestLouisiana

11 ................. lOOOO t ooooo Chloride, mg/L

1oooooo

Fig. IlL Br-C1 systematics for several sedimentary basins. The solid black line is the sea water evaporation trend (Fig. 4). Sources of data given in text. These relations fail at salinities above approximately 300g 1-1, where saturation with respect to halite is reached. Here, Na concentrations become progressively smaller with increasing salinity as a result of the increase in C1. There is an abrupt increase in K + as buffering shifts from a strictly silicate-carbonate assemblage to one involving halite as well.

Chloride and bromide Not all components in sedimentary formation waters are buffered by ambient mineral phases. The single most important example is chloride. Subsurface waters having salinities of 300 g 1- i o r less are undersaturated with respect to halite, sylvite, and other chloride-bearing mineral phases, and chloride concentrations in these waters are controlled primarily by mass transport processes of advection and dispersion, not by thermodynamic equilibrium with respect to one or more mineral phases. Even in the US Gulf Coast where a compelling case can be made for the origin of chloride by subsurface dissolution of salt domes and by burial of previously halite-saturated brines, most waters have salinities well below the >300g 1-~ levels expected at saturation. An important exception includes the Michigan Basin, where some pore waters are clearly halite-saturated (Wilson & Long 1992). Bromine is similarly non-buffered, and Br-CI systematics, although potentially complex (Hanor 1988), are thus more likely to provide clues as to the identity of saline endmember water types than other pairs of solutes. The high Br/C1 ratios of waters in the Smackover Formation, USA, for example, support the hypothesis that Br-rich, subaerially produced brines are

important saline endmembers in this natural system (Moldovanyi & Walter 1992) (Fig. 10). The low Br/C1 values of waters in South Louisiana, USA, in contrast, are consistent with the idea that high salinity is derived from the dissolution of halite-dominatred salt domes, which typically have low bulk Br/C1 ratios. The waters of the Norwegian Shelf have yet a third signature, one that supports the conclusion of Egeberg & Aagaard (1989) that there has been at least some contribution from subaerially produced brines.

Bicarbonate and sulphate The alkalinity of most of the formation waters shown in Fig. 11 is contributed predominantly by bicarbonate and organic acid anions. The latter species will be discussed in more detail later in this paper. Alkalinity in general decreases with increasing salinity. There are three probable reasons for this. First, the organic acid anions are associated primarily with low-salinity waters possibly derived by dewatering of organic-rich mudstones. Second, in a calcium carbonatebuffered system, carbonate alkalinity should decrease with an increase in dissolved calcium. Finally, the increase in H + with increasing salinity shifts dissolved carbonate and bicarbonate toward carbonic acid: HCO3- q- H + ~ H2C03 °

(6)

Sulphate shows little or no systematic variation with total dissolved solids. Recent work by McManus & Hanor (1988, 1993) provides clues as to why this is the case in the South Louisiana Gulf Coast. The isotopic composition of massive carbonate and pyrite cemented sands near the West Hackberry salt stock just south of Calcasieu Parish supports the hypothesis that the sulphur in the pyrite was derived from salt stock anhydrite and the carbon in the carbonate is from a methane source. The net reaction has involved the thermogenic reduction of sulphate and oxidation of methane and the consequent precipitation of iron sulphides and calcium carbonate. The concentrations of dissolved sulphate in this area thus controlled by relative rates of: (1) release of sulphate through salt dissolution, (2) solute transport and (3) removal by reduction. The later process, of course, is further dependent upon the availability of suitable reducing agents, such as hydrocarbons.

Dissolved aluminium and silica Dissolved Al concentrations in most subsurface waters are generally less than I mg 1-1 . There are

ORIGIN OF SALINE FLUIDS 10000

Alkalin LIty

~• •

_1 1000 O)

E

" "Z~

"-#

-

i

.................... ii i "

1000

..*.. '~'2"'Wll--i ~1~ ........d"J '~"ii,"........... *4"

__..1 e~

100

Sulfate

10000

.............. , ......

-

"

•0

i

165

100

....................... +

. . . . . . . . . . . . . . . . . . . . . . .

i

.,t ~]~'t"~

!-~,~

~

Q• '.o-11

L~

'

. . . .

.........i..................

r¢i

"

, . - -

10

i

1

........

1000

lOO

•

i

i

. . . . . . . .

10000

"•7-

i

.

--

100000

......

1000000

mg/L

TDS,

1

1 lOO

........ i ........ i . . . . . . . lOOO

lOOOO

lOOOOO 1000000

TDS, mg/L

Fig. 11. Variation in alkalinity and dissolved sulfate as a function of TDS for typical saline waters.

1000

J

Amorphous Silica Saturation

~

lOO

E 0 oo

.

buffered by silicate assemblages such as kaolinite-smectite.

..~

..ai" , ; ~

.I,,,~..~ ~•~i~7• lO

~- . .

uua~ m,. Saturation

•

/

end• ° o

o

T, deg C Fig. 12. Variation is dissolved silica versus temperature for typical saline waters. Curves show concentrations at quartz and amorphous silica saturation.

insufficient data of high quality at present to establish the systematics of dissolved A1 in deep sedimentary basins. However, simply based on the observation that Al-minerals are in general more soluble in low and in high pH fluids, and that there is a decrease in pH with salinity, one might expect that greater solubilization and transport of A1 could take place in waters of either very low or very high salinity, i.e., fresh waters associated with the dehydration of mudstones and halite-saturated brines. Many authors have noted that although quartz is a nearly ubiquitous phase in sedimentary basins, most basinal waters are not in thermodynamic equilibrium with quartz (Fig. 12). It is not known whether silica values are kinetically controlled or to what extent silica may be

Rates o f reaction. The absolute rates of diagenetic reactions, R, in basinal settings are in general not well known. Recent summaries of work on a number of important chemical systems are included in the collection of papers edited by Kharaka & Maest (1992). The evidence above, however, suggests that the rates of some hydrolysis reactions involving major species are sufficiently rapid, even at modest sedimentary temperatures, that the drive toward thermodynamic equilibrium between formation waters and ambient silicate and carbonate mineral phases is an important control on composition of formation waters. Isotopic composition

Recent investigations, unfortunately beyond the scope of this review, have established the importance of a wide variety of isotopic systems in establishing the sources of components in subsurface waters and the rates of fluid flow. Examples include isotopes of H, O, C, S, B, Sr, Nd, Sm, U, Th, C1, and I (see Banner et al. 1989, 1990 for examples). Of the many isotopic systems of interest in formation waters, 6D-51so systematics has been the most extensively used in problems of determining marine versus meteoric sources for H20 and for investigating the degree of water-rock exchange, particularly of oxygen. The H- and O-is•topic composition of formation waters has been reviewed by Sheppard (1986). The characteristic shift of values to the right of the meteoric water line, and the increase in 6180 values of formation waters

166

(.)

J.S. HANOR 20

(b)

/••,SMOW #

,.

-1000

q,

E =- -2ooo

o.'.-=.~. •

Oe 0

•

o

-40

•

-3000 /

-60

oOo

" - ..'.z

*~qP"**'' "

MWL . /

-20

0

/.

•

•

0

•

nnn

•

:'.".,o

""

° •

¢

-8O

-

o•

-4000"

• ••o

,

,

-5

0 8180

-5000

5

10

,

-10

0 ,5180

Fig. 13. D-O isotopic compositions of saline waters from master data set (a). O-is•topic composition of these waters as a function of elevation and, hence, increasing temperature (b).

with increasing temperature (Fig. 13) are thought to reflect exchange with and partial buffering of ~ 8 0 by silicate and carbonate minerals. In some sedimentary basins, there is a covarience in ~D and ~180 which may reflect mixing of light waters having significant meteoric component with heavier waters derived either from: (1) marine or evaporated marine precursors or (2) from waters having undergone isotopic exchange with ambient minerals. Other isotopic systems provide additional information on the degree of rock-water exchange. The 87Sr/86Srratios of formation water, for example, are typically higher than connate sea water values for the host sediments, reflecting in most cases the introduction of radiogenic Sr from the dissolution of silicates. There is a general decrease in 8VSr/86Srvalues of formation waters in the data sets utilized here with increasing Sr (not shown). It is possible that the Sr in the Sr-enriched waters, which are also the most saline waters, has been preferentially derived from carbonates, which should have less radiogenic values than siliciclastic minerals.

Chloride as a master variable in diagenesis The importance of the chemical potential of chloride, or, alternatively, HC1, in the thermodynamic interpretation of processes involving natural aqueous solutions was recognized over 20 years ago by Helgeson (1970). This general concept has been utilized by Giggenbach (1984) and others as a charge balance constraint in the analysis of high-temperature hydrothermal

processes. With the exception of work by Hanor (1988) in the Gulf Coast and Michard & Bastide (1988) in the Paris Basin, it has not attracted as much attention in lower-temperature basinal studies. In the specific examples discussed by Helgeson, CI activity was held constant and pH allowed to vary. In sedimentary basins, however, the case can be made on the basis of the discussion above that chloride is the more important master variable and that pH is more likely to be buffered. Figure 14 illustrates an example of the potential effect that changing chloride concentration through dilution, dissolution of a halide phase, or by mixing of waters of different salinity can have on driving diagenetic exchange of major solutes. Even though on log-log plots there is a striking linearity between the concentrations of major cations and TDS (and hence chloride), the variation in concentration of a rock-buffered cation with chloride is actually non-linear. In the example shown, waters A and B are mixed, producing a water of intermediate salinity and cation composition. On the assumption that this mixed water will attempt to achieve partial equilibrium with its surroundings, substantial amounts of Na will be dissolved and K precipitated. Dispersive dilution of a K-rich brine during regional fluid migration, for example, may be a reasonable hypothesis for the K-metasomatism, which occurred during Alleghenian regional fluid migration in eastern North America (Hearn & Sutter 1985). The amounts of the various solids which should be dissolved or precipitated as a result of changes in chloride concentration can be estimated using reactionpath calculations or mass balance techniques.

ORIGIN OF SALINE FLUIDS 120000

167

30000 @

100000

o °

•

80000

Oo

•

20000

-..:

_J tD

E 60000 d Z 40000

I

20000

¢"° •• IPCb

•

•

•

.

~')

//

A

E

• • •

,'.

/ •

"-

"" Q

0

..

,oooo

•

0

,

, 100000

, 200000

!

0J 300000

CI, m g / L

,, 0

• --~JSm~

' i

•

I

llLk, k*MSP. 100000 200000

300000

CI, m g / L

Fig. 14. Diagrams showing observed variations in dissolved Na and in K with chloride for saline waters. Changes in chloride concentrations by dilution, dissolution of salt, or by dispersive mixing (for example, mixing of waters A and B to produce C) should induce chemical disequilibrium and rock-water mass transfer. In the example shown, K-silicates should be precipitated and Na-silicates destroyed.

B r i n e s a n d organic m a t t e r

A number of dissolved constituents of subsurface aqueous fluids in sedimentary basins are profoundly influenced by the burial diagenesis of organic matter. Most important are those which are either direct or indirect participants in redox reactions. In turn, aqueous fluids can play a critical role in the redistribution of organic matter, particularly hydrocarbon fluids, in sedimentary basins (England et al. 1987). The details of the geochemical evolution of organic matter during burial diagenesis are beyond the scope of this chapter. We will briefly review, however, some of the net effects that reactions involving organic compounds can have on aqueous fluid geochemistry.

A q u e o u s organic species A wide variety of dissolved organic compounds have been found in sedimentary formation waters. These include minor amounts of neutral organic molecules of hydrocarbons, such as the alkanes, cycloalkanes, and aromatics (McAuliffe 1980) and charged anionic species, such as the mono- and dicarboxylic acid anions (Carothers & Kharaka 1978).

Dissolved hydrocarbons. As noted by McAuliffe (1980), there is a marked decrease in aqueous solubility with increasing carbon number for each class of hydrocarbons present in petroleum. Trends are nearly linear on a plot of log solubility versus carbon number. For the nalkanes, which have the general formula

C,,H2,,+2, there is an expotential decrease in solubility of at least six orders of magnitude from pure methane (n = 1) to pure dodecane (n = 12). There is a pronounced flattening out of solubilities beyond a carbon number of 12, with n-alkanes having C numbers from 12 to 36 being accommodated in water to approximately the same degree, i.e., 0.001--0.01 mg 1-1 for pure compounds. The cycloalkanes are slightly more soluble than the n-alkanes for a given carbon number (McAuliffe 1980), and the aromatics are roughly two orders of magnitude more soluble than the n-alkanes over the range n = 6 to 12. Solubilities of most dissolved hydrocarbons increase with increasing temperature, but decrease with increasing salinity (Price 1976).

Organic acid anions. Many aspects of the systematics of organic acid anion production and generation are still incompletely known. Within the data sets considered in this paper, acetate, CH3COO-, is the dominant dissolved organic species and exists in concentrations of up to approximately 2000mg 1-1. The concentrations of the other organic anions are in general much less than 100 mg 1-1 . Although concentrations of acetate exceeding 10000ppm have been reported by Surdam et al. (1984), some published high values appear to be analytically suspect (Hanor et al. 1993). There is a pronounced decrease in the range of concentration of acetate with increasing salinity. The apparent preferential association of acetate with low salinity waters may reflect their production in organicrich mudstones undergoing both mineral dehydration and maturation of organic matter. Based

168

J.S. HANOR

(a)

(b) 0

2000

+ o

Formate Acetate

0o -1000 ~ _~9, o

x Propionate A i-Butyrate

1500

[]

E il~"~ X X ~ o ff -200o~° d o o o

Butyrate

"~1~ ~= 1000

.0

o

.~

o

500

Oo

o

o

o

o~

o oO

0

200000

300000

400000

o o

-5000 / 0

TDS, mg/L

o

o q,Oo%°o

OoO

0

o , 1000

500

o cbo o

O oOO

%0® -4000 "~

==100000

o oO oo o~b

,:~oo OOo

,,,

0~ 0

Io o

-3o0o,

o

_

oo '1~O

o

Acetate,

1500

2000

mg/L

Fig. 15. Scatter plots showing variation in organic acid anions as a function of salinity (a) and dissolved acetate

as a function of depth (b).

on the data portrayed in Fig. 15, it would be a mistake to conclude that acetate is a major constituent in typical subsurface brines having salinites in excess of 100000 mg !-1 . Initial work by Carothers & Kharaka (1978) on VFAs in oil-field waters from California and Texas suggested that there is a temperature control on the distribution of VFAs. Subsequent work by these and other authors, however, has demonstrated that a several order of magnitude range in concentrations exists at all temperatures below 200 ° C. This spread is well illustrated in a plot of the variation in concentration in acetate versus elevation for the data sets considered here (Fig. 15).

Redox state o f subsurface brines Although free oxygen is normally not present in measurable quantities in deep subsurface environments, it is often convenient to write redox reactions with 02 as the electron acceptor and to relate redox state to the fugacity of oxygen, fo2. This will be convention followed here for many of the reactions considered. Most solid phases and waters deposited in continental and open marine sedimentary environments contain elemental components in relatively high oxidation states. Examples include Fe 3÷ in Fe-bearing mineral phases produced by weathering and S6÷ in dissolved sulphate. A notable exception is C in biologically-derived carbon compounds other than CO2. The nominal valence state for C in most such compounds and their burial derivatives varies from - 4 to 0, and organic carbon thus exists as a potential electron source for the

biologically-mediated or inorganic reduction of other elemental components in the system. An important example is the thermogenic reduction of sulphate to sulphide with hydrocarbons as the electron source (McManus & Hanor 1988). The approximate redox state of a sediment and its included pore water can sometimes be inferred from the presence of various authigenic mineral assemblages and/or the activities of dissolved species which may be in homogeneous or heterogeneous redox equilibrium. The presence of authigenic magnetite or pyrrohotite (McManus & Hanor 1988), for example, sets upper limits on probable oxygen fugacities of 10 -4° bars or less under normal basinal temperatures (Henley et al. 1984). Shock (1988) discusses evidence which suggests that there is metastable equilibrium between dissolved carboxylate and carbonate species in the oil-field brines he considered: CH3COO-

+ 202 = 2HCO3- + H +

(7)

and between dissolved carboxylate species: 2CH3CH2COO-

+ 02

=3CH3COO- +H +

(8)

but marked disequilibrium between carboxylate species and low carbon number atkanes, such as methane: CH3COO- + H + #COz + CH4

(9)

Helgeson et al. (1993) discuss evidence for the existence of metastable equilibria between higher molecular weight hydrocarbons, here represented by the liquid alkane C,,H2,,+2, and aqueous organic species, here represented by acetate, C H 3 C O O - :

ORIGIN OF SALINE FLUIDS 2C,H2,,+z(L) + (n+l)O2(g) = nCH3COO- + H + + 2H20

(10)

The above reactions, as noted by these authors, could be coupled to mineral-fluid equilibria through such reactions as: 2C,H2,,+2(L) + [(3n+ 1)/2102(8) + nCa2+ = nCaCO3 + 2nil + + nH20

(11)

and CH3COO- + 202 + 2Ca + = 2CACO3 + 4H +

(12)

As illustrated in the reactions above, many redox processes involve the production or consumption of hydrogen ions. In acidunbuffered systems, production of hydrogen ion may induce acid hydrolysis (equation 5). In pH-buffered systems, however, mineral-fluid equilibria may instead play an important role in buffering fo2, if stable or metastable equilibria exists between reduced and oxidized carbon compounds.

169

inclusions in ore and gangue minerals. Chemical analyses indicate that many of the fluids responsible for precipitating main stage mineralization in MVT deposits were Na-Ca-CI, low-sulphate brines having salinities in the range of 100-300 g kg -1 (Hanor 1979; Sverjensky 1984, 1986). It has generally been assumed that to qualify on mass balance grounds as a potential ore-forming fluid, a basinal fluid must contain on the order of 1 mg 1- i or more of the dissolved metal in question.

Metals in subsurface brines Within the data sets evaluated here Pb, Zn or Cu values are most likely to exceed 1 mg 1-1 in waters having salinities in excess of 200000mg 1-1 (Fig. 16). Such waters also have high chloride concentrations and low pHs. Possible reasons for the preferential association of metals with highly saline brines may thus include the solubilization of metals by chloride complexing and low pH, as for example: CuO + 2C1- + 2H + ~ CHCI2° 4- H2O (13)

Exchange with organic liquid phases Maturation of organic material often produces organic liquid and gas phases which may exchange components with a coexisting aqueous phase. One example is introduction of aqueous carboxylic acids by the oxidation of liquid alkanes, as noted in equation 10 above. Another is the loss of dissolved volatiles, such as methane, carbon dioxide, and hydrogen sulphide, through pressure release and degassing (Hanor 1988).

Ore-forming components in sedimentary waters

Ore-forming fluids Sedimentary formation waters have long been invoked as ore-forming fluids in a number of distinctly different geological settings. Although ore deposit classification schemes vary from author to author, general types of ore deposits which have been genetically associated with basinal fluids include (Force et al. 1991): (1) Mississippi-Valley type Pb, Zn, Cu, Ba, and F deposits; (2) shale-hosted Pb, Zn, Ba deposits; (3) rift-basin and redbed Cu deposits; (4) sandstone-hostedUdeposits. Much of the information available on the nature of the fluids which were involved in the genesis of these deposits comes from physical and chemical measurements made on fluid

High salinity by itself, however, does not guarantee a high dissolved metal content. Many of the most saline brines in which metals were measured have metal concentrations well below 1 mg 1-I. Considering the potential importance of the subject, there has been surprisingly little work done on the systematics of the distribution of metals in subsurface waters. An exception is the well-documented occurrence of Pb and Zn enriched brines of the Mississippi Gulf Coast (Carpenter et al. 1974; Kharaka et al. 1987), which provides clues as to other factors operating in the deep sedimentary environment which may control dissolved metal concentrations. A plot of modal concentrations of dissolved Pb and Zn versus stratigraphic interval (Fig. 17) supports the hypothesis of Carpenter and colleagues that these metals have been derived by reduction of iron oxide grain coatings in the Hosston Formation, a Lower Cretaceous red bed unit, possibly during reduction of iron oxide grain coatings which release Fe and other metals. Hanor (1988) has proposed that the configuration of the concentration gradients reflects vertical dispersive transport of these metals above and below this source bed. Low values in the Jurassic section below may reflect the presence of a sulphide sink. Many of the areas of the Upper Jurassic in central Mississippi contain HzS-rich natural gases, and it is possible that Pb and Zn are in the active process of being precipitated out of solution. It was early noted by White (1965) and Sawkins (1968) that the average K/Na ratio of

170

J.S. H A N O R

(b) 1000

300

o o: ;,o

100

o-. ~. %

10

°

._J

o %o,O. ,.o

.01 o

o°

o

a

200

~

s~o%~ ~%1m% ~=oO i ~ .ooi

E

• Pb (mg/L) o Zn (mg/L) Cu (mg/L)

.-I

" 100 • o

Pb Zn

&

Cu

r

.001 0

100000

200000

300000

TDS,mg/L

400000

200

0

400

600

Acetate,mg/L

800

1000

Fig. 16. Variation in dissolved metals in formation waters as a function of salinity (a) and dissolved acetate (b). Metals are generally more abundant in Ci-rich waters than in waters having high concentrations of organic acids. These data support the hypothesis that complexing by chloride is a more important control on metal concentrations.

o) 0

B

0

D~ o

Chloride

-~Wash.-Fred.-~

i/

,.- 0

Pb

[\

~~N~ n

o 0

Sligo ~ : . . v . : ~ . : , . . - . . .

.!i!!i!!!ii,i,i~to.~:i:Fimi:i,i.

3

1 /JJ

....

"Cottonvaiiey :::::::::::::::::::::

'::: "

'

'

'""

L

.............:.:.:.:.:.::::

!iiiiiiiiiiii:i:: _li>;ouano.,i!t °

I

100

200

g/I-

i k I

300 0

•

i

100

•

!

200

•

I

300

mg/L

Fig. 17. Vertical stratigraphic section showing variation in modal values of dissolved Pb and Zn by formation in the central Mississippi salt basin, USA. Data from Carpenter et al. (1974).

MVT ore fluids appeared to be higher than that of average subsurface brines. Hanor (1979) noted, however, that the elevated K/Na ratios consistent with K/Na ratios of more saline, deeper, and hotter fluids in basins. This is reflected as well in the present data set, in which the most saline waters have elevated K/Na ratios.

Organics as metal-complexing agents Interest in the role of dissolved organics as complexing agents for metals such as Pb, Zn and Cu has been prompted by the purported low solubility of metals in what have been taken to be typical formation waters. A plot of dissolved metals versus acetate for waters in the data set

ORIGIN OF SALINE FLUIDS 1000

precipitation would have sulphate concentrations less than 200mg 1-1 but could have salinities ranging anywhere from 10000 to 300000 mg 1-1.

l:~slope ! •

:.

i

100

.,i

E

d

..~ ~.~- 8 r,...e-,,&_.

10

Summary

lid

.........................

1

..°T~

1

171

10 100 S04, mg/L

1000

10000

Fig. 18. Variation in dissolved Ba as a function of dissolved sulfate for typical saline fluids. Note inverse relation reflecting possible buffering of fluid compositions by barite.

used in this paper, however, show a pronounced inverse correlation between metal content and acetate concentrations (Fig. 16). Metal concentrations well above I mg 1-1 occur only in low acetate waters. At acetate concentrations greater than approximately 50mg 1-1, metal concentrations are negligible. Because acetate is the most abundant organic anion in most of these waters, it is unlikely that the occurrence of high metal concentrations is in any way related to the concentration of acetate or any of the other organic species analyzed (see also Kharaka et al. 1987; Moldovanyi & Walter 1992). There is a simple explanation for this inverse correlation. High concentrations of organic acids are associated with less saline fluids; high concentrations of metals are associated with the most saline fluids. It is thus more likely that metal solubilities are enhanced by high chloride concentrations and low pHs than by organic anions, at least for the present data set.

Barium in subsurface brines There is no significant trend between barium and salinity, and barium, unlike its group IIA cousins Mg, Ca and Sr, is not typically buffered by silicate-carbonate equilibrium reactions. There is, however, a very rough inverse correlation between Ba and sulphate (Fig. 18) which is consistent with the hypothesis that equilibrium with respect to barite (BaSO4) may be one factor controlling Ba concentrations (see also Macpherson 1989). Not surprisingly, the highest Ba waters are low in sulphate. Fluids that would have the potential for transporting > 10mg 1-1 barium to the site of ore-mineral

Subsurface saline waters can be divided into three groups based on their salinity and anionic composition: (1) waters with anions other than C1 dominant. With possible rare exceptions, these are waters having salinities of less than 10000 mg 1-1. These include Na-HCO3 and Na-acetate waters; (2) Cl-dominated, halite-unsaturated waters. Typical salinity range 10000-250000300000 mg 1-1. These include Na-C1 waters and, at higher salinities, Na-Ca-CI waters; (3) Cl-dominated, halite-saturated waters. Typical salinities in excess of 300000mg 1-1. Ca and K become increasingly important solutes with increasing salinity. Na decreases with increasing salinity. Subaerial evaporation of marine and continental waters and the subsurface dissolution of evaporites have the potential for producing brines having both the range of salinities and dissolved chloride concentrations of most subsurface brines, but not their major cation compositions. No convincing field example of the formation of highly saline brines by membrane filtration has yet been documented. Although the concentrations of major solutes within subsurface saline waters range by a factor of ten or more at a given salinity, the broad systematic increase in dissolved Na, K, Mg, Ca and Sr and decrease in pH and alkalinity with increasing salinity support the hypothesis that the approach toward thermodynamic buffering by silicate-carbonate + (halide) mineral assemblages is a first-order control on subsurface fluid compositions, even at temperatures well below 100° C. The increasing dominance of dissolved Ca over Na and the increasing importance of K with increasing salinity, for example, can be explained simply by the stoichiometry of mineral-fluid hydrolysis, the thermodynamic nonideality of brines, and the buffering of Na concentrations by halite saturation at extreme salinities. The chemical potential of chloride or, alternatively, the aqueous concentration of anionic charge, is a master variable which ranks in importance with such other variables as pressure and temperature in driving fluid-rock exchange and controlling bulk fluid compositions. This variable in turn is controlled

172

J.S. H A N O R

largely by physical processes of fluid advection and dispersion. W h e r e the c o m p o s i t i o n of the fluid is largely r o c k - b u f f e r e d , its ultimate origin and its pathway of chemical evolution m a y be o b s c u r e d , at least in t e r m s of its m a j o r solute c o m p o s i t i o n , by its most r e c e n t P - T - X e n v i r o n m e n t . S o m e n o n - b u f f e r e d c o m p o n e n t s , such as C1 a n d Br, are m o r e likely to retain i n f o r m a t i o n on original e n d - m e m b e r fluid compositions. Dissolved organic acid anions are associated primarily with low salinity waters, but dissolved metals are preferentially f o u n d in brines having salinities in excess of 200 g 1-1 . T h e high chloride c o n c e n t r a t i o n and low p H of these waters may e n h a n c e solubilization of metals t h r o u g h chloride complexing. This work was supported in part by NSF Grants EAR-8803889 and EAR-9019342. I thank G. Bagnetto-Waters for her help with data reduction, J.A. Nunn and A. Sarkar for many valuable discussions on fluids in sedimentary basins, and A. Carpenter and W. Sanford for their helpful reviews of this script.

References BANNER, J.L., WASSERBURG,G.J., CHEN, J.H. & MOORE, C.H. 1990. 234U-238U-23°Th-232Th systematics in saline groundwaters from central Missouri. Earth and Planetary Science Letters, 101,296-312. - - , DOBSON, P.F., CARPENTER,A.B. & Moore, C.H. 1989. Isotopic and trace element constraints on the origin and evolution of saline groundwaters from central Missouri. Geochimica et Cosmochimica Acta, 53,383-398. BENNETT, S.S. & HANOR, J.S. 1987. Dynamics of subsurface salt dissolution at the Welsh Dome, Louisiana Gulf Coast. In: LERCH, I. & O'BRIEN, J.J. (eds) Dynamical Geology of salt and related structures. Academic Press, Orlando, 653-677. BERNER, E.K. & BERNER,R.A. 1987. The global water cycle, geochemistry and environment. PrenticeHall, Inc., Englewood Cliffs. BOWERS, T.S., JACKSON, K.J. & HELGESON, H.C. 1984. Equilibrium activity diagrams. SpringerVerlag, New York. BRAY, R.B. & HANOR, J.S. 1990. Spatial variations in subsurface pore fluid properties in a portion of southeast Louisiana: Implications for regional fluid flow and solute transport. Gulf Coast

Association of Geological Societies Transactions, 40, 53-64. CAROXHERS,W.W. & KnARAKA,Y.K. 1978. Aliphatic acid anions in oil-field waters - Implications for origin of natural gas. American Association of Petroleum Geologists Bulletin, 62, 2441-2453. CARPENTER,A.B. 1978. Origin and chemical evolution of brines in sedimentary basins. Oklahoma Geological Survey Circular, 79, 78-88.

& MILLER, J.C. 1969. Geochemistry of saline subsurface water, Saline County, Missouri. Chemical Geology, 4, 135-167. , TROUT, M.L. & PICKET, E.E. 1974. Preliminary report on the origin and chemical evolution of lead- and zinc-rich oilfield brines in central Mississippi. Economic Geology, 69, 1191-1206. CASE, L.C. 1945. Exceptional Silurian brine near Bay City, Michigan. American Association of Petroleum Geologists Bulletin, 29,567-570.

CONNOLLY,C.A., WALTER,L.M., BAADSGAARD,H. LONGSTAFFE, F.J. 1990. Origin and evolution of formation waters, Alberta Basin, Western Canada Sedimentary Basin. I. Chemistry. Applied Geochemistry, 5,375-396. DEMIR, 1. 1988. Studies of smectite membrane behavior: electrokinetic, osmotic, and isotopic fractionation processes at elevated temperatures. Geochimica et Cosmochimica Acta, 52,727-737. DESITrER, L.U. 1947. Diagenesis of oil-field brines.

American Association of Petroleum Geologists Bulletin, 31, 2030-2040. EDMUNDS,W.M., KAY,R.L.F., MILES,D.L. & COOK, J.M. 1987. The origin of saline groundwaters in the Carnmenellis granite, Cornwall (U.K.): further evidence from the minor and trace elements. Geological Association of Canada Special Paper, 33,127-144. EGEBERG, P.K. & AAGAARD, P. 1989. Origin and evolution of formation waters from oil fields on the Norwegian shelf. Applied Geochemistry, 4, 131-142. ENGLAND,W.A., MACKENZIE,A.S., MANN,D.M. & OUIGLEY, T.M. 1987. The movement and entrapment of petroleum fluids in the subsurface. Journal of the Geological Society, London, 144, 327-347. EUGSTER, H.P. & HARDIE,L.A. 1978. Saline lakes. In: LERMAN, A. (ed.) Lakes- Chemistry, Geology, Physics. Springer-Verlag, New York, 237-293. & JONES, B.F. 1979. Behavior of major solutes during closed basin brine evolution. American Journal of Science, 279,609-631. FISHER, J.B. & BOLES, J.R. 1990. Water-rock interaction in Tertiary sandstones, San Joaquin Basin, California, U.S.A.: Diagenetic controls on water composition. Chemical Geology, 82, 83-101. FORCE, E.R., EIDEL, J.J. & MAYNARD, J.B. 1991.

Sedimentary and diagenetic mineral deposits: a basin analysis approach to exploration. Reviews in Economic Geology, 5. FRAPE, S.K. & FRITZ, P. 1987. Geochemical trends from groundwaters from the Canadian Shield.

Geological Association of Canada Special Paper, 33, 19-38. GIGGENBACH, W.F. 1984. Mass transfer in hydrothermal alteration systems - a conceptual approach. Geochimica et Cosmochimica Acta, 48, 2693-2711. GRAE, D.L. 1982. Chemical osmosis, reverse chemical osmosis, and the origin of subsurface brines. Geochimica et Cosmochimica Acta, 46, 14311448. , MEENTS,W.F. & SHIMP,N.F. 1966. The origin of

ORIGIN OF SALINE FLUIDS

-

saline formation waters, III: Calcium chloride waters. Illinois Geological Survey Circular, 397, 60. HANOR, J.S. 1979. The sedimentary genesis of hydrothermal fluids. In: BARNES, H.L. (ed.) Geochemistry of Hydrothermal Ore Deposits (2nd ed.). John Wiley & Sons, New York, 137-172. 1983. Fifty years of development of thought on the origin and evolution of subsurface sedimentary brines. In: BOARDMAN,S.J. (ed.) Revolution in the Earth Sciences - A d v a n c e s in the Past Half Century. Kerdall/Hunt, Dubuque, 99-111. 1984. On the non-importance of membrane filtration as a mechanism for producing subsurface brines in the Louisiana Gulf Coast. Clay Mineral Society, 1984 Annual Meeting, Program with Abstracts, 55. 1987. History of thought on the origin of subsurface sedimentary brines. In: LANDA, E.R. & INCE, S. (eds) The History of Hydrology. American Geophysical Union, Washington, 81-91. -1988. Origin and migration of subsurface sedimentary brines. Society of Economic Paleontologists and Mineralogists Short Course 21. 1994. Physical and chemical controls on the composition of waters in sedimentary basins. Marine and Petroleum Geology, 11, 31-46. --, BAILEY, J.E., ROGERS, M.C. & MILNER, L.R. 1986. Regional variations in physical and chemical properties of South Louisiana oil field brines. Gulf Coast Association of Geological Societies Transactions, 36,143-149. -& EVANS, W.R. 1988. Subsurface brines from rain water: the salinization of the Murray Basin, Australia. Geological Society of America Abstracts with Program, 20, A364. -& SASSEN, R. 1990. Evidence for large-scale vertical and lateral migration of formation waters, dissolved salt, and crude oil in the Louisiana Gulf Coast. In: SCHUMACHER, D. & PERKINS, B.F. (eds) Gulf Coast Oil and Gases: Their characteristics, origin, distribution, and exploration and production significance. Proceedings 9th Annual Research Conference, Gulf Coast Section Society of Economic Paleontologists and Mineralogists Foundation, 293-296. HARVIE, C.E., MOLLER, N. & WEARE, J.H. 1984. The prediction of mineral solubilities in natural waters: The Na-K-Mg-Ca-H-CI-SO4-OHHCO3-H20 system to high ionic strengths at 25°C. Geochimica et Cosmochimica Acta, 48, 723-751. HEARN, P.P. & SUTI'ER,J.F. 1985. Authigenic potassium feldspar in Cambrian carbonates: evidence of Alleghenian brine migration. Science, 228, 15291531. HELGESON, H.C. 1970. Description and interpretation of phase relations in geochemical processes involving aqueous solutions. American Journal of Science, 268,415-438. - - , KIRKHAM,D.H. & FLOWERS,G.C. 1981. Theoretical prediction of the thermodynamic behavior of aqueous electrolytes at high pressures and -

173

temperatures, Part IV. American Journal of Science, 281, 1249-1516. - - , KNox, A.M., OWENS, C.E. & SHOCK,E.L. 1993. Petroleum, oil field brines, and authigenic mineral assemblages: are they in metastable equilibrium in hydrocarbon reservoirs? Geochimica et Cosmochimica Acta, 57, 3295-3340. HENLEY, R.W., TRUESDELL, A.H., & BARTON, P.B. 1984. Fluid-mineral equilibria in hydrothermal systems. Reviews in Economic Geology, 1. HITCHCOCK, E. 1845. Elementary Geology. Newman, New York. HUNT, T.S. 1879. Chemical and geological essays. Osgood and Co., Boston. HUTCHEON, I., SHEVALIER,M. & ABERCROMBIE,H.J. 1993. pH buffering by metastable mineral-fluid equilibria and evolution of carbon dioxide fugacity during burial diagenesis. Geochemica et Cosmochimica Acta, 57, 1017-1028. JONES, B.F., HANOR, J.S. & EVANS, W.R. 1993. Sources of dissolved salts in the central Murray Basin, Australia. Chemical Geology, I l l , 135154. KHARAKA,Y.K. & MAEST, A.S. (eds) 1992. Proceedings of the 7th International Symposium on Water-Rock Interaction. Balkema, Rotterdam. - - , CAROTHERS,W.W., LAW, L.M., LAMOTHE, P.J. & FRIES, T.L. 1987. Geochemistry of metalrich brines from central Mississippi Salt Dome basin, U.S.A. Applied Geochemistry, 2,543-562. LAND, L.S. 1987. The major ion chemistry of brines in sedimentary basins. American Institute of Physics Conference Proceedings, 154,160-179. 1992. Saline formation waters in sedimentary basins: Connate or diagenetic? In: KHARAKA,Y. & MAEST, A. (eds) Water-rock interaction. Balkema, Rotterdam, 865-868. ~, MACPHERSON, G.L. & MACK, L.E. 1988. The geochemistry of saline formation waters, Miocene, offshore Louisiana. Gulf Coast Association of Geological Societies Transactions, 38,503-511. LANE, A.C. 1908. Mine waters and their field assay. Geological Society of America Bulletin, 19, 501512. LUNDEGARD, P.D. & TREVENA, A.S. 1990. Sandstone diagenesis in the Pattani Basin (Gulf of Thailand): history of water-rock interaction and comparison with the Gulf of Mexico. Applied Geochemistry, 5,669-686. MACPHERSON, G.L. 1989. Sources of lithium and barium in Gulf of Mexico formation waters, USA. In: MILES, D.L. (ed.) Water rock interaction (WRI-6). Balkema, Rotterdam, 453-456. 1992. Regional variations in formation water chemistry: major and minor elements, Frio Formation fluids, Texas. American Association of Petroleum Geologists Bulletin, 76, 740-757. MARINE, I.W. & FRITZ, S.J. 1981. Osmotic model to explain anomalous hydraulic heads. Water Resources Research, 10, 89-107. MCAULIFFE, C.D. 1980. Oil and gas migration: chemical and physical constraints. American Association of Petroleum Geologists Studies in Geology, 10, 89-107.

174

J.S. H A N O R

MACPHERSON, G.L. 1989. Geochemistry of formation water, Plio-Pleistocene reservoirs offshore Louisiana. Transactions of the Gulf Coast Association of Geological Societies, 39, 421-430. MCCAFFREY, M.A., LAZAR, B. & HOLLAND, H.D. 1987. The evaporation path of seawater and the coprecipitation of Br- and K + with halite. Journal of Sedimentary Petrology, 57,928-937. MCMANUS, K.M. & HANOR, J.S. 1988. Calcite and iron sulfide cementation of Miocene sediments flanking the West Hackberry salt dome, Southwest Louisiana, U.S.A. Chemical Geology, 74, 99-112. - & ~ 1993. Diagenetic evidence for massive evaporite dissolution, fluid flow, and mass transfer in the Louisiana Gulf Coast. Geology, 21, 727-730. MERINO, E. 1975. Diagenesis in Tertiary sandstones from Kettleman North Dome, California - II. Interstitial solutions: distribution of aqueous species at 100°C and chemical relation to the diagenetic mineralogy. Geochimica et Cosmochimica A cta, 39, 1629-1645. MICHARD, G . & BASTIDE, J.-P. 1988. Etude geochimique de la nappe du Dogger du Bassin Parisien. Journal of Volcanism and Geothermal Resources, 35, 151-163. MILLS, R.V. & WELLS, R.C. 1919. The evaporation and concentration of waters associated with petroleum and natural gas. United States Geological Survey Bulletin, 693. MOLDOVANYI,E.P. & WALTER,L.M. 1992. Regional trends in water chemistry, Smackover Formation, southwest Arkansas: Geochemical and physical controls. American Association of Petroleum Geologists Bulletin, 76,864-894. NESBI1"r, H.W. 1985. A chemical equilibrium model for the Illinois Basin formation waters. American Journal of Science, 285,436-458. PHILLIPS, S.L., IGBENE,A., FAIR, J.A., OZBEC~:,H. & TAVANA, M. 1981. A technical databook for geothermal energy utilization. Lawrence Berkeley Laboratory Report 12810, University of California, Berkeley. PITZER, K.S. 1987. Thermodynamic model for aqueous solutions of liquid-like density. Reviews in Mineralogy, 17, 97-142. POSEY, H. & KYLE, R. 1988. Fluid-rock interactions in the salt dome environment: an introduction and review. Chemical Geology, 74, 1-24. PRICE, L.C. 1976. Aqueous solubility of petroleum as applied to its origin and primary migration.

American Association of Petroleum Geologists Bulletin, 60, 213-244. RICHARDSON, G.B. 1917. Note on Appalachian oil field brines. Economic Geology, 12, 37-41. RI~ENIJOtJSE, G. 1967. Bromide in oilfield waters and its use in determining possibilities of origins of these waters. American Association of Petroleum Geologists Bulletin, 51, 2430-2440. RUSSELL, W.L. 1933. Subsurface concentration of chloride brines. American Association of Petroleum Geologists Bulletin, 17, 1213-1228. SANFORt), W.E. & WOOD, W.W. 1991. Brine evolution and mineral deposition in hydrologically open evaporite basins. American Journal of Science, 291,687-710. SAWKINS, F.J. 1968. The significance of Na/K and C1/SO4 ratios in fluid inclusions and subsurface waters with respect to Mississippi Valley type ore deposits. Economic Geology, 63,935-942. SHEPPARD,S.M.F. 1986. Characterization and isotopic variations in natural waters. Reviews in Mineralogy, 16, 165-180. SHOCK, E.L. 1988. Organic acid metastability in sedimentary basins. Geology, 16,886-890. SMITh, J.T. & EHRENBERG,S.N. 1989. Correlation of carbon dioxide abundance with temperature in clastic hydrocarbon reservoirs: relationship to inorganic chemical equilibrium. Marine and Petroleum Geology, 6,129-135. STUEBER, A.M. & WALTER,L.M. 1991. Origin and chemical evolution of formation waters from Silurian-Devonian strata in the Illinois Basin, USA. Geochimica et Cosmochimica Acta, 55, 309-325. SURDAM, R.C., BOESE, S. & CROSSLY, L.J. 1984. The chemistry of secondary porosity. American Association of Petroleum Geologists Memoirs, 37, 127-149. SVERJENSKY,D.A. 1984. Oil field brines as ore-forming solutions. Economic Geology, 79, 23-37. -1986. Genesis of Mississippi Valley Type ore deposits. Annual Reviews of Earth and Planetary Sciences, 14, 177-199. WASHBURNE,C.W. 1914. Chlorides in oil-field waters. American Institute of Mining Engineers Transactions, 48,687-693. WHITE, D.H. 1965. Saline waters in sedimentary rocks. American Association of Petroleum Geologists Memoirs, 4,342-366. WILSON,T.P. & LONG, D.T. 1992. Geochemistry and isotope chemistry of Michigan Basin brines: Devonian formations. Applied Geochemistry, 7 81-100.

Organic ligand distribution and speciation in sedimentary basin brines, diagenetic fluids and related ore solutions THOMAS

H. G I O R D A N O

1 & Y O U S I F K. K H A R A K A 2

I Department o f Geological Sciences, New Mexico State University, Box 3AB, Las Cruces, New Mexico 88003, USA 2 United States Geological Survey, 345 Middlefield Road, Menlo Park, California 94025, USA Abstract: The nature, distribution, and interactions of dissolved organic ligands in the shallower zones of sedimentary basins are highly variable and poorly understood. Better understood is the chemistry of dissolved aqueous carboxylic acid species in the deeper regions of sedimentary basins. In most oil-field brines, acetate is generally the dominant organic ligand and is followed in level of concentration by longer chained aliphatic monocarboxylicacid anions. The total concentration of dicarboxylic acid anions is probably lower than previously reported and likely less than 500 mg I-1 with succinate and glutarate being the dominant species. The highest concentrations of organic acid anions are present in formation waters at 80--120° C. Based on the available field and laboratory evidence, the concentrations of organic anions are controlled mainly by the rates of their generation from kerogen and their destruction thermally or by bacteria. An extensive compilation of experimentally determined stability constants and other thermodynamic data for a large number of non-humic organic ligands is available for temperatures near 25° C. However, corresponding high-temperature data are available for only a few species of interest; most notably, acetate complexes of some rock- and ore-forming metals and protonated species of several carboxylate ligands. Simulations of speciation in oil-field brines show that significant amounts of Ca, Mg, Fe, and Al can be complexed by carboxylate ligands, in the pH range 4-6, if ligand concentrations are on the order of 10-100 mg l-I. Calculated speciation in model ore fluids for Mississippi Valley-type deposits and red-bed related base metal deposits show that optimum conditions for Pb and Zn transport by organic ligand complexation are oxidized ore fluids with total reduced inorganic sulphide concentration less than 10 -9 molal.

The accumulation and alteration of organic matter within sedimentary basins and subsequent hydrocarbon generation are well documented processes (Tissot & Welte 1984). Less well understood are the roles of organic matter in the basinal processes of diagenesis and ore formation. The diverse interactions of organic matter in sedimentary diagenesis are discussed in some detail in two recent works (Gautier et al. 1985; Gautier 1986), while specific types of diagenetic interactions involving dissolved organic acids have been succinctly outlined by Kharaka et al. (1985) and Surdam & Crossey (1987). Dissolved organic acids can play an important role in sedimentary diagenesis because of the following chemical properties (Kharaka et al. 1985): (1) they can be the dominant source or sink for protons (H+); and thus, directly or indirectly control the pH and buffer capacity of subsurface waters; (2) as organic ligands, they form stable aqueous complexes with metals and other inorganic

species, thus enhancing their transport and modifying their behavior in diagenetic reactions; (3) they behave as reducing agents controlling the oxidation potential (Eh, fo2) of subsurface waters and the concentration of multivalent elements; and (4) they can be decarboxylated, thermally or with the aid of bacteria, to form carbon dioxide and hydrocarbon gases. The current interest in organic matter as an active chemical agent in ore-forming processes is based primarily on the presence of condensed organic compounds in a variety of sedimenthosted ore deposits formed at temperatures near or below 250 ° C (Giordano 1993; Parnell et al. 1993) and the inadequacy, in many cases, of inorganic mechanisms alone to satisfactorily account for transport and deposition of these ores. Most of these deposits are hosted by cratonic sedimentary rocks and were formed by hydrochemical processes within sedimentary basins. In Table 1 are listed five major ore deposit types which typically contain visible

From PARNELL,J. (ed.), 1994, Geofluids: Origin, Migrationand Evolution of Fluids in Sedimentary Basins, Geological Society Special Publication No. 78,175-202.

175

176

T.H. GIORDANO & Y.K. KHARAKA

Table 1. Major types of hydrochemical ore deposits containing visible organic matter and hosted by cratonic basinal sedimentary rocks (from Giordano 1993)

Ore deposit type

Tectonic environment

Host rocks

Syngenetic chemical

Intracratonic basins

Sandstone uranium

Intracratonic basins

Red-bed copper

Intracratonic basins

Mississippi Valleytype

Intracratonic basins

Sediment-hosted base metal

Incipient rift basins

Near shore marine: siliciclastic Continental: siliciclastic Continental and marginal marine: siliciclastic Shallowmarine and shoreline: carbonate and siliciclastic Shallow marine: carbonate and siliciclastic

Approx. deposition temperature (o C)

Ref.

25

1

50

2

50-100

3

75-200

4

100-275

3

( 1) Vine & Tourtel ot 1970; Tourtelot 1979; Coveney et al. 1987; Gra uch & H uyck 1990; Ripl ey et al. 1990; Schultz 1991; (2) Nash et al. 1981; Turner-Peterson & Fishman 1986; Turner-Peterson et al. 1986; Maynard 1991a; Hansley & Spirakis 1992; (3) Gustafson & Williams 1981 ; Boyle et al. 1989; Heroux et al. 1989; Landais & Meyer 1989; Ho etal. 1990; Maynard 1991b; (4) Macqueen & Powel11983; Gize 1984; Macqueen 1986; Sverjensky 1986; Price & Kyle 1986; Gize & Barnes 1987; Leventhal 1990; Henry et al. 1992. organic matter and that are unequivocally linked to sedimentary basins and related diagenetic processes. Specific roles of organic matter in ore-forming processes are similar to those for diagenesis and have been recently reviewed and outlined by Saxby (1976), Giordano (1985, 1993), Leventhal (1986), and Manning (1986). Giordano (1993) divides the potential roles of non-living organic matter in ore-forming processes into five categories: (1) aqueous complexation, (2) non-aqueous complexation, (3) substrate for microbial processes, (4) reducing or oxidizing agent, and (5) modifying the pH and other physicochemical parameters of the environment. Dissolved organic acids in ore fluids are involved to a greater or lesser extent in all five of the above processes. In this contribution, our main focus is on the aqueous complexing and protonation behavior of organic acid-ion ligands in sedimentary diagenetic fluids and related ore solutions. Organic matter may function up to magmatic temperatures as a reductant or oxidant. However, most of the roles for organic matter outlined above, including complexation and protonation, are limited to less than about 200°C (Barton 1982). The observational evidence discussed in Tissot & Welte (1984) and Giordano (1993) strongly suggests that organic matter is not a significant active agent in diagenetic and ore-forming processes operating near or above 350 + 50° C. For this reason, as well as thermal stability evidence discussed below, it is not likely that organic ligands

significantly contribute to metal and proton speciation in basinal fluids with temperatures greater than about 250 ° C. On the other hand, there is usually sufficient and often compelling evidence suggesting that organic matter is an active agent during normal diagenetic processes and in the genesis of those deposit-types listed in Table 1, as well as other low temperature (<50°C) and moderate temperature (50250°C) deposits. In many cases, inorganic mechanisms alone are insufficient to satisfactorily describe the chemistry in these lower temperature systems. In the following sections we will focus on basinal diagenetic environments and those ore-forming environments developed within or peripheral to cratonic sedimentary basins. The material covered in this contribution is divided into four major topics. First, to clarify the meaning of the terms ligand and complex, a brief discussion of coordination compounds is presented. This is followed by a rather detailed status report on current knowledge about organic ligands and metal-organic complexes of possible importance in diagenetic and oreforming processes. We then discuss the thermodynamic data base for protonation and complexation equilibria involving the organic ligands of interest and, finally, we present several chemical models of diagenetic and ore-forming fluids. The results of these models, although provisional, clearly establish the relative importance of specific organic ligands in these solutions and shed some light on the

ORGANIC LIGANDS IN SUBSURFACE WATERS

177

LIGAND SPECIES Humic and Fulvic Acids

OH

Tetrapyrroles

A m i n o Acids

OH

H O N O R

CZH$

(glycine)

H --C--C -- OH I NH2

o~C"OH...... C'OH I~ O ,H O | °)C~cNzo ~c ~ c ~ C~ .oOH)nc~:* OH

H

CIHI

CH$~"~CH$

OH

II

-- c -- c --c I

CH~

C H $ ~

(alanine)

CzH5

I

NH 2

(metal-porphyrin complex)

~.,? ?,.

%o

(cysteine)

H--C--C--C

(partial structure)

Carboxylic Acids

H NH2 0

Thiols (Ivlercaptans)

Hydroxy Aromatic C o m p o u n d s

~ H--C--C--

(acetic acid)

OH

I H

~ Ho©

H-

C'~OH

C - - C --

(oxalic acid)

OH

, s.

(mercaptoacetic acid)

d/

~o

OH

H

OH

(salicylic acid) SH H

I

I

OH (malonic acid)

\c_ lc_c /

OH

/

H--C--C--C (catechol)

HO

(3-mercaptopropanoic acid)

OH

\

H

H

I

I

c--c--c--c

OH

/

(succinic acid)

Fig. 1. Idealized structures for selected organic ligand species found in sedimentary environments (from Giordano 1993). processes of metal transport, pH control, and porosity development.

Coordination compounds Coordination compounds are substances in which a central atom is bonded to surrounding atoms or groups of atoms called ligands. Two types of coordination compounds involving central metal atoms and organic ligands are recognized: organometallic compounds and metal-organic complexes. Organometallic compounds are species in which a central metal atom is bonded covalently to at least one carbon atom of a surrounding organic molecule (e.g., tetraethyllead and methylmercury). Although most organometallic compounds are artificially produced and are not found in nature, some are known to be synthesized biogenically and a few may be produced abiogenically in natural systems (Stumm & Morgan 1981; Gill & Bruland 1990). The lack of evidence for the presence of dissolved organometallic compounds in natural aqueous systems other than rare occurrences in surface water and shallow groundwater strongly suggests that these metal-bearing coordination

compounds are not involved in metal-transport processes except perhaps in the fixation and accumulation of metals in organisms (Trudinger 1976). Metal-organic complexes are structures in which a metal cation is attached to one or more organic ligand by direct bonding to electrondonor atoms other than carbon, most commonly oxygen, sulphur and nitrogen (Langmuir 1979). If the ligand does not replace water molecules in the first hydration sphere of the metal cation, a weak outer-sphere complex (ion-pair) is formed. If a water molecule in the first hydration sphere is replaced by the organic ligand, a strong inner-sphere complex is formed. Strong complexes called chelate compounds are formed if two or more donor atoms from the same ligand bind the metal cation. Non-metal cations such as the ammonium ion (NH4+) can form complexes analogous to metal-organic complexes. Unlike organometallic compounds, metal-organic complexes are known to be present in a wide variety of geochemical environments, including soil solution, surface water, and groundwater (Thurman 1985; Aiken et al. 1985; Buffle 1988); oil-field brines (Lundegard & Kharaka 1990; MacGowan & Surdam 1990a); and ore fluids

178

T.H. GIORDANO & Y.K. KHARAKA

(Giordano & Barnes 1981; Giordano 1985; Drummond & Palmer 1986; Hennet et al. 1988a; Giordano 1990, 1993). Idealized structures of selected ligand species are illustrated in Fig. 1 for many of the geochemically important organic ligands considered below.

have low thermal stabilities under hydrothermal conditions (Schnitzer & Khan 1978; Aiken et al. 1985; Boles et al. 1988), they are not likely to be important complexing agents in ore fluids and diagenetic fluids having temperatures much greater than 50° C.

Nature of important organic ligands Amino Humic and fulvic acids

Humic and fulvic acids (Fig. 1) are the dominant dissolved organic constituents in surface waters, shallow subsurface waters derived by infiltration, and interstitial waters in young subaqueous sediments (Nissenbaum & Swaine 1976; Reuter & Perdue 1977; Thurman 1985; Aiken et al. 1985; Rashid 1985). These highly oxidized polymers occur in low temperature (less than about 50° C) fresh waters at concentrations up to about 100 mg 1-1 (10 -4 molal based on average molecular weight of about 1000) and in sea water typically to about 1 mg 1-I (10 -6 molal). Single molecules of these polyelectrolyte ligands have molecular weights typically between 500 and 5000 and contain a number of oxygen-bearing functional groups (especially, carboxyl and phenolic hydroxyl groups) capable of bonding with metal cations. Thus, humic and fulvic acids form strong metal-organic complexes (chelate compounds) with most polyvalent metal cations (Reuter & Perdue 1977; Schnitzer & Khan 1978; Jackson et al. 1978; Stevenson 1983; Aiken et al. 1985). Metalhumate and metal-fulvate complexes probably contribute significantly to metal transport and speciation in interstitial waters of subaqueous sediments and in shallow sediments (less than about 1000 m) undergoing early diagenesis. It is, therefore, likely that these complexes are involved in supergene mobilization of metals and ore-forming processes responsible for syngenetic fixation of metals in young sediment as well as epigenetic deposition of metals from low temperature (less than about 50° C) ore fluids to depths of about 1000 meters. Aqueous humate and fulvate complexes of gold (Baker 1978; Varshal et al. 1984; Severson et al. 1986; Vlassopoulous et al. 1990; Coel et al. 1991), platinum (Wood 1990), palladium (Wood 1991), and the metals Cu, Zn, A1, Fe (Kribek et al. 1977) are thought to be important in supergene environments. Humate complexes are also thought to be important in the genesis of certain sandstone uranium deposits (Turner-Peterson et al. 1986) and may possibly transport metals in ore fluids responsible for red-bed copper and related deposits. Because humic and fulvic acids

acids

Amino acids (Fig. 1) are a ubiquitous but minor dissolved organic constituent in surface waters, shallow groundwaters and interstitial waters of sediments (Reuter & Perdue 1977; Jackson et al. 1978; Thurman 1985). Within the water column and during early diagenesis, amino acids are subject to intense microbial degradation and, therefore, their concentrations are normally low compared to the relatively inert humic and fulvic acids residing in the same environment. Typical maximum concentrations of total free amino acids in fresh water and marine water are approximately 1 mg1-1 (10 -5 molal) and 0.1 mg 1-1 (10 -6 molal), respectively, about an order of magnitude lower than concentrations of dissolved humic substances. Degens et al. (1964) and Rapp (1976) report concentrations of total amino acids of up to about 0.3 mg 1-1 (7 × 10-6 molal) in oil-field brines, with glycine, alanine, and serine typically the dominant species. These ligands, as well as other naturally occurring aqueous amino acids (e.g., aspartic acid, leucine, glutamic acid, cystine, and cysteine), form strong complexes with most polyvalent metal cations except the alkaline earth cations (Martell & Smith 1974). Because ligand concentrations of amino acids are typically an order of magnitude lower than concentrations of humic and fulvic acids in interstitial waters, metal-amino acid complexes are probably less important than humate or fulvate complexes in most low temperature diagenetic and ore-forming environments. However, if concentrations of amino acids are sufficiently high in specific environments, amino acid complexation may play an important role in metal transport and deposition (Veitch & McLeroy 1972; Saxby 1976). Although theoretical calculations by Shock (1990) show that amino acids may survive metastably to elevated temperatures, observational evidence from the field (Degens et al. 1964; Rapp 1976; Haberstroh & Karl 1989) and from degradation experiments (Bernhardt et al. 1984; White 1984; Miller & Bada 1988; Bada et al. 1991) strongly suggests that amino acids are probably not present in natural waters above 100°C at concentration levels high enough to significantly affect metal speciation.

ORGANIC LIGANDS IN SUBSURFACE WATERS

Organosulphur ligands Organosulphur compounds (Fig. 1) containing reduced sulphur are present as minor constituents in marine sediments (Tissot & Welte 1984; Vairavamurthy & Mopper 1989; Kiene & Taylor 1989) but concentrations of specific species dissolved in interstitial waters are not widely documented. Interestingly, the classes of organosulphur compounds found in young marine sediments are the same as those found in petroleum and include thiols (mercaptans), sulphides, disulphides, and thiophene derivatives. To our knowledge, specific organosulphur compounds have not been identified in oil-field brines, but it is likely that such compounds partition into the interstitial aqueous phase of carbonaceous rocks during diagenesis (Tissot & Welte 1984). Kharaka et al. (1979) measured sulphide concentrations in oil-field brines from the Gulf coast by two methods and suggested the discrepancy in results may be due in part to organic sulphur species. Possibly important organosulphur ligands dissolved in interstitial waters in young sediments include sulphurbearing amino acids (e.g., cystine and cysteine) and mercaptocarboxylic acids. Mercaptocarboxylate ligands and similar thiol compounds are found in crude oil (Tissot & Welte 1984) and are probably also released to deep, hot formation waters during the initial stages of kerogen degradation. Mercaptocarboxylate ligands are known to form highly stable complexes with most ore-forming and rock-forming metal cations, but only weak complexes with cations of the alkaline earth metals (Martell & Smith 1977). Although not strictly an organosulphur compound, the thiocarbonate ligand CO2S2- may be an important complexing agent of Pt, Pd, Zn, and Ni in moderate temperature (50-250 ° C) hydrothermal ore fluids (Hennet et al. 1988b). An important advantage of organosulphur ligands over other organic complexing agents is the ability to transport both metals and reduced sulphur in the same fluid (Saxby 1976). This gives rise to an attractive version of the single ore-fluid hypothesis. Metals and sulphide are transported together in one solution as soluble metal-organic sulphide complexes (e.g., metalmercaptocarboxylate complexes). Metals and sulphur could be transported in sufficient amounts to form economic deposits if concentrations of such complexes were greater than about 10 -5 molal. Subsequent breakdown of these complexes at the site of deposition could yield the necessary metals and sulphide to form

179

galena, sphalerite, and other minerals (Saxby 1976; Barnes 1983; Gize & Barnes 1989). The evidence reviewed above suggests that cystine, cysteine, mercaptocarboxylic acids, and other organosulphur ligands may contribute to the speciation of metals and sulphur in diagenetic processes responsible for syngenetic deposition of metal sulphides in low-temperature subaqueous sediments. To our knowledge, the thermal stability of thiols, organic sulphides and thiophene derivatives have not been determined under a wide range of hydrothermal conditions. Nevertheless, these compounds are found in petroleum, which is typically generated between 80 and 120°C (Tissot & Welte 1984) and, therefore, it is possible that they have a similar thermal stability under hydrothermal conditions. If they are sufficiently stable at elevated temperatures, thiols and other organosulphur ligands could conceivably contribute to both metal and sulphur mobilization in moderate-temperature (50-250 °C) hydrothermal fluids.

Tetrapyrrole ligands Porphyrins and related tetrapyrrole ligands (Fig. 1) form metal-organic complexes in which a central metal cation is bonded to four nitrogen atoms of an aromatic tetrapyrrole structure (Lewan & Maynard 1982). Although such complexes are thermally stable and moderately inert to temperatures above 350 ° C (Saxby 1976; Tissot & Welte 1984; Lewan 1984), they are strongly partitioned into condensed phases of organic matter relative to any coexisting aqueous phase. Thus, with increasing maturity of sediment-hosted organic matter, metal tetrapyrrole complexes are concentrated in humic substances, kerogen, and finally bitumens, including expelled petroleum phases, which subsequent to generation in a source rock can migrate to sites of ore deposition as well as petroleum reservoirs. Tetrapyrrole ligands form strong complexes with nickel (Ni 2+) and vanadium (VO 2+) and it is now thought that such complexes are responsible for the fixation of nickel, vanadium and perhaps other metals in young sediments during early diagenesis (Lewan & Maynard 1982). The amounts of individual metals complexed by tetrapyrrole ligands during diagenesis depend strongly on the composition of the condensed organic matter and minerals in the sediment, as well as the chemistry of the coexisting interstitial water, most importantly pH, Eh, and the concentrations of competing ligands and reduced inorganic sulphur (Lewan 1984; Lipiner et al. 1988; Kotova et al. 1987).

180

T.H. GIORDANO & Y.K. KHARAKA

It has been suggested that petroleum, containing dissolved tetrapyrrole complexes of V, Ni, Co, Fe, Cu, Zn, and Pb, should be considered a potentially important ore-transporting agent (Manning 1986; Price & Kyle 1986). In a multiphase system comprising a mobile petroleum phase and an aqueous ore fluid, metals may be released to the ore fluid along a flow path or at the site of deposition by thermal, oxidative, or biological degradation of metal-tetrapyrrole complexes or by favourable partitioning of metals from the petroleum phase (containing metals in tetrapyrrole complexes) to the aqueous phase (containing metals in organic and inorganic complexes) as a result of changes in ore fluid chemistry (Manning 1986; Hennet et al. 1988b). Gize & Barnes (1989) suggested that metalloporphyrin complexes may be the source of nickel and cobalt found in Mississippi Valleytype deposits. To properly evaluate the affects of metal-tetrapyrrole complexation on metal transport and deposition, a better understanding is required of (1) the stability and concentration of such complexes in humic substances, kerogen, and oil and (2) the nature of metal cation partitioning between tetrapyrrole complexes in these phases and complexes in a coexisting aqueous phase.

Carboxylic acid anions Individual carboxylic acids (Fig. 1) and their corresponding acid anions (carboxylate ions) are found in surface and shallow subsurface waters where they always form a minor component of the dissolved organic matter (Thurman 1985). The concentration of carboxylate ions in surface and shallow subsurface interstitial waters are sufficiently low to preclude the importance of carboxylate species as significant agents of pH control and metal transport during early diagenesis and in low-temperature ore fluids responsible for supergene mineralization, syngenetic deposition in anoxic environments, and low temperature epigenetic ores such as sandstone uranium deposits. However, the observed elevated concentration of organic acids in oil-field brines, as discussed below, strongly suggest that metal-carboxylate complexes may contribute significantly to metal transport and pH control in diagenetic fluids and ore fluids with temperatures ranging from roughly 50 to 150°C. In deeper formation waters, carboxylate species are the dominant dissolved organic constituents and in some deep subsurface waters carboxylate anions are present at sufficient concentrations to dominate the total alkalinity (Kharaka et al. 1986).

Carboxylate ions form metal--organic complexes of moderate strength with most polyvalent metal cations (Martell & Smith 1977) and if present in formation waters at concentrations near or above 10-4 molal, such ligands can be important complexing agents for metals in carbonaceous source rocks and basinal aquifers.

Distribution, origin, and interactions of organic ligands in subsurface waters Organic species in shallow subsurface waters The concentration of dissolved organic carbon in uncontaminated shallow ground waters ranges from 0.2 to 15 mg1-1, with a median value of 0.7mg1-1 (Leenheer et al. 1974; Thurman 1985). The high molecular weight humic and fulvic acids are the dominant organic species in these waters comprising 20-40% and up to 90% of this dissolved carbon. The bulk of the remaining carbon comprises volatile and nonvolatile fatty acids, phenols, amino acids, peptides and carbohydrates (Thurman 1985). The polymeric fuivic and humic acids as well as the more simple components of dissolved carbon contain reactive functional groups with oxygen (especially carboxylic and phenolic groups) and nitrogen (amine and amide groups) that can form strong metal-organic complexes and chelates (Schnitzer & Khan 1978; Aiken etal. 1985). Thus, as mentioned above, these organic species are likely to play some role in the mobilization and precipitation of metals as well as other water-rock interactions in the shallow subsurface environment. The concentration of dissolved organic carbon in pore waters of young marine and lake sediments generally is much higher than that of ground water and may reach values of about 400 mg 1-~ in anaerobic sediments. Fulvic and humic acids are generally the most abundant components in these waters, but volatile fatty acids, especially acetic, formic and butyric may predominate especially with increasing depth of burial (Barcelona et al. 1980; Thurman 1985). These organic species play an important role during early diagenesis in all the zones of buried young sediment, especially in the speciation, mobilization, and transport of metals.

Organic species in oil-field waters The concentrations of organic species in formation waters associated with petroleum and obtained from reservoirs at temperatures of

ORGANIC LIGANDS IN SUBSURFACE WATERS about 20-200 ° C are generally much higher than in other natural waters (Lochte et al. 1949; Kartsev 1974; Lundegard & Kharaka 1990). The origin, distribution, and interactions of organic acid anions in these waters have become an intensively studied field in geochemistry since their widespread occurrence in formation waters of many sedimentary basins was first documented by Carothers & Kharaka (1978). These organic species are present in waters as anions, acids and complexes, but are generally referred to as organic acid anions because anions generally predominate in oil-field waters which have p H values of about 5-7 (Kharaka et al. 1985). High concentrations (up to 10000mg1-1 as acetate) of monocarboxylic (mainly acetate, propionate and butyrate) and dicarboxylic (mainly oxalate, malonate and succinate) acid anions are reported from many sedimentary basins worldwide, with the highest values present in relatively young (Cenozoic age) petroleum reservoir rocks at subsurface temperatures of 80-120°C (Zinger & Kravchick 1973; Kartsev 1974; Germanov & Mel'kanovitskaya 1975; Kharaka et ai. 1986; H a n o r & Workman 1986; Lundegard & Land 1986; Means & Hubbard 1987; Fisher & Boles 1990; MacGowan & Surdam 1990b; Barth 1991). Geochemical interest in these organic species stems mainly from their important role in mineral diagenesis in sedimentary basins (Willey et al. 1975; Crossey et al. 1986; Kharaka et al. 1986; MacGowan & Surdam 1990a). In particular, these species act as sources or sinks of protons, as a source of CO2 and as p H and Eh buffering agents (Carothers & Kharaka 1980; Lundegard & Kharaka 1990). They also form complexes with cations and metals such as Ca, A1, Fe, Pb, and Zn (Giordano & Barnes 1981; Giordano 1985; Kharaka et al. 1985; Fein 1991 a; Harrison & Thyne 1992). Organic acid anions have been studied for use as proximity indicators in petroleum exploration (Kartsev 1974; Carothers & Kharaka 1978), as indicators of fluid flow (Hanor 1987) and as possible precursors for a substantial portion of natural gas (Kharaka et al. 1983; D r u m m o n d & Palmer 1986). Field m e t h o d s The presence of organic acid anions in oil-field waters can be detected by routine field procedures described by Carothers & Kharaka (1978). These procedures include acidification of the samples for cation analysis to pH of about 2.0; lack of significant effervescence could indicate the presence of organic anions, especially in the case of samples with alkalinities >1000 mg 1-2 (as bicarbonate). Also, the acidified samples with organic anions generally give the characteristic rancid odor of butyric acid. The most definitive test in the field, however, is obtained by

2~

I

I 20

I

I 40

181

i

l 60

l

I 80

I

I00

VOLUME OF 0.050 N H2SO4 ADDED Fig. 2. Titration curves for sodium bicarbonate,

sodium acetate and an oil field sample. Note that the inflection point for oil field water is at pH ~ 3 which is similar to that of acetate but different from that of bicarbonate (from Carothers & Kharaka 1978).

examining the inflection points of the alkalinity titrations (Fig. 2). Samples with organic acid anions have an inflection point near pH 3.0 as compared to values near pH 4.5 for samples with no organic anions. The volume of acid titrant added between pH 4.5 and the inflection point of the sample gives some indication of organic alkalinity and of the concentrations of organic acid anions. Organic alkalinity, it is worth noting here, generally is much higher than bicarbonate alkalinity in some formation waters, especially those with temperatures from 80° to 120° C (Carothers & Kharaka 1978; Lundegard & Kharaka 1990). Water samples for analysis of dissolved organic species require special handling and treatment in the field (Lico el al. 1982; Kharaka etal. 1985). In order to minimize contamination, only glass, metal (e.g. stainless steel, copper) or teflon containers and tubing should be used for collection, filtration and storage of these samples. To reduce the loss of organic species by volatilization, agitation or aeration of water prior to analysis should be minimized, especially in samples with pH values lower than about 5.0. Bacterial degradation, however, is of greater concern and is likely to affect these samples unless preventive measures are taken (Kharaka et al. 1985; MacGowan & Surdam 1988). The measures should include filtering the sample through 0.1 p~m (or at least 0.45 ~m) teflon- or silver-filter paper and storage in sterilized amber glass bottles. A bactericide such as mercuric chloride (40 mg1-1 Hg), sodium azide or zephrin chloride should be added to the bottle before storage in a refrigerator at temperatures of about

182

T.H. GIORDANO & Y.K. KHARAKA

10,000 f

,

'

',I

~

r

r

I

'

r

,

-T"'

J 5000 i-

d ;ll~ \

Zone 1

Zone 2

•

-I

.J E .~ 1000-

~

soo

!

t

o

.~_ ~ 100

°° o~

5O

~.~

!

104 "ll'l

EXPLANATION ,

© • A ÷

Texas (Oligocene) California (Miocene to Cretaceous) Alaska (Permian to Triassic) Texas (Pleistocene)

I'I'L 60

i ~ ~

160 200 Subsurface

Temperature

220

(°C)

Fig. 3. Distribution of organic acid anions in oil-field waters from USA. Note that the highest concentrations are in waters from reservoirs at temperatures of 80-100°C (from Kharaka et al. 1988).

4°C. The samples should be analysed as soon as possible after returning to the laboratory. M o n o c a r b o x l i c acid a n i o n s Data on the concentration of monocarboxlic acid anions in formation waters from many sedimentary basins worldwide were reviewed by Lundegard & Kharaka (1990) to investigate the major controls on the distribution of these species. The samples were obtained from reservoir rocks ranging in age from Pennsylvanian to Pleistocene, although most were from clastic reservoirs of Cenozoic age with temperatures from 50° to 160° C. The main conclusions from this review were: (1) concentrations of acid anions in these waters are highly variable with respect to the major known controls, including subsurface temperature, age of the reservoir rock and the type and amount of kerogen in source rock; (2) the highest concentration of acetate reported is by MacGowan & Surdam (1988) at 10000 mg 1-1 (0.17 molal), but values higher than 5000 mg 1-1 are very rare; (3) maximum concentrations vary with temperature and are generally present in waters at 80°-120°C; (4) at similar temperatures, concentrations are generally lower in waters from older reservoir rocks. While variations in source material influence organic acid

anion concentrations, the main control on these concentrations appears to be the kinetics of decarboxylation (Lundegard & Kharaka 1990). The three temperature zones established by Carothers & Kharaka (1978) appear still useful in understanding the geochemistry of organic anions (Fig. 3), although some modifications are warranted, especially if data from other brines are included (Lundegard & Kharaka 1990). The concentrations of organic acid anions in zone 1 (temperature <80 ° C) are low, generally <100mg1-1, but may increase to about 500 mg 1-I as the higher temperature limit of this zone is approached (Kharaka et al. 1986; Fisher 1987). The highest concentrations are present in waters at the lower temperatures of zone 2 (80°-100° C) and decrease with increasing temperatures reaching the value of zero at about 220 ° C, the low temperature limit of zone 3. Water samples from two geothermal wells with temperatures >250 ° C at Salton Sea, California showed no detectable organic acids, confirming their absence in the waters of zone 3. The concentrations of monocarboxylic acid anions in selected formation water samples from sedimentary basins in North America are listed in Table 2. In the waters of zone 2, acetate is by far the most abundant organic acid anion comprising approximately 80-90% of the total; the abundance of other acid anions generally decrease with increasing carbon number for an overall order: acetate > > propionate > butyrate > valerate (Carothers & Kharaka 1978; Lundegard & Kharaka 1990). Formate appears to be either absent (Carothers & Kharaka 1978) or to be present at very low concentrations in these waters (MacGowan & Surdam 1988; Fisher & Boles 1990). The relative abundance of acetate in zone 1 is highly variable with values that range from 0 to 90% of the total organic anions, with propionate being the dominant anion in waters with low acetate. The generally low concentrations of organic acid anions and the low relative abundance of acetate in the waters of zone 1 are generally attributed to the bacterial consumption of these anions, especially the acetate (Carothers & Kharaka 1978). Bacteria are known to survive at temperatures of up to about 130°C (Ehrlich 1990), but their operation is likely limited to about 80°C in the subsurface (Davis 1967; Carothers & Kharaka 1980). In a few petroleum fields the low total acid anions in formation waters from zone 1, which have anion proportions similar to zone 2, may result from dilution of waters high in anions that originated and flowed upward from zone 2 and mixed with shallow and depleted waters of zone 1 (Kharaka

50400

1186 88 9 17 bd bd 1300

42800

Acetate Propionate i-Butyrate n-Butyrate i-Valerate n-Valerate Total organic anions TDS 10000

271 35 13 5 5 bd 330

Lower McAdams Eocene 3520 141

Kettleman North Dome, CA 323-21J

912-2"

25200

bd 16 2 6 bd bd 24

Miocene 675 56

Olcese

113-28

Wheeler Ridge, CA

74WR-2"

14900

bd 11 1 3 4 5 24

Miocene 507 50

Valve

434-8

Wheeler Ridge, CA

74WR-6"

10900

4100 336 131 157 110 30 4860

Miocene 3337 149

Reef Ridge

21-15

San Emidio Nose, CA

74SEN-3"

157000

264 31 4 3 na na 302

Cretaceous 1965 68

Stanley

McCullough et al. Gas Unit 1

Soso, NI

84-MS-3 *

Pleistocene 3719 98 1525 226 26 34 19 12 1842 91000

Pleistocene 1876 55 12.4 1.9 0.0 0.0 11 0.0 25 80000

Jurassic 4350 121 15 1 2 0.0

18 333000

* From Carothers & Kharaka (1978) * From Kharaka et al. (1987). * From Kharaka et al. (1986).

na

na

Sand Sand

A-45-1

A-11-B

Jessie Allen No. 1-N

Norphlet

> Z

High Island, TX

High Island, TX

E. Nancy, NI

>

>

Z

> Z

t"

©

83-TX-12'

83-TX-2'

83-MS-15'

bd, below detection limit; na is not analysed; TDS is calculated inorganic and organic total dissolved solids; TR is trace detected.

195 TR TR TR bd bd 195

Oligocene 2624 94

Lower Weiting Oligocene 3960 154

Evans 'A' no. 1

Age Depth (m) Temp. (o C)

Bernard No. 6

Well name

Alto Loma, TX

Frio 'A'

Chocolate Bayou, TX

Field name Location

76-GG-30"

Production zone

76-GG-1 *

Sample No.

Table 2. Concentration (mg l -t) of monocarboxylic acid anions from selected petroleum fields in USA

184

T.H. GIORDANO & Y.K. KHARAKA

et al. 1986; Fisher 1987). The low concentrations may also result from low rates of generation of organic acid anions from the source rock (Lundegard & Kharaka 1990). The decrease in the concentration of organic acid anions with increasing temperature in zone 2 is generally attributed to decarboxylation of these anions (see later section for more details) to CO2 and natural gas (Kharaka et al. 1983; Drummond & Palmer 1986). The overall reaction for acetate is:

CH3COOH ~- CH4 + C02

(1)

The relative abundance of C2-C5 organic acid anions in zone 2 is likely controlled mainly by their rates of generation from precursor molecules within the parent kerogen, because the decarboxylation rates for these anions are rather similar (Palmer & Drummond 1986; Bell 1991). The decarboxylation rate for formate is much higher than that of acetate (Crossey 1991) which can explain the low concentrations of this anion in formation waters. Shock (1988, 1989) computed the stability fields of acetic and propionic acids and concluded that the concentrations of these and other acids in formation waters at temperatures >80 ° C were controlled not by decarboxylation, but by metastable redox reactions involving 02 and CO2. The reaction proposed between acetic and propionic acids was" 2C2HsCOOH + 02 ~ 3CH3COOH

(2)

Using selected values from Carothers & Kharaka (1978), Shock (1989) observed that a plot of aCH3COOHand aC2H5COOHhad a slope of 3/2 and plotted in the equilibrium stability field required by reaction (2). Using more extensive data sets for similar plots, Bell (1991) and Lundegard & Kharaka (1993) were unable to confirm the equilibrium of reaction (2) as they observed variable trends that rarely displayed a slope of 3/2 even for samples from the same sedimentary basin. Equilibrium was also not attained between acetic and propionic acids generated from hydrous pyrolysis of crude oils (Kharaka et al. 1993a). Because oxygen fugacities carry very large uncertainties for experimental as well as field situations, Kharaka et al. (1993a) computed the departure from equilibrium for the reaction: C2H5COOH + H2 ~- CH3COOH + CH4 (3) The possible attainment of stable or metastable equilibrium between organic acid anions and other aqueous and gaseous components of formation waters clearly requires additional

testing. Based on the available evidence, the concentrations of organic acid anions, in our opinion, are controlled mainly by the rates of their generation from kerogen and destruction thermally or by bacteria. Dicarboxylic acid anions In comparison with monocarboxylic anions, data on the concentrations of dicarboxylic acid anions in formation waters from sedimentary basins are much more limited, some reported values are controversial, and the total values reported range widely from 0 to 2640 mg 1-1 (Surdam et al. 1984; Kharaka et al. 1986, 1993b; Barth 1987; MacGowan & Surdam 1988, 1990b; Fisher & Boles 1990). Because of these uncertainties in the reported concentrations and because dicarboxylic acid anions generally form stronger complexes with A1, Fe, and other cations than do monocarboxylic acid anions, this section is covered in some detail. The first reported concentrations of dicarboxylic acid anions in oil-field waters were by Surdam et al. (1984). They reported values of up to 11 mg 1-1 for malonic acid and up to 26 mg 1-1 for maleic acid, but very little geological or geographical information was given. Kharaka et al. (1985, 1986) reported the concentrations of organic and inorganic species in formation waters from sandstone reservoirs of Pleistocene age from 11 petroleum wells in the High Island field, offshore Texas. They reported succinate (up to 63 mg 1-I), glutarate (up to 36 mg 1-1) and trace concentrations (1--6 mg 1-1) of C6 to C10 dicarboxylic acid anions (Table 3). The concentrations of oxalate and malonate in these waters were below the detection limit of about 1 mg 1-1. Methyl succinate and 2-methyl glutarate were identified by the GC-MS analysis employed, but these were not quantified. Detailed inorganic and organic chemical analyses of 20 formation water samples from six petroleum fields in central Mississippi Salt Dome basin were reported by Kharaka et al. (1987). The samples were obtained from sandstones and limestones of Cretaceous and Jurassic age. The total concentrations of monocarboxylic acid anions in these waters are low with less than 302 mg 1-1. Four samples were analysed by GC-MS techniques for oxalate and other (C3-C10) dicarboxylic acid anions; the concentrations obtained were all below the detection limit of about 1 mg 1-1. The concentrations of oxalate reported by Lundegard & Trevena (1990) were also below that detection limit in water samples from oil wells in the Gulf of Thailand. No other dicarboxylic acid anions were determined in these waters which contain

185

ORGANIC LIGANDS IN SUBSURFACE WATERS Table 3. Concentration (mg l-l) of dicarboxylic acids and acetate information waters from High Island field, offshore Texas (Kharaka et al. (1986) Well

Temp. (° C)

Acetate

A-10A A-14B A-45-1

66 77 98

148 222 1525

Sample 83-TX-3 83-TX-10 83-TX-12

Butanedioic Pentanedioic Hexanedioic Heptanedioic Octanedioic Nonanedioic Decanedioic acid acid acid acid acid acid acid 1.8 0.2 63

1.7 1.2 36

0.5 0.5 bd

bd 0.6 0.2

bd 0.7 5.0

bd 1.4 6.0

bd bd 1.3

bd, Below detection limit.

Table 4. Concentration (mg 1-1) of dicarboxylic acid anions and acetate in watersfrom North Coles Levee (NCL) and Paloma (P) fields, California (Kharaka et al. (1993b) Sample 92NCL-10 92NCL- 11 92P-13 92P-14

Well

Field

Production zone

Temp. (° C)

Acetate

34--31 12-31 66X-2 24X-11

NCL NCL P P

Stevens Stevens Stevens Stevens

93 95 ! 16 114

3296 2945 1281 595

abundant monocarboxylic anions (up to 1500mg1-1 acetate) and that were obtained from reservoirs with relatively high temperatures (120-177 ° C). Barth (1987) reported 38 mg 1-1 oxalate and 10 mg 1-~ malonate in one sample of water from North Sea oil wells. Because this sample was nearly devoid of monocarboxylic acid anions and dicarboxylic anions were not detected in other samples, she concluded that the measured dicarboxylic anions were not present in natural formation water, attributing them instead to contamination by organics in drilling mud. The highest concentrations of dicarboxylic acid anions are those reported by MacGowan & Surdam (1988, 1990b) for water samples from about 40 petroleum wells located mainly in the San Joaquin, Santa Maria and northern Gulf of Mexico basins. They reported values of up to 494 mgl -~ oxalate, 2540 mg 1-1 malonate and 66 mg 1-1 maleate. Well 12-31 in the North Coles Levee field, San Joaquin basin, California has the distinction of producing water with the highest reported concentration of dicarboxylic acid anions (2640 mg 1-1) (MacGowan & Surdam 1988). Because of the high organic content, this well and several wells from the North Coles Levee and the nearby Paloma field, were re-sampled by several other investigators (MacGowan & Surdam 1990b; Fisher & Boles 1990; Kharaka et al. 1993b). For samples from well 12-31, MacGowan & Surdam (1990b) reported much lower values for the concentrations of oxalate (17 mg 1-1), malonate (42 mg 1-1) and maleate (below detection limit). Fisher & Boles (1990) detected no oxalate or malonate in the

Methyl Oxalate Malonate Succinate succinate 0.07 0.16 0.10 0.05

<0.05 0.08 <0.05 0.07

<0.02 <0.2 2.5 2.2

14 22 5.5 3.6

waters from this well or other wells sampled in the area. The concentrations of dicarboxylic acid anions from this well and the other three wells sampled by Kharaka et al. (1993b) are very low with methyl succinate values ranging from 3.6--22 mg 1-1 (Table 4). Glutarate concentrations were higher with 95 mg 1-1 (92 NCL-11) and 15mg1-1 (92P-13) in the two samples analysed. Small amounts of succinate (2.2 and 2.5 mg 1-1) were measured in the waters from the Paloma field. Only traces of oxalate (up to 0.2 mg 1-1) and malonate (up to 0.1 mg 1-~) were measured. The concentrations of methyl malonate, maleate, adipate and pimelate were below the detection limit of about 0.5 mg 1-~ We believe that the low concentrations of dicarboxylic acid anions reported by MacGowan & Surdam (1990b), Fisher & Boles (1990) and Kharaka et al. (1993b) are more representative of the concentrations of these organics in formation waters. The concentrations of dicarboxylic acid anions are probably limited by several factors, including a rapid rate of thermal decomposition (MacGowan & Surdam 1988; Crossey 1991) and low solubilities of calcium oxalate and calcium malonate (Table 5; Kharaka et al. 1986; Harrison & Thyne 1992). Other reactive organic species. Data on the concentration and nature of organic species other than the mono- and dicarboxylic acid anions are very few. However, some measured concentrations of dissolved organic carbon in formation waters are higher than those expected from dissolved acid anions in oil-field waters

186

T.H. GIORDANO & Y.K. KHARAKA

Table 5. Solubilities of calcium salts of selected carboxylic acids (from Giordano 1993) Composition Ca(CzH302)2 Ca(C3H5Oz)2"H20 Ca(CsH9Oz)z CAC204 CaC3H204-4H20 CaC4H604.3H20 Ca(CTH502)2"3HzO Ca(CTHsO3)2.2H20

Name

Temperature (o C)

Solubility* (molality)

Ca acetate Ca propionate Ca valerate Ca oxalate Ca malonate Ca succinate Ca benzoate Ca salicylate

100 100 100 96 100 80 80 25

1.5 6.18 1.16 0.0001 0.0228 0.0274 0.2776 0.1288

* Morrison & Boyd (1966), Weast (1977).

suggesting that other species may be present (Kharaka et al. 1986; Fisher 1987). Interpretation of the excess measured concentrations of DOC in oil-field waters, however, are difficult because the excess could be due to entrained oil or to soluble components of oil, including benzene and toluene (Kharaka et al. 1986). Degens et al. (1964) and Rapp (1976) identified several amino acids, including serine, glycine, alanine, and aspartic acid, but the concentrations were low at <0.3 mg 1-1. In formation water from the High Island field, offshore Texas, Kharaka et al. (1986) identified but did not quantify a number of species, including phenol, 2-, 3-, and 4-methylphenol, 2-ethylphenol, 3-, 4-, and 3-, 5- dimethylphenol, cyclohexanone, and 1-, 4-dimethylbenzene. Fisher & Boles (1990) measured the dissolved organic carbon in two formation water samples from the San Joaquin basin by GC-MS analysis of combined acid, base, and neutral methylene chloride extracts. They identified various polar aliphatics (fatty acids to C9 with various methyl and ethyl substituents), cyclics (phenols and benzoic acids), and heterocyclics (quinolines). They were able to quantify, at the ppm or sub-ppm level, phenol, methyl-substituted phenols, and benzoic acid. Lundegard & Kharaka (1993) report data collected by Kharaka in 1987 on formation water from oil and gas wells in the Sacramento Valley, California, giving the following organic species: phenols (up to 20 mg 1-1), 4-methyl phenol (up to 2mgl-1), benzoic acid (up to 5mgl-1), 4-methyl benzoic acid (up to 4mgl-1), 2hydroxybenzoic acid (up to 0.2mgl-a), 3hydroxybenzoic acid (up to 1.2mgl-~), 4hydroxybenzoic acid (up to 0.2mgl-l), and citric acid (up to 4 mg 1-1). Additional dissolved organic species likely will be detected in formation waters as analytical procedures improve.

Origin o f m a j o r reactive species

The low concentrations of humic and fulvic acids, carbohydrates, peptides, amino acids and other organic species in shallow ground waters are derived mainly from the associated organic matter and organisms, but an important portion may originate in surface water or in the soil zone from decomposition of relatively modern organic matter (Thurman 1985). Bacteria play a major role both in the generation of dissolved organic species and their selective modification and destruction in these waters (Ehrlich 1990). The highest concentrations of dissolved organic species are present in formation waters at temperatures of 80-120 ° C, a thermal regime considered above that of normal bacterial action (Davis 1967; Carothers & Kharaka 1980; Lundegard & Kharaka 1990). The most likely source of these acid anions as well as the source for petroleum is thermocatalytic degradation of aliphatic acids incorporated in kerogen (Kvenvolden & Weiser 1967; Hunt 1979; Kharaka et al. 1983). Oxygen functional groups in kerogen, and thus organic acid anions, are lost from kerogen at thermal ranks lower than those of principal hydrocarbon generation (Robin & Rouxhet 1978; Lundegard & Kharaka 1990). The amount of oxygen in the functional groups of kerogen is high (up to 30% in type III kerogen) and more than enough to account for the highest concentrations of organic acid anions reported in formation waters (Tissot & Welte 1984; Lundegard & Kharaka 1990). Laboratory experiments have confirmed the generation of these organics from the hydrous pyrolysis of kerogens and shales (Surdam et al. 1984; Kawamura et al. 1986; Lundegard & Senftle 1987; Barth et al. 1989). Lundegard & Kharaka (1990) were able to model the generation of organic acid anions from kerogen and obtain field distribution profiles similar to those observed for the modeled basins. Hydrous pyrolysis experiments of crude oils generated relatively large concentrations of mono- and dicarboxylic acid anions (Fig. 4) with relative abundances generally similar to those observed in sedimentary basins (Kharaka et al. 1993a). The amounts of reactive organic species that can be generated from crude oils are potentially high because oils, especially the biodegraded and immature ones with low API gravity, can contain relatively high concentrations of low and high molecular weight (C1-C40) normal, aromatic and heterocyclic mono- and dicarboxylic acids and other oxygenbearing compounds (Seifert 1975; Tissot & Welte 1984). Analysis of all of the pertinent

ORGANIC LIGANDS IN SUBSURFACE WATERS

Organic acid anions ~. FORM*10 120 A ACET*0.6 i -- o- - PROP _.100 ? ~ - - o - - OXAL*50 - - []-- SUCC*50 a~ . ' ~-MSUC*50 Eso F \

187

l f

-

O

=

so-

l

"-'~oc:40 ~

,

/

~,~s~

-"

. . . .

-0

" "'~" "-:2 . . . . . . . .

---.-

sc~---

0

20 o

~

0

.

500

~ ~

1000 1500 T i m e (h)

2000

2500

Fig. 4. Aqueous concentrations of selected organic acid anions generated from crude oil from the Midway Sunset field, California at 300°C. Note the relatively low concentrations of dicarboxylic acid anions (from Kharaka et al. 1993a).

data, however, shows that the bulk of organic acid anions in formation waters are most likely generated by thermal alteration of kerogen in source rocks (Kharaka et al. 1993a). This conclusion was reached by Kharaka et al. (1993a) because the oxygen content of oil (0-1 wt%) is about 20 times lower than that of kerogen (Tissot & Welte 1984), the organic acid anion yields per weight are approximately two orders of magnitude lower in oil than in kerogen experiments (Lundegard & Senftle 1987; Barth etal. 1989; Kharaka etal. 1993a), oil is much less abundant than kerogen in sedimentary basins, and high concentrations of organic acid anions have been reported from gas fields where liquid petroleum probably never existed (Lundegard & Kharaka 1990). Under favorable conditions, nonetheless, a significant portion (10-30%) of acid anions in formation waters associated with oil can be liberated from in situ alteration of reservoired oil. Destruction o f organic acid anions

The highest concentrations of organic acid anions are present in formation waters at 80-120 ° C (Fig. 3). This suggests that at higher and lower temperatures these organics are either produced in lower amounts from the source

rocks or that they are destroyed thermally or by bacterial action. Acetic and other organic acids are thermodynamically unstable in the subsurface and ultimately will decarboxylate to CO2 and alkanes (Kharaka et al. 1983; Shock 1988). Decarboxylation rates for acetic acid in the absence of minerals were measured experimentally by Kharaka et al. (1983) and Palmer & Drummond (1986). Results showed that the rates obtained follow a first-order rate law, but were highly variable depending mainly on the catalytic surface effects of the vessels employed. The half-lives calculated at 100 ° C ranged from about 10 years with stainless steel vessels to more than 1014years with catalytically less active gold and titanium vessels. With stainless steel vessels Kharaka et al. (1983) measured much higher decarboxylation rates for acetic acid than for acetate. Decarboxylation rates for acetic acid and acetate in the presence of geologically relevant mineral surfaces have recently been measured by Bell (1991). Results showed that: (1) quartz, amorphous silica and pyrite had minimal catalytic effects on decarboxylation rates; (2) calcite had a small effect giving a half-life at 100°C of 2.5 x 107 a; (3) calcium montmorillonite had a moderate catalytic effect giving a half-life of 2 x 104 a at 100 ° C; and (4) magnetite was the

188

T.H. GIORDANO & Y.K. KHARAKA

most effective catalyst, but only one experiment was conducted at 335 ° C. Rapid rates of decomposition of acetic acid were observed with hematite and an iron-bearing montmorillonite, but the rates did not follow a first-order rate law and the gas composition indicated that the acetic acid was being oxidized, not decarboxylated, with these minerals (Bell 1991). Decarboxylation rates for organic anions other than acetate are limited, but the available data show generally higher rates of destruction than those for acetate, probably explaining the dominance of acetate at subsurface temperatures higher than about 80°C (Kharaka 1986). Boles (1986) showed that the decarboxylation rates for gallic acid give a half-life value of 150 hours at 100° C and concluded that this and other aromatic carboxylic acids, if generated, will be destroyed rapidly in sedimentary basins at temperatures higher than 80 ° C. Crossey (1991) measured decarboxylation rates for oxalic acid at 80°C that gave half-lives ranging from 2.5 × 103 to 2.8 × 104 a at pH values 5-7. She compared her results for oxalic and formic acids with literature values and concluded that the stability order is: acetate > > formate > oxalate > gallate > malonate. The decarboxylation rates measured for butyric acid were similar to those for acetic acid indicating that the abundance of C2-C5 monocarboxylic acids in formation waters may be controlled mainly by their generation from precursor compounds in kerogen (Palmer & Drummond 1986). While laboratory studies of decarboxylation have established the important controls and the relative destruction rates, these data are not directly applicable to field situations where complex mineral assemblages with variable amounts of multiple catalysts are present. Decarboxylation rates applicable to field situations, however, can be estimated at temperatures higher than about 80° C from the reported concentrations of organic acid anions, the subsurface temperatures and the ages of the reservoir rocks (Kharaka 1986). It is assumed in these calculations that decarboxylation rates follow a first-order rate law and that generation rates from kerogen are equal. Using the data base from Carothers & Kharaka (1978), Kharaka (1986) estimated a range of 2751 × 106 a for half-lives of acetate decomposition at 100°C in San Joaquin Valley, California, and Gulf Coast, Texas and Louisiana, basins. Using data from the temperature range 80-120 ° C from Gulf Coast basin, Lundegard &

Land (1989) estimated a half-life value of 60 × 106 a. Field rates discussed above are approximations that probably give maximum half-lives. The greatest unknown is the residence time of the dissolved organic acid at a particular temperature, which depends on the geological, thermal and hydrological history of the basin. Kharaka (1986) attempted to estimate the residence times and calculated an average half-life value of 20 × 106 a at 100° C. These estimations indicate that the concentrations of acetate in formation waters from Cenozoic rocks could have been higher initially, but that values could have been reduced thermally only by a factor of two to three. Combining the field-estimated decarboxylation rates for acetate with experimental results for acetate and other organics discussed above, it follows that the concentrations of C2-C5 monocarboxylic acid anions in formation waters from Cenozoic rocks could have been reduced by a factor of two or three. Because of much higher decarboxylation rates, however, the concentrations of formate and all the dicarboxylic and aromatic acid anions could have been reduced to a much greater extent thermally. Thermal decarboxylation of organic acids in the subsurface at temperatures less than 80°C also occurs, but bacterial destruction is probably a more important process responsible for the low concentrations at these temperatures. Rates of bacterial destruction are rapid, but these have not been adequately quantified, though bacteria are selective and destroy acetate, amino acids and carbohydrates at much faster rates than other organic anions (Thurman 1985; Ehrlich 1990).

Thermodynamic data for complexing and protonation reactions To evaluate the significance of a particular organic ligand species in diagenetic and oreforming processes, the theoretically estimated concentration of that species in the aqueous phase must be determined by calculation. Various types of information are required to perform such a calculation including chemical and physical parameters of the aqueous fluid and its surrounding geochemical environment, composition and thermodynamic data of all possibly important non-aqueous phases, and stoichiometry and thermodynamic data for all possibly important aqueous species. This latter category will be discussed here, with specific attention given to organic acid ligands and their corresponding metal-organic complexes and protonated species. To calculate acid-ion ligand

ORGANIC LIGANDS IN SUBSURFACE WATERS

189

Table 6. Experimentally determined stability constants for selected acetate, oxalate, malonate, and succinate species under hydrothermal conditions

Reaction

Log K ( T , o C)

Ref.

Acetate (Ac- = C H 3 C O 0 - )

H + + Ac- = HAc Na + + Ac- = NaAc Ca 2+ + Ac- = CaAc + Mg 2+ + Ac- = MgAc ÷ AI 3+ + AcAIAc 2+ AP ÷ + 2Ac- = AI(Ac)2 + AP + + 3Ac- = AI(Ac)3 Fe 2+ + Ac- = FeAc + Fe 2+ + 2Ac- = Fe(Ac)2 Pb 2+ + Ac- = PbAc + Pb z+ + 2Ac- = Pb(Ac)~ Zn 2÷ + Ac- = ZnAc + Zn 2÷ + 2Ac- = Zn(Ac)2 Zn 2+ + 3Ac- = Zn(Ac)3=

4.79 -0.12 1.2 1.3 3.06 5.1

(50) (45) (80) (80) (50) (50)

4.94 0.03 2.53

(100) (275) (200)

5.52 0.29 3.72

(200) (300) (300)

6.66 (300) 0.498 (320) 4.59 (350)

2.9 4.8

(80) (80)

3.94 6.4

(100) (100)

1.1 2.1 2.5 3.8 1.9 3.4 4.1

(50) (50) (40) (40) (50) (50) (50)

1.6 (100) 2.6 (100) 2.56 (55) 3.98 (55) 2.3 (100) 4.0 (100) 4.7 (100)

2.4 4.0 2.66 4.18 3.5 5.9 6.6

(200) (200) (70) (70) (200) (200) (200)

5.08 (150) 7.9 (150) 11.3 (150) 3.7 (300) 6.82 (300) 2.78 (85) 4.38 (85) 5.3 (300) 8.7 (300) 9.4 (300)

1 2 3 4 5 5 5 6 6 7 7 8 8 8

4.9

(50)

5.06

5.32

(150)

5.64

(200)

9

1.358 4.408 6.7 16.5 5.38 7.52

(50) (50) (100) (80) (45) (45)

1.463 (75) 4.792 (100)

1.581 (100) 5.264 (125)

1.709 (125) 5.529 (150)

5.44 7.51

5.53 7.23

5.81 7.80

10 10 11 12 13 13

Propionate (Pr- = C3H502-)

H + + Pr- = HPr

(100)

Oxalate (Ox 2- = C2042-)

H ÷ + HOx- = H2Ox H + + Ox 2- = HOxAP + + Ox z- = AIOx + A13÷ + 3Ox 2- = Al(Ox)33Pb 2+ + Ox 2- = PbOx Pb 2+ + 2Ox 2- = Pb(Ox)22-

(65) (65)

(85) (85)

(91) (91)

Malonate (Ma 2- = C3H2042-)

H + + HMa- = H2Ma H + + Ma 2- = HMa-

2.879 (50) 5.805 (50)

2.944 (75) 5.958 (75)

3.02 (100) 6.138 (100)

14 14

4.19 (40) 5.654 (40)

4.18 (50) 5.680 (50)

4.13 (74) 5.726 (74)

15 15

Succinate (Suc 2- = C4H4042-

H + + HSuc- = H2Suc H ÷ + Suc 2- = HSuc-

(l) Mesmeretal. 1989; (2) 45 ° C: De Robertis etal. 1985; 275o-320 ° C: Oscarson etal. 1988; (3) 80° C: Fein 1991a. 200-350 ° C: Seewald & Seyfried 1991 ; (4) Fein 1991a; (5) 80° C: Fein 1991a; 50 °, 100° and 150° C: Palmer & Bell 1993; (6) Palmer & Drummond 1988; (7) best fit values: data from Hennet et al. 1988a and Giordano 1989; (8) Giordano & Drummond 1991; (9) Ellis (1963); (10) Kettler et al. 1991; (11) Thyne etal. 1992; (12) Fein 1991b; (13) Hennet et al. 1988a; (14) Kettler et al. 1992; (15) Jones & Soper 1936 and Pinching & Bates 1950. speciation, stability constants or free e n e r g y data for p r o t o n a t e d organic ligand species and specific m e t a l - o r g a n i c c o m p l e x e s are r e q u i r e d . With respect to the acid-ion ligands discussed in the previous section, t h e r m o d y n a m i c data bases are the least d e v e l o p e d for h u m a t e and fulvate species. T r u e t h e r m o d y n a m i c equilibrium constants are difficult to obtain for reactions involving h u m a t e and fulvate species b e c a u s e of the n o n - u n i f o r m n a t u r e of naturally occurring h u m i c substances and the difficulty in characterizing individual h u m a t e and fulvate ions (Tipping & H u r l e y 1992). H o w e v e r , b e c a u s e of the i m p o r t a n c e of such c o m p o u n d s in the soil e n v i r o n m e n t , conditional stability constants have b e e n g e n e r a t e d at t e m p e r a t u r e s n e a r 25 ° C for a large n u m b e r of fulvate and h u m a t e c o m p l e x e s involving metal cations typically

f o u n d in oxidized e n v i r o n m e n t s n e a r the E a r t h ' s surface. Several excellent reviews a n d compilations of m e t a l - h u m a t e and metal-fulvate c o m p l e x e s are available (e.g., Schnitzer & K h a n 1978; Fitch & S t e v e n s o n 1983; P e r d u e 1985; P e r d u e & Lytle 1986; Tipping & H u r l e y 1992). Martell & Smith (1974, 1977, 1982) and Smith & Martell (1975, 1989) have d e v e l o p e d an extensive c o m p i l a t i o n of stability constants and o t h e r t h e r m o d y n a m i c data for a large n u m b e r of n o n - h u m i c organic ligands including a m i n o acids, carboxylic acids, and o r g a n o s u l p h u r species. N e a r l y all the data in this c o m p i l a t i o n and in o t h e r r e c e n t compilations of stability constants (e.g., Perrin 1979) are for t e m p e r a tures b e l o w 100 ° C and mostly for t e m p e r a t u r e s n e a r 25 ° C. T h e s e data as well as those available for h u m a t e and fulvate c o m p l e x e s are sufficient

190

T.H. G I O R D A N O & Y.K. K H A R A K A

Table 7. Stability and protonation constants for selected organic ligand equilibria Log K(T, ° C) Reaction

25

50

100

150

200

4.76 -0.39 -0.18 1.12 1.28 2.75 4.61 0.87 1.89 2.25 3.50 3.72 1.7 3.1 3.85

4.79 -0.13 -0.16 1.15 1.16 3.06 5.15 1.10 2.10 2.49 3.74 3.67 1.9 3.4 4.1

4.94 -0.13 -0.03 1.42 1.18 3.93 6.44 1.60 2.60 2.76 4.52 3.66 2.3 4.0 4.7

5.18 -0.13 0.15 1.86 1.42 5.07 7.91 2.00 3.20 3.15 5.52 3.77 2.8 4.83 5.5

5.52 -0.13 0.38 2.41 1.79 6.36 9.52 2.40 4.00 3.65 6.66 4.02 3.5 5.9 6.6

1.27 4.28 0.70 0.90 3.00 3.47 5.19 11.00 15.0 4.40 5.36 7.46 4.90

1.36 4.41 0.70 0.90 3.00 3.47 5.58 11.00 15.6 4.50 5.49 7.58 4.90

1.58 4.79 0.70 0.90 3.00 3.47 6.35 11.00 17.0 4.90 5.92 8.10 5.10

1.84 5.26 0.70 0.90 3.00 3.47 7.15 11.00 18.3 5.20 6.58 8.82 5.40

2.85 5.70 0.74 2.51 2.90 3.56 2.80 2.54 3.85

2.88 5.81 0.74 2.90 3.19 5.73 3.24 2.97 4.14

3.02 6.14 0.74 3.39 3.58 8.22 3.73 3.47 4.55

Ref.

Acetate (Ac- = C H 3 C O 0 - ) H + + Ac- = H A c K ÷ + Ac- = KAc Na ÷ + Ac- = NaAc Ca 2+ + Ac- = CaAc ÷ Mg ~÷ + Ac- = MgAc + AP ÷ + Ac- = AIAc 2÷ A! 3÷ + 2Ac- = AI(Ac)+2 Fe 2+ + Ac- = FeAc + Fe z+ + 2Ac- = Fe(Ac)2 Pb 2+ + Ac- = PbAc ÷ Pb 2+ + 2Ac- = Pb(Ac)2 Pb 2+ + 3Ac- = Pb(Ac)-3 Zn 2+ + Ac- = Z n A c ÷ Zn 2+ + 2Ac- = Zn(Ac)2 Zn 2+ + 3Ac- = Zn(Ac)-~

1 2 3 4 4 5 5 6 6 7 7 2 8 8 8

Oxalate (Ox 2- = C20 2-) H ÷ + H O x - = H2Ox H + + Ox 2- = H O x K + + Ox 2- = KOxNa + + Ox 2- = NaOxCa 2+ + Ox 2- = CaOx Mg 2+ + Ox 2- = MgOx AP ÷ + Ox 2- = AIOx ÷ AP ÷ + 2Ox 2- = Al(Ox)5 AP + + 3Ox 2- = Al(Ox) 3Fe 2+ + Ox 2- = FeOx Pb 2÷ + Ox 2- = PbOx Pb 2+ + 2Ox 2- = Pb(Ox) 2Zn 2+ + Ox 2- = ZnOx

2.16 5.81 0.70 0.90 3.00 3.47 7.90 11.00 19.6 5.80 7.40,' 9.68 5.90

9 9 2 2 2 2 10 2 10 2 11 11 2

0.74 4.95 4.90 15.30 4.98 4.72 6.01

12 12 13 14 14 14 14 14 14

Malonate (Ma 2- = C3H2024-) H + + H M a - = H2Ma H + + Ma 2- = H M a Na + + Ma 2- = N a M a Ca 2+ + Ma 2- = CaMa Mg 2÷ + Ma 2- = MgMa AI 3+ + Ma 2- = AlMa + Fe 2+ + Ma 2- = FeMa Pb 2÷ + Ma 2- = PbMa Zn 2+ + Ma 2- = Z n M a

to a d e q u a t e l y m o d e l o r g a n i c l i g a n d s p e c i a t i o n in s u p e r g e n e e n v i r o n m e n t s , e a r l y d i a g e n e t i c syst e m s , a n d l o w - t e m p e r a t u r e o r e fluids, if all o t h e r p e r t i n e n t d a t a a n d i n f o r m a t i o n are a v a i l a b l e . T o a c c u r a t e l y c a l c u l a t e o r g a n i c l i g a n d specia t i o n in h y d r o t h e r m a i fluids, t h e r m o d y n a m i c d a t a at t h e s p e c i f i e d h y d r o t h e r m a l t e m p e r a t u r e are r e q u i r e d . A t p r e s e n t , e x p e r i m e n t a l l y d e t e r m i n e d h i g h - t e m p e r a t u r e s t a b i l i t y c o n s t a n t s are available for only a few species of interest; most notably, acetate and oxalate complexes of some rock-forming and ore-forming metals and for p r o t o n a t e d species o f s e v e r a l c a r b o x y l a t e

0.74 4.09 4.16 11.51 4.36 4.09 5.18

l i g a n d s ( T a b l e 6). F o r m o s t o t h e r o r g a n i c acids and metal-organic complexes of interest, experimentally d e t e r m i n e d stability constants are a v a i l a b l e o n l y at t e m p e r a t u r e s n e a r 2 5 ° C . A l t h o u g h e x p e r i m e n t a l d a t a m a y be l a c k i n g , high t e m p e r a t u r e stability constants can be estimated using reference thermodynamic data at s o m e l o w e r t e m p e r a t u r e , u s u a l l y 25 ° C, a n d one of several extrapolation m e t h o d s ( K h a r a k a et al. 1988). In a d d i t i o n to t h e v a n ' t H o f f a p p r o a c h (SUPCRT92, J o h n s o n et al. 1992), t h e r e are m e t h o d s based on the volumetric properties o f w a t e r ( M a r s h a l l 1970; M e s m e r et al. 1988;

ORGANIC LIGANDS IN SUBSURFACE WATERS

Table 7.

191

Continued.

Log K(T,°C) Reaction

25

50

100

150

200

Ref.

4.2I 5.64 0.30 0.30 2.00 2.05 5.30 10.00 2.20 3.50 2.46

4.I8 5.68 0.30 0.30 2.00 2.05 5.30 10.00 2.40 3.70 2.73

4.25 5.89 0.30 0.30 2.00 2.05 5.30 10.00 2.80 4.30 3.33

4.44 6.25 0.30 0.30 2.00 2.05 5.30 10.00 3.30 5.00 4.02

4.80 6.89 0.30 0.30 2.00 2.05 5.30 10.00 3.90 5.90 4.85

15 15 2 2 2 2 2 2 2 2 2

Succinate (Suc e- = C4H40~-)

H + + HSuc- = H2Suc H ÷ + Suc2- = HSucK + + Suc2- = KSucNa + + Suc2- = NaSucCa 2+ + Suc2- = CaSuc Mgz+ + Suc2- = MgSuc AP ÷ + Suc2- = AISuc ÷ AP ÷ + 2Suc 2- = Al(Suc)~Fe 2+ + Suc2- = FeSuc Pb 2+ + Suc2- = PbSuc Zn 2÷ + Suc2- = ZnSuc

(1) Mesmer et al. 1989; (2) Data from SOLMINEQ. 88, Kharaka et al. 1988; (3) extrapolated based on low temperature data from Archer & Monk 1964 and DeRobertis et al. 1985; (4) extrapolated based on low temperature data from Archer & Monk 1964; (5) Palmer & Bell 1993; (6) Palmer & Drummond 1988; (7) extrapolated based on 25-85° C data from Giordano 1989; (8) Giordano & Drummond 1991; (9) Kettler et al. 1991; (10) Thyne etal. 1992; (11) extrapolated based on 25-90 ° C data from Hennet et al. 1988a; (12) Kettler etal. 1992; (13) 25° C constant cited by Martell & Smith 1982; (14) Harrison & Thyne 1992; (15) extrapolated based on 25-74 ° C data from Jones & Soper 1936 and Pinching & Bates 1950. Anderson et al. 1991), those based on the isocoulombic principle (Lindsay 1980; Hennet et al. 1988a; Mesmer et al. 1988; Giordano & D r u m m o n d 1991), and the predictive method DQUANT which uses the standard enthalpy and entropy of reaction (Helgeson 1967, 1969; Harrison & Thyne 1992). If experimentally determined values for stability constants are completely lacking for a particular metalorganic complex of interest, it may be possible to estimate values using a variety of correlation methods (Langmuir 1979; Helgeson et al. 1981; Tanger & Helgeson 1988; Shock & Helgeson 1990; Shock et al. 1992). In the speciation models presented in the next section, activities for metal-organic complexes involving four acid-ion ligands (acetate, oxalate, malonate, and succinate) were calculated using experimental or predicted high-temperature equilibrium constants (Table 7) for all acid ionization reactions; acetate complexes of Na +, Ca 2+, Mg 2+, AI 3+, Fe 2÷, Pb 2+, and Zn2+; oxalate complexes of A13+, Fe 2+, Pb 2+, and Zn2+; malonate complexes of Ca 2+ , Mg 2+ , AP ÷ , Fe 2÷, Pb 2÷, and Zn2+; and succinate complexes of Fe 2+, Pb z+, and Zn 2÷. Low-temperature data near 25°C (Table 7) were used for all other complexes involving these four organic ligands and the cations Na +, Ca 2÷, M f +, A13+, Fe 2+, Pb 2÷, and Zn 2+. The available high-temperature data from the literature were used for reactions involving the inorganic ligands chloride, carbonate, sulphate, sulphide, and hydroxyl ions.

Propionate is an important ligand in many subsurface waters but high temperature data is lacking for its complexes. Because propionate complexes have stabilities similar to acetate complexes, we assign, in our chemical models, the concentration value of acetate plus propionate as total acetate concentration and consider only acetate complexation.

Chemical modelling Diagenesis models

The concentration of organic acid anions in subsurface waters can be high, especially in oil-field waters at temperatures of 80-120 ° C, where the total may be up to 10000mg1-1. Because acetate and other organic acid anions are proton donors or acceptors, depending on the p H of water as well as the anion, these anions play an important role in all pH-dependent homogeneous and heterogeneous reactions in water-rock systems (Willey et al. 1975; Lundegard & Kharaka 1990). Also, these organic species can form strong complexes with A1, Fe, and other cations that control mineral diagenesis (Surdam & Crossey 1987; Lundegard & Kharaka 1990; Harrison & Thyne 1992). In order to quantify the role of organic and inorganic complexes in mineral diagenesis we used the geochemical code SOLMINEQ (Kharaka et al. 1988) and modeled ligand speciation in a

192

T.H. GIORDANO & Y.K. KHARAKA

Table g. The percentage of cations complexed with added organic acid anions in formation waterfrom Pleasant Bayou no. 2 well, Texas. Percentage of cation complexed with Constituent Na K Ca Mg Fe AI Pb Zn CI HCO3 504 H2S SiO2 pH TD S

mg 1-~

Acetate

Oxalate

Maionate

38000 840 9100 660 62 0.01 1.1 1.5 80000 365 5.4 0.5 120 4.92 132000

1.4 0.6 25.5 19.5 12.3 0.3 3.4 0.2

<0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1

<0.1 <0.1 0.2 0.5 0.1 3.3 <0.1 <0.1

Succinate <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1

Anion concentration (mg 1-1): acetate 10000; oxalate 10; malonate 100; succinate 100. Sample No: 79GG204, Production zone: Frio Sand. Depth: 4462 m. Temperature: 150° C. Pressure 700 bar.

modern oil-field brine. This code has thermochemical data for various metal complexes involving acetate, oxalate, and succinate, with an option to add the complexes of two additional organic anions for a total close to 100 organic species. Where necessary the dissociation constants of cation-organic complexes in SOLMINEO were updated using the data in Table 6. The chemical composition and the subsurface temperature and pressure of water from Pleasant Bayou No. 2 well, Texas, used by Lundegard & Kharaka (1990) were also used in this test case (Table 8). The concentration of organic acid anions used in this test case are very high, but possible, and were selected to investigate the maximum complexing by organic anions. The concentration of oxalate (10 mg 1-~) and other dicarboxylic acid anions used are lower than those reported by MacGowan & Surdam (1988) and lower than those used by other investigators (e.g., Lundegard & Kharaka 1990; Harrison & Thyne 1992) in similar studies, because a recent investigation by Kharaka et al. (1993b) indicated that the values in Table 8 are probably the highest likely concentrations in modern oil-field waters. Results of simulations at pH 4.92, the calculated pH at subsurface conditions for this sample, are comparable to those obtained by Lundegard & Kharaka (1990) and show that important percentages of Ca, Mg, and Fe are complexed mainly with acetate (Table 8). Only a small amount of A1 (c. 4%) and even smaller

percentages of Na and K are complexed. Higher amounts of A1 are complexed (c. 50%) mainly with malonate when the simulations are run using a pH value of 3.92, but less than 0.1% AI is complexed at pH values of 5.92 and higher. Because organic acid anions become protonated at low pH values, the amounts of cations, other than Ai, complexed with organic anions increase and decrease as the pH values used are increased and decreased from 4.92, respectively. The saturation states of selected minerals with and without the added organic anions to the waters of Pleasant Bayou No. 2 well at subsurface conditions (Table 8) are shown in Table 9. The difference between the saturation states of halite are very small, reflecting the fact t h a t Na-organic complexes are very weak (Table 8). The differences between the saturation states of minerals containing A1, Ca, Fe, and other multivalent cations are much larger and geochemically important because they form stronger complexes with organic anions. In the case of calcite and siderite results show that the water is more undersaturated with these minerals in the absence of organic anions (Table 9). These results are surprising because important portions of Ca and Fe are complexed with organic anions (Table 8) and the pH of water is held constant at 4.92. Careful examination of the computer-generated printout of the aqueous species shows that the increased saturation states of calcite and siderite result from increased activity of CO32- species (aco32-) which

ORGANIC LIGANDS IN SUBSURFACE WATERS Table 9. Saturation states of selected minerals with and without added organic anions to waterfrom Pleasant Bayou no. 2 well, Texas.

Minerals Halite Albite Kaolinite Anhydrite Calcite Siderite

AG(with AG(without) organics) kcal/mole -2.93 -4.38 -6.77 -3.95 -0.56 - 1.88

-2.94 -4.41 -6.83 -4.16 0.56 -0.59

is greater than the decreased a Ca 2+ and a F e 2+ caused by complexing with organic anions. The increased aco3 2 - results from decreased complexation of CO32- with Ca, Fe, and other cations which are strongly complexed with organic anions. In the computation of metal-organic complexes and mineral saturation states discussed above, the aim is not to model all potential chemical shifts to be expected in subsurface waters, but to indicate when organic complexing can be important geochemically. These computations can be extended further to indicate porosity gain in specific petroleum reservoirs from increased mineral dissolution in the presence of organic ligands (Surdam & Crossey 1987; Harrison & Thyne 1992). The results, of course, will be different for different organic and inorganic composition as well as the temperature and pressure conditions, but the test cases show that organic species should be included in the chemical analysis of the water and in geochemical modeling. Results from such computations will become more reliable as the thermochemical data on cation-organic complexes are improved, especially at elevated temperatures. Ore f l u i d m o d e l s

To transport significant quantities of metal in ore solutions as a metal-organic complex, the participating organic ligand must meet the following criteria (Giordano 1985; Drummond & Palmer 1986). The ligand must (1) be present as a dissolved constituent of the ore fluid at a sufficient level of concentration, (2) have ionization constants that permit significant proportions of the anionic form at ore-fluid pHs, (3) have sufficient thermal stability to survive at least one pass of the ore fluid through the system, and (4) form metal-organic complexes

193

of sufficient strength to favorably compete with inorganic and other organic ligands. This latter criterion is discussed below in terms of competing organic ligands and their effect on metal speciation. Several lines of evidence (Barnes & Czamanske 1967; Barnes 1979; Anderson 1983) indicate that concentrations of base and ferrous metals in ore-forming solutions must be at least 1-10ppm (approximately 10-5-10 -4 molal) to form a typical hydrothermal ore deposit. Minimum ore-fluid concentrations for rarer metals (e. g. Hg, Au, Ag) are probably several orders of magnitude lower. It follows that the total concentration of the ligand or ligands involved in significant ore-metal transport must also be near or above these same minimum concentrations. Giordano (1993) calculated concentrations of Na, Ca, Mg, A1, Fe, Pb, and Zn as acetate, oxalate, malonate, and succinate complexes in three reconstructed Mississippi Valley-type (MVT) ore fluids and a model composite ore-fluid for red-bed related base metal (RBRBM) deposits. It is now widely accepted that these two deposit types are related to sedimentary basinal processes and that their ore fluids were probably evolved, sedimentary basin brines (Sverjensky 1989). Based principally on this relationship, it is assumed that the most important organic ligands in these ore fluids are the same as those in sedimentary basin brines (i.e., acetate, propionate, oxalate, malonate, and succinate). The speciation models presented by Giordano (1993) show (1) that significant amounts of lead and zinc cannot be mobilized as metal-organic complexes involving acetate or other carboxylate ligands in ore fluids containing greater than 10 -5 molal reduced sulphur as hydrogen sulphide or bisulphide and (2) that small but significant quantities of dissolved lead and zinc can be transported by carboxylate complexes in oxidized ore fluids containing less than 10 -9 molal reduced sulphur as sulphide or bisulphide. In Table 10, results are presented for revised speciation calculations for the model RBRBM ore fluid at 100° C. These new results are based in part on an updated set of values for protonation and stability constants for equilibria involving organic ligands (Table 7). The composite RBRBM ore fluid is moderately oxidized and falls well above the sulphide-sulphate redox boundary at 100° C. It is slightly acid (pH = 6) and contains 10 -2 molal total sulphur, principally in the form of sulphate, with reduced sulphur concentrations well below 10 -5 molal. The model parameters used for this composite ore fluid are based on those proposed by Rose (1976, 1989), Sverjensky (1987, 1989), and

194

T.H. GIORDANO & Y.K. KHARAKA

Table 10. Calculated metal speciation in model ore-fluid for red-bed related base metal deposits Percent metal in free ion and indicated complex*

Metal Na Ca Mg Pb Zn Fe AI

Total (ppm) 3.82 × 3.40 × 8.74 x 2.61 × 1.78 × 1.16 × 2.92 ×

10 4 10 3

102 101 102 10° 10-3

Free ion 85 77 61 0.57 0.88 35 <0.01

Chloride Hydroxide 13 6.0 26 86 90 47 na

<0.01 <0.01 <0.01 8.0 5.6 0.23 99

Acetate

Oxalate

Malonate

Succinate

2.0 17 12 5.2 3.2 16 0.10

<0.01 0.03 0.14 0.12 0.04 1.1 <0.01

<0.01 0.46 0.85 <0.01 0.08 0.47 <0.01

0.01 0.09 0.11 0.08 0.02 0.24 <0.01

Parameters for composite ore fluid include: T = 100° C, pH = 6, logao2 = -50, total sulphate = 10.2.0molal, total carbon = 10.2.4 molal (Giordano 1993). Total chloride = 2.0 molal, total acetate = 0.1 molal, total oxalate = 0.0001 molal, total succinate = 0.0009 molal, total malonate = 0.001 molal, na indicates not calculated. Branam & Ripley (1990) for sediment-hosted copper-rich deposits. The ore fluid is also assumed to be saturated with respect to galena, sphalerite, hematite, calcite, quartz and potassium feldspar. The concentrations of acetate + propionate (10000 mg 1-1), oxalate (10 mg 1-~), malonate (100 mg 1-~), and succinate (100 mg 1-~) used in the model are based on our revised estimates for the highest likely concentrations of these acid ions in oil-field brines. The results presented in Table 10 illustrate that significant amounts of lead, zinc, iron, and aluminium can be mobilized in the model R B R B M ore fluid, and that competition among cations for acetate, and to a lesser extent the other acid ions, is an important factor in controlling the speciation in this ore solution. In this oxidized and slightly acid solution, the ligands acetate, oxalate, malonate, and succinate are present primarily as free ions, singly protonated species, and complexes of Na, Ca, and Mg. Although carboxylate complexes of Pb, Zn and AI have a greater thermodynamic stability than carboxylate complexes of Na, Mg, and Ca, the free ion activities of the former group of metals are too low to allow significant competition with the weakly complexing, but more abundant, ions of Na, Mg and Ca. Calculated metal speciation is given in Table 10 as percent metal bound in the indicated species. More than 80% of the the lead and zinc is in chloride complexes but a significant 3-5% of these metals is bound in acetate complexes. Calcium and magnesium are present primarily as free ions or in chloride complexes, but approximately 12-17% of these metals is present in acetate complexes, with 1% or less present in complexes involving oxalate, malonate, and succinate. Iron is present primarily as a free ion (35 %) and in chloride complexes (47 % ). Similar

to calcium and magnesium, about 16% of the iron is in acetate complexes, with less than about 1% in oxalate, malonate, and succinate complexes. Aluminium is primarily in the form of hydroxide complexes, but the model shows that significant amounts of aluminium may be complexed with monocarboxylate ligands (0.1% in acetate complexes). In the future, the availability of high temperature thermodynamic data for aluminium-dicarboxylate complexes should allow a more reliable evaluation of the effects of oxalate, malonate, and succinate complexing on aluminium speciation in oil-field brines and ore fluids. The lack of thermodynamic data for Cu+-carboxylate complexes precluded an evaluation of copper speciation in this R B R B M ore fluid. The calculated speciations presented in Table 10 illustrate the relative importance of inorganic and organic complexing and the significant effects of carboxylate complexing on the speciation of ore-forming and rock-forming metals for the tested ore fluid. However, these results should be considered with caution. More reliable estimates of speciation can be made when additional high-temperature thermodynamic data for metal-organic complexes and additional information regarding pertinent organic ligands and their concentrations in ore fluids become available.

Summary and conclusions The review presented in this contribution summarizes our current understanding of the nature, distribution, and interactions of organic ligands in sedimentary basins and related diagenetic and ore-forming systems. In water-saturated sediment and other shallow subsurface environments, humic and fulvic acids, amino acids, and

ORGANIC LIGANDS IN SUBSURFACE WATERS organosulphur species are the dominant organic compounds affecting early diagenetic reactions and related ore-forming processes. However, the nature, distribution and interactions of these dissolved ligands in the shallower zones of sedimentary basins are poorly understood. Better understood is the chemistry of dissolved aqueous carboxylic acid species in the deeper regions of sedimentary basins. In most oil-field brines, acetate is generally the dominant organic ligand and is followed in level of concentration by propionate and longer chained aliphatic carboxylate ions. Some deep basinal waters contain relatively high concentrations of difunctional carboxylate ions, especially succinate. Although higher concentrations have been reported, our best estimate of actual maximum concentrations of the dominant acid ions in modern oil-field brines are 10000 mg1-1 (acetate + propionate), 10mg1-1 (oxalate), 100 mg 1-1 (malonate), and 100 mg 1-1 (succinate). The highest concentrations of organic acid anions are present in formation waters at temperatures between 80°-120 ° C. Based on the available field and laboratory evidence, the concentrations of organic acid anions are controlled mainly by the rates of their generation from kerogen and their destruction thermally or by bacteria. In addition, concentrations of oxalate and malonate may be limited by the low-solubilities of their calcium salts. The relatively low concentrations of organic acid anions and low relative abundance of acetate in waters less than 80 ° C are generally attributed to bacterial consumption of these anions, especially acetate. The decrease in concentration of organic acids with increasing temperature is generally attributed to thermal decarboxylation of these anions to CO2 and natural gas. Martell & Smith (1974, 1977, 1982) and Smith & Martell (1975, 1989) have developed an extensive compilation of stability constants and other thermodynamic data for a large number of non-humic organic ligands. Unfortunately, nearly all the data in these compilations are for temperatures near 25° C. These data as well as those available for humate and fulvate complexes are sufficient to adequately model organic ligand speciation in low temperature ore fluids and early diagenetic systems, if all other pertinent data are available. To accurately calculate organic ligand speciation in hydrothermal fluids, thermodynamic data at the specified hydrothermal temperatures and pressures are required. At present, experimentally determined high temperature stability constants are available for only a few species of interest; most

195

notably, acetate and oxalate complexes of some rock-forming and ore-forming metals and for protonated species of several carboxylate ligands. Chemical models are useful in illustrating the significance of organic ligand interactions on the speciation of cations in ore-forming and diagenetic fluids. General conclusions can be made in this regard by examining the results of speciation calculations for oil-field brines and model ore-fluids. Simulations of speciation in oil-field brines show that significant amounts of Ca, Mg, Fe, and A1 can be complexed by carboxylate ligands, in the pH range 4-6, if ligand concentrations are of the order of 10-100 mg1-1. An important consequence of such complexation is enhanced mineral solubilities and increasing porosity development with increasing organic ligand concentrations. Calculated speciation in model ore fluids for Mississippi Valley-type deposits and red-bed related base metal deposits show that carboxylate complexing can contribute significantly to the speciation of Na, Ca, Mg, and A1 in ore fluids over a wide range of oxidation states and total reduced sulphur concentration; however, optimum conditions for Fe, Pb, and Zn transport by organic ligand complexation are oxidized ore fluids with total reduced inorganic sulphide concentration less than 10 -9 molal. The details of such model calculations, however, should be considered with caution. More reliable estimates of speciation and solution-mineral interactions can be made when additional hightemperature thermodynamic data for cationorganic complexes and additional information regarding pertinent organic ligands and their concentrations in basinal fluids become available.

References AIKEN, G.R., MCKNIGHT,D.M. & WERSHAW,R.L. 1985. Humic Substances in Soil, Sediment and Water, New York, John Wiley & Sons. ANDERSON,G.M. 1983. Some geochemical aspects of sulfide precipitation in carbonate rocks: In: KESVARSANYI,G., GRANT,S.K., PRATT,W.P. & KOENIC, J.W. (eds) International Conference on Mississippi Valley Type Lead-Zinc Deposits. Proceedings volume. University Missouri-Rolla,

61-76. CASTET,S., SCHOTr, J. & MESMEa, R.E. 1991. The density model for estimation of thermodynamic parameters of reactions at high temperatures and pressures. Geochimica et Cosmochimica A cta, 55, 1769-1779. ARCHER, D.W. & MONK, C.B. 1964. Ion-association constants of some acetates by pH (glass electrode)

--,

196

T.H. G I O R D A N O & Y.K. K H A R A K A

measurements. Journal of the Chemical Society, 60, 3117-3122. BAKER, W.E. 1978. The role of humic acid in the transport of gold. Geochimica et Cosmochimica Acta, 42,645-649. BADA, J.L., ZHAO, M. & MILLER, S.L. 1991. Alanine stability in aqueous solutions at 250 ° C. Geologi-

cal Society of America Abstracts with Programs, 23, A25. BARCELONA, M.J., LILJESTRAND, H.M. & MORGAN, J.J. 1980. Determination of low molecular weight volatile fatty acids in aqueous samples. Analytical Chemistry, 52,321-325. BARNES, H.L. 1979. Solubilities of ore minerals. In: BARNES, H.L. (ed.) Geochemistry of Hydrothermal Ore Deposits, 2rid ed. New York, John Wiley & Sons, 405-461. 1983. Ore-deposition reactions in Mississippi Valley-type deposits. In: KISVARSANY, G., GRANT, S.K., PRATr, W.P. & KOENIG,J.W. (eds)

International Conference on Mississippi Valley Type Lead-Zinc Deposits, Proceedings Volume. University of Missouri-Rolla, 75-85. & CZAMANSKE, G.K. 1967. Solubilities and transport of ore minerals. In: BARNES, H.L. (ed.)

Geochemistry of Hydrothermal Ore Deposits, Holt, Rinehart, and Winston, New York, 334381. BARTH, T. 1987. Quantitative determination of volatile carboxylic acids in formation waters by isotachophoresis. Analytical Chemistry, 59, 22322237. 1991. Organic and inorganic ions in water from petroleum reservoirs. Norwegian continental shelf. Applied Geochemistry, 6, 1-15. ~, BORGUND, A.E. & HOPLAND, A.L. 1989. Generation of organic compounds by hydrous pyrolysis of Kimmeride oil shale - Bulk results and activation energy calculations. Organic Geochemistry, 14, 69-76. BARTON, P.B. 1982. The many roles of organic matter in the genesis of mineral deposits. Geological Society of America Abstracts with Programs, 14, 440. BELL, J.L. 1991. Acetate decomposition in hydrothermal solutions. PhD thesis, The Pennsylvania State University. BERNHARDT, G., LUNDEMANN, H.D., JAENICKE, R., KONI6, H. & STErrER, K.O. 1984. Biomolecules are unstable under 'black smoker' conditions. Naturwissenschaften, 71,583-586. BOLES, J.S. 1986. Aqueous thermal degradation of

naturally occurring aromatic organic acids and the synthetic chelating agent disodium EDTA. PhD thesis, Princeton University. , CRERAR, D.A., GRISSOM, G. & KEY, T. 1988. Aqueous thermal degradation of gallic acid. Geochimica et Cosmochimica Acta, 52,341-344. BOYLE, R.W., BROWN, A.C., JEFFERSON, C.W., JOWETT, E.C. & KIRKHAM,R.V. 1989. SedimentHosted Stratiform Copper Deposits. Geological Association of Canada Special Papers, 36. BRANAM, T.D. & RIPELY, E.M. 1990. Genesis of

sediment-hosted copper mineralization in southcentral Kansas: sulfur/carbon and sulfur isotope systematics. Economic Geology, 85,601-4521. BVVFLE, J. 1988. Complexation Reactions in Aquatic Systems. New York, John Wiley & Sons. CAROTHERS,W.W. & KHARAKA,Y.K. 1978. Aliphatic acid anions in oil-field waters - Implications for origin of natural gas. American Association of Petroleum Geologists Bulletin, 62, 2441-2453. & 1980. Stable carbon isotopes of HCO3- in oil-field waters - Implications for the origin of

C02. Geochimica et Cosmochimica Acta, 44, 323-332. COLE, R.J., CROCI~,J.G. & KYLE, J.R. 1991. Biogeo-

chemical studies of gold in a placer deposit, Livengood, Alaska. US Geological Survey OpenFile Report 91-142. COVENEY,R.M., LEVENTHAL,J.S., GLASCOCK,M.D. & HATCH, J.R. 1987. Origins of metals and organic matter in the Mecca Quarry Shale Member and stratigraphically equivalent beds across the midwest. Economic Geology, 82,915-933. CROSSLY, L.J. 1991. Thermal degradation of aqueous oxalate species. Geochimica et Cosmochimica Acta, 55, 1515-1527. ~ ' , SURDAM,R.C. t~ LAHANN,R. 1986. Application of organic/inorganic diagenesis to porosity prediction. In: GAUTIER, D.L. (ed.) Roles of Organic Matter in Sedimentary Diagenesis, Society of Economic Paleontologists and Mineralogists Special Publications, 38,147-155. DAVIS, J.B. 1967. Petroleum microbiology. Elsevier Scientific Pub. Co., New York. DEGENS, E.T., HUNT, J.M., REUTER, J.H. & REED, W.E. 1964. Data on the distribution of amino acids and oxygen isotopes in petroleum brine waters of various geologic ages. Sedimentology, 3, 199-225. DEROBERTIS, A., DESTEFANO, C., RIGANO, C., SAMMARTANO,S. & SCARCELLA,R. 1985. Studies on acetate complexes. Part 1. formation of proton, alkali-metal, and alkaline-earth-metal ion complexes at several temperatures and ionic strengths. Journal of Chemical Research, 5, 42-43. DRUMMOND, S.E. & PALMER, D.A. 1986. Thermal decarboxylation of acetate. Part II. boundary conditions for the role of acetate in the primary migration of natural gas and the transportation of metals in hydrothermal systems. Geochimica et Cosmochimica Acta, 50,825-833. EHRLICH, H.L. 1990. Geomicrobiology. Marcel Dekker, Inc., New York. ELLIS, A.J. 1963. The ionization of acetic, propionic, n-butyric, and benzoic acid in water, from conductance measurements up to 225 ° C. Journal of the Chemical Society, 59, 2299-2310. FEIN, J.B. 1991a. Experimental study of aluminium-, calcium-, and magnesium-acetate compleTing at 80° C. Geochimica et Cosmochimica Acta, 55, 955-964. 1991b. Experimental study of aluminiumoxalate compleTing at 80° C" Implications for the

ORGANIC LIGANDS IN SUBSURFACE WATERS formation of secondary porosity within sedimentary reservoirs. Geology, 19, 1037-1040. FISHER, J.B. 1987. Distribution and occurrence of aliphatic acid anions in deep subsurface waters. Geochimica et Cosmochimica Acta, 51, 24592468. & BOLES, J.R. 1990. Water-rock interaction in Tertiary sandstones, San Joaquin basin, California, USA.: Diagenetic controls on water composition. Chemical Geology, 82, 83-101. FrrcH, A. & STEVENSON,F.J. 1983. Stability constants of metal-organic matter complexes: theoretical aspects and mathematical models. In: THEOPHRASTUS, S.A. (ed.) Significance of Trace Elements in Solving Petrogenetic Problems. Athens, National Technical University of Athens, 645-669. GAUTIER, D.L. 1986. Roles of Organic Matter in Sedimentary Diagenesis. Society of Economic Paleontologists and Mineralogists Special Publications, 38. --, KHARAKA, Y.K. & SURDAM, R.C. 1985. Re-

Valley-type ore genesis. 28th International Geological Congress Abstracts, 1,557-558. GRAUCH, R.I. & HUYCK, H.L.O. 1990. Metalliferous Black Shales and Related Ore Deposits, Proceed-

ings, 1989 United States Working Group Meeting, International Geological Correlation Program Project 254. United States Geological Survey

-

lationship of Organic Matter and Mineral Diagenesis. Society of Economic Paleontologists and Mineralogists Short Course No. 17. GERMANOV, A.I. & MEL'KANOVITSKAYA,S.G. 1975. Organic acids in hydrothermal ores of polymetallic deposits in groundwaters. Doklady Akedemii Nauk SSSR, 225, 182-184. GILL, G.A. & BRULAND, K.W. 1990. Mercury speciation in surface freshwater systems in California and other areas. Environmental Science and Technology, 24, 1392-1400. GIORDANO, T.H. 1985. A preliminary evaluation of organic ligands and metal-organic complexing in Mississippi Valley-type ore solutions. Economic Geology, 80, 96-106. 1989. Anglesite (PbSO4) solubility in acetate solutions: the determination of stability constants for lead acetate complexes to 85° C. Geochimica et Cosmochimica Acta, 53,359-366. 1990. Organic iigands and metal-organic complexing in ore fluids of sedimentary origin. United States Geological Survey Circular, 11158, 31--41. 1993. Metal transport in ore fluids by organic ligand complexation. In: LEWAN, M.D. & PITTMAN,E.D. (eds) The Role of Organic Acids in Geological Processes. Springer-Verlag, New York, (in press). & BARNES, H.L. 1981. Lead transport in Mississippi Valley-type ore solutions. Economic Geology, 76, 2200-2211. & DRUMMOND, S.E. 1991. The potentiometric determination of stability constants for zinc acetate complexes in aqueous solutions to 295 ° C. Geochimica et Cosmochimica Acta, 55, 24012415. GIZE, A.P. 1984. The organic geochemisty of three Mississippi Valley-type ore deposits. PhD thesis, the Pennsylvania State University. t~ BARNES,H.L. 1987. The Organic geochemistry of two Mississippi Valley-type lead-zinc deposits. Economic Geology, 82,457-470. 1989. Organic processes in Mississippi -

-

-

-

&

-

-

197

Circular, 1058. GUSTAFSON, L.B. & WILmAMS, N. 1981. Sedimenthosted stratiform deposits of copper, lead, and zinc. In: SKINNER, B.J. (ed.) Economic Geology 75th Anniversary Volume, 139-178. HABERSrROH, P.R. & KARL, D.M. 1989. Dissolved free amino acids in hydrothermal vent habitats of the Guaymas basin. Geochmica et Cosmochimica Acta, 53, 2937-2945. HANOR, J.S. 1987. Origin and migration of subsurface sedimentary brines. Society of Economic Paleontologists and Mineralogists Short Course No. 21. & WORKMAN, A.L. 1986. Distribution of dissolved volatile fatty acids in some Louisiana oil-field brines. Applied Geochemistry, 1, 37-46. HANSLEY,P.L. & SPIRAKIS,C.S. 1992. Organic matter diagenesis as the key to a unifying theory for the genesis of tabular uranium-vanadium deposits in the Morrison Formation, Colorado Plateau. Economic Geology, 87,352-365. HARRISON, W.J. & THYNE, G.D. 1992. Predictions of diagenetic reactions in the presence of organic acids. Geochimica et Cosmochimica Acta, 565-586. HELGESON, H.C. 1967. Thermodynamics of complex dissociation in aqueous solution at elevated temperatures. Journal of Physical Chemistry, 71, 3121-3136. -1969. Thermodynamics of hydrothermal systems at elevated temperatures and pressures. American Journal of Science, 267,729-804. - - , KIRKHAM,D.H. & FLOWERS,G.C. 1981. Theoretical prediction of the thermodynamic behavior of aqueous electrolytes at high pressures and temperatures: IV. Calculation of activity coefficients, osmotic coefficients, and apparent molal and standard and relative molal properties to 600 ° C and 5 kb. American Journal of Science, 281, 1249-1516. HENNET, R.J.C., CRERAR, D.A. & SCHWARTZ, J. 1988a. Organic complexes in hydrothermal systems. Economic Geology, 83,742-764. -- & ~ 1988b. The effect of carbon dioxide partial pressure on metal transport in lowtemperature hydrothermal systems. Chemical Geology, 69,321-330. HENRY, A.L., ANDERSON, G.M. & HEROUX, Y. 1992. Alteration of organic matter in the Viburnum Trend lead-zinc district of southeast Missouri. Economic Geology, 87,288--309. HEROUX, Y., MICHOUX, D., DESJARDINS, M. & SANGASTER,D. 1989. Petrographic et geochimie des matieres organiques des sequences plombo zinciferes d'age Carbonifere, Bassin Salmon River, Nouvelle - Ecosse, Canada. Organic Geochemistry, 14,253-268.

-

-

56,

198

T.H. GIORDANO & Y.K. K H A R A K A

Ho, E.S., MEYERS, P.A. & MAUK, J.L. 1990. Organic geochemical study of mineralization in the Keweenawan Nonesuch Formation at White Pine, Michigan. Organic Geochemistry, 16, 229234. HUNT, J.H. 1979. Petroleum Geochemistry and Geology. Freeman and Co., San Francisco. JACKSON, K.S., JONASSON,I.R. & SKIPPEN,G.B. 1978. The nature of metals-sediment-water interactions in freshwater bodies, with emphasis on the role of organic matter. Earth Science Reviews, 14, 97-146. JOHNSON, J.W., OELKERS, F.H. & HELGESON, H.C. 1992. SUPCRT92: A software package for calculating the standard molal thermodynamic properties of minerals, gases, aqueous species, and reactions from 1 to 5000 bars and 0 to 1000° C. Computers & Geosciences, 18,899-947. JONES, I. & SORER, F.G. 1936. The effect of temperature on the ionization constants of some dibasic acids. Journal of the Chemical Society (London), 133-137. KARTSEV, A.A. 1974. Hydrogeology of oil and gas deposits. National Technical Information Service Report. TT73-58022. KAWAMURA, K., TANNENBAUM,E., HUIZINGA, B. & KAPALAN, I.R. 1986. Volatile organic acids generated from kerogen during laboratory heating. GeochemicalJournal, 20, 51-59. KETTLER, R.M., PALMER,D.A. & WESOLOWSKI,D.J. 1991. Dissociation quotients of malonic acid in aqueous sodium chloride media to 100° C. Journal of Solution Chemistry, 21,883-900. --, WESOLOWSKI, D.J. & PALMER, D.A. 1992. Dissociation quotients of oxalic acid in aqueous sodium chloride media to 175°C. Journal of Solution Chemistry, 20,905-927. KHARAKA, Y.K. 1986. Origin and evolution of water and solutes in sedimentary basins. In: HITCHON, B., BACHU, S. • SAUVEPLANE,M. (eds) Hydro-

D.F. 1986. Role of organic species dissolved in formation waters from sedimentary basins in mineral diagenesis. In: GAUTIER,D.L. (ed.) Roles

of Organic Matter in Sedimentary Diagenesis. Society of Economic Paleontologists and Mineralogists Special Publications, 38, 111-122. , LIco, M.S., WRIGHT,V.A. & CAROTHERS,W.W. 1979. Geochemistry of formation waters from Pleasant Bayou No. 2 well and adjacent areas in coastal Texas. Proceedings of the Fourth U.S.

Gulf Coast Geopressured-Geothermal Energy Conference: Research and Development, University of Texas at Austin, 168-199. --,

LUNDEGARD,P.D., AMBATS, GIL, EVANS, W.C. & BISCHOFF, J.L. 1993a. Generation of aliphatic acid anions and carbon dioxide by hydrous pyrolysis of crude oils. Applied Geochemistry, 8, 317-324. --, MAEST, A.S., CAROTHERS,W.W., LAW, L.M., LAMOTHE,P.J. & FRIES, T.L. 1987. Geochemistry of metal-rich brines from central Mississippi Salt Dome basin, U.S.A. Applied Geochemistry, 2, 543-561. KIENE, R.P. & TAYLOR, B.F. 1989. Metabolism of acrylate and 3-mercaptopropionate. In: SALTZMAN, E.S. & COOPER, W.J. (eds) Biogenic Sulfur in the Environment. American Chemical Society Symposium Series, 393,222-230. KOTOVA, A.V., BAKIROVA,S.F., YAG'YAYERA,S.M. & LEONOV, I.D. 1987. Effects of hydrogen sulfide on petroleum porphyrin complexes. Geochemistry International, 24, I 11-116. KRIBEK, B., KAIGL, J. & ORUZINSKY, V. 1977. Characteristics of di and trivalent metal-humic acid complexes on the basis of their molecularweight distribution. Chemical Geology 19, 73-81. KVENVOLDEN,K.A. & WEISER,D. 1967. A mathematical model of a geochemical process: normal paraffin formation from normal fatty acids. Geochimica at Cosmochimica Acta, 31, 12811309. geology of Sedimentary Basins: Application to Exploration and Exploitation. National Water LANDAIS, P. & MEYER, A.J. 1989. Temperature evolution assessment and indirect dating of Well Association, Dublin, OH, 173-195. mineralization in a sedimentary ore deposit: --, AMBATS, G. & THORDSEN, J.J. 1993b. Distriorganic matter, fluid inclusions, fission track, and bution and significance of dicarboxylic acid anions computer modeling. 28th International Geologiin oil-field waters. Chemical Geology, 107, 499cal Congress Abstracts, 2,256. 501. LANGMUIR,D. 1979. Techniques of estimating thermo, CAROTHERS,W.W. & ROSENBAUER, R.J. 1983. dynamic properties for some aqueous complexes Thermal decarboxylation of acetic acid: Impliof geochemical interests. In: JENNE, E.A. (ed.) cations for origin of natural gas. Geochimica et Chemical Modeling in Aqueous Systems. AmeriCosmochimica Acta, 47,397-402. can Chemical Society Symposium Series, 93, , GUNTER, W.D., AGGARUAL, P.K., PERKINS, E.H. & DEDRAAL,J.D. 1988. SOLMINEQ.88, A 353-387. computer program for geochemical modelling of LEENHEER,J.A., MALCOLM,R.L., MCKINLEY,P.W. & ECCLES, L.A. 1974. Occurrence of dissolved water-rock interactions. United States Geological organic carbon in selected groundwater samples Survey Water-Resources Investigation Report in the United States. U.S. Geological Survey 88--4227. Journal of Research, 2,361-369. - - , HULL, R.W. & CAROTHERS,W.W. 1985. WaterLEVENTHAL,J.S. 1986. Roles of organic matter in ore rock interactions in sedimentary basins. In: deposits. In: DEAN, W.E. (ed.) Organics and Ore GAUTIER, D.L., KHARAKA,Y.K. & SURDAM,R.C. Deposits. Proceedings of the Denver Region (eds) Relationship of Organic Matter and Mineral Exploration Geologists Society Symposium on Diagenesis. Society of Economic Paleontologists Organics and Ore Deposits, 7-20. and Mineralogists Short Courses, 17, 79-176. 1990. Organic matter and thermochemical sulfate --, LAW, L.M., CAROTHERS,W.W. & GOERLITZ,

ORGANIC LIGANDS IN SUBSURFACE WATERS reduction in the Viburnum Trend, Southeast Missouri. Economic Geology, 85,622-632. LEWAN, M.D. 1984. Factors controlling the proportionality of vanadium and nickel in crude oils. Geochimica et Cosmochimica Acta, 48, 22312238. LEWAN, M.D. & MAYNARD, J.B. 1982. Factors controlling enrichment of vanadium and nickel in the bitumen of organic sedimentary rocks. Geochimica et Cosmochimica Acta, 46, 2547-2560. Lico, M.S., KHARAKA,Y.K., CAROTHERS, W.W. & WRIGHT, V.A. 1982. Methods for collection and analysis of geopressured geothermal and oil-field waters. United States Geological Survey WaterSupply Paper 2194. LIPINER, G., WILLNER, I. & AIZENSHTAT, Z. 1988. Correlation between geochemical environments and controlling factors in the metallation of porphyrins. Organic Geochemistry, 13,747-756. LINDSAY, W.T. 1980. Estimation of concentration quotients for ionic equilibria in high temperature water: The model substance approach. In: Proceedings International Water Conference. Engineering Society of Western Pennsylvania, 41, 284-294. LoCHTE, H.L., BURNAM, C.W. & MEYER, H.W.H. 1949. Organic acids produced by high pressure gas wells. Petroleum Engineer, 21, C34-C40. LUNDEGARD, P.D. & KHARAKA, Y.K. 1990. Geochemistry of organic acids in subsurface waters. In: MELCHIAR, D.C. & BASSETr, R.L. (eds) Chemical Modeling of Aqueous Systems H. American Chemical Society Symposium Series, 416,170-189. & - 1993. Distribution of organic acids in subsurface waters. In: LEWAN, M.D. & PITMAN, E.D. (eds) The Role of Organic Acids in Geological Processes. Springer-Verlag (in press). & LAND, L.S. 1986. Carbon dioxide and organic acids: Their role in porosity enhancement and cementation, Paleogene of the Texas Gulf coast. In: GAUTIER, D.L. (ed.) Roles of Organic Matter in Sedimentary Diagenesis. Society of Economic Paleontologists and Mineralogists Special Publications, 38, 129-146. & -1989. Carbonate equilibria and pH buffering by organic acids - Response to changes in PCO2. Chemical Geology, 74, 277-287. -& SENFTLE, J.T. 1987. Hydrous pyrolysis: a tool for the study of organic acid synthesis. Applied Geochemistry, 2,605-612. & TREVENA, A.S. 1990. Sandstone diagenesis in the Pattani Basin (Gulf of Thailand): history of water-rock interaction and comparison with the Gulf of Mexico. Applied Geochemistry, 5, 669685. MACGOWAN, D.B. & SURDAM, R.C. 1988. Difunctional carboxylic acid anions in oilfield waters. Organic Geochemistry, 12,245-259. & - 1990a. Importance of organic-inorganic reactions to modeling water-rock interactions during progressive clastic diagenesis. In: MELCHIAR,D.C. & BASSETT,R.L. (eds) Chemical Modeling of Aqueous Systems H. American

199

Chemical Society Symposium Series, 416, 494507. & 1990b. Carboxylic acid anions in formation waters, San Joaquin Basin and Louisiana Gulf Coast, U.S.A. - implications for clastic diagenesis. Applied Geochemistry, 5,687-701. MACQUEEN, R.W. 1986. Origin of Mississippi Valleytype lead zinc ores by organic matter-sulfate reactions: The Pine Point example. In: DEAN, W.E. (ed.) Organics and Ore Deposits. Proceedings of the Denver Region Exploration Geologists Society Symposium on Organics and Ore Deposits, 151-158. & POWELL, T.G. 1983. Organic geochemistry of the Pine Point lead-zinc ore field and region, Northwest Territories, Canada. Economic Geology, 78, 1-25. MANNING, D.A.C. 1986. Assessment of the role of organic matter in ore transport processes in low-temperature base-metal systems. Transactions of the Institution of Mining and Metallurgy, B95, B195-B200. MARSHALL, W.L. 1970. Complete equilibrium constants, electrolyte equilibria, and reaction rates. Journal of Physical Chemistry, 74,346-355. MARTELL, A.E. & SMITH, R.M. 1974. Critical Stability Constants, vol.1, Amino Acids. Plenum, New York. -&~ 1977. Critical Stability Constants, v. 3, Other Organic Ligands. Plenum, New York. -& -1982. Critical Stability Constants, v. 5, First Supplement. Plenum, New York. MAYNARD, J.B. 1991a Uranium: syngenetic to diagenetic deposits in foreland basins. In: FORCE, E.R., EIDEL, J.J. & MAYNARD, J.B. (eds) Sedimentary and Diagenetic Mineral Deposits: A Basin Analysis Approach to Exploration. Society of Economic Geologists Reviews in Economic Geology, 5,187-197. -1991b. Copper: product of diagenesis in rifted basins. In: FORCE, E.R., EIDEL, J.J. & MAYNARD, J.B. (eds) Sedimentary and Diagenetic Mineral Deposits: A Basin Analysis Approach to Exploration. Society of Economic Geologists Reviews in Economic Geology, 5,199-207. MEANS, J.L. & HUBBARD, N. 1987. Short-chain aliphatic acid anions in deep subsurface brines: A review of their origin, occurrence, properties, and importance and new data on their distribution and geochemical implications in the PaiD Duro Basin, Texas. Organic Geochemistry, 11, 177-191. MESMER, R.E., MARSHALL, W.L., PALMER, D.A., SIMENSON, J.M. & HOLMES, H.F. 1988. Thermodynamics of aqueous association and ionization reactions at high temperatures and pressures. Journal of Solution Chemistry, 17,699-718. - - , PATTERSON,C.S., BUSEY,R.H. & HOLMES, H.F. 1989. Ionization of acetic acid in NaCI (aq) media: A potentiometric study to 573 K. Journal of Physical Chemistry, 93, 7483-7490. MILLER, S.L. & BADA, J.L. 1988. Submarine Hot Springs and the origin of life. Nature, 334, 609-611.

200

T.H. G I O R D A N O & Y.K. K H A R A K A

MORRISON, R.T. & BOYD, R.N. 1966. Organic Chemistry. Boston, Allyn and Bacon, Inc. NASH, J.T., GRANGER, H.C. & ADAMS, S.S. 1981. Geology and concepts of genesis of important types of uranium deposits. In: SKINNER,B.J. (ed.) Economic Geology 75th Anniversary Volume, 63-116. NISSENBAUM, A. • SWAINE, D.J. 1976. Organic matter-metal interactions in recent sediments: the role of humic substances. Geochimica et Cosmochimica Acta, 40,809-816. OSCARSON, J.L., GILLESPIE, S.E., CHRISTENSEN,J.J., IzA-rr, R.M. & BROWN, P.R. 1988. Thermodynamic quantities for the interaction of H + and Na + with C2H302 and CI- in aqueous solutions from 275 to 320 ° C. Journal of Solution Chemistry, 17,865-885. PALMER, D.A. & BELL, J.L.S. 1993. Aluminium speciation and equilibria in aqueous solution: IV. A potentiometric study of aluminium acetate complexation in acidic NaCI brines to 150° C. Geochimica et Cosmochimica Acta, (in press). - - & DRUMMOND, S.E. 1986. Thermal decarboxyiation of acetate. Part I. the kinetics and mechanism of reaction in aqueous solution. Geochimica et Cosmochimia Acta, 50,813-823. - & -1988. Potentiometric determination of the molal formation constants of ferrous acetate complexes in aqueous solutions to high temperatures. Journal of Physical Chemistry, 92, 67956800. PARNELL, J., HUCItA, H. & LANDAIS, P. 1993. Bitumens in Ore Deposits. Springer-Verlag, New York. PERDUE,E.M. 1985. Acidic functional groups of humic substances. In: AIKEN, G.R., McKNIGHT, D.M. & WERSHAW, R.L. (eds) Humic Substances in Soil, Sediment, and Water. New York, John Wiley & Sons, 493-526. & LYTLE, C.R. 1986. Chemical equilibrium modeling of metal complexation by humic substances. In: CARLISLE,D., BERRY,W.L., KAPLAN, I.R. & WATrERSON,J.R. (eds) Minerals Exploration: Biological Systems and Organic Matter. Prentice-Hall, Englewood Cliffs, NJ, 428-444. PERRIN, D.D. 1979. Stability constants of metal-ion complexes part B, organic ligands. IUPAC chemical data series No. 22. Pergamon Press, Oxford. PINCHING, G.D. & BATES, R.G. 1950. Second dissociation constant of succinic acid from 0° to 50° C. National Bureau of Standards Journal of Research, 45,322-328. PRICE, P.E. & KYLE, J.R. 1986. Genesis of Salt Dome hosted metallic sulfide deposits: The role of hydrocarbons and related fluids. In: DEAN, W.E. (ed.) Organics and Ore Deposits. Proceedings of the Denver Region Exploration Geologists Society Symposium on Organics and Ore Deposits, 171-182. RAPP, J.B. 1976. Amino acids and gases in some springs and an oil field in California. Journal of Research, United States Geological Survey, 4, 227-232.

RASHID, M.A. 1985. Geochemistry of Marine Humic Compounds. Springer-Verlag, New York. REUTER, J.H. & PERDUE, E.M. 1977. Importance of heavy metal-organic matter interactions in natural waters. Geochimica et Cosmochimica Acta, 41,325-334. RIPELY, E.M., SHAFFER,N.R. & GILSTRAP,M.S. 1990. Distribution and geochemical characteristics of metal enrichment in the New Albany Shale (Devonian-Mississippi), Indiana. Economic Geology, 85, 1790-1807. ROSIN, P.L. & ROtJXHET, P.G. 1978. Characterization of kerogens and study of their evolution by infrared spectroscopy: carbonyl and carboxyl groups. Geochimica et Cosmochimica Acta, 42, 1341-1349. ROSE, A.W. 1976. The effect of cuprous chloride complexes in the origin of red-bed copper and related deposits. Economic Geology, 71, 10361048. 1989. Mobility of copper and other heavy metals in sedimentary environments. In: BOYLE, R.W., BROWN, A.C., JEFFERSON,C.W., JOWETT,E.C. & KIRKHAM,R.V. (eds) Sediment-Hosted Stratiform Copper Deposits. Geological Association of Canada Special Papers, 36, 97-110. SAXBY,J.D. 1976. The significance of organic matter in ore genesis. In: WOLF, K.H. (ed.) Handbook of Strata-Bound and Stratiform Ore Deposits. Elsevier, New York, 1, 111-133. SCHNrrZER, M. & KHAN, S.U. 1978. Soil Organic Matter. Elsevier, New York. SCHULTZ, R.B. 1991. Metalliferous black shales: accumulation of carbon and metals in cratonic basins. In: FORCE, E.R., EIDEL, J.J. & MAYNARD, J.B. (eds) Sedimentary and Diagenetic Mineral Deposits: A Basin Analysis Approach to Exploration. Society of Economic Geologists Reviews in Economic Geology, 5,171-176. SEEWALD,J.S. & SEYFRIED,W.E. 1991. Experimental determination of portlandite solubility in H20 and acetate solutions at 100-350 ° C and 500 bars: constraints on calcium hydroxide and calcium acetate complex stability. Geochimica et Cosmochimica Acta, 5 5 , 6 5 9 - 6 6 9 . SEIFERT, W.K. 1975. Carboxylic acids in petroleum and sediments. Progress in the Chemistry of Organic Natural Products, 32, 1-49. SEVERSON, R.C., CROCK, J.G. & MCCONNELL, B.M. 1986. Processes in the formation of crystalline gold in placers. In: DEAN, W.E. (ed.) Organics and Ore Deposits. Proceedings of the Denver Region Exploration Geologists Society Symposium on Organics and Ore Deposits, 69-80. SHOCK, E.L. 1988. Organic acid metastability in sedimentary basins. Geology, 16,886-890. 1989. Corrections to 'organic acid metastability in sedimentary basins'. Geology, 17,572-573. 1990. Do amino acids equilibrate in hydrothermal fluids? Geochimica et Cosmochimica Acta, 54, 1185-1189. - & HELGESON, H.C. 1990. Calculation of the thermodynamic and transport properties of aqueous species at high pressures and temperatures:

ORGANIC LIGANDS IN SUBSURFACE WATERS standard partial molal properties of organic species. Geochimica et Cosmochimica Acta, 54, 915-945. , OELKERS, E.H., JOHNSON, J.W., SVERJENSKY, D.A. & HELGESON,H.C. 1992. Calculation of the thermodynamic properties of aqueous species at high pressures and temperatures: effective electrostatic radii, dissociation constants, and standard partial molal properties to 1000° C and 5 kb. Journal of the Chemical Society (London). Faraday Transactions, 88, 803-826. Surrn, R.M. & MARTELL,A.E. 1975. Critical Stability Constants, v. 2, Amines. Plenum, New York. -& -1989. Critical Stability Constants, v. 6, Second Supplement. Plenum, New York. STEVENSON, F.J. 1983. Trace metal-organic matter interactions in geologic environments. In: THEOPHRASTUS, S.A. (ed.) Significance of Trace Elements in Solving Petrogenetic Problems. National Technical University of Athens, 671691. STUMM,W. & MORGAN,J.J. 1981. Aquatic Chemistry. John Wiley & Sons, New York. SURDAM, R.C. & CROSSLY, L.J. 1987. Integrated diagenetic modeling: A process-oriented approach for clastic systems. Annual Review of Earth and Planetary Science, 15,141-170. --, BOESE, S.W. t~ CROSSEY, L.J. 1984. The chemistry of secondary porosity. In: McDONALD, D.A. & SURDAM, R.C. (eds) Clastic Diagenesis. American Association of Petroleum Geologists Memoirs, 37,127-151. SVERJENSKY,D.A. 1986. Genesis of Mississippi Valleytype lead-zinc deposits. Annual Review of Earth and Planetary Science, 14, 177-199. 1987. The role of migrating oil fields brines in the formation of sediment-hosted Cu-rich deposits. Economic Geology, 82, 1130-1141. 1989. Chemical evolution of basinal brines that formed sediment-hosted Cu-Pb-Zn deposits. In: BOYLE, R.W., BROWN, A.C., JEFFERSON, C.W., JOWETT, E.C. & KIRKHAM,R.V. (eds) SedimentHosted Stratiform Copper Deposits. Geological Association of Canada Special Papers, 36, 127134. TANGER, J.C. IV & HELGESON, H.C. 1988. Calculation of the thermodynamic and transport properties of aqueous species at high pressures and temperatures: revised equations of state for the standard partial raDial properties of ions and electrolytes. American Journal of Science, 288, 19-98. THURMAN, E.M. 1985. Organic Geochemistry of Natural Waters. Martinus Nijhoff, Boston. THYNE, G.D., HARRISON, W.J. 8~ ALLOWAY,M.D. 1992. Experimental study of the stability of the Al-oxalate complexation at 100°C and calculations of the effects of complexation on clastic diagenesis. In: KHARAKA, Y.K. & MAEST, A.S. (eds) Water-Rock Interaction. A.A. Balkema, Rotterdam, 1,353-357. TIPPING, E. 8/; HURLEY, M.A. 1992. A unifying model of cation binding by humic substances. Geochimica et Cosmochimica Acta, 56, 3627-3641. -

-

-

-

201

TlSSOT, B.P. & WELTE, D.H. 1984. Petroleum Formation and Occurrence. Springer-Verlag, New York. TOURTELOT, H.A. 1979. Black shale-its deposition and diagenesis. Clays and Clay Minerals, 27, 313-321. TRUDINGER, P.A. 1976. Microbiological processes in relation to ore genesis. In: WOLF, K.H. (ed.) Handbook of Strata-Bound and Stratiform Ore Deposits. Elsevier, New York, 2,135-190. TURNER-PETERSON, C.E. & FISHMAN, N.S. 1986. Geologic synthesis and genetic models for uranium mineralization in the Morrison formation, Grants Uranium Region, New Mexico. In: TURNER-PETERSON, C.E., SANTOS, E.S. & FISHMAN,N.S. (eds) A Basin Analysis Case Study: The Morrison Formation, Grants Uranium Region, New Mexico. American Association of Petroleum Geologists, Studies in Geology Series, 22,357-388. - - , HATCHER, P.G. & SPIKER, E.C. 1986. Origin of organic matter in sandstone uranium deposits of the Morrison Formation, New Mexico: geologic and chemical constraints. In: DEAN, W.E. (ed.) Organics and Ore Deposits. Proceedings of the Denver Region of Exploration Geologists Symposium on Organics and Ore Deposits, 185-195. VAIRAVAMURTHY,A. • MOPPER, K. 1989. Mechanistic studies of organosulfur (thiol) formation in coastal marine sediments. In: SALTZMAN,E.S. & COOPER, W.J. (eds) Biogenic Sulfur in the Environment. American Chemical Society Symposium Series, 393,231-242. VARSHAL, G.M., VELYNKHANOVA,T. & BARANOVA,N. 1984. Geochemical role of gold (III) fulvate complexes. Geochemistry International, 21,139145. VEITCH, J.D. & MCLEROY, D.G. 1972. Organic mobilization of ore metals in low-temperature carbonate environments. Geological Society of America Abstracts with Programs, 7, 110-111. VINE, J.D. & TOURTELOT,E.B. 1970. Geochemistry of black shale deposits - a summary report. Economic Geology, 65,253-272. VLASSOPOULOUS,D., WOOD, S.A. t~zMUCCI, A. 1990. Gold speciation in natural waters: II. the importance of organic complexing - experiments with some simple model ligands. Geochimica et Cosmochimica Acta, 54, 1575-1586. WEAST, R.C. 1977. Handbook of Chemistry and Physics, 58th ed. CRC press, Cleveland, Ohio. WHITE, R.H. 1984. Hydrolytic stability of biDmolecules at high temperatures and its implication for life at 250 ° C. Nature, 310,430-432. WILLEY, L.M., KHARAKA,Y.K., PRESSER,T.S., RAPP, J.B. & BARNES,I. 1975. Short chain alipha'ic acid anions in oil field waters and their contribution to the measured alkalinity. Geochimica et Cosmochimica Acta, 39, 1707-1711. WOOD, S.A. 1990. The interaction of dissolved Pt with fulvic acid and simple organic acid analogues in aqueous solutions. Canadian Mineralogist, 28, 665-673.

202

T.H. G I O R D A N O & Y.K. K H A R A K A 1991. The interaction of Pd 2÷ with fulvic acid and simple organic acids - solubility and spectroscopic studies. Geological Society of America Abstracts with Programs, 23, A214.

ZINCER, A.S. & KRAVCHIK,T.E. 1973. Simpler organic acids in groundwater of the lower Volga region (genesis and possible use in prospecting for oil). Akademiya Nauk SSSR Doklady, 202,218-221.

Origin and nature of trace element species in crude oils, bitumens and kerogens: implications for correlation and other geochemical studies ROYSTON

H. FILBY

Department of Chemistry, Washington State University, Pullman, WA 99164-4630, USA Abstract: Crude oils and bitumens contain trace elements whose abundances range from

less than a part per billion (ppb) to more than several thousand parts per million (ppm). Nickel and vanadium are usually the most abundant metals with values ranging up to 200 ppm Ni and 2000 ppm V. Abundances of other trace elements are generally less than 100 ppm. Nickel and vanadium occur in oils and source-rock bitumens as metalloporphyrins and non-porphyrin species. The nickel and vanadium porphyrins are predominantly DPEP or etio complexes of Ni 2÷ or VO 2÷ although smaller amounts of benzo and other complex porphyrins are found. The metalioporphyrins are formed during early diagenesis of source rocks and the relative abundances of nickel and vanadium are related to the depositional environment. The origin of the non-porphyrin nickel and vanadium complexes and the complexes of other trace elements in crude oils may be primary or secondary. Primary processes include release of metal complexes from kerogen during categenesis and formation of metal-organic complexes by mineral-kerogen reactions during catagenesis. Secondary processes such as interaction of oils with mineral matter or formation waters during migration, maturation or biodegradation may also affect the concentrations of some elements in crude oils, e.g. Hg.

Crude oils and source-rock bitumens* consist primarily of hydrocarbons and other nonhydrocarbon organic compounds with minor amounts (generally less than 1%) of inorganic constituents, including organic complexes of metal ions. The inorganic component of a crude oil is important for two rzasons. In the first case, small amounts of metal-organic species, particularly nickel, vanadium and arsenic compounds, have deleterious effects during processing and upgrading of crude oils, particularly through poisoning of and build-up of metal deposits on catalysts used for cracking, reforming and hydrogenation. Secondly, inorganic or metalorganic species may b,z used as geochemical biomarkers in oil--oil or oil-source rock correlation, identification of depositional environment of source rocks and in the determination of thermal maturity levels of oils or source rocks. The geochemical information provided by most biomarkers is contained in the skeletal structure of the organic molecule and its :elationship to a biological precursor. Metal complexes may provide additional information, particularly on depositional or post-sedimentary environments, because the oxidation states of transition metal

* The term bitumen in this paper refers to solventextractable organic matter in a sedimentary rock

ions are indicators of redox conditions at the time of formation of the metal complex. The major types of metal species found in crude oils are shown in Table 1. Except for the contaminants introduced by drilling and production, each type of metal species may provide information on the origin of the oil, thermal maturation level, migration pathways, reservoir processes, alteration processes, and interaction with ore-forming fluids. Most geochemical information, however, has been provided by the discrete metal complexes, principally the geoporphyrins, found in crude oils and bitumens (Filby & Branthaver 1987). This review is concerned primarily with the first four types listed in Table 1. The abundances of trace elements in crude oils and source-rock bitumens vary widely depending on (a) the depositional environment of the source rock, (b) the nature of the organic material in the source rock, (c) the geological setting, (d) the maturity of the oil, (e) the diagenetic history of the source rock, and (f) the chemical composition of the oil. The effects of these factors on trace element abundances are, however, not sufficiently understood. Most condensate oils contain negligible amounts of trace elements (typically ppb level or lower) whereas, at the other extreme, asphaltic oils of low API gravity and source-rock bitumens which

FromPARNELL,J. (ed.), 1994, Geofluids:Origin, Migrationand Evolution of Fluidsin SedimentaryBasins, Geological Society Special Publication No. 78,203-219.

203

204

R.H. FILBY

Table 1. Types o f metal species found in crude oils and source rock bitumens Chemical/physical species

Elements found and compound types

Elemental species dissolved in oil or associated with polar constituents Discrete and extractable metal-organic complexes

S°, Se°,Hg °

Organometallic complexes with metal-carbon bonds Metal complexes or metal ions associated with or bonded in polar components of oils Entrained formation waters Entrained mineral matter Drilling/production contaminants

Ni, V, (Fe), porphyrin complexes; metal chlorins; metal naphthenates methyl and phenyl arsonic acids; (methyl mercury compounds?) Ni, V, Fe and other transition metal ions in asphaltene moieties; As, Se, possibly in molecular structure Na ÷, K +, Ca > , Mg2+, I-, CI-, Br-, 8042AI, K, Si, Mg, Ca, Na in clay minerals; trace elements in clay minerals, e.g. Sc, REE Ba as BaSO4 in drilling muds; As in bactericides; Hg in oils flashed using liquid Hg

Table 2. Variation of nickel, vanadium and iron abundances in oils and bitumens with source-rock type and depositional environment

Oil type and location Condensate; Kapuni, Taranaki Basin, NZ (CretaceousCenozoic) (ref. 1) Light oil (non marine); Tariki, Taranaki Basin, NZ (Cretaceous-Cenozoic) (ref. 1) Crude oil (marine); Gilby, Mannville Group, Alberta Basin (L. Cretaceous) (ref. 2) Heavy biodegraded oil (marine); Boscan, Venezuela (Cretaceous) (ref. 3) Bitumen, New Albany shale, Indiana (marine; MississippianDevonian) (ref. 3) Bitumen, Green River shale Mahogany Zone, CO, (lacustrine; Eocene) (ref. 4)

Ni(ppm)

<1 1.2

V(ppm)

<1

0.4

Ni!V

Fe(ppm)

-

6

3.0

0 0.8

11

32

2.9

97

1200

12.4

16

2560

2700

1.1

353

59.1

41.2

1.4

79.0

References: 1. Frankenberger et al. (1993); 2. Hitchon & Filby (1983); 3. Mercer et al. (1992); 4. Van Berkel (1987)

are rich in asphaltenes and other polar constituents generally contain the highest concentrations of trace metals, particularly nickel and vanadium (see Table 2). In general, oils of marine origin have higher concentrations of trace elements than oils derived from terrestrial sources (Barwise 1990). Nickel and vanadium (and for some oils, iron) are typically the most abundant metals in crude oils and related source-rock bitumens and their concentrations may range from less than 1 ppb in light condensate oils to greater than 2000ppm V and

200 ppm Ni in asphaltic, low API gravity, oils (Filby & Branthaver 1987), with even higher values in some source-rock bitumens (Filby & Van Berkel 1987; Mercer et al. 1992). Abundances of other trace elements are generally much lower than for nickel and vanadium (or iron) and large variations in concentration for a given element (and from element to element) occur, even among oils from a given depositional basin. These variations in concentration among oils from a given basin are illustrated for the West Canada Basin in Table 3 (Hitchon & Filby 1983;

TRACE ELEMENTS IN CRUDE OILS Table 3. Mean concentrations and ranges (ppm) of selected trace elements in crude oils from the West Canada Basin (Hitchon & Filby 1983)

Element S(%) V Ni Fe Co Cr Mn Zn As Sb Se Hg Na CI Br I

Mean concentration (ppm) 0.83 13.6 9.38 10.8 0.054 0.093 0.010 0.459 0.111 0.006 0.052 0.051 3.62 39.3 0.491 0.719

Range (ppm) 0.05-3.9 0.1-177 0.1-74.1 0.1-254 0.0002-2.0 0.005-1.68 0.003-3.85 0.025-5.92 0.002-1.99 0.0001-0.035 0.003-0.511 0.002--0.399 0.01-64.7 0.1-1010 0.002-12.5 0.01-9.0

Hitchon et al. 1975). Oils analysed for this basin ranged from paraffinic oils of generally low trace element content to heavy, highly biodegraded, asphaltic oils containing high concentrations of nickel, vanadium and other metals. Source-rock bitumens also contain suites of trace elements similar to those found in crude oils, as might be expected from the genetic relationship between a migrated crude oil and the bitumen generated in the original source rock. Nickel and vanadium are typically the most abundant trace elements in bitumens, as for crude oils, and a wide variation in composition is also seen depending on the nature of the source rock, the parent kerogen and the degree of maturity (Filby & Van Berkel 1987; Odermatt & Curiale 1991). Source-rock bitumens are generally enriched in asphaltenes and other polar compounds compared to their corresponding crude oils, and this is reflected in the relatively large abundances of nickel and vanadium in source rock bitumens. If it is assumed that catagenesis of kerogen is the source of an evolving bitumen in a source rock (Tissot & Welte 1984), and thus of an expelled crude oil, it might be expected that the kerogen contains a trace element fingerprint which is related to that of the evolving bitumen. Unfortunately kerogens isolated from source rocks by dissolution of the silicate-carbonate matrix by hydrochloric-hydrofluoric acid digestion contain residual minerals that are insoluble in this acid mixture. Pyrite, marcasite (FeS2), chalcopyrite

205

(CuFeS2), zircon (ZrSiO4) and rutile (TiO2) are the major minerals that are insoluble in HC1-HF and they cannot be completely removed from the isolated kerogen without recourse to chemical procedures which may change kerogen composition. Thus the trace element abundances in the in situ kerogen can only be determined reliably if corrections are made for contributions from the residual mineral fraction. These corrections are large for most elements, except for nickel and vanadium which are normally present at low concentrations in the residual minerals and for which the mineral contributions are usually small (Mercer et al. 1991, 1993). Table 4 shows trace element data for some kerogens (corrected for mineral contributions) and the associated bitumens isolated from Mississippian-Devonian shales (New Albany and Woodford) and the Eocene Green River shale. It is evident that there is an approximate correlation between the nickel and vanadium concentrations in the bitumen and those in the corresponding kerogen. It should be noted that for all shales the Ni/V ratio in the kerogen is higher than that in the associated bitumen. This phenomenon is also observed for the bitumenkerogen fractions isolated from the Monterey shale by Odermatt & Curiale (1991) and has not been adequately explained.

The nature of metal species in oils and bitumens The geochemical origin of metal species in crude oils and source-rock bitumens probably involves both primary and secondary processes. Primary processes can be defined as those that occur in the source rock during the generation of the bitumen. Secondary processes include migration of the oil from the source rock (primary migration), migration to a trap (secondary migration), maturation in the reservoir, biodegradation, and interaction with mineral deposits, metal-rich fluids, or formation waters. These secondary processes may modify existing metal species in an oil or introduce new complexes. The composition and structure of metal complexes in crude oils, their mode of origin, and the influence of secondary processes (e.g. migration, maturation, biodegradation) on metal species must be known in order to use these compounds as biomarkers in oil-oil and oil-source rock correlations and to characterize depositional environments of source rocks (Branthaver & Filby 1987). Unfortunately, the chemical species of many trace elements in crude

206

R.H. FILBY

Table 4. Trace element concentrations (in ppm) in kerogens and bitumens of the New Albany, Woodford and Green River Shales

Element V Ni Ni/V Fe Co Cr Mn Zn As Se Mo

NewAlbany kerogen**

NewAlbany bitumen*

Woodford kerogent*

Woodford bitumen*

Green River kerogen**

GreenRiver bitumen*

1130 2160 1.91 115 71.9 36.4 175 2400

2700 2560 0.95 353 8.63 1.04 0.32 65.7 75.8 79.5 28.7

3260 353 0.11 3.92 41.6 7.29 215 215

5440 335 0.06 606 10.5 5.00 18.3 811 5.36 15.8 15.8

28.3 41.5 1.47 11.7 17.0 37.6 37.6

41.2 59.1 1.43 79.0 1.47 0.33 0.28 5.06 22.4 3.70 3.70

* Mercer etal. (1993) * Van Berkel (1987) * Concentrations in kerogens corrected for mineral content (Mercer et al. 1992; Van Berke11987)

oils have not been identified and their geochemical origin is thus impossible to determine. The most abundant metal complexes conclusively identified in crude oils and bitumens are the metalloporphyrins in which nickel and vanadium are the chelated metals (Baker & Louda 1986; Filby & Van Berkel 1987). These important biomarkers have been studied extensively and their geochemistry is sufficiently well understood for these compounds to be used in geochemical exploration.

The metalloporphyrins The metalloporFhyrins are found in crude oils, asphalts, carbonaceous sedimentary rocks and in some coals. In porphyrins isolated from crude oils and mature sediments, the metal ion is either Ni 2+ or VO 2+ (i.e. vanadyl ion) although Cu 2+ (Baker & Louda 1984; Verne-Mismer etal. 1990; Bennett et aI. 1993) and possibly Fe z+ (Eckardt et al. 1989) porphyrins have been identified in immature sediments. Since the isolation and identification of the geoporphyrins with chlorophyll a by Treibs (Treibs 1934), considerable progress has been made in elucidating the structure of the Ni(II) and VO(II) porphyrins found in crude oils and carbonaceous sediments, primarily through advances in mass spectrometric and N M R methods for porphyrin characterization. Structures of the major metalloporphyrin types which have been identified in crude oils and sediments are shown in Fig. 1. In most oils and sediments, the major structural types are metallated deoxophylloerythroetioporphyrin (DPEP) type 2, and etioporphyrin, type 1, although important amounts of

benzo, types 8 and 9, tetrahydrobenzo, types 6 and 7 and cycloalkanoporphyrins, e.g. types 3-5, may be present (Filby & Van Berkel 1987; Chicarelli et al. 1987; Keely et al. 1990; Callot et al. 1990; Lash 1993). For most of the porphyrin structural types shown in Fig. 1, the R-groups attached to the porphin macrocyclic ring may be H-, CH3-, CH3CH2-, or other alkyl groups, which gives rise to pseudo-homologous series within each structural type. In immature sediments, free-base porphyrins and metallated porphyrin acids (e.g. R 6 or R 7 = -CHzCOOH in etio structure type 1 (Fig. 1)) have also been identified (Baker & Louda 1986; Callot et al. 1990). The Treib's scheme of defunctionalization and metallation reactions for the conversion of chlorophyll a to the Ni(II) and VO(II) porphyrins in maturing sediments (Treibs 1934) is shown in Fig. 2 and is accepted to be essentially correct (Filby & Van Berkel 1987; Keely et al. 1990) although the sequence of reactions may not be the same as that postulated by Treibs in all depositional environments. In particular, the reaction pathways for the formation of the etio porphyrins from chlorophylls are not well established. It is now generally accepted that chlorophylls other than chlorophyll a, including chlorophyll b, cholorphylls c, the bacteriochlorophylls a--e and others may be important porphyrin precursors, particularly of the benzo, tetrahydrobenzo, cycloalkano (nonDPEP), and other minor porphyrins (Chicarelli et al. 1987; Callot et al. 1990) which have no obvious structural relationship to chlorophyll a. Although incompletely understood, the geochemical evolution of the macrocyclic porphyrin structure in organic sediments is well

TRACE ELEMENTS IN CRUDE OILS

207

Chlorophyll-a

DEMETALLATION

]

pheophytin-a

HYDROLYSIS

N

Z pyropheophorbide-a

/

!

[..,

coos

AROMATIZATION

coos

;9 Z

COOH

DECARBOXYLATION

CHELATION

¢0OH

( M = Ni2+,V O 2+)

Fig. 1. Major metalloporphyrin skeletal types found in crude oils and sediments (M = Ni 2+ or VO2+).

208

R.H. FILBY

R2

R

~

R5

R'

R3

R ~ R

Rs

"R8

R2

3

R-3

R2

#

s

4

6

5

R2

l

~

2

1

R2

R

R3

R2

\M /

,>~...R 3 R4

R5

R

7

~..R3 R~

8

9

Fig. 2. Modified Treibs' scheme for the evolution of the geoporphyrins (Treibs 1934; Filby & Van Berkel 1987).

documented, thus providing a framework for the use of the metalloporphyrins as geochemical biomarkers.

The non-porphyrin metal complexes The Ni(II) and VO(II) porphyrins account for only a fraction of the total nickel and vanadium present in most crude oils (Filby & Van Berkel 1987). The nature of the 'non-porphyrin' complexes of nickel and vanadium, which occur primarily in the asphaltic or polar fraction of the oil, has been investigated extensively because of

implications that these metal species poison or foul catalysts during upgrading of petroleum heavy ends (Branthaver & Filby 1987; Filby & Van Berkel 1987). However, despite this interest, no well-characterized non-porphyrin complexes of either metal have been identified. Attempts to separate and identify the nonporphyrin complexes from crude oils or residual oils by combined chromatographic/spectroscopic techniques have not been successful because both nickel and vanadium tend to be distributed among all, or most, of the fractions separated from a crude oil (Pearson & Green

TRACE ELEMENTS IN CRUDE OILS 1989, 1993). There is evidence that the nonporphyrin nickel and vanadium complexes may be, at least in part, metalloporphyrins which are strongly associated or incorporated into the asphaltenes in crude oils and which cannot be separated chromatographically or by solvent extraction. Tooulakou & Filby (1987), for example, showed that after separation of the asphaltenes in a 'porphyrin-free' state from the Athabasca bitumen vanadyl porphyrins are still associated with the asphaltenes. Exhaustive solvent extraction of these asphaltenes removed 47% of the 'non-porphyrin' vanadium in the asphaltenes as VO(II) porphyrins and the extracted porphyrins were similar in structural type and carbon number distribution to those isolated from the bitumen by conventional extraction. Goulon et al. (1984) used XANES spectroscopy to show that the ligand environment of vanadium in asphaltenes from the Boscan, Venezuela, oil is consistent with porphyrinic forms that are entrained in, or associated with, the asphaltenes. Non-porphyrin species that have been suggested include metal naphthenates (Fish et al. 1987), nickel and vanadium complexed into the asphaltene structure by heteroatom functionalities (Nguyen & Filby 1987) or metal species produced by decomposition of the metalloporphyrins during oil maturation. The latter possibility would be consistent with the fact that geologically stable nickel and vanadium complexes other than the metalloporphyrins have not been identified. Evidence that the non-porphyrin species are porphyrin degradation products is that the total nickel and vanadium contents of crude oils show similar dependence on depositional conditions (i.e. pH, Eh) as do Ni(II) and VO(II) porphyrin abundances (Barwise 1990). During maturation of a crude oil in a reservoir or a bitumen in a source rock, decomposition of the metalloporphyrins would lead to a free cation, e.g. Ni 2÷ or VO 2+, which would either be complexed by an asphaltene functionality, complexed by other polar molecules, e.g. naphthenic acids, or form a metal sulphide molecule that would be trapped in an asphaltene micelle. Metal ions complexed in this fashion would be distributed among polar components of the oil and hence among chromatographic factions which separate on the basis of molecular weight or asphaltene functionality. The observation that there is a correlation between vanadium and sulphur contents of heavy crude oils has been interpreted as evidence for vanadium complexes with sulphur-containing ligands present in resin or asphaltene components of crude oils (Tissot & Welte 1984). However, the vanadium-sulphur

209

correlation is probably the result of a depositional environment of the source rock which favours accumulation of organic matter rich in reduced sulphur, S(II), and which also favored the formation of vanadyl porphyrins over the corresponding nickel species (Lewan 1984). There is much less information on the chemical forms of elements other than nickel and vanadium because the low concentrations of most trace elements in crude oils and bitumens (see Table 3) makes speciation very difficult if not impossible. Contamination, particularly from formation waters or clay minerals from the source or reservoir rock, and from drilling fluids may also be significant (Olsen & Filby in prep.). However, in the case of mercury and selenium, elemental species have been identified in some oils and bitumens. Thus oils from the Cymric field in California which are associated with mercury mineralization contain elemental mercury (Filby 1975; Bailey etal. 1961) and selenium has recently been shown to occur partly as the SeS7 molecule in the New Albany shale bitumen (Mercer et al. 1992). In both of these cases, secondary processes may have resulted in the incorporation of these elemental species. The only documented identification of true organometallic (i.e. containing carbon-metal bonds) compounds in petroleum is that of the arsenic compounds, methyl- and phenylarsonic acids, in the bitumen of the Green River oil shale (Fish & Brinckman 1983; Fish et al. 1987). These compounds may have formed by bacterial action in the shale, or alternatively, may be the result of the reaction of kerogen catagenesis products with the mineral skutterudite, (Co,Fe,Ni)As4, and safflorite, (Co,Fe)As2, which have been identified in the shale by Jagarnathan et al. (1986). These authors explained the presence of organoarsenic compounds in retorted shale oil by the reaction of kerogen breakdown products with the arsenic-bearing minerals.

Evolution of the metal complexes The formation of the metal species found in crude oils (other than entrained mineral matter or formation waters) is related to the mode of origin of the oil. It is generally accepted (Tissot & Welte 1984) that petroleum forms from organic matter preserved in sediments deposited from marine or lacustrine environments. Diagenesis of this organic matter results in a kerogen which undergoes catagenesis with increasing burial and temperature to form a bitumen, which if formed in large enough amounts may migrate from the source rock and accumulate as a crude oil in a reservoir. During

210

R.H. FILBY

the formation of petroleum, metal species can form by a number of primary process, including (a) accumulation in the evolving bitumen of metal complexes which may have formed in the sediment during early diagenesis; (b) release of metal species in the kerogen to the bitumen during catagenesis, or (c) formation of metal complexes via kerogen-mineral interactions in the source rock during catagenesis. Mechanisms (a) and (b) appear to be the major mechanisms involved in the origin of the geoporphyrins (Baker & Louda 1983, 1986; Filby & Van Berkel 1987). However, the diagenetic reactions involved in the porphyrin or porphyrin precursor metallation, the influence of the depositional environment and the role of the source rock kerogen or clay minerals on the metallation process are not as well understood as the evolution of the macrocyclic carbon skeleton of the porphyrin molecule. The predominance of nickel and vanadium as the chelated metal in porphyrins from crude oil and source-rock bitumens is at first somewhat surprising given the large number of very stable metal porphyrins known, particularly of trivalent and tetravalent metals. However Lewan (1984) and others (e.g. Quirke 1987) have explained the prevalence of nickel and vanadium on the basis of thermodynamic stability of the Ni(II) and VO(II) porphyrins and on the pH and Eh control on the availability of metallating ions in marine sedimentary environments. The extremely stable porphyrins of Si4+, Ge 4+, Sn 4+, Ti 4+, and Zr 4+ are difficult to synthesize in the laboratory because, in aqueous solutions, these tetravalent metals exist either as anionic species (e.g. SiO44-) or as very stable hydrolysed ions (e.g. TiO2+). In order to metallate the free-base porphyrin, the M-O bond must be broken to insert the tetravalent cation into the macrocycle. Hence in the synthesis of these metalloporphyrins, metal compounds that do not contain M-O bonds, e.g. TiCI4, other halides or metal alkyls, must be used. In the marine environment, the concentrations of the tetravalent cations, and most trivalent cations (AI 3+, Ga 3+, etc), in sea water in the pH range of 4-8 are negligible and the metalloporphyrins do not form, even though the total concentrations of certain of these elements (e.g. Si, AI, Ti) are relatively high. The concentrations of many divalent metals that form stable metaUoporphyrins are extremely low in seawater under reducing conditions, in which H2S is present, because of the precipitation of very insoluble sulphides, e.g. CdS, PbS, HgS. These factors combined with the very low abundances of many metals in seawater inhibit the formation of most metalloporphyrins

during sedimentation, either in solution or on mineral surfaces. However, for nickel and vanadium, the pH and Eh ranges of marine sediments are conducive to formation of the Ni(II) and VO(II) porphyrins, although to different extents depending on the pH and redox conditions. Strongly reducing, HzS-rich, sedimentary environments favour speciation of vanadium as VO 2+ in the pH 4-8 range, under which conditions Ni 2+ is precipitated as NiS. In less anoxic, H2S-poor, environments Niz+ is the predominant nickel species in solution whereas vanadium is present primarily as V(V) which, in anionic form (e.g. VO3), is capable of metallating free-base porphyrins. Thus oils formed from sediments deposited in highly reducing, sulphurrich, depositional environments have porphyrin assemblages dominated by VO(II) porphyrins and those formed in less anoxic, sulphur-poor environments, show Ni(II) porphyrins as the dominant species. Thus a transition from a porphyrin assemblage dominated by Ni(II) porphyrins to that dominated by VO(II) porphyrins has been observed in the Lower Toarcian in Germany in which a change from an oxic to an anoxic depositional environment occurred (Moldowan et al. 1986). A similar trend in the Ni/(Ni + V) ratio was also observed in shales of the Meade Peak member of the Phosphoria shale deposited under different redox conditions (Sharata & Filby 1989; Sharata 1993). Although exceptions occur, the dominant factor controlling metal ion incorporation into the porphyrin macrocycle in sediments, and source-rock bitumens and crude oils, appears to be the pH-Eh conditions during sedimentation. Barwise (1990) has used the abundances of nickel and vanadium and the Ni/V ratio to classify the types of source rocks of crude oils based partly on the data of Lewan (1984). Barwise (1990) showed that the nickel and vanadium concentrations, as well as the Ni(II) and VO(II) porphyrin distributions appear to be controlled by the depositional environment of the source rock. Although the depositional environment appears to control the type of metallating ion in the geoporphyrins, and nickel and vanadium abundances in source rock bitumens and crude oils, it has not been established when metallation occurs in the diagenetic conversion of chlorophyll to metalloporphyrin, or the mechanism by which the reaction takes place. Based on the distribution of porphyrins and porphyrin precursors in marine sediments collected in the Deep Sea Drilling Project, Baker and co-workers (Baker & Louda 1983) have proposed that metallation of free-based porphyrins occurs in early diagenesis. This would suggest that the

TRACE ELEMENTS IN CRUDE OILS buried sediment may not be in open contact with the water column above the sediment, as assumed by Lewan (1984). If this is the case, pH-Eh conditions during sedimentation may in fact determine metal ion speciation in the accumulating sediment and this speciation is preserved for incorporation into free-base porphyrins later in the diagenetic process. Metallation of a free-base porphyrin or porphyrin acid is unlikely to take place in true solution given that the formation of the free-base porphyrin, or porphyrin acid, occurs after the immature sediment has been compacted and the interstitial water is probably not in contact with the water column above the sediment. Also, the low solubility of free-base porphyrins and porphyrin acids in water would suggest that the metallation of the porphyrin is a heterogeneous process which may occur on an active mineral surface. In an immature sediment, the free-base porphyrin may be absorbed on the insoluble organic matter in the sediment or on clay minerals which are abundant in most sediments. There is substantial evidence that mineral surfaces and the kerogen in sediments play an important role in metailoporphyrin geochemistry. Bergaya & Van Damme (1982), Cady & Pinnavia (1978) and more recently Day & Filby (1992) have shown that montmorillonite and other clay minerals may act as metallation sites for porphyrins in sediments. For example, Day & Filby (1992) have shown that demetallationmetallation of VO(II) and Ni(II) porphyrins on montmorillonite is reversible and that the reaction is dependent on the 'acidity' of the surface (as measured by the Hammett function, Ho) which is a function of the degree of dehydration of the clay. They also found that adsorption of free-base porphyrins from an organic solvent onto the clay (measured as a distribution coefficient) was much greater than the adsorption of the corresponding Ni(II), VO(II) or Cu(II) species. Because clays may retain high concentrations of metal ions in ion-exchange sites, these metal ions may be sterically oriented to metallate an adsorbed free-base porphyrin. The metalloporphyrin which has a lower affinity than the free-base porphyrin for the mineral surface may then desorb into the bitumen forming at the source rock-kerogen interface by catagenesis (Regner & Filby 1992; Bergaya & Van Damme 1982). Regner & Filby (1992) have shown that the DPEP/etio ratio of Ni(II) and VO(II) porphyrins in the bitumen extracted from the New Albany shale depends on the polarity of the extracting solvent (Table 5). In toluene, which is a relatively non-polar solvent, the DPEP/etio

211

Table 5. DPEP/etio ratios of the Ni(ll) and VO(II) porphyrins in the bitumen extracted from the New Albany Shale (Regner & Filby 1992)

Extracting solvent Toluene Toluene-methanol Dichloromethanemethanol Chloroform Pyridine Bitumen II

VO(II) porphyrin DPEP/etio ratio*

Ni(II) porphyrin DPEP/etio ratio*

0.31 1.46

0.73 1.03

1.52 1.57 1.45 8.11

1.25 1.11 1.15 0.89

* DPEP/etio ratio determined by summation of ion intensities of DPEP and etio species in EI mass spectrum

ratios for extracted Ni(II) and VO(II) porphyrins are considerably lower than when more polar solvents (i.e. toluene-methanol, methylene chloride, pyridine) are used for extraction. If the solvent-extracted shale is demineralized and the kerogen re-extracted a small amount of highly polar material (bitumen II) is obtained. This bitumen II is material, located at kerogenmineral interfaces, which is inaccessible to the original extracting solvent and the VO(II) porphyrins in this material have much higher DPEP/etio ratios than in the original extracted bitumen (Table 5). These data indicate that the DPEP species adsorb more strongly to the mineral (probably clay) surface than do the etio species, a conclusion that is confirmed by batch adsorption studies of the metalloporphyrins from toluene and chloroform solutions on montmorillonite surfaces (Regner & Filby 1992). The surface activity of clay minerals also appears to influence the evolution of metalloporphyrins in bitumens and crude oils as shown in the data of Regner et al. (in prep.) on the distribution of VO(II) porphyrins released during pyrolysis of the New Albany shale, Indiana (Mississippian-Devonian). They showed that the DPEP/etio porphyrin ratio of the VO(II) porphyrins generated during hydrous pyrolysis at 300°C was similar to that of the VO(II) porphyrins in the bitumen of the shale whereas pyrolysis of the shale in toluene, or in the absence of solvent, generated VO(II) porphyrins with much lower DPEP/etio ratios. The data were interpreted to show that in the presence of water, the clays are deactivated (hydrated) and thus have little effect on the

212

R.H. FILBY

Table 6. Concentrations of trace elements in pyrolysates of the New Albany Shale (S) and kerogen ( K) under different pyrolysis conditions (Mercer et al. 1992; Fitzgerald & Filby 1992" Element

Kerogen

Bitumen I

S-toluene

S-no solvent

S-hydrous

K-hydrous

V Ni Ni/V As Se Co Cr Fe Mo

1130 2160 1.91 36.4 175 115 35.2 0+ 2400

2700 2560 0.95 75.8 79.5 8.63 1.04 353 28.7

1880 1710 0.91 536 43.8 13.4 1.65 992 54.6

365 896 2.45 86.3 18.9 14.3 1.75 556 17.4

1540 2760 1.80 64.5 215 10.6 0.87 744 6.24

1430 3290 2.30 18.1 32.2 26.2 22.2 237 4.79

* Pyrolysis of shale and kerogen carried out at 300° C for 5.5 hr in presence of no solvent, toluene, or H20 (hydrous) + Kerogen values corrected for mineral (FeS2) contribution. All Fe assumed to be in FeS2form metalloporphyrins released in hydrous pyrolysis, whereas during pyrolysis in toluene or 'dry' the clay surfaces are active and adsorb (or decompose) the DPEP porphyrins at a higher rate than the etio species. Thus the composition of the Ni(II) and VO(II) porphyrin species accumulating in a source-rock bitumen may be influenced by the composition of the mineral matrix of the shale. The role of kerogen in the diagenetic reactions in the Treibs scheme for the evolution of the porphyrins has not been established. However, several studies have inferred (Mackenzie et al. 1980) or shown experimentally (Van Berkel & Filby 1987, 1988; Beato et al. 1991; Sundararaman & Moldowan 1993) that Ni(II) and VO(II) porphyrins are released from kerogens or shales during catagenesis or simulated catagenesis (pyrolysis). Recent work has also shown that the porphyrins generated during pyrolysis of a nickel and vanadium-rich kerogen (i.e simulated catagenesis) are similar in type, distribution and structure to those present in the inherent bitumen (Van Berkel et al. 1989a, b; Beato et al. 1991; Concha et al. 1991). There is no evidence that the metalloporphyrins released during kerogen pyrolysis were bound via C-C bonds to the kerogen structure and it appears probable that these metal complexes are adsorbed on, or occluded in, the kerogen structure (Van Berkel et al. 1989). It has also been shown that simulation of bitumen generation in a source rock by either controlled pyrolysis of the kerogen or of the parent shale under different conditions produces pyrolysates containing similar suites of trace elements to those found in the inherent bitumen of the shale (Mercer et al. 1992; Fitzgerald & Filby 1992). However, the abundances of the trace elements in the pyrolysates show wide

variation with pyrolysis conditions and are different to those of the source rock bitumen. Table 6 compares data on the trace element abundances in pyrolysates of the New Albany shale and the New Albany kerogen with the corresponding concentrations in the bitumen and kerogen of the shale. The concentrations of all trace elements in the pyrolysates are clearly dependent on the conditions of pyrolysis, i.e. the pyrolysis mode (solvent versus dry) and the pyrolysis solvent. For nickel and vanadium, the Ni/V ratio in the pyrolysates reflects more closely the N i N ratio in the kerogen rather than that of the bitumen, except for the toluenepyrolysate. For other trace elements, there is greater variation in concentration than is observed for nickel and vanadium, although the composition of the hydrous pyrolysate approximates the bitumen in terms of trace element abundances more closely than other pyrolysates. This finding is consistent with the claim that hydrous pyrolysis of a source rock closely simulates natural bitumen generation (Lewan et al. 1979). Comparison of the compositions of the hydrous pyrolysates generated from the shale and the kerogen indicates that Ni, V and Mo concentrations are similar for the two pyrolysates whereas values for As, Se, Cr, Co, and Fe are significantly different. This suggests that inorganic--organic interactions may play a role in determining the nature and abundances of non-porphyrin metal complexes in source rock bitumens as has been demonstrated for the metalloporphyrins. The factors controlling the primary incorporation of metal ions or complexes in a source-rock bitumen have not been completely explained but, as suggested by the pyrolysis studies, are clearly related to kerogen composition, the composition and type of the oil being generated

TRACE ELEMENTS IN CRUDE OILS (i.e solvent type) and mineral reactivity during pyrolysis or catagenesis (i.e. nature of the source rock). Thus the abundances of trace elements in crude oils may ultimately be related to depositional conditions, as has been demonstrated for the metalloporphyrin complexes (Lewan 1984; Moldowan et al. 1986).

Secondary processes There is considerable evidence that the concentrations of trace elements or metalloporphyrins in bitumens and crude oils may be modified by secondary processes such as maturation, migration and biodegradation. There are also documented cases of the interaction of crude oils with mineral deposits, metal-rich fluids, or formation waters which may also modify trace element abundances in oils. The effect of thermal maturation on the absolute abundances of metalloporphyrins in source-rock bitumens is a function of two competing processes. Thermal degradation of the Ni(II) and VO(II) porphyrins in a maturing bitumen, either by demetallation of the metalloporphyrin or by direct decomposition, will result in decreasing concentrations of these species with increasing maturity. However this effect may be masked by the release of metalloporphyrins during kerogen catagenesis as has been inferred for shales of the Toarcian Basin by Mackenzie et al. (1980) and shown experimentally in several pyrolysis studies (Van Berkel et al. 1989; Filby & Van Berkel 1987). It has been well established, however, that the DPEP/etio ratio of geoporphyrins decreases with increasing thermal maturity. Several studies have shown that the ratio of the DPEP to etio structural types of the vanadyl porphyrins (or total porphyrins) decreases with increasing thermal maturity of a source rock or sediment (Didyk et al. 1975; Mackenzie et al. 1980; Barwise & Roberts 1984; Barwise 1987). A similar observation was made by Barwise & Park (1983) for porphyrins from oils from a common source rock but reservoired at different depths and hence exposed to different thermal regimes. Didyk et al. (1975) initially explained the variation in DPEP/etio with maturity by the thermal conversion of the DPEP porphyrin to the corresponding etio species via scission of the exocyclic five-membered ring on the DPEP molecule. However, on the basis of more recent evidence, Barwise & Park (1983) and Barwise (1987) concluded that DPEP to etio conversion did not occur and that the decrease in the DPEP/etio ratio with increasing thermal maturity was the

213

result of a slower degradation rate for the etio species relative to that of the DPEP metalloporphyrins in the maturing sediment or oil. Increasing maturity has also been shown to change the distribution of VO(II) etio porphyrins towards lower carbon numbers (Sundararaman & Moldowan 1993; Gallango & Cassani 1992) indicating that for a given porphyrin type, the thermal degradation rate is higher for higher carbon number species. Although thermal degradation rates of the DPEP and etio porphyrins differ slightly, radiation degradation rates of the DPEP and etio species appear to be the same. Sundararaman & Dahl (1993) examined the vanadyl porphyrins from the Alum Shale in Sweden and found that total vanadyl porphyrin contents decreased with increasing uranium content of the shale whereas all samples had similar DPEP/etio ratios. The effect of maturity on the abundances of trace elements, other than nickel and vanadium is less clear although, in general, the abundances of most trace elements decrease with increasing oil maturity because of the decrease in asphaltene content (Branthaver & Filby 1987). The effects of petroleum migration on trace element and metal complex abundances are not well documented. In the case of the metalloporphyrin complexes, migration may change the DPEP/etio porphyrin ratios of the Ni(II) or VO(II) complexes by chromatographic effects. During migration the DPEP and etio species which have different polarities may be fractionated at the solvent (petroleum) - substrate (rock) interface. For a given carbon number, the DPEP species is more polar than the corresponding etio species (Quirke 1987), hence migration through a shale or sandstone formation may result in depletion of the DPEP porphyrins relative to the etio species in early migration stages. As an example, Chakhmakhchev et al. (1985) have shown a decrease in the DPEP/etio ratio in VO porphyrins in laboratory simulations of petroleum migration and have interpreted the changes in DPEP/etio ratios of West Surgut, Siberia, oils as due to migration of the oils from southwest to northeast and the preferential retention of the more polar DPEP species during migration. Although such chromatographic effects based on polarity may occur during expulsion of the oil from the source rock in primary and secondary migration they may be obliterated if a significant amount of the oil from the source rock has migrated long distances and reached a trap (i.e. equivalent to total elution from a chromatographic column). There is little evidence that biodegradation

214

R.H. FILBY

changes Ni(II) or VO(II) porphyrin distributions (Palmer 1983; Strong & Filby 1987; Sundararaman & Hwang 1993), as might be expected given the high molecular weight and aromatic nature of these complexes which would make them resistant to bacterial attack. As Sundararaman & Hwang (1993) have shown, the VO(II) porphyrin distributions in highly degraded oils can be used as geochemical biomarkers when other, principally hydrocarbon, biomarkers have been completely degraded. The many associations of crude oils or solid bitumens with mineral deposits (e.g. of Au, Cu, Hg, U, Th) or mineralization implies processes that involve either the interaction of hydrocarbon fluids with mineral deposits during migration, interactions of hydrocarbon and metalrich aqueous fluids, or interaction of metalbearing aqueous fluids with hydrocarbon deposits. The role of hydrocarbons, hydrocarbon precursors (kerogen, organic acids etc) or other crude oil components in mineral paragenesis appears to be by either facilitating metal ion transport in aqueous fluids as metal-organic complexes or as a reducing agent in the reduction of mobile higher oxidation states of metals such as U and V to lower, less soluble oxidation states. Although the role of organic matter in mineral genesis has been extensively investigated, the influence of mineral paragenesis on metal complexes in crude oils and solid bitumens is less well understood. An example of enhancement of the metal content of crude oils by secondary processes is provided by oils from the Cymric field, Kern County, California (Eocene-Pleistocene) which may contain up to 23 ppm Hg as metallic Hg (Filby 1975; White et al. 1970; Bailey et al. 1961), a value which approaches the solubility of metallic Hg in hydrocarbon solvents. In the California coastal range region the occurrence of bituminous nodules and inclusions in cinnabar deposits also demonstrates the interaction of metalliferous fluids and organic matter, including petroleum (Peabody & Einaudi 1992). Anomalously high V concentrations are found in Cretaceous and Tertiary oils of the Maracaibo Basin, Venezuela (e.g. Boscan oils) and are associated with a regional geochemical enrichment of V (Kapo 1978). In this case, however, it is likely that the enrichment of V in the oils is related to the Eh-pH conditions under which the source rocks were deposited and is thus of primary origin.

Metal species and geochemical correlations Metal complexes in crude oils and source rock

bitumens have been used as biomarkers to provide information on the depositional environment of source rocks, determine the type of organic source material, estimate sediment or oil maturity, or to correlate crude oils with other oils or potential source rocks. Most of these studies have utilized the metalloporphyrins, either the vanadyl porphyrins or the total demetallated porphyrin fraction of the oil or sediment. Thus the DPEP/etio, DPEP/(DPEP + etio), or etio/[DPEP + etio] ratios have been used to estimate source rock (Didyk et al. 1975; Mackenzie et al. 1980; Barwise & Roberts 1984; Barwise 1987) or oil maturity (Barwise & Park 1983). Sundararaman and co-workers (Sundararaman et al. 1988; Aizenshtat & Sundararaman 1990; Sundararaman & Raedeke 1993; Sundararaman & Moldowan 1993) have refined the use of porphyrin structural parameters and have proposed a porphyrin maturity parameter (PMP) defined as the C28etio/[C28etio + C32DPEP] ratio for the vanadyl porphyrins. They have shown that the PMP increases as sediment or oil maturity increases and that the PMP can be related to other maturity parameters such as vitrinite reflectance (Sundararaman & Teerman 1991), rock-eval parameters (Sundararaman et al. 1988), or sterane-steroid biomarker indices (Sundararaman & Moldowan 1993). Gallango & Cassini (1992) have used the C27etio/C~setio and Czsetio/Czgetio as maturity indexes for biodegraded oils from the Maracaibo Basin, Venezuela, and Sundararaman & Moldowan (1993) have used the ratio of low molecular weight etio (LMWE) porphyrins to higher molecular etio (LMWE + C28etio) porphyrins as a maturity index (PMP-2) for highly mature oils from the Oriente Basin, Ecuador. In both of these studies the correlations are based on the increase in thermal stability of the etio porphyrins with decreasing carbon number. Thus maturities can be estimated for very mature oils that contain very little, or no, DPEP porphyrins. The metalloporphyrins are of particular value for estimating the maturities or correlating oils that have been biodegraded and for which other hydrocarbon biomarker indices using steranes, hopanes or terpanes have been altered by biodegradation. Sundararaman & Hwang (1993) have shown that even for highly biodegraded oils and tars from the Phosphoria formation in Wyoming in which hydrocarbon maturity indices have been altered, the PMP index is unaffected and can be used to estimate maturity. The DPEP/etio ratios of vanadyl porphyrins have been used by Chakhmakhchev et al. (1985) to correlate among migrated oils from the West Surgut Basin,

TRACE ELEMENTS IN CRUDE OILS

215

Table 7. Classification of crude oil types based on organic matter type and depositional environment of source rocks (modified from Barwise 1990)

Oil class

Organic matter type forming kerogen

Depositional environment Sulphur Total Ni and V Ni/V ratio of source rock content of oil content

A

Phytoplankton/bacteria

B

Phytoplankton/bacteria

Marine carbonates; other non-siliciclastics Marine siliciclastics

C

Phytoplankton/bacteria

Lacustrine

Moderate sulphur Low sulphur

D E

Terrestrial plants Terrestrial plants

Non-marine Non-marine

Low sulphur Low sulphur

Siberia, and the DPEP-etio ratio of the nickel porphyrins has been used to correlate terrestrial oils of the Jianghan Basin, China (Yang et al. 1988; C h e n & Philp 1991). Because the redox conditions (Eh) influence the speciation of nickel and vanadium in the depositional environments of petroleum source rocks, VO(II) porphyrins predominate in more anoxic environments and Ni(II) porphyrins predominate under more oxic conditions (Lewan 1984). Hence, the Ni(II) porphyrin/ [Ni(II) porphyrin + VO(II) porphyrin] ratio has been used to characterize changes in the redox conditions of the depositional environment of petroleum source rocks from the Lower Toarcian shales of SW Germany (Moldowan et al. 1986), the Posidonia Shale (Toarcian) of northern Germany (Sundararaman et al. 1993) and the Meade Peak Member of the Phosphoria Shale, Idaho (Sharata & Filby 1989; Sharata 1993). In the study by Moldowan et al. (1986), it was shown that the Middle Lias shales deposited under oxic shallow water conditions contain only Ni(II) porphyrins whereas the Lower Lias shales deposited under more anoxic, deeper, conditions have a porphyrin assemblage that is almost 100% VO(II) porphyrins. The transition zone between the Lower and Middle Lias shales contains both porphyrin types. In the Phosphoria Basin, a similar transition from a high Ni/V ratio to a low Ni/V ratio was observed in the transition from sediments deposited under shallow more oxic conditions at the edge of the basin to sediments deposited under more anoxic conditions in deeper parts of the basin (Sharata & Filby 1989; Sharata 1993). The relative abundances of the nickel and vanadyl porphyrins have also been used by Gransch & Eisma (1970) to correlate oils from the Maracaibo Basin, Venezuela with the Cretaceous La Luna shale as the probable source rock. The nickel and vanadium contents of crude

High sulphur

High (>50 ppm) High-medium (1-10 ppm) Low (<10 ppm) Low (<5 ppm) Low (<5 ppm)

<0.5 <0.5 >2-10 >5 >5

oils have been used extensively in oil-oil and oil-source rock correlations and in determining the type of organic matter of source-rocks (see review by Branthaver & Filby 1987). A systematic classification of crude oils based on total nickel and vanadium contents and the Ni/V ratio has been made by Barwise (1990) who proposed the classification for use as a preliminary exploration tool. This classification (summarized in Table 7) is based on field observations but is also consistent with the Eh-pH model proposed by Lewan (1984). Several authors have successfully used multielement distributions in oil-oil correlation studies (Connor & Gerrild 1971; Saban et al. 1984; Hitchon & Filby 1984; Ellrich et al. 1985; Hirner 1987; Curiale 1987). In these studies it is assumed that the trace element fingerprint in the oil is inherited from the bitumen of the source rock and that migration, maturation, and other secondary effects may affect absolute concentrations of metal species but do not affect metal/metal ratios. This is probably justified because the metal species in crude oils, including the metalloporphyrins, are associated primarily with the asphaltene or polar component of the oil and hence may vary in concentration as the asphaltene concentration varies while metal/ metal ratios remain constant. Multivariate statistical techniques have been used in a number of correlation studies. Connor & Gerrild (1971) classified oils from the Elk Hills National Petroleum Reserve using stepwise multiplediscriminant function analysis, and Hitchon & Filby (1984) extended this technique to classify Devonian to Cretaceous oils from the West Canada Basin into three families. Curiale (1987) used cluster analysis of transition metal data to show that the North Slope, Alaska, oils could be classified into two groups. Recently, Frankenberger et al. (1993) used factor analysis to classify the terrestrial oils and condensates of the

216

R.H. FILBY

Taranaki Basin, New Zealand on the basis of the concentrations of sixteen trace elements. A few studies on oil-source rock correlations have been made (e.g. Odermatt & Curiale 1991). An interesting and potentially very useful geochemical tool for oil-source rock correlation is the variation in the 143Nd/144Nd ratio that results from the radiogenic production of ~43Nd by alpha decay of 1475m (half life = 1.06 x 109 years) in rocks containing both samarium and neodymium (Manning et al. 1991; Stille et al. 1993). Source rocks of different ages and initial Sm/Nd ratios will have different 143Nd/laaNd ratios and oils derived from such source rocks thus inherit the 143Nd/144Nd ratio of the source rock. Manning et al. (1991) showed that pyrolysates of source rocks had the same 143Nd/144Nd ratios as the source rocks and proposed this isotopic ratio as a correlation tool. Stille et al. (1993) have shown that the ~43Nd/~a4Nd ratio can be used for oil reservoir-source rock correlations when other biomarkers have been completely degraded. Caution must obviously be used in interpreting correlations based on trace elements because the geochemical controls on their distributions and the relative importance of primary and secondary processes are not as well understood as is the case for the metalloporphyrins. Conclusions

Crude oils and source rock bitumens contain a large suite of trace elements of which nickel and vanadium are typically the most abundant. Of the metal complexes present in petroleum, only the Ni(II) and VO(II) complexes of the geoporphyrins have been characterized and these have been derived principally from chlorophylls deposited in sedimentary environments. The relative abundances of the porphyrin types, e.g. DPEP, etio, and the Ni(II) porphyrin to VO(II) porphyrin ratio provide information on the depositional environment of the petroleum source rock, the maturity of the oil or sediment and thus provide the basis for oil-oil and oil-source rock correlations. Both the mineral components and the kerogen of source rocks play a role in the evolution of the porphyrins, either through surface mediated reactions (e.g. metallation) or by adsorption and subsequent release in catagenesis. Less is known of the nature of other metal species in petroleum but kerogen-mineral reactions during catageneis also appear to play an important role in establishing the trace metal distributions in source rock bitumens and crude oils. Although

the interaction of petroleum with metal-rich fluids or ore bodies has been shown to be an important process in ore genesis, only limited information on the effect of such processes on trace metal distributions in crude oils is available. In the California coastal range region, the interaction of petroleum with ore fluids containing mercury or mercury deposits has resulted in enrichment of some crude oils in elemental mercury. References

AIZENSHTAT, Z. ¢~ SUNDARARAMAN,P. 1990. Maturation trend in oils and alsphalts of the Jordan Rift: utilization of detailed vanadylporphyrin analysis. Geochimica et Cosmochimica Acta, 5 3 , 3185-3188. BAILEY, E.H., SNAVELY,P.D. & WHITE, D.E. 1970. Chemical analysis of brines and crude oil, Cymric Field, Kern County, California. US Geological Survey Professional Paper 424D. BAKER,E.W. & LOUDA,J.W. 1993. Thermal aspects of chlorophyll geochemistry. In: BJOROY, M. (ed.) Advances in Organic Geochemistry 1981. John Wiley, London, 401-421. - & -1984. Highly dealkylated copper and nickel porphyrins in marine sediments. Organic Geochemistry, 6, 183-192. - & 1986. Porphyrins in the geological record. In: JOHNS, R.B. (ed.) Biological Markers in the Sedimentary Record. Elsevier, Amsterdam, 125225. BARWISE, A.J.G. 1987. Mechanisms involved in altering deoxophylioerythroetioporphyrin-etioporphyrin ratios in sediments and oils. In: FILBY, R.H. & BRANTHAVER,J.F. (eds) Metal Complexes in Fossil Fuels. American Chemical Society Symposium Series, 344, 100-109. BARWISE,A.J.G. 1990. Role of nickel and vanadium in petroleum classification. Energy and Fuels, 4, 647-652. - - & PARK,P.J.D. 1983. Petroporphyrin fingerprinting as a geochemical marker. In: BJOROYM. (ed.) Advances in Organic Geochemistry 1981. John Wiley, London, 668-674. - & ROBERTS,I. 1984. Diagenetic and catagenetic pathways for porphyrins in sediments. Organic Geochemistry, 6,167-176. BEATO, B.D., YOST, R.A., VANBERKEL, G.J., FILBY, R.H. & QUIRKE, J.M.E. 1991. The Henryville Bed of the New Albany Shale - III: tandem mass spectrometric analyses of geoporphyrins for the bitumen and kerogen. Organic Geochemistry, 17, 93-105. BENNET, B., ECKARDT,C.B. • MAXWELL,J.R. 1993. Tetrapyrrole pigments as indicators of the extent of oxygen depletion in paleo water bodies. (Abs.) 16th International Meeting on Organic Geochemistry, Stavanger, Norway, 35. BERGAYA, F. & VAN DAMME, H. 1982 Stability of metalloporphyrins adsorbed on clays: a comparative study. Geochimica et Cosmochimica Acta 46, 349-360.

TRACE ELEMENTS IN CRUDE OILS BRANTHAVER, J.F. & FILBY, R.H. 1987. Application of metal complexes in petroleum to exploration geochemistry. In: FILBY, R.H. & BRANTHAVER, J.F. (eds) Metal Complexes in Fossil Fuels. American Chemical Society Symposium Series, 344, 84-99. CADY, S. & PINNAVIA,T.J. 1978. Porphyrin intercalation in mica-type silicates. Inorganic Chemistry, 17, 1501-1507. CALLOT, H.J., OCAMPO, R. & ALBRECHT, P. 1990. Sedimentary porphyrins: correlations with biological precursors. Energy and Fuels, 4,635-643. CHAKHMAKHCHEV,V.A., BURKOVA,V.N., ZHARKOV, N.I., PUNANOVA,S.A., SEREBRENNIKOVA,O.V. & TITOV,V.I. 1985. Composition changes in vanadyl porphyrins in oil filtering through porous media. Geokhimiya, 1985, 381-386. CHEN, J.H. & PHILP, R.P. 1991. Porphyrin distributions in crude oils from the Jianghan and Biyang basins, China. Chemical Geology, 91, 139-152. CHICARELLI, M.I., KAUR, S. & MAXWELL,J.R. 1987. Sedimentary porphyrins: unexpected structures, occurrence and possible origins. In: FILBY, R.H. & BRANTHAVER,J.F. (eds) Metal Complexes in Fossil Fuels. American Chemical Society Symposium Series, 344, 40-67. CONCHA,M.A., QUIRKE,J.M.E., BEATO,B.D., YOST, R.A., MERCER, G.E. & FILBY, R.H. 1991. The Henryville Bed of the New Albany Shale, IV: tandem mass spectrometric analyses of the geoporphyrins from the bitumen of the demineralized shale. Chemical Geology, 91,153-168. CONNOR, J.J. & GERRILD, P.M. 1971. Geochemical differentiation of crude oils from six Pleistocene sandstone units, Elk Hills, U.S. Naval Petroleum Reserve No. 1, California. The American Association of Petroleum Geologists Bulletin, 55, 18021813. CURIALE,J.A. 1987. Distribution of transition metals in North Alaskan oils. In: FIERY, R.H. & BRANTHAVER, J.F. (eds) Metal Complexes in Fossil Fuels. American Chemical Society Symposium Series, 344, 135-145. DAY, J.W. & FILBY, R.H. 1992. Role of clay mineral acidity in the evolution of the nickel, vanadyl and copper geoporphyrins. Preprints American Chemical Society, Division of Petroleum Chemistry, 37, 1399. DIDYK, B.M., ALTURKI, Y.I.A., PILLINGER, C.T. & EGLINTON,G. 1975. Petroporphyrins as indicators of geothermal maturation. Nature, 256,563-565. ECKARDT, C.B., WOLF, M. & MAXWELL,J.R. 1989. Iron porphyrins in the Permian Kupferschiefer of the Lower Rhine, N.W. Germany. Organic Geochemistry, 14,659-666. ELLRICH, J., HIRNER, A.V. & STARK, H. 1985. Distribution of trace elements in crude oils from Southern Germany. Chemical Geology, 48, 313323. FILBY, R.H. 1975. The nature of metals in petroleum. In: YEN, T.F. (ed.) Role of Trace Metals in petroleum. Ann Arbor Science Publishers, Ann Arbor, MI, 31-58.

217

& BRANTHAVER,J.F. (eds) 1987. Metal Complexes in Fossil Fuels. American Chemical Society Symposium Series, 344. VAN BERKEL, G.J. 1987. Geochemistry of metal complexes in petroleum, source rocks and coals: An Overview. In: FILBY, R.H. & BRANTHAVER, J.F. (eds) Metal Complexes in Fossil Fuels. American Chemical Society Symposium Series, 344, 2-39. FISH, R.H. & BRINCKMAN,F.E. 1983. Isolation and Identification of Organoarsenic and Inorganic Arsenic Compounds in the Green River Formation Shale. Preprints American Chemical Society Division of Petroleum Chemistry, 28, 177180. , REYNOLDS, J.G. & GALLEGOS, E.J. 1987. Molecular Characterization of Nickel and Vanadium Nonporphyrin Compounds found in Heavy Crude Petroleums and Bitumens. In: FILBY, R.H. & BRANTHAVER,J.F. (eds) Metal Complexes in Fossil Fuels. American Chemical Society Symposium Series, 344, 332-349. FITZGERALD, S.L. & FILBY, R.H. 1992. The Effect of Minerals on Trace Element Distributions in Hydrous Pyrolysates of Kerogens from the New Albany Shale. Preprints American Chemical Society Division of Fuel Chemistry, 37,1754-1760. FRANKENBERGER, A., BROOKS, R.R., VARELAALVAREZ, H., COLLEEN, J.D., FILBY, R.H. & FITZGERALD, S.L. 1993. Classification of some New Zealand crude oils and condensates by means of their trace element contents. Applied Geochemistry, 9, 65-72. GALLANGO,O. t~ CASSANI,F. 1992. Biological marker maturity parameters of marine crude oils and rock extracts from Maracaibo Basin, Venezuela. Organic Geochemistry, 18,215-224. GOULON, J., RETOURNARD,A., FRIENT, P., GOULONGINET, C., BERTHE, C., MULLER, J.F., PONCET, J.L., GUILARD, R., ESCALIER, J.C. & NEFF, B. 1984. Structural characterization by X-ray absorption spectroscopy (EXAFS/XANES) of the V chemical environment in Boscan asphaltenes. Journal of the Chemical Society, Dalton Transactions, 1984, 1095-1103. GRANSCH, J.A. & EISMA, E. 1970. Geochemical aspects of the occurrence of porphyrins in West Venezuelan mineral oils and rocks. In: HOBSON, G.D. & SPEERS,G.C. (eds) Advances in Organic Geochemistry 1966. Pergamon Press, Oxford, 69-86. HIRNER, A.V. 1987. Metals in crude oils, asphaltenes, bitumen and kerogen in Molasse Basin, Southern Germany. In: FILBY, R.H. & BRANTHAVER,J.F. Metal Complexes in Fossil Fuels. American Chemical Society Symposium Series, 344, 146153. HITCHON, B. & FILBY, R.H. 1983. Trace Elements in Alberta Crude Oils. Open File Report, 1983-02, Alberta Research Council, Edmonton. & 1984. Use of trace elements for classification of crude oils into families - example from Alberta, Canada. American Association of Petroleum Geologists Bulletin, 68,838-849.

218

R.H. FILBY

, & SHAH, K.R. 1975. Geochemistry of trace elements in crude oils, Alberta, Canada. In: YEN, T.F. (ed.) Role of Trace Metals in Petroleum. Ann Arbor Science Publishers, Ann Arbor, 111-121. JAGARNATHAN, J., MOnAN, M.S. & ZINGARO, R.A. 1986. Identification of arsenic-bearing minerals in a sample of Green River Oil Shale. Fuel, 65, 266-269. KAPO, G. 1978. Vanadium; key to Venezuelan fossil hydrocarbons. In: CHILINGARIAN,G.V. & YEN, T.F. (eds) Bitumens, Asphalts and Tar Sands. Elsevier, Amsterdam, 213-238. KELLY, B.J., PROWSE,W.G. & MAXWELL,J.R. 1990. The Treibs hypothesis: an evaluation based on structural studies. Energy & Fuels, 4,628--634. LASH, T.D. 1993. Geochemical origins of sedimentary benzoporphyrins and tetrahydrobenzoporphyrins. Energy & Fuels, 7,166-171. , BALASUBRAMANIAM,R.P. & I2 OTHERS 1990. Influence of carbocyclic rings on porphyrin cyclizations: synthesis of geochemically significant geoporphyrins. Energy and Fuels, 4,668-4574. LEWAN, M.D. 1984. Factors controlling the proportionality of vanadium to nickel in crude oils. Geochimica et Cosmochimica Acta, 48, 22312238. , WINTERS, J.C. & MCDONALD, J.H. 1979. Generation of oil-like pyrolyzates from organicrich shales. Science, 203,897-899. MACKENZIE,A.S., OUIRKE,J.M.E. & MAXWELL,J.R. 1980. Molecular parameters of maturation in the Toarcian Shales, Paris Basin, France - II. evolution of metalloporphyrins. In: DOUGLAS,A.G. t~ MAXWELL, J.R. (eds) Advances in Organic Geochemistry 1979. Pergamon Press, Oxford, 239-248. MANNING, L.K., FROST, C.D. & BRANTHAVER,J.F. 1991. A neodymium isotopic study of crude oils and source rocks: potential applications to petroleum exploration. Chemical Geology, 91,125138. MERCER,G.E., FITZGERALD,S.L., DAY,J.W. & FILBY, R.H. 1991. Determination of organic/inorganic associations of trace elements in oil shale kerogens. Preprints American Chemical Society Division of Fuel Chemistry, 36, 1180-1186. MERCER, G.E., FITZGERALD,S.L., DAY,J.W. & FILBY, R.H. 1993. Determination of oraganic and inorganic associations of trace elements in the kerogen of the New Albany Shale. Fuel, 72, 1187-1195. MERCER, G.E., REGNER, A.J. & FILBY, R.H. 1992. Trace element distributions in kerogen, bitumen and pyrolysates isolated from the New Albany Shale. Preprints American Chemical Society Division of Fuel Chemistry, 37, 1761-1768. MOLDOWAN,J.M., SUNDARARAMAN,P. & SCHOELL,M. 1986. Sensitivity of biomarker properties to depositional environment and/or source input in the Lower Toarcian of SW Germany. Organic Geochemistry, 10,915-926. NGtr~EN, S.N. & FILBY, R.H. 1987. Interaction of Ni(II) complexes with Athabasca asphaltenes. In:

FILBY, R.H. & BRANTHAVER,J.F. (eds) Metal Complexes in Fossil Fuels. American Chemical Society Symposium Series, 344,384-401. ODERMATT, J.R. & CURIALE, J.A. 1991. Organically bound metals and biomarkers in the Monterey Formation of the Santa Maria Basin, California. Chemical Geology, 91, 99-114. PALMER, S.E. 1983. Porphyrin distributions in degraded and non-degraded oils from Colombia. Abstracts 186th American Chemical Society National Convention, Geochemistry Division, August 1993, Washington, DC. PEABODY, C.E. d~ EINAUDI, M.T. 1992. Origin of petroleum and mercury in the Culver-Baer cinnabar deposit, Mayacmas District, California. Economic Geology, 87, 1078-1103. PEARSON, C.D. • GREEN, J.B. 1989. Comparison of processing characteristics of Mayan and Wilmington heavy residues. 2. characterization of vanadium and nickel complexes in acid-baseneutral fractions. Fuel, 68,465-474. & - 1993. Vanadium and nickel complexes in petroleum resid acid, base, and neutral fractions. Energy & Fuels, 7,338-346. QUIRKE, J.M.E. 1987. Rationalization for the predominance of nickel and vanadium porphyrins in the geosphere. In: FILBY, R.H. & BRANTHAVER, J.F. (eds) Metal Complexes in Fossil Fuels. American Chemical Society Symposium Series, 344, 74-83. REGNER, A.J. & FILBY, R.H. 1992. A comparison of metailoporphyrin distributions from bitumens of the New Albany Shale using different extraction solvents. Preprints American Chemical Society Division of Petroleum Chemistry, 37, 1393-1394. SABAN, M., VITOROVIC, O. & VITOROVIC, D. 1984. Correlation of crude oils from Vojdovina (Yugoslavia) based on trace elements. In: Symposium on Characterization of Heavy Crude Oils and Petroleum Residues. Editions Technip, Paris, 122-127. SHARATA,S.M. 1993. Relation of vanadium and nickel in bitumen to the depositional environment of the Meade Peak Member, Phosphoria Formation, southeastern Idaho. PhD University of Idaho. Moscow, Idaho. & FILBY, R.H. 1989. The distribution of nickel and vanadium in the Meade Peak Member of the Phosphoria Shale. Abstract, 197th American Chemical Society Meeting, Dallas, TX. STILLE, P., GAUTHIER-LAFAYE,F. & BROS, R. 1993. The neodymium isotope system as a tool for petroleum exploration. Geochimica et Cosmochimica Acta, 57, 4521-4525. STRONG, D. & FILBY, R.H. 1987. Vanadyl porphyrin distribution in the Alberta Oil Sand Bitumens. In: FILBY, R.H. & BRANTHAVER,J.F. (eds) Metal Complexes in Fossil Fuels. American Chemical Society Symposium Series, 344, 154-172. SUNDARARAMAN,P. & DANE, J.E. 1993. Depositional environment, thermal maturity and radiation effects on porphyrin distribution: Alum Shale, Sweden. Organic Geochemistry, 20,333-337. & HWANG, R.J. 1993. Effect of biodegradation

TRACE ELEMENTS IN C R U D E OILS on vanadyl porphyrin distribution. Geochimica et Cosmochimica Acta, 57, 2283-2290. & MOLDOWAN, J.M. 1993. Comparison of maturity based on steroid and vanadyl porphyrin parameters: a new vanadyl porphyrin maturity parameter for higher maturities. Geochimica et Cosmochimica Acta, 57, 1379-1386. t~Z RAEDEKE, L.D. 1993. Vanadylporphyrins in exploration: maturity indicators of source rocks and oils. Applied Geochemistry, 8,245-254. & TEERMAN,S.C. 1991. Porphyrins in extracts as indicators of maturity and depositional environment of Tertiary source rocks from Central Sumatra. In 19th Annual Indonesian Petroleum Association Convention (Jakarta, Indonesia) Vol. 1,621-635. , BIGGS, W.R., REYNOLDS,J.G. & FETZER,J.C. 1988. Vanadylporphyrins, indicators of kerogen breakdown and generation of petroleum. Geochimica et Cosmochimica Acta, 52, 2337-2341. ~, SCHOELL, M., LITrKE, R., BAKER, D.R., LErrHAEUSER, D. & RULLKOX'rER,J. 1993. Depositional environment of Toarcian shales from northern Germany as monitored with porphyrins. Geochimica et Cosmochimica Acta, 57, 42134218. TxSSOT, B.P. & WELTE, D.H. 1984. Petroleum Formation and Occurrence. 2nd Ed. Springer Verlag, Berlin. TOOULAKOU, D. & FILBY, R.H. 1988. Separation of nickel and vanadium complexes from the Athabasca Oil Sand asphaltenes. In: YEN, T.F. & MOLDOWAN, J.M. (eds) Geochemical BiDmarkers. Harwood Acad. Publishers, Chur, 221240. TREIBS, A. 1934. Chlorophyll- und h~imindervate in bitumin6sen gesteinen, erdOlen, erdwachsen und asphalten. Annalen Chemie, 510, 42-62. VAN BERKEL, G.J. 1987. The role of kerogen in the evolution of the nickel and vanadyl porphyrins. PhD Dissertation, Washington State University, Pullman, WA.

219

& FILBY, R.H. 1987. Generation of nickel and vanadyl porphyrins from kerogen during simulated catagenesis. In: FILSY, R.H. & BRANTHAVER, J.F. (eds) Metal Complexes in Fossil Fuels. American Chemical Society Symposium Series, 344, 110-134. & - 1988. Determination of the mineral-free Ni and V contents of Green River Oil-Shale kerogen. In: YEN, T.F. & MOLDOWAN,J.M. (eds) Geochemical Biomarkers. Harwood Publishers, Chur, 89-114. VAN BERKEL, G.J., QUIRKE, J.M.E. & FILBY, R.H. 1989a. The Henryville Bed of the New Albany Shale - I. preliminary characterization of the nickel and vanadyl porphyrins in the bitumen. Organic Geochemistry, 14, 119-128. -& -1989b. The Henryville Bed of the New Albany Shale - II. comparison of the nickel and vanadyl porphyrins in the bitumen with those generated from the kerogen during simulated catagenesis. Organic Geochemistry, 14,129-144.

VERNE-MISMER,J., OCAMPO,R., BAUDER,C., CALLOT, H.J. & ALBRECHT,P. 1990. Structural comparison of nickel, vanadyl, copper, and free base porphyrins from Oulad Abdoun Oil Shale (Maastrichtian, Morocco). Energy & Fuels, 4,635-638. WHITE, D.E. 1958. Liquid of inclusions in sulfides from Tri-State is probably connate in origin, (abs). Geological Society of America Bulletin, 69, 1660-166l. ~, HINKLE, M.E. & BARNES, I. 1970. Mercury contents of natural thermal and mineral fluids. U.S. Geological Survey Professional Paper, 713D, 25-28. YANG, Z., TONC, Y. & FAN, Z. 1988. Some features of porphyrins and other biomarkers in crude oil and source rock from continental salt-like sediments in China. In: YEN, T.F. & MOLDOWAN, J.M. (eds). Geochemical Biomarkers. Harwood Publishers, Chur, 275-292.

Fluid chemistry and hydrological regimes in geothermal systems: a possible link between gold-depositing and hydrocarbon-bearing aqueous systems KEITH NICHOLSON

Environment Division, School o f Applied Sciences, The Robert Gordon University, Aberdeen AB1 1HG, UK Abstract: Geothermal fluids form by the infiltration, heating and circulation of meteoric waters in the Earth's crust. Rock-water reactions, boiling and condensation processes influence the chemistry of the fluids. The composition of geothermal fluids ranges from that of gold-depositing, dilute waters to saline, oil-field brines; as such they form a link between these two aqueous systems. Techniques developed to examine the processes active in geothermal systems may find valuable application in gold exploration and hydrocarbon reservoir modelling. The concept of hydrological regime is introduced and applied to recognising deposits of geothermal activity, the level of preservation and permeable, potentially gold-bearing zones in ancient epithermai systems. Mixing models, such as Cl-enthalpy plots, and geothermometry can aid identification of breakthrough of edge waters and casing damage. Differences in the equilibration time of geothermometers may enable the thermal profile or history of the reservoir to be developed. Statistical models enable major controls or sources of the solutes to be isolated, and aid in the correlation of formation waters between wells.

Fluid geochemistry has concerned earth scientists in many disciplines. Economic geologists, attempting to describe the composition of long gone 'ore-forming solutions' and petroleum geologists, in evaluating reservoir history, all benefit from an appreciation of geofluid chemistry. As more aqueous geochemical systems, both ancient and modern, are explored, so the commonality of geochemical processes and fluid types becomes more evident, and for some fluids the link between modern observable fluids and palaeomineralising systems has been established (Fig. 1). Saunders & Swan (1990) and Sverjensky (1984), for example, illustrate how oil-field brines can evolve through water-rock reactions to metal-rich solutions capable of depositing Mississippi Valley Type Pb-Zn deposits. The saline high-temperature geothermal fluids debouched on the seafloor at several modern active ridge systems are now known to be the modern equivalents of ore solutions which deposited sedimentary exhalative Fe-Mn oxide and massive sulphide mineralization (Cyamex 1979; Hannington et al. 1991). Epithermal gold deposits also find present-day equivalents in high-temperature geothermal systems in silicic volcanic terrains (Henley & Ellis 1983; Henley 1985, 1991). Geothermal fluids show a range in composition from dilute, high-temperature waters

( < 1 molal, > 150 ° C to saline, low-temperature waters ( > 1 molal, <150°C). As such they represent a bridging compositional link (Fig. 1) between gold-depositing ore fluids (Brown 1986) and oil-field brines (Matray & Fontes 1993). As it is possible to sample and directly study geothermal waters and gases, they provide the opportunity to examine geochemical processes from the perspective of fluid chemistry and relate how this impacts upon rock chemistry through alteration and deposition reactions. This approach contrasts with that often employed in mineral exploration where it is usual to focus attention on the rock chemistry. However, geothermal fluid chemistry provides an insight into reaction processes in aqueous systems, which may enable reconstruction of the hydrodynamics of a palaeosystem, and additionally permits the development of geochemical tools which may find valuable application to other present-day geofluids. This work aims primarily to highlight potential cross-links from geothermal fluids to golddepositing and hydrocarbon-bearing aqueous systems. A fluid-based approach is used to summarize the variety and chemistry of geothermal fluids and the hydrological regimes developed in geothermal systems. This is then applied to the recognition and reconstruction of gold-depositing palaeosystems. Finally,

From PARNELL,J. (ed.), 1994, Geofluids: Origin, Migrationand Evolution of Fluids in Sedimentary Basins, Geological Society Special Publication No. 78,221-232.

221

222

K. NICHOLSON Pb Zn Cu massive sulphides Fe Mn oxides

I

saline

high-temperature

L GEOTHERMAL

FLUIDS

dilute

high-temperature

I

EPITHERMAL DEPOSITS

I

I

: Au

saline low-temperature

I

Oil-field brines

I I MISSISSIPPIVALLEY I TYPE DEPOSITS

I

Pb Zn Fig. 1. Generalised inter-relationship between geothermal fluid composition, oil-field brines, massive sulphide deposits and epithermal gold mineralization.

suggestions are made for the possible application of geothermal geochemical methods to oil-field waters.

Geothermal systems Geothermal systems are varied in nature and are classified or divided by a series of descriptive terms. They are commonly referred to as liquid or vapour dominated, low or high temperature (enthalpy), sedimentary or volcanic hosted etc. Although some terms are self-explanatory, the significance of reservoir equilibrium, fluid type and reservoir temperature needs to be emphasised. Reviews of explored geothermal systems are described by Ellis & Mahon (1977), Edwards et al. (1982) and Armstead (1983). The reservoir equilibrium state is the fundamental division between geothermal systems

and is based on the circulation of the reservoir fluid and the mechanism of heat transfer. Systems in dynamic equilibrium are continually recharged by water entering the reservoir. The water is heated and then discharged out of the reservoir, either to the surface or to underground permeable horizons. Heat is transferred through the system by convection of the fluid. Systems in static equilibrium (also known as stagnant or storage systems) have only minor or no recharge to the reservoir and heat is transferred only by conduction. The reservoir fluid can be composed mainly of liquid water (liquid-dominated) or steam (vapour-dominated). In most reservoirs, both steam and liquid water exist in varying proportions as two-phase zones. Liquid-dominated systems are most common, and may contain a steam cap which can expand or develop on

GEOTHERMAL FLUID CHEMISTRY As

Sb

Hg

NH 3 OUTFLOW ZONE (lalelal flow, dilul~n)

UPFLOW ZONE RECHARGE ZONE i r . V~\\ ~ =- ~

bicarbonal, ~

acid-sulphale springs, mud pools

223

boiling chl~ide spd~9, sinter

sleaming 9round

dilule-chkxide ± bicarbonate spdng

±condensafion~

",

BOILING ZONE [_ Ikm

"°"

~kc. :

•

~

",

I"?

k~* •

PITHERMN_

QOQ ""~ Q •

0 l~Opylil¢O 0 0 ~'0 0 0 V~O00 Oalla'lll~1 0 0 0 0 o o o o/ o o o o o "~

u

,cNoddewalers

I I I I I

- 2km mck.wate¢ reactions

meteoric wal.

rechmge

(--.=~u=os)

magnatic

ZONE

I I 1

H20 CO2 H2S NaCl

/..

\

I I I I I I

....-.

heal (? + fluid & solutes)

soufc~s 0[ metals coml~exin9

9ands

maQm,

/

IZl two phase zone (liquid ~ s~eam÷gas)

Fig. 2. Hydrological structure of a high-enthalpy geothermal system developed in gently sloping terrain (after Henley & Ellis 1983; Henley 1991). Note inter-relationship of fluid-type, depth and alteration style. See Tables 1 and 2.

exploitation. Systems which discharge only steam are rare: the best known are Larderello, Italy and The Geysers, USA. The reservoir temperature (or enthalpy) of geothermal reservoirs is an important discriminator in terms of potential resource usage. Systems are commonly described as lowtemperature (< c. 150° C) or high-temperature ( > c. 150 ° C). The discriminatory temperatures are not rigid, and some workers also use the term 'intermediate' to indicate reservoir temperatures in the 120-180 ° C range. Low-temperature systems can only be used for 'direct-use' applications (e.g. heating), while high-temperature systems can be used for electricity generation as well as direct-use applications. The heat source for the system is a function of the geological and tectonic setting. If the driving heat flux is provided by a magma, then such systems are termed volcanogenic. They are invariably high-temperature systems. Heat does not, however, have to be supplied by a magma, and a geothermal system can be generated in areas of tectonic activity. For example, heat may

be supplied by the tectonic uplift of hot basement rocks, or water can be heated by unusually deep circulation created by folding of a permeable horizon or faulting. These are termed non-votcanogenic systems and can contain high or low-temperature reservoirs.

Geothermal fluids Geothermal fluids are classified according to the dominant anions. Although not a formal genetic scheme, this descriptive classification does permit some generalizations to be made on the likely origins of the waters. The compositional range and variety of geothermal fluids is primarily a consequence of their different formational processes. Primary (deep) fluids evolve through rock-water reactions, while secondary fluids are produced from steam which is formed by boiling of the deep fluid. Characteristics of the three dominant types of geothermal fluid are discussed below; more detailed consideration of geothermal fluid chemistry may be found in books by Ellis & Mahon (1977), Henley et al. (1984)

224

K. NICHOLSON

and Nicholson (1993). The conceptual structure of geothermal systems, and the inter-relationship of the fluid types is illustrated in Fig. 2.

The primary fluid The most common type of fluid found at depth in high-temperature geothermal systems is of nearneutral pH, with chloride as the dominant anion. Other waters encountered within the profile of a geothermal field are commonly derived from this deep fluid as a consequence of chemical or physical processes.

Chloride fluids. The evolution of the primary chloride fluid in dynamic, liquid-dominated systems can be summarised as follows. Meteoric waters penetrate the crust through permeable zones and circulate to depths of up to around 5-8km (Henley 1985). As they descend, they are heated, react with the host rocks and rise by convection. At depth, the fluids typically contain 1000-10 000 mg kg -1 CI at temperatures of about 350 °C. The 'soluble-group' elements are leached from the host rocks by the waters, along with other elements whose solubilities are controlled by temperature-dependent reactions (Ellis & Mahon 1977). These reactions change the primary mineralogy of the host rocks to a distinctive alteration assemblage characteristic of the fluid and its temperature. The fluids are retained within a permeable horizon forming a reservoir in which mineral-fluid equilibria are approached, and a suite of secondary alteration minerals is established. Chloride concentrations are usually in the thousands of milligrams per kilometre range, up to about 10000mg kg -I. Less commonly, CI levels exceed 100000rag kg -~ in more saline systems (e.g. Salton Sea, California, USA). In such systems formation waters or seawater may have mixed with the original chloride fluid. Other main constituents include sodium and potassium (often in a c. 10:1 concentration ratio) as the principal cations, with significant concentrations of silica (higher concentrations with increasing temperature at depth) and boron. Sulphate and bicarbonate concentrations are variable, but are commonly several orders of magnitude less than that of chloride. Carbon dioxide, and much lower levels of hydrogen sulphide, are the main dissolved gases. Areas which contain hot, large-flow springs with the greatest C1 concentration are fed more directly from the deep reservoir, and identify permeable zones within the field. However, these areas may not necessarily overlie the major upflow zone since the local topography and

lateral permeable horizons can exert a significant control on the hydrology. Such discharge features are invariably surrounded by silica sinter, which provides a guide to extinct hot springs, pools and geysers, and to subsurface temperatures in excess of c. 200°C. High-gas fields with significant bicarbonate concentrations can deposit an intimate mixture of sinter and travertine. Sub-surface silicification is commonly produced by chloride fluids, while the alteration assemblage is of argillic-propylitic type, with the following characteristic minerals: silica (amorphous silica, cristobalite, quartz) albite, adularia, illite, chlorite, epidote, zeolites, calcite, pyrite, pyrrhotite and base-metal sulphides (Browne 1978, 1991).

Origin of water and solutes. The water constituting the geothermal fluid can be derived from a number of sources. It may represent surface (meteoric) water which has gained depths of several kilometres through fractures and permeable horizons, or it can be water which was buried along with the host sediments (formation or connate waters). Other sources of water in geothermal systems have been suggested; these include waters evolved in metamorphism (metamorphic waters) and from magmas (juvenile waters), but the importance of these sources of water is uncertain (White 1957). Magmas were initially thought to be the source of the water, solutes and heat of geothermal systems. However, this model was radically changed in the early 1960s when it was demonstrated that the fluids were of dominantly meteoric origin, and solutes could be derived from rock-water reactions. Work on the isotopic signature of the fluids by Craig (1963) showed that they had the same deuterium signature as that of local meteoric water and could not be magmatic. In a series of now classic studies in aqueous geochemistry, Ellis & Mahon (1964, 1967) and Mahon (1967) demonstrated that all the solutes in geothermal fluids could be derived from reactions between the meteoric groundwater and the host lithologies. Later experiments with seawater and basalt (e.g. Bischoff et al. 1981) produced solutions of similar chemistry to seawater-influenced geothermal systems such as those in Iceland. Rock-water reaction is therefore thought to be the major source for many of the solutes, although they may also be contributed by mixing with formation waters, seawater or magmatic fluids. While there is no doubt that the geothermal fluids are of a predominantly meteoric origin, there is sufficient latitude in the isotope data to permit 5-10% of the fluid to be from an

GEOTHERMAL FLUID CHEMISTRY alternative source, possibly a magmatic fluid. Mixing with even a small amount of magmatic fluid would significantly affect the chemistry of the final geothermal fluid, and isotope determinations cannot discount a magmatic contribution subsequently diluted by meteoric waters. However, mass balance considerations using typical values for water-rock ratios, the chloride content of host rocks, the concentration of chloride in waters from high-temperature systems, the aerial extent of geothermal systems, and the duration of geothermal activity indicate that unrealistically large volumes of rock would have to be leached over the lifetime of a geothermal system. A small but significant magmatic contribution to the geothermal fluid is therefore thought to be likely. Density differences would, however, preclude any intimate mixing between meteoric waters and a magmatic fluid. If small pulses of magmatic fluid did enter the geothermal convection cell then, while not detectable isotopically, they would make a major contribution to the solute composition. Such fluids would be at temperatures in excess of 400°C and be rich in solutes such as C1, SO2 and CO2. Although the extent to which mixing may occur is uncertain, recent analytical advances now make it possible to distinguish between the isotopes of chlorine and boron. Information from these isotopes may enable a model of magmatic fluid-meteoric water mixing to be derived.

Secondary fluids As the chloride fluids leave the reservoir and ascend to the surface they may boil to create a two-phase (steam + liquid) boiling zone. The residual chloride water can discharge at the surface in hot springs or travel laterally to finally emerge many kilometres from the upflow zone. The vapours from this boiling zone may migrate to the surface independently of the liquid phase and discharge as fumaroles. Alternatively, the vapours may dissolve in groundwaters or condense in the cooler ground to form steamheated, acid sulphate and/or bicarbonate waters.

Sulphate waters. Also known as 'acid-sulphate waters', these are invariably superficial fluids formed by the condensation of geothermal gases into near-surface, oxygenated groundwater. The gases, with steam and other volatiles, were originally dissolved in the deep fluid but separated from the chloride waters following boiling at depth. They are found on the margins of a field some distance from a major upflow area, at

225

topographic levels high above the water table, in perched water tables and over boiling zones. Although usually found near the surface ( < c. 100m), sulphate waters can penetrate to depth through faults into the geothermal system. Here they are heated, take part in rock alteration reactions and mix with the ascending chloride fluids. Other sources of sulphatechloride mixed fluids are discussed by Ellis & Mahon (1977) and Nicholson (1993). Sulphate is the principal anion, and is formed by the oxidation of condensed hydrogen sulphide H2S(g) + 202(aq) = 2H+(aq) + SO42-(aq). This reaction, and the condensation of carbon dioxide, CO2(g) + H20(1) = HzCO3(aq) = H+(aq) + HCO3-(aq) = 2H+(aq) + CO32-(aq), produces protons, creating acid waters. Chloride occurs in trace amounts. Bicarbonate is either absent or at low concentrations since in very acid waters the dissolved carbonate is usually lost from solution as carbon dioxide gas (back-reaction of above equilibrium). Other volatile constituents which separate from the deep fluid on boiling may also condense into these waters (e.g. NH3, As, B) and attain significant concentrations. Nearsurface reactions between the acid waters and the surrounding county rocks may leach silica and metal cations (Na, K, Mg, Ca, A1, Fe etc.) which can thereby attain high concentrations in the waters. Acid-sulphate waters react rapidly and leach the host rocks to produce advanced argillic alteration, with kaolinite, halloysite, crystobalite and alunite as diagnostic minerals (Browne 1978, 1991). Extensive leaching of surface lithologies by these waters, or acidic steam, can produce a silica residue. This must be differentiated from silica sinter, which is a product of depositional, not alteration, processes. Anhydrite, hematitie, dickite, jarosite, pyrite, goethite-hematite mixtures and native sulphur are also commonly found.

Bicarbonate. These waters, which include those termed CO2-rich fluids and neutral bicarbonatesulphate waters, are the product of steam and gas condensation into poorly-oxygenated subsurface groundwaters. Such fluids can occur in an umbrella-shaped perched condensate zone overlying the geothermal system, and are common on the margins of fields. Bicarbonate waters found in non-volcanogenic, hightemperature systems (e.g. Hungary, Turkey and Africa) may constitute the deep reservoir fluid. The waters are of near-neutral pH as reaction with the local rocks (either in the shallow reservoir or during lateral flow) neutralises the initial acidity of these waters (see above

226

K. NICHOLSON

carbonate equilibrium). Loss of protons in such reactions produces near neutral waters with bicarbonate and sodium as the principal constituents. Sulphate may be present in variable amounts, with chloride at low concentrations or absent (Mahon et al. 1980). These waters are highly reactive, and their corrosive action on well casings needs to be taken into account in the development of a field (Hedenquist & Stewart 1985). These waters can precipitate extensive deposits of travertine (CaCO3), and can be. indicative of subsurface temperatures of below c. 150°C. An argillic alteration assemblage is typical of these waters, with the formation of clays (kaolinite, montmorillonite) and mordinite; some calcite and minor silicification may also occur (Browne 1978, 1991).

Steam formation: boiling point-depth relations. As a geothermal fluid ascends towards the surface, the pressure imposed upon it by the overlying column of water (hydrostatic pressure) will decrease. Eventually, the pressure will drop to a level which permits the dissolved gases and steam to separate from the liquid phase. This phase separation is commonly referred to as 'boiling'. It is one of the most important processes controlling the chemistry of liquid and vapour (i.e. water and steam) discharges. The relationship between boiling point and depth has been described by Haas (1971) and is illustrated in Fig. 3. The curve indicates the maximum temperature which water can attain at any given depth (or pressure), and therefore shows the depth at which a reservoir water at a given temperature will commence boiling. From this depth the boiling, or two-phase zone, can extend upwards towards the surface. The curve assumes that only hydrostatic pressure acts upon the fluid. In practice, however, it has been found that hydrodynamic pressures exist at depth in a geothermal system at about 10% above hydrostatic pressure. This excess pressure is necessary to maintain flow through the system. It is created by the buoyancy of hot water relative to cold water recharge and by a hydrostatic head in recharge waters from areas of greater relief (Grant et al. 1982; Henley 1985). This means that higher temperatures can exist at shallower depths than indicated by the curve, and therefore that boiling will occur at shallower depths. An increase in the salinity of water lowers the vapour pressure of water, raises the curve and prevents boiling until shallower depths are attained (Sutton & McNabb 1977). However, for most geothermal systems, the fluids are dilute and small changes in salinity will not

Temperature (*C) 1 O0

200

300

400

800

1200

1600

4.4% C 0 2

I

\

(1.0m) .]~

Fig. 3. Boiling-point-depth relationship under hydrostatic conditions (Henley 1985). Note that increases in salinity and gas content have opposite effects on the boiling-point-depth profile. For a fluid at a given temperature, increasing the salinity prevents the fluid boiling until shallower depths are attained; by contrast, increasing the gas content permits the fluid to boil at greater depths. Note too that increasing the salinity has only a slight effect on the boiling profile, while relatively small increases in the gas content of the fluid significantly alters the boiling-depth relationship.

significantly alter the boiling point-depth profile of the system. More significant however, is the gas content of the fluid. The presence of several weight percent gas in the fluid will depress the isotherms in a system below the usual boiling point-depth curve. This means that boiling zones in high-gas systems will appear at far greater depths than for gas-poor systems, which follow the relationship for pure water (Fig. 3).

Hydrological profile of geothermal systems An idealized thermal, hydrological and chemical profile of a high-temperature liquid-dominated dynamic system in low-relief terrain is illustrated in Fig. 2. The hydrology and distribution of discharge features is controlled by topography and vertical or horizontal permeable structures, e.g. faults and the conduits produced by hydrothermal eruptions. Vapour-fed discharge features occupy higher ground than the chloridewater springs, and include fumaroles, hot springs discharging steam-heated waters and steaming ground. These can form above boiling zones where carbon dioxide and hydrogen

GEOTHERMAL FLUID CHEMISTRY

227

Table 1. Geochemical-mineralogicalcharacteristicsof hydrological regimes within high-enthalpy geothermal systems Regime

Fluids

Palaeosurface

CI HCO3

Steam zone

SO4, HCO3

Boiling zone (two-phase)

CI + gas

Deep fluid zone

CI

Minerals/deposits Silica sinter carbonate sinter eruption breccia Kaolin, hematite silica residue, cristobalite Quartz, adularia bladed calcite * Silicification, albite adularia t, chlorite epidote, zeblites

Geochemical enrichment As, Sb precipitates Sr As, Sb, Hg, NH3

SiO2, K

* May be pseudomorphed by quartz. 3 Adularia present in greatest amounts in permeable structures. sulphide dissolve into groundwaters or steam condensates. Bicarbonate waters (CO2-rich steam heated waters) are also common on the margins of many fields. Systems in mountainous terrain show a similar profile, but the Cl-fluid rarely attains the surface near the upflow zone and undergoes up to tens of kilometres of lateral flow; additionally, the steam zone will be deeper and more extensive. The dynamic nature of the system is shown by the cycle: meteoric water descent, geothermal fluid formation, (replenished by meteoric waters descending from the recharge zone) and surface discharge of geothermal waters and vapours through springs and fumaroles (Fig. 2). However, fluid flow from depth is unlikely to follow the idealised vertical path shown in Fig. 2, and some degree of lateral flow is probable. Local topographic and permeability influences will exert an important control on the direction of flow and area of discharge. Many systems display lateral flow structures created by strong hydraulic gradients. These in turn are formed due to high relief, often with a near-surface low-permeability horizon. Cooling by conduction and groundwater mixing are reflected in the chemistry of the discharges. Even in low-relief (< c. 250 m) settings, including those typical of silicic volcanic terrain (e.g. Taupo Volcanic Zone, New Zealand), nearsurface lateral flows can extend for several kilometres. This is greatly extended in terrain of high relief ( > c . 1000m), typical of andesitic volcanoes, where flows are 10-50 km in length.

Hydrological regimes and their use in gold exploration In epithermal deposits gold is commonly found in, or just above, boiling zones and zones of fluid

mixing. Both processes lead to cooling of the fluid, loss of HzS and dissociation of the gold-complexing sulphide ligand (Seward 1973, 1991). Recognition of such zones in both geothermal and epithermal systems, together with other parallels, has bound these systems together as modern and ancient equivalents. As the geothermal fluid rises from depth it may cool adiabatically and boil, forming a two-phase zone composed of liquid water, water vapour and gases ('steam'). Since gold commonly deposits in or near the boiling zone, a favourable exploration target requires that this zone and the overlying steam zone are preserved. However, since the ore-depositing zones are relatively shallow features (often < 1 km depth), ancient geothermal systems (especially pre-Tertiary) may have been eroded to depths below the ore zone. Gold exploration programmes therefore need to be tailored towards three objectives: recognition of remnants of geothermal deposits; identifying the level of preservation of the system; location of the ore zone. These objectives can be achieved by reconstruction of the palaeohydrology of the system through hydrological regime recognition. To do this requires an appreciation of the hydrological profile of a geothermal system derived from the physical and chemical process acting on the geothermal fluid as it ascends to the surface. These processes produce hydrological zones in which each fluid type leaves diagnostic mineralogical and geochemical signatures in the surrounding lithologies through mineral-fluid reaction products (the alteration mineral assemblage), elemental enrichments and mineral deposits (Tables 1, 2). Criteria diagnostic of each hydrological regime are summarized in Table 1. Recognition of these signatures enables

228

K. NICHOLSON

Table 2. Characterbticwater-rock interaction products for contrastinggeothermalfluids CI fluid Amorphous silica, quartz, adularia, albite, illite, chlorite, epidote, zeolites, sulphides

SO4 waters

HCO3 waters

Kaolin alunite, smectite, hematite, sulphates

Kaolin, calcite, illite, montmorillorite

Note relative positions of the fluids in the hydrological structure of the system (Fig. 3). Direction of change within upflow or outflow structures Increasing depth in a system or approach to permeable upflow structure. ancient geothermal deposits to be identified and the relative position in the hydrological profile to be established.

Palaeosurface recognition If the palaeosurface can be recognised, this indicates a high level of preservation and greatly assists interpretation. This can often only be confidently done through discovery and recognition of geothermal deposits, notably silica sinters and hydrothermal eruption breccias.

Hydrothermal eruption breccias. These often present conduits for geothermal fluids and are valuable guides to locating epithermal gold exploration targets. The characteristics of these breccias have been detailed by several authors (Nairn & Solia 1989; Collar 1985; Hedenquist & Henley 1985; Nelson & Giles 1985). They are typically matrix-dominated with free-floating angular to sub-rounded clasts of local lithologies. Silica sinters. The presence of a thick sinter deposit not only indicates a palaeosurface but demonstrates that the palaeo-reservoir temperature was in excess of c. 200°C and marks areas of chloride water discharge (the golddepositing fluid, Brown 1986). Textural and geochemical characteristics which may be used to discriminate sinters from other siliceous lithologies have been described elsewhere (Nicholson 1988, 1989, 1993; Nicholson & Aquino 1989; White et al. 1989; Nicholson & Parker 1990; Parker & Nicholson 1990; Browne 1991). Surface sinter textures include striations, mushrooms, banding, ripples, terraces, circular rims, overhangs, plant material and boxworks,

while in cross-section columnar textures are distinct. Thin sections of sinter may show rods and filament textures from biogenic activity. Geochemical enrichments in B, C1, As, Sb and Hg may also be diagnostic.

Permeable (steam) zone recognition Surficial deposits such as sinters and eruption breccias commonly do not survive; furthermore, the boiling zone represents a narrow exploration target. Given the more pervasive nature of steam-rock reactions and the greater aerial extent of the steam zone (Fig. 2) it is this hydrological horizon which presents a better prospect for exploration surveys. Steam from the boiling zone may migrate to the surface independent of the liquid phase and is often able to take a more direct ascent to the surface through horizons which are impermeable to liquid water. Steam reacts with the rocks in three ways: directly, by condensation, and by heating/ dissolving in near-surface meteoric waters. Steam condensates and steam-heated waters form the acidic sulphate and bicarbonate waters found near-surface (SO4-HCO3-type, steam zone, Fig. 2; Tables 1,2). More subtle indicators of permeable zones, the consequence of fluid-rock interaction, may also be detected. Although geochemical surveys are generally less useful than mineralogical mapping in epithermal gold exploration, they are of value in the identification of permeable, upflow structures. Vapours and gases migrating directly to the surface over the boiling upflow zone can accumulate in the steam zone on the altered near-surface lithologies to form geochemical enrichments in As, Sb, NH3, Hg and possibly B (Nicholson 1993). Permeable structures can be identified by areas of the highest concentration of these species, and flow directions within a lithology may also be traced by the use of iso-concentration maps (flow direction is towards lower concentrations as more volatiles are lost from the fluid during early boiling).

Overprinting Fluctuations in the water-table, descent of sulphate fluids and expansion of the steam zone will all lead to an overprinting of alteration mineralogies (typically of sulphate and chloride fluids), complicating interpretation and hydrological reconstruction. However, using the hydrological model and criteria outlined (Tables 1, 2; Fig. 2), the level of preservation can be assessed and permeable upflow structures, which may have contained zones of boiling and

GEOTHERMAL FLUID CHEMISTRY

a

229

b

SP

I11

C o n a ~ etmling

Chloride (mg/kg)

~I~M ~

oG Chloride

"HS

(mg/kg)

Fig. 4. Example of a chloride-enthalpy mixing model. SP, steam point; R, reservoir fluid; HS, hot spring, cooling by boiling; SH, steam-heated waters; M, meteoric waters; C, C1, cooling by condensation, Well discharges which have experienced dilution plot on R-SH or R-M mixing lines.

Given that so much information on the geothermal reservoirs is obtainable from the fluid chemistry, is it possible to apply these geochemical techniques to oil-field waters to assist in the modelling of hydrocarbon reservoirs? The areas where geothermal geochemical techniques may be transferable are (i) identification of water mixing and water source correlation through the application of mixing models (ii) temporal and/or depth thermal profiles in the reservoir through the use of geothermometers (iii) correlation of formation waters.

non-producing formations through casing damage. In geothermal systems, mixing models enable dilution of the primary reservoir fluid, and the diluting fluid, to be recognized. They are used to monitor well chemistry during production to identify intrusion of waters from other formations. The most commonly employed model is a Cl-enthalpy plot (Fig. 4), but others such as silica-enthalpy, carbonate-based and statistical models have also been developed. Figure 4 shows the dilution lines to meteoric waters and steam-heated waters, both of which are lower in CI content and enthalpy than the reservoir fluid, on which diluted well chemistries would plot. Dilution trends in wells across a field can be identified by mixing models as well as temporal dilution changes in a given well. Although oil-field waters may not have a boiling component, the figure could still be used to illustrate mixing-dilution processes between waters of differing CI concentration and temperature.

Mixing models

Geothermometry

Mixing models applied to oil-field brines commonly use plots of species against Br (e.g. C1 v. Br) or oxygen-hydrogen isotope ratios to identify the composition of pre-mixing end-member waters (Egeberg & Aagaard 1989; Connolly et al. 1990). The geothermal mixing models could be similarly applied in hydrocarbon reservoirs to identify breakthrough of waters on the field margins ('edge water') and water intrusion from

The concentration of gaseous, solute and isotopic species involved in temperaturedependent equilibria in geothermal fluids have been calibrated to act as geothermometers (see Nicholson 1993 for a review). These can also be applied to oil-field water well discharges. Possible applications are: identification of water intrusion; reservoir monitoring during production and thermal profile modelling. Monitoring

gold deposition, located. This approach has been used with some success to identify palaeogeothermal systems and gold prospects in Australia and Scotland (White et al. 1989; Nicholson 1988, 1989).

Geothermal geochemical techniques and oil-field waters

230

K. NICHOLSON

Table 3. Equilibration times of some geothermometers expressed as to.5

Exchange

1 ROCK ADDITION K

t05 (at 250°C)

AlSO SO4-H20 AlSO COz--HzO A~80 H2Oo)-H20~v) AD H2-H20(1) AD CH4-H2 AD HzO~)-CH4

AI3CCH4-CO2 A13CCO2-HCO3 A34S SO4-H2S SiO2 geothermometer NaK geothermometer NaKCa geothermometer

c. 1 year 1-100 s 1-100 s 1-20 weeks 1-20 weeks 1-1000 years | 000--10 000 years 1-100 hours > 1000 years 1-100 hours 0.3 years 0.3 years

Si02 ~• IA

•B

Na

BICARBONATE S TEAM-HEATED WATERS

CI CHLORIDE WATERS

• Mg

I

I Ca

;4

• HCO3

SO4

ACID SULPHATE WATERS

the geothermometry of the produced waters may enable breakthrough of non-producing formation waters or edge waters to be recognized by unusual variations in geothermometry temperatures. The temperature indicated by geothermometers is that of the last equilibration, which is not necessarily that of the reservoir. Table 3 lists common isotope and solute geothermometers, together with estimates of equilibration times. In old reservoirs,

Fig. 5. Factor analysis model to illustrate major sources and control on solutes in geothermal well discharges from fields in the Taupo Volcanic Zone, New Zealand. Plot shows Factors 1 and 4 with interpreted sources/controls labelled (from Salvania & Nicholson 1990).

2

• RIVER WATERS

A SP~NG WATERS •

RESERVOIR WATERS

Fig. 6. Principal component analysis of freshwater chemistry. Plot of scores of Principle Components 1 and 2 discriminates between river waters, spring waters and surface reservoir waters (from Tse Tig Cheong & Nicholson 1993).

GEOTHERMAL FLUID CHEMISTRY and reservoirs with little or no dynamic recharge, all geothermometers should have attained equilibrium and each should indicate the reservoir temperature. However, as geothermometers with large t0.5 values ' r e m e m b e r ' temperatures longer than geothermometers which equilibrate rapidly, then in young fields and 'flushed' reservoirs differences in geotherm o m e t e r equilibration times may enable the thermal history of the reservoir to be established. A more general application of geothermometers, however, may be found during production from a field. Here differences in equilibration time can be used to develop a thermal profile either with depth in the area of a well, or across a field using data from several wells.

Correlation o f f o r m a t i o n waters Statistical models to identify common sources of solutes and geothermal waters was developed by Salvania & Nichoison (1990). These were based on multivariate analytical methods applied to the discharge chemistry of wells from several fields. Cluster analysis was able to group waters of common origin, while factor analysis identified the major sources and controls on solutes (Fig. 5). Furthermore, by applying factor analysis to freshwater chemistry, Tse Tig Cheong & Nicholson (1993) were able to effectively discriminate spring waters, river waters and surface reservoir waters (Fig. 6). This statistical approach would also be valuable in monitoring changes in well discharge chemistry, identification of mixing and may be applicable for the correlation of formation waters over large areas such as the North Sea. I am grateful to S. Arnorsson for valuable comments on the original manuscript.

References ARMSTEAD,H.C.H. 1983. Geothermal energy. Second edition, Spon, London. BIscnoFr, J.L., RADTKE, A.S. & ROSENBAUER,R.J. 1981. Hydrothermal alteration of greywacke by brine and seawater: roles of alteration and chlorite complexing on metal solubilities at 200° C and 350° C. Economic Geology, 79, 659--676. BROWN, K. 1986. Gold deposition from geothermal discharges in New Zealand. Economic Geology, 81,979-983. BROWNE, P.R.L. 1978. Hydrothermal alteration in active geothermal fields. Annual Reviews of Earth and Planetary Science, 6,229-250. 1991. Mineralogical guides to interpreting shallow paleohydroiogy of epithermal mineral depositing environments. In: FREESTONE, D.H.,

231

BROWNE,P.R.L. & SCOTT,G.S. (eds) Proceedings 13th New Zealand Geothermal Workshop, 263269. COLLAR, R.J. 1985. Hydrothermal eruptions in the Rotokawa geothermal system, Taupo Volcanic Zone, New Zealand. Geothermal Institute Report 14, University of Auckland, New Zealand. CONNOLLY,C.A., WALTER,L.M., BAADSGAARD,H. LONGSTAFFE,F.J. 1990. Origin and evolution of formation waters, Alberta Basin, Western Canada Sedimentary Basin. I. Chemistry. Applied Geochemistry, 5,375-395. CRAIG, H. 1953. The geochemistry of stable carbon isotopes. Geochimica et Cosmochimica Acta, 3, 53--92. CYAMEX, 1979. Massive deep-sea sulphide deposits discovered in the East Pacific Rise. Nature, 277, 523-528. EDWARDS, L.M., CHILINGAR,G.V., RIEKE, H.H. & FERTL, W.H. 1982. Handbook of geothermal energy. Gulf Publishing Co., Houston. EGEBERG, P.K. & AAGAARD, P. 1989. Origin and evolution of formation waters from oil fields on the Norwegian Shelf. Applied Geochemistry, 4, 131-142. ELLIS, A.J. & MAHON,W.A.J. 1964. Natural hydrothermal systems and experimental hot-water/rock interactions. Geochimica et Cosmochimica Acta, 28, 1323-1357. & -1967. Natural hydrothermal systems and experimental hot-water/rock interactions (Part II). Geochimica et Cosmochimica Acta, 31, 519-538. & - - 1977. Chemistry and geothermal systems. Academic Press, New York. GRANT,M.A., DONALDSON,I.A. & BIXLEY,P.F. 1982. Geothermal reservoir engineering. Academic Press, New York. HAAS, J.L. 1971. The effect of salinity on the maximum thermal gradient of a hydrothermai system at hydrostatic pressure. Economic Geology, 66,940-946. HANNINGTON,M.D., HERZIG,P. & ScoTT, S.D. 1991. Auriferous hydrothermal precipitates on the modern seafloor. In: FOSTER, R.P. (ed.) Gold metallogeny and exploration. BIackie, Glasgow, 250-282. HEDENQUIST, J.W. • HENLEY, R.W. 1985. Hydrothermal eruptions in the Waiotapu geothermal system, New Zealand: Their origin, associated breccias and relation to precious metal mineraliNation. Economic Geology, 80, 1640-1668. & 1985. Natural CO~-rich steam-heated waters in the Broadlands-Ohakki geothermal system, New Zealand: Their chemistry, distribution and corrosive nature. Geothermal Resources Council Transactions, 9,245-250. HENLEY, R.W. 1985. The geothermal framework for epithermal deposits. In: BERGER,B.R. & BETHKE, P.M. (eds) Geology and geochemistry of epithermal systems. Reviews in Economic Geology, 2, Society of Economic Geologists, 1-24. 1991. Epithermal gold deposits in volcanic

232

K. NICHOLSON

terranes. In: FOSTER,R.P. (ed.) Gold metallogeny and exploration. Blackie, Glasgow, 133-164. & ELLIS, A.J. 1983. Geothermal systems ancient and modern: A geochemical review. Earth Science Reviews, 19, 1-50. --, TRUESDELL, A.H. & BARTON, P.B. 1984. Fluid-mineral equilibria in hydrothermal systems. Reviews in Economic Geology l, Society of Economic Geologists. MAHON, W.A.J. 1967. Natural hydrothermal systems and the reaction of hot water with sedimentary rocks. New Zealand Journal of Science, 10, 206-221. --, KLYEN, L.E. & RHODE, M. 1980. Neutral sodium/bicarbonate/sulphate hot waters in geothermal systems. Chinetsu (Journal Japan Geothermal Energy Association), 17, 11-24. MATRAY, J.M. & FONTES, J.CH. 1993. Multi-sourced formation waters from the Dogger aquifer; Paris basin. In: PARNELL,J., RUFFELL, A.H. & MOLES, N.H. (eds) Geofluids '93,304-308. NAIRN, I.A. & SOLIA, W. 1980. Late Quaternary hydrothermal explosion breccias at Kawerau geothermal field New Zealand. Bulletin of Volcanology, 43, 1-13. NELSON, C.E. & GILES, D.L. 1985. Hydrothermal eruption mechanisms and hot spring gold deposits. Economic Geology, 80, 1633-1639. NICHOLSON, K. 1988. Geothermal deposits in ancient terrain as a tool in epithermal gold exploration: examples from Scotland. In: MCKIBBIN, R. (ed) Proceedings lOth New Zealand Geothermal Workshop, Auckland, 151-153. 1989. Early Devonian geothermal systems in northeast Scotland: Exploration targets for epithermal gold. Geology, 17,568-571. -1993. Geothermal fluids: Chemistry and exploration techniques. Springer-Verlag, Berlin, 268pp. & AQUINO, C. 1989. Life in geothermal systems. A key to sinter formation and recognition? In: BROWNE,P.R.L. & NICHOLSON,K. (eds) Proceedings llth NZ Geothermal Workshop, Auckland University, 143-148. & PARKER,R.J. 1990. Geothermal sinter chemis-

-

-

-

-

-

-

try: Towards a diagnostic signature and a sinter geothermometer. In: HARVEY, C.C., BROWNE, P.R.L., FREESTONE, D.H. & SCOTT, G.L. (eds) Proceedings 12th NZ Geothermal Workshop, Auckland University, 97-102. PARKER, R.J. & NtCHOLSON, K. 1990. Arsenic in geothermal sinters: Determination and implications for mineral exploration. In: HARVEY, C.C., BROWNE, P.R.L., FREESTONE, D.H. & ScoxT, G.L. (eds) Proceedings 12th NZ Geothermal Workshop, Auckland University, 35-39. SALVANIA, N.V. ,~ NICHOLSON, K. 1990. Chemometrics applied to the fluid chemistry of geothermal fields in the Taupo Volcanic Zone, New Zealand. In: HARVEY, C.C., BROWNE, P.R.L., FREESTONE, D.H. &ScoTr, G.L. (eds) Proceedings 12th NZ Geothermal Workshop, Auckland, 157-163. SAUNDERS, J.A. & SWAN, C.T. 1990. Trace metal content of Mississippi oil field brines. Journal of Geochemical Exploration, 37, 171-183. SEWARD,T.M. 1973. Thio complexes of gold and the transport of gold in hydrothermal ore solutions. Geochimica et Cosmochimica Acta, 37,379-399. 1991. The hydrothermal geochemistry of gold. In: FOSTER, R.P. (ed.) Gold metallogeny and exploration. Blackie, Glasgow, 37-62. SUTTON, F.M. & McNABB, A. 1977. Boiling curves at Broadlands geothermal field, New Zealand. New Zealand Journal of Science, 20, 333-337. SVERJENSKY,D.A. 1984. Oil field brines as ore forming solutions. Economic Geology, 79, 23-37. TSE TIG CHEONG,M. • NICHOLSON,K. 1993. Identification of element associations and controls on freshwater chemistry in Grampian Region, Scotland (Abs). llth European Symposium on Environmental Geochemistry and Health, Aberystwyth, April, 1993. WHtTE, D.E. 1957. Thermal waters of volcanic origin. Geological Society of America Bulletin, 68, 16371658. WHITE, N.C., WOOD, D.G. & LEE, M.C. 1989. Epithermal sinters of Paleozoic age in north Queensland, Australia. Geology, 17,718-722.

An integrated approach to the study of primary petroleum migration ULRICH

MANN

Forschungszentrum Jiilich GmbH, Institut fiir Erd61 und Organische Geochemie, D 52425-Jiilich, Germany Abstract: The precise mechanism of primary petroleum migration has been elusive despite intensive investigation and discussion. There is more than one mechanism, and the pathways and efficiencies of individual mechanisms vary from case history to case history due to the variable abundances of micropores, macropores and fractures in source rocks as well as different sources for the build-up of a pressure gradient. Primary migration probably proceeds as diffusion through source rock micropores via mesopores to macropores and fractures. From there, a petroleum bulk phase develops, and moves along the macropore and fracture system, possibly together with aqueous solutions. In fact, many parameters influence petroleum transport out of a source rock, and all of them have seldom been checked by exploration geologists as most mature source rocks are inaccessible. In order to collect information about primary petroleum migration that is as complete as possible, an integrated approach consisting of sedimentology, petrophysics, organic geochemistry and numerical modelling should furnish geologically acceptable results: (i) by sedimentological methods, all potential primary migration pathways are identified, (ii) with the help of petrophysical methods, non-effective migration pathways are separated from effective pathways and excluded, (iii) organic geochemical results provide control and direct evidence for petroleum expulsion stage and efficiency, and (iv) numerical modelling finally quantifies all observed effects within the framework of the geological situation.

Primary petroleum migration is one of the principal processes in the formation of petroleum accumulations. Considerable progress has been achieved in the recognition of the effects and causes of primary migration effects within the last 15-20 years. Nevertheless, the process of primary migration is still under intensive scientific debate. Several contributing mechanisms exist which vary from source rock to source rock and from sedimentary basin to sedimentary basin, and as primary migration occurred in the geological past it is difficult to prove. Besides petroleum geochemistry text books (Hunt 1979; Tissot & Welte 1984), petroleum migration in source rocks has been the subject of several extensive reviews, including Durand (1983, 1988), England et al. (1987), Ungerer (1990), Price (1989), and Welte (1987). The approach for the review presented here follows the applied analytical view of the geochemist rather than that of the exploration geologist. First, a short review of the possible modes of the primary migration process itself should familiarize the reader with the present state of knowledge. Second, several relevant research topics, such as experimental simulation methods, will be addressed as keys to the understanding of the petroleum migration process. Thereafter, most

of this review is dedicated to the methodological approach: how petroleum migration in source rocks can be analysed in order to find out if, when, where and how much hydrocarbons did or did not migrate. Certainly, this approach is not limited to petroleum migration alone and includes many similarities to the analysis of other migrating subsurface fluids. It intends to demonstrate that not only modern organic geochemistry but also other geosciences disciplines are necessary to understand primary migration sensu lato. Such an integrated approach is required especially for a more precise quantitative understanding of petroleum migration in source rocks. A reliable approach has to integrate sedimentology, petrophysics, organic geochemistry, and numerical simulation in such a way that the individual methods fit the relevant case study. In order to provide an overview of the experimental techniques, analyses and individual steps of numerical simulation (according to the four disciplines mentioned), Table 1 summarizes several of the most common methods of investigation. Definition

Movement of petroleum from the source via carrier bed to the reservoir rocks is called

From PARNELL,J. (ed.), 1994, Geofluids: Origin, Migrationand Evolution of Fluids in Sedimentary Basins, Geological Society Special Publication No. 78,233-260.

233

234

U. MANN

Table 1. Experimental and analytical techniques and numerical verification steps for an integrated approach to petroleum migration in source rocks Sedimentology

Petrophysics

Organic geochemistry

Numerical simulation

Inorganic petrography

Helium pycnometry

Organic petrography

Cathodo-luminescence

Gas adsorption

Fluid inclusion analysis

Absolute permeability versus stress Mercury porosimetry

Total organic carbon, pyrolysis Thermodesorption GC

Source potential, temperature history Overpressure development Flow modelling

SEM, Cryo-SEM Carbon and oxygen isotopes

Liquid and gas chromatography Wettability, contact angle Gas chromatographymeasurement mass spectrometry

migration. It is subdivided into primary migration, which is defined as the movement of oil and gas through and out of the fine-grained source rocks, and secondary migration, the movement through wider pores to the trap. The loss of petroleum from the trap is called dismigration. This paper reviews the migration which takes place directly after petroleum generation, the process of primary migration. In this review, petroleum and hydrocarbons also include NSO-compounds if not explicitly excluded. Bitumen is applied as a generic term to natural inflammable substances of variable colour, hardness, and volatility, composed principally of a mixture of organic compounds (mainly hydrocarbons). It is applied in order to indicate a residual part of petroleum which has been lost during migration (residual oil, shows). Soluble organic matter may represent that part of the bitumen which can be dissolved by an organic solvent, but in general, it represents the organic solvent extract of any rock sample.

Modes of primary migration Mechanisms Several mechanisms seem to work alone or together in different types of source rocks and sedimentary basins, depending on the source rock properties like richness, sedimentological and petrophysical attributes, and the prevailing pressure-temperature conditions as well as the potential of the geological conditions for overpressure formation and microfracturing. According to most authors, the most important form of primary migration during the main phase of oil and gas formation seems to be discrete hydrocarbon phase movements (Dickey 1975; Hunt 1979; Momper 1978; Tissot & Welte 1984; Welte 1987; Durand 1988; Ungerer 1990).

Fracture modelling Migration pathway modelling, sensitivity analysis

However, as true hydrocarbon flow cannot take place in micro- and mesopores of a source rock, yet petroleum has to reach the macropores, diffusion may be the dominating process at the beginning of the primary migration of a hydrocarbon molecule. A sketch of the possible mechanisms from generation to expulsion is presented in Fig. 1. Under standard geological conditions, diffusion alone seems several orders of magnitude too small in order to fill a petroleum reservoir (Thomas & Clouse 1990). For the same reason, the postulation by Stainforth & Reinders (1990) that active diffusion of bitumen molecules through organic matter networks represents the rate-limiting petroleum migration process may only hold for short migration distances and/or relatively poor source rocks. According to Hunt's (1979) calculations, based on Fick's law, migration paths for diffusion should be short, between 'tens to hundred of feet', in order to account for a commercial oil accumulation by diffusion in a reasonable time scale such as 10 million years. However, fluid movement can augment a diffusion process, therefore two or three migration mechanisms acting in parallel would increase the distance of migration. Diffusion is thus the most likely initial mechanism, directly after petroleum generation up to the time a pressure-driven flow takes over. Hunt (1973) and Dickey (1975) both have suggested that water may limit the free pore space sufficiently to permit oil-phase migration. Since an oil droplet would be restricted to the bulk water phase, it is conceivable that at some depth the oil in the smaller pores would occupy enough of the bulk space to be expelled by capillary forces assisted by the fluid potential gradient into a coarser-grained (= lower capillary pressure) zone of a source rock. Water is not needed to explain oil-phase migration in

PRIMARY PETROLEUM MIGRATION

235

Pathway

Pore pressure solution feature

Kerogen

Diffusion

Deaorption

Aggregation

Bulk flow

Mechanism

Fig. 1. Primary migration mechanisms for petroleum in source rocks (after Mann et al. 1991). very organic-rich rocks, as such rocks may be partially oil-wet (Hunt 1979). Completely oilwet pores in source rocks will hardly be encountered, as the estimated minimum organic content for this situation is about 30% (Byramjee 1967). Therefore, most petroleum source rocks should provide a mixed wettability. The ratio of petroleum to water during primary migration was estimated for the Western Canada Basin by Hunt (1977). He calculated that the expelled water masses should have contained about 300-1000ppm oil which requires oil-phase migration, but a few parts per million hydrocarbon in water can be explained by a solution mechanism. Data by McAuliffe (1966, 1979) and Price (1976) for molecular solution limit efficient primary migration by solution in water to the most soluble light hydrocarbon fraction such as methane, ethane and benzene. Nevertheless, the fact that aromatics show a much better solubility than naphthenes, and naphthenes a better one than paraffins, can be used to prove whether solution in water may have contributed to specific primary migration avenues. Price (1976) could show a pronounced temperature effect for the solubility of oil in water: it increases from 25 to 100° C gradually by a factor of about two, but from 100 to 180°C by a factor of about five. Hydrocarbon transport in gaseous solution (proposed by Sokolov 1948) was re-evaluated by Price (1989) and Leythaeuser & Poelchau (1991). They advocated a mechanism that overcomes the often encountered low bitumen saturation in oxygen-rich (kerogen type III) organic matter. The hydrocarbon gathering problem is solved by a diffusive partitioning of hydrocarbons by aqueous solution through shale pore water to dispersed hydrocarbon gas bubbles.

From a theoretical point of view, the transportation of individual oil droplets seems possible, as colloidal and micellar solutions have been proposed in the past by several authors (Baker 1959; Meinschein 1959; Cordell 1972). However, 'physicochemical, geochemical and geological considerations make individual droplets and bubbles or colloidal and micellar solutions highly unlikely as an effective means of transport during primary migration' (Tissot & Welte 1984, p. 323). A consideration of the mobilization of oil in primary migration must take all the major components in the pore fluids into account. The precursors and by-products of oil generation from sedimentary organic matter may contribute significantly to the pore fluids. The experimental data by Bray & Foster (1980) support a migration process where carbon dioxide and hydrocarbon gases mobilize the liquid hydrocarbons so that they can leave the source rock with any water expelled during compaction. This process should be operable in source rocks that have pore water as the principal pore fluid. Carbon dioxide from great depth may also support primary migration of hydrocarbons (Kvenvolden & Claypool 1980). Micro fracturing

In view of the fact that source rock lithologies are generally regarded as having a low permeability, geologists have considered other potential petroleum pathways. Several authors suggested that oil generation may induce microfractures in source rocks (Snarsky 1962; Tissot & Pelet 1971; Momper 1978; Palciauskas & Domenico 1980; du Rouchet 1981; Ozkaya 1988; Lehner 1991; Diippenbecker & Welte 1992). The term 'hydraulic fracturing' has been borrowed from the petroleum industry, where it

236

U. MANN

is known as a technique of well stimulation by means of pressure applied in bore holes, and it is used to describe the lithostatic failure by high pore pressure in overpressured reservoir regimes. For differentiation, a more specific and genetic term like 'expulsion fracturing' seems preferable. According to Palciauskas & Domenico (1980), a relationship may exist between the microfracture development and hydrocarbon-maturation kinetics. They conclude that, once initiated, the average rate of microfracture propagation would keep pace with the rate of sediment burial. However, microfractures can seldom be confirmed as a migration pathway by direct observations because in most clay-shale source rocks they will probably reheal completely. This reason, the necessary excess pressure for fracturing in relation to the petrophysical properties as calculated by Sandvik & Mercer (1990), and the higher ductility of source rocks (in contrast to reservoirs), might be the main arguments why many exploration geologists are still sceptical about the concept of microfractures. On the other hand, in a thick, organic-rich, rapidly heated source rock which generates petroleum at high rates, the pore pressure may be in excess of the mechanical strength of the rock and open fractures. Long geological times may greatly compensate for low permeabilities and low flow rates. Therefore, primary and secondary porosity of a source rock may be sufficient to serve as migration pathway.

Potential driving forces for primary migration To obtain petroleum flow within macropores or fractures of a source rock, a pressure gradient is needed. Several kinds of geological processes may contribute to varying degrees. The weight of the overburden can partly be transferred from the solid grains to the pore fluids. During generation, the transformation of part of the kerogen to hydrocarbons results in the partial transmission of the geostatic load from the solid phase to the liquid phase. This situation arises when the permeability of sediments is so low that the pore fluids cannot be squeezed out fast enough. Compaction seems to represent the most common case in order to provide the driving force for expulsion (Meissner 1978; Hunt 1979; du Rouchet 1981). However, several other processes may contribute to pressure generation or influence the direction of the pressure gradient. In deeper petroleum systems, flow directions of primary

migration may be controlled by the configuration and internal pressures of individual, seal-bounded fluid compartments (Vandenbrouke et al. 1983; Demaison 1984). The deeper petroleum systems, where the oil is generated, do not exhibit one basin-wide compartment with hydrostatic pressure such as in most shallow systems, but consist of a series of individual fluid compartments that are not in hydraulic pressure communication with each other nor with the overlying hydrodynamic regime (Hunt 1990). Oil and gas migrates only downwards from a source rock to an underlying permeable rock which releases pressure and allows the petroleum potential (for migration) to continually decrease. The petroleum potential is given as the sum of the overpressure, the buoyancy potential and the capillary potential (England et al. 1987). Therefore, the overpressure must be sufficiently high in order that downward flow becomes possible. The gradient required to overcome buoyancy is such that it is unlikely that petroleum is able to migrate more than 300--500 m downwards from a source rock (Mann & Mackenzie 1990). The volume increase of organic matter due to the solid-liquid conversion during hydrocarbon generation provides an internal pressure source in the source rock (Snarsky 1962; Sokolov et al. 1964; Hedberg 1974; Momper 1978). Ungerer et al. (1983) and Goff (1983) estimated that this may be as high as between 10 and 20% of the initial volume. Although clay mineral dehydration contributes a remarkable water volume, most water leaves a siliciclastic source rock prior to petroleum expulsion (Burst 1969; Perry & Hower 1972). Nevertheless, it is tempting to propose a genetic connection between smectite dehydration (illitization) and hydrocarbon migration, not only because enormous volumes of water are involved, but especially because the temperature necessary for intense oil generation in source rocks appears, in many areas, to roughly coincide with the zone of abrupt illitization. This relationship may be fortuitious, but on the other hand, a genetic relationship may exist for selected basins such as the Gulf of Mexico (Bruce 1984). However, the existence of clay-poor carbonate source rocks indicates that water from clay mineral dehydration is certainly not a requirement for primary migration (e.g. Grabowsky 1984). If oil droplets block the pore openings in a source rock, then the water trapped behind the oil will expand as the temperature rises. The increased pressure caused by the thermal expansion of water has been termed aquathermal

PRIMARY PETROLEUM MIGRATION

pressuring by Barker (1972). According to Barker's diagram (1972, fig. 1) the pressure increase because of complete isolation is significant. Although results from subsurface and compaction data from the Gulf Coast (Magara 1975) seem to demonstrate the possible presence of aquathermal pressuring effects, Daines (1982) concludes that this mechanism could be responsible for abnormal pore pressures only in shallow, impermeable sediments, particularly in areas of high geothermal gradients. Local seals of carbonate cement may form directly above and below solution seams, and may act as a pressure reinforcement for expulsion (di Primio 1990). This would point to an accelerated expulsion process from solution seams and supports Jones's (1984) thesis for an 'explosive migration' in many carbonate source rocks. Other local sealing effects formed by cementation during diagenesis may help to bring about an internal pressure build-up within a source rock and support petroleum expulsion.

Keys to the process of primary migration The simulation of migration in the laboratory represents an alternative approach for elucidating the effects of petroleum migration from geological case studies. The advantages in the laboratory are a better control over maturation, generation, degradation, organofacies and mixing effects which generally complicate the interpretation of migration effects in natural settings. On the other hand, flow or diffusion experiments in the laboratory reflect a unique optimised system in which a small quantity of oil with a well-known composition is passed through a well-defined mineral phase path at a relatively constant rate and for a very short time. It must be expected that very subtle changes in the parameter configuration of a system may change the results tremendously. All laboratory experiments are unreal, and extrapolations to the geological situation have to be accepted for what they are. As a tool to simulate generation and expulsion of petroleum at the same time, hydrous pyrolysis has become a widely used laboratory technique (introduced by Lewan et al. 1979). The technique maintains a liquid-water phase in contact with potential petroleum source rocks while they are heated at subcritical water temperatures. Although the generated and expelled oil which accumulates on the surface of the water can be quantitatively collected, the relationship of the amount and the composition of the expelled oil phase to the petrophysical properties of the rock

237

(during the actual expulsion process) and the petroleum flow itself cannot be investigated because mineralogy and petrophysical properties may change during heating. However, in contrast to many small-scale pyrolysis methods which, in general, are limited to the evaluation of the generation characteristics, hydrous pyrolysis as applied by Lewan (1985, 1987, 1992) can also consider rock properties relevant for expulsion, as for example mineral matrix or textural effects. If generation and migration are simulated together, the conclusions of numerous studies indicate that a minimum of effluents or water pressure is necessary to allow natural conditions to be reproduced and to promote hydrogenation reactions via hydrogen transfer. Observing the behaviour of a selected hydrocarbon compound or oil during permeation of a selected mineral or rock matrix with (= bulk flow) or without a pressure gradient (= molecular flow) represents one principal approach for understanding petroleum migration in source rocks. In principle, the effects of the individual expulsion parameters like pressure-temperature conditions, maturity stage of organic matter, water content, kerogen and migrating compound type, mineral matrix, petrophysical properties and others, should be investigated separately. In spite of the idealization of the system and the chance to vary only one parameter, most experiments are complicated by the fact that several mechanisms work hand-in-hand. In fact, petroleum migration in many source rocks may represent the net effect of bulk flow, diffusion (in solution or gas phase diffusion), solution (water and hydrocarbons), sorption and swelling effects (organic/organic and organic/inorganic) and possibly other processes. Therefore, although most of the quality of a migration experiment is controlled by the ability to monitor one single migration mechanisms, this makes the system increasingly unreal compared to the geological situation. In order to elucidate primary migration mechanisms, the analysis of the petroleum composition within pores of different size may represent another promising approach. Based on a relatively general approach to petroleum migration (not only source rocks) by Beletskaya and co-workers (Beletskaya 1972; and later references cited by Sajgo et al. 1983; Brukner & Vet6 1983) and by Sajgo et al. (1983), Ropertz (1993), re-evaluated this concept in more specific terms. His SPEX-method (Selective Pore EXtraction) provides step by step the individual extracts from large, small and closed pores of the same rock sample. Compositional differences were interpreted in accordance with

238

U. MANN

standard maturity parameters for oils known from organic geochemistry such as the MPI (Radke & Welte 1983) or from the relative composition of biomarkers. It was found that the composition differed especially during expulsion at the early maturity stage of source rocks, and that the maturity (according to the composition of the soluble organic matter content) increased with decreasing pore size. Ropertz (1993) concluded that at least during the early expulsion stage, petroleum migration predominantly takes place through a macropore system. This confirms earlier considerations by Momper (1978) and Honda & Magara (1982) who claimed that petroleum migration in source rocks takes place through the largest available pore throats, through so-called 'atypical' pores. Especially during early maturation of source rocks, this may produce mismatches in oilsource rock correlations (Price & Clayton 1992). With progressive generation and increasing oil saturation of the pore space, petroleum seems to move through the whole pore system, whereas the petroleum of the closed pores is hardly influenced by progressive generation (Ropertz 1993). This would mean that if no fractures or pressure solution features are present in a source rock, primary migration takes place as a channelling process via macropores. This would provide distinctive advantages for petroleum flow. The existence of such migration channels would enhance the potential for a more oil-wet expulsion pathway and better flow conditions. In addition, if selected petroleum precursor compounds, like organic acids, possess the capability to influence oil wettability in a positive way, then such channels which have experienced more intensive flow of water with the precursor compounds in solution should be most suited for petroleum expulsion.

Well logs From the explorationist's point of view, it is important to decipher first the correct direction of petroleum migration in order to interpret the drainage conditions of a potential trap correctly. The principal direction of primary migration can be either vertical or horizontal: according to Magara (1980), in relatively young sedimentary basins strong vertical potential differences are common in the intermediate depth range (1000-2500 m). In the older sedimentary rocks, such differences may have been dissipated, but shale-porosity profiles sometimes indicate conditions of undercompaction. England et al. (1987) have argued that most fluids (oil, gas, water) in rocks with permeabilities greater than

1 mD move laterally whereas most fluid movement in rocks with permeabilities less than 1 mD is vertical. However, porosity and permeability within the same source rock may vary considerably due to lithofacies changes or due to specific pathways like fractures or solution seams. This leads to an internal hydrocarbon redistribution within individual units of a source rock and/or different expulsion times for individual source beds. Diagenetically formed pathways may determine the direction of petroleum alone or in combination with the pore network, which is known from reservoir engineering as a dual porosity model. Based on comparisons of shale porosity curves from sonic logs with hydrocarbon extract data, Magara (1980) was able to provide evidence for intervals with depleted hydrocarbons which coincided with undercompacted zones or were in close association with reservoir sections. A practical model, the 'A log R technique', was developed by Passey etal. (1990) for organic richness from porosity and resistivity logs, but this method may also have capability to reveal individual hydrocarbon-depleted (= former migration avenues) or -enriched (= active migration avenues) intervals within a source rock formation. The method of Passey et al. (1990) uses an easily implemented curve overlay that is calibrated for organic richness and maturity. Direct evidence for hydrocarbon-depleted source rock intervals may be recognized from lowered resistivity values, as shown by Mann (1991) from a secondary porosity zone from the Lower Toarcian. In principle, well logs provide a fast and therefore economic quick-look method, because logs are available soon after completion or drilling. However, in many case studies, as for example in narrow interbedded shale-sand or carbonate--evaporite sequences, the use of log data for revealing migration trends will probably be insufficient due to limited depth resolution and accuracy. Especially for testing the effects of diagenetic migration pathways, a detailed approach using sedimentological and petrophysical techniques and methods will often be unavoidable. A combination of a porositypermeability and a geochemical parameter should be applied in order to reveal the correct direction for migration.

Sedimentological approach Lithofacies, as well as texture and bedding of the sedimentary rock, are the primary controlling parameters determined by and inherited from the specific depositional environment. They

PRIMARY PETROLEUM MIGRATION control the potential fluid migration pathways through the primary interparticle pore space of the rock. Compaction, chemical diagenesis, and fracturing are secondary controlling parameters during basin evolution which control the formation of specific migration pathways. Due to much more variation of the lithofacies in carbonate, evaporite and siliceous source rocks as well as due to a more complicated compaction and diagenetic history in contrast to siliciclastic source rocks, carbonate, evaporite and siliceous source rocks possess much more potential to develop migration pathways. According to the two genetically different types of migration avenues for petroleum migration in source rocks, first individual lithofacies aspects of the different source rock lithologies will be discussed, followed by various possibilities for the formation of specific migration pathways.

Lithology and lithofacies variation in source rocks During macro- and microscopic petrography, first of all the lithology-specific sedimentological properties have to be recorded. Shales, marls and argillaceous carbonates generally convey an impression of being homogeneous, dense and impermeable, except for an occasional fracture or other discontinuities. Scanning electron micrographs, however, often reveal a remarkably open fabric beside the unresolved microand mesopores, consisting of a network of regular and irregular macropores and vugs. Both detrital and secondary minerals may be present in a corroded or partly dissolved stage. Detailed optical investigations by normal light microscopy (e.g. Littke et al. 1988), incident-light fluorescence microscopy (e.g. Soeder 1990), and by scanning electron microscopy (e.g. Lindgreen 1987b) are a very effective way to gain insight by direct observation into the most permeable part of a source rock, and to reveal which features and minerals contribute to source rock porosity and permeability. Generally, nearly all detrital non-clay minerals (quartz, calcite, pyrite, dolomite siderite, halite, feldspars, barite etc.) provide permeability because their silt- and sand-size grains form larger pores. Secondary crystals of quartz or calcite commonly provide their own permeable haloes. Clusters or laminations of microfossils and silt lenses in source rocks are generally areas of relatively large pores. In carbonate and evaporite source rocks, pathways created by diagenetic redistribution of carbonate (see below) are much more common

239

and seem to provide the dominant expulsion avenues (Grabowski 1984; Pollastro & Scholle 1986; Comer & Hinch 1987; Denham & Tieh 1990; Lee 1991 ; Hofmann 1992). Good to excellent biogenic siliceous source rocks are known from the Monterey Formation from the San Joaquin basin in California (e.g. Isaacs et al. 1983). Diagenetically altered and unaltered diatoms make up silica, porcellanites (opal-A and opal-CT) and quartz cherts (e.g. Graham & Williams 1985). Such sedimentary constituents provide generally a high porosity, but at the same time a relatively low permeability. This explains why many Monterey source rocks are distinctive in yielding high S1 values during Rockeval pyrolysis (Graham & Williams 1985), an indicator of fluid hydrocarbons and/or solid bitumen, and why petroleum migration through the macropore network has not been reported from the Monterey Formation so far. Primary migration in the Monterey shales has recently been evaluated by Lee (1991). He observed oil-stained solution seams and fractures and concluded that these are the pathways for primary migration. This is very similar to the reservoirs in the Monterey Formation, where production depends on the content of detrital material (Issacs 1980) and the resulting degree of fracturability (Hornafius 1991).

Migration path ways by diagenetic dissolution In order to check potential diagenetic pathways for primary migration, the diagenetic evolution of each individual lithological and lithofacies source rock unit has to be retraced. By luminescence investigations, (trace) element analyses and by analyses of stable carbon or oxygen isotopes, it is possible to support microscopic observations about the genetic history. The composition and mineralogy of carbonate cement is controlled by the composition of pore water and the availability of particular cations. For instance, carbonate produced from the degradation of organic matter can have a distinctive mineralogy (Curtis 1978). Carbonate produced from bacterial sulphate reduction is usually iron-poor, and carbonate produced by bacterial fermentation or thermal decarboxylation can be iron-bearing or iron-rich (Irwin et al. 1977; Coleman et al. 1979). Furthermore, the iron content of carbonate cements in shaly sequences often increases with depth (Boles & Franks 1979; Irwin 1980; Matsumoto & Iijima 1981; Irwin & Hurst 1983).

240

U. MANN

Table 2. References for pressure solution features as primary migration pathways for petroleum

Formation, rock unit

Age

Locality

Lithology

Reference

Niobrara

Late Cretaceous

Denver Basin, USA Chalk, calc. shale

Austin Chalk Sunniland Limestone Niagaran

Late Cretaceous Early Cretaceous

Chalk Limestone Limestone

Budai et al. (1984)

Woodford

Shale, chert

La Luna

Late Devonian, Early Mississippian Late Cretaceous

Texas, USA SouthFlorida Basin, USA Michigan Basin, USA Oklahoma, USA

Pollastro & Scholle (1986) Grabowski (1984) Palacas (1984)

Limestone

Smackover

Late Jurassic

Maracaibo Basin, Venezuela Gulf Coast, USA

unknown formation Monterey

Triassic Mid- to Late Miocene Early Oligocene

Comer & Hinch (1987) Talukdar et al. (1987) Sassen et al. (1987), Denham & Tieh (1990) di Primio (1990) Lee (1991)

S-Unit

Silurian

Northern-Italy California, USA Mulhouse Basin, France

The carbon isotopic composition can be used to identify the origin of carbonate cements, and the oxygen isotopic composition for deduction of the precipitation temperature of the cement and the isotopic composition of the pore water at the time of precipitation (Irwin & Hurst 1983). Composite views of mudrock evolution during burial diagenesis have been proposed by Hower et al. (1976), Foscolos & Powell (1980) and Curtis (1983). According to Curtis (1983), calcite, K-feldspar, kaolinite and possibly other fine-fraction clays are potential candidates for dissolution during the 65-95°C catagenetic phase, and may thus provide secondary porosity for primary migration. Other possible causes for dissolution of carbonate may be the interaction with organic acids (Barth et al. 1988) or cooling formation waters (Bj0rlykke 1984; Giles & de Boer 1989). Due to the higher chemical reactivity of carbonate and evaporite minerals in contrast to clay minerals, several cementation, dissolution and re-precipitation generations are common in carbonate and evaporite source rocks and thus provide a great potential for primary petroleum migration pathways created by secondary porosity. A further cause for carbonate dissolution may be the intensified flow of formation waters in front of permeability barriers (Mann et al. 1990), because flow is 'concentrated through the more permeable horizons, out of the more porous mudrocks' (Curtis 1983). The nature of dewatering during

Lime mudstone Carbonates Siliciceous and calcaceous shale Evaporites

Hofmann (1992)

compaction undoubtedly occurs upwards, although not uniformly, but relative to any stratigraphic marker (Bonham 1980). Lindgreen (1985) investigated carbonate bands as primary migration pathways from the Kimmeridgian in the Central Graben of the North Sea. He found that ' . . . once the mobile phase has reached the large pores in the claystone, flow transport out of the source rock through these pores and the large ones in the carbonate bands should be the dominant process'. By a combination of X-ray diffraction, thermal analysis, M6ssbauer spectroscopy, thinsection studies, scanning microscopy, electron microprobe, and gas adsorption studies, Lindgreen (1985) was able to show the better permeability of the carbonate bands in comparison to the adj acent clay-shale. Similar carbonate bands were described by Prozorovich et al. (1973) from Volgian source rocks in western Siberia, and by Irwin (1980) from the Kimmeridge Clay of Dorset in England. Fibrous calcite veins (so-called 'beef') lying parallel to the bedding of certain sedimentary rocks may also indicate potential as primary migration pathway (Stoneley 1983). Another type of diagenetically-formed migration pathway from the Lower Toarcian in Germany was investigated by Mann (1990, 1991). Probably due to the uplift of the source rock during basin inversion, cementation of the original pathways did not take place, and it was

PRIMARY PETROLEUM MIGRATION

241

0 0

0

C a C 0 3 (%) 0 0

0

CaCo3(%) Pre-compaction

Post-compaction

Fig. 2. Stylolite formation due to re-distribution of carbonate: pre- and post-compaction stages (after Scholle & Halley 1985). possible to study relatively open migration pathways. Based on mineralogical, electronoptical and petrophysical investigations, it was possible to differentiate four diagenetically modified lithofacies types: a calcareous clayshale, a marlstone, a cemented marlstone, and a marlstone with secondary porosity. Although all facies had the same hydrocarbon generation potential, they differed in the extent of hydrocarbon expulsion which was controlled by their petrophysical properties. Pressure solution seams deserve special attention as primary migration pathways. They are known to occur in carbonate, evaporite and siliceous source rocks from the Palaeozoic to the Tertiary (Table 2). Pressure solution seams as primary migration pathways were first described by Ramsden (1952). He suspected stylolites as the principal petroleum migration avenues in carbonate source rocks from the Middle East. Stylolites can be identified on the basis of stylolitic geometry, the truncation of fossils or a general offset of marker horizons. They often develop at contacts of contrasting grain size or along pre-existing fractures (Kulander et al. 1990). There are horizontal stylolites (mainly due to overburden stress) and vertical ones (due to compressional tectonics). Stylolites with a low relief are generally called solution seams. Oilstained solution seams provide the most convincing evidence for active migration paths. But

very hydrocarbon-poor solution seams can also be diagnosed as depleted palaeopathways if hydrocarbon inclusions can be identified (see below). Before petroleum expulsion takes place, solution seams have acted as the place for concentrating organic material (together with clay minerals) which improves the source rock potential drastically (Sassen et al. 1987). Figure 2 presents such a redistribution from the pre- to the post-compaction stage. Because of these redistribution processes, it is often difficult to answer the question about the actual petroleum expulsion stage without a quantification of the compaction stage, because the initial hydrocarbon generation potential remains speculative. For mass balance of the organic material as well as for petroleum quantification during numerical simulation, the extent of the diagenetic concentration process has to be separated from primary litho- and organofacies effects. Suitable methods for determination of the volume reduction by compaction are the measurement of the maximum stylolite amplitude (Stockdale 1926; Mossop 1972), reconstruction of fossils (Ricken 1986), or comparison of the amounts of pyrite in stylolite and rock matrix (di Primio 1990). Fractures

Fractures have been reported as migration

242

U. MANN

conduits for petroleum expulsion from source rocks in case studies from the Bakken Shale (Meissner 1978), from the Lower Toarcian (Littke & Rullk6tter 1987; Littke et al. 1988; Leythaeuser et al. 1988; Mann 1990), from the La Luna formation (e.g. Talukdar et al. 1987) and from several other locations. For a quantitative approach, the number of fractures, fracture width, distribution and orientation must be recorded during the sediment-petrographical characterisation. Jochum (1988) and Jochum et al. (1991) determined the accumulated fracture width of horizontal to sub-horizontal fractures (parallel to sub-parallel to bedding) of the Lower Toarcian shale in Germany as between 1 and 3% of the total source rock thickness. Thereafter, the origin of the fractures must be assessed in order to establish a possible link to the timing of petroleum expulsion. In principle, two genetic types of fractures are differentiated: expulsion fractures, created as a result of petroleum generation, and tectonic fractures, related to the stress of a certain tectonic event. Expulsion fractures probably completely re-heal. Therefore, they will hardly be recognized at later geological times if they are not kept open due to pressure release, e.g. during later uplift of the source rock, or preserved by secondary crystallization as in the carbonate bands reported by Lindgreen (1985). Tectonic fractures as primary petroleum migration pathways may be particularly abundant in shear zones (e.g. Azevedo 1990). They may open and close several times due to activation by processes like seismic pumping (Sibson et al. 1975; Sibson 1981). Completely mineral-filled fractures should be checked for fluid inclusions (see below) and for the timing of the open fracture stage before cementation relative to petroleum migration. Cemented intersections of fractures allow the relative dating of individual fracture generations as well as the delineation of the respective stress directions (Kulander et al. 1990). In this way, the open stage of a certain fracture generation can be related to the time of a tectonic event and tied to a possible coincidence with petroleum expulsion. Nevertheless, fracture formation may be the result of a combination of effects like halokinetic doming, overpressure formation, and with possible contributions from source rock maturation, as for example in chalks from the Norwegian North Sea (Watts 1983). Lindgreen (1987b) investigated fractures from the Upper Jurassic claystone and from the Cambrian Alum shales by scanning electron microscopy and by detailed gas adsorption experiments. He found fractures parallel to, inclined to and normal to the bedding

which he related to overpressure formation and tectonics at different compaction stages. Capuano (1993) provides direct evidence of fluid flow in microfractures in geopressured shales of the Oligocene Frio Formation from Texas. Although not in a source rock, he showed a paragenetic relationship between a calcium sulphate fracture fill and later deposited organic material, and the more extensive alteration of the fracture margins which indicates that these microfractures developed in situ and supported fluid flow.

Petrophysical approach After all potential expulsion pathways have been identified by sedimentological methods, petroleum migration along these pathways has to be quantified in order to separate effective from non-effective migration pathways. This is generally based on permeability data for a source rock. The absolute gas permeability of sediments and sedimentary rocks ranges over more than 13 orders of magnitude from 10-1°mz to about 10-23m2 (Brace 1980). Fine-grained source rocks will certainly show no better permeability than 10-1Sm~ (compare Table 3). For permeability determinations of relatively tight rocks like source rocks, pulsed permeability tests are recommended due to the much shorter measurement times (Brace 1980). The values vary according to the elasticity of the rocks in relation to the overburden pressure (0-500 bars) generally between 2-4 orders of magnitude (Table 3). This sensitivity of permeability to effective stress depends upon stress history. Rocks during compaction (inelastic) are much more sensitive than rocks which are decompressed during uplift (about 10:1). This was also reported by Sandvik & Mercer (1990) when commenting on the observations of Shi & Wang (1986) and Morrow et al. (1984). Pore size and fracture width represent the main controlling agents for petroleum flow as permeability is proportional to several powers of pore size or fracture width, and it is known that very few large pore canals need be present to have a dramatic effect on permeability (e.g. Khanin 1968). Thus, it may become essential to determine the relative amounts of pore size classes present in a specific source bed, especially to differentiate between the low and higher permeability domains of a source rock, as different types of migration mechanisms may be involved at different pore size levels. Such a quantification is possible by a combination of

243

PRIMARY PETROLEUM MIGRATION Table 3. Permeability (p.D, helium) of selected source rock formations from various basins versus confining pressure for different lithologies (after Ropertz 1993) Fm

Depth (m)

Lower Saxony Basin LT <100 <100 <100 <100 Magellan Basin SH 1685 1685 1685 1685 Williston Basin BS 767 787 1055 1055 1055 1055 1055 1055 Molasse Basin FS 2645 2645 2025 2025

Lithology

k, 50bar

k, 100bar

k, 200bar

Marlstone Marlstone Clay-shale Clay-shale

30 5.2 12 24

23 4.4 5.3 1.6

6.5 3.0 2.0 -

3.5 1.5 0.23 -

1.1 0.91 -

Clay-shale Clay-shale Clay-shale Clay-shale

483 125 253 3.7

314 44 112 1.5

149 8.0 25 0.49

28 0.52 3.1 -

6.5 0.63 -

Clay-shale Clay-shale Clay-shale Clay-shale Clay-shale Siltstone Clay-shale Clay-shale

57.5 2.5 1.2 7.6 3.0 275 783 174

24.8 1.5 0.65 3.0 0.98 265 539 95

6.7 0.67 0.18 1.5 0.39 170 313 17

0.94 86 149 2.2

23 70 0.29

Marlstone Marlstone Marlstone Marlstone

9.7 15 5.1 12.6

1.3 2.7 0.46 1.1

0.22 0.49 0.19

5.0 6.9 1.7 5.3

k, 400bar

k, 600bar

-

LT, Lower Toarcian; SH, Springhill; BS, Bakken Shale; FS, Fischschiefer

microscopy, mercury porosimetry and gas adsorption measurements. According to the 1962 I U P A C classification (Singh 1976), the volume contributions of submicropores (d < 8/~), micropores (8 < d < 20/~), mesopores (20 < d < 500A) and macropores (d > 500A) are generally differentiated. Such an approach is known from the characterisation of coal pore networks (e.g. Seewald 1982; Janowsky 1984). Even the shape of the pores within a source rock can be determined. Based on the adsorption-desorption isotherm type (according to the B D D T classification after Gregg & Singh 1967), Lindgreen (1987a) was able to characterize the micropores in U p p e r Jurassic claystone source rocks as having a slit-like shape.

Low permeability domain Diffusion of hydrocarbons through the watersaturated pore space is accepted as a contributing process of primary migration (Sokolov et al. 1964; Watts 1963; Hinch 1980; Huc & H u n t 1980; Krooss & Leythaeuser 1988). Based on the calculations of Leythaeuser et al. (1982), the rate of mass transport for gas can be sufficiently high to account for commercial-sized gas fields. In

contrast to flow processes, diffusion predominates especially in the low permeability domain, i.e. in the micropore network of a source rock. Diffusion processes are studied quantitatively using Fick's law. Water release due to the collapse of smectite layers occurs prior to oil generation (Perry & Hower 1972), and Eberl (1984) suggested that such created pore space could increase the pore volume of the micropore system. Effective diffusivity in a water-saturated rock is about 10 -4 less than in the pure gas phase and related to the permeability-porosity product of the source rock (Pandey et ai. 1974). Diffusion through organic matter networks represents an alternative concept to explain primary migration (Stainforth & Reinders 1990). In fact, the activation energies measured for the diffusion of light hydrocarbons in tight rocks are much higher than for diffusion in water (e.g. Witherspoon & Saraf 1965). Krooss (1988) obtains a better fit of his experimental results for C1 to C6 hydrocarbons with a model of pore diffusion plus adsorption, rather than diffusion alone. Stainforth & Reinders (1990) presented a quantitative approach for organic matter diffusion which is based on thermally activated diffusion. Besides temperature, a major control

244

U. MANN

M

"-1

60

M

50

S

40-

S

v

(D

E "-1

o> 30-

after experiment

o

a_ 2010-

.ore I

!

100

I I flll

I

i

I

I IIIII

101

I

I

I

~ IIIII

102

I

I

[

I

I II,!

104

103

Pore radius equivalents (nm) Clay-Shale before

Marlstone

after

before

after I

13.1

2.39 i

Spec. surface area (m2g -1) •

.

19.0

14.6 ! J

4.28 i

i

18.7 7.88 J

51.___.~4

41.8 I

57.4 I

64.4

Macr°p°re v°l" (Pl g-l)

4046

5193

439 =

Largest pore radius (nm) =

13.6

24.1 ,

464 •

i

Atypical pore radius (nm) ,

•

99.3

34-9 i

I

Fig. 3. Selected permeability related petrophysical properties of a Lower Toarcian clay-shale and marlstone sample before and after a triaxial pressure experiment (after Mann et al. 1991).

of their migration model is represented by a characteristic length of the primary migration system, which takes into account the tortuosity of the petroleum pathway. However, the mi-

gration of hydrocarbons through a threedimensional organic matter network does not apply to source rocks with low organic matter content.

PRIMARY PETROLEUM MIGRATION

Higher permeability domain With increasing pore size and permeability, the dominance of diffusive transport gives way to the Darcy-dominated flow regime. Flow can be quantified by using Darcy's law with the most effective migration pathways of a given source rock being those of highest permeability during expulsion. Relatively permeable and therefore effective pathways exist in source rocks as ' . . . it is not unusual for the peak hydrocarbon generation and migration intervals to take less than 4 my' (Hunt 1991). In order to obtain total permeabilites of pores plus fractures, fractures in representative rock units have to be counted and the variation must be checked by statistical methods. Correct up-scaling methods from the thin section via the rock specimen to the rock unit would save much work but have not yet been developed. Figure 3 gives an example of the petrophysical properties before and after a triaxial stress experiment which was performed with a clay-shale and a marlstone sample from the Lower Toarcian of NW Germany in order to simulate fracture formation. In spite of a differing compaction during the experiment, for both lithologies the size of the fracture apertures created in this way (as represented by the additional modality in the pore size distribution) is one to two orders of magnitude larger than the original pore size. Such a formation of larger pore throats reduces the capillary resistance for petroleum flow considerably, and expulsion can start at lower pressure. It is unclear if the permeability of pressure solution seams can be quantified like pores or fractures by Darcy's law during the time of petroleum expulsion. Bearing insoluble organic matter, clay minerals and further residues, they represent no really clean pathway and, therefore, adsorption by clay minerals and kerogen may reduce expulsion efficiency.

Saturation, sorption and wettability Different types of water are present within a rock under subsurface conditions (Hunt 1979; Honda & Magara 1982; Mann 1989): structural water within certain minerals, hydration or bound water on mineral surfaces or as interlayers, capillary water in small pores and mobile water in large pores. Similar interaction exists between kerogen, minerals and petroleum: petroleum may be absorbed by kerogen, adsorbed on the kerogen surface or trapped within the kerogen network, and residual bitumen may remain in small pores whereas the mobile part of

245

petroleum appears to migrate predominantly in large pores (Lindgreen 1985; Mann 1989; Price & Clayton 1992; Ropertz 1993). Although at present the importance of most absorption, adsorption and wettability effects between kerogen, minerals and petroleum during the petroleum expulsion process is under debate, some information is given below. Following reservoir engineering principles, a critical oil saturation relative to water has to be reached before oil begins to flow. By analogy with low permeability reservoirs, Durand (1983) first applied the notion of relative permeability to source rocks using two examples of sedimentary basins having different sedimentation styles, the first of which had a low sedimentation rate and discrete organic-rich layers, and the second having a high sedimentation rate and numerous organic-poor layers. The relative permeability is several orders of magnitude lower in the second basin than in the first one, although Durand (1983) applied largely arbitrary values for the relative permeabilities versus oil saturation extrapolated from a production study. Ropertz (1989) determined the oil saturation necessary for the beginning of expulsion between 14 and 24% for an organicrich source rock like the Lower Toarcian. The variation is due to the range of porosity which can act as a buffer and retard petroleum expulsion for a limited time. The range of these saturation values confirms earlier theoretical estimations from Welte (1987), but these values will vary locally. Pyrolysis data from different case histories by various authors show that a clay mineral matrix of a source rock retains hydrocarbons much longer than a carbonate matrix (Espitali6 et al. 1980; Tannenbaum et al. 1986; Langford & Blanc-Valleron 1990; Heinen 1991). More TOC is necessary to achieve the same source potential in claystones than in limestones: backward extrapolations of the $2 from Rock-Eval pyrolysis versus total organic carbon content provide intersections with the TOC-axis at around 0.2% for limestones, at around 0.4% for marlstones and between 0.7% and 0.8% for claystones (Fig. 4). If these data from the high temperature conditions can be translated to the geological environment, in a clay-shale and a carbonate source bed (with the same amount and type of organic matter and at the same geological conditions), hydrocarbon migration in the clayshale would be retarded compared to the carbonate source rock. An alternative interpretation is that the generally increasing inertinite content from carbonates to marlstones to shales might be responsible for this effect. However,

246

U. MANN

50

/

0

/

!

o _

03 . m

t@

O 0)

S S°

2K

o

s SS

03

¢D -7-

j

Limestone~ s s s s/'f~

03

,, "

@

03

~ ,, 0

't

0

s

~

,all

~

L

Marlstone** ..'"_OO00" . _..~.o.,~Claystlone , • _~ _

'

t

_ _,A,.,A,

t

TOC (wt%)

(~

after Langford and Blanc-Valleron,1990) Heinen, 1991

"after

Fig. 4. Rock matrix adsorption effects as exemplified by backwards extrapolations of trends of source rock potential v. TOC content for limestone, marlstone and claystone (after Langford & Blanc-Valleron 1990 and Heinen 1991).

volumetric estimates for the main maceral groups in shale-marlstone sequences from the Lower Toarcian show that the inertinite content changes insignificantly from marlstone to shale (Littke & Rullk6tter 1987). The wettability of petroleum on mineral grains represents a further critical physicochemical parameter for hydrocarbon migration in macropores. Better wetting conditions facilitate hydrocarbon migration within the large pores because they reduce the necessary pressure for expulsion due to the increase of the contact angle. Wettability characteristics of source rocks have been studied for example by Lemke (1992). He confirmed that the contact angle between hexadecane (used as oil substitute) and a source rock section increases with progressive catagenesis, which can be explained as the effect of increasing bitumen content. He

also found that carbonates should have considerable advantages in contrast to siliciclastic source rocks during expulsion as calcite gives the widest contact angles. He also investigated the effects of different electrolytes on wettability: brines with high contents of MgSO4 tend to improve wettability. Selected results from Lemke (1992) are summarized in Table 4. It is possible that further physicochemical effects for hydrocarbon expulsion may arise from the sorption of oil by kerogen and a possible swelling of the kerogen.

Organic geochemical approach Organic geochemical methods and analyses can identify migration pathways by tracing the redistribution of soluble organic matter within representative litho- and organofacies intervals,

247

PRIMARY PETROLEUM MIGRATION Table 4. Selected wettability characteristics for source rocks: effects of minerals, electrolytes and petroleum saturation on contact angles (degree) for hexadecane after ageing for 72 hours at 80°C (after Lemke 1992) Mineral Contact angle Electrolyte Contact angle with quartz

quartz 58 H20 27

Petroleum saturation of Lower Toarcian marlstone at: Contact angle

kaolinite 83 NaCI 27

muscovite 86 Na2SO4 29

albite 108

CaCI2 39

AICI3 42

dolomite 122

calcite 138

Fe(NO3)2 55

MgSO4 75

0.53% Rm, solvent extracted

0.53% Rm, original bitumen content

0.88% Rm, original bitumen content

119

133

146

fractures or solution seams which exhibit sedimentologically and petrophysically the best potential for petroleum expulsion. Organic geochemical methods and analyses help to determine the prevailing migration processes and to define how much and what type of hydrocarbons have been already expelled or would be expelled at a higher maturation stage of the kerogen. According to the state of evolution of the kerogen, different types of petroleum compounds, i.e. hydrocarbons with different structures and chemical and physical properties are generated and are available for primary migration. Most of the various compounds can also be transported by primary migration processes, documented by the great variety of reservoir oils which contain the same type of compounds as found in source rocks.

Identification o f migration p a t h w a y s

Migration pathways generally exhibit a residual oil which may be normal oil-staining (Talukdar et al. 1987), highly viscous (Mann 1990), or partly or completely solidified (Parnell & Eakin 1987). In order to achieve a well-resolved 3D documentation of the distribution of liquid bitumen phases within a source rock specimen and especially in relation to the various cement generations, the combination of a scanning electron microscope and a liquid nitrogen cooled specimen stage seems to represent an excellent tool (Burchard et al. 1991; Neisel et al. 1991). With the help of this technique Mann et al. (1993) have been able to document details of the petroleum-rock interface, not only in reservoir rocks but also in source rocks. For example, it was possible to detect bubble-like spherical

holes within the residual bitumen of a clay-shale source rock which could be interpreted as formerly gas-filled cavities formed probably by retrograde condensation during uplift of the source bed. This method has excellent research potential, e.g. in order to document the bitumen distribution at progressing stages during hydrous pyrolysis experiments. Direct evidence for migration pathways can be provided by hydrocarbons in inclusions. The fluorescence colours of inclusions point to the respective composition of the expelled petroleum phase and the catagenetic stage when the expulsion took place. Detailed information about the classification of organic inclusions and their relations to oil and gas occurrences are given by Roedder (1984) and Shiet al. (1988). However, it is important to find enough inclusions for a reproducible result, and to avoid inclusions with post-trapping alterations, because they exhibit a re-equilibrated stage and a correspondingly altered hydrocarbon composition. A general example from Ropertz (1993) for the compositional relationship between hydrocarbon inclusions from fracture cement, residual oil from the open centre of the fracture and the depleted neighbouring source rock matrix is presented in Fig. 5. The n-alkanes obtained via thermodesorption microanalysis exhibit the following distribution: the fracture shows a predominance of the long-chained n-alkanes, i.e. the composition of the residual oil, the inclusions show a predominance of the shorter-chained hydrocarbons, i.e. a composition similar to the expelled oil, whereas the rock matrix shows a composition with a maximum content in the medium chain length, i.e. the composition of the organic source material plus residual oil.

248

U. MANN

Rock source -I- generated bitumen

jJ VV v

Fracture residual oil

Inclusions expelled oil

Fig. 5. Relative distribution of the n-alkane content of rock matrix, fracture and inclusions from adiacent Lower Toarcian rock samples which exhibit the distribution of source plus generated oil, residual oil and expelled oil respectively (after Ropertz 1993).

PRIMARY PETROLEUM MIGRATION

249

A

0 0 I--

Q -I

C 0 0 o "I"

® 500.0

® 250.0

-I

6

®

o.o

n-C15

n-'C20

n-C25

n-C30

n-C35

Fig. 6. Cts+ n-alkane content of selected samples from the same rock unit representing different types and phases of petroleum expulsion: 1, recrystallized carbonate bed; 2, siltstone bed; 3, rock matrix from unmodified carbonate bed; 4, stylolite from unmodified carbonate bed. Regular depth trends of the amount and composition of organic compounds can likewise identify the drainage direction from a source rock. Increasing hydrocarbon depletion towards the edges of source rocks have been documented for example by Leythaeuser et al. (1984) and by Mackenzie et al. (1987). In order to prove that a horizontal fracture had acted as an expulsion pathway, Leythaeuser et al. (1988) investigated the hydrocarbon content of the fracture and the adjacent shale above and below. They found the shale samples very similar, but a strong decrease in the proportion of saturated and aromatic hydrocarbons as well as an increase of the percentage of NSO-compounds in the fracturefill sample together with a low abundance of n-alkanes but similar isoprenoid/n-alkane ratios compared to the adjacent shale samples. The concentration and the composition are interpreted to reflect the net result of two processes, the drainage of petroleum from the adjacent shale matrix into the fracture, and subsequent lateral petroleum flow along the fracture.

Determination o f the type o f processes involved On the molecular level, a separate phase flow of a bulk oil expulsion should exhibit no fractionation, whereas hydrocarbon transport by

solution in water or gas as well as diffusion show characteristic fractionation effects. The residual soluble organic matter in the source rock provides hints about which type of process(es) might have taken place. In order to arrive at such conclusions, soluble organic matter has to be extracted from the rock material and separated by liquid and gas chromatography into compound groups, homologous series, isomers and individual molecules. Then, concentrations of selected hydrocarbon compound groups or individual molecules with distinctive chemical and physical properties have to be compared. For example, saturated hydrocarbons have to be ratioed to aromatics or both to NSOcompounds, more water-soluble aromatics have to be ratioed to less water-soluble ones, shortchain to long-chain hydrocarbons or n-alkanes to isoprenoids. Figure 6 shows selected examples for hydrocarbon-depleted rock samples from the same source bed. The envelope of the C15+ n-alkane composition of a depleted solution seam (4) exhibits a very equal distribution with no signs of fractionation relative to the hydrocarbon chain length. The respective adjacent rock matrix (3) still contains a much higher concentration of hydrocarbons than the solution seam, especially between C15 and C2o: bulk petroleum flow may have flushed the solution seam, but no or little expulsion seems to have occurred at the adjacent

250

U. MANN

rock matrix. A rock sample which represents a more silty layer (2) exhibits a very sudden break at C21, before Czl relatively low concentrations with a regular decrease with shorter chain length, and beyond C21 still a quite high concentration of n-alkanes: a mechanism with a strong fractionation tendency must be involved, possibly migration by gaseous solution participates in primary migration. Finally, a recrystallized limestone layer (1) exhibits a high concentration of n-alkanes with a very uniform distribution and maximum concentrations at C19: only the very short-chain hydrocarbons have left this source bed. Compared to the other samples, most n-alkanes have been preserved which seems reasonable due to the strong cementation by carbonate which occurred prior to primary migration. However, care has to be taken before drawing conclusions about the final expulsion process: other geochemical indications (e.g. from the composition of aromatic hydrocarbons) must be found to be in accordance with the above observations. Furthermore, a possible mechanism should be checked versus the gas/oil ratio in the subsurface which is known to increase with temperature because the less stable higher molecular weight components of oil are cracked into the more stable components of gas. In fact, only if the temperature, pressure and diagenetic history of the source rock is understood reasonably well, limits can be estimated for the involved expulsion mechanisms.

Defining relative and absolute expulsion efficiency In order to define expulsion efficiency, relative and absolute comparisons of individual source rock units as well as identified migration avenues have to be performed. In any case, for comparisons of different source rock samples, effects of source rock richness and organic matter type must be excluded. The following observations have been made in the past. Generally, better drainage at source rock edges close to the carrier bed or at source rock/reservoir rock contacts can be anticipated (Vandenbrouke 1972; Leythaeuser & Schwarzkopf 1986) as well as elevated expulsion efficiencies in interbedded source/carrier bed systems (Leythaeuser et al. 1984). The question of whether or not fractures or solution seams are actually working as expulsion pathways has always to be answered relative to the adjacent rock matrix (see above) because they might be in a preceding, active or post-expulsion stage, and accordingly such pathways may be enriched or

depleted in hydrocarbons (relative to the rock matrix). At the same catagenetic stage, expulsion efficiencies for adjacent lithologic units can differ drastically due to different petrophysical properties. A relationship between petrophysical properties and hydrocarbon content due to different expulsion efficiencies is presented in Fig. 7 (after Mann et al. 1991). Four adjacent source rock units (clay-shale, marlstone, cemented marlstone and marlstone with secondary porosity) from the same Lower Toarcian section possess a more or less identical hydrocarbon generation potential, but different pore networks. In particular the amount of pores with throats greater than 10nm is very different, which causes, together with different porosity and specific surface area, a permeability variation between 0.13 and 321xD. The respective bitumen compositions reflect the better drainage conditions of the more permeable units by a lower content of heptane and of C15+-soluble organic matter. The pristane/heptadecane ratio indicates the depletion stage according to Leythaeuser & Schwarzkopf (1986). A higher ratio of saturated plus aromatic hydrocarbons to NSO-compounds reflects more intensive lateral hydrocarbon flow along the marlstone with secondary porosity, and the phenanthrene/ methyl-phenanthrene ratio indicates that some of the more water-soluble aromatic hydrocarbons have been removed by a parallel water flow. A comparison of the relative expulsion efficiencies versus permeability between the total C15+-soluble organic matter content and the concentration of the C15_n-alkane and the C27-n-alkane provides clear evidence that each pore network produces specific fractionation characteristics. Besides the determination of petroleum masses generated, Cooles et al. (1986) proposed an algebraic scheme for the calculation of petroleum expulsion efficiencies. It is based on a simple model which envisages kerogen as consisting of a reactive and an inert fraction. The reactive fraction is further subdivided into a labile component which generates both oil and gas, and a refractory component which generates chiefly gas at higher maturities. The inert proportion generates neither oil nor gas. Based on TOC measurements, Rock-Eval pyrolysis yields and solvent extraction yields, it is possible to define a petroleum generation index (= (generated + initial petroleum) : total petroleum potential) as well as a petroleum expulsion efficiency (= expelled petroleum : (generated + initial petroleum)). As a prerequisite of all other comparisons in respect to

PRIMARY PETROLEUM MIGRATION Petrophysical properties

Pore size distribution

40-

P

251

S

~

M

C

P

7.25 SSA (m2g -1)

..-&

30-

-~ 20-

PV(pig-')

6.65 ,220 ~ 2.25!

HI (nm)

_ 2.3

I

c ~_

10-

S

PHI 0

........

100

]

. . . .

, ,,,

10~ 102 Porethroatequivalents(nm)

k~(~O)

0.13

Bitumen composition S

M

Bitumen expulsion C

P

142 n-C7 (mg/g TOC)

n.a,

2

1.9

C15-n100-

C~s+-SOM 80-

Cls+-SOM (mg/g TOC)

~

l-I

101

!

82

96

!

.

I

g "5 Pristane/n-C~r

0.96 r---!

1.48 1.55 r ~ i 1

i

cr

604020-.! -

|

M C

P

00.1

Phen. / M.-Phen.

C27-n-alkane

¢

]

[

1

10

100

N-(pD)

Fig. 7. Interrelationship between lithology, petrophysical properties, content of soluble organic matter and relative expulsion efficiencyof total Cls+ -soluble organic matter, C,5-n- alkane and CzT-n-alkaneof the Lower Toarcian Posidonia shale source rock. (S, clay-shale; M, marlstone; C, cemented marlstone; P, maristone with secondary porosity.) petroleum expulsion efficiency, this approach assumes identical kerogen type, but is independent of the original organic matter content. By applying such mass balances to various source rocks, Cooles et al. (1986) calculated expulsion efficiencies up to 90% for source rocks with an excellent initial potential like the Kimmeridge Clay from the North Sea or the Brown Limestone from the Gulf of Suez. But even source rocks with a potential of lower than 10kg hydrocarbons per ton of rock could still expel more than 50%. Numerical

approach

Numerical modelling of petroleum expulsion can address all interrelated processes leading to

the observations discussed above. Values for parameters controlled by the depositional environment and by the diagenetic history can be acquired by petrographical, petrophysical and organic-geochemical investigations. A complete history of petroleum expulsion, governed by the pressure and temperature changes in the source rock during the evolution of a sedimentary basin can only be attained by integrating all effects into a numerical simulation of the geological history. Several numerical models integrate the simulation of heat transfer, compaction, hydrocarbon generation and migration, generally known as basin modelling (Yiikler et al. 1978, Welte & Ytikler 1981; Ungerer et al. 1983, 1984; Yalqin & Welte 1988; Burnham & Braun 1990; Waples et al. 1992a).

252

U. MANN

Nevertheless, it is in fact too simple to use the modelling results of hydrocarbon generation (volume, perhaps also the composition) as the starting point for numerical modelling of secondary migration, and to neglect the above mentioned effects of primary migration. Of course, numerical modelling of petroleum migration in source rocks is coupled to petroleum generation, and several critical parameters for generation are therefore also critical for expulsion.

Critical prerequisites The critical prerequisites for a successful numerical modelling of petroleum expulsion are sufficient data for source rock potential, temperature and compaction history as well as for the formation of overpressure. If the source rock potential is incorrect, the continuous mass and volume balancing of the various source rock components during the changing PVTconditions with time would be incorrect as well as the pressure build-up due to volume expansion during solid/liquid conversion of the organic matter. A wrong temperature history would lead to a wrong timing of the expulsion event. This may result from an insufficient effort in respect to calibration to (time- and) temperaturesensitive parameters such as subsurface temperature measurements, vitrinite reflectance, geochemical maturity data, data from apatite fission tracks, microthermometry of fluid inclusions, or others. Modelling carried out in the absence of any measured thermal indicator to constrain the palaeotemperature history means accepting significant uncertainty in prediction of the generation and expulsion process (Waples et al. 1992a). On the other hand, it is still unclear how far thermal short-time events (heat 'spikes') influence the kinetics of petroleum formation and the subsequent timing of primary migration (e.g. Leischner et al. 1993). Significant errors in the heat flow history often result from an insufficient geological knowledge about unconformities (what was the original overburden?) and consequently uncertainty about palaeothermal gradients. The same problems exist with the diagenesis of carbonates (early or late diagenetic cementation?) and a subsequent wrong estimation of their thermal properties. In order to avoid misinterpretations, similar to the calibration of the temperature history, calibrations for compaction as well as for diagenesis would help to create a better conceptional model for numerical simulation. Such calibrations can only be attained by a better integration of sedimentpetrographical and petrophysical investigations.

Flow modelling Fluid flow in petroleum source rocks can be described by applying the relative permeability concept of a three-phase fluid flow between oil, gas and water. However, even a two-phase fluid flow adaptation of Darcy.'s law (Nakayama 1987; Ungerer et al. 1990; Ozkaya 1991) can only represent an approximation in such tight rocks as petroleum source rocks (Legait & Kalaydjian 1987). Therefore, Stainforth & Reinders (1990) proposed an alternative quantitative approach based on a temperature-activated diffusion model. Diippenbecker (1992) included for the first time the relative contribution of different pore sizes into a mathematical flow concept. His idea was to connect pore classes in ascending order of their size, and a respective overflow from pore class to pore class after sufficient pressure has been reached. However, this implies that nearly the whole pore network would have to become oil-saturated before expulsion can start. Single phase flow along a network of larger, interconnected pores, known from reservoir engineering as a fingering petroleum network, allows petroleum flow at much lower saturation. In order to quantify petroleum expulsion, pore fluid pressures have to be predicted which result from the distribution of different rock types within a sedimentary basin, their burial rates and several possibilities for overpressure formation (see also above petrophysical approach). Examples of the various factors influencing overpressure have been discussed by Bredehoeft & Hanshaw (1968), Bradley (1975), Chapman (1980), Ozkaya (1986) and by Mann & Mackenzie (1990). Overpressures occur if pore fluids support some of the weight of the overlying rock column in addition to the fluid weight. These rocks are in a compaction disequilibrium. The situation arises when the permeability of the overlying sediments is too low to let fluids escape in equilibrium with added overburden. The accuracy to which pore pressures and the resulting petroleum expulsion may be predicted seems to be much more determined by the availability and the accuracy of the geological input rather than by the approximations of a numerical model.

Fracture modelling There are two basic parameters which control fracturing of a rock: first, physical rock properties such as strength, Poisson's Ratio and thermal expansion, and second, differential rock matrix stress. These parameters are affected by a

PRIMARY PETROLEUM MIGRATION variety of controlling geological factors. Besides tectonic fractures, the numerical treatment of expulsion fractures (microfractures, hydraulic fractures in association with hydrocarbon generation) is crucial for source rocks. Concepts and theoretical approaches for the numerical treatment of such fractures have been proposed by Du Rouchet (1981), Ozkaya (1984, 1988), Ungerer et al. (1990), Lehner (1991) and Dtippenbecker & Welte (1992). A principal part of most models concerns the contemporaneous development of source rock and petroleum physical properties and the application of fracture mechanisms. The present situation is that most numerical approaches for fracture simulation require better quantitative information on the PVT properties of the kerogen as well as of the changing composition of the generated bitumen. More research is needed in that area, and more field data of in situ stress measurements and fluid pressures should be made publicly available. Expulsion pathway modelling and sensitivity analysis As shown above, several migration pathways are possible for petroleum to leave the source rock: (i) the pore network of the rock which consists of a primary porosity component according to the depositional environment of the source rock, and a secondary porosity component, created by compaction and diagenesis, (ii) pressure solution seams, created according to the composition of the source rock and the specific stress conditions, and (iii) fractures, formed due to tectonic stress or in association with hydrocarbon generation. In order to consider the parameters from which those migration pathways arise, it is necessary to integrate into a numerical simulation compaction and temperature history and especially the effects of carbonate diagenesis. As carbonate re-distributes relatively easily and quickly, it is most suited to open or to block possible petroleum expulsion pathways. Further potential candidates for integration into a numerical approach for petroleum migration are sediment-stratigraphical models for basin filling and tectonic models. Computer simulations are ideal for investigating hierarchical levels of stratigraphic cyclicity, especially those caused by eustatic variations or by fluctuations in clastic input (Aigner et al. 1990; Lawrence et al. 1990) or in intraplate stress (Cloetingh 1986; Kooi & Cloetingh 1989).

253

A further integrated numerical model necessitates (i) more decision levels in order to establish the optimal conceptional model, (ii) additional parameters for calibrations beside the temperature history, and (iii) more emphasis on sensitivity analyses, in order to quantify the uncertainty of the geological information. The process of establishing a geologically adequate conceptional model and its verification is of more importance than the whole numerical simulation. Calibrations for numerical expulsion modelling should involve proving primary migration along a specific pathway. Sensitivity analyses in basin modelling have been applied by Yfikler et al. (1978), Ritter (1988), Hermannrud et al. (1990) and Waples et al. (1992b) but they still represent a very neglected area. These studies are confined mainly to the modelling of heat transport and hydrocarbon generation, and do not discuss explicitly the factors affecting hydrocarbon expulsion from source rocks. For calibrations, 'real' measured data are required, but there are generally less geological data available than would be necessary for a satisfactory calibration. Therefore, it is the duty of every geologist to run sensitivity analyses for the most critical parameters of his program. A probability distribution should be assigned to those parameters, and the sensitivity should be investigated by varying each parameter within the assigned ranges as shown by Hermannrud et al. (1990).

Summary Both primary sedimentological conditions in the depositional environment of the source rock, and secondary diagenetic processes during evolution of a sedimentary basin, have to be considered as critical controls of primary petroleum migration. In order to accommodate most critical parameters, an integrated approach based on sedimentology, petrophysics, organic geochemistry, and numerical modelling should furnish geologically acceptable results. First, the lithofacies of the source rock has to be characterized in order to deduct petrographic evidence for potential migration pathways. By establishing the diagenetic, tectonic and compaction history of the individual rock units of the petroleum source bed, all sedimentologically reasonable expulsion pathways can be identified. Chemical diagenesis is crucial because it controls permeability. With the help of reservoir engineering principles, effective pathways are separated from non-effective migration, and the true expulsion potential in respect to the permeability of the

254

U. MANN

source rock unit can be quantified. The level of fluid pressure and especially the formation of overpressure represent the crucial factors for the driving mechanism of petroleum migration in source rocks. Organic geochemical techniques can be used to trace the redistribution of organic matter within representative litho- and organofacies intervals and especially along the effective migration pathways, in order to define relative and absolute expulsion efficiencies. Understanding the sensitivity of individual parameters within the framework of a quantitative model and verifying a primary migration pathway concept (i.e. the relevant migration pathway, embedded into a specific geological setting at the most realistic palaeo-temperature and -pressure conditions) by numerical simulation is the final challenge for an integrated approach to petroleum migration in source rocks. However, as more and more parameters become integrated into a numerical simulation, more attention must be given in order to include additional calibrations of the diagenetic and the compaction history besides the standard calibration to the temperature history. Furthermore, cross-checks with organic-geochemical determinations of the expulsion stage should be performed and incorporated as a standard concept for an integrated approach to petroleum expulsion. Geoscientists must accept the fact that their models cannot be unique solutions, and must be prepared to provide alternatives when a match cannot be made. Where calibration to real data is insufficient or impossible, a combination of deterministic and stochastic modelling should mirror the actual knowledge of the geological situation and provide the necessary background for decisions.

Perspective At present, petroleum expulsion volumes are estimated by applying standardized bulk model concepts. In the future, quantitative models with integrated concepts for specific expulsion pathways are desirable. These models prove by a combination of various methods of geosciences which type of migration pathway of all potential pathways has been effective. By this method, expulsion models consider only those petroleum expulsion pathways which indeed have been significant. Such models should become possible by further improved data management systems. Much future potential for a quantitative description for primary migration may be the application of fractals (Mandelbrot 1975; de Gennes 1985). Fractal approaches to primary

migration have not been investigated so far for their potential to reflect primary migration processes, although they were successfully applied to secondary migration and reservoir characterization (e.g. Feder 1988). Of course, the exploration geologist is most interested in mass balances and expulsion stages of his case history. Therefore, organic geochemical checks for expulsion stages of source rocks compare predominantly the hydrocarbon content from immature rock to mature rock samples. However, this neglects potential molecular sieving effects for petroleum during the transition from kerogen via the micropore to the macropore (or fracture) system. More basic research, with focus on the microscale and a clear differentiation between the micro- and the macropore system of a source rock as well as of its hydrocarbon content, could help to better understand the reason why differences in the primary migration mechanism(s) occur from case history to case history. I thank H. Lindgreen and an anonymous reviewer for constructive comments and B. Horsfield and H.S. Poelchau for linguistic aid. The presented progress on understanding of primary migration of petroleum in source rocks would not have been achieved without the analytical support from many colleagues at KFA during my own research, and without support by providing appropriate sample suites from BEB Erdgas Erd61, Hannover; Canadian Hunter, Calgary; Deminex, Essen; Maraven, Carracas; Preussag AG, Hannover; RAG, Wien; Wintershall AG, Kassel. Continuous interest and support by D.H. WeRe is gratefully acknowledged.

References AIGNER, T., BRANDENBURG, A., VILET, g . VAN, DOYLE, M., LAWRENCE,D. & WESTRICH,J. 1990. Stratigraphic modelling of epicontinental basins: two applications. Sedimentary Geology, 69, 167190. AZEVEDO, R.P. DE 1990. Shear zones and primary hydrocarbon migration. Boletim de Geoci6ncias da Petrobr~s, 4, 217-220. BAKER, E.G. 1959. Origin and migration of oil. Science, 129, 871-874. BARKER, C. 1972. Aquathermal pressuring - role of temperature in development of abnormalpressure zones. American Association of Petroleum Geologists Bulletin, 56, 2068-2071. BARTH, T., BORGUND, A.E., HOPLAND, A.L. & GRAUE, A. 1988. Volatile organic acids produced during kerogen maturation - Amounts, composition and role in migration of oil. Organic Chemistry, 13,461-466. BELETSKAYA,S.N. 1972. A study of distribution of the disseminated bitumens in the pore system of rocks related to the assessment of the state of migration processes. A comparative study of the gaseous

PRIMARY PETROLEUM MIGRATION extracts from non-crushed rocks. Doklady Geokh. Nefii i Gaza, 16, 39-45. BJORLYKKE, K. 1984. Formation of secondary porosity: How important is it? In: MCDONALD,D.A. & SURnAM,R.C. (eds) Clastic Diagenesis. American Association of Petroleum Geologists Memoirs, 37,277-286. BOLES, J.R. & FRANKS,S.G. 1979. Clay diagenesis in Wilcox sandstones of Southwest Texas: implications of smectite diagenesis on sandstone cementation. Journal of Sedimentary Petrology, 49, 55-70. BONHAM, L. 1980. Migration of hydrocarbons in compacting basins. American Association of Petroleum Geologists Bulletin, 64,549-567. BRACE, W.F. 1980. Permeability of crystalline and argillaceous rocks. International Journal of Mechanics and Mineral Sciences & Geomechanical Abstracts, 17,241-251. BRADLEY, J.S. 1975. Abnormal formation pressure. American Association of Petroleum Geologists Bulletin, 59,957-973. BRAY, E.E. & FOSTER, W.R. 1980. A process for primary migration of petroleum. American Association of Petroleum Geologists Bulletin, 64, 107-114. BRUKNER,A. & VETO, I. 1983. Extracts from open and closed pores of an Upper Triassic sequence from W. Hungary: a contribution to studies of primary migration. In: BJOROY,M. et al. (eds) Advances in Organic Geochemistry. Wiley, New York, 175182. BREDEHOEFT, J.D. & HANSHAW, B.B. 1968. On the maintenance of anomalous fluid pressures, I. Thick sedimentary sequences. Geological Society of America Bulletin, 79, 1097-1106. BRUCE, C.H. 1984. Smectite dehydration - its relation to structural development and hydrocarbon accumulation in northern Gulf of Mexico. American Association of Petroleum Geologists Bulletin, 68, 673-4583. BUDAI, J.M., CERCONE, K. & WILTSCHKO,D. 1984. The role of tectonic stylolites in the rock deformation. In: Stylolites and Associated Relevance to Hydrocarbon Reservoirs. Abu Dhabi National Reservoir Research Foundation, Special Publications, 221-236. BURCHARD,W.-G., CLOOTH,G. & NEISEL, J.D. 1991. Das Kryo-REM Verfahren: Methode zur Untersuchung und Darstellung von Feuchtigkeit in mineralischen Baustoffen. Jahresberichte aus dem Forschungsprogramm Steinzerfall-Steinkonservierung, 1, Ernst & Sohn, Berlin. BURNHAM,I.C. & BRAUN,R.L. 1990. Development of a detailed model of petroleum formation, destruction, and expulsion from lacustrine and marine source rocks. Organic Geochemistry, 16, 27-39. BURST, J.R. 1969. Diagenesis of Gulf coast clayey sediments and its possible relation to petroleum migration. American Association of Petroleum Geologists Bulletin, 53, 73-93. BYRAMJEE, R. 1967. Probl~me g6ochimique de la gen~se et de la migration du p6trole. Revue de L'Institute Fran~ais Pdtrole, 22,791-796.

255

CAPUANO, R.M. 1993. Evidence of fluid flow in microfractures in geopressured shales. American Association of Petroleum Geologists Bulletin, 77, 1303-1314. CHAPMAN, R.E. 1980. Mechanical versus thermal cause of abnormally high pore pressures in shales. American Association of Petroleum Geologists Bulletin, 64, 2179-2183. CLOETINGH,S. 1986. Intraplate stresses: a new tectonic mechanism for fluctuations of relative sea level. Geology, 14,617-620. COLEMAN, M.L., CURTIS, C.D. & IRWIN, H. 1979. Rate of burial important for source and reservoir potential. World Oil, 188, 83-92. COMER, J.B. & HINCH, H.H. 1987. Recognising and quantifying expulsion of oil from the Woodford Formation and age-equivalent rocks in Oklahoma and Arkansas. American Association of Petroleum Geologists Bulletin, 71,844-858. COOLES, G.P., MACKENZIE, A.S. & QUIGLEY, T.M. 1986. Calculation of petroleum masses generated and expelled from source rocks. Organic Geochemistry, 10, 235-245. CORDELL, R.J. 1972. Colloidal soap as proposed primary migration medium for hydrocarbons. American Association of Petroleum Geologists Bulletin, 57, 1618-1643. CURTIS, C.D. 1978. Possible links between sandstone diagenesis and depth-related geochemical reactions occurring in enclosing mudstones. Journal of the Geological Society, London, 135,107-118. -1983. Geochemistry of porosity enhancement and reduction in clastic sediments. In: BROOKS,J. (ed.) Petroleum geochemistry and exploration of Europe. Geological Society, London, Special Publications, 12, 113-126. DAMES, S. 1982. Aquathermal pressuring and geopressure evaluation. American Association of Petroleum Geologists Bulletin, 66,931-939. DE GENNES, P.-G. 1985. Partial filling of a fractal wetting structure by a fluid. In: ADLER, D. et al. (eds) Physics of disordered materials. Plenum, New York, 227-241. DEMAISON,G. 1984. The generative basin concept. In: DEMAISON, G. & MURRIS, R.J. (eds) Petroleum geochemistry and basin evaluation. American Association of Petroleum Geologists Memoirs, 35, 1-14. DENHAM, M.E. & TIEH, T.T. 1990. Stylolites in the upper Smackover Formation of north Louisiana (abstract). American Association of Petroleum Geologists Bulletin, 74,640. DICKEY, P.A. 1975. Possible primary migration of oil from source rocks in oil phase. American Association of Petroleum Geologists Bulletin, 59, 337345. DI PRIMIO, R. 1990. Organisch-geochemische und mikroskopische Untersuchungen zur Kohlenwasserstoffgenese und primiiren Migration in karbonatischen Muttergesteinen der Trias Italiens. Diplomarbeit. RWTH Aachen. DUPPENBECKER,S.J. 1992. Genese und Expulsion von Kohlenwasserstoffen in zwei Regionen des Niedersdchsischen Beckens unter besonderer

256

U. MANN

Beriicksichtigung der Aufheizraten. Dissertation, RWTH Aachen. Dt3PPENBECKER,S.J. & WELTE, D.H. 1992. Petroleum expulsion from source rocks - insights from geology, geochemistry and computerized numerical modelling. Proceedings of the 13th World Petroleum Congress. Wiley, 165-177. DURAND, B. 1983. Present trends in organic geochemistry in research on migration of hydrocarbons. In: BJOROY, M. et al. (eds) Advances in Organic Geochemistry. Wiley, New York, 117128. 1988. Understanding of HC migration in sedimentary basins (present state of knowledge). Organic Geochemistry, 13,445-459. Du ROUCHET, J. 1981. Stress fields, a key to oil migration. American Association of Petroleum Geologists Bulletin, 65, 74-85. EBERL, D.D. 1984. Clay-mineral formation and transformation in rocks and soils. Philosophical Transactions of the Royal Society, London, A311, 241-257. ENGLAND, W.A., MACKENZIE,A.S., MANN, D.M. & QUIGLEY,T.M. 1987. The movement and entrapment of petroleum fluids in the subsurface. Journal of the Geological Society, London, 144, 327-347. ESPITALII~,J., MADEC,M. & TISSOT, B. 1980. Role of mineral matrix in kerogen pyrolysis: influence on petroleum generation and migration. American Association of Petroleum Geologists Bulletin, 64, 59-66. FEDER,J. 1988. Fractals. Plenum Press, New York and London. FoscoLos, A.E. & POWELL,T.G. t980. Mineralogical and geochemical transformation of clays during catagenesis and their relation to oil generation. Canadian Society of Petroleum Geologists Memoirs, 6,153-172. GILES, M.R. & DE BOER, R.B. 1989. Secondary porosity: creation of enhanced porosities in the subsurface from the dissolution of carbonate cements as a result of cooling formation waters. Marine and Petroleum Geology, 6,261-269. GOFE, J.C. 1983. Hydrocarbon generation and migration from Jurassic source rocks in the E. Shetland Basin and Viking Graben of the Northern North Sea. Journal of the Geological Society, London, 140,445-474. GRABOWSKI, G.J. 1984. Generation and migration of hydrocarbons in Upper Cretaceous Austin Chalk, south-central Texas. In: PALACAS, J.G. (ed.) Petroleum geochemistry and source rock potential of carbonate rocks. American Association of Petroleum Geologists Studies in Geology, 18, 97-115. GRAHAM, S.A. & WILLIAMS, L.A. 1985. Tectonic, depositional, and diagenetic history of Monterey Formation (Miocene), Central San Joaquin Basin, California. American Association of Petroleum Geologists Bulletin, 69, 385-411. GREGG, S.J. & SINGH, S.W. 1967. Adsorption, surface area, and porosity. Academic Press, London. HEDBERG,H.D. 1974. Relation of methane generation -

-

to undercompacted shales, shale diapirs and mud volcanoes. American Association of Petroleum Geologists Bulletin, 58,661--673. HEINEN, V. 1991. Der Einflufl der Lithofazies im Lias epsilon (Unteres Toarcium) auf die Erd6labgabe. Diplomarbeit RWTH Aachen. HERMANNRUD,C., EGGEN,S., JACOBSEN,T., CARLSEN, E.M. & PALLESEN, S. 1990. On the accuracy of modelling hydrocarbon generation and migration: The Egersund Basin oil find, Norway. Organic Geochemistry, 16,389-399. HINCH, H.H. 1980. The nature of shales and the dynamics of hydrocarbon expulsion in the Gulf Coast Tertiary section. In: ROBERTS, W.H. & CORDELL, R. (eds) Problems of petroleum migration. American Association of Petroleum Geologists Studies in Geology, 10, 1-18. HOFMANN, P.M. 1992. Sedimentary facies, organic facies, and hydrocarbon generation in evaporite sediments of the Mulhouse Basin, France. PhD thesis, The City University, New York. HONDA, H. & MAGARA, K. 1982. Estimation of irreducible water saturation and effective pore size of mudstones. Journal of Petroleum Geology, 4,407--418. HORNAFIUS,J.S. 1991. Facies analysis of the Monterey Formation in the northern Santa Barbara Channel. American Association of Petroleum Geologists Bulletin, 75,894-909. HOWER, J., ESLINGER, E.V., HOWER, M.E. & PERRY, E.A. 1976. Mechanism of burial metamorphism of argillaceous sediment: 1. Mineralogical and chemical evidence. Geological Society of America Bulletin, 87,725-737. Huc, A. & HUNT, J.M. 1980. Generation and migration of hydrocarbons in offshore south Texas gulf coast sediments. Geochimica et Cosmochimica Acta, 44, 1981-1989. HUNT, J.M. 1973. An examination of petroleum migration processes. Doklady Geokh. Nefti i Gaza, 1,219-229. 1977. Ratio of petroleum to water during primary migration in Western Canada Basin. American Association of Petroleum Geologists Bulletin, 64, 434-435. 1979. Petroleum geochemistry and geology. Freeman & Co., San Francisco. 1990. Generation and migration of petroleum from abnormally pressured fluid compartments. American Association of Petroleum Geologists Bulletin, 74, 1-12. 1991. Generation and migration of petroleum from abnormally pressured fluid compartments: reply. American Association of Petroleum Geologists Bulletin, 75,336-338. IRWIN, H. 1980. Early diagenetic carbonate precipitation and pore fluid migration in the Kimmeridge Clay of Dorset, England. Sedimentology, 27, 577-592. & HURST, A. 1983. Applications of geochemistry to sandstone reservoir studies. In: BROOKS, J. (ed.) Petroleum geochemistry and exploration of Europe. Geological Society, London, Special Publications, 12, 127-146. -

-

-

-

PRIMARY PETROLEUM MIGRATION IRWIN, H., CURTIS, C.D. & COLEMAN, M.L. 1977. Isotopic evidence of source of diagnostic carbonates formed during burial of organic-rich sediments. Nature, 269,209-213. ISAACS, C.M., PiscioTro, K.A. & GARRISON, R.E. 1983. Facies and diagenesis of the Miocene Monterey Formation, California: A summary. In: IIJIMAA, A., HEIN, J.R. & SILVER, R. (eds) Siliceous deposits of the Pacific region. Elsevier, New York, 247-282. JANOWSKY, U. 1984. Experimentelle Untersuchungen zum StrOmungs- und Sorptionsverhalten yon Wasser und Gasen in Steinkohle und Ableitung eines Porenmodells. Dissertation Universit~t Essen, Essen. JOCHUM, J. 1988. Untersuchungen zur Bildung und Karbonatmineralisation yon Kli~ften im Posidonienschiefer (Lias epsilon) der Hilsmulde. Diplomarbeit, RWTH Aachen. , LEYTHAEUSER, D. & LVITKE, R. 1991. Oilbearing fluid inclusions in calcite-filled horizontal fractures from mature Posidonia Shale (Hils Syncline, NW Germany). In: MANNING, D.A. (ed.) Organic Geochemistry, advances and applications in the natural environment. Manchester University Press, New York, 160-162. JONES, R.W. 1984. Comparison of carbonate and shale source rocks. In: PALACAS,J.G. (ed.) Petroleum geochemistry and source rock potential of carbonate rocks. American Association of Petroleum Geologists Studies in Geology, 18, 163-180. KHANIN, A.A. 1968. Evaluation of sealing abilities of clayey cap rocks in gas deposits. Geologiya Nefti i Gaza, 8, 17-21 [in Russian]. Kool, H. ~£ CLOETINGH, S. 1989. Some consequences of compressional tectonics for extensional models of basin subsidence. In: POELCHAU, H.S. & MANN, U. (eds) Evolution of Sedimentary Basins. Geologische Rundschau Special Volume, 78/1, 171-182. KROOSS, B. 1988. Experimental investigation of the molecular migration of C1-C6 hydrocarbons: kinetics of hydrocarbon release from source rocks. Organic Geochemistry, 13,513-524. & LEYTHAEUSER, D. 1988. Experimental measurements of the diffusion parameters of light hydrocarbons in water-saturated sedimentary rocks - II. Results and geochemical significance. Organic Geochemistry, 12, 91-108. KULANDER, B.R., DEAN, S.L. & WARD, B.J. 1990. Fractured core analysis: interpretation, logging, and use of natural and induced fractures in core. American Association of Petroleum Geologists Methods in Exploration, 8. KVENVOLDEN,K.A. & CEAYPOOL,G.E. 1980. Origin of gasoline-range hydrocarbons and their migration by solution in carbon dioxide in Norton Basin, Alaska. American Association of Petroleum Geologists Bulletin, 64, 1078-1086. LANGFORD, F.F. & BLANC-VALLERON, M.-M. 1990. Interpreting Rock-Eval pyrolysis data using graphs of pyrolizable hydrocarbons vs. total organic carbon. American Association of Petroleum Geologists Bulletin, 74,779-804.

257

LAWRENCE, D.T., DOYLE, M. & AIGNER, T. 1990. Stratigraphic simulation of sedimentary basins: concepts and calibrations. American Association of Petroleum Geologists Bulletin, 74,273-295. LEE CHUNG-I. 1991. Generation and migration in the Miocene Monterey Formation, southern San Joaquin Basin, California. PhD thesis. Texas A&M University, Houston, Texas. LEGAIT, B. • KALAYDJIAN,F. 1987. Origin and limits of laws for two-phase flow in porous media: a review of recent progresses performed at Institut Francais du P6trole. In: DOLIGEZ, B. (ed.) Migration of hydrocarbons in sedimentary basins. Editions Technip, Paris, 205-222. LEHNER, F.K. 1991. Pore-pressure induced fracturing of petroleum source rocks. Implications for primary migration. In: IMARISIO,G., FRIAS, M. & BEMTGEN, J.M. (eds) The European oil and gas conference, a multidisciplinary approach in exploration and production R&D proceedings. Graham & Trotman, London, 142-154. LEISCHNER, K., WELTE, D. & LITrKE, K. 1993. Fluid inclusions and organic maturity parameters as calibration tools in basin modelling. In: DORE et al. (eds) Basin Modelling: Advances and Applications. Norwegian Petroleum Society, Special Publications, 3,161-172. LEMKE, H. 1992. Untersuchungen zur Benetzbarkeit yon ErdGlreservoir- und Muttergesteinen dutch Kontaktwinkelmessungen an Mineralien und Gesteinen. Dissertation, RWTH Aachen. LEWAN, M.D. 1985. Evaluation of petroleum generation by hydrous pyrolysis experimentation. Philosophical Transactions Royal Society of London, A315, 123-134. 1987. Petrographic study of primary petroleum migration in the Woodford Shale and related rock units. In: DOLIGEZ, B. (ed.) Migration of hydrocarbons in sedimentary basins. Editions Technip, Paris, 113-130. 1992. Primary oil migration and expulsion as determined by hydrous pyrolysis. Proceedings of the 13th World Petroleum Congress, Buenos Aires, Argentina 1991, 2, 215-223. ~, WINTERS, J.C. & MCDONALD, J.H. 1979. Generation of oil-like pyrolysates from organicrich shales. Science, 203, 896-899. LEYTHAEUSER,D. & POELCHAU,H.S. 1991. Expulsion of petroleum from type III kerogen source rocks in gaseous solution: modelling of solubility fractionation. In: ENGLAND, W. & FLEET, A.J. (eds) Petroleum Migration. Geological Society, London, Special Publications, 59, 33-46. -& SCHWARZKOPF,T.H. (1986) The pristane/nheptadecane ratio as an indicator for recognition of hydrocarbon migration effects. Organic Geochemistry, 12, 191-198. ~, LITTKE, R., RADKE, M. & SCHAEFER,R.G. 1988. Geochemical effects of petroleum migration and expulsion from Toarcian source rocks in the Hiis syncline area, NW-Germany. Organic Geochemistry, 13,489-502. - - , MACKENZIE, A.S., SCHAEFER, R.G. & BJOROY, M. 1984. A novel approach for recognition and

258

U. MANN

quantification of hydrocarbon migration effects in shale-sandstone sequences. American Association of Petroleum Geologists Bulletin, 68, 196219. ~, RADKE, M. & WILLSCH, H. 1988. Geochemical effects of primary migration of petroleum in Kimmeridge source rocks from Brae field area, North Sea. II: Molecular composition of alkylated naphthalenes, phenanthrenes, benzoand dibenzothiophenes. Geochimica et Cosmochimica Acta, 52, 2879-2891. , SCHAEFER, R.G. & Yf0KLER, A. 1982. Role of diffusion in primary migration of hydrocarbons. American Association of Petroleum Geologists Bulletin, 66,408-429. LINDGREEN, H. 1985. Diagenesis and primary migration in Upper Jurassic claystone source rocks in North Sea. American Association of Petroleum Geologists Bulletin, 69,525-536. 1987a. Experiments on adsorption and molecular sieving and inferences on primary migration in Upper Jurassic claystone source rocks, North Sea. American Association of Petroleum Geologists Bulletin, 71,308--321. 1987b. Molecular sieving and primary migration in Upper Jurassic and Cambrian claystone source rocks. In: BROOKS, J. & GLENNIE, K.W. (eds) Petroleum geology of northwest Europe, 1. Graham & Trotman, London, 357-364. LITrKE, R. & RULLKOTrER, J. 1987. Mikroskopische und makroskopische Unterschiede zwischen Profilen unreifen und reifen Posidonienschiefers aus der Hilsmulde. Facies, 17,171-180. - - , BAKER,D.R. & LEYTHAEUSER,D. 1988. Microscopic and sedimentologicevidence for the generation and migration of hydrocarbons in Toarcian source rocks of different maturities. Organic Geochemistry, 13,549-560. MACKENZIE, A.S., PRICE, I., LEYTHAEUSER, D., M(~LLER, P., RADKE,M. & SCHAEFER,R.G. 1987. The expulsion of petroleum from Kimmeridge clay source rocks in the area of the Brae field, UK continental shelf. In: BROOKS,J. & GLENNIE, K. (eds) Petroleum Geology of North Western Europe. Graham & Trotman, London, 865-877. MAGARA,K. 1975. Importance of aquathermai pressuring effects in Gulf Coast. American Association of Petroleum Geologists Bulletin, 59, 20372045. MAGARA, K. 1980. Evidences of primary migration. American Association of Petroleum Geologists Bulletin, 64, 2108-2117. MANDELBROT, B.B. 1975. Les objets fractale: forme, hasard et dimension. Flammerion, Paris. MANN, D. & MACKENZIE, A.S. 1990. Prediction of pore fluid pressures in sedimentary basins. Marine and Petroleum Geology, 7, 55-65. MANN, U. 1989. Revealing hydrocarbon migration pathways. In: POELCHAU,H.S. & MANN,U. (eds) Geological Modelling. Geologische Rundschau Special Volume, 78/1,337-348. 1990. Sedimentologicaland petrophysical aspects of primary petroleum migration pathways. In: HELING, D., ROTHE, P., FORSTNER, U. & -

-

-

-

-

-

STOFFERS, P. (eds) Sediments and environmental geochemistry. Springer, New York, 152-178. 1991. A multidisciplinary approach to primary migration pathways of petroleum - the Lower Toarcian of NW-Germany as a model source rock. In: IMARISIO, G., FRIAS, M. & BEMTGEN, J.M. (eds) The European oil and gas conference, a multidisciplinary approach in exploration and production R&D proceedings. Graham & Trotman, London, 142-154. ~, DUPPENBECKER,S., LANGEN,A., ROPERTZ,B. & WELTE, D.H. 1990. Pore network evolution of the lower Toarcian Posidonia Shale during petroleum generation and expulsion - A multidisciplinary approach. Zentralblatt Geologie & Paliiontologie Teil I, 8, 1051-1071. , JOCHUM, J., Dt PRIMIO, R. & ROPERTZ, B. 1991. Expulsion pathways from petroleum source rocks. 3rd Conference European Association of Petroleum Geoscientists, Florence 1991, preprint. ~, NEISEL, J.D., BURCHARD,W.-G., HEINEN,V. WELTE, D.H. 1993. Fluid-rock interfaces as revealed by cryo-scanning electron microscopy. First Break, in press. MATSUMOTO, R. & IIJIMA, A. 1981. Origin and diagenetic evolution of Ca-Mg-Fe carbonates in some coal fields of Japan. Sedimentology, 28, 239-260. MCAULIFFE,C.D. 1966. Solubility in water of paraffin, cycloparaffin, olefin, acetylene, cycloolefin and aromatic hydrocarbons. Journal of Physical Chemistry, 70, 1267-1275. 1979. Oil and gas migration - chemical and physical constraints. American Association of Petroleum Geologists Bulletin, 63,761-781. MEINSCHEIN,W.G. 1959. Origin of petroleum. American Association of Petroleum Geologists Bulletin, 43,925-943. MEISSNER, F. 1978. Petroleum geology of the Bakken Formation, Williston Basin, North Dakota and Montana. In: Williston Basin proceedings 1978: Montana Geological Society, Billings, 207-227. MOMPER, J.A. 1978. Oil migration limitations suggested by geological and geochemical considerations. In: ROBERTS,W.H., t~ CORDELL,R. (eds) Physical and chemical constraints on petroleum migration. American Association of Petroleum Geologists Geological Course Notes, 8, B1-B60. MORROW, C.A., SHI, L.Q. & BEYERLEE,J.D. 1984. Permeability of fault gauge under confining pressure and stress. Journal of Geophysical Research, 89, 3193-3200. MossoP, G.D. 1972. Origin of the peripheral rim, Redwater reef, Alberta. Bulletin of Canadian Petroleum Geology, 20,238-280. NAKAYAMA,K. 1987. Hydrocarbon-expulsion model and its application to Niigata area, Japan. American Association of Petroleum Geologists Bulletin, 71,810-821. NEISEL, J.D., BURCHARD, W.-G., CLOTH, G. & WIEGERSHAUS, J. 1991. Mobile Einrichtung ftir Kryofixation, Kryolagerung und Kryotransport feuchter mineralischer Baustoffe. Beitriige -

-

PRIMARY PETROLEUM MIGRATION

-

elektronenmikroskopischer Direktabbildung von Oberfliichen, 24, 17-22. (3ZKAYA, I. 1984. Computer simulation of hydraulic fracturing in shales - influences on primary migration. Journal of Petroleum Technology, 24, 826-828. 1986. Analysis of factors influencing excess heads in shales during burial. Marine and Petroleum Geology, 3, 74-78. 1988. A simple analysis of oil-induced fracturing in sedimentary rocks. Marine and Petroleum Geology, 5,293-297. 1991. Computer simulation of primary oil migration in Kuwait. Journal of Petroleum Geology, 14, 37--48. PALACAS, J.G. 1984. Carbonate rocks as sources of petroleum: geological and chemical characteristics and oil-source correlations. Proceedings of the llth World Petroleum Congress, 2, 31-43. PALCIAUSKAS,V.V. & DOMENICO,P.A. 1980. Microfracture development in compacting sediments: relation to hydrocarbon-maturation kinetics. American Association of Petroleum Geologists Bulletin, 64,927-937. PANDEY, G.N., RASIN TEK, M. & KATZ, D.L. 1974. Diffusion of fluids through porous media with implications in petroleum geology. American Association of Petroleum Geologists Bulletin, 58, 291-303. PASSEY, Q.R., CREANY, S., KULLA, J.B., MORETTI, F.J. & SVROUD,J.D. 1990. A practical model for organic richness from porosity and maturity logs. American Association of Petroleum Geologists Bulletin, 74, 1777-1794. PARNELL, J. & EAKIN, P. 1987. The replacement of sandstones by uraniferous hydrocarbons: significance for petroleum migration. Mineralogical Magazine, 51,505-515. PERRY, E.A., JR & HOWER, J. 1972. Late-stage dehydration in deeply buried pelitic sediments. American Association of Petroleum Geologists Bulletin, 56, 2013-2021. POLLASTRO,R.M. & SCHOLLE,P.A. 1986. Exploration and development of hydrocarbons from lowpermeability chalks - an example from the upper Cretaceous Niobrara Formation, Rocky Mountains region. In: SPENCER,C.W. • MASTS, R.F. (eds) Geology of tight gas reservoirs. American Association of Petroleum Geologists Studies in Geology, 24,129-141. PRICE, L.C. 1976. The solubility of petroleum as applied to its origin and primary migration. American Association of Petroleum Geologists Bulletin, 60, 213-244. 1989. Primary petroleum migration from shales with oxygen-rich organic matter. Journal of Petroleum Geology, 12,289-324. & CLAYTON,J.L. 1992. Extraction of whole versus ground source rocks: fundamental petroleum geochemical implications including oilsource rock correlation. Geochimica et Cosmochimica Acta, 56, 1213-1222. PROZOROVICH, G.E., SOKOLOVSKrY,A.P. & MALYKH, A.G. 1973. Regenerated minerals in Volgian -

-

-

-

-

259

oil-bearing clay of the Salym field, western Siberia. Doklady Akademii Nauk SSSR, 208, 945-947, (in Russian). RADrE, M. & WEE'rE, D.H. 1983. The Methylphenanthrene Index (MPI): a maturity parameter based on aromatic hydrocarbons. In: BJOROYet al. (eds) Advances in Organic Geochemistry. Wiley, New York, 504-512. RAMSDEN, R.M. 1952. Stylolites and oil migration. American Association of Petroleum Geologists Bulletin, 11, 2185-2186. RICKEN, W. 1986. Diagenetic bedding. Lecturer Notes in Earth Sciences, 6,206. RrrrER, U. 1988. Modelling of hydrocarbon generation patterns in the Egersund Sub-Basin, North Sea. Organic Geochemistry, 13, 165-174. ROEODER, E. 1984. Fluid inclusions. Mineralogical Society of America Reviews in Mineralogy, 12. ROPERTZ, B. 1989. Petrophysikalische, organischchemische und organisch-petrographische Untersuchungen zur Verteilung des organischen Materials in Ton- und Mergelsteinen in zwei ausgewiihlten Katagenesestadien (Unteres Toarc, Hilsmulde, NW-Deutschland). Diplomarbeit, RWTH Aachen. -1993. Wege der primiiren ErdOlmigration: Eine Untersuchung iiber Porennetze, Kliifte und Kerogennetzwerke als Leiterbahnen fiir den Kohlenwasserstofftransport. Dissertation RWTH Aachen. SAJGO, C.S., MAXWELL, J.R. & MACKENZIE, A.S. 1983. Evaluation of fractionation effects during the early stages of primary migration. Organic Geochemistry, 5, 65-73. SANDVIK, E.I. & MERCER,J.N. 1990. Primary migration by bulk hydrocarbon flow. Organic Geochemistry, 16, 83-89. SASSEN, R., MOORE, C.H. & MEENDSEN, F.C. 1987. Distribution of hydrocarbon source potential in the Jurassic Smackover Formation. Organic Geochemistry, 11,379-383. SCHOLLE, P.A. & HALLEY, R.B. 1985. Burial diagenesis: out of sight, out of mind! In: SCHNEIDERMANN, N. & HARRIS, P.M. (eds) Carbonate cements. Society for Sedimentary Geology, Special Publications, 36,309-334. SEEWALD, H. 1982. Charakterisierung von Mikroporen und ihre Bedeutung fiJr die Kohletechnik. ErdOl und Kohle- Erdgas- Petrochemie, 35, 418--427. SHI, J., LI, B., LIu, D. & PENG, P. 1988. Organic inclusions and their relations to oil and gas occurrences. Organic Geochemistry, 13, 11011107. Snl, Y. t~ WANG,C.Y. 1986. Pore pressure generation in sedimentary basins: overloading versus aquathermal. Journal Geophysical Research, 91, 2153-2162. SmsoN, R.H. 1981. Fluid flow accompanying faulting. In: Earthquake prediction and an International review. Maurice Ewing Series, Washington, 4, 593--603. , MORE, J. MCM. & RANKIN, A.H. 1975. Seismic pumping: a hydrothermal fluid transport

2~

U. MANN mechanism. Journal of Geological Society, London, 131,653-659.

Smcn, K.S.W. 1976. Surface characterization: physical. In: PARFrr, G.D. & SIN~H, K.S.W. (eds) Characterization of powder surfaces. Academic Press, London, 1-56. SNARSKV, A.N. 1962. Die prim/ire Migration des Erd61s. Freiberger Forschungshefte, C123, 63-73. SOEDER,D.J. 1990. Application of fluorescence microscopy to study of pores in tight rocks. American Association of Petroleum Geologists Bulletin, 74, 30-40. SO~:OLOV, V.A. 1948. Essays on the genesis of petroleum. Gostoptekhizdat, 447. ~, ZHAUSE, T.P., VASSOYEWCH,N.B., ANTONOC, P.O., GRmORrVEV, G.G. & KOZLOV, V.P. 1964. Migration processes of gas and oil, their intensity and directionality. Proceedings of the 6th World Petroleum Congress, Frankfurt 1963, 493-505. STAINFORTH,J.G. & REINDERS,J.E.A. 1990. Primary migration of hydrocarbons by diffusion through organic matter networks, and its effect on oil and gas generation. Organic Geochemistry, 16, 61-74. STOCKDAEE, P.B. 1926. The stratigraphic significance of solution in rocks. Journal of Geology, 34, 399-414. STONELEY, R. 1983. Fibrous calcite veins, overpressures, and primary migration. American Association of Petroleum Geologists Bulletin, 67, 14271428. TALUKDAR, S., GALLANGO, O., VALLEJOS, C. & RUCCIERO, A. 1987. Observations on the primary migration of oil in the La Luna source rocks of the Maracaibo Basin, Venezuela: In: Doligez, B. (ed.) Migration of hydrocarbons in sedimentary basins. Editions Technip, Paris, 59-78. TANNENBAUM, E., BRADLEY, J.H. & KAPLAN, J.R. 1986. Role of minerals in thermal alteration of organic matter- II: A material balance. American Association of Petroleum Geologists Bulletin, 70, 1156-1165.

THOMAS, M.M. & CLOUSE, J.A. (1990). Primary migration by diffusion through kerogen: III. Calculation of geologic fluxes. Geochimica et Cosmochimica Acta, 54, 2793-2797. TISSOT, B. & PELET, R. 1971. Nouvelles donn6es sur ies m&hanismes de gen6se et de migration du p6trole, simulation math6matique et application fi la prospection. 8th World Petroleum Congress, Moscow, 4. Wiley, Chichester, 35-46. & WELTE, D.H. 1984. Petroleum formation and occurrence, 2nd edition. Springer, New York. UNCERER, P. 1990. State of the art of research in kinetic modelling of oil formation and expulsion. Organic Geochemistry, 16, 1-26. , BEHAR, E. & DISCAMPS, D. 1983. Tentative calculation of the overall volume expansion of organic matter during hydrocarbon genesis from geochemistry data. Implications for primary migration. In: BJOROY, M. et al. (eds) Advances in Organic Geochemistry, 10, 129-135. , BESSlS, F., CHENET, P.Y., DURAND, B., NOGARET, E., CHIARELLI, A., OUDIN, J.L. &

PERRIN, J.F. 1984. Geological and geochemical models in oil exploration; principles and practical examples. In: DEMAISON, G. & MURRIS, R.J. (eds) Petroleum Geochemistry and Basin Evol-

ution. American Association of Petroleum Geologists Memoirs, 35, 53-78. --, NGOKWEY, J.M., NOGARET, E. & PERRIN, J.F. 1983. Geological deterministic models and oil exploration: principles and practical examples. American Association of Petroleum Geologists Bulletin, 67,185. ~, BURRUS, J., DOLIGEZ, B., CHENET, P.Y. & BESSIS, F. 1990. Basin evolution by integrated 2D modelling of heat transfer, fluid flow, hydrocarbon generation and migration. American Association of Petroleum Geologists Bulletin, 74, 309335. VANDENBROUCKE, M. 1972. Etude de la migration primaire: Variation de la composition des extraits de roche a un passage rochem~re/r6servoir. Organic Geochemistry, 5,547-565. ..... , DURAND, B. & OUDIN, J.L. 1983. Detecting migration phenomena in a geological series by means of C1-C35 hydrocarbon amounts and distributions. In: Advances in Organic Geochemistry 1981, 147-155. WAPLES, D.W., KAMATA,H. & StJlZU, M. 1992a. The art of maturity modelling, Part l: finding a satisfactory geological model. American Association of Petroleum Geologists Bulletin, 76, 31-46. ~, SUIZO, M. ff~zKAMATA, H. 1992b. The art of maturity modelling. Part 2: alternative models and sensitivity analysis. American Association of Petroleum Geologists Bulletin, 76, 47-66. WATTS, H. 1963. The possible role of adsorption and diffusion in the accumulation of crude oil deposits; a hypothesis. Geochimica et Cosmochimica Acta, 27,925-928. WATTS, N.L. 1983. Microfractures in chalks of Albuskjell Field, Norwegian sector, North Sea: possible origin and distribution. American Association of Petroleum Geologists Bulletin, 67,201-234. WITHERSPOON,P.A. & SARAF,D.N. 1965. Diffusion of methane, ethane, propane and n-butane in water from 25 to 43° C. Journal of Physical Chemistry, 69, 3752-3755. WELTE, D.H. 1987. Migration of hydrocarbons. Facts and theory. In: DOLIGEZ, B. (ed.) Migration of hydrocarbons in sedimentary basins. Editions Technip, Paris, 393-413. & YUKLER, M.A. 1981. Petroleum origin and accumulation in basin evolution - a quantitative model. American Association of Petroleum Geologists Bulletin, 65, 1387-1396. YAL•IN, M.N. & WELTE, D.H. 1988. The thermal evolution of sedimentary basins and significance for hydrocarbon generation. Bulletin of Turkish Petroleum Geologists, 1, 1-12. YUKLER, A.M., CORNFORD,C. • WELTE, D.H. 1978. One-dimensional model to simulate geologic, hydrodynamic and thermodynamic development of a sedimentary basin. Geologische Rundschau, 67,960-979.

Organic matter alteration and fluid migration in hydrothermal systems BERND

R.T. SIMONEIT

Petroleum and Environmental Geochemistry Group, College of Oceanic and Atmospheric Sciences, Oregon State University, Corvallis, OR 97331, USA Abstract: Hydrothermal systems associated with oceanic spreading centres are now recognized as relatively common phenomena, and the organic chemistry occurring there is geologicallyrapid and novel. Especially in the marine hydrothermal systems at water depths >1.5 km, this chemistry proceeds at high temperatures (up to >400 ° C) and high pressures (>150 bar) in an aqueous open flow medium. Continental systems may also be of interest, as for example failed or dormant rifts and regions around piercement volcanoes. Organic matter alteration to petroleum hydrocarbons by reductive reactions occurs in hydrothermal systems over a wide temperature window (c. 60 to >400 ° C), under high pressure, and in a brief geological period of time (years to hundreds of years). Thus, the products are considered to be in a metastable equilibrium state during their brief formation and residence times at high temperatures. These conditions are conducive to organic chemistry which yields concurrent products by reduction (due to mineral buffering), oxidation (high thermal stress) and synthesis. Therefore, this brief time interval coupled with the wide range of hydrothermal petroleum compositions indicates that organic matter maturation and petroleum generation, expulsion and migration occur as an overlapping continuous process during hydrothermal activity. The behavior of organic matter (CH4-C40+) in near to supercritical water, in many cases with associated high concentrations of COz and methane, needs to be further characterized in order to understand the implications of this novel phenomenon in geological and geochemical processes.

Hydrothermal systems are found along the active tectonic areas of the earth and are defined here as fluid flow regimes with thermal gradients at elevated temperatures. The different types of hydrothermal systems under study and their fluid temperatures are summarized in Table 1. These systems are comprised of spreading ridges as well as off-axis flanks and basins, back-arc activity, hot spots, volcanism, and subduction. There are currently about 100 locations with known hydrothermal activity and associated mineralization at various seafloor spreading centres (divergent plate boundaries) (see reviews by Rona 1984, 1988). Those hydrothermal systems with associated organic matter alteration are indicated on the tectonic sketch map in Fig. 1. These compilations (Table 1 and Fig. 1) are expected to expand as exploration continues. Two continental hydrothermal systems are also shown in Fig. 1. Petroleums generated in high temperature and rapid fluid flow regimes are defined here as hydrothermal because the agent of thermal alteration and mass transfer is hot circulating water (temperature range warm to >400 °C). This water is responsible for organic matter alteration (generally reductive) and product expulsion and migration from the source rocks

or unconsolidated sediments in a continuous single process (Didyk & Simoneit 1989, 1990; Simoneit 1983). The process occurs in systems with high water-rock ratios. These hydrothermal oils generally contain disequilibrium reaction products comprised of reduced and oxidized species (e.g., methylcyclopentane v. benzene, cholestane v. Diels' hydrocarbon; Kawka & Simoneit 1987, 1990; Simoneit et al. 1990a, 1992a, b). In contrast, conventional oils are products derived from basin evolution and are generated contemporaneously with sediment compaction and heating (temperature window: warm to c. 150° C; Hunt 1979; Tissot & Welte 1984). Generation of hydrothermal oils and gases is a geologically rapid process, taking place within hundreds to thousands of years (Peter et al. 1991; Simoneit & Kvenvolden 1994), whereas conventional oils are generated at a rate that is tied to basin subsidence and increasing geothermal gradient occurring over millions of years (Hunt 1979; Tissot & Welte 1984).

Experimental methods The experimental methods for sample preservation, headspace gas extraction, bitumen extraction and separation, and instrumental

FromPARNELL,J. (ed.), 1994, Geofluids:Origin, Migrationand Evolution of Fluidsin SedimentaryBasins, 261 Geological Society Special Publication No. 78,261-274.

262

Table 1.

B.R.T. SIMONEIT Types o f hydrothermal systems

Typical discharge temperatures (°C)

Examples studied

References

Marine (recharge sea water)

Bazylinski et al. 1988; Kvenvolden et al. 1986; Michaelis et al. 1990; Simoneit & Lonsdale 1982; Simoneit 1985; Simoneit et al. 1987; Davis et al. 1992 Brault & Simoneit 1989; Brault et al. 1985, 1989

Sediment-covered spreading ridge

Guaymas Basin, Escanaba Trough, Middle Valley, Red Sea

Warm up to c. 400° C

Mid-Ocean ridge (no sediment) Off-axis flanks and basins Back-arc

East Pacific Rise, MidAtlantic Ridge

Warm up to c. 350° C

Bransfield Strait

No discharge (< 150° C)

Brault & Simoneit 1988; Whiticar et al. 1985

Hot spots Subduction

-Oregon Margin

-Ambient

Kulm et al. 1986

<96 ° C

Clifton et al. 1990

65-80 ° C <100 °C

Tiercelin et al. 1989 Czochanska et al. 1986

Continental (recharge meteoric water)

Hot spots

Yellowstone National Park Lake Tanganyika Waiotapu, NewZealand

Rift valleys Volcanism

- - Organic matter not yet studied.

, ..,; "" .~'

'

"~'. ,

a

~ C kinawa

Park

, "";'""'~, ~"""~:'~.~

~

• !.

.

- - ~

,

'

¢'_

--.l:;':

~'

.'j _

.

Strait

Fig. 1. General location map of the hydrothermal vent fields discussed here with a sketch of the global tectonics. analyses for molecular compositions by gas c h r o m a t o g r a p h y and gas c h r o m a t o g r a p h y - m a s s s p e c t r o m e t r y have b e e n published elsewhere ( K v e n v o l d e n et al. 1986; K a w k a & Simoneit 1987; Simoneit et al. 1988).

Geological locales with hydrothermal organic products Sedimented

submarine

systems

G u a y m a s Basin (Fig. 1) is an active oceanic spreading centre (2000 m water d e p t h in the rifts) w h e r e s e d i m e n t a t i o n is rapid ( > 2 m ka -1)

HYDROTHERMAL PETROLEUM IOO% AROMATIC HC

ET GB

10o%'"'~ ALIPHATIC HC

• •

(GB-DSDP o)

"~100% NON-HYDROCARBONS

Fig. 2. Ternary diagram representing gross (C15+) compositions of hydrothermal petroleum as percentages of each of three major compound classes determined gravimetrically. GB, Guaymas Basin; ET, Escanaba Trough; MV, Middle Valley. Typical conventional petroleum falls within hachured area (Tissot & Welte 1984). Adapted from Kawka & Simoneit (1987), Kvenvolden & Simoneit (1990), and Simoneit et al. (1992). covering the rift floors to a depth of c. 300-500 m (Curray et al. 1982). The organic matter of these unconsolidated recent sediments is derived primarily from diatomaceous and microbial detritus, and the sediments average about 2% organic carbon (Curray et al. 1982). Numerous hydrothermal mounds rise to 20-30 m above the south rift floor, and most mounds are actively discharging vent fluids with water temperatures ranging from warm up to 350°C at c.200 bars (Lonsdale 1985; Lonsdale & Becker 1985). The mounds are composed of complex deposits of sulphide, sulphate, silicate and carbonate minerals, as well as petroleum (Bazylinski et al. 1988; Peter 1986; Koski et al. 1985; Simoneit 1985a, b; Simoneit & Lonsdale 1982; Kawka & Simoneit 1987). They are covered with colonies of tube worms, bacterial mats, and many other chemosynthetic organisms (Bazylinski et al. 1989; Jones 1985; Tunnicliffe 1991, 1992). Typical hydrothermal oils from this basin are depleted in volatile aliphatic hydrocarbons compared to conventional crude oils (Didyk & Simoneit 1990). The normalized percentage compositions of the fractions from liquid chromatography of each sample are plotted as a ternary diagram (Fig. 2), where the gross (C15+) composition of all analysed petroleum-bearing samples from Guaymas Basin, Escanaba Trough, and Middle Valley are shown (Kawka & Simoneit 1987; Kvenvolden & Simoneit 1990; Simoneit et al. 1992a, b). Most of the Guaymas Basin and all of the Escanaba Trough samples fall outside the

263

field of typical conventional petroleums due to their enhanced contents of aromatic and polar components. The Middle Valley sample and a few of the Guaymas Basin samples are within the field due to an enrichment in aliphatic and/or naphthenic hydrocarbons. The n-alkane distributions (e.g. Fig. 3a) generally fall within the range from CH4 to >C4o, with usual carbon number maxima in the mid-C20 region and no carbon number predominance (CPI* = 1.0) (Simoneit & Lonsdale 1982; Kawka & Simoneit 1987). A significant amount of an unresolved complex mixture (UCM) of branched and cyclic naphthenic hydrocarbons is also present. The generation of the complete suite of saturated biomarkers from their biological precursors (e.g., cholestanes from cholesterol) is supportive evidence for the strongly reductive process operating during initial organic matter alteration. The major resolved peaks in the aromatic/naphthenic fraction are unsubstituted polynuclear aromatic hydrocarbons (PAH, Fig. 3b). P A H become the dominant species due to their high thermal stability as well as their enhanced solubility in near- and supercritical water (e.g., Sanders 1986). The aromatic/naphthenic fractions of the Guaymas oils also contain significant amounts of N, S, and O hetero-PAH (e.g., Gieskes et al. 1988), and Diels' hydrocarbon (Simoneit et al. 1992a). The chemical compositions of the aromatic fractions indicate an origin from oxidative alteration of precursor compounds at high temperatures in the system (>300 ° C). The volatile and most of the soluble compounds (mainly C1-C10 hydrocarbons) are not effectively retained with the heavy petroleum as it solidifies at the vents on the seafloor of Guaymas Basin. Upon exiting at the seabed the fluids are often saturated with a broad range of volatile hydrocarbons (CH4 to CmH22) as well as lower concentrations of heavy ends (>C15) (Simoneit et al. 1988). Thus, the headspace gases of a Guaymas Basin mound sample (1629-A3, Fig. 4a) can be compared with the hydrocarbon content of a 308 ° C vent water, which is highly enriched in the lower alkanes (
* Carbon Preference Index (modified): for hydrocarbons it is expressed as a summation of the odd carbon number homologues over the range present, divided by a summation of the even carbon number homologues over the same range (Cooper & Bray 1963; Simoneit 1978)

264

B.R.T. SIMONEIT

b)

GB - 1172- 4

ET-659-RI

20

25 27

18 31

27

UCM TIME

-

TIME

C)

GB-TD-ZB

~

1

I.

d)

ET-$20-2

°

,

. . . .

TIME e)

,5j~6 i ] / i p h

20

22 24 119 21 26

17

i

858C-2 H-CC

1618 ,

I/

TIME

Pr Ph

"

•

856A-2H-3,2983-85cm

28 II

I

u 31 32

.t

;zz

c

I

31

~

28

[

35

32 33

3.0

UCM ,

i

,

i

,

I

i

i

,

i

,

i

,

i

,

Time

•

i

.

i

.

|

.

,

Time

,

I

a

I

J

!

,

~,,

Fig. 3. Gas chromatograms of saturated (a, b), aromatic (c, d) and total (e, f) hydrocarbons in: (a, c) Guaymas Basin, (b, d) Escanaba Trough, and (e, f) Middle Valley oils (Kvenvolden & Simoneit 1990; Davis et al• 1992; Simoneit 1994)• (Numbers refer to carbon chain length of n-alkanes; Pr, pristane; Ph, phytane; asterisk, other isoprenoids; UCM, unresolved complex mixture; PAH are labelled).

soluble) versus aliphatic hydrocarbons (Fig. 4b). Interstitial gas in sediments of Deep Sea Drilling Project (DSDP) cores from Guaymas Basin consists of biogenic methane overprinted by thermogenic C H 4 t o C5 hydrocarbons near the sills and, to a lesser extent, at increasing

sub-bottom depths (Galimov & Simoneit 1982; Simoneit & Galimov 1984; Whelan et al. 1988). The sub-bottom hydrocarbons are of a similar composition to the venting volatile hydrocarbons. Escanaba Trough in the northeastern Pacific

HYDROTHERMAL PETROLEUM

Oil- Headspace Gas

n8

nlO

~-

..,."

== •-' n

!

I I I I

! I

fr

Mcs n

I

DMC5 .

a

!

I

g:

6

1'249

,, g

n9

i,k

t-

-~

265

/ ,', d q

'ac

.L 1

~x

~ ~6.~. ~ ~

In7 M.Cf6

T

I

I

!

I

II

~n8

, ,Ll

ox,

"

EB

!n9

=nlO

oX

Time ---,

Fig. 4. Gas chromatograms of headspace analyses for comparison of the volatile hydrocarbons from: (a) hydrothermal mineral/oil crust, 1629-A3; (b) hot venting water, 1620-2C (T = 308°C) (Simoneit et al. 1988). Numbers refer to carbon chain length with n, normal; i, iso- (2-methyl-); a, anteiso- (3-methyl-) of corresponding chain length. Other acyclic compounds are: q, 2,6-dimethylheptane; r, 2,3-dimethylheptane; s, 2,6-dimethyloctane. Cyclic compounds are: C, cyclo-; MC, methylcyclo-; DMC, dimethylcyclo-alkanes.The DMC5 triplet contains the c, cis-1,3-; d, trans-l,3-; e, trans-1,2-dimethyl isomers. Other individual alkylcyclopentanesare f, 1,1,3-trimethyl-; g, 1,2,4-trimethyl-; h, 1,2,3-trimethyl-. The aromatics are: B, benzene; EB, ethylbenzene; T, toluene and X, xylenes (with (o) ortho-, (m) meta- and (p) para-isomers). * Indicates a coeluting unknown and the symbol 1/2x reflects a signal attenuation by a factor of 2.

(Fig. 1) is an active oceanic spreading centre about 300 km long and bounded on the north and south by the Blanco and Mendocino fracture zones respectively. It is filled with up to 500 m of Quaternary turbidite sediments (Kvenvolden et al. 1986) which have lower organic carbon contents (average - 0 . 8 % ) than sediments in Guaymas Basin. The petroleum in the sediments and mineral ores occurs as veins or cement and is derived from hydrothermal alteration of sedimentary organic matter primarily from terrestrial sources (Kvenvolden et al. 1986, 1990). The bulk compositions of these hydrothermal petroleums are highly aromatic with a significant NSO-compound content, well outside the field of conventional oils (Fig. 2). Examples of the compositions of the saturated and aromatic hydrocarbon fractions of a hydrothermal petroleum from Escanaba Trough are shown in Fig. 3c,d. The n-alkanes range from C14 to C4o, with a carbon number maximum at n-C27

and a significant odd carbon number predominance >n-C25 (CPI = 1.25), compared to conventional crude oils (CPI = 1.0). The UCM is very low. This composition is typical of a catagenetic product from organic matter of a terrestrial origin (Kvenvolden et al. 1990; Kvenvolden & Simoneit 1990). The P A H are more concentrated relative to the UCM in the aromatic fraction when compared to the example from Guaymas Basin (Fig. 3b) although the relative yields are similar. This may be due to the deposition of more oxidized terrestrial organic detritus and precursor biomarkers in Escanaba Trough than in Guaymas Basin. Volatile hydrocarbons are not found in the oils from Escanaba Trough. Middle Valley is another sediment-covered hydrothermal system in the northeastern Pacific (Fig. 1), with associated hydrothermal organic matter alteration (Simoneit et aI. 1992b). These oils are also highly aromatic/polar in

266

B.R.T. SIMONEIT

composition (Fig. 2). This system was recently drilled by the Ocean Drilling Program (Davis et al. 1992; Simoneit 1994) and hydrothermal petroleum was recovered in some of the core sections. The bitumen in the thermally unaltered sections reflects the admixture of marine autochthonous compounds (e.g., pristane-Pr, phytanePh and n-alkanes C23 from terrestrial vascular plant wax (Fig. 3e). This sample is from 6.4 m below the seafloor and has a CPI26-35 of 2.84. The compositional signatures of the total hydrocarbon fractions of the hydrothermal petroleums in the various core sections were very diverse, comprised of oils consisting of either UCM, aromatics, condensate/volatiles ( C m a x = 17), to aliphatic higher molecular weight mixtures. The example shown in Fig. 3f is a solid bitumen at ambient temperature from 13 m below seafioor and has a CPI26-35 of 0.46, with an n-alkane range of c. C~4 to C3s. The strong even carbon number predominance is intriguing and occurs in numerous hydrothermal petroleums from ODP 139 in Middle Valley (Simoneit 1994). The Bransfield Strait, Antarctica (Fig. 1) is a typical example of a heavily sedimented backarc rift, which is tectonically active with extensional features such as dip-slip faults and intrusive rocks (Whiticar et al. 1985; Suess et al. 1994). Hydrothermat activity is evidenced by mineral alteration and a slight petroliferous odor in some sediment sections. However, the bitumen compositions indicate only mild and localized heating from intrusions, which resulted in accelerated diagenesis and limited product migration (Brault & Simoneit 1988). This heat regime has also been confirmed by amino acid racemization analysis (Silfer et al. 1990). The Atlantis II Deep in the Red Sea (Fig. 1) contains stratified brine layers, where the deepest is presently at a temperature of 62 ° C. Bulk organic matter and hydrocarbons have been analysed in two sediment cores from the Deep, and the results indicate mild hydrothermal alteration (Simoneit et al. 1987). The reductive products, i.e. saturated hydrocarbons, are predominant and the oxidative products, i.e. PAH, are not detectable, confirming that hydrothermal alteration commences at relatively low temperatures. Data on hydrothermal petroleums from the Kebrit and Shaban Deeps of the Red Sea have also been reported; however, these systems were interpreted to be at higher temperatures (Michaelis et al. 1990). Sediment-starved submarine systems

The examples of sediment-starved hydrothermal

systems are located in basaltic rift areas as for example at 26° N on the Mid-Atlantic Ridge and at 13° and 21 ° N on the East Pacific Rise (Fig. 1). Hydrocarbons from metalliferous sediments have distributions characteristic of immature organic matter, which has recently been biosynthesized and microbiologically degraded, as might be expected in the low temperature environment of the surrounding talus at a vent system (Brau|t et al. 1985, 1988; Simoneit et al. 1990a). Thermally mature n-alkanes with no carbon number predominance (CPI = 1.0) and biomarkers (e.g. 17ot(H)-hopanes, steranes) are present at trace levels. Various samples of massive sulphides from vent chimneys at 21°N on the EPR have hydrocarbon contents which are extremely low but originate from hydrothermal alteration (Brault et al. 1989). All samples contain n-alkanes with no carbon number predominance (CPI = 1.0) and PAH, supporting evidence for an origin from high temperature alteration. A sample with pyritized tube worm residues also contains hydrothermally-altered derivatives (e.g., cholestenes, hopenes) of biomarkers characteristic of this vent biota. The Trans-Atlantic Geotraverse (TAG) hydrothermal field on the Mid-Atlantic Ridge crest at 26 ° N (Fig. 1) is an active vent system on a slow-spreading mid-oceanic ridge (Rona et al. 1984). Various hydrothermal mineral ores were dredged from the area (TAG 1985-1) and four types of samples have been examined for lipid/bitumen content (Brault & Simoneit 1989). A ferric oxide sample contained no hydrocarbons attributable to hydrothermal alteration of organic matter, probably due to oxidative loss. However, three other samples (consisting of mainly anhydrite, sphalerite and chalcopyrite, respectively) did contain trace amounts of lower molecular weight (C10-C22) hydrothermal petroleums, consisting of both n-alkanes (with UCM) and PAH. Continental systems

Continental hydrothermal systems occur in volcanic or failed and dormant rift terranes, as for example Yellowstone National Park, Lake Tanganyika, and Waiotapu (Fig. 1). In most cases, the hydrothermal processes cause remobilization of organic matter in the form of bitumen as illustrated by examples of oils from Yellowstone National Park (Love & Good 1970; Clifton et al. 1990). The following two cases are examples of petroleums generated within continental hydrothermal systems. In the Waiotapu geothermal region of New Zealand, small amounts of oil are currently

HYDROTHERMAL PETROLEUM being generated from volcanic sedimentary rocks of Lower Pleistocene age (Czochanska et al. 1986). The source material is terrigenous organic matter present in vitric tuff which has been rapidly buried by volcanic overburden. The associated breccias serve as regional aquifers and surround the tuff with high temperature water. The generated oil, however, lacks the high temperature reaction products, e.g., PAH, present in typical hydrothermal petroleums. Massive sulphides and petroleum occur in the north Tanganyika trough of the East African Rift (Tiercelin et al. 1989, 1993). Hydrothermal fluids pass through about 2 km of organic-rich lacustrine sediments (algal detritus), mobilizing asphaltic petroleum and venting at the lake bed in a water depth of c. 20 m with temperatures of 65-80 ° C. The site described is in close proximity to shore; vents at higher temperatures are suspected to occur in deeper water of the lake. The vent waters also contain thermogenic hydrocarbons (Tiercelin et al. 1989). Hydrothermal activity can generate and migrate petroleum from continental source material in both lithified rocks and unconsolidated sediments. The invasion of hydrothermal fluids into mature source rocks will result in migration by remobilization with some alteration of the bitumen in the formation. Invasion of hydrothermal fluids into unconsolidated recent sediments will result in organic matter alteration as observed in the marine systems, except the effective temperature may be lower due to the lower confining pressure of the overburden.

Hydrothermal petroleumexpulsion/ extraction/migration Generally, the volatile hydrocarbon mixtures in hydrothermal fluids exhibit large variations in character in terms of carbon number range (CH4-C10+), structural diversity (relative contents of the normal, branched and cyclic components) and polarity (aliphatic versus aromatic components) (Simoneit et al. 1988). This character is controlled by a number of factors, primarily temperature, aqueous solubility, biodegradation, and water-washing. Migration of these volatile hydrocarbons occurs through dispersion in vent fluids and as a bulk phase in the sediments and vein systems. The more soluble and volatile hydrocarbons are released into the water column by rapidly venting fluids, rising in some cases as large plumes (Merewether et al. 1985; Simoneit et al. 1990b), and by aqueous remobilization from some exposed hydrothermal mounds.

267

Although direct measurement of the oil flow rate at the vent sites of the Guaymas Basin has not yet been feasible, oil globules have been collected under 'in situ' conditions (200 bar, 2-3 ° C). These samples had a gas to oil ratio ranging from approximately 5 to 155 at standard temperature and pressure (Simoneit et al. 1988; Simoneit, unpublished results). The low water temperatures at the sea floor contribute to condensation/precipitation and retention of some of the oil, generating oil impregnations in fluid inclusions (Peter et al. 1990) and on inorganic substrates (Carranza-Edwards et al. 1990) of the hydrothermal vent system. In general, the low pour point (<18 ° C, a consequence of the high liquefied hydrocarbon gas content, Didyk & Simoneit 1990) of the hydrothermal oil allows it to remain fluid at these bottom temperatures. Volatile hydrothermal oil is also trapped in fluid inclusions, especially in amorphous silica of veins and conduits of the vent/mound systems (Peter et al. 1990). The oils in Escanaba Trough and Middle Valley are emplaced as higher temperature fluids (>80 ° C, bulk phase migration) and solidify in the mineral matrix as the temperatures approach ambient. Conventional petroleum formation is believed to occur in the temperature window of c. 60-150°C during basin subsidence over long geological time, and above that temperature the organic compounds are inferred to convert to CH4 and graphite (Hunt 1979; Tissot & Welte 1984). Geologically 'instantaneous' organic matter alteration in hydrothermal systems is a widespread process occurring over a temperature range from c. 60° C to >400 ° C. Formation of hydrothermal petroleums commences under low temperature conditions, generating products from weaker bonds of the generally immature organic matter, and, as the temperature regime rises, additional products are derived from more refractory organic matter and are even 'reformed' (e.g., PAH). The products are continuously expelled/extracted and removed by fluid flow. The process progresses from reductive to more oxidative reactions of the residual organic matter as the temperature increases. At very high temperatures, organic matter is only partly destroyed, probably because the thermogenic products are soluble in the ambient fluid (Connolly 1966; Price 1976; Price et al. 1983; Sanders 1986) and are thus rapidly removed from the hot zone by convection. The thermal alteration products of organic matter in hydrothermal systems can be considered to be in a metastable equilibrium state (e.g. Shock 1988, 1989, 1990) during their brief

268

B.R.T. SIMONEIT

Boiling point (at 1 bar) 25oC I

I000

200 °

I

I

Critical temperature 500° 1400* II I i

500°

I

I

I I

I

I

I

I Near- t Super- Super-heated icritical ' critical I super-critical water I water I water

Normal water 1.0 0.5

I

I •

o

r-

I00 -

~ ~

50

o~

o

.o >,

20 f

~

I

I

I

I

I

I

I I

I

I

"

t I I

~

?

i I

, i I I

I

I

I

"it t ? I

I

I I I

I I

I0

0

ii1,1

---

j

-1-

600°C I

I

I

I

II

Fig. 5. Properties of water as a function of temperature at 200-300 bar pressure (adapted from Josephson 1982). formation and residence times at high temperatures. In this state not all of the stable equilibrium species are present due to kinetic constraints. Thus, Guaymas Basin vent fluids for example, concurrently contain reduced species (e.g., hydrogen, hydrogen sulphide and CH4C40 hydrocarbons), and oxidized species (e.g., CO2, acetate, PAH). The alteration rates of organic matter to petroleum in hydrothermal systems are geologically rapid. For example, carbon-14 dates have been obtained for hydrothermal petroleum from the southern trough of Guaymas Basin (Peter et al. 1991; Simoneit & Kvenvolden 1994). The ages range from c. 3200 to 6600 years BP, mean 4690 years BP (before present, referenced to the year At) 1950 and using the 14C half life of 5570 years). These are not true ages, but rather they reflect the age of carbon within these materials. The age of the 'unaltered' organic carbon in seabed sediments of Guaymas Basin has not been determined, but extrapolation from DSDP cores indicates 0-1000 years BP (Spiker & Simoneit 1982). Additional 14C data on the aliphatic and aromatic hydrocarbon fractions of an oil sample from this area yielded a similar age

(c. 4500 years BP), indicating that the P A H are generated from the same carbon pool as the saturated hydrocarbons at a subseafloor depth of c. 12-30 m. These results demonstrate that late phase, high temperature products such as P A H are derived from shallow depth just as the aliphatic material from lower temperature alteration. Hydrothermal petroleum from the northern trough of Guaymas Basin is 7400 years BP, from Escanaba Trough 17000 years BP, Middle Valley 29000 years BP, and from Lake Tanganyika 25 000 years BP (Simoneit & Kvenvolden 1994), confirming this rapid geological process. Deposition, precipitation and/or trapping of hydrothermal petroleum occur as the migrating fluid experiences reduced temperatures. Phase separation of the oil from water is a consequence of a temperature reduction of the fluid to about 200-300 ° C. Trapping of oil during migration is observed in hydrocarbon fluid inclusions with a mean homogenization temperature of 120°C (Peter et al. 1990). The Guaymas Basin oils are liquid at lower temperatures (e.g., <18°C, Didyk & Simoneit 1990) than are those from Escanaba Trough or Middle Valley due to their high volatile hydrocarbon content (CH4-C10+).

HYDROTHERMAL PETROLEUM Thus, in the temperature window from ambient to c. 300° C the hydrothermal oils are partitioned between bulk phase, microdroplet emulsion and true solution, where the predominance shifts to the former as the temperature decreases. Because these systems are semi-open not all products are trapped. Deposition or precipitation of the heavy components of the oils (> c. C~0-asphalt) occurs at the seafloor as the migrating fluid comes into contact with cold seawater (c. 3° C). This process happens in the mineral mounds and chimneys where the heavy petroleum deposits as a filler in voids of the mineral matrix.

Fluid interactions The interactions of hydrothermal fluid media in terms of chemistry and solvent properties are not well understood. The dominant fluid component is water, and in the example locales of Guaymas Basin and Escanaba Trough, it is at temperatures approaching 350° and 400 ° C, respectively, and under pressures exceeding 200 and 300 bar, respectively. The reduced density of hydrothermal fluids due to heating results in convective circulation, which in effect makes hydrothermal systems semi-open (a flowthrough system) rather than closed as in most laboratory simulation experiments to date. These temperature and pressure conditions are in the near-critical domain of water (Fig. 5; Bischoff & Pitzer 1989; Bischoff & Rosenbauer 1984, 1988; Chen 1981; Josephson 1982; Pitzer 1986). Supercritical water has enhanced solvent capacity for organic compounds and reduced solvation properties for ionic species due to its loss of aqueous hydrogen bonding (Fig. 5; Connolly 1966; Shaw et al. 1991; Siskin & Katritzky 1991; T6dheide 1982). Supercritical water is also a reactive medium for either reductive or oxidative reactions (Ross 1984; Ross et al. 1986; Leif et al. 1991, 1992). Thus, the nearcritical domain of water in hydrothermal systems is expected to aid reaction rates and enhance the solvation capacity for organic matter. Fluids in hydrothermal systems also contain large concentrations of CH4 and CO2 (Sakai et al. 1990; Simoneit & Galimov 1984; Simoneit et al. 1988; Welhan & Lupton 1987). These gases, as well as many other possible trace components, are supercritical under the temperature and pressure conditions of typical hydrothermal systems (CO2 Tcrit = 31 ° C at 73 bar, C H 4 Tcrit = - 8 2 ° C at 46 bar), and their effects on the critical point of seawater are not known. Phase separation of CO2 from water at reduced temperatures has been proposed for liquid CO2 vents in a

269

back-arc hydrothermal system (Sakai et al. 1990). Carbon dioxide liquid (supercritical) is also an excellent solvent for organic compounds (e.g., Paulaitis et al. 1982; Polak et al. 1989). Thus, hydrothermal fluids are efficient solvents for scavenging hydrothermally generated organic compounds (e.g., petroleum) from the source and for migrating them away from the hot zone.

Implications Hydrothermal petroleum formation is a rapid, continuous, and overlapping process consisting of rapid diagenesis of the source organic matter, petroleum generation, expulsion, and migration. In terms of tectonics, hydrothermal systems are particularly active during the early rifting of ocean basins along continental margins (Lonsdale 1985). Thus, geological locales where this process should be considered in exploration are, for example, split rift basins, failed or dormant rifts with hemipelagic or lacustrine sediments, pull apart basins, and rifts overridden by continental drift. Remobilization of petroleum by hydrothermal fluids from magmatic activity affecting conventional sedimentary basins is another aspect for consideration in exploration. A schematic representation of the prevailing conditions existing in the Guaymas Basin vent systems is shown in Fig. 6a (adapted from Scott 1985). Hydrothermal fluids driven by a deep heat source permeate through an open, finegrained body of recent sediments causing organic matter alteration and discharge directly into the water column. The oil discharged with the hydrothermal fluids partially adsorbs or condenses/precipitates on inorganic substrates cooled by sea water (c. 3° C) surrounding the vents. The major part of the oil plume above the vent area dissipates into the water column mainly by dispersion, dissolution, and eventual biodegradation. Another scenario could be postulated, as for example in Fig. 6b, where a similar hydrothermally generated oil is discharged into a porous sediment body, with a finite retention time for the fluids (Didyk & Simoneit 1989, 1990). There, the hydrothermal oil-water mixtures can undergo phase separation as the temperature decreases and petroleum can eventually accumulate if adequate sedimentary and tectonic features are available to constitute a reservoir. Such a scenario could possibly lead to a hydrothermal oil accumulation which would h~ve a potential for exploration.

270

B.R.T. SIMONEIT O Model

of Guaymas

Hydrothermal

b

Basin

Other

Possible

Scenario

Discharges

Sediments or Water Column

o

~

~

.

2OOOm Water Column

"

"" Condensate .. - . i .........-~plume • ";.....: : '" :"

-.""-'.

Phase ~

U-~iii -~

sediments---~\

1

~

~

+iii~+' ~ -]

~

..~/ ~

.

.

.

_

I

~~.-P°rs°~lSme ~

nt s (PossibleTrap)

"

r ' (Sou ce)

v

Hydrothermol " F l u

f l ~ j ~

~-- - - ~ " " - Fine - gr Omed . . . . ~ii~iiii~i!i!!!i!i!:!!iiiiiii~i~i~i!i!!-~J " s-ed~rnenis (Low e'ermeaomTW

I Accumulati°n Separa Fine-groined

_

Dissipation Hydrothermal fluid discharges

/

~Fine-grained

Hydrot~'se~i'menls l- lula

I Energy I Source

~

I Energy I Source

• Open

Sedimentary System • Hydrothermol Fluids Discharge into Water Column • Short or No Retention Time

• Closed Sedimentary System • Hydrothermal Fluids Discharge into Porous Medium • High Retention Time

* Hydrothermal Oil Dissipates and Biodegrades

• Hydrothermal Oil Con Coalesce

• No Preservation of Hydrothermol Oil and no Accumutotion Results in: No Accumulation of Oil

• Possible Preservation and Accumulation of Hydrothermal Oil Results in: Possible Oil Accumulation Preservation of Organics Homogeneous Organics

Only Organic Impregnations around Vent Diversity of Organics

Fig. 6. Schematic models for hydrothermal petroleum generation and migration scenarios (adapted from Didyk & Simoneit 1989, 1990): (a) Guaymas Basin open system; (b) hypothetical closed system.

Table 2. Overview of hydrothermal processes affecting organic matter alteration

Hydrothermal petroleum generation Alteration reactions (relative importance): (a) reduction (primary) (b) oxidation (minor) (c) synthesis (trace) Hydrothermal petroleum composition Source organic matter: (a) marine (e.g. Guaymas Basin) (b) Terrigenous (e.g. Middle Valley)

Hydrothermal fluids as solvents Fluid: (a) water (b) methane (c) carbon dioxide

Compound type: aliphatic hydrocarbons PAH, alkanones thioheterocyclic compounds Products: gas (CH4-CI0) oil (C8--C40+) asphalt (>C40) gas (CH4-C10) oil (C8-C40+) asphalt (>C40)

Relative importance state: dominant/near-critical minor/supercritical minor/supercriticai

Relative concentration: high intermediate minor trace high intermediate Effect: enhanced oil solubility complete oil solubility complete oil solubility

Hydrothermal petroleum migration (a) effective in systems with high water to rock ratios (unlike conventional sedimentary basins) (b) migration bulk phase (trapped in filled veins and voids) emulsion (trapped in fluid inclusions) solution (precipitated as wax crystals in chimneys and fluid conduits)

HYDROTHERMAL PETROLEUM

271

Summary

References

In general, hydrothermal oil generation processes differ significantly from the conventionally accepted scenario for petroleum formation in sedimentary basins, where organic matter input, subsidence, geothermal maturation, oil generation, and oil migration are discrete and successive steps that occur over a long period of geological time. Conversely, in hydrothermal petroleum formation several of the steps of oil generation occur simultaneously and have been shown to complete the organic matter maturation to oil-generating processes in short periods of geological time (Table 2). At seafloor spreading axes, hydrothermal systems active under a sedimentary cover (e.g., Guaymas Basin, Escanaba Trough, etc.) generate petroleum from the generally immature organic matter in the sediments. The alteration reactions are primarily reductive and to a lesser extent oxidative, with trace formation of synthetic products (Table 2). The process occurs in systems with high water to rock ratios, where the interplay of high temperature water, carbon dioxide and methane under pressure represents the effective and efficient thermal driving force for reactions and solvent for product extraction and migration (Table 2). Migration of hydrothermal petroleum has been observed to occur as bulk phase, emulsion and solution (Table 2). It is not known to what extent the hydrothermal petroleum generation process has contributed to the formation of crude oil accumulations, but the process appears to be highly efficient for organic matter maturation, oil generation and its migration. Hydrothermal oil generation should be evaluated thoroughly to document its occurrence and implications, and to yield a better understanding of the geological and geochemical constraints for the process.

BAZYLINKSI, D.A., FARRINGTON,J.W. & JANNASCH, H.W. 1988. Hydrocarbons in surface sediments from a Guaymas Basin hydrothermal vent site. Organic Geochemistry, 12, 547-558. ~, WIRSEN, C.O. & JANNASCH,H.W. 1989. Microbial utilization of naturally occurring hydrocarbons at the Guaymas Basin vent site. Applied Environmental Microbiology, 55, 2832-2836. BISCHOFF,J.L. & ROSENBAUER,R.J. 1984. The critical point and two-phase boundary of sea water, 200--500° C. Earth and Planetary Science Letters, 68, 172-180. & 1988. Liquid-vapor relations in the critical region of the system NaCI-HzO from 380 to 415 ° C: A refined determination of the critical point and two-phase boundary of seawater. Geochimica et Cosmochimica Acta, 52, 21212126. -& PrrZER, K.S. 1989. Liquid-vapor relations from the system NaCI-HzO: Summary of the P-T-x surface from 300° to 500° C. American Journal of Science, 289,217-248. BRAULT, M. & SIMONEIT, B.R.T. 1988. Steroid and triterpenoid distributions in Bransfield Strait sediments: Hydrothermaliy-enhanced diagenetic transformations. In: Advances in Organic Geochemistry 1987. Organic Geochemistry, 13, 697705. &~ 1989. Trace petroliferous organic matter associated with hydrothermal minerals from the Mid-Atlantic Ridge at the Trans-Atlantic Geotraverse 26°N site. Journal of Geophysical Research, 94, 9791-9798. --, MARTY, J.C. & SALIOT, A. 1985. Les hydrocarbures dans le syst6me hydrothermal de la ride Est-Pacifique, ~ 13° N. Comptes Rendus de l'Acad~mie des Sciences, Paris, 301, H, 807-812. -- & 1988. Hydrocarbons in waters and particulate material from hydrothermal environments at the East Pacific Rise, 13°N. Organic Geochemistry, 12,209-219. -& SALIOT, A. 1989. Trace petroliferous organic matter associated with massive hydrothermal sulfides from the East Pacific Rise at 13° N and 21° N. Oceanologica Acta, 12,405--415. CARRANZA-EDWARDS, A., ROSALES-HOz, L., AGUAYO-CAMARGO, J.E., LOZANO-SANTACRUZ, R. & HORNELAS-OROZCO,Y. 1990. Geochemical study of hydrothermal core sediments and rocks from the Guaymas Basin, Gulf of California. In: SIMONEIT,B.R.T. (ed.) Organic Matter Alteration in Hydrothermal Systems - Petroleum Generation, Migration and Biogeochemistry. Applied Geochemistry, 5, 77-82. CHEN, C.-T.A. 1981. Geothermal systems at 21° N. Science, 211,298. CLIFTON, C.G., WALTERS,C.C. & SIMONEIT, B.R.T. 1990. Hydrothermal petroleums from Yellowstone National Park, Wyoming, U.S.A. In: SIMONEIT,B.R.T. (ed.) Organic Matter in Hydrothermal Systems - Petroleum Generation, Migration and Biogeochemistry. Applied Geochemistry, 5,169-191.

I thank the Deep Sea Drilling Project and the National Science Foundation for access to DSDP and ODP samples and participation on the DSV Alvin cruises; J. Baross, M. Brault, C.G. Clifton, B.M. Didyk, E.M. Galimov, J.M. Gieskes, W.D. Goodfellow, P.D. Jenden, O.E. Kawka, K.A. Kvenvolden, R.N. Leif, P.F. Lonsdale, A. Lorre, J.D. Love, M.A. Mazurek, J.M. Peter, R.P. Philp, P.A. Rona, E. Ruth, A. Saliot, M. Schoell, K.A. Sundell, J. Tiercelin and C.C. Waiters for samples, data and assistance, and the reviewers for helpful comments and suggestions to improve this manuscript. Funding from the National Science Foundation, Division of Ocean Sciences (Grant OCE-9002366) and the National Aeronautics and Space Administration (Grant NAGW-2833) is gratefully acknowledged.

272

B.R.T. SIMONEIT

of Guaymas Basin, Gulf of California. Journal of CONNOLLY, J.F. 1966. Solubility of hydrocarbons Geophysical Research, 90, 6695-6707. in water near the critical solution temperatures. KULM, L.D., SUESS, E., MOORE, J.C., CARSON, B., Journal of Chemical Engineering Data, 11, 13-16. LEWIS, B.T., RITGER,S.T., KAOKO,D.C., THORNCOOPER,J.E. & BRAY,E.E. 1963. A postulated role of BURG, T.M., EMBLEY, R.W., RUGH, W.D., fatty acids in petroleum formation. Geochimica et MASSOTH, G.J., LANGSETH, M.G., COCHRANE, Comochimica Acta, 29, 1113-1127. G.R. & SCAMMAN,R.L. 1986. Oregon subduction CURRAY,J.R., MOORE,D.G. & ~7 OTHERS.1982. Initial zone: Venting, fauna and carbonates. Science, Reports of the Deep Sea Drilling Project, 64. US Government Printing Office, Washington, D.C. 231,561-566. KVENVOLDEN, K.A. & SIMONEIT, B.R.T. 1990. CZOCHANSKA, Z., SHEPPARD, C.M., WESTON, R.J., WOOLHOUSE,A.D. & COOK, R.A. 1986. Organic Hydrothermally derived petroleum: Examples geochemistry of sediments in New Zealand, Part from Guaymas Basin, Gulf of California and I. A biomarker study of the petroleum seepage at Escanaba Trough, Northeast Pacific. American Association of Petroleum Geologists Bulletin, 74, the geothermal region of Waiotapu. Geochimica et Cosmochimica Acta, 50,507-515. 223-237. RAPP, J.B. & HOSTETTLER, F.D. 1990. HydroDAVIS, E., MOTrL, M. & SHIPBOARDSCIENTIFICPARTY ~ , 1992. Proceedings of the Ocean Drilling Program, carbon geochemistry of hydrothermallyInitial Reports, 139. Ocean Drilling Program, generated petroleum from Escanaba Trough, College Station, TX. offshore California. In: SIMONEIT, B.R.T. (ed.) DIDYK,B.M. & SIMONEIT,B.R.T. 1989. Hydrothermal Organic Matter Alteration in Hydrothermal Sysoil of Guaymas Basin and implications for tems - Petroleum Generation, Migration and Biogeochem~try. Applied Geochemistry, 5, petroleum formation mechanisms. Nature, 342, 65-69. 83-91. - & - - 1990. Petroleum characteristics of the oil --, MORTON, J.L., KING, J.D. & in a Guaymas Basin hydrothermal chimney. In: CLAYPOOL,G.E. 1986. Petroleum associated with SIMONEIT,B.R.T. (ed.) Organic Matter A Iteration polymetallic sulfide in sediment from Gorda in Hydrothermal Systems - Petroleum GenerRidge. Science, 234, 1231-1234. LEIF, R.N., SIMONHT, B.R.T & KVENVOLDEN,K.A. ation, Migration and Biogeochemistry. Applied Geochemistry, 5, 29-40. 1991. Simulation of hydrothermal petroleum GALIMOV,E.M. & SIMONEIT,B.R.T 1982. Geochemisgeneration by laboratory hydrous pyrolysis. In: MANNING, D.A.C. (ed.) Organic Geochemistry, try of interstitial gases in sedimentary deposits of the Gulf of California, Deep Sea Drilling Project Advances and Applications in the Natural EnLeg 64. In: CURRAY, J.R., MOORE, D.G. et al. vironment. Manchester University Press, Man(eds) Initial Reports of the Deep Sea Drilling chester, 300-303. Project, 64, US Government Printing Office, ,-- & 1992. Hydrous pyrolysis of n-C32H66 Washington, D.C., 781-788. in the presence and absence of inorganic comGIESKES, J.M., SIMONEIT,B.R.T., BROWN,T., SHAW, ponents. American Chemical Society, Division of Fuel Chemistry, 204th National Meeting, PreT., WANG, Y.-C. • MAGENHEIM, A. 1988. Hydrothermal fluids and petroleum in surface prints, 37(4), 1748-1753. sediments of Guaymas Basin, Gulf of California: LONSDALE, P. 1985. A transform continental margin rich in hydrocarbons, Gulf of California. AmeriA case study. The Canadian Mineralogist, 26, 589-602. can Association of Petroleum Geologists Bulletin, HUNT, J.M. 1979. Petroleum Geochemistry and Ge69, 1160--1180. ology. W.H. Freeman and Company. & BECKER, K. 1985. Hydrothermal plumes, hot springs, and conductive heat flow in the Southern JONES, M.L. (ed.) 1985. Hydrothermal vents of the Trough of Guaymas Basin. Earth and Planetary Eastern Pacific: An overview. Bulletin of the Science Letters, 73, 211-225. Biological Society of Washington, 6, 1-566. JOSEPHSON, J. 1982. Supercritical fluids. EnvironLOVE, J.D. & GOOD, J.M. 1970. Hydrocarbons in mental Science and Technology, 16,548A-551A. thermal areas, northwestern Yellowstone National KAWKA, O.E. ~¢. SIMONEIT,B.R.T. 1987. Survey of Park, Wyoming. US Geological Survey Prohydrothermally-generated petroleums from the fessional Paper, 7644-B.' Guaymas Basin spreading center. Organic GeoMEREWETHER, R., OLSSON, M.S. & LONSDALE, P. chemistry, II, 311-328. 1985. Acoustically detected hydrocarbon plumes rising from 2-km depths in Guaymas Basin, Gulf - &~ 1990. Polycyclic aromatic hydrocarbons in hydrothermal petroleums from the Guaymas of California. Journal of Geophysical Research, Basin spreading center. In: SIMONE|T, B.R.T. 90, 3075-3085. (ed.) Organic Matter Alteration in Hydrothermal MICHAELIS,W., JENISCH,A. & RICHNOW,H.H. 1990. Systems - Petroleum Generation, Migration and Hydrothermai petroleum generation in Red Sea Biogeochemistry. Applied Geochemistry, 5, sediments from the Kebrit and Shaban Deeps. In: 17-27. SIMONHT, B.R.T (ed.) Organic Matter Alteration KOSKL R.A., LONSDALE, P.F., SHANKS, W.C., in Hydrothermal Systems - Petroleum GenerBERNDT, M.E. & HOWE, S.S. 1985. Mineralogy ation, Migration and Biogeochemistry. Applied and geochemistry of a sediment-hosted hydroGeochemistry, 5, 103-114. thermal sulfide deposit from the southern trough PAULAITUS, M.E., KRUKONIS,V.J., KURNIK, R.T. &

H Y D R O T H E R M A L PETROLEUM REID, R.C. 1982. Supercritical fluid extraction. Reviews in Chemical Engineering, l, 179-250. PETER, J.M. 1986. Genesis of hydrothermal vent

deposits in the southern trough of Guaymas Basin, Gulf of California: A mineralogical and geochemical study. MSc thesis, University of Toronto. ~,

SIMONEIT,B.R.T., KAWKA,O.E. & ScoTr, S.D. 1990. Liquid hydrocarbon-bearing inclusions in modern hydrothermal chimneys and mounds from the southern trough of Guaymas Basin. In: SIMONErr, B.R.T. (ed.) Organic Matter Alteration

in Hydrothermal Systems - Petroleum Generation, Migration and Biogeochemistry. Applied Geochemistry, 5, 51--63. ~,

PELTONEN,P., SCOTI', S.D., SIMONEIT,B.R.T. & KAWKA, O.E. 1991. Carbon-14 ages of hydrothermal petroleum and carbonate in Guaymas Basin, Gulf of California - implications for oil generation, expulsion and migration. Geology, 19,253--256. PITZER, K.S. 1986. Large-scale fluctuations and the critical behavior of dilute NaCI in H20. Journal of Physical Chemistry, 90, 1502-1504. POLAr, J.T., BAEABAN, M., PEPLOW, A. & PHLIPS, A.J. 1989. Supercritical carbon dioxide extraction of lipids from algae. In: JOHNSTON, K.P. & PENNINGER, J.M.L. (eds), Supercritical Fluid Science and Technology. American Chemical Society, Symposium Series, Washington, DC, 4 0 6 , 4 4 9 - 4 6 6 .

water. Industrial Engineering Chemistry Fundamentals, 25, 169-171. ScoT, S.D. 1985. Seafloor polymetallic sulfide deposits: Modern and ancient. Marine Mining, 5, 191-212. SHAW, R.W., BRILL, T.B., CLIFFORD, A.A., ECKERT, C.E. & FRANCK,E.U. 1991. Supercritical water, a medium for chemistry. Chemical Engineering News, Dec. 23, 26--38. SHOCK, E.L. 1988. Organic acid metastability in sedimentary basins. Geology, 16, 886-890. 1989. Corrections to 'Organic acid metastability in sedimentary basins'. Geology, 17,572-573. 1990. Chemical constraints on the origin of organic compounds in hydrothermal systems.

-

-

Origins of Life and Evolution of the Biosphere, 20, 331-367.

SILFER,J.A., ENGEL,M.H. & MAcro, S.A. 1990. The effect of hydrothermal processes on the distribution and stereochemistry of amino acids in Recent Antarctic sediments. In: SIMONEIT, B. R.T. (ed.) Organic Matter A Iteration in Hydro-

thermal Systems - Petroleum Generation, Migration and Biogeochemistry. Applied Geochemistry, 5, 159-167. SIMONEIT, B.R.T. 1978. The organic chemistry of marine sediments. In: RILEY, J.P. & CHESTER, R. (eds) Chemical Oceanography, 7, Academic -

Press, 233-311. 1983. Organic matter maturation and petroleum genesis: Geothermal versus hydrothermal. In:

-

PRICE, L.C. 1976. Aqueous solubility of petroleum as applied to its origin and primary migration.

The Role of Heat in the Development of Energy and Mineral Resources in the Northern Basin and Range Province. Geothermal Research Council,

American Association of Petroleum Geologists Bulletin, 60, 213-244. , WENGER,L.M., GING, T. & BLOUNT, C.W. 1983. Solubility of crude oil in methane as a function of pressure and temperature. Organic Geochemistry, 4,201-221. RONA, P.A. 1984. Hydrothermal mineralization at seafloor spreading centers. Earth Science Reviews, 20, 1-104. 1988. Hydrothermal mineralization at oceanic ridges. The Canadian Mineralogist, 26, 431-465. ~, THOMPSON, G., Mot"rE, M.J., KARSON, J.A., JENKINS,W.J., GRAHAM,D., MALLETTE,M., VON DAMM, K. & EDMOND, J.M. 1984. Hydrothermai activity at the Trans-Atlantic Geotraverse hydrothermal field, Mid-Atlantic Ridge Crest at 26 ° N. Journal of Geophysical Research, 89, 1136511377. Ross, D.S. 1984. Coal conversion in carbon monoxide - water systems. Coal Science, 3,301-337. ~, HUM, G.P., MIIN, T.-C., GREEN, T.K. & MANSINI, R. 1986. Supercritical water/CO liquefaction and a model for coal conversion. Fuel Process Technology, 12,277-285. SAKAI, H., GAMO, T., KIM, E.-S., TSUTSUMI, M., TANAKA,T., ISHIBASHI,J., WAKITA,H., YAMANO, M. & OOMORI, T. 1990. Venting of carbon dioxide-rich fluid and hydrate formation in midOkinawa Trough backarc basin. Science, 248, 1093-1096. SANDERS, N.D. 1986. Visual observation of the solubility of heavy hydrocarbons in near-critical

273

-

-

-

-

1

9

Special Reports, 13. Davis, California, 215-241. 1985a. Hydrothermal petroleum: Genesis, migration and deposition in Guaymas Basin, Gulf of California. Canadian Journal of Earth Sciences, 22, 1919-1929. Hydrothermal petroleum: Composition and utility as a biogenic carbon source. In: JONES, M.L. (ed.) Hydrothermal Vents of the Eastern 8

5

b

.

Pacific: An Overview. Bulletin of the Biological Society of Washington, 6, 49-56. 1994. Lipid/bitumen maturation by hydrothermal activity in sediments of Middle Valley, ODP Leg 139. In: MOTTL, M., DAVIES, E. et al. (eds)

Proceedings of the Ocean Drilling Program, Scientific Results, 139, in press. GALIMOV, E.M. 1984. Geochemistry of interstitial gases in Quaternary sediments of the Gulf of California. Chemical Geology, 43, 151166. & KVENVOLDEN,K.A. 1994. Comparison of 14C ages of hydrothermal petroleums. Organic Geochemistry, in press. & LONSDALE, P.F. 1982. Hydrothermal petroleum in mineralized mounds at the seabed of Guaymas Basin. Nature, 295, 198-202. --, BRAULT,M. & SALIOT,A. 1990a. Hydrocarbons associated with hydrothermal minerals, vent waters and talus on the East Pacific Rise and Mid-Atlantic Ridge. In: SIMONEIT,B.R.T. (ed.)

-

-

-

-

Organic Matter Alteration

in Hydrothermal

274

B.R.T. SIMONEIT

Systems - Petroleum Generation, Migration and Biogeochemistry. Applied Geochemistry, 5, 115124. , GOODFELLOW,W.D. & FRANKLIN,J.M. 1992b. Hydrothermal petroleum at the seafloor and organic matter alteration in sediments of Middle Valley, Northern Juan de Fuca Ridge. Applied Geochemistry, 7,257-264. - - , GRIMALT, J.O., HAYES, J.M. & HARTMAN,H. 1987. LOWtemperature hydrothermal maturation of organic matter in sediments from the Atlantis II Deep, Red Sea. Geochimica et Cosmochimica Acta, 51,879-894. ~, KAWKA, O.E. & BRAULT, M. 1988. Origin of gases and condensates in the Guaymas Basin hydrothermal system. In: Origins of Methane in the Earth (ed. M. Schoell). Chemical Geology, 71, 169-182. - & WANG, G.-M. 1992a. Biomarker maturation in contemporary hydrothermal systems, alteration of immature organic matter in zero geological time. In: MOLDOWAN,J., PHILP, R.P. & ALBRECHT, P. (eds) Biological Markers in Sediments and Petroleum. Prentice Hall, Englewood Cliffs, NJ, 124-141. --, LONSDALE, P.F., EDMOND, J.M. & SHANKS, W.C., III. 1990b. Deep-water hydrocarbon seeps in Guaymas Basin, Gulf of California. In: SIMONEIT, B.R.T (ed.) Organic Matter Alteration in Hydrothermal Systems - Petroleum Generation, Migration and Biogeochemistry. Applied Geochemistry, 5, 41-49. SISKIN, M. & KATRITZKY,A.R. 1991. Reactivity of organic compounds in hot water: Geochemical and technological implications. Science, 254, 231-237. SPIKER, E.C. • SIMONEIT, B.R.T 1982. Radiocarbon dating of Recent sediments from Leg 64, Gulf of California. In: CURRAY,J.R., MOORE,D.G. et al. (eds) Initial Reports of the Deep Sea Drilling Project. US Government Printing Office, Washington, DC, 757-758. SUESS, E., SIMONErr, B.R.T, WEFER,G., WHITICAR, M.J., FISK, M., TORRES, M., HAN, M.W.,

WI~STOCK, R. & LABAN, C. 1994. Hydrothermalism in the Bransfield Strait, Antarctica. Geologische Rundschau, in preparation. TIERCELIN, J.-J., BOUL~GUE,J. & SIMONEIT, B.R.T. 1993. Hydrocarbon, sulphide and carbonate deposits related to sublacustrine hydrothermal seeps in the North Tanganyika Trough, East African Rift. In: PARNELL, J., KUCHA, H. & LANDAIS, P. (eds) Bitumens in Ore Deposits, Springer Verlag, Berlin, 96-113. - - , THOURIN, C., KALALA,T. & MONDEGEUR, A. 1989. Discovery of sublacustrine hydrothermal activity and associated massive sulfides and hydrocarbons in the north Tanganyika trough, East African Rift. Geology, 17, 1053-1056. TISSOT, B.P. & WELTE, D.H. 1984. Petroleum Formation and Occurrence: A New Approach to Oil and Gas Exploration 2nd ed. Springer Verlag, Berlin. TODHEIDE, K. 1982. Hydrothermal solutions. Berichte der Bunsengesellschaft fiir Physikalische Chemie, 8 6 , 1005-1016. TUNNICLIFFE, V. 1991. The biology of hydrothermal vents: Ecology and evolution. In: BARNES, M. (ed.) Oceanography and Marine Biology Annual Reviews, 29. Aberdeen University Press, 319407. 1992. Hydrothermai-vent communities of the deep sea. American Scientist, 80, 336-349. WELHAN, J.A. & LUPTON, J.E. 1987. Light hydrocarbon gases in Guaymas Basin hydrothermal fluids: Thermogenic versus abiogenic origin. American Association of Petroleum Geologists Bulletin, 71,215-223. WHELAN,J.K., SIMONErr, B.R.T. & TARAEA,M. 1988. C1-C8 hydrocarbons in sediments from Guaymas Basin, Gulf of California - comparison to Peru Margin, Japan Trench and California Borderlands. Organic Geochemistry, 12, 171-194. WHITICAR, M.J., SUESS, E. & WEHNER, H. 1985. Thermogenic hydrocarbons in surface sediments of the Bransfield Strait, Antarctic Peninsula. Nature, 314, 87-90.

Hydrocarbons and other fluids: paragenesis, interactions and exploration potential inferred from petrographic studies JOHN PARNELL

Department of Geology, Queen's University of Belfast, Belfast BT7 INN, UK Abstract: Residues of hydrocarbons (bitumens sensu lato) are found in a wide range of settings in addition to reservoir porosity, including microfractures, megafractures, and several types of vein porosity. Paragenetic relationships between hydrocarbons and inorganic minerals provide information on the relative timing of hydrocarbon migration and the migration of other fluids. Hydrocarbons in reservoir rocks and mineralized fractures are paragenetically late as they are in pre-existing fluid pathways. Some other types of hydrocarbon occurrence are paragenetically earlier, especially proximal to the source rock. Different types of ore mineralization (notably base metal and uranium) in which hydrocarbons occur exhibit varying parageneses which represent distinct burial/thermal histories responsible for the ore types. Co-migration of hydrocarbons and silicon-bearing fluids is evinced by the occurrence of large quantities of authigenic silicate minerals within hydrocarbon residues. Interactions between metals and hydrocarbons may be exploited in the exploration for both metals and hydrocarbons. Metal-rich fossil fuels can preserve a signature of anomalous metal concentrations in the vicinity of ore deposits. Hydrocarbons precipitated by irradiation-solidification processes about radioelement-bearing phases help to identify hydrocarbon migration pathways and in some circumstances allow isotopic dating of migration.

There are numerous records of the paragenesis of hydrocarbons (oil or oil residues; bitumens sensu lato) and inorganic mineral phases, particularly in siliciclastic reservoir rocks and to a lesser extent in mineralized veins. The enormous database of the diagenesis of sandstones prior to hydrocarbon migration, recorded to assess poroperm constraints, is also a record of the composition of the fluid which migrated through the rock before the hydrocarbon fluid. The sequence of hydrocarbons and inorganic phases represents the changing composition of fluids passing through a sampling point, but not necessarily the sequence of fluid generation, as the origins of successive fluid pulses may be at widely different depths. It is reasonable to expect that parageneses recorded proximal to the hydrocarbon source rock would include hydrocarbons at a relatively early stage compared to parageneses in a more distal setting. As an introduction to this account I review the paragenesis of hydrocarbon and other fluids in a wide variety of settings (Fig. 1), including intraformational veinlets within the source rock, deposits in overpressured systems along bedding planes, reservoir/migration pathways at proximal and distal points and involving interaction with metalliferous groundwaters, hydrothermal systems which pass through the source rock, large bitumen veins which usually originate close

to the source rock, and faults (generally thrusts) in which hydrocarbon lubrication may have facilitated displacement. The exploration potential of hydrocarbon-inorganic interactions is subsequently reviewed, particularly in the light of the paragenetic relationships between hydrocarbons and other (mostly metalliferous) fluids.

Paragenetic context of hydrocarbons in varied settings

Fracture-hosted bitumens Solid veins of bitumen (A in Fig. 1) occur in successions which have very high hydrocarbon source rock potential (high TOC, and hydrogen indices often above 500 mg g-a) and rapid burial rates, causing the generation of substantial volumes of hydrocarbon over a geologically short time period (Monson & Parnell 1992). The veins are thought to be a consequence of this hydrocarbon generation, which is in excess of that which can migrate away through intergranular pathways. This causes a pressure build-up which is ultimately released by fracturing and migration through vein systems. As hydrocarbons are the cause of the fractures, they are in most cases the first fluids to enter the fractures: solid bitumen veins generally show

FromPARNELL,J. (ed.), 1994, Geofluids:Origin, Migrationand Evolution of Fluidsin SedimentaryBasins, 275 Geological Society Special Publication No. 78,275-291.

276

J. PARNELL

~bA

..... . E ~reservoJr ~ F .... \~,,~ JJ rock

Itumen

,11

ein

/

/

groundwater

E SOURCE

hydrothermal fluid circulation

B

intraformational bitumen

G

~.

C

overpressuring

/

Fig. 1. Schematic diagram showing settings in which paragenesis of hydrocarbons (bitumens) and other phases is discussed (letter codes in text). clean contacts with their wallrocks. However in some examples there is evidence for a very early, and volumetrically limited, stage of silica precipitation; quartz crystals lining the fractures or even suspended (coeval) in the bitumen. Examples of early authigenic quartz suspended in bitumen from Utah (Monson & Parnell 1992), Oklahoma (Monson pers. comm.) and Scotland yield fluid inclusion homogenization temperatures of about 100°C. Generally, these large bitumen veins are very pure as they have not functioned as pathways for subsequent fluids. Exceptions are some carbonate-rich successions where the bitumen has been followed and vigorously shattered/displaced by calcite precipitation; examples are seen in veins in the Neuquen basin, Argentina and at Bentheim, Germany. Vein bitumens also have very limited metal contents, excepting vanadium and nickel contents which may be high due to inheritance of metal in organic molecules from the source rock. Examples are amongst the vein bitumens in the Ouachita Mountains, Oklahoma, which were formerly mined for bitumen and as a source of vanadium (Ham 1956). Other metals are not enriched within the bitumens because of lack of opportunity for them to be introduced by other fluids. Many sequences which contain source rocks contain minor intraformational fracture-fillings of bitumen (B in Fig. 1), which are generally pure or associated with calcite. The high pressures developed during excessive hydrocarbon generation may allow the pooling of hydrocarbons along bedding planes as their fluid pressure supports overburden (overpressuring, C in Fig. 1) and holds bedding surfaces apart. In such instances, pools cause source rocks to function also as reservoir rocks, as in the West Siberian Basin (Peterson & Clarke 1991). Upon

uplift, these hydrocarbon pools result in bedding-parallel bitumen veins. As in the large bitumen veins, hydrocarbons were generally the first, causative, fluid, or sometimes coeval with/followed shortly by carbonate precipitation. The calcite in these veinlets commonly exhibits spherular interfaces with the bitumen, indicating coeval precipitation and poisoning of crystal growth by the bitumen. In other cases, the carbonate can be fibrous (Fig. 2), a petrographic feature interpreted as additional evidence for overpressuring (Stoneley 1983). The generation of hydrocarbons, with or without the development of overpressuring, may cause lubrication of the source and other rocks along migration pathways. Lubrication allows structural deformation, including thrusting (D in Fig. 1) in compressional regimes. Low-angle veins are commonplace in thrusted rocks (de Roo & Weber 1992), particularly rocks which are finely laminated as are many hydrocarbon source rocks. The veins contain traces of the hydrocarbons, and some veins along thrust planes are predominantly bitumen (Fig. 3). Hydrocarbons may also enter the thrust planes at a later stage, either during deformation (e.g. Roberts 1991) or after the cessation of deformation in which case they would not have a causative role. Small- (centimetre-) scale bitumen veinlets along decollement surfaces are also often pure, or a mixture of bitumen and calcite. Again the hydrocarbons, having a causative role, are an early component of veins, predating later carbonates and quartz.

Reservoir bitumens Hydrocarbons entering reservoir rocks (E in Fig. 1) displace aqueous pore fluids, which themselves may have been subject to long-term

PETROGRAPHIC STUDIES

~:~+ } + ~ +:i:e; ; +++' + +++ ~+:++ ; +~'+++++ C,+~++ +i ~ +++~' ....>

~'~+II~..+ +

_~'+'

.... '++++ .... +~<+:+-r:,++; ++: + ~

: . . . . + + ++, + + . +'+ ~ .,+~." +~+~.+,.,+ .... ..... + ~ , ++ +~ ++ .,++ + ~. .+: . +++++ + , ' : ,+~+ ,+ ' + ,,+-+:m+ +. . ++ .

++. . . . . . . . . . . . . . .

277

+~+~,.+"....

~+"e++++~ + + ~ ~. ~ . ~ + ¢ ,+ + +++ ,.++

++~ +++-. +,:,~. . . .+++~ . +.+...... +~++ +.++++++++~+ ++:.+~++++++~+++r+:+: +:~: . . . . + ++ +++++ ++rk+~+. :~,:++++++++.+:++ ++:+++:+ +~,:+.~ ++~++~+. ~......... ~++::-+ .++++ + ......... ++ ..... • . . . . +~+::+++ .

F i g . 2 . Dolomitic laminite with bedding-parallel seam of bitumen (black) associated with fibrous dolomite• Thin section photomicrograph; field width 3 mm; Carboniferous, Granton, Scotland•

m

j

BITUMEN

1500

1000

C,,

0

500

~PPep

-

"

1000 m

Fig. 3. Cross-section in SE Turkey showing bitumen veins related to fault planes showing both reverse and normal throws (after Lebkuchner et al. 1972).

chemical evolution. T h e h y d r o c a r b o n s m a y t h e r e f o r e post-date a succession of authigenic m i n e r a l phases; the study of which c o m m a n d s m u c h attention because of the c o n s e q u e n c e s for porosity and permeability. W h e r e the hydrocarbons o c c u p y s e c o n d a r y pores, these pores m a y be relatively y o u n g if they w e r e f o r m e d by a decarboxylation-decarbonatization mechanism just prior to h y d r o c a r b o n g e n e r a t i o n . C o m m o n l y , h y d r o c a r b o n s are r e s e r v o i r e d at shall o w e r levels than their source, at a level w h e r e porosities and the w a t e r c o n t e n t of the rocks are

higher. W h e r e reservoir rocks are i n t e r b e d d e d with source rocks, g r e a t e r burial depths generally m e a n that m o r e w a t e r has b e e n expelled, porosities are l o w e r and the h y d r o c a r b o n s m a y be even later in the p a r a g e n e t i c s e q u e n c e . It is well k n o w n that oil or gas e m p l a c e m e n t into a reservoir can 'freeze' d i a g e n e t i c processes by excluding the w a t e r n e e d e d to t r a n s p o r t ions. C o m p a r i s o n of diagenetic evolution in c o m p a r able oil-bearing and w a t e r - b e a r i n g sandstones can t h e r e f o r e place oil migration within a diagenetic s e q u e n c e .

J . PARNELL depth

(m) 500

,,

S

-

O

3000m !

- - - - palaeo-GWC Illite-bearing sandstone

- - -- present GWC [Illite-free sandstone

Fig. 4. Section through South Morecambe gas field, showing present gas-water contact below palaeo-gaswater contact (GWC) represented by limit of occurrence of platy illite (after Macchi et al. 1990). MMG, Mercia Mudstone Group; KWS, Keuper WaterstonesIKeuper Sandstones; SBS, St. Bees Sandstone.

A consequence of the inhibition of diagenesis in a hydrocarbon-charged reservoir is that the distribution of diagenetic phases can be used to reconstruct the distribution and attitude of palaeo-reservoirs, i.e. former hydrocarbonwater contacts can be inferred from boundaries between absence/presence of diagenetic phases. In particular, the distribution of authigenic clay phases is informative, as has been shown in the Triassic sandstones of the Irish Sea basin (Bushell 1986; Woodward & Curtis 1987; Macchi et at. 1990). Sandstones of the Keuper sandstones and Waterstones (Shenvood Sandstone Group) hold gas below a Mercia Mudstone Group seal (Fig. 4). Above the present gaswater contact, a palaeo-gas-water contact, now slightly tilted, can be inferred from a boundary between illite-free sandstone above illitebearing sandstone (Fig. 4). The illite-free sandstone represents the gas column present during illitization, as the gas excluded the ion flux necessary for illitization. The recognition that illite development is suspended by reservoired hydrocarbons offers the possibility of constraining the timing of reservoir charging through radiometric dating of the illite. Authigenic illites in reservoir systems are routinely dated by K-Ar and Ar-Ar isotopic methods. An interesting example from the Huldra Field, North Sea yielded two dates from different depths (Glasmann et af. 1989). The younger date was from a lower depth, indicating that illite growth had continued for longer at that depth until charging of the reservoir brought the oil-water contact down to the lower depth (Fig. 5 ) , i .e. the data indicate progressive filling of the reservoir. In order to assess the relative timing of hydrocarbon migration through carrier beds, where diagenesis is not suspended, it is possible

to examine the small quantities of bitumen 'fixed' onto detritaVauthigenic radioelementbearing minerals, particulary thorite (see below), and relate them to the diagenetic sequence for the host rock. In most cases where this has been attempted (e.g. in Carboniferous, North West Irish Basin: Parnell & Monson 1990), the bitumen is relatively late in the diagenetic sequence, post-dating quartz and illite precipitation. This type of bitumen does not occur in thermally immature basins: in basins where the bitumen is lacking due to immaturity, the host rocks exhibit substantial quartz overgrowth, clay precipitation, carbonate cementation and feldspar dissolution. Where sandstone-hosted ore deposits contain traces of bitumen, they are generally assumed to be a product of an interaction between reservoired hydrocarbons and metalliferous groundwaters (G in Fig. 1). The ores involved are particularly copper, uranium and vanadium, i.e. those metals mobile in oxidising conditions, and liable to precipitation by reductant hydrocarbons. Microprobe studies show that the metalliferous minerals are paragenetically late, and have often precipitated within the intergranular bitumens of reservoirs, nucleated on grainbitumen interfaces (Fig. 6, from the Heumul uranium deposit, Argentina, where uranium from volcanics precipitated in Cretaceous hosted oil reservoirs; Ferreyra & Lardone 1990). In these cases, metalliferous fluids clearly were able to penetrate into the oil in the reservoir.

Bitumens in ore deposits Bitumens are found in a wide range of other ore deposits (G in Fig. 1) in addition to mineralized oil reservoirs, and have been used as evidence for an organic role in ore genesis and a source of

PETROGRAPHIC STUDIES

279

depth (m)

///~

3680-

///,,~

~

3700"

oil-water contact 1

"///

illite 1

. . " ......... 37;~0" .....................

1 58Ma

3740///i

....

oil-water contact 2

///>

illite 2

3760 ..................................................... • 38Ma 3780

;-'-".~. "//A

/IA

": / /I //l, , I

://2

3800

///1

//A

8'o

4'0

K/Ar age (Ma)

Fig. 5. Plot of illite (< 0.1 ixm)-measured K-Ar ages against depth for two samples from the Huldra field, North Sea, representing progressive filling of reservoir (adapted from Glasmann et al. 1989).

Fig. 6. Backscattered electron micrograph of bitumen (black)-bearing sandstone, with uraninite (bright) precipitated at grain-bitumen interfaces; Huemul uranium deposit, Argentina. Field width 500 ~m. information about the geochemical/thermal conditions of mineralization (Gize 1993). Petrographic studies help to understand the paragenesis of bitumens and ore minerals which

varies between deposits. Some deposits contain bitumens which are coeval with the main ore stage, some contain bitumens which predate mineralization, and others contain late-stage bitumens (Fig. 7). It is important to take paragenesis into account when making inferences about the role/significance of organic matter. In some cases, late-stage bitumens may be completely unrelated to the mineralization process and simply utilised the same migration pathway (Parnell 1991). Pre-mineralization bitumens are the least commonly encountered of the possible parageneses, but are fundamental where the metals have mineralized a pre-existing oil reservoir. In many cases, hydrocarbons post-date the main ore stage; this is the case in many hydrothermal base metal deposits, including carbonate-hosted lead-zinc deposits (e.g. Jones & Brand 1986; Marikos e t al. 1986). However in some examples of Mississippi Valley Type deposits, hydrocarbons are found within fluid inclusions in sphalerite, indicating that they were present during mineralization. Bitumens are coeval with ore minerals in cases where mineralization is a consequence of the mixing of hydrocarbons and metalliferous fluids, as in the case of hydrocarbons leaking from a reservoir into a metalliferous groundwater system (see below). Several uranium deposits contain apparently coeval uraninite and bitumen. Epithermal mercury ores also commonly show coeval cinnabar and bitumen, an association which partly reflects their precipitation under the same (low) temperature conditions (Peabody 1993). The close paragenetic relationship

280

J. PARNELL UBi. SCOTLAND

Time

Quartz Hematite

I

I

I

i

l

i

nimimimmi i

I

i

i

Uraninite Dolomite HYDROCARBON Chalcopyrite Bismuth

i I

I l

I i

tieR i

i |

l i

l l

|

| I

|

J I

I

i

$bAu. NEVADA HYDROCARBON Quartz + Au

I

I

Kaolinite Stibnite Sulphates

Hematite

PbZn. IRELAND Calcite

Pyrite Sphalerite

I

l l

l

I

l

~

l

I

l

•

~

l

I

•

Galena

Barite HYDROCARBON

Fig. 7. Examples of parageneses for ore deposits showing variable context of bitumens (hydrocarbons) which may be syn-ore (Southwick, Scotland, after Miller & Taylor 1966), pre-ore (Alligator Ridge, Nevada, after Ilchik et al. 1986) or post-ore (Ballinalack, Ireland, after Jones & Brand 1986). between cinnabar and bitumen has led to the use of bitumen in exploration for mercury (Vershkovskaya et al. 1972).

Continental scale migration On a continental scale, several workers have drawn attention to the possible roles in mineralization and hydrocarbon migration of fluids expelled from compressive continental margins. Fluids migrating towards the foreland basin and continental interior due to either squeezing out beneath thrust sheets (Oliver 1986) or the topographical relief of the orogen (Bethke & Marshak 1990; Bethke etal. 1991) are likely to be greatly modified through interaction with rocks in those regions. Topographical relief seems the more likely driving mechanism, as compression and thrusting would cause fluid flow too slow to explain the inferred flow rates (Ge & Garven 1989). Heat transferred with the fluid from the orogenic zone could mobilise further inorganic and organic components from the foreland basin. The temporal evolution of these fluids is a matter for conjecture: the heat

flow in the orogenic zone would enhance the mobility of metals and hydrocarbons over a wide area and over a long time period. Particular attention has been paid to the role of fluids from the Ouachita orogen, southern USA. These fluids have been invoked as northwards carriers of oil to Kansas (Oliver 1986) and the metals to Missouri, Arkansas and Oklahoma (Leach et al. 1984; Leach & Rowan 1986). Accounts of Mississippi Valley Type lead-zinc deposits in the central USA including the Tri-state area indicate that bitumens are commonly present within them. Where paragenesis is determinable, the bitumen is often later than the main phase of ore mineralization (e.g. Marikos et al. 1986; Niewendorp & Clendenin 1993), as discussed above, although in some instances there are also hydrocarbon-bearing fluid inclusions which evince an earlier presence of hydrocarbons.

Authigenic minerals in bitumens Recent studies of bitumens in fracture systems show that they can contain significant quantities of authigenic mineral matter, including illitic

PETROGRAPHIC STUDIES

281

mobilization or immobilization of metals. These include physical adsorption of metals onto organic materials, chemisorption of metals into organic materials, precipitation of organometallic compounds by reaction of metals with organic ligands, mobilization/redeposition of metals by formation/destabilization of organometallic complexes, and change in oxidation state including reduction of metals by organic materials and consequent mineral precipitation (Saxby 1976). Some interactions have potential in the exploration for either hydrocarbons or metals.

Hydrocarbon exploration

Fig. 8. Backscattered electron micrograph of fracture-hosted bitumen containing authigenic crystals of clay (elongate) and pyrite framboids, Carboniferous, NW Ireland. Field width 500 la,m.

clays, feldspars and pyrite framboids, which appear to have precipitated in situ within the bitumens (Parnell 1991). The minerals are assumed to be authigenic because some are too delicate to have survived transport, they are suspended in an equidistant manner rather than clotted together, and they are mostly monomineralic and euhedral (Fig. 8). This implies that the bitumens were deposited from fluids with a substantial aqueous component to transport the necessary ions for the mineral phases. An implication is that in some circumstances hydrocarbons in reservoir rocks might also include authigenic minerals not related to the host rock, but this has not been demonstrated conclusively. Bitumens in base metal and uranium-bearing ore deposits additionally contain authigenic inclusions of ore minerals, particularly minerals of the ore metals of the host deposit, which tend to be enriched in the bitumens as noted above.

Metal-organic interactions and exploration potential Several types of interaction between dissolved metals and organic materials could result in the

Metalliferous organic material is not only created by metal precipitation within a preexisting oil reservoir: it may be a product of an interaction between two migrating fluids, one metal-bearing and the other hydrocarbonbearing. An important cause of fluid mixing in sedimentary basins is the meeting of gravitydriven meteoric waters/shallow groundwaters which are moving down-dip with deeper groundwaters being expelled up-dip by burial-compaction. Such mixing can bring together metalbearing fluids derived from surface waters flowing off hinterland basement rocks and hydrocarbons migrating from source rocks in the deeper parts of a basin. Precipitation of metals with or without accompanying organic residues is most likely to be a product of redox processes. Many such interactions do leave metalliferous organic materials (e.g. Roberts 1980; Parnell 1988; Disnar & Sureau 1990). Consequently, there is potential to use these materials as tracers of either metal migration or hydrocarbon migration. The precipitation of metals by redox processes is particularly likely where migrating hydrocarbons interact with metalliferous groundwaters in red beds, as the potential for large changes in Eh is high. Interactions in red beds could follow the roll-front model which involves distinct populations of bacteria on either side of an ore-precipitating Eh front (e.g. Rackley 1976). Alternatively hydrocarbons could enter the groundwater pathway by leakage along fractures from depth, either directly from the hydrocarbon source rock or from an intermediate reservoir (Roberts 1980). A distinctive product from such interactions in red beds is the metalliferous organic cores found in some reduction spheroids. These are generally on a millimetre-scale, and consist of a spheroidal bituminous mass rich in metals and incorporating, by partial replacement, the detrital grains

282

J. PARNELL

~

.,~ ~ !

+ +

+

+

+

+

+

4-~

~

,

~

~

-

~

~ gneiss "~

'

"

~

J

,

*

veins

(bitumen

in fau.s) ~...p

~

~

. = j ~ ~ I r - . ; . ~ d l

' ' '

~,

mudrock v

~

'

~

^

=

~ =~=---~--(oilsource) ~ ~

.

A

breccia

Fig. 9. Schematic drawing showing interaction between uraniferous groundwaters derived from granitic basement and hydrocarbons leaking along faults from sandstone reservoir, to form uraniferous bitumen nodules. Based on Devonian, northern Scotland, over horizontal distance 15 km. of the country rock. The metals concentrated in the spheroid cores are those which are typical of red bed ore deposits and which are mobile in oxidizing conditions (uranium, vanadium, copper) and additionally precious metals including gold, silver and platinum group elements (Hofmann 1990, 1991). Not all of the metalliferous cores are organic-rich: those which are tend to be uraniferous, while vanadium concentrations are more typical of cores which lack organic matter. The spheroids represent sites in red beds where iron oxides were dissolved in the vicinity of reductant (organic matter) point sources. The origin of the metalliferous cores is attributed to bacterially-enhanced low-temperature sulphate reduction (Hofmann 1990) to precipitate sulphides along with other metalliferous phases. The organic-rich cores can be a useful indication of the leakage of hydrocarbons from depth. In a small-scale test study, the distribution of such organic-bearing uraniferous spheroids in Devonian red beds in northern Scotland was used to predict reservoired hydrocarbons lower in the section, the existence of which was subsequently confirmed by detailed fieldwork (Parnell & Eakin 1987). The rocks beneath the red beds and above the reservoir sandstones exhibit brittle faults, some of which are infilled with solid bitumen, indicating leakage of hydrocarbons from the reservoir into the metalliferous red beds (Fig. 9). Another example of interaction between metalliferous fluids and hydrocarbon fluids is the precipitation of thoriferous bitumen nodules. This is a recently recognized phenomenon but

.

.

.

.

.

.

Fig. 10. Backscattered electron micrograph of bitumen nodule (black) containing thorite inclusions (bright), Carboniferous Rough Rock Group, Ramsbottom, Lancashire. Field width 800 microns.

one which appears to be widespread in basins where there are plutonic rocks in the hinterland to supply the thorium (Parnell et al. 1990). The nodules are sub-millimetre size and consist of

PETROGRAPHIC STUDIES Distribution of S a m p l i n g Areas and T h o r i f e r o u s B i t u m e n Nodule Localities in British and Irish Carboniferous

283

Lower Carboniferous

~_,o ~ ~b-"

Upper Carboniferous

~,

Thoriferous bitumen nodule locality

•

Barren sample area

N

0

CMV

D

S 0

0

•

km t

200 I

Fig. 11. Distribution of thoriferous bitumen nodules recorded during a reconnaissance sampling programme in Carboniferous sandstones of the British Isles. AC, Ayrshire Coalfield: CMV, central Midland Valley; EM, East Midlands; NWlB, North West Irish Basin; S, Solway.

bitumen with inclusions of thorite/thorianite (Fig. 10). The interaction is an example of the irradiation-induced processes which cause bitumen precipitation (see below), and similarly forms bitumen coatings around detrital grains of thoriferous monazite (Rasmussen et al. 1989) and uraniferous zircon (McKirdy & Kantsler 1980). Due to mineral fragmentation/ displacement processes within the nodules it is not always clear if the mineral inclusions were detrital or authigenic. In some rocks these bitumen nodules/coatings may be the only evidence for hydrocarbon

migration through the rocks, and they are therefore valuable sources of information. Despite the alteration to hydrocarbon composition which is caused by radiation damage from the thoriferous minerals, it may still be possible to relate them to conventional shows of hydrocarbon elsewhere using organic geochemistry. Rasmussen et al. (1989) have demonstrated a relationship between bitumen coatings around monazite grains and crude oil in the Perth Basin, Western Australia, using gas chromatography. The thoriferous bitumen nodules have only been recorded in pale-coloured sandstones,

284

J. PARNELL 6'0

'

8b

'

oo

'

-40

2b

~1~_~O'er" ~ . ~ r - , , ~ t , ~ -

0.,.~ ilackburn

.

'~

40-

I , ,L,~s

""

,,. ,Ib a,-., .@

O0

-

1

O, /v

>10

20-

~,'~- Hudder sfield

\

Q / F ~
40

'

" "-'~ 6 oo-

"~ ,Or) `

~

\

~"~/~,~ J.Sheffield -

80_0

nodules present

•

nodules absent

-~o

, Q,~

"'~

~.Vl \

_

,,~

Stoke

on Trent •

( •

60-

~

-

O"

~

-40

..400

F

10

20

30

I

I

I

1Derby

km

D

6t0

I

8t0

I

00

I

210

i~

4j0

Fig. 12. Outcrop (black) of Upper Carboniferous Rough Rock Group sandstones in northern England and occurrences of thoriferous bitumen nodules at sample sites. All nodule occurrences are in region of low quartz/feldspar (Q/F) ratios, suggesting importance of mineralogically immature host sediments. (Basemap and Q/F ratios from Bristow 1988).

while uraniferous bitumen cores in reduction spheroids only occur in red beds. This host rock control and the mutually exclusive occurrence of the two types is readily explained by the Eh conditions which favour either thorium or uranium mobility, depending on reducing or oxidising conditions respectively. As the thoriferous nodules clearly necessitate passage of hydrocarbons through the rock their distribution can be used to infer hydrocarbon migration pathways and, at the most basic level,

show that hydrocarbons have been generated. A reconnaissance study of the distribution of nodules was undertaken in Carboniferous basins in the British Isles, chosen because the Carboniferous includes hydrocarbon source rocks and the basins developed on a surface which included several Caledonian phtons. The survey showed that the nodules occur in many basins (Fig. 11), including those for which evidence of hydrocarbon migration was already established (NW Irish basin, central Midland Valley,

PETROGRAPHIC STUDIES

285

21 2°Tpb/2°4pb

19

18 ~ , ~ " ~ G 2

17

~

~

T

G

Age = 256+23Ma

3

16 .~TG6~GTG5 TG4 15 ~, 18

2°6Pb/ 2°4Pb,.t 28

3'8

48

5'8

6'8

7'8

8'8

9'8

168

118

Fig. 13. Cross plot of lead isotope ratios for six uraniferous bitumen samples from Ty Gwyn (TG) copper deposit (from Parnell & Swainbank 1990 with additional data). Age determined from slope of line of least squares fit is interpreted as time of hydrocarbon migration/interaction with uranium. Solway, East Midlands), but also where hydrocarbons had not been discovered (Ayrshire Coalfield). Work undertaken on sandstones at a particular stratigraphic level show that the distribution of nodules is dependent upon the provenance/mineralogy of the host sandstones. Samples from the Rough Rock Group (upper Namurian) in northern England yield thoriferous bitumen nodules from most sites, except where the mineralogical maturity of the sandstones is high (Fig. 12), as evinced from the quartz/feldspar ratios of Bristow (1988). This suggests that the availability of thorium is controlled either by the detrital component in the rocks or by variations in the sand provenance which are reflected by differences in the maturity levels. Further work will help to resolve this matter.

Dating of hydrocarbon migration One of the most commonly encountered interactions between hydrocarbons and metals is the precipitation of bitumens rich in uranium, and less commonly thorium, through the effects of irradiation upon migrating fluid hydrocarbons. Radiation promotes condensation and polymerisation reactions, which enhance solidification, such that solid bitumens may be precipitated around radioelement-bearing sources in settings where other traces of hydrocarbons are lacking. The mechanisms involved in these processes are adequately discussed elsewhere (Schidlowski 1981; Willingham et al. 1985). The interactions commonly involve uptake of uranium or thorium (possibly mobilized from detrital grains) by the solid bitumen, to a substantial degree, and

subsequent precipitation of radioelement minerals, particularly uraninite and thorite. Assuming that this mineral precipitation takes place quite rapidly after interaction (evidence for this includes disintegration and displacement of the mineral phase by polymer expansion in the bitumen), dating of the mineral phase would also date the migrating hydrocarbons. The radioactive decay of the radioelements offers some possibility of this dating, although in many cases migration of the daughter lead has upset the isotopic equilibrium (e.g. Aberg et al. 1985). This is particularly likely in the case of thorium minerals which are more likely to be metamict and allow element diffusion. A method which avoids the problem of differential migration of lead and uranium is Pb-Pb dating. The ratios between different lead isotopes should not be significantly affected even if some lead has migrated. The method depends upon the determination of lead isotope compositions in a set of uraniferous bitumen samples in which there is a variable mixture of radiogenic lead and common (background) lead. Different contents of uraninite inclusions in the samples give correspondingly different proportions of radiogenic lead. The variable mixtures allow determination of an age for the precipitation of the uraninite. A 2°Tpb/2°6pb age is given by the slope of the line on the lead isotope cross-plot: the example (Fig. 13) from a bitumen-bearing copper ore deposit, North Wales, yielded an earliest Triassic age (from Parnell & Swainbank 1990, with additional data), reasonably consistent with the probable Triassic commencement of hydrocarbon generation in an adjacent gas field (Bushell 1986).

286

J. PARNELL Metal exploration

a~at)/2aaU 0.70' I °"6°1

1

o so

-

S

0.40

0.30-

/

~

0.20-

//

~

uDD~ intercept = 228r'"/_,,uo

o.Io- . , , ~ ~//

0

2267Ma

760Ma

210

+ /_~Ma _ =760+*'

• == Lower intercept

4:0

6:0

I ! d.o 10.0 12.0

2°rpb/~U

Fig. 14. U-Pb concordia plot for uraniferous carbon data from Witwatersrand gold deposits, South Africa (from Allsopp et al. 1986; Robb et al. 1993), in which upper intercept age of 2267 Ma is interpreted as time of hydrocarbon migration/interaction with uranium.

Success has also been claimed using U-Pb isotopic analyses of uraniferous bitumens. Data from the uraniferous carbon of the Witwatersrand gold deposits yields a 2°6pb/238U versus 2°7pb/235U concordia plot (Fig. 14, from Allsopp et al. 1986; Robb et aI. 1993) with an upper intercept age of 2267 Ma. The authors interpret the age to represent the time that the U-Pb isotopic system was set in the carbon, which may reflect the timing of hydrocarbon migration in the Witwatersrand basin. In some circumstances, U-Pb dating may also be possible by a chemical age dating technique, in which the ratio of lead and uranium contents in uraninite, determined by microprobe analyses, can be directly converted into the time since uraninite precipitation (Bowles 1990). This approach has been tested on uraninite in uraniferous bitumen (Parnell 1993a), but will only work in young sequences (in which the contribution from thorium can be ignored) where no element migration has occurred. In the case of the Witwatersrand carbon, which yields an age of about 2300Ma (Fig. 14), the chemical ages are younger, in the range 1080-1750Ma (Bowles 1990), which are less likely ages for hydrocarbon migration.

Regardless of the paragenesis of hydrocarbons and metalliferous minerals, the bitumens in ore deposits are commonly enriched in metals, particularly the metals which form the ore deposit. This suggests that absorption of metal can take place after bitumen deposition, and that residual/remobilized metal is available to late-stage hydrocarbons. The Lower Carboniferous of the British Isles, which is both widely mineralized by limestone-hosted sulphides and includes hydrocarbon source rocks, was the subject of a pilot survey of metal-rich bitumens in ore deposits (Parnell 1992, 1993b). The survey showed that there is some correlation between the metals present as inclusions within the bitumens and the metals present in the related host ore deposits, and so suggested that the metal chemistry of bitumens may be of value in ore exploration. However, there would be no exploration value in metal-rich organics if it were necessary to sample them only from within ore deposits. More detailed studies were undertaken around two limestone-hosted base metal sulphide deposits in the Lower Carboniferous of Ireland to determine if bitumens exhibit regional anomalies of metal concentration which could be detected in an exploration sampling programme. The full results remain confidential at present, but the study did show that regional anomalies do exist for some elements, and that in comparison with published studies of whole rock metal anomalies around Carboniferoushosted lead-zinc ore deposits, the bitumen metal anomalies are evident over several times the distance from the deposit. Figure 15 shows an example of enrichment factors for barium in bitumen samples around the Tynagh lead-zincbarite deposit, relative to a threshold anomaly value determined from analysis of 'background' bitumen samples distant from the deposit. In this case, bitumens were sampled from boreholes up to 4 km from the deposit. Some elements are still in anomalous quantities in bitumen, at this distance, while whole rock values in the host Carboniferous limestones are at background levels at only 1 km distance (Clifford etal. 1988). This approach has been used to a limited extent on a commercial basis in support of exploration. The fact that bitumens in lead-zinc deposits are enriched in metals, even where the paragenetic relationship in a deposit shows the bitumens to post-date the ore minerals, suggests that metalliferous fluid is still available during hydrocarbon migration. On a larger scale, regional metal anomalies in fossil fuels can also be a signature of some types

PETROGRAPHIC STUDIES

287

(~)0.9 ~ __

~-

(

/

f

--

i

--'~

1.6

(~)3.1

08

13.6 ........... L..~y/(~, , ,--.(~) ~ . . . . ore pit ,, ) 1 7 . 9 k~ c~, ............ .-

9.9~ \ 20 4 "~'i.':6.....

Ranamacken

~E) 2 . 8

11.0

4 6 1 _ ~

--,..

110 1.1

~

~

_.i

I" i ® 'lO ~

Killimor

II

®

2.5

0.6

® 1.8 •

0 [

Tynagh

km ,

2 l

Fig. 15. Enrichment factors for barium in bitumens from borehole samples around the Tynagh lead-zinc ore deposit, Ireland. Values are multiples of one and a half times a mean background value; values averaged from several samples in most boreholes. Contour line represents limit of enrichment factor of 10.

Trinidad

/'<

"

",

Venezuela

""

Columbia i

f'",, ," ;-GOre

deposits

•

Coal rich in V, Cu

•

Oil rich in V, Ni

'

N

;

0

300km /

Fig. 16. Distribution of vanadium/nickel-rich petroleum and vanadium/copper-rich coal in basinal regions, relative to vanadium/nickel/copper ore deposits in basin hinterlands, Venezuela (after Kapo 1978).

of ore deposit in the hinterland rocks. One of the largest regional anomalies is the high level of vanadium and nickel in oils and bitumens in much of South America, including Venezuela (Fig. 16). Kapo (1978) suggests that groundwaters flowing through ore deposits of vanadium, nickel and copper transported the metals and reprecipitated them in organic-rich sediments, including coals and oil source rocks.

The result is coals rich in vanadium and copper, and oil rich in vanadium and nickel (Fig. 16). Early migration of hydrocarbons along mineralized fracture systems helps to explain the occurrence of trace but anomalous quantities of hydrocarbon gases in such systems. The detection of these anomalies (either anomalous quantities or anomalous ratios between components) has been proposed, and attempted on a

288

J. PARNELL

Carboniferous ©

basement

[ 0I

C ~ orebody

1I

2I

3km J

~ g a s >25ppb ~ g a s 15-25ppb [----~gas<15ppb

Fig. 17. Gas anomalies in the immediate vicinity of Silvermines lead-zinc ore bodies (after Carter & Cazalet 1984 and Thermasearch promotional literature).

commercial basis, as an exploration aid (Carter & Cazalet 1984; Carter et al. 1988, Ferguson 1988). For example, determinations of light hydrocarbon gas levels at 155 sites around the Silvermines lead-zinc deposit, Ireland, show a marked association of maxima in the immediate vicinity of the orebodies and an adjacent basin-bounding fault (Fig. 17, Carter & Cazalet 1984). The gas anomalies are believed to be a consequence of reactions during the mineralization process or maturation of organic matter in the country rock by enhanced heat flow. Studies of minerals in the Pennine lead-zinc deposits showed that they contain fluid inclusions rich in light hydrocarbon gases, derived from fluids similar to oilfield brines; consistent with an origin for the ore deposits from basinal brines (Ferguson 1988). This indicates that the gases were present and trapped during the mineralization process. The high level of mercury in some surficial gas/water samples has been used for some years as an indication of hydrocarbon distribution. Samples of gas, oil, coal and oil shale themselves can also contain high mercury concentrations. The association is attributed to a common dependency of hydrocarbons and mercury upon high geochemical gradients, high fluid fluxes and structural focussing of fluids (Ozerova et al.

1992; Peabody 1993). Work by Russian geologists in particular has suggested that in addition to a hydrocarbon exploration tool the association can be used to explore for mercury ore deposits.

Summary In general, bitumens in mineralized fracture systems tend to be paragenetically late because they are taking advantage of pre-existing fluid pathways. Similarly, bitumens in reservoir rocks, even where proximal to the source rock, are paragenetically late because the pores have had a long history of fluid evolution before hydrocarbon migration. Most other cases, in large bitumen veins, bedding-parallel overpressure veins, bituminous thrust surfaces and intraformational veinlets, are generally proximal to the source rock: hydrocarbon migration has a causative role in fracture formation in many cases and is paragenetically early. Coeval carbonate and other authigenic mineral phases are commonly present in bitumen, indicating that a significant aqueous component may be present, even in apparently pure proximal hydrocarbon fluids. Metal-hydrocarbon interactions offer potential exploitation in the exploration for both

PETROGRAPHIC STUDIES

289

European Community. Springer-Verlag, Berlin, 406-427. CLIFFORD, J.A., KUCHA,H. & MELDRUM,A.H. 1988. Lithogeochemistry, its applicability to base metal exploration in a carbonate environment. In: BOISSONNAS, J. • OMENETrO, P. (eds) Mineral Deposits within the European Community. Springer-Verlag, Berlin, 391-405. DE RoD, J.A. & WEBER,K. 1992. Laminated veins and hydrothermal breccia as markers of low-angle faulting, Rhenish Massif, Germany. Tectonophysics, 208,413-430. Acknowledgement is made to the Donors of The DISNAR, J.R. & SUREAU,J.F. 1990. Organic matter in Petroleum Research Fund (grant no. 25024-AC2), ore genesis: progress and perspectives. Organic administered by the American Chemical Society, for Geochemistry, 16,577-599. the partial support of this research. Technical support FERGUSON, J. 1988. The nature and origin of light was kindly provided by the QUB Electron Microscopy hydrocarbon gases associated with mineralization Unit and G. Alexander. The manuscript benefitted in the Northern Pennines. Marine and Petroleum considerably from reviews by D. Ettner and R. Geology, 5,378-384. Kinghorn. Editorial handling: A.H. Ruffell. FERREYRA,R.E. & LARDONE,L.E. 1990. Stratabound uranium deposits in the Argentinian Andes. In: FONTBOTE, L., AMSTUTZ, G.C., CARDOZO, M., References CEDILLO, E. & FRUTOS,J. (eds) Stratabound Ore Deposits in the Andes. Springer-Verlag, Berlin, ABERG, G., LOFVENDAHL,R., NORD, A.G. & HOLM, 671-680. E. 1985. Radionuclide mobility in thucholitic GE, S. & GARVEN, G. 1989. Tectonically induced hydrocarbons in fractured quartzite. Canadian transient groundwater flow in foreland basins. In: Journal of Earth Sciences, 22,959-967. ALLSOPP, H.L., EVANS,I.B., GIus'n, L., HALLBAUER, PRICE, R.A. (ed.) The Origin and Evolution of Sedimentary Basins and Their Energy and Mineral D.K., JONES, M.Q.W. & WELKE, H.J. 1986. Resources. AGU Geophysical Monograph, 48, U-Pb dating and isotopic characterization of carbonaceous components of Witwatersrand 145-157. reefs. In: Abstracts, Geocongress '86. Geological GIZE, A.P. 1993. The analysis of organic matter in ore deposits. In: PARNELL,J., KUCHA,H. & LANDAIS, Society of South Africa, Johannesburg, 85-88. P. (eds) Bitumens in Ore Deposits. SpringerBETHKE, C.M. & MARSHAK,S. 1990. Brine migrations across North America - The plate tectonics of Verlag, Berlin, 28-52. GLASMANN,J.R., CLARK, R.A., LARTER,S., BRIEDIS, groundwater. Annual Review of Earth and Planetary Sciences, 18,287-315. N.A. & LUNDEGARD,P.D. 1989. Diagenesis and , REED, J.D. & OLTZ, D.F. 1991. Long-range hydrocarbon accumulation, Brent Sandstone (Jupetroleum migration in the Illinois Basin. Amerirassic), Bergen High Area, North Sea. American Association of Petroleum Geologists Bulletin, 73, can Association of Petroleum Geologists Bulletin, 75,925-945. 1341-1360. BOWELS,J.F.W. 1990. Age dating of individual grains HAM, W.E. 1956. Asphaltite in the Ouachita Mountains. Oklahoma Geological Survey Mineral Reof uraninite in rocks from electron microscope analyses. Chemical Geology, 83, 47-53. ports No. 30. BRISTOW, C.S. 1988. Controls on the sedimentation of HOFMANN, B. 1990. Reduction spheroids from northern Switzerland: mineralogy, geochemistry the Rough Rock Group (Namurian) from the Pennine Basin of northern England. In: BESLY, and genetic models. Chemical Geology, 81, B.M. & KEELING, G. (eds) Sedimentation in a 55-81. 1991. Mineralogy and geochemistry of reduction Synorogenic Basin Complex: the Upper Carbonspheroids in red beds. Mineralogy and Petrology, iferous of Northwest Europe. Blackie, Glasgow, 114-131. 44,107-124. BUSHELL, T.P. 1986. Reservoir geology of the More- IECHIK, R.P., BRIMHALL,G.H. & SCHULL,H.W. 1986. Hydrothermal maturation of indigenous organic cambe field. In: BROOKS, J., GOFF, J.C. & VAN matter at the Alligator Ridge gold deposits, HOORN, B. (eds) Habitat of Palaeozoic gas in Nevada. Economic Geology, 81, 113-130. N.W. Europe. Geological Society, London, JONES, G.V. & BRAND,S.F. 1986. The setting, styles of Special Publications, 23,189-208. CARTER, J.S. ~g CAZALET,P.C.D. 1984. Hydrocarbon mineralization and mode of origin of the Balligases in rocks as pathfinders for mineral explornalack Zn-Pb deposits. In: ANDREW, C.J., CROWE, R., FINLAY,S., PENNELL, S. & PYNE, J. ation. In: Prospecting in Areas of Glaciated (eds) The Geology and Genesis of Mineral Terrain, 1984. Institute of Mining and Metallurgy, London, 1-20. Deposits in Ireland. Irish Association for Economic Geology, Dublin, 355-375. - - , - - & FERGUSON,J. 1988. Light hydrocarbon gases and mineralization. In: BOISSONNAS,J. & KAPO, G. 1978. Vanadium: Key to Venezuelan fossil hydrocarbons. In: CHIEINGARIAN,G.V. & YEN, OMENETrO, P. (eds) Mineral Deposits within the

metals and hydrocarbons. Metal-rich organic matter can be an indicator of metal anomalies in the proximity of an ore deposit, even in the case of bitumen which post-dates the main ore stage. Solid bitumens which have been precipitated with/about uranium or thorium minerals provide evidence for hydrocarbon migration pathways, and may even permit dating of the migration/ interaction process.

290

J. PARNELL

T.F. (eds) Bitumens, Asphalts and Tar Sands. Elsevier, Amsterdam, 155-190. LEACI-I, D.L. & ROWAN, E.L. 1986. Genetic link between Ouachita foldbelt tectonism and the Mississippi Valley-type lead-zinc deposits of the Ozarks. Geology, 14,931-935. ~, VmTS, J.G. & ROWAN,E.L. 1984. AppalachianOuachita orogeny and Mississippi Valley-type lead-zinc deposits. Geological Society of America Abstracts with Programs, 16,572. LEBKUCHNER, R.F., ORHUN, F. & WOLF, M. 1972. Asphaltic substances in southeastern Turkey. American Association of Petroleum Geologists Bulletin, 56, 1939-1964. MACCHI,L., CURTIS,C.D., LEVISON,A., WOODWARD, K. & HUGHES, C.R. 1990. Chemistry, morphology, and distribution of illites from Morecambe Gas Field, Irish Sea, offshore United Kingdom. American Association of Petroleum Geologists Bulletin, 74,296-308. McKmDY, D.M. & KANTSLER, A.J. 1980. Oil geochemistry and potential source rocks of the Officer Basin, South Australia. APEA Journal, 20, 68--86. MARIKOS, M.A., LAUDON, R.C. & LEVENTHAL,J.S. 1986. Solid insoluble bitumen in the Magmont West orebody, southeast Missouri. Economic Geology, 81, 1983-1988. MILLER,J.M. & TAYLOR,K. 1966. Uranium mineralization near Dalbeattie, Kirkcudbrightshire. Bulletin of the Geological Survey of Great Britain, 25, 1-18. MONSON, B. & PARNELL, J. 1992. The origin of gilsonite vein deposits in the Uinta Basin, Utah. In: FOUCH,T.D., NuccIo, V.F. & CmDSEY, T.C. (eds) Hydrocarbon and Mineral Resources of the Uinta Basin, Utah and Colorado. Utah Geological Association, Salt Lake City, 257-270. NIEWENDORP, C.A. & CLENDENIN,C.W. 1993. Paragenetic link between organic matter and latestage ore deposition in the Sweetwater mine, Viburnum Trend, Southeast Missouri. Economic Geology, 88,957-960. OLIVER, J. 1986. Fluids expelled tectonically from orogenic belts: Their role in hydrocarbon migration and other geological phenomena. Geology, 14, 99-102. OZEROVA, N.A., MASHYANOV,N.R., RYZHOV,V.V., SVESHNIKOV,G.B., PIKOVSKY,YU. I., LEONTYEV, I.A., DOBRYANSKY,L.A. & GRUZDEVA, M.A. 1992. Naphthametallogeny of mercury. Proceedings of International Symposium on Unconventional Hydrocarbon Sources, VNIGRI, St. Petersburg, 114-115. PARNELL, J. 1988. Metal enrichments in solid bitumens: A review. Mineralium Deposita, 23, 191199. 1991. Timing of hydrocarbon-metal interactions during basin evolution. In: PAGEL,M. & LEROY,J. (eds) Source, Transport and Deposition of Metals. Balkema, Rotterdam, 573-576. 1992. Metal enrichment in bitumens from Carboniferous-hosted ore deposits of the British Isles. Chemical Geology, 99, 115-124.

1993a. Chemical age dating of hydrocarbon migration using uraniferous bitumens, CzechPolish border region. In: PARNELL,J., KUCHA,H. & LANDAIS, P. (eds) Bitumens in Ore Deposits. Springer-Verlag, Berlin, 510-517. 1993b. Metal enrichments in bitumens from the Carboniferous of Ireland: Potential in exploration for ore deposits. In: PARNELL,J., KUCHA,H. & LANDAIS, P. (eds) Bitumens in Ore Deposits. Springer-Verlag, Berlin, 475-489. & EAKIN,P. 1987. The replacement of sandstones by uraniferous hydrocarbons: Significance for petroleum migration. Mineralogical Magazine, 51,505-515. & MONSON,B. 1990. Sandstone-hosted thoriumbitumen mineralization in the Northwest Irish Basin. Sedimentology, 37, 1011-1022. & TOSSWILL, R.J. 1990. Petrography of thoriferous hydrocarbon nodules in sandstones, and their significance for petroleum exploration. Journal of the Geological Society, London, 147, 837-842. SWAINBANK, I.G. 1990. Pb-Pb dating of hydrocarbon migration into a bitumen-bearing ore deposit, North Wales. Geology, 18, 10281030. PEABODY,C.E. 1993. The association of cinnabar and bitumen in mercury deposits of the California Coast Ranges. In: PARNELL, J., KUCHA, H. & LANDAIS, P. (eds) Bitumens in Ore Deposits. Springer-Verlag, Berlin, 178-209. PETERSON, J.A. & CLARKE, J.W. 1991. Petroleum Geology and habitat of the West Siberian Basin. AAPG Studies in Geology, 32. RACKLEY, R.I. 1976. Origin of Western-states type uranium mineralization. In: WOLF, K.H. (ed.) Handbook of Stratabound and Stratiform Ore Deposits, 7. Elsevier, Amsterdam, 89-152. RASMUSSEN, B., GLOVER,J.E. & ALEXANDER,R. 1989. Hydrocarbon rims on monazite in PermianTriassic arenites, northern Perth Basin, Western Australia: Pointers to the former presence of oil. Geology, 17,115-118. ROBB,L.J., LANDAIS,P., MEYER,F.M. & DAVIS,D.W. 1993. Nodular kerogen in granites: Implications for the origin of carbonaceous matter in the Witwatersrand Basin, South Africa. In: PARNELL, J., RUFFELL, A.H. & MOLES, N.R. (eds) Geofluids '93, Torquay 4-7 May 1993, 446-449. ROBERTS, G. 1991. Structural controls on fluid migration through the Rencurel thrust zone, Vercors, French Sub-Alpine Chains. In: ENGLAND, W.A. & FLEET, A.J. (eds) Petroleum Migration. Geological Society, London, Special Publications 59,245-262. ROBERTS, W.H. 1980. Design and function of oil and gas traps. In: ROBERTS,W.H. & CORDELL,R.J. (eds) Problems of Petroleum Migration. AAPG Studies in Geology, 10,217-240. SAXBY,J.D. 1976. The significance of organic matter in ore genesis. In: WOLF, K.H. (ed.) Handbook of Stratabound and Stratiform Ore Deposits, Volume 2: Geochemical Studies. Elsevier, Amsterdam, 111-133.

PETROGRAPHIC STUDIES SCHIDLOWSKI, M. 1981. Uraniferous constituents

of the Witwatersrand conglomerates: Oremicroscopic observations and implications for Witwatersrand metallogeny. United States Geological Survey Professional Paper, 1161. S'rON~LEV, R. 1983. Fibrous calcite veins, overpressures, and primary oil migration. American Association of Petroleum Geologists Bulletin, 67, 1427-1428. VERSI-IKOVSKAYA, O.V., PIKOVSHIY, YU. I. & SOLOV'YEV, A.A 1972. Dispersed carbonaceous material in rocks and ores of the Plamenoye antimony-mercury deposit. Doklady Akademiya Nauk USSR (Earth Sciences Section), 205, 220222.

291

WILLINGHAM,T.O., NAGY,B., NAGY,L.A., KRINSLEY, D.H. & MOSSMAN,D.J. 1985. Uranium-bearing stratiform organic matter in paleoplacers of the lower Huronian Supergroup, Elliot Lake-Blind River region, Canada. Canadian Journal of Earth Sciences, 22, 1930-1944. WOODWARD, K. & CURTIS, C.D. 1987. Predictive modelling for the distribution of productionconstraining illites - Morecambe Gas Field, Irish Sea, Offshore UK. In: BROOKS,J. & GLENNIE,K. (eds) Petroleum Geology of North West Europe. Graham & Trotman, London, 205-215.

The role of geopressure zones in the formation of hydrothermal Pb-Zn Mississippi Valley type mineralization in sedimentary basins A N T H O N Y D. F O W L E R

Ottawa Carleton Geoscience Centre and Department of Earth Sciences, University of Ottawa, 770 King Edward Avenue, Ottawa, Ontario, Canada K1N 6N5 Abstract: Geopressure zones immediately subjacent to platform carbonate rocks can be modelled to serve as proximal sources of hydrothermal fluids for epigenetic Pb-Zn deposits in sedimentary basins. In some geopressure zones of the Gulf of Mexico, geothermal gradients can be as high as 10°C 100m-I. This arises because the water-saturated geopressured shale masses act as thermal insulators. Geopressure zones may have sufficient fluid pressure to rupture overlying strata, providing a vertical conduit for hot mineralized brine to migrate directly into carbonate host rocks.

A complete understanding of ore formation in sedimentary basins is a difficult and unresolved problem of economic geology. One of the more enigmatic types of deposit, in terms of its origins, is the Mississippi Valley type (MVT). The realization that hydrothermal mineralization 'happened' in these deposits without igneous activity has long driven deposit modellers to find an alternative source of heat. Chief among the explanations are migration mechanisms that involve fluid scavenging of heat from deep within the basin during their migration over hundreds of kilometers to sites of mineralization at basin margins. An alternative model (Fowler & Anderson 1991, 1985) relates the mineralization to nearby sources of hydrothermal fluids in geopressure zones. The purpose of this paper is to review the nature of geopressure zones and how they might be related to Pb-Zn mineral deposits.

Characteristics of Pb-Zn ore deposits Although Pb-Zn mineralization is found in a variety of settings in sedimentary systems, herein I consider only those with a demonstrable epigenetic character. With notable exceptions (e.g. Laisvall, Sweden) the bulk of these deposits are formed in carbonate host rocks (Schrijver 1992). Famous among these are the Mississippi Valley type deposits. This term has been used worldwide to classify deposits which on a broad scale appear to be very similar. This apparent commonality of characteristics has further led to the application of similar genetic models for the formation of the deposits. Although the deposits are indeed related by common features, genetic mechanisms may

have been extremely varied between the individual deposits and mining districts. Mississippi Valley type deposits are found within carbonate rocks often close to major facies changes at the present day edges of sedimentary basins (Anderson & Macqueen 1982). Ore is dominated by galena and sphalerite found as open space fillings; in karst cavities, faults or collapse breccias. It is associated with a white sparry dolomite and the gangue minerals pyrite, calcite, quartz, barite, fluorite, and in some cases bitumen. Non-equilibrium textures such as dendritic and bladed galena, and oscillatory zoned sphalerite are common within the ores. Mineral banding has been interpreted to be correlative over long distances, of the order of hundreds and even thousands of metres (McLimans et al. 1980). Fluid inclusions are saline and have low Na/K ratios relative to oilfield brines (Roedder 1976). Geothermometry based upon fluid inclusion data, stable isotope data, and vitrinite reflectance data indicate that the fluids responsible for the mineralization were hot with respect to the host rocks (Anderson & Macqueen 1982). Finally, the deposits are thought to form at shallow depths of approximately a kilometre or less (Anderson & Macqueen 1982).

Models of origin Modern workers (e.g. Jackson & Beales 1967) first interpreted the mineralization to have arisen from the compaction-driven migration of metal-rich brines into HzS-bearing carbonate rocks. The fluids and metals were thought to have been derived from the diagenesis of clastic sediments. The model did not account for the

FromPARNELL,J. (ed.), 1994, Geofluids: Origin, Migrationand Evolution of Fluidsin Sedimentary Basins, Geological Society Special Publication No. 78,293-300.

293

294

A.D. FOWLER

ki I00

basement

km

300

"v-

~

shale

Fig. 1. Schematic diagram of the gravity driven cross-formational fluid flow model. Uplift in the topographically high region provides an energy source for sub-surface fluid flow. Fluids travel down through shales to the carbonate formation whereupon they are focused and travel horizontally, eventually reaching the site of ore deposition which is at a lower elevation than the recharge area in the uplifted region. Fluids are modelled to migrate rapidly such that they can advect heat from deep within the basin. Permeability limits the rate at which the fluid can migrate.

fact that the mineralization temperatures were often in the 100--200° C range, and in some cases significantly higher than the temperatures recorded by host rocks. This led to the idea that hot mineralizing fluids were derived from distal sources deep (c. 5 kilometres) within the basins, in response to compaction. Fluids migrated laterally along hundreds of kilometres through gently-dipping basinal aquifers to shallower (c. 1 km) and cooler host rocks at the basin margins. However, it has been shown that basinal compaction rates are slow and, as a consequence, fluids thermally re-equilibrate with their host rocks during long-distance migration, arriving only at ambient temperature. Accordingly, numerous models have been recently proposed in order to drive fluids more rapidly, and thus conserve heat from depth. Sharp (1978) and Cathles & Smith (1983) proposed that geopressure zones could serve as an energy source for rapid fluid migration. Cathles & Smith postulated that rupture of geopressure zones deep within a basin could cause fluid migration down into homogeneous and permeable aquifers. The rapid fluid flow

would be driven by the pressure drop in the geopressure zone and would enable fast transport over hundreds of kilometres, conserving heat. Garven (1985) and Bethke (1986a) proposed that rapid fluid migration can be driven by gravity potential. Tectonic uplift causes a hydraulic head, and hence an energy source for subsurface flow (see Fig. 1). Garven (1985) used this concept to model the genesis of the Pine Point deposit, North West Territories, Canada. In essence, the fluids are proposed to have migrated into the middle Devonian host carbonate rocks in response to uplift of the Rocky mountains in post-Cretaceous time. Thus the fluids are modelled to have flowed approximately 500 km through and across shales, sandstones and carbonate rocks. Fowler (1986) argued that the permeability of the strata was insufficient to support the rapid migration. Moreover, it has been recently demonstrated on the basis of precise Rb-Sr dating of sphalerite, that the Pine Point deposit was formed shortly after the deposition of its host rocks, i.e. in middle Devonian rather than Eocene time (Nakai et al. 1993).

GEOPRESSURE ZONES AND Pb-Zn MINERALIZATION The gravity flow model has also been applied to the deposits of the Ozark Mountains, USA. For instance, Leach & Rowan (1986) related the Ozark mineralization to the migration of fluids through a foreland basin as a result of a hydraulic head established by uplift of the Ouachita Mountains during Permian plate collision. They cite evidence of broad-scale heating, cooling in the down-hydraulic gradient direction, largescale correlations of mineral banding, and age data as being consistent with the gravity driven fluid flow model (e.g. Bethke 1986a). Additionally, Pan et al. (1990) and Symons & Sangster (1991) have dated many deposits within the area by palaeomagnetic means, and conclude that the data are consistent with the timing of uplift of the Ouachita Mountains. The field relationships of Clendenin & Duane (1990) confirm the timing of mineralization. However these authors cite and interpret field evidence to be indicative of multiple fluid sources, and invoke seismic pumping (e.g. Sibson et al. 1975) as a drive mechanism. Subsequent comment by Leach & Rowan (1991) and reply by Clendenin (1991) and the recent Rb-Sr age dating indicating a Mid-Palaeozoic age date for these deposits shows that the drive mechanism(s) for the Ozark Mississippi Valley type deposits are unresolved. Bethke (1986a) studied mineralization in the Illinois basin and concluded that although a compaction-driven flow mechanism cannot account for the temperatures of the Mississippi Valley type mineralization, gravity flow produced by a 700 m uplift across the 700 km basin was a feasible mechanism. His calculations demonstrate that permeability is the critical transport velocity factor for gravity flow, whereas for compaction-driven flow it is the compaction rate.

295

Geopressure zones Under hydrostatic conditions fluid within a saturated sediment or rock has a pressure given by

v , : gf o,(z) az

o)

where Pf is the hydrostatic pore fluid pressure or neutral stress, and pf = fluid density, z = depth, and g = acceleration due to gravity. The lithostatic pressure (or overburden stress) in the absence of a deviatoric stress field is given by

P,=~f pt(z)dz

(2)

where 9t = bulk rock density. The difference between equations 1 and 2 (P, - Pr) gives the intergranular stress or pressure, i.e. the stress transmitted from grain to grain. Under hydrostatic conditions the fluid fills a continuously connected matrix of voids such that the pressure is defined by the weight of the overlying fluid column, as given in equation 1. In contrast, geopressured fluids have pressures greater than those predicted by equation 1 and, surprisingly, in some cases pressures greater than Pt calculated from equation 2 (Fertl & Timko 1970). In contrast to hydrostatically pressured fluids, geopressure zone fluids support a load greater than that of just the overlying fluid column. Frequently this arises when relatively impermeable sediments are rapidly deposited. As the load is increased with continued sedimentation, the already deposited impermeable sediments are unable to compact by expulsion of their contained pore water. Consequently fluids within the intergranular porosity of the matrix support stress that under hydrostatic conditions would be supported by grain-to-grain contact in the rock matrix. Fluid overpressure may build Geopressure zones as proximal sources of up to exceed the sum of the least principal stress h y d r o t h e r m a l fluids and tensile stress of the matrix so that rupture The fact that permeability and compaction rates occurs. Although mineral exploration geologists may are the critical rate-limiting factors for long distance flow led Fowler & Anderson (1991, not be familiar with the magnitude of the 1985) to propose a proximal source model for pressure within geopressure zones, petroleum the hydrothermal fluids. A key observation for exploration geologists are, because geopressure this model is the fact that the fluids of some zones represent a distinct drilling hazard. Geopressure zones can sustain large differences in geopressure zones may be anomalously hot. Below, I review the nature of geopressure zones, pressure. For instance a difference in pore drawing heavily on observations from the Gulf pressure of almost 350 atm over a 100 m thick of Mexico, because geopressure zones are shale section has been reported, and cases where widespread there and have been intensively Pf = Pt and also Pc> Pt ('superpressure') are studied. I then present and discuss the proximal known (Fertl & Timko 1970, 1972). Fluid overpressure can form in a variety of source model.

296

A.D. FOWLER

different environments (e.g. igneous hydrothermal systems). However, the relevant systems for Mississippi Valley type deposits are those of clastic sedimentary rocks. The basin in which geopressuring has been most studied is the Gulf of Mexico. Here a variety of mechanisms that may cause or enhance the development of geopressuring has been proposed in addition to the rapid deposition of impermeable sediments (e.g. Hunt 1990). Magara (1975) has proposed that there is a net gain in volume during the diagenetic transition from smectite to illite, which involves expulsion of structurally bound water and may cause an increase in the pressure of enclosed formations. It has also been demonstrated that geopressuring can also result as a consequence of the formation of hydrocarbons. Barker (1972) concluded that the coefficient of thermal expansion of water is sufficient to cause a significant pressure increase. He calculated that a temperature change of 50° C in an isolated system could produce an excess pore pressure of approximately 300 atmospheres. Bethke (1986b) has calculated that the sediments of shale-rich basins in which the subsidence rate is greater than approximately 1 mm a -1 are likely to be geopressured. Shale-poor basins and those subsiding at less than 0.1 mm a -1 are not likely to develop geopressure. Irrespective of the mechanism or mechanisms of formation of excess pressure it is certain that it must be contained within impermeable rocks in order to be maintained. One outstanding feature of many of the geopressure zones of the Gulf of Mexico is the fact that they are also geothermal zones. During the energy crisis of the 1970s several of the on-shore geopressure-geothermal zones of the US Gulf Coast were extensively studied as they were perceived as potential sources of geothermal energy. Although there are areas where geopressure zones are not geothermal zones, many geopressure zones of the Gulf Coast are exceptional. Jones (1970) has shown that the average geothermal gradient in the normally pressured sediments of the Gulf Coast varies from 20 to 40° C km -1. Within tens of metres of the tops of geopressure zones, the geothermal gradient often increases abruptly, reaching values sometimes in excess of 100°C km -1. His data show that the 120° C isotherm is associated with geopressure zones as shallow as 2.5km, whereas in a hydrostatic pressured-section it is found at approximately 3.5 km. Lewis & Rose (1970) proposed that the steep geothermal gradients associated with geopressure zones are due to the insulating properties of the impermeable sealing strata. They attributed

the increased insulation to the fact that geopressure zones are more porous and hence contain more water (which has a low coefficient of thermal conductivity) than their surrounding normally pressured counterparts. The low permeability of the geopressured shales prevents widespread advection of the pore fluid, hence there is a build up of heat, due to the change from advective to the less effective conductive heat transfer mechanisms. Thus the rocks above the insulator have lower than average heat flow and temperatures, whereas within the insulator the geotherms converge and higher temperatures prevail.

A proximal geopressure zone model of origin for Pb-Zn deposits Figure 2 captures the essence of the model of Fowler & Anderson (1991). The model depicts a section through a prograded basin in which rapidly deposited deltaic shales, siltstones and sandstones are overlain by carbonates. It is particularly appealing for those deposits hosted in shale-rich basins. The shale is porous, yet impermeable because of the large surface areas and ramified pore-choking structure of its constituent clay minerals. It is overpressured and contains within it sandstone lenses of significant size and permeability (1 D). The hypothetical geothermal gradients of Fig. 2 reveal that the section at the top of the geopressure zone is anomolously hot whereas the overlying carbonates are correspondingly anomalously cool. Sands within geopressure zones can be extremely permeable, again because they are under-compacted. For instance, Fowler & Anderson (1991) reported permeabilities greater than 1 D for geopressured sandstones recovered from a depth of 5 km from the Miocene Corsair trend of the Texas Gulf Coast. It can be shown that fluids within sand bodies may freely convect if they lie within a steep geothermal gradient. Calculations based on the model of Bj~rlykke et al. (1988) show that fluids within a large (1 km 3) km permeable (1D) sandstone within a high geothermal gradient of 10°C 100 m -1 will spontaneously convect. Thus the minerals within the sands undergo diagenetic reactions at shallower depths than their normally pressured counterparts because of the relatively high temperatures and fluid convection. Albitization of K-feldspar which has significant amounts of Pb substituting in the K-site enriches the fluid in Pb. Zn is enriched within the fluid as a result of the lower cation exchange capacity of illitic clays in comparison to their precursor smectitic clays.

GEOPRESSURE ZONES AND Pb-Zn MINERALIZATION

297

km

km

~basement

v

[~]

limestone

shale

I

-

- I siltst°ne

sandstone

Fig. 2. Schematic diagram depicting the geopressure zone model of Pb-Zn mineralization of Fowler & Anderson (1991, 1985). Here, siltstones grade into shales at a facies change (jagged curve). Geopressured shales containing geopressured sands underlie carbonate rocks. The shale is impermeable and has a fluid pressure exceeding hydrostatic pressure. The isotherms are warped due to the insulating effect of the shale. Thus the large sand mass buried within the geopressure zone is a potential reservoir wherein diagenetic reactions (e.g. albitization of K-feldspar, and smectite-illite transition) capable of release Pb and Zn may proceed at relatively shallow depths. Hydraulic fracturing leads to vertical expulsion of the fluids into the overlying carbonate rock, causing hydrothermal karstification and mineralization. Brecciation is in part due to hydraulic fracturing, and collapse as a result of volume loss in the subjacent geopressure zone. Fluids within the sand of the permeable siltstone are at hydrostatic pressure and 'normal' temperature, and as a consequence this sand is not a source of mineralizing fluid. Fluid pressure builds up by one or more of the mechanisms above until the geopressured system leaks. This may occur catastrophically by the formation of self-induced hydraulic fractures, or through the exploitation of existing channelways such as faults. Hydraulic fractures in the absence of an imposed far field (i.e. tectonic) stress are vertical, and occur when the fluid pressure exceeds the sum of the least compressive stress and the tensile stress of the rock. Generally this pressure is less than the overburden or lithostatic stress (Hubbert & Willis 1957). As rocks are weak in tension, fracturing will generally occur at fluid pressures less than the total pressure given in equation 2. The hot fluid rapidly migrates a short vertical distance into relatively cool carbonate host rocks, thus explaining the temperature anomaly associated with the deposits. The porosity of the host is increased through brecciation associated with the hydraulic fracturing, brecciation as a result of collapse due to volume loss within the

underlying geopressure zone, and dissolution of the carbonate along pre-existing joints, so-called hydrothermal karstification (Sass-Gustkiewicz et al. 1982).

Discussion Although this model for the origin of P b - Z n deposits is conjectural it may explain the presence of hydrothermal mineralization under conditions of restricted permeability, which long distance migration models cannot. The parameters critical to the model are the pressure, temperature, and fluid volume of geopressured zones. Undoubtedly overpressured sections exist as witnessed by many hapless drillers and the observation of features such as sandstone dykes and cone-in-cone structures. However, it is important to investigate the possibility of overpressure having naturally given rise to hydraulic fracturing on a large scale in the geologic past.

298

A.D. FOWLER

Although the data are scant, and there is no proof, recent work is wholly consistent with the idea. For instance Cartwright (1994) has used 3-D seismic data to map out a set of vertical normal faults found over tens of thousands of square kilometres within Palaeocene to midMiocene age mudrocks of the UK, Norwegian and Danish sectors of the North Sea. The faults originate and terminate within the Palaeocene to mid-Miocene age mudrocks and are organized into cellular networks. They are distributed at intervals of approximately 100-500m, have throws of 10-100m, and random orientations within the vertical plane. Cartwright interpreted the faults as being due to hydraulic fracturing associated with fluid expulsion from extensively developed overpressure zones. Capuano (1993) documented the existence of altered fractures within cores of the geopressured Oligocene Frio Formation of Texas, USA. The fractures are interpreted to 'provide direct evidence of hydraulic fracturing and significant fluid flow in the shales'. Significantly, Capuano (1993) used the 'cubic law' and the Darcy equation to calculate shale permeabilities on the order of 10-13m 2, similar to those of the Frio sandstone reservoirs. Carey & Parnell (1993) studied the bitumen veins of the Neuquen Basin Argentina, and although the area is structurally complex and some of the veins can be attributed to igneous activity, Carey & Parnell have been able to demonstrate through detailed field mapping that other bitumen veins are associated with hydraulic fracturing. The fracturing was related to pressure build up during 'the generation of hydrocarbons within the succession as a result of rapid burial'. Clearly, any model of formation must be capable of accounting for the production of hot mineralizing fluids. Indeed the impetus for the gravity drive and other long distance migration models was the explanation of a mechanism to bring heat from the deep and central portions of the basin to shallow depths near the basin margin. In the proximal source model the fluids are hot because of the insulating properties of the geopressure zone. This is not to suggest that the fluids of all geopressured zones are anomalously hot. However, the important factor for the mineralization is that the mineralizing fluids were hot with respect to the host rocks at the time of mineralization. Due to the steep geothermal gradients observed in some geopressure zones the model does not require large vertical separation of host and source as is the case for models that consider normal geothermal gradients. Therefore one may expect that the host and source could be separated in the vertical plane

by as short a distance as a few hundred metres. Aside from deposits where evidence of near surface processes such as vadose karstification are present, it appears that the idea in which deposits form close to surface (c. 1 km) and near basin margins stems from the low temperatures recorded by the host rocks. Given that a basin margin is ephemeral (Walther's facies rule) it is entirely possible that the deposits are formed deep (c. 72 km) in rocks cooler than expected, at what are now the exposed paleo-margins of basins. Geopressured shale masses in the Gulf of Mexico are large (hundreds of cubic kilometres); however, their contained individual sands are generally small, on the order of cubic kilometres or less. Paine et al. (1979) have shown a small reservoir in the Abbeville, Louisiana area to contain of the order of 4 x 10111 of brine. Cannon & Craze (1937) reported that a well in Allen Parish, Louisiana produced approximately 7 million barrels (c. 1091) of formation water without a reduction in flow rate. In many cases the sandstones of geopressured reservoirs over-produce, or have a reduced rate of pressure reduction compared to predictions based upon their volume, porosity, pressure, and fluid drive mechanism (Atwater 1967). The discrepancy has been explained by fluid flow from overpressured shales into their contained sands during production. It is estimated that sand reservoirs drain fluids from adjacent sediments a distance of two to seven times their radii (Glezen & Lerche 1985). Moreover, the recent evidence of Capuano (1993) that geopressured shales may have substantially increased permeability as a result of hydraulic fracturing and the observations of Cartwright (1994) on the widespread extent of hydrofracturing in the North Sea suggests that geopressure zones can supply a vast amount of fluid.

Conclusions Geopressure zones are large masses of impermeable undercompacted sediment that have fluid pressures in excess of hydrostatic. Due to the insulating effects of the geopressure zones they may trap heat and be geothermal zones. The geopressure zone is a source of hot mineralizing fluid formed directly subjacent to host rocks for mineralization. Permeability is the key parameter in order to initiate or conserve geopressure. Once initiated the development of the overpressure becomes maybe self-propagating because of feedback in the controlling parameters. For instance, the insulating properties of the overpressure zone

GEOPRESSURE ZONES AND Pb-Zn MINERALIZATION affect the local heat flow regime and hence temperature, which effect pressure-regulating temperature-sensitive diagenetic reactions. The permeability of the system is tremendously enhanced through hydraulic fracturing such that a large volume of rock can be quickly drained of its fluid into overlying host rocks. The comments of N. Moles, J. Parnell and an anonymous reviewer are greatly appreciated. The author also appreciates the financial support of the Natural Science and Engineering Research Council of Canada.

References ANDERSON, G.M. & MACQUEEN,R.W. 1982. Ore deposit models -6. Mississippi Valley-type leadzinc deposits. Geoscience Canada, 9, 108-117. ATWATER, G.I. 1967. Geopressured gas reservoir performance in the Hollywood Field Terrebonne Parish, Louisiana. In: FERRELL,R.E. & HISE, B.R. (eds) Proceedings of the first symposium on abnormal subsurface pressure. Louisiana State University, Baton Rouge, Louisiana, 11-17. BARKER,C. 1972. Aquathermal pressuring - role of temperature in development of abnormalpressure zones. American Association of Petroleum Geologists Bulletin, 56, 2068-2071. BETHKE, C.M. 1986a. Hydrologic constraints on the genesis of the Upper Mississippi Valley District from Illinois Basin brines. Economic Geology, 81,233-249. - 1986b. Inverse hydrogeologic analysis of the distribution and origin of Gulf Coast-type geopressured zones. Journal of Geophysical Research, B6, 91, 6535-6545. BJORLYKKE,K., Mo. A. & PALM,E. 1988. Modelling of thermal convection in sedimentary basins and its relevance to diagenetic reactions. Marine and Petroleum Geology, 5,338-351. CANNON, G.E. & CRAZE, R.C. 1937. Excessive pressures and pressure variations with depth of petroleum reservoirs in the Gulf Coast region of Texas and Louisiana. Journal of the Institute of Petroleum Technology, 23, 31-38. CAPUANO, R.M. 1993. Evidence of fluid flow in microfractures in geopressured shales. American Association of Petroleum Geologists Bulletin. 77. 1303-1314. CAREY,P.F. & PARNELL,J. 1993. Solid bitumen veins related to overpressuring and thrusting in the Neuquen basin, Argentina. In: PARNELL, J., RUFFELL,A.H. & MOLES,N.R. (eds) Geofluids '93, Torquay 4-7 May 1993, 181-185. CARTWRIOHT, J.A. 1994. Episodic basin-wide hydrofracturing of overpressured early Cenozoic mudrock sequences in the North Sea basin. Marine and Petroleum Geology, (in press). CATHLES, L.M. & SMmt, A.T. 1983. Thermal constraints on the formation of Mississippi Valleytype lead-zinc deposits and their implications for

299

episodic basin dewatering and deposit genesis. Economic Geology, 78,983-1002. CLENDENIN, C.W. 1991. Comment and reply on focused flow and Ozark Mississippi Valley-type deposits. Geology, 19,190-191. & DUANE, M.J. 1990. Focused fluid flow and Ozark Mississippi Valley-type deposits. Geology, 18, 116-119. FERTL, W.H. & TIMKO, D.J. 1970. Overpressured formations - 1. Occurrence and significance of abnormal - pressure formation. Oil and Gas Journal, Jan. 5, 97. & 1972. How downhole pressures, temperatures affect drilling. World Oil, Sept, 45-5-50. FOWLER, A.D. 1986. The role of regional fluid flow in the genesis of the Pine Point deposit, Western Canada sedimentary basin - a discussion. Economic Geology, 81, 1014-1015. & ANDERSON,M.T. 1991. Geopressure zones as proximal sources of hydrothermal fluids in sedimentary basins and the origin of Mississippi Valley type deposits in shale rich sequences.

Transactions of the Institute of Mining and Metallurgy, 100, B14-B18. & 1985. A general model relating carbonate hosted lead zinc deposits and overpressure zones.

Geological Association of Canada, Mineralogical Association of Canada joint annual meeting program and abstracts, Fredericton N.B. A19. GARVEN,G. 1985. The role of regional fluid flow in the genesis of the Pine Point deposit, Western Canada Sedimentary Basin. Economic Geology, 80,307-324.

GLEZEN, W.H. & LERCHE, I. 1985. I. A model of regional fluid flow: sand concentration factors and effective lateral and vertical permeabilities. Mathematical Geology, 17,297-315. HUBBERT, M.K. & WILLIS, D.G. 1957. Mechanics of hydraulic fracturing. Transactions American In-

stitute of Mechanical Engineers, Society of Petroleum Engineers, 210, 153-168. HUNT, J.M. 1990. Generation and migration of petroleum from abnormally pressured fluid compartments. Bulletin of the American Association of Petroleum Geologists, 74, 1-12. JACKSON,S.A. & BEALES,F.W. 1967. An aspect of sedimentary basin evolution: The concentration of Mississippi Valley-type ores during late stages of diagenesis. Bulletin of Canadian Petroleum Geology, 15,383-433. JONES,P.H. 1970. Geothermal resources of the northern Gulf of Mexico basin: U.N. symposium on the development and utilization of geothermal resources, Pisa. Geothermics special issue, 2, 14-26. LEACH, D.L. & ROWAN, E.L. 1986. Genetic link between Ouachita foldbelt tectonism and the Mississippi Valley-type lead-zinc deposits of the Ozarks. Geology, 14,931-935. -& -1991. Comment and Reply on Focused flow and Ozark Mississippi Valley-type deposits. Geology, 19,190-191. LEwis, C.R. & ROSE, S.C. 1970. A theory relating high temperatures and overpressures. Journal of Petroleum Technology, 11-16.

300

A.D. FOWLER

MAGARA, K. 1975. Reevaluation of montmorillonite dehydration as cause of abnormal pressure and hydrocarbon migration. American Association of Petroleum Geologists Bulletin, 59, 292302. MCLIMANS, R.K., BARNES,H.L. & OHMOTO,H. 1980. Sphalerite stratigraphy of the upper Mississippi Valley zinc district, southwestern Wisconsin. Economic Geology, 75,351-361. NAKAI, S., HALLIDAV, A.N., KESLER, S.E., JONES, H.D., KYLE, R.J. & LANE, T.S. 1993. Rb-Sr dating of spha|erites from Mississippi Valley-type (MVT) ore deposits. Geochimica et Cosmochimica Acta, 57,417--427. PAINE, W.R., KINSLAND, G.L., DUHON, M.P. & DUNGAN, J.R. 1979. Subsurface and seismic investigation of the geopressured - geothermal potential of south Louisiana, part 1: The Abbeville area. In: DORFMAN, M.H. & FISHER, W.L. (eds) Proceedings of the fourth United States Gulf Coast geopressured-geothermal energy conference. Austin University Center for Energy Studies, 3, 1160-1193. PAN, H., SYMONS, D.T.A. & SANGSTER,D.F. 1990. Paleomagnetism of the Mississippi Valley-Type ores and host rocks in the northern Arkansas and Tri-state districts. Canadian Journal of Earth Sciences, 27,923-931.

ROEDDER, E. 1976. Fluid inclusion evidence in the genesis of ores in sedimentary and volcanic rocks. In: WOLF,K.H. (ed.) Handbook of strata-bound and stratiform ore deposits, 4, 67-110. SASS-GUSTKIEWICZ,M., DZULYNSKI,S. & RIDGE, J.D. 1982. The emplacement of zinc-lead sulfide ores in the Upper Silesian District - A contribution to the understanding of Mississippi Valley-Type deposits. Economic Geology, 77, 1982, 392-412. SCHRIJVER,K. 1992. Basinal brines and groundwaters as possible metal carriers in the formation of sandstone-hosted lead-zinc deposits. Mineralium Deposita, 27,109-114. SHARP,J.M. JR. 1978. Energy and momentum transport model of the Ouachita basin and its possible impact on formation of economic mineral deposits. Economic Geology, 73, 1057-1068. SIBSON, R.H., MOORE, J.MCM. & RANKIN, A.H. 1975. Seismic pumping - a hydrothermal fluid transport mechanism. Journal of the Geological Society, London, 131,653-659. SYMONS, D.T.A. & SANGSTER,D.F. 1991. Paleomagnetic age of the central Missouri barite deposits and its genetic implications. Economic Geology, 86, 1-12.

Fluid-rock interactions during continental red bed diagenesis: implications for theoretical models of mineralization in sedimentary basins R. METCALFE, C.A. ROCHELLE, D. SAVAGE' & J. W. HIGGO Fluid Processes Group, British Geological Survey, Keyworth, Nottingham NG12 5GG, UK Present address: Intera, 47 Burton Street, Melton Mowbray, Leics LE13 1A F, UK Abstract: Continental red beds are first-cycle, immature sediments which are deposited in oxidizing conditions and owe their red colouration to the early diagenetic development of hematite. Previously published work on the fluidlrock interactions which occur during the diagenesis of such sediments highlights that pH and redox are critical fluid parameters which control diagenesis and ore formation. However, except in modern surface water and shallow groundwater systems, these parameters cannot be measured directly, and must be estimated from theoretical considerations. These suggest that mixing between fluids of different redox states is likely to be a critical control on heavy metal mobility. Many diagenetic features can be explained largely by the chemical heterogeneity of red beds, and by the diagenetic ranges of redox conditions and pH which are among the greatest for any type of sediment. Such features include: the formation of red bed-hosted ore deposits; the extensive development of hematite; and the development of early diagenetic non-ferroan carbonate and late diagenetic ferroan carbonate cements. In order to develop theoretical models of fluidlrock interactions during such diagenesis, it is important to consider the interrelationships between fluid flow, mineral dissolution and precipitation, and sorption. At the present time such models are at an early stage of development.

Continental red beds originate as first-cycle, immature sediments which are principally derived from the erosion of crystalline rocks, and which contain abundant labile minerals. They are typically deposited in extensional tectonic settings, and form from fluvial, lacustrine, paralic and aeolian sediments, which are classically ascribed to deposition within rapidly subsiding, syn-sedimentary, fault-bounded basins (e.g. Turner 1980; Frostick & Reid 1987). Such sedimentary deposits may also form in a number of other environments, such as along continental margins and in clastic wedges along the edges of mountain belts (Walker 1976; Turner 1980). Their accumulation in continental settings allows them to remain in an oxidizing environment for relatively long periods following deposition. This leads to the breakdown of unstable detrital ferromagnesian constituents, such as olivine, pyroxene and amphibole, which produces hematite and imparts the characteristic colouration to red bed sediments. The depositional conditions of continental red beds are often considered to have been arid to semi-arid (Walker 1967), and the common occurrence of evaporites in many sequences is

evidence for this. However, red beds may also form during the diagenesis of sediments deposited in humid climates, provided that the sediments remain under oxidizing conditions for a sufficiently long period following deposition (Walker 1974; Turner 1980). Red beds play two important roles in the occurrence of sedimentary basin resources: firstly, they act as a source of heavy metals which are either bound in the solid phase, or which are located within metalliferous formation waters; and secondly, owing to their relatively high mean porosities and permeabilities, they act as conduits for groundwater flow during basin evolution and may become 'reservoirs' for mineralising fluids, hydrocarbons or groundwater. Red beds may also act as chemical 'conditioning agents' for groundwaters passing through them. As a result of these factors, red beds play a fundamental part both in the genesis and spatial distribution of natural resources, and are important hosts of hydrocarbon accumulations and deposits of metalliferous minerals worldwide. For example, sediment-hosted copper deposits within red bed sequences are exceeded in

From PARNELL, J. (ed.), 1994, GeoJEuids:Origin, Migration and Evolution of Fluids in Sedimentary Basins, Geological Society Special Publication No. 78,301-324.

301

302

R. METCALFE ET AL.

importance only by porphyry copper deposits, Previous studies of diagenesis and and account for about 25% of the world's copper mineralization production and reserves (Kirkham 1989). Red bed sandstones are also important sources of Diagenesis uranium, and account for more than 75% of uranium production in the USA, (OECD 1986). Early diagenesis (eodiagenesis). This occurs Furthermore, single red bed formations may under conditions which are fundamentally simicontrol the distribution of a wide variety of lar to those occurring at the Earth's surface, and economic resources. For example, the Triassic its effects have been constrained by using red beds of the Sherwood Sandstone Group in modern red beds as an analogue for ancient red the UK may, in different places, host major oil bed sequences (Walker 1967, 1976, 1989; and gas reservoirs (North Sea, Wessex and Irish Walker & Honea 1969; Walker & Runnells 1984; Walker et al. 1978; Zielinski et al. 1983, Sea basins) Cu-Pb-Zn mineralization (Cheshire and Lincolnshire basins) commercially valuable 1986; Flint 1987). This has shown that the early evaporites, including sylvite (Cleveland), an- diagenetic waters are generally alkaline and hydrite (Cheshire, Cumbria and Notting- oxidizing and cause in situ weathering and dissolution of unstable ferromagnesian detrital hamshire) and halite (Cleveland, Cheshire). Additionally, these red beds may act as major minerals (pyroxenes, amphiboles, magnetite, potable water sources (Cheshire, Lancashire biotite and olivine). Walker (1989) reported that and Nottinghamshire), and geothermal reser- this leads to a reduction in amphibole abundances in modern red bed fanglomerates to c. 2%, voirs (Wessex Basin). The regional distribution, detailed geometry from c. 4% in the precursor igneous rocks. In and chemical composition of such economic parallel with this process, iron oxyhydroxides resources are functions of a complex interplay precipitate and thus the red beds acquire their between fluid flow paths and the chemical characteristic coloration. With progressive diaevolution of the fluid phase and the host red bed. genesis mineral dissolution releases K, Fe, Mg, In order to understand how this interplay AI, Si into the porewaters and eventually causes controls the distribution of economic resources, chlorite and smectite to replace ferromagnesian it is necessary to understand the relationship minerals, and illite-smectite to replace feldspar. Studies of the Sherwood Sandstone Group of between fluid chemistry and red bed diagenesis. A great deal has been published on the the UK have suggested that variations in detrital mineralogical/petrological aspects of this prob- mineralogy are not a major control on early lem, by a large number of previous authors (e.g. diagenetic mineral parageneses, but that these Walker 1967, 1974, 1976, 1989; Waugh 1978; are instead controlled by the fluid chemistry, in Turner 1980; Holmes et al. 1983; Zielinski et al. turn controlled by the depositional environment 1983; Burley 1984; Bath et al. 1987; Flint 1987) (Burley 1984). A feature of early diagenetic conditions are and the associated ore deposits have also received considerable attention (e.g. Turner the formation of intergranular and replacive 1980; Nash et al. 1981; Ixer & Vaughan 1982; carbonates, and often extensive calcretes (Bath Merino et al. 1986; Sverjensky 1984, 1987, 1989; et al. 1987; Strong & Milodowski 1987). The Kirkham 1989; Walker 1989; Bechtel & P/ittman oxidizing groundwater conditions mean that 1991; Sawlowicz 1991). However, with some carbonates are characteristically non-ferroan notable exceptions (Rose 1976, 1989; Svejensky since iron is in its ferric state and is not easily 1984, 1987, 1989) previously published works incorporated into the carbonate mineral struchave tended to concentrate on the evolution of ture. Rarely, zeolites including clinoptilolite, eriored bed mineral assemblages during diagenesis, and to date there has been no comprehensive nite and analcime may occur when groundwaters review of the state of knowledge regarding the are oxidizing and at near-neutral pH, or when chemical evolution of the fluid phase, or of the alkaline brines are evaporated in closed basins theoretical basis for modelling red bed dia- (Surdam & Sheppard 1978; Hartley et al. 1991). genesis. The aim of the present paper is Zeolite formation appears to be controlled therefore to fill this gap by providing an partly by the presence of suitable detritus, such overview of fluid evolution during red bed as volcanic grains of suitable composition. In modern red beds the interdigitation of diagenesis. We aim to identify the major controls on fluid chemistry, to identify areas alluvium with fanglomerates and sabkha evapwhere our understanding is poor, and to point to orite deposits is common. This leads to a gross approaches which might be used to address these chemical heterogeneity of the deposit which in turn leads to the development of a number of areas.

CONTINENTAL RED BED DIAGENESIS

303

Basin Evaporation

J

Inter-Basin Highland

Evaporation l

~

..... , oana uunes

~

{'N,.Cl brines'T ~,noe-- "~---"NaCI'7~"~aSO4 k ~ s . p .... 3 o o ~ , - 1 ~ ~

~ (

Saline Lake . o bnno:T 25°c ' IpH= c.10;Eh . . . . 200 mV]

-- . . ~ e c h a r g e

CaM~(COa)o - ~ ~ - ~ C a - N a - I - l ¢ O ""--

~

~

~

°o

3 fresh..... Y= 50°C~ I kTDS
)

,_..",J \ \ "

Y~

--:::::::::::::ii:.,v'/

/

....

I

/-,

~'~ 'o'~ ~

I--1 ~ .. [___/ ~anos~one Conglomerate

I_..,

10s- 100s km

,.._1

Fig. 1. Schematic model for the evolution of red bed formation waters showing how contrasting formation water chemistries may occur in different places within a red bed sequence. The exact evolution of the formation waters will depend upon a wide variety of factors, such as the composition of the crystalline basement source areas of the sediments, occurrence of marine incursions, and fluid mixing. One possible evolutionary pathway for fresh, atmosphere-equilibrated recharge waters (A) is illustrated. In this, evaporative concentration of porewaters leads to calcite or dolomite precipitation thereby increasing Na/Ca, Na/Mg and (SO4 + Cl)/ (CO3 + HCO3) ratios, leading eventually to gypsum precipitation. This increases Na/Ca and C1/SO4ratios, and possibly causes development of Na-Cl waters. TDS = Total Dissolved Solids (based on Jowett 1989).

different formation water types (Fig. 1). For example, calcium carbonate waters may develop where meteoric waters weather silicate and carbonate minerals, but calcium sulphate and sodium chloride waters may evolve where there is dissolution of evaporite minerals. In general halite is the dominant constituent of both marine and non-marine evaporites with gypsum also occurring in both environments. Carbonate minerals are only a minor feature of marine evaporites because seawater has CI- and SO42as the dominant anions, only c. 120mg1-1 HCO3- and Ca 2+ + Mg 2+ > HCO3-, so that evaporating seawater never becomes bicarbonate-rich. However, continental evaporites often precipitate from HCO3--dominated waters leading to the precipitation of a wide range of carbonate minerals including trona (NaHCO3.NazCO3.2H20) and burkeite (Na2CO3.2Na2SO4) (Eugster 1980). It is important for comparative purposes to quantify the redox conditions of eodiagenesis. The redox conditions of diagenetic waters are often expressed in terms of redox potentials (Eh) or oxygen fugacities (/'o2). However,

natural waters rarely approach electrochemical redox equilibrium so that Eh values are difficult to interpret in terms of overall system redox states. Additionally, foz values do not always have a clear physical meaning, and although in groundwaters Eh values can often be measured, fo2 values must generally be calculated from analytical data. Furthermore, foe and Eh values do not show a simple correspondence, and for a given aqueous system, both parameters will change with temperature making comparisons between redox conditions at different temperatures imprecise. In spite of these problems, both Eh and fo2 are of practical value if used as semi-quantitative indices of redox, and enable general comparisons to be made between redox conditions in different environments. Eodiagenetic waters will include some of the most oxidizing red bed formation waters, and organic-free waters at 25°C will have redox states up to approximately Eh ~ +500 mV, due to buffering by the atmosphere, with a redox state represented by 10~eO2,bars = - 0 . 7 . However, it is likely that eodiagenetic waters originating in density-stratified saline lakes may

304

R. METCALFE E T AL.

become locally very reducing, with Eh = - 4 0 0 to - 5 0 0 m V , due to the trapping of organic matter in bottom brines (Sonnenfeld 1984). The pH conditions are likely to be quite variable, generally lying in the range 6-10, which is typical of near-surface waters (Stumm & Morgan 1981), but possibly reaching >10 in continental evaporating lakes (e.g. Sonnenfeld 1984; Darragi & Tardy 1987). In such continental settings, HCO3- is often the most abundant anion and evaporation may lead to the precipitation of carbonates and smectites, thereby reducing the Ca and Mg content of waters to the point where they are N a - K - H C O 3 - dominated. When CO2 is independently buffered, for example by the atmosphere, then the pH is governed by: HCO3- + H + = CO2(g) +

H20

6- '

~

Q

u

a

r

t

z

.

Saturation

5-

7" 4

~K-feldspar t

\,,

E +

M 3

........"\-'~,

2

A~

'

[

", g?o

i

o

"',,

(1) ]Pyrophyllite]

fc°2[H20]

[r~CO3-1[U+l

= Constant

(2)

This means that pH may reach >10 because as [HCO3-] increases during evaporation, [H +] must decrease. Where waters were of marine origin, and in the absence of microbial activity, pH conditions will generally have decreased as ionic strengths increased during evaporation, owing to changes in the activity coefficients of the carbonate species which buffer pH (Krumgalz 1980). For example, during atmosphere-equilibrated seawater evaporation at 25 ° C, the pH falls from 8.2 to 6.7 at the point when halite saturates. However, microbial processes, such as sulphate reduction may lead to an increase in pH to 8-9 (Sonnenfeld 1984). Burial diagenesis (mesodiagenesis). As a red bed is buried, it becomes isolated from near surface conditions and undergoes burial diagenesis (mesodiagenesis) with formation waters becoming progressively more reducing and of generally lower pH. It is characterized by continuing framework grain dissolution, notably of K-feldspar, plagioclase, biotite etc, transformation of early diagenetic minerals to higher temperature minerals, and precipitation of higher temperature authigenic minerals. Early diagenetic smectite transforms to iilite, leading to the liberation of aqueous silica, Fe and Mg, according to reactions such as: 1.84Ca0.025N aoiK02Mg1-15Fe2+o,sFe3+0.2AI3+ 1.25Si3.50,o(OH)2

[Smectite] + 0.232K+ + 6.72H+ = K0.6Mg025AI23Si3sO,0(OH)2 [Inite] + 2.94SiO2(aq) + 0.92Fe2+0.368Fe3+ + 1.866Mg2++ 0.046Ca2+ + 0.184Na+ + 4.2H20. (3)

0

~

-6

r

~

1

-5

-4 -3 Log [SiO2(aq)]

-2

-1

Fig. 2. Logarithmic mineral stability diagram for the K20-SiO2-AI203-H20 system, showing how variations in K ÷ and SiO2 (aq) activities, and pH in diagenetic porewaters can be a major control on authigenic phases during red bed diagenesis. Phase boundaries were calculated for 25° C and 100° C using the SUPCRT92 code (Johnson et al. 1991). As K ÷ activities increase due to mineral dissolution during burial, red bed formation waters will move generally from the kaolinite stability field (A) to the muscovite stability field (B), and then to the K-feldspar stability field (C). Subsequent influx of dilute meteoric, or acidic basinal waters may then move the fluid composition back towards the kaolinite stability field. This leads to precipitation of authigenic quartz and feldspar overgrowths by the formation waters (Fig. 2; Bath et al. 1987; Strong & Milodowski 1987; McBride 1989). The formation of quartz cement in red beds is important because it is a major control of porosities and permeabilities, and hence influences fluid flow during diagenesis. In quartz-rich sandstones authigenic quartz precipitation may be the main factor causing porosity reduction (Leder & Park 1986). The various factors affecting quartz overgrowth formation have been reviewed by McBride (1989) who concluded that the inter-relationships between burial depths, temperatures and diagenetic mineral transformations generally cause quartz precipitation at temperatures of 60-100 ° C and depths of 1-2 km. Such depths coincide with the maximum theoretical formation water flow velocities of the order of 10 cm a -1 (Leder & Park 1986).

CONTINENTAL RED BED DIAGENESIS A consideration of the mechanisms of formation of such cements also aids our understanding of the volumes of fluid which flux through red beds during diagenesis. A major problem in explaining the formation of quartz cements in sandstones in general (not just red bed sandstones) is that there is often >10% imported silica and it is difficult to explain the large fluid fluxes required to achieve this. This subject has received considerable discussion in the literature (e.g. Leder & Park 1986; Morad & Aldahan 1987; Houseknecht 1987; Dutton & Land 1988; McBride 1989). The origins of the silica are generally not known, but likely contenders include dissolution of quartz, for instance by pressure solution, and liberation of silica during mineral reactions such as (3) above. At present there is no generally accepted source for the silica, but there is some agreement that very large volumes of water must flux through a given volume of rock and estimates range from 104-106 pore volumes. Additionally, it seems that silica must be transported over distances on the scale of kilometres in order to link possible sites of silica release to solution, to sites of cement production (McBride 1989). Burial diagenetic formation waters become progressively more reducing, and frequently reduce ferric iron. This is particularly the case when the red bed acts as a hydrocarbon reser~,oir, when conditions may become sufficiently reducing to cause the formation of pyrite from early hematite (Burley 1984). As the crystal structures of dolomite and calcite will accommodate Fe 2+ more readily than Fe 3÷, this causes late burial diagenetic carbonates to be typically ferroan, in contrast to early diagenetic carbonates (e.g. Burley 1984; Strong & Milodowski 1987). Modern deep groundwaters from red beds can place useful constraints upon the chemistry of mesodiagenetic waters. Waters from PermoTriassic red beds in the Wessex Basin, UK, reach temperatures of more than 75°C (Smith 1986) and can be considered diagenetic waters. Most deep waters in Permo-Triassic red beds of the UK are of Na-C1 type, while salinities are highly variable, but may approach halite saturation (>300 gl -a total dissolved solids (TDS); Edmunds 1986). The dominant origin of the salinity appears to be halite dissolution within the Permo-Triassic (Edmunds 1986), and it seems likely that this process has been a major control of water salinity throughout diagenesis. Neglecting localized redox variations, within individual organic-rich beds for instance, red bed sequences as a whole are likely to encounter the most reducing conditions during late burial

305

diagenesis, due to introduction of hydrocarbonbearing waters which originate in hydrocarbon source-rocks outside the red bed sequence (Fig. 1). The redox state of such waters can be estimated by assuming equilibria involving silicates to control the pH, and that equilibria involving light organic acids buffer the redox state. For example, if the activity of K ÷ is known, then the pH can be constrained by using equilibria such as: 2KA12(A1Si3Oa0)(OH)2 + 2H + + 3H20 Muscovite = 3AI2SizOs(OH)4 + 2K + Kaolinite

(4)

Using the calculated pH, the redox state of the water can be constrained with equilibria such as:

CH3COOH(aq) = CH3COO-(aq) + H+(aq)

(5)

C2HsCOOH(aq) ~- C2HsCOO-(aq) q- H+(aq)

(6)

and 3CH3COOH~aq) = 2C2UsCOOU(aq) + O2(g) (7) Appropriate concentrations of K + can be estimated from published analyses of oilfield brines (e. g. Gulf Coast USA data from Carpenter et al. 1974). The activity coefficients of K + and light organic species can then be calculated using a suitable equation of state such as the HKF equation (Helgeson et al. 1981), and the equilibrium constants for the relevant equilibria can be obtained using the computer code SUPCRT92 (Johnson et al. 1991), and thermodynamic data for organic species from Shock & Helgeson (1990). This suggests that at a temperature of 125 ° C, reducing late diagenetic waters will be slightly acidic, with a pH of c. 5 and a redox state represented by 1OgfO2,bars = C. --50 (approximately equivalent to Eh = - 2 0 0 mV). Movement and mixing of formation waters from different parts of a red bed sequence during mesodiagenesis may exert an important effect upon diagenetic mineral assemblages. For instance influxes of dilute meteoric waters which have low K+/H + activity ratios will tend to produce diagenetic kaolinite, whereas higher temperature waters have higher K + concentrations due to mineral dissolution and will produce diagenetic illite (represented by muscovite: Fig. 2). Cooling of high temperature muscovite-equilibrated waters (point C for 100°C on Fig. 2) may give rise to authigenic K-feldspar (e.g. at point C for 25°C on Fig. 2) during ascent. Diagenesis during uplift (telodiagenesis) Telodiagenesis is characterized by progressively decreasing temperatures, and formation water

306

R. METCALFE ET AL.

salinities and by progressively more oxidising conditions. This is due to the influx of dilute neutral to low pH meteoric waters which displace basinal brines and lead to evaporite dissolution and the removal of sulphate and halite cements (e.g. Burley 1984; Parnell 1992). This also causes carbonate cement and feldspar dissolution, and leads to the precipitation of hematite due to oxidation of Fe 2÷ which is liberated from dissolving ferroan carbonates. With increasing dilution, SiO2(aq) and K+(aq) activities are depressed, leading to the precipitation of kaolinite (Burley 1984; Bath et al. 1987; Strong & Milodowski 1987; Parnell 1992; Fig. 2). Modern groundwater chemistry can be used as an indicator of late diagenetic mineral transformations. Edmunds et al. (1984) and Bath et al. (1987) investigated groundwaters in the Triassic Sherwood Sandstone red bed aquifer of the East Midlands, UK. They found that invasion of meteoric waters had flushed any pre-existing Na-C1 dominated saline waters from the aquifer, and that the groundwater chemistry was dominated by reactions with carbonate cement, detrital dolomite and sulphate minerals. The pH of pure water which is equilibrated with modern atmospheric levels of CO: is c 5.6 (Stumm & Morgan 1981) and can be used as an approximate estimate of the minimum pH of the late diagenetic waters. However away from outcrop, the pH will rise owing to the interaction between the meteoric water and diagenetic carbonates in the red bed. In the Sherwood Sandstone, pH increased from c. 7 to c. 8 at greater distances from the outcrop (Edmunds et al. 1984). However, in spite of this flushing, there is still a considerable redox gradient down-dip. Edmunds et al. (1984) found that only the waters in the unconfined part of the aquifer are sufficiently oxidizing to contain dissolved oxygen, and down dip from the outcrop, Fe 2÷ and then F e 2÷ and HS- become important in controlling redox potentials, as Eh falls from +300 mV to 0 mV.

Types o f mineralization associated with red beds

Economically, copper mineralization is perhaps the most important type of red bed ore deposit. However, other heavy metals also occur within red bed sequences, including important ore deposits of U, Pb, Zn and Ag.

The occurrence and tectonic setting of 'sediment-hosted stratiform copper deposits' has been reviewed by Kirkham (1989). He concluded that ore deposits contained wholly within red beds are of relatively minor importance, and that most major ore bodies comprise disseminated sulphides in reducing sediments, such as organic-rich shales which lie immediately above continental clastic red beds. These reducing sediments acted as traps, stripping metals from relatively oxidizing, evaporite-derived brines which had previously extracted heavy metals from the red bed sediments themselves. An important control on the formation of such ore deposits is thought to be rifting (Jowett 1989; Kirkham 1989) because this acts to: (1) create an initial source of heavy metals through riftrelated volcanism; (2) cause rapid faultcontrolled sedimentation; (3) allows the development of an internal closed drainage, causing the formation of saline brines and evaporites; and (4) leads to the development of a basin and range topography and the establishment of a hydraulic regime likely to channel oxidizing brines towards overlying reduced strata during late diagenesis. In red bed-hosted ore deposits there may be evidence for a redox-related zonation of ore minerals. In general, ore deposits have a more reduced mineral assemblage than the initial host red bed, reflecting the relative solubilities of sulphide ore minerals. For example, Cu deposits may show a zonation which reflects the sequence of increasing solubility of sulphides, similar to: hematite-bearing barren rock-native copperchalcocite (Cu2S)-bornite (CusFeS4)-chalcopyrite (CuFeS2)-galena (PbS)-sphalerite (ZnS)-pyrite (FeS2) (e.g. Gustafson & Williams 1981; Kirkham 1989; Jowett 1992), although there are variations in this pattern. In red bed mineral deposits, Cu is often accompanied by Ag, and sometimes by Co, Pb, Zn and other metals. However, Gustafson & Williams (1981) comment that Pb-Zn deposits are not nearly as well linked to red beds as are Cu-deposits, and that major Pb-Zn associated with Cu only occurs at Mt Isa, Australia. Red bed ore deposits have metal abundances: C u > Z n , Pb which contrast with those of Mississippi Valley Type (MVT) ore deposits (Pb > Zn -> Cu). Uranium deposits are also an important feature of some red bed sequences and stratiform, mostly tabular deposits of the Colorado Plateau, USA are hosted by red bed sandstones (Kimberley 1979; Durrance 1986, Chapter 6).

CONTINENTAL RED BED DIAGENESIS

Distribution of metals in red beds Distribution o f metals in whole rocks Relatively few whole-rock analyses of red bed sediments have been published. Zielinski et al. (1983,1986) analysed modern red bed sediments and showed that heavy metal variations in sediments which have undergone only early diagenesis are primarily controlled by differences in sources and/or mineral fractionation during deposition. From their analyses they concluded that in young (Holocene-Pliocene) red beds in northern Baja California, there is an early diagenetic redistribution of metals on an intergranular scale, from detrital host silicates to authigenic Fe-oxides. The abundances of these Fe-oxides were found to increase with increasing sediment age, so that although there is desorption of metals from Fe-oxides as the oxides become more crystalline with age, the rock remains closed to metal movement. The Feoxide abundances are also lithology-dependent, with finer sediments having relatively high concentrations of heavy metals (Wedepohl 1978; Zielinski et al. 1983, 1986). Additionally, the early diagenetic transformation of labile minerals other than Fe-oxides, such as the alteration of plagioclase, hornblende and biotite to smectite may also be a control on the retention of heavy metals by the bulk rock. The proportion of the metal content which is mobilized, and the distance over which it is transported, is a function of fluid flow in the red bed, which in turn is porosity- and permeabilitydependent. Thus it is to be anticipated that coarse, relatively porous and permeable sediments would liberate a higher proportion of their metals to formation waters than finer sediments. This is supported by the results of experimental heavy metal leaching from red bed sediments, which suggest that during burial diagenesis, under equivalent conditions, coarse sediments and young red beds probably undergo relatively rapid removal of metals, although older red beds may be a greater source of metals due to their higher Fe-oxide abundances (Zielinski et al. 1983, 1986). Further information regarding the integrated movement of metals through red beds during burial and uplift-related diagenesis (mesodiagenesis and telodiagenesis) can be gained from whole-rock analyses of ancient red beds. Some data for the bulk content of heavy metals in red bed sediments are presented in Wedepohl (1978), and from these it seems that red bed sandstones have mean abundances of heavy metals of: Pb = 9.9 ppm (533 samples); Cu =

307

10.0ppm (382 samples); Zn = 30.9ppm (64 samples); and U abundances generally <2 ppm. Recently, Haslam & Sandon (1991) have published a large number of analyses of red bed sediments from the U.K. They show that relative to mean crustal concentrations Cu is usually more depleted than Pb and Zn, and that occasionally Cu-depleted sandstones are also U-depleted. Some lithologies, such as some components of the Upper Triassic Mercia Mudstone Group have relatively Fe-poor (reduced) horizons with enhanced Cu. By means of a statistical analysis of their data, Haslam & Sandon (1991) concluded that variations in A1, K contents of the red beds reflect variations in clay (illite) content, and that Si (and to some extent Zr) reflect variations in the arenaceous content of the sediment. In contrast, Mg, Ca, Mn, Sr, and Ba generally vary independently and were concluded to reside mainly in authigenic minerals.

Distribution o f metals in minerals The abundances of heavy metals in various detrital minerals are given in Wedepohl (1978) and these data are summarized in Table 1. From this it can be seen that generally both Cu and Zn reside mainly in relatively unstable ferromagnesian minerals, whereas Pb occurs mainly in more stable feldspar. It is also apparent that whereas Cu, Pb and Zn occur in minerals which together form a significant proportion of a sediment, U is concentrated in minerals such as zircon and titanite which usually occur only as accessory phases. These are normally much more resistant to dissolution and alteration than are the Cu-, Pb- and Zn-hosting minerals. This mineralogical distribution of these metals means that the dissolution and alteration of the Cu-, Pb- and Zn-hosting phases will also exert an important control upon the gross chemical character of the formation waters, and will be related to the various diagenetic mineral transformations. In contrast, the dissolution and alteration of U-bearing minerals will not buffer the overall chemical characteristics of the formation water, but will instead occur in response to the gross chemical characteristics of the formation water which are imposed by different mineral-fluid equilibria. There are few data for the stabilities of these accessory phases, but it does appear that the order of stability of the various minerals is strongly controlled by the pH of the diagenetic waters. Apatite, garnet, titanite and spinel are expected to be less stable under acidic conditions ( p H < 6 ) than under more

308

R. METCALFE E T AL.

Table 1. The mean contents of Cu, Pb, Zn and U in minerals found within immature red bed sediments Cu (ppm) Mineral Pyroxene Amphibole Magnetite Biotite Muscovite Kaolinite Plagioclase K-feldspar Allanite Epidote Apatite Garnet Zircon Titanite Monazite

Mean

No. samples

130 95 81 39 80 251 227 752 36 12 23 13 71 108 Range <1-20 nr nr nr nr nr nr 33 61 nr nr nr nr nr nr

Pb (ppm)

Zn (ppm)

Mean

No. samples

Mean

No. samples

6 16 < 10 40 26 nr 19 140 nr nr nr nr nr nr nr

20 76 13 512 33 nr 61 791 nr nr nr nr nr nr nr

248 823 563 545 54 59 17 5 nr nr nr nr nr nr nr

28 60 93 672 19 50 23 3 nr nr nr nr nr nr nr

U (ppm) Mean

No. samples

3.6 nr 7.9 nr Range 1-30 8.1 nr Range 2-8 nr nr 2.7" nr 2.7* nr 200 nr 43 nr 65 nr Range 6-30 1330 nr 280 nr 3000 nr

Data from Wedepohl (1978). nr not reported. * Value reported for a combination of K-feldspar and plagioclase.

alkaline conditions (c. 6-11) (Nickel 1973; Morton 1984). Since oxidizing conditions favour U transportation (see below; Fig. 5d) these observations are consistent with U-transport in near surface low p H waters. The heavy metals released from unstable detrital ferromagnesian minerals during red bed diagenesis are initially co-precipitated with Fe-oxide coatings which form over detrital grains. The metals are both incorporated into the structure of the Fe-minerals to form solid solutions, and are sorbed onto the oxide surfaces. However, at present there are few data on the distribution of the heavy metals between surface sorption and structural crystallographic sites. Gerth (1990) investigated the incorporation of heavy metals into goethite in highly alkaline (sodium hydroxide) solutions and concluded that the proportion of incorporated metal decreases in the order Cu = Co > Cd > Zn >> Ni > Th >> U, and that up to c. 10% of the Fe-sites in the goethite may be substituted by Cu. However, these results may not be relevant to the distribution of metals in the actual Fe-oxides which occur in red beds (depending upon the stage of diagenesis Fe-oxides may be more or less crystalline than goethite), under the much lower p H conditions of burial and uplift related red bed diagenesis. The available data do, however, allow some conclusions about the proportion of the total metal content of a red bed which is located on

leachable sites rather than structural crystallographical sites. By conducting leaching experiments, in which whole-rock samples were exposed to progressively more stringent chemical leaching agents, Zielinski et aI. (1983) concluded that 10-63% of the metal content in the red beds is located in leachable sites on secondary Fe-oxides, and to some extent on minor Mn oxides, and clays. The abundances of metals on leachable sites increased in the order: Cr(10) < V(15) < Pb (t>20) < Zn(22) < U(27) < Fe(30) < Ni(41) < Cu(44) < Co(47) < Mn(63), where values in parentheses are average percentages of metals which were leachable. Their experiments suggest that distribution of metals between sorbed sites and the aqueous phase depends upon the crystallinity of the Fe-oxides, with aqueous partitioning being favoured by increasing crystallinity.

The nature of metal-transporting formation waters The characteristics of the waters required for the mobilization of copper and other heavy metals in red beds have been evaluated by Rose and co-workers (Rose 1976, 1989; Rose & BianchiMosquera 1985), Zielinski et al. (1983, 1986), Eugster (1985), Sverjensky (1984, 1987, 1989), Nash et al. (1981), Durrance (1986, Chapter 6), and Hofmann (1990, 1991, 1992). Geological

CONTINENTAL RED BED DIAGENESIS evidence suggests that the mineralizing fluids in the Kupferschiefer ore deposits were late diagenetic basinal brines which migrated through the Rotliegendes sandstones, leached metals from Fe-oxide mineral coatings, and ascended from flanks of basement highs. The brines deposited their heavy metal content as a result of encountering reducing conditions in the organicrich, pyrite-bearing Kupferschiefer shales and Zechstein limestone above. Jowett (1986) calculated that a groundwater velocity of 13 cm a -~ and a copper solubility of I000 mg 1-1 in 20-30% Ca-Na-C1 brines could have formed the Cu deposit at Lubin, Poland, in less than 6 m.y., but that convective recycling of water must have occurred in order to produce the deposit with more realistic Cu solubilities of a few milligrams per litre. The mineral zonation which occurs in the Kupferschiefer concurs with progressive increase of H2S within a relatively oxidizing fluid containing Cu, Fe and possibly Pb and Zn. This may occur through a combination of the addition of H2S, for example through fluid mixing, or by reducing SO42- which is already present in the solution. Textures indicate that organic matter or pyrite are common reductants (Gustafson & Williams 1981; Sawlowicz 1992), and water rock interactions involving these might act both to reduce oxidized sulphur already in solution, and to add reduced sulphur to the solution. Uranium deposits form where oxidizing meteoric waters encounter more reducing groundwaters as they move down-dip through sandstone aquifers, for example where there is organic-rich detritus. In general, under the relatively oxidizing conditions which favour U mobility, Fe is locked in hematite and the rock is characteristically coloured red. In contrast, Fe is usually mobilized as Fe 2+ under reducing conditions which favour U ore precipitation, leading to U deposits being hosted by drab grey to white coloured sediments. The redox boundaries which are associated with sediment colour changes in U ore deposits may extend for hundreds of metres. However, more localized variations in redox conditions are responsible for commonly observed, hematitedepleted, high U 'bleached' spheroids, generally from a few millimetres to a few tens of centimetres in diameter. The origin of such spheroids has been a subject of some debate, and localized reduction has been variously ascribed to organic detritus (e.g. Durrance 1986), and to bacterially mediated redox reactions involving dissolved reductants such as organic acids and methane (e.g. Hofmann 1990, 1991, 1992). The temperatures of ore-forming fluids in red

309

beds are not as well constrained as for other types of base metal ore deposits, such as MVT Pb-Zn deposits. However, a number of lines of evidence point to low-moderate temperatures of less than c. 150°C. The exsolution of chalcopyrite upon heating bornite above 75°C was used by Rose (1976) to infer low temperatures (<100 ° C) for Cu ore-forming fluids in red beds. Similar temperatures are indicated by the replacement of early diagenetic framboidal Fe-sulphides by Cu-sulphides in the Kupferschiefer (Sawlowicz, 1992), and by limited fluid inclusion data, which suggest temperatures as low as 60-70°C (Naylor et al. 1989). Slightly higher temperatures of c. 130 ° C are indicated by oxygen isotopic data for illites associated with Kupferschiefer mineralisation (Bechtel & Hoernes 1993). Indirect evidence for the likely temperatures of mineralizing fluids comes from fluid inclusions in other types of sedimenthosted ore deposits, such as MVT deposits which are thought to have a similar origin in basinal brines (e.g. Sverjensky 1989). These MVT deposit data suggest temperatures in the range 100-150 ° C, and occasionally up to 200 ° C (Roedder 1984). Burial depths during ore deposition in many sedimentary basins such as the Permo-Triassic basins of the UK often suggest maximum formation water temperatures between 100 and 200°C. In contrast, U-deposits which are associated with red beds are often formed at near surface temperatures. The theoretical constraints on the likely compositions of mineralising fluids have been assessed by Rose (1976, 1989). In summary, he concluded that the aqueous speciation of copper in the system C u - O - H - S and under oxidizing (Eh > 300mV) conditions, is dominated by Cu 2+, but the solubility was predicted to be >1 mg1-1 only at pH <6.2. He considered that, the carbonate complexes, CuCO3(aq) (pH 7.4-9.3), or Cu(CO3)22- (pH >9.3) may account for appreciable copper solubility in the presence of CO2. For Pco2 ranges of 10-15-10 -2.5 bars, malachite was predicted to be the stable mineral and at higher levels of Pco2 azurite (2CuCO3.Cu(OH)2) was predicted to be stable. Due to the abundance of halite-bearing evaporites in many red bed sequences, the dominant Cu-complexing ligand will be CI-. However, the C1- complexation of Cu 2÷ is very weak, and the solubility of Cu in oxidizing (Eh > 300 mV), Cl-rich fluids may be reduced due to the precipitation of atacamite (Cu4CI2(OH)6). In contrast, CI- forms very strong complexes, CuCl2-, CuCl3 2- with Cu + (the cuprous ion). These complexes are stable under more reducing conditions than Cu(II) minerals (malachite

310

R. METCALFE ET AL. I'lL

1.0

Io

I"

~e°2

I

I

I

I

-" 02

~ o ~

0.75 ~-

~

I

I I I PCO 2 = 0.01 b a r -

acu ___1]0-5

CuC]+

ac] =

/ 0.25 0

2

3

4

5

6

7

8

9 10 11

2

pH Fig. 3. Eh-pH diagram for Cu at 25° C. tenorite = CuO; cuprite = Cu20; chalcocite = Cu2S; covellite = CuS. The shaded field represents the most likely chemical limits of red bed brines which are capable of transporting Cu as Cu+-Cl - complexes. The field is defined by the area where hematite is stable, and is bounded by 9 > pH > 5 (considered a reasonable range for ore-forming fluids) and the fields of copper and chalcocite stability. 'A' is tlie area where hematite dissolution is possible at a pH within the range most likely to be encountered during diagenesis. The diagram was constructed using data from the EO3/6 thermodynamic database (Wolery & Daveler 1992; Wolery 1992a, b). (CuCO3.Cu(OH)2, tenorite (CuO), brochantite (Cu4(SO4)(OH)6), atacamite), and more oxidizing than native Cu or Cu-sulphides. Therefore, these cuprous Cl-complexes are probably the most important Cu-transporting species over most of the range of redox conditions encountered during red bed diagenesis. Complexes with CN-, F-, NH3(aq) are unimportant, whereas 5032- and 52032-complexes may be important in weathering conditions. The complex Cu(HS)32is important only at very high HS- concentrations because Cu-sulphides have low solubilities. Complexes with oxalate, formate, acetate and fulvate may be important in organic-rich pore fluids. The E h - p H conditions of most Cu-bearing red bed formation waters are compared with the fields of stability of common Cu-ore minerals in Fig. 3. The figure is constructed for a temperature of 25 ° C since reliable higher temperature data are unavailable for the aqueous Cu species which are shown. The shaded field on the diagram approximates the most common ranges of pH and Eh over which Cu is likely to be mobilized as Cl-complexes and is not intended to represent the total possible field of Cu mobility.

The lower pH limit is taken as 5, which is near the value predicted from silicate equilibria such as equation 4. above, and the upper pH limit is taken as 9, which is near the maximum value found in most groundwater systems (e.g. Hem 1989). The exact boundaries of the field are dependent upon water composition and temperature, but nevertheless it serves as an aid in comparing the conditions which favour Cu transport, with mineral stabilities in different compositional systems. Therefore, the field of Cu-mobility as cuprous chloride has been superimposed on E h - p H plots which illustrate Fe, Pb, Zn and U speciation (Fig. 4). Evaluation of the complexation behaviour of other heavy metals reveals that Fe is insoluble as hematite under most conditions required for Cu mobility (Fig. 4a), whereas Pb is soluble as PbC12(aq) (at pH <7; Fig. 4b), and Zn is soluble as ZnSOa(aq) (at pH <8.1; Fig. 4c). Silver behaves differently in being soluble under only the most oxidizing conditions as AgC12- (Eh > 200 mV). Therefore, Ag-rich Cu deposits may only form under more oxidising conditions. Uranium mobility is similarly favoured by oxidizing conditions, which allow formation of uranyl (UO22÷) complexes rather than uranous (U 4÷) complexes (Langmuir 1979; Nash et al. 1981; Durrance 1986, Chapter 6). However, in contrast to Cu, Pb, Zn etc., the most common ligand involved in U transport is generally carbonate (Fig. 4d), particularly in the presence of meteoric waters, although phosphate, vanadate, fluoride and silicate complexes may also be important. An important feature is that although common U-ore minerals are oxides, such as uraninite (UO2; and the variety pitchblende, U3Os), they form under pH-redox conditions which are similar to those under which Cu-, Pb- and Zn-sulphides form (cf. Fig. 3 and Fig. 4d). Sverjensky (1984, 1987, 1989) has considered the chemical evolution of brines during diagenesis and the possible effects upon metal mobilisation and transport. His thesis is that a single basinal brine may evolve chemically to become an ore-forming fluid for a range of different types of heavy metal sulphide deposits (Fig. 5). Central to this idea are the chemical buffering properties of the aquifer along which the basinal brine migrates. Sverjensky considered a basinal brine initially saturated with galena, chalcopyrite, muscovite, kaolinite and quartz but undersaturated with respect to sphalerite, with up to 1 mg1-1 Pb, 0.1 mg1-1 Cu and 5 mgl -I Zn. Passage of this fluid through a carbonate-cemented aquifer would produce a mineral deposit characterised by high Zn/Pb and

CONTINENTAL RED BED DIAGENESIS a.

311

be I ""-..I.....

Io

]

"'2o .......... ~..~r)

1.0 I"

I

I

I

....."7:,?...cf..°2"~o2,

I

I

I

I

l

r".-.l.

................... -~s =oo~/

/

0.75 R n~L~

..................aFe= 10- "1

Hematite

Io

1.0 I-

asiderite,calcite = 0.01-] PCO9 = 0.01 b a r ]

...................].

~--~ 025

I

I

.........";;.'.?/..f°2--o.. -20 ............ 2..2~)

/

I

I

I

I

/

apb 105 / acl = ] / =

.....1"~.....

/

0.75 I--

I

PCO2 =0.01 bar

I/

................a.Xs = 0.03--I

D__.°5Lebcl2(aq)l~7-~.~,!EI/ ,! . . . . . . .,4. . . . . FJAg#I -! ~0"25

"0

L

0

...

!

-0.25 V

......~2"~::'-'..,....Gale~a ~

I

-0"25 _05. r~Y~'"g~7"~'215 0CI i / ~ 1 ~~ . . ] Ferr°an . ~ [

2

3

~

~'-' ~';~ ...........

25oc -0.5~ 25°C 2 3 4 5 6 7 8 9 1 0 1 1 1 2

b(CO3)

4

,

,

5

6

pH

!

, ........................ , ........]..........i..........P'"'...~..

7

8

9

10 11 12

pH

C.

d.

.........L I o I I I 1.0 ".......... ~ fro. ;q;'3"-...~-' 0.26,~, .......... ~

I

.....................

. . . . . . . . . . 0.75 ZnSO4(aq)

l

o,

I

I

I

PCO,~ = 0 01 bar~a z n ="t V,,,-5 aCl = 1

1.0

azs -- .o3..................

0.75

0.2

~""....... o~,~

r-

, o .~

-.-~- ~

a~,oo.-~.-'o-

eo T..... ' r) "'co 1 " - ~ { ~ ) ......

I

o, LF

~r

UO22+

a ~ c = 10.3

, "............ [

Q

-

.;.._.].............

:

...............

-

0.2

-0.25

-0.25 . . . . . . . . . . . . . . . . . . .... .

-o.s

-05

2

3

4

5

6

7

8

pH

9

10 l i

12

~2ra.

-

I-

25°c

I

,

,

,

,

,

........ _.< .......... :::till e,

,

, ....... I...........

2

3

4

5

6

7

8

9

10 11 12

o

pH

Fig. 4. Eh-pH diagrams for Fe, Zn, Pb and U at 25° C. The field of 'Cu transport as CuCI3:-', which is superimposed on each diagram is the same as the shaded field in Fig. 3. (a) Fe (constructed using the EQ3/6 thermodynamic database; Wolery & D aveler 1992; Wolery 1992a, b); (b) Pb; phosgenite = PbCO3.PbCI2; cerussite = PbCO3; galena = PbS (adapted from Rose 1989); (c) Zn; smithsonite = Z n C O 3 ; sphalerite = ZnS (adapted from Rose 1989); (d) U; uraninite = UOz (adapted from Brookins 1984).

(Zn + Pb)/Cu ratios. Transport of the same fluid through a quartz-cemented sandstone could exhaust the buffering capacity of the aquifer relatively quickly such that the resulting mineral deposit would be galena-rich. Maintenance of the oxidation state near the hematite-magnetite buffer would prevent mobilization of copper (solubility of Cu in a basinal brine is c. 0.07 mgl-1 ; Fig. 6). However, if the basinal brine were to migrate through a red bed containing

hematite and anhydrite, the oxidation state and SO42-/H28 ratio of the fluid would be increased by reactions such as: CaSO4

= C a 2+ + S O 4 2 -

4Fe203 + HzS + 14H + = 8Fe 2+ + SO42- + 8H20

(8)

(9)

Migration of this more oxidizing fluid through a red bed could scavenge the necessary Cu, Zn

312

R. METCALFE E T A L .

Examples

~

Upper Mississippi Valley Zn-rich ore 4 Pine Point Zn > Pb >> Cu McArthur River

~Qu

I

Basinal Brine

Laisvall IPb-rich ore 4 Southeast Missouri P b > Zn >> Ca

artz Sandstone Aquifer System

-../!i/////////~/~/,

Source

to Reducing

i

7//////////,~/..

~/1

Lubin

I Water in ~/'~ Burial Diagenetic ~ ' / / C u > Pb + Zn [.., Fig. 8 _ ~ Porewater Modelled in Fig. 7 / ~ 4 l~"

~

I ~'=

White Pine Creta

~I

Fig. 5. Diagrammatic representation of the reaction of a single source of basinal brine with three different aquifer systems, resulting in Zn-rich, Pb-rich and Cu-rich deposits (adapted from Sverjensky 1989).

y

i -42 -44 -46

o

ha

-48 -50 -52

°

"-

-54

-

61 -32

i -28

I1(I

i

-24 -20 log fs2,bars

I

I

-16

-12

T = 125°C Fig. 6. Logfsz,ba,sv e r s u s Iogfo2,barsat 125°C (adapted from Sverjensky 1987). The relative stabilities of Feand Cu-bearing solid phases are shown (hematite = Fe203; magnetite = Fe304; pyrite = FeSz; chalcocite = Cu2S; bornite = CusFeS4; chalcopyrite = CuFeS2; and covellite = CuS). The solid diamond indicates thefoz-fs2 composition of a hypothetical basin brine. At physical conditions in the shaded area, at least 35 mg 1-1 Cu and SO42- can be transported by fluids saturated with respect to hematite and chalcocite.

and Pb to form a mineral deposit with much greater Cu contents [Cu > (Pb + Zn)]. Redox conditions of logfoz,bar~ > --46 at 125°C are necessary to mobilise copper as cuprous chloride complexes (Fig. 6). Bleaching of red bed sandstones is a feature of some red bed hosted ore deposits (e.g. Ixer & Vaughan 1982) and is a result of hematite de-stabilization and dissolution. Dissolution of hematite requires fluids more reducing and/or of lower pH than most of those waters capable of mobilising significant cuprous chloride (compare Figs 3 and 4a). Similarly, bleaching due to the reduction of hematite to magnetite would require conditions more reducing than the field of significant Cu mobility (logfo~,ba,s < c. - 5 0 . 3 ; Fig. 6). Relative to mean continental crustal concentrations, some red beds are depleted in Cu, but are undepleted in Zn, Pb and Co (Rose & Bianchi-Mosquera 1985; Haslam & Sandon 1991), while ore deposits which are associated with red beds contain variable proportions of these metals (Gustafson & Williams 1981). These observations contrast with thermodynamic predictions which suggest that Cu, Pb, Zn and Co should all be soluble at similar pH and in brines with similar C1- concentrations, and implies that additional factors must have influenced metal mobility. One possibility is that sorption exerts a control on metal behaviour, and this possibility is considered further below. Since red beds are important hydrocarbon reservoirs in many parts of the world, organic species are likely to have been an important

CONTINENTAL RED BED DIAGENESIS component of red bed diagenetic fluids. The organic fluid component is likely to exert a major control on red bed ore deposits by directly controlling mineral stability, and by controlling the speciation of metal cations in solution. Increasingly, attention is being drawn to controis on cation mobility and secondary mineral stability by these organic components (e.g. Surdam et al. 1984; Manning 1986; Hennet et al. 1988; Thornton & Seyfried 1987; Seewald et al. 1990). Liquid hydrocarbons may transport cations such as V, Ni, Cu and Zn, (Manning 1986), while Ca, Mg, Fe, AI, Sr, Mn, U, Th, Pb, Cu and Zn may be transported as complexes with light mono- and di-carboxylic acids (e.g. Giordano & Drummond 1991; Holm & Curtiss 1990; Fein 1991; Harrison & Thyne 1992). Furthermore, it appears that the inorganic chemistry of aqueous solutions will exert a control on the way in which the organic component behaves with, for example, high Ca concentrations leading to formation of Caoxalate and prevention of oxalate complexes transporting metal cations (Hennet et al. 1988; Huang & Longo 1992). The role of organic acids in controlling mineral stability has been a subject of some debate, and it is uncertain, for example, whether the primary role of the acids is to increase reaction rates (e.g. Stoessell & Pittman 1990), or whether metal-organic complexing enhances mineral dissolution (e.g. Surdam et al. 1984; Surdam et al. 1989). However, it seems that organic acids may buffer the pH and redox states of fluids (e.g. Lundegard & Land 1986; Shock 1988; Huang & Longo 1992), and therefore these acids may have the capacity to influence metal-inorganic ligand complexing even where metal-organic complexing is unimportant.

Discussion The importance o f p H and redox state in red bed diagenesis and mineralisation In general, the diagenesis of silicate and carbonate minerals is influenced greatly by pH, and only to a relatively small degree by redox state, whereas the distribution of heavy metals is controlled by both pH and redox state. This is particularly important in the case of red beds because during diagenesis the formation waters of these sediments acquire a wide range of pH and redox states. However, unlike the solute content of formation waters which can be constrained from groundwater analyses from

313

deep boreholes, and diagenetic mineral assemblages which can be examined directly, pH and redox states are usually inaccessible to direct measurement. This is because of problems associated with making measurements in highly saline waters, and the inherent difficulty of collecting groundwaters without perturbing their in situ pH and redox states, due to factors such as degassing and interaction with steel drilling equipment. The pH and redox states of red bed formation waters must therefore usually be estimated theoretically and/or by using experimental data. In order to illustrate the problems which are encountered when doing this, and to assess the significance of fluid mixing as a pH and redox state control during red bed diagenesis, three important processes are considered: mineral-fluid equilibria; fluid mixing; and sorption.

Mineral-fluid equilibria Equilibrium thermodynamic models are commonly used to simulate mineral-fluid interactions. As an example, the code EQ3/6 (Wolery 1992a, b; Wolery & Daveler 1992) has been used to model red bed diagenesis according to the scheme in Fig. 7. The diagenetic modelling used an approach similar to that of Bruton (1989), and the initial, pre-diagenetic sediment composition was estimated by assuming that a typical red bed sequence is a representative sample of the bulk continental crust. Accordingly, estimates of the bulk crustal mineral composition, with mean crustal abundances of 25 ppm of Cu, 20 ppm of Pb and 70 ppm of Zn (from Taylor & McLennan 1985) were used. Tenorite (CuO), cerussite (PbCO3) and zincite (ZnO) were considered to contain these metals because there are only limited thermodynamic data for other oreforming minerals. The initial compositions of the diagenetic waters were considered to be buffered by the atmosphere, which constrained their initial CO2 content at a value represented by logfco2,bars = -3.5. The initial pH was taken to be 8, near the middle of the range observed for near surface waters in modern arid environments. In common with the majority of natural saline waters, the initial waters were taken to be chloride dominated. In such waters Na + and Ca ~+ are the most common cations, and fluid inclusion studies of a wide range of MVT ore deposits and red bed evaporites suggest that ore fluids also typically have Na and Ca as the dominant cations (e.g. Roedder 1984). Therefore, the initial computer simulations used a 1:1 (molar ratio)

314

R. METCALFE E T A L .

Late Diagenesis

With Anhydrite

Late • Diagenesis

Earl'

Solids

I Late Burial I Diagenetic Porewater

Fig. 7. Schematic illustration of the approach adopted when modelling red bed formation waters during diagenesis, and calculating the composition of a late burial diagenetic red bed porewater. N a : C a chloride solution in order to approximate the initial diagenetic water. The proportions of rock and water will play a major role in controlling diagenesis. The simplest model assumes that the system is closed, and that the porewater-rock system reaches equilibrium. In the example, the water/ rock volume ratio was taken to be the same as the percentage porosity of the rock (a value of 25% was used). In nature open system conditions will generally prevail, but models of these require significantly more complex computer codes than are required to model closed systems. A major uncertainty is the degree of equilibration between any moving fluid, the red bed aquifer through which it passes, and any minerals which it precipitates. This is a function of reaction kinetics and the rate of fluid movement through the rock, which are coupled since mineral precipitation or dissolution will be an important control on rock permeability. It can be concluded that diagenetic temperatures during ore deposition are usually < c. 150°C (see above). Since maximum burial depths can often be inferred from stratigraphical evidence and diagenetic mineral assemblages to be <4 km, the maximum burial pressures are often < c. 1 kbar. Although many codes, such as EQ3/6, cannot use pressure as an independent variable, mineral equilibria during red bed diagenesis will be relatively insensitive to pressure. In the example, early diagenesis was simulated for a temperature of 25° C and 1 bar, whereas later diagenesis was simulated at a

temperature of 125°C and a pressure of 2.32 bars (the pressure being fixed by the liquid water/water vapour equilibrium curve at 125° C). Equilibrium modelling which neglects reaction kinetics and which allows all thermodynamically stable phases to 'precipitate' during a simulation, tends to produce highly unrealistic mineral assemblage. For example, in the absence of user-defined constraints on precipitation, the EQ3/6 model predicted numerous geologically unreasonable zeolites to form during early diagenesis. A knowledge of the actual mineral phases which occur in red beds, must therefore be used to decide what minerals are actually allowed to 'precipitate' during a simulation. The only minerals which were allowed to appear in the early diagenetic mineral assemblage in this case were those listed in Table 2, plus tenorite (CuO), cerussite (PbCO3), zincite (ZnO), cuprite (Cu20), chalcocite (Cu2S), covellite (CuS), galena (PbS) and sphalerite

(ZnS). In order to simulate later, deeper, higher temperature diagenesis in the presence of red bed evaporites, the modelled early diagenetic waters were 'equilibrated' with anhydrite. The early diagenetic mineral assemblages were 'reacted' at this higher temperature with the resulting model porewater to yield a model late diagenetic porewater (Fig. 7; Table 3). This approach gave a reasonable representation of early (eodiagenetic), and early burial diagenetic (early mesodiagenetic) mineral assemblages, characterized predominantly by

CONTINENTAL RED BED DIAGENESIS

315

Table 2. Summary of the example EQ3/6 model of the mineralogical evolution of red beds during eodiagenesis

and mesodiagenesis

Mineral component

Initial detrital sediment composition (moles)

Sediment composition following eodiagenesis (25 ° C) (moles)

Sediment composition following mesodiagenesis (125 ° C) (moles)

9.2 28.7 3.2 1.6 6.2 1.1 0.3 0.2 0.6 (Ba in feldspar) np np 0.9 np

1.8 39.3 np np np 6.0 np np np 3.0 x 10-2 6.2 1.3 3.7 2.9

np* 39.3 np* np np 6.0 np np np 3.0 × 10 -2 6.2 1.3 3.7 2.9

Albite Quartz K-feldspar Annite Anorthite Muscovite Clinochlore-14A Tremolite Hedenbergite Barite Calcite Dolomite Hematite Na smectite

The listed abundances of minerals were produced during closed-system diagenesis of a water-saturated mass of 8 kg of red bed with a porosity of 25%. np, not present. * Mesodiagenetic feldspars are not predicted, possibly owing to model formation waters being dilute. Table 3. Summary of the initial water compositions, and product fluid compositions for the example EQ3/6

model of mixing between late burial diagenetic red bed porewater and reducing oilfield water to the point where all Cu is stripped from the solution (waters mixed in the proportions 95:5)

Component Na, mg 1-1 Ca, mg !-~ K, mg 1-1 Mg, mg 1-1 SIO2, mg 1-1 CI, mg 1-1 HCO3, mg !-1 S, mg 1-1 Fe, mg 1-1 Ba, mg 1-1 A1, mg 1-1 pH

Iogfo2,bars

Late burial diagenetic porewater at 125° C (A) composition 24385 t09 81 3 81 40239 395 738 ~1 0.3 0.04 6.5 -0.70

Reducing oilfield water at 125° C (B) Constraint

Composition

Product water on mixing (A and B in proportions 95 : 5)

Albite Calcite Illite Clinochlore-14A Quartz Charge balance Oilfield brine Pyrite Daphnite-14 Oilfield brine Anorthite Silicate equilibria (4) Organic acid equilibria (5-6)

591 3586 0.5 × 10-5 531 37 8669 4000* 0.25 9 55 t 2 5 -52

23700 200 78 14 80 39300 473 716 0.04 0.33 0.02 6.2 -46

* Acetate analysis from Carothers & Kharaka (1978) , From Carpenter et al. (1974)

quartz and calcite, with smaller quantities of h e m a t i t e , mica and clay phases (Table 2). H o w e v e r , certain features of the m o d e l are inconsistent with o b s e r v e d m i n e r a l parageneses. N o t a b l y , late burial diagenetic feldspar overgrowths w e r e not predicted. This is b e c a u s e the waters n e v e r a t t a i n e d realistic salinities, which m e a n t that feldspars n e v e r r e a c h e d saturation.

T h e m o d e l b e c o m e s increasingly i n a c c u r a t e for waters with ionic strengths > c. 1 molal, although for the range of w a t e r compositions c o n s i d e r e d h e r e , r e a s o n a b l e results w e r e given for ionic strengths up to c. 2 molal. V a r y i n g the salinity of the starting w a t e r up to this value had little effect on the d i a g e n e t i c m i n e r a l assemblage. H o w e v e r , in late burial diagenesis,

316

R. METCALFE ET AL. Reducing Oilfield Water

Published Oilfield Brine Analyses

Concentrations

Organic Acid

rlllllll~ K + (aq) II Concentrations

[

Alkalinity: 4000 mg !-1 Acetate Ba: 55 mg 1-1

Silicate Equilibria

Log fo2, bars = -52, pH = 5

Mudrock Mineral Asseml

Fig. 8. Schematic illustration of the approach adopted when modelling reducing organic-rich oilfield waters,

such as those which might enter a red bed aquifer. evaporite dissolution may result in ionic strengths significantly >5 molal (at halite saturation at 25 ° C ionic strengths are c. 7 molal) and therefore late burial diagenesis cannot be modelled accurately. Although the specific ion interaction model (Pitzer 1973) provides a theoretical framework for modelling such concentrated waters, the necessary data are extremely limited. A lack of Pitzer coefficients for silica means that silicate equilibria cannot be simulated. Furthermore, the data are applicable only over limited temperature ranges, with data for carbonates, for example, being limited to 25 ° C. When pH and redox states were allowed to drift from their initial values, it was found that Iogfo2,bar~ remained almost constant during burial diagenesis, whereas pH fell by about 1.5 units. This is due to the model calculating that all Fe 2÷ in the initial sediment was oxidised during early diagenesis under atmosphere-buffered conditions, so that during later diagenesis oxygen was not consumed by Fe-oxidation. In nature it is more likely that a significant proportion of the Fe will remain in the ferrous state until after burial has isolated the sediment from the atmosphere, and more reducing conditions will evolve as Fe-oxidation depletes oxygen in the formation waters. A further major problem is the prediction of geologically reasonable concentrations of metals in solution. In the example, only very small proportions of the available copper in the red bed were predicted to be released to solution (c. 1% for 1 M C1- solution), whereas all the

available Pb and Zn were liberated from the model rock. This contrasts with the experimental simulations of diagenesis performed by Zielinski et al. (1986) which indicated that up to c. 45% of Cu, but only c. 20% of Pb and Zn should be available to the solution. These experimental data suggest that 1 kg of sandstone could liberate up to c. 70 mg Cu, c. 20 mg Pb and c. 40 mg Zn (taking values for metal concentrations in sandstones from Gustafson & Williams 1981). One possible explanation is that because the model formation waters remained highly oxidizing, as discussed above, Cu is effectively locked in tenorite (Fig. 3). For this reason, additional models were run with more realistic reducing redox states (to lo~fo2,bars = - 4 0 ) being specified for the later diagenetic waters. However, these gave similar results whatever redox states were selected, with Cu concentrations remaining at c. 0.01 mg1-1, and Pb and Zn concentrations remaining at c. 20 mg 1-1 and c. 70 mg 1-1 respectively. Alternatively, the inconsistency with experimental results may possibly be due to: a poor knowledge of the actual phases which control the solubility of Pb, Zn and Cu during diagenesis; and a lack of thermodynamic data, especially high temperature data, for potential phases which might control Pb, Zn and Cu solubility. Fluid m i x i n g In order to assess the importance of fluid mixing it is necessary to estimate the proportions of fluid of different redox states which are necessary to

CONTINENTAL RED BED DIAGENESIS produce significant changes in the aqueous chemical speciation and precipitated mineral assemblages. It is instructive to consider the amount of reducing fluid which is required to mix with an oxidised, metal-bearing red bed formation water, in order to strip it of heavy metals owing to sulphide formation. The most reducing burial diagenetic water which can be envisaged is an oilfield water with a redox state imposed by organic matter. Such a redox state is appropriate for many reduced sediments which are actually associated with red bed Cu mineralisation, for example the Kupferschiefer. These conditions were modelled as illustrated in Fig. 8, by assuming that silicate equilibria controlled pH, organic acid equilibria controlled redox states (see equations 4-7 above), and that the water was equilibrated with a typical mudrock mineral assemblage such as might occur in a hydrocarbon source rock: plagioclase; illite; chlorite; calcite; quartz; galena; sphalerite; pyrite; and chalcopyrite. Salinity was specified by making the simple assumption that the solution was charge-balanced with CI-. Other parameters were fixed by using oilfield brine analyses as a guide to reasonable values. For instance, alkalinity was fixed by assuming 4000 mg 1-1 acetate (near the maximum concentration reported for oilfield brines by Carothers & Kharaka 1978). This gave a model reducing oilfield water of the composition shown in Table 3 and with a metal content of: 1.3 x 10 -3 mgl -~ Pb; 1.8 x 10 -2 mgl -~ Zn; and 6.3 x 10 -9 mgl -~ Cu. When the formation water is at its most oxidized the maximum amount of such a reducing oilfield water will be required to remove heavy metals completely. This limiting case was investigated by using the EQ6 code to model progressive addition of the reducing oilfield water to the model atmosphere-equilibrated, late burial diagenetic porewater described above. The model was terminated when sphalerite appeared as a product since the common mineral zonation in red bed ore deposits suggests that sphalerite generally occurs late in the sulphide paragenesis (e.g. Gustafson & Williams 1981). The results of this model are illustrated in Fig. 9 and Table 3, and show a sharp redox boundary when the reducing oilfield water constituted about 5% of the mixture. At this point, all of the Cu was removed from the solution, reflecting the insolubility of Cu-sulphides compared to Pb- and Zn-sulphides (Fig. 9a). This coincided with a redox boundary at which the water became dramatically more reducing, with Iogfo2,bars decreasing from c. - 5 to c. - 4 5 (Fig. 9b). The concentrations of Pb and

317

1oo!" ~

8o"

=

60-"

"~

40 :

I

cul

I

Pbl

a

I

20, o 100

10

g 80 -6 .=_ 60

I

.~

20

-cu

30

40

50 0

log fO 2

-]0 -20 ~

¢~ 4o

-30

20

-40

0 ~ # 3e-3 0 t~

. 10

20

~h muartz uscovite ematite lbite [ ] dolomite

~J

2e-3.

-°~'X:I3.---74>-~ -50 30 40 50 C

.¢///JJS

le-3. o

~0e+0 0

10 20 30 40 50 % of reducingoilfieldwater in mixture

Fig. 9. Illustration of the output from an EQ6 model for the mixing of evaporite-equilibrated late burial diagenetic porewater, and reducing organic-rich oilfield water. (a) Cu is stripped from the solution, and then Pb, followed by Zn begin to be stripped from solution as progressively larger amounts of reducing oilfield water are added to the mixture; (b) There is a dramatic change in redox state and decrease in Cu content when only c. 5% of the mixture is reducing, organic-rich oilfield water; (e) the model predicts a gangue mineral assemblage which is dominated by dolomite and barite, with trace quantities of quartz, muscovite, hematite and albite. The moles of minerals precipitating are those produced by a notional 1 kg of solution. Zn decreased only after all the Cu had been stripped from the solution, and the onset of Pb removal from solution commenced before the onset of Zn removal. Also, initially (during addition of the first 3% of reducing oilfield water) metal concentrations fell as a consequence of simple dilution (Fig. 9a). In the model, dolomite and barite were dominant predicted gangue minerals while quartz, muscovite, albite and hematite formed a minor proportion of the mineral assemblage (Fig. 9c). Trace amounts of albite and muscovite

318

R. METCALFE E T AL.

were present throughout the paragenesis, while trace quantities of hematite were predicted by the model initially, but disappeared when about 30% of the mixture was reducing oilfield water. Thus, the mixing model can reproduce the major features of many red bed-hosted copper deposits. It is significant that for this limiting case only 5% of reducing oilfield water is needed to strip all the Cu from the oxidized late burial diagenetic porewater. This implies that only a small amount of mixing between different fluids could be a major control of heavy metal mobility and ore deposit formation. The unrealistically small Cu concentrations in the initial late burial diagenetic porewater do not invalidate this conclusion because Cu-sulphides are relatively insoluble and in sulphur-rich fluids, such as those which are typical of evaporite-bearing red bed sequences, sulphur speciation is a major control on metal mobility. It is likely that formation water redox states more reducing than the atmosphere, and outside the stability field of Cu oxides (Fig. 3), would mobilize more Cu than in the model. However, such waters will also lie closer to the field of sulphide stability, so that smaller quantities of reducing water will be required to cause sufficient reduction of the system to result in Cu-sulphide precipitation.

Sorption

Equilibrium thermodynamic models are of limited applicability because they generally assume that mineral dissolution/precipitation equilibria govern the composition of the aqueous phase. However, equilibria between dissolved and sorbed species may be a major control on the composition of diagenetic waters, particularly in the case of trace metal constituents. The dominant sorbents in nature are the common hydrous oxides of Fe, AI, Mn and Si, and since Fe-oxides are a common constituent of red beds, sorption is likely to be particularly important in controlling metal mobility in red beds. Oxide and hydroxide surfaces are capable of adsorbing or dissociating H + from a surface O H group to acquire a net surface change, according to: S - OH + H + = SOH2 +

(10)

S - OH = SO- + H +

(11)

where S - O H indicates surface sites occupied by OH. These charged sites affect the electrostatic attraction for other ions on or near the surface

of the solid and create sites for further reaction for example by: S-OH+Me

+=SO-Me ++H +

(12)

where Me + is a metal cation. Ions bound only by the electrostatic forces are said to be non-specifically sorbed whereas those bound at surface sites are specifically sorbed. The abundance of these SOH2 + and SO- charged sites is solution pHsensitive, and hence the proportion of a metal cation which is specifically sorbed is a function of pH (Fig. 10), which in turn is a function of other equilibria, such as 4. Specific adsorption of cations onto oxide surfaces increases from low values to complete adsorption over a relatively small pH range to form an 'adsorption edge'. At low pH where the surface sites are generally positively charged sorption of metal cations is low, whereas at higher pH sorption increases and can result in the 100% retention of metals. At intermediate pH values sorption may cause retardation, but not complete retention of a metal. Factors other than pH which affect sorption are: (1) concentration of the metal, which will affect sorption as the number of occupied sites approaches a significant proportion of the total sites; (2) competing cations which will reduce the number of sites available to the cation of interest; (3) redox state, which controls aqueous speciation and the solubility of sorbing phases; (4) temperature, sorption generally decreasing as temperatures rise; (5) the sorbent/sorbate ratio. The importance of sorption during red bed diagenesis has been evaluated by calculating the position of the sorption edge for Cu, Pb, Zn, Co and Ag when these are sorbed by hydrous ferric oxide (FeOOH). The H Y D R A Q L code (Papelis et al. 1988) was used in conjunction with a diffuse layer model, surface complexation data from Dzombak & Morel (1990) and the HYD R A Q L database for solution species (Papelis et al. 1988). An oxide surface area of 600 m 2 g-~ was assumed and the surface densities of strong and weak sites were taken to be 0.005 and 0.2molmol -~ Fe respectively (Dzombak & Morel 1990). The effects of varying trace-metal concentration, ionic strength and surface site concentrations on the proportion of metal which is sorbed were investigated. Redox conditions were within the stability field of Cu 2+, owing to a lack of data for the cuprous ion. Under these conditions, increasing the ionic strength from

CONTINENTAL RED BED DIAGENESIS

i ~.;~,;~;~;;;I;;~S.;;;~;;~;;~;~;;;~;;;;1;;

1.0 : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : ii!iiiii!?iiiiiiiiiiiiiii?iiiiiiiiiiii~ii :::::::::::::::::::::::::::::::::::::::::

0.5

.

.

.

.

.

.

.

.

.

.

.

.

I

I

I

I

I

4.0

5.0

6.0

7.0

8.0

1

..~

frp/ /

~40t/

,,/ ~

/

/

/

/ ,

.

7 7

4.0

/

5.0

6.0

/

~40

/

1(89g/l Goethite) 10.1 M NaCI

i

i

8.0

SurfaceArea= 53.4m2/1 (0.89g/l Goethite)

.

,' // /

/

~

0.1MNaCI

co/

,

/'

20

9.0

T = 25°C

Pb p~C~_¢~ u277' / Z n / ~J /

x~60-

--n---n /,~'

7.0

.

80"

] ISurface Area= 53400m2/1

i,'~ p

120

100I C

I

9.0

~'

.of; / j 3.0

Approx~ate F,e~d

o T = 25 C

/

z4

As CuC132-

77-71 AgCl2Field

120-r t b

1001 •-~ 801

Field r ~ Approximate Copper Transport ~ 1Silver Transport

T=25°C aAg=1 aZS = 0.03 acu = 10-5 -0.5 PCO2= 0.01 barac1= 1 3.0

319

g

0 3.0

4.0

5.0

6.0

7.0

8.0

9.0

pH Fig. 10. A comparison between the fields of aqueous Cu- and Ag-chloride mobility in the absence of sorption (a), and calculated sorption edges for Pb, Cn, Zn, Co and Ag (b and c). The 'Approximate Field Copper Transport as Cu + - C1-1' in (a) is the same as the shaded field in Fig. 3. Sorption was modelled for Eh conditions consistent with Cu 2+ stability; Cu + is considered likely to sorb in a similar fashion to Ag +. It is evident that under the range of pH required for Cu- and Ag-chloride stability (a), significant proportions of Cu and Ag may be rendered immobile (b and c). In the presence of iron oxides, the fields of Cu and Ag mobility may be significantly smaller than the approximate fields of Cu- and Ag-chloride transportation in (a). 0.1 M to 1 M and varying the metal concentration by four orders of magnitude had no significant effect on the sorption curves. However increasing the surface area, or the density of sites available for sorption lowers the pH at which 100% of the metal is sorbed. Figure 10 shows the calculated graphs for sorption of Pb, Cu, Zn, Co and Ag (5 × 10 -7 M)

in 0.1 M NaC1 onto goethite with surface areas of 53400 and 53.4m z per litre of solution. The position of the pH edge is partly dependent on the sorbate/sorbent ratio. Similar graphs are obtained in each case, but the adsorption edge is shifted up by about 1.5 pH units in the lower sorbate/sorbent ratio case (Fig. 10b compared to Fig. 10c).

320

R. METCALFE E T AL.

Variable redox conditions during red bed diagenesis will not greatly affect sorption of Pb, Zn, Co and Ag, because the oxidation states of these metals will be mostly constant. In contrast, Cu can form Cu 2÷ and Cu + complexes, and redox may control the sorption of this metal, although a lack of data for Cu + prevented modelling this. Sorption of Cu + may be weaker than for Cu 2÷, because Cu ÷ probably behaves like Ag ÷ which forms strong chloride complexes (AgCI2-, AgCI32- and mgCl43-) and which is weakly sorbed at acid and neutral pH (Fig. 10b). Recently Rose & Bianchi-Mosquera (1993) carried out a series of experimental simulations of red bed conditions, and investigated sorption of Pb, Zn, Co, Ni, Cu and Ag by goethite under a range of redox conditions. Their experimental adsorption curves were similar to the modelled curves, but compared to Pb, the copper adsorption edge was at slightly lower, rather than slightly higher pH value, possibly because of differences in Eh, and hence speciation. From Fig. 10 it is clear that under conditions of pH similar to those bounding the field of Cu transport as cuprous chloride, goethite is likely to sorb all Cu 2+, and if Cu + behaves like Ag + and sorbent/sorbate ratios are suitable, may sorb significant amounts of Cu ÷. It is likely that Cu ÷ and Ag ÷ are likely to be more mobile if the adsorbing phase is hematite rather than goethite (Rose & Bianchi-Mosquera 1993). This is consistent with the release of heavy metals to solution due to the transformation of goethite to hematite during burial diagenesis.

migrate through a red bed sequence. This is because the temperature of metal mobilization is relatively unimportant in comparison with Eh-pH conditions. Current theoretical models are capable of reproducing the major features of the mineral parageneses actually observed in ore deposits, including: Cu-bearing sulphides which pre-date Pb- and Zn- sulphides; hematite which formed before the main sulphide-forming event; 'bleached' sandstones adjacent to sulphide mineralization, owing to diminished quantities of hematite; quartz overgrowths in sandstones which predated the main phase of sulphide formation; and the formation of late barite. An important conclusion from these models is that extremely small amounts of fluid mixing can exert a critical control on both diagenesis and upon ore formation. This means that even when there is limited fluid flow within a red bed sequence, due to small hydraulic head gradients and/or low permeabilities, sufficient fluid mixing might occur to cause significant variations in mineral parageneses. Sorption may be a significant control on metal cation partitioning between the rock and the fluid phase, and variations in the pH of the diagenetic fluids may cause variations in this partitioning which in turn lead to fractionation of heavy metals in solution. Our conceptual understanding of the major processes governing red bed diagenesis is relatively good, but there is a lack of realistic computer models and there are deficiencies in the data necessary for computer modelling, especially:

Conclusions

(1) uncertainty about the identities of mineral phases which control heavy metal solubility; (2) limited thermodynamic data at >25°C for likely metal solubility-controlling minerals and aqueous species; (3) a limited Pitzer coefficient dataset, in particular for silicon-bearing aqueous species, but also for all other aqueous species for the temperatures of interest; (4) a lack of kinetic data; (5) alack ofsorption data.

Red bed sequences are chemically heterogeneous, and although they largely have low initial organic contents, they may act as conduits and reservoirs for hydrocarbon-bearing fluids. Together these factors result in a unique diagenetic evolution of authigenic mineral assemblages and pore-water chemistries. Diagenesis occurs over ranges of redox and pH conditions (broadly, logfo2.bars = --0.7 to C. --50; and Eh ~ +500 mV to - 2 0 0 m V (although Eh may reach c. -500 mV in low temperature, alkaline environments); and pH c. 10 to c. 5) which are among the largest for any type of sedimentary sequence. There is a complex interplay between the evolution of red bed porosity and permeability, red bed mineralogy, and the chemical compositions of diagenetic fluids. Formation waters may evolve from metal-transporting fluids to metal precipitating fluids and back again as they

Additional uncertainties are due to current models being limited to simple scenarios such as closed system water-rock interactions which take little account of surface sorption or reaction kinetics. Future work should be directed towards developing open-system, coupled fluid flow-chemical transport models, which can incorporate a treatment of reaction kinetics and sorption.

CONTINENTAL RED BED DIAGENESIS The authors extend their thanks to P. Turner and B.A. Hofmann for helpful reviews of the manuscript. M.B. Crawford is thanked for running the H Y D R A Q L computer sorption models. The work reported here was carried out as part of the Cheshire Basin Project of the British Geological Survey. This paper is published with the permission of the Director of the British Geological Survey.

References BATH, A.H., MILODOWSKI, A.E. & STRONG, G.E. 1987. Fluid flow and diagenesis in the East Midlands Triassic sandstone aquifer: In: GOFF, J.C. • WILLIAMS, B.P.J. (eds) Fluid flow in sedimentary basins and aquifers. Geological Society, London, Special Publications, 34, 127-140. BECHTEL, A. & HOERNES, S. 1993. Stable isotopic variations in clay minerals: a key to the understanding of Kupferschiefer-type mineralization, Germany. Geochimica et Cosmochimica Acta, 57, 1799-1816. t~ POTTMANN, W. 1991. The origin of the Kupferschiefer-type mineralization in the Richelsdorf Hills, Germany, as deduced from stable isotope and organic geochemical studies. Chemical Geology, 91, 1-18. BROOKINS, D.G. 1984. Geochemical aspects of radioactive waste disposal. Springer-Verlag, New York. BRUTON, C.J. 1989. Predicting mineral dissolution and precipitation during burial: synthetic diagenetic sequences. In: JACKSON, K.J. & BOURCIER,W.L. (eds) Proceedings of the Workshop on Geochemical Modelling, CONF-8609134. Lawrence Livermore National Laboratory, Livermore, California, 111-119. BURLEY, S.D. 1984. Patterns of diagenesis in the Sherwood Sandstone Group (Triassic), United Kingdom. Clay Minerals, 19,403-440. CAROTHERS, W.W. & KHARAKA,Y.K. 1978. Aliphatic acid anions in oil-field waters-implications for origin of natural gas. American Association of Petroleum Geologists Bulletin, 12, 2441-2453. CARPENTER,A.B., TROUT,M.L. & P1CKETr,E.E. 1974. Preliminary report on the origin and chemical evolution of lead- and zinc-rich oilfield brines in central Mississippi. Economic Geology, 69,. 1191-1206. DARRAGI, F. & TARDY, Y. 1987. Authigenic trioctahedral smectites controlling pH, alkalinity, silica and magnesium concentrations in alkaline lakes. Chemical Geology, 63, 59-72. DURRANCE, E.M. 1986. Radioactivity in Geology: principles and applications. Ellis Horwood, Chichester. DUTrON, S.P. & LAND, L.S. 1988. Diagenetic history of a well-cemented quartz arenite, Lower Cretaceous, Travis Peak Formation, East Texas. Geological Society of America Bulletin, 100, 1271-1282. DZOMBAK, D.A. & MOREL, F.M.M. 1990. Surface Complexation Modelling. Wiley Interscience, New York.

321

EDMUNDS, W.M. 1986. Geochemistry of geothermal waters in the U.K. In: DOWmNG, R.A. & GRAY, D.A. (eds) Geothermal energy: the potential in the United Kingdom. Her Majesty's Stationery Ofrice, London, 111-122. ~, MILES, D.L. & COOK, J.M. 1984. A comparative study of sequential redox processes in three British aquifers. In: ERIKSSON, E. (ed.) Hydrochemical balances of freshwater systems. IAHS Publications, 50, 55-70. EUGSTER, H.P. 1980. Geochemistry of evaporitic lacustrine deposits. Annual Reviews in Earth and Planetary Sciences, 8, 35-63. -1985. Oil shales, evaporites and ore deposits. Geochimica et Cosmochimica Acta, 49,619-635. 1989. Geochemical environments of sedimenthosted Cu-Pb-Zn deposits. In: BOYLE, R.W., BROWN, A.C., JEFFERSON,C.W., JOWETT,E.C. & KIRKHAM,R.V. (eds) Sediment-hosted Stratiform Copper Deposits. Geological Association of Canada Special Papers, 36, 111-126. FEIN, J.B. 1991. Experimental study of aluminium-, calcium- and magnesium-acetate complexing at 80° C. Geochimica et Cosmochimica Acta, 55, 955-964. FLINT, S. 1987. Diagenesis of Tertiary playa sandstones of Northern Chile; implications for Andean uplift and metallogeny. Sedimentology, 34, 11-29. FROSTICK, L.E. & REID, I. 1987. Tectonic control of desert sediments in rift basins ancient and modern. In: FROSTICK,L. & REID, I. (eds) Desert sediments: ancient and modern. Geological Society, London, Special Publications, 35, 53-68. GERTH, J. 1990. Unit cell dimensions of trace metalassociated goethites. Geochimica et Cosmochimica Acta, 54,363-371. GIORDANO, T.H. & DRUMMOND, S.E. 1991. The potentiometric determination of stability constants for zinc acetate complexes in aqueous solutions to 295 ° C. Geochimica et Cosmochimica Acta, 55, 2401-2415. GUSTAFSON, L.B. & WILLIAMS, N. 1981. Sedimenthosted stratiform deposits of copper, lead and zinc. Economic Geology, 75th Anniversary Volume, 139-178. HARRISON, W.J. & THYNE, G.D. 1992. Predictions of diagenetic reactions in the presence of organic acids. Geochimica et Cosmochimica Acta, 56, 565-586. HARTLEY, A., W.J., FLINT, S. ~¢~ TURNER, P. 1991. Analcime: a characteristic authigenic phase of Andean alluvium, northern Chile. Geological Journal, 26, 189-202. HASLAM, H.W. 8£ SANDON, P.T.S. 1991. The geochemistry of some red bed formations in the United Kingdom. British Geological Survey Technical Report WP/90/2. HELGESON, H.C., KIRKHAM,D.H. & FLOWERS, G.C. 1981. Theoretical prediction of the thermodynamic behaviour of aqueous electrolytes at high pressures and temperatures IV: calculation of activity coefficients, osmotic coefficients, and apparent molal and standard and relative partial

322

R. METCALFE ET AL.

molal properties to 5 kb and 600° C. American genesis of sediment-hosted stratiform copper Journal of Science, 281, 1241-1516. deposits. In: BOYLE, R.W., BROWN, A.C., HEM, J.D. 1989. Study and interpretation of the JEFFERSON, C.W., JOWETT, E . C . • KIRKHAM, chemical characteristics of natural water. United R.V. (eds) Sediment-hosted Stratiform Copper States Geological Survey Water-Supply Paper Deposits. Geological Association of Canada 2254. Special Papers, 36, 3-38. HENNET, R.J.C., CRERAR,D.A. & SCHWARTZ,J. 1988. KRUMGALZ, B. 1980. Salt effect on the pH of Organic complexes in hydrothermal systems. hypersaline solutions. In: NISSENBAUM,A. (ed.) Economic Geology, 83,742-764. Hypersaline brines and evaporitic environments. HOFMANN, B.A. 1990. Reduction spheroids from Developments in Sedimentology, 28, Elsevier, northern Switzerland: mineralogy, geochemistry Amsterdam. 73-82. and genetic models. Chemical Geology, 81,55-81. LANGMUIR, D. 1979. Uranium geochemistry. In: 1991. Mineralogy and geochemistry of reduction KIMBERLEY, M.M. (ed.) Shortcourse in uranium spheroids in red beds. Mineralogy and Petrology, deposits: their mineralogy and origin. Mineralogi44,107-124. cal Association of Canada Short Course Hand1992. Isolated reduction phenomena in red beds: books, 3, 17-29. a result of porewater radiolysis? In: KHARAKA, LEDER, F. & PARK, W.C. 1986. Porosity reduction in Y.K. & MAEST, A.S. (eds) Water-Rock Intersandstone by quartz overgrowth. American Asaction, 7. Balkema, Rotterdam, 503-506. sociation of Petroleum Geologists Bulletin, 70, HOLM, T.R. & CURTISS, C.D. 1990. Copper com1713-1728. plexation by natural organic matter in ground- LUNDEGARD,P.D. & LAND,L.S. 1986. Carbon dioxide water. In: MELCHIOR,D.C. & BASSETT,R.L. (eds) and organic acids: their role in porosity enhanceChemical modelling in aqueous systems If. Ameriment and cementation, Palaeogene of the Texas can Chemical Society, Washington DC, 509-518. Gulf Coast. In: GAUTIER, D.L. (ed.) Roles of HOLMES, L., CHAMBERS,A.D., IXER, R.A., TURNER, Organic matter in sediment diagenesis. Society of P. & VAUGHAN,D.J. 1983. Diagenetic processes Economic Paleontologists and Mineralogists and the mineralization in the Triassic of central Special Publications, 38, 129-147. England. Mineralium Deposita, 18,365-377. MANNING, D.A.C. 1986. Assessment of the role of HOUSEKNECHT, D.W. 1987. Intergranular pressure organic matter in ore transport processes in solution in four quartzose sandstones. Journal of low-temperature base-metal systems. TransSedimentary Petrology, 58,228-246. actions of the Institution of Mining and MetalHUANG, W-L. & LONGO, J.M. 1992. The effect of lurgy, 95, B 195-B200. organics on feldspar dissolution and the develop- McBRIDE, E.F. 1989. Quartz cement in sandstones: a ment of secondary porosity. Chemical Geology, review. Earth Science Reviews, 26, 69-112. 98,271-292. MERINO, E., MOORE, C., ORTOLEVA,P. & RIPELY, E. IXER, R.A. & VAUGHAN,D.J. 1982. The primary ore 1986. Mineral zoning in sediment-hosted coppermineralogy of the Aiderley Edge deposit, iron sulphide deposits-a quantitative kinetic apCheshire. Mineralogical Magazine, 46,485-492. proach. In: FRIEDRICH, D.H. (ed.) Geology and JOHNSON, J.W., OELKERS, E.H. & HELGESON, H.C. metallogeny of copper deposits: proceedings of the copper symposium, 27th International Geological 1991. SUPCR792: a software package for calculating the standard radial thermodynamic Congress. Special Publications of the Society for properties of minerals, gases, aqueous species, and Geology Applied to Mineral Deposits, 4, 559reactions from 1 to 5000 bars and 0° to 1000° C. 571. University of California, Berkeley. MORAD, S. & ALDAHAN, A.A. 1986. Diagenetic JOWETT, E.C. 1986. Genesis of Kupferschiefer Cu-Ag replacement of feldspars by quartz in sandstones. deposits by convective flow of Rotliegende brines Journal of Sedimentary Petrology, 57,488-493. during Triassic rifting. Economic Geology, 81, MORTON, A.C. 1984. Stability of detrital heavy 1823-1837. minerals in Tertiary sandstones from the North - 1989. Effects of continental rifting on the location Sea Basin. Clay Minerals, 19,287-308. and genesis of stratiform copper-silver deposits. NASH, J.T., GRANGER, H.C. & ADAMS, S.S. 1981. In: BOYLE, R.W., BROWN, A.C., JEFFERSON, Geology and concepts of genesis of important C.W., JowE'rr, E.C. & KIRKnAM, R.V. (eds) types of uranium deposits. Economic Geology, Sediment-hosted Stratiform Copper Deposits. 75th Anniversary Volume, 63-116. Geological Association of Canada Special NAYLOR, H., TURNER, P., VAUGHAN, D.J., BOYCE, Papers, 36, 53-66. A.J. & FALLICK, A.E. 1989. Genetic studies of - 1992. Role of organics and methane in sulfide ore red bed mineralization in the Triassic of the formation, exemplified by Kupferschiefer Cu-Ag Cheshire Basin, northwest England. Journal of deposits, Poland. Chemical Geology, 99, 51-63. the Geological Society, London, 146, 685-699. KIMBERLEY, M.M. 1979. Uranium geochemistry. In: NICKEL,E. 1973. Experimental dissolution of light and KIMBERLEY,M.M. (ed.) Short course in uranium heavy minerals in comparison with weathering deposits: their mineralogy and origin. Mineralogiand interstratal solution. Contributions to Sedical Association of Canada Short Course Handmentology, l, 1-68. books, 3,339-375. OECD NUCLEARENERGYAGENCY& INTERNATIONAL KIRKHAM, R.V. 1989. Distribution, settings, and ATOMIC ENERGY AGENCY 1986. Uranium:

CONTINENTAL RED BED DIAGENESIS

resources, production and demand. OECD, Paris. PAPELIS, C., HAYES, K. & LECKIE, J.O. 1988. HYDRA QL: A program for the computation of chemical equilibrium composition of aqueous batch systems including surface-complexation modelling of ion adsorption at the oxide~solution interface. Department of Civil Engineering, Stanford University, Stanford, California. Technical Report, 306. PARNELL, J. 1992. Hydrocarbon potential of Northern Ireland: Part III-reservoir potential of the PermoTriassic. Journal of Petroleum Geology, 15, 51-70. PrrzER, K.S. 1973. Thermodynamics of electrolytes I: Theoretical basis and general equations. Journal of Physical Chemistry, 77,268-277. ROEDDER, E. 1984. Fluid inclusions. Mineralogical Society of America, Reviews in Mineralogy, 12. ROSE, A.W. 1976. The effect of cuprous chloride complexes in the origin of red-bed copper and related deposits. Economic Geology, 71, 10361048. - 1989. Mobility of copper and other heavy metals in sedimentary environments. In: BOYLE, R.W., BROWN, A.C., JEFFERSON,C.W., JOWETI', E.C. & K~R~HAM, R.V. (eds) Sediment-hosted stratiform copper deposits. Geological Association of Canada, Special Papers, 38, 97-110. & BIANCHI-MOSQUERA,G.C. 1985. Adsorption of Cu, Ag and other metals on Fe-oxides as a control on elemental composition of stratiform and red bed copper deposits. Geological Society of America, Abstracts with Programmes, 17,702. -&~ 1993. Adsorption of Cu, Pb, Zn, Co, Ni and Ag on goethite and hematite as a control on metal mobilisation from redbeds into stratiform copper deposits. Economic Geology, 88, 12261236. SAWLOWICZ,Z. 1991. The relationship between copper mineralisation and organic matter in the Polish Kupferschiefer. In: PAGEL, M. & LEROY, J.L. (eds) Source, Transport and Deposition of Metals. Balkema. Rotterdam, 589-592. - 1992. Primary sulphide mineralisation in Cu-Fe-S zones of Kupferschiefer, Fore-Sudetic monocline, Poland. Transactions of the Institution of Mining and Metallurgy, 101, B1-B8. SEEWALD,J.S., SEYFRIED,W.E. & THORNTON,E.C. 1990. Organic-rich sediment alteration: an experimental and theoretical study at elevated temperatures and pressures. Geochimica et Cosmochimica Acta, 55,659-669. SHOCK, E.L. 1988. Organic acid metastability in sedimentary basins. Geology, 16, 886-890. & HELGESON, H.C. 1990. Calculation of the thermodynamic and transport properties of aqueous species at high pressures and temperatures: Standard partial molal properties of organic species. Geochimica et Cosmochimica Acta, 54,915-945. SMITH, I.F. 1986. Mesozoic basins. In: DOWNING, R.A. & GRAY, D.A. (eds) Geothermal energy the potential in the United Kingdom. Her Majesty's Stationery Office, London, 42-83.

323

SONNENFELD, P. 1984. Brines and evaporites. Academic Press, London. STOESSELL, R.K. & P1TI?MAN,E.D. 1990. Secondary porosity revisited: the chemistry of feldspar dissolution by carboxylic acids and anions. American Association of Petroleum Geologists Bulletin, 74, 1795-1805. STRONG, G.E. & MILODOWSKI,A.E. 1987. Aspects of the diagenesis of the Sherwood sandstones of the Wessex Basin and their influence on reservoir characteristics. In: MARSHALL,J. (ed.) Diagenesis of Sedimentary Sequences. Geological Society, London, Special Publications, 36,325-337. STUMM, W. • MORGAN, 1981. Aquatic chemistry. John Wiley & Sons, New York. SURDAM, R.C. & SHEPPARD, R.A. 1978. Zeolites in alkaline-lake deposits. In: SAND, L.B. & MUMIrrON, F.A. (eds) Natural zeolites: Occurrence, properties, use. Pergammon, New York, 145-174. --, Boese, S.W. & CROSSLY, L.J. 1984. The chemistry of secondary porosity. In: MCDONALD, D.A. & SURDAM,R.C. (eds) Clastic Diagenesis. American Association of Petroleum Geologists Memoirs, 37, 127-149. --, CROSSLY, L.J., HAGEN, E.S. 8z HEASLER, H.P. 1989. Organic-inorganic interactions and sandstone diagenesis. American Association of Petroleum Geologists Bulletin, 73, 1-23. SVERJENSKY,D.A. 1984. Oil field brines as ore-forming solutions. Economic Geology, 79, 23-37. -1987. The role of migrating oil field brines in the formation of sediment-hosted Cu-rich deposits. Economic Geology, 82, 1130--1141. - 1989. Chemical evolution of basinal brines that formed sediment-hosted Cu-Pb-Zn deposits. In: BOYLE, R.W., BROWN A.C., JEFFERSON, C.W., JOWET'r, E.C. & KmKHAM,R.V. (eds) Sedimenthosted Stratiform Copper Deposits. Geological Association of Canada Special Papers, 36, 127134. TAYLOR, S.R. & MCLENNAN,S.M. 1985. The continental crust: its composition and evolution. Blackwell Scientific Publications, Oxford. THORNTON, E.C. & SEYFRIED,W.E. 1987. Reactivity of organic-rich sediment in seawater at 350 ° C, 500 bars: experimental and theoretical constraints for the Guaymas Basin hydrothermal system. Geochimica et Cosmochimica Acta, 51, 19972010. TURNER, P. 1980. Continental red beds. Developments in Sedimentology, 29. Elsevier, Oxford. WALKER,T.R. 1967. Formation of red beds in modern and ancient deserts. Geological Society of America Bulletin, 78,353-368. - 1974. Formation of red beds in moist tropical climates. Geological Society of America Bulletin, 85,633-638. - 1976. Diagenetic origin of continental red beds. In: FAERY, H. (ed.) The continental Permian in central, west and south Europe. Reidel Publishing Co., Dordrecht, 240-282. - 1989. Application of diagenetic alterations in redbeds to the origin of copper in stratiform

324

R. METCALFE ET AL.

copper deposits. In: BOYLE,R.W., BROWN,A.C., JEFFERSON, C.W., JOWETT, E.C. & KIRKHAM, R.V. (eds) Sediment-hosted Stratiform Copper Deposits. Geological Association of Canada Special Papers, 36, 85-96. & HONEA, R.M. 1969. Iron content of modern deposits in the Sonoran desert: a contribution to the origin of red beds. Geological Society of America Bulletin, 80,535-544. (~z RUNNELLS, D.D. 1984. Potential role of smectite-illite conversion in the formation of hematite pigment in red beds. Geological Society of America Abstracts with Programs, 16,685. , WAUGH, B. (~ CRONE, A.J. 1978. Diagenesis in first-cycle alluvium of Cenozoic Age, southwestern United States and northwestern Mexico. Geological Society of America Bulletin, 89, 19-32. WAUGH, B. 1978. Authigenic K-feldspar in British Permo-Triassic sandstones. Journal of the Geological Society, London, 135, 51-56.

WEDEPOHL,K.H. (ed.) 1978. Handbook of geochemistry. Springer-Verlag, Berlin. WOLERY, T.J. 1992a. EQ3/6 package overview and installation guide. Lawrence Livermore National Laboratory Report. 1992b. EQ3NR, Theoretical manual and user's guide. Lawrence Livermore National Laboratory Report. (~¢DAVELER,S.A. 1992. EQ6 Theoretical manual and user's guide. Lawrence Livermore National Laboratory Report. ZIELINSKI, R.A., BLOCH, S. & WALKER, T.R. 1983. The mobility and distribution of heavy metals during the formation of first cycle red beds. Economic Geology, 78, 1574-1589. & -1986. Mobility and distribution of heavy metals during the formation of first cycle red beds. In: MUMVrON, F.A. (ed.) Studies in Diagenesis. United States Geological Survey, Bulletin, 1578, 52-59. -

-

Recognition of the thermal effects of fluid flow in sedimentary basins I . R . D U D D Y 1, P . F . G R E E N 1, R . J . B R A Y 2 & K . A . H E G A R T Y

~

Geotrack International Pty Ltd 1 PO Box 4120, Melbourne University, Vic, 3052 Australia 2 30 Upper High Street, Thame, Oxfordshire O X 9 3EX, UK Abstract: Moving fluids are capable of transporting a large amount of heat over long distances in sedimentary basins, but the effects are often ignored when modelling the thermal evolution of sedimentary basins. If significant, the passage of these heated fluids through the sediment pile will leave a thermal signature which can be measured at the present-day using palaeotemperature determinants such as vitrinite reflectance and apatite fission track analysis (AFTATM). The interpretation of palaeotemperature-depth profiles, particularly the slope of a palaeotemperature-depth profile, i.e. the palaeogeothermal gradient, allows fluid-induced temperature profiles to be distinguished from those due to simple conduction of basal heat flow. Steady-state systems, typified by large-scale lateral fluid flow in foreland basins or where low-temperature hydrothermal circulation systems occur associated with intrusions and thick volcanic piles, are characterized by dog-length geothermal gradients, with high-interval gradients near the surface, above the shallowest aquifer, and lower interval gradients beneath an aquifer. It is argued that the classic occurrence of near vertical vitrinite reflectance-depth profiles observed in many ancient foreland basins can result from this mechanism. Thermal effects of transient fluid flow are most easily observed at the present-day, where they are characterized by low or negative geothermal gradients beneath aquifers in active geothermal systems, and the same criterion enables their recognition in ancient situations. Similarly, shallow-level igneous intrusion into porous and permeable sediments can produce observable thermal signatures further from the intrusion than will result from simple conduction, but which can be readily explained by movement of fluids heated by the intrusion. Complex steady-state profiles that may be difficult to distinguish from transient profiles without additional geological data may arise where multiple aquifers are separated by aquitards. The thermal consequences of fluid flow in two of these situations, a foreland basin and igneous intrusion into porous sediments, are illustrated by examples of thermal history reconstruction from the Pliocene Papua New Guinea Fold Belt and the Canning Basin of Australia, respectively.

Fluids, water and hydrocarbons, are fundamental constituents of sedimentary basins. They have the ability to transfer large amounts of heat (e.g. Smith & Chapman 1983) and may therefore influence the thermal evolution of the sediments within the basin, with consequent effects on maturation of source rocks (e.g. Person & Garven 1992) and also diagenetic alteration of reservoirs (e.g. Summer & Verosub 1989). The prime causes of large-scale heated fluid flow appear to be gravity-driven fluid flow (e.g. Bethke 1986; Garven 1989) and the interaction of igneous intrusions with water-saturated sediments (e.g. Einsele et al. 1980; Summer & Verosub 1989). These mechanisms seem to offer the best ways of maintaining the temperatures and moving fluids over geological time scales in excess of a million years, which are necessary for

the formation of large ore deposits and hydrocarbon accumulations and to produce a steadystate perturbation of crustal temperature profiles. Migration of ore fluids is not specifically discussed in this paper, but the methodology of thermal history reconstruction described here has been applied to Pb-Zn Mississippi Valleytype and other ore deposits (e.g. Arne 1992; Arne et al. 1991). Person & Garven (1992), in a study of the Rhine graben, modelled the influence of heated fluids on hydrocarbon source rock maturation patterns and concluded that a regional topographically driven ground water flow system could adequately explain the present-day heatflow data and the measured thermal maturation data. In a series of papers, Summer & Verosub (e.g.

From PARNELL,J. (ed.), 1994, Geofluids: Origin, Migrationand Evolution of Fluids in Sedimentary Basins, Geological Society Special Publication No. 78, 325-345.

325

326

I.R. DUDDY ET AL.

1987, 1989, 1992) have documented widespread thermal effects attributed to fluid flow in sediments under thick Tertiary volcanic sequences in Oregon. They also recognized the importance of vertical maturation profiles, discussed further in this paper, in the interpretation of a fluid flow mechanism for heating and the mistakes in overburden estimation that would result if lateral heat transfer was not recognised. Differences in vitrinite reflectance profiles have also been recorded between ground water recharge and discharge areas in the Uinta basin by Wilier & Chapman (1987). Despite the importance of fluids in heat transfer, the most common approach to modelling the thermal development of sedimentary basins for maturation assessment assumes basal heat flow as the sole source of heat. Overlooking the thermal consequences of lateral fluid flow can lead to gross inaccuracies in the assessment of basal heat flow and therefore in the reconstruction of the thermal, burial, structural and hydrocarbon generation histories of a region. We propose that the key to recognizing the thermal effects of fluid flow starts with reconstruction of the thermal history in the preserved section at a site, using a combination of techniques, the principal among them being apatite fission track analysis (AFTA TM) and vitrinite reflectance (VR). Bray et al. (1992), in a study of the UK East Midlands Shelf, have shown that determination of the palaeogeothermal gradient at the time of maximum palaeotemperatures using AFTA and VR data can provide insight into the cause of heating and subsequent cooling. For example, if heating was due solely to a period of high heat flow, the geothermal gradient at the time of maximum temperature would be higher than the present-day gradient, whereas if heating was caused solely by deep burial followed by uplift and erosion with no change in thermal gradient, the palaeogeothermal gradient would be the same as the present gradient. In this paper, we extend the interpretation of palaeogeothermal gradients to the identification of other causes of heating, specifically heating due to steady-state and transient fluid flow, illustrated with examples from two hydrocarbon provinces: the Papuan Fold Belt and the Canning Basin, Western Australia.

Thermal history reconstruction The basis of assessing the transfer of heat by the passage of fluids in a basin is reconstruction of the thermal history framework. Integration of palaeogeothermal gradients with the timing of

cooling from maximum palaeotemperatures allows reconstruction of the major features of the thermal history. Our approach to thermal history reconstruction has been described by Duddy et al. (1991), and involves three principal steps: (1) determination of maximum palaeotemperatures; (2) measurement of the time of cooling from maximum palaeotemperatures; (3) determination of palaeogeothermal gradient at the time of maximum palaeotemperatures.

Determination o f m a x i m u m palaeotemperatures Vitrinite reflectance and AFTA are the two prime techniques which can provide quantitative estimates of maximum palaeotemperature for use in thermal history reconstruction. Vitrinite reflectance measurement. The main technique for palaeotemperature measurement is provided by determination of the reflectance of the coal maceral vitrinite. The approach of Cook (1989) is used as briefly described below. Indigenous vitrinite identification is made primarily on textural grounds rather than by arbitrary analysis of histograms disconnected from the measurement process, as this allows an independent assessment to be made of the 'in situ' population from 'caved' and 're-worked' vitrinite populations. Ro(max) data is collected for both coal and dispersed organic matter (DOM). The determination of Ro(max) is more exacting than Ro(random) measurements for DOM and requires an accurately centred rotating stage, but it has the key advantage that pairs of measurements at the maximum reflectance positions (180 ° opposed) allows the surface flatness to be assessed, a key factor which may affect the quality of VR determinations. Paired Ro(max) readings which differ by more than ___5% (relative) usually indicate excessive surface relief or tilting; such readings then being rejected and the problem corrected. Simultaneous observation of the fluorescence characteristics of liptinite macerals in a VR sample significantly aids assessment of the indigenous level of maturity and hence the maximum palaeotemperature experienced by the sample. Interpretation of VR data in terms of maximum palaeotemperature. In this study, vitrinite reflectance data are interpreted on the basis of the distributed activation energy model describing

RECOGNIZING HOT FLUID FLOW IN SEDIMENTS the evolution of VR with temperature and time, described by Burnham & Sweeney (1989). In a considerable number of wells from around the world, in a variety of settings in which A F T A has been used independently to constrain the thermal history (e.g. Bray etal. 1992), we have found that the Burnham & Sweeney (1989) model gives good agreement between predicted and observed VR data (unpublished results). As in the case of fission track annealing, it is clear from the chemical kinetic description embodied in equation 2 of Burnham & Sweeney (1989) that temperature is more important than time in controlling the increase of vitrinite reflectance.

AFTA measurement. A F T A is a kinetic method of thermal history analysis applicable to both sediments and basement rocks (e.g. Green et al. 1989a). It provides an estimate of the maximum palaeotemperature to which a rock was subjected in the range up to c. l l 0 ° C for typical geological heating rates (e.g. Green et al. 1989b). Interpretation of A F T A data from a sediment sample requires knowledge of both the present temperature (equilibrium temperature) at which the sample resided and its stratigraphic age. Comparisons between the measured confined track length distribution and present temperature, and the fission track age and the stratigraphic age enable assessment of whether a particular sample has been hotter in the past than it is at the present day, and if so, what the maximum palaeotemperature was. As discussed further below, A F T A also provides a direct estimate of when cooling from the higher palaeotemperature took place. A F T A involves measurement of fission tracks in detrital apatite grains. Fission tracks are linear zones of radiation damage within the apatite crystal lattice, created by spontaneous fission of uranium atoms, usually present at the ppm level. Several points are important in the data collection. For example, the external detector method should be used as it enables the determination of fission track ages on individual apatite grains. The lengths of confined tracks contain much more information on thermal history than projected lengths (e.g. Laslett et al. 1994) and are preferred. Further details of the collection of A F T A data are provided in Green etal. (1989a). Interpretation of AFTA data in terms of maximum palaeotemperature. Fission tracks in apatite form with a narrow distribution of lengths (around c. 16 ~m long), but once formed fission tracks begin to shorten because of the gradual repair of the radiation damage which

327

constitutes the unetched tracks. In effect, the tracks shrink from each end in a process which is known as fission track 'annealing'. The final length of each individual track is essentially determined by the maximum temperature which that track has experienced. A time difference of an order of magnitude produces a change in fission track parameters that is equivalent to a temperature change of only c. 10° C, so temperature is by far the dominant factor in determining the final fission track parameters. As temperature increases, all existing tracks shorten to a length determined by the prevailing temperature, regardless of when they were formed. If the temperature decreases, all tracks formed prior to the thermal maximum are 'frozen' at the degree of length reduction they attained at that time. Thus the length of each track can be thought of as a maximum-reading thermometer, recording the maximum temperature to which it has been subjected. At temperatures greater than c. l l 0 ° C all fission tracks are totally annealed, so that in these cases AFTA provides only a minimum estimate of maximum palaeotemperature. The maximum palaeotemperature to which a sediment has been subjected is determined using a forward modelling approach based on the quantitative description of fission track annealing derived by Laslett et al. (1987), extended to the range of apatite compositions found in sediments (unpublished data). A 'default' thermal history based on the preserved stratigraphy and the present-day temperature of the sample is used as the basis for this forward modelling, but with the addition of episodes of elevated palaeotemperatures as required to explain the data. A F F A parameters are modelled iteratively through successive thermal history scenarios in order to identify thermal histories that can account for observed parameters. The maximum palaeotemperature is adjusted until it is sufficient to account for the degree of fission track age and track length reduction shown by those tracks formed prior to the onset of cooling. Palaeotemperatures between c. 20 and l l 0 ° C can be estimated to a precision of + 10°C at lower values and + 5 ° C at values between c. 70 and 110 ° C (Green et al. 1989b).

Measurement o f the time o f cooling f r o m m a x i m u m palaeotemperatures using apatite fission track analysis AFTA offers the only direct radiometric method for quantitative determination of the time of cooling from maximum palaeotemperature up

328

I.R. DUDDY ET AL.

to c. 110° C for use in thermal history reconstruction in sedimentary basins. Such timing information is critical in determining the relative importance of multiple unconformities in a sequence and evaluating the time of oil generation and migration in relation to trap formation.

Interpretation of time of cooling from A FTA data. Since new fission tracks are produced continually through time, different tracks experience different proportions of the thermal history. Therefore the distribution of track lengths observed at the present day reflects the variation of temperature through time, while the number of tracks measures the total time over which tracks have been accumulating. The most direct information on the time of cooling of a sediment is provided by the AFTA age parameters of those samples in which the fission tracks were totally annealed after deposition. In this case only those tracks formed after cooling will be present, and if cooling occurred rapidly to temperatures less than c. 50°C, the measured fission track age will closely approximate the time of cooling. Typically the time of cooling can be estimated with an accuracy of c. + 10% for this style of thermal history. If the cooling was more protracted and/or the present temperature is greater than c. 50°C, the measured fission track age will underestimate the time of cooling, but this can be compensated for by allowance for the measured track length reduction. In those samples in which fission tracks were not totally annealed after deposition (i.e. those subjected to temperatures less than 110° C), assessment of the relative proportions of short and long tracks in the track length distribution allows the time of cooling to be determined, although with increasingly lower precision with decreasing maximum palaeotemperature. In practice, palaeotemperatures and time of cooling are estimated at the same time in an iterative process by forward modelling the fission track age and length parameters using the approach described in a series of papers (Green etal. 1986, 1989a; Laslett etal. 1987; Duddyetal. 1988). Examples of the application of the technique are given in Bray et al. (1992), Green etal. (1989b) and Miller & Duddy (1989). Integration of AFTA and VR data. Duddy et al. (1991) showed that if the Burnham & Sweeney (1989) distributed activation energy model for vitrinite reflectance is expressed in the form of an Arrhenius plot (a plot of the logarithm of time versus inverse absolute temperature), the slopes

of lines defining contours of equal vitrinite reflectance in such a plot are very similar to those describing the Laslett et al. (1987) kinetic description of annealing of fission tracks in Durango apatite. This feature of the two quite independent approaches to palaeotemperature determination means that for a particular sample, a given degree of fission track annealing in apatite of Durango composition will be associated with the same value of vitrinite reflectance, regardless of the heating rate experienced by a sample. Thus, palaeotemperature estimates based on either A F T A or VR data sets should be equivalent, regardless of the duration of heating. One practical consequence of this relationship between AFFA and VR is that, for example, a VR value of 0.7% is associated with total annealing of all fission tracks in apatite of Durango composition, and total annealing of all fission tracks in apatites of more chlorine-rich composition is accomplished between VR values of 0.7 and c. 0.9%. Furthermore, because vitrinite reflectance continues to increase progressively with increasing temperature, VR data allow direct estimation of maximum palaeotemperatures in the range where fission tracks in apatite are totally annealed (generally above c. l l0°C) and where therefore AFTA only provides minimum estimates. Maximum palaeotemperature estimates based on vitrinite reflectance data from a well in which most AFTA samples were totally annealed will allow constraints on the palaeogeothermal gradient which would not be possible from A F r A alone. In such cases, the AFTA data should allow tight constraints to be placed on the time of cooling and the cooling history, since AFTA parameters will be dominated by the effects of tracks formed after cooling from maximum palaeotemperatures. Even in situations where AFTA samples were not totally annealed, integration of AFTA and VR can allow palaeotemperature control over a greater range of depth, e.g. by combining AFTA from sand-dominated units with VR from other parts of the section, thereby providing tighter constraint on the palaeogeothermal gradient.

Determination o f palaeogeothermal gradient at the time o f m a x i m u m palaeotemperatures Estimation of palaeogeothermal gradient has been discussed by Duddy et al. (1991) and Bray et al. (1992) and proceeds from calculation of the slope of a plot of maximum palaeotemperatures

RECOGNIZING HOT FLUID FLOW IN SEDIMENTS Temperature (*C) 0

li~'c 20

40

60

80

100

120

140

\='!,

%,

t

x

.\

E

J Pa]eogcotherma]

/ , . gradi=,-3,.

-..

•-..

~ . LS O> E3

d' Prcsmat-day / -

gradient -33"CAm

"

.,. -

, ..,

\

i ... \

'it

'.,,

'.,,, o

O V---I •

180

TYPICAL-1

~

\

160

¶•

Corrected BEI" tmasurement P a l e o t ~ t u r e data Range estimate

TD 2600 m

Point estimate Minimum estimate

329

Estimation of the amount of eroded section and heat flow. Calculating the amount of eroded section from thermal data requires either detailed knowledge of the lithologies of the removed section, or assumptions about either the thermal conductivity and heat flow variation or the palaeogeothermal gradient through the missing section. In the absence of this information, and if the palaeogeothermal gradient in the preserved section is in the normal crustal range, then a straight line projection of the palaeogeothermal gradient to an appropriate palaeo-surface temperature value will provide a reasonable estimate of the amount of eroded section (Fig. 2). This procedure contains the implicit assumptions that the lithologies in the removed section were similar to those in the preserved section (i.e. large thicknesses of lithologies with extremes of thermal conductivity like salt or coal were not formerly present), and importantly

[ Maximum estimate

Fig. I. Palaeotemperature profile constructed using observed VR and AFTA data in a typical well section which has cooled from maximum palaeotemperatures at some time since the Carboniferous. In a real case, AFTA can be used to determine the time of cooling. determined from AFTA and/or VR data in a suite of samples in a vertical depth sequence (Fig. 1). This palaeogeothermal gradient is applicable to the time immediately before the onset of cooling from maximum paleotemperatures. In addition, A F T A also allows the cooling paths of a vertical suite of samples to be determined, and thus the possibility exists of constraining a range of palaeogeothermal gradients from the time of maximum palaeotemperatures to the present day. In practice, the constraints on palaeotemperatures during the cooling history may be quite wide, so only broad, but nevertheless useful, limits may be placed on the variation of palaeogeothermal gradient during cooling. The depth range over which samples are analysed is generally the biggest factor controlling the accuracy with which the palaeogeothermal gradient can be constrained. Palaeotemperature estimates over more than 1 km, preferably 2 or 3 km, will give the best constraints on geothermal gradient, and consequently the best constraints on uplift and erosion estimates. Thus, it is important when collecting samples for thermal history reconstruction that VR samples in particular are collected from suitable lithologies from surface to total depth, not simply in target source rock horizons.

P~loo-sur~ce

temperature = 15*C •. : . 4 - - E r o s i o n e s t i m a t e = 1.2 km

g

~

I I I I

0.5

C,m,,~l 0

¢,2o,.

4o.

- Temperature (°C) "~o_ 8o, loo, 12o 14o 16o 18o '--I

-

I

I

i,

I

I

TYPICAL-I

\

0.5

~.~

. PaJeogeothezmal

1.0 0

A o',

-

1.5

.',

-\

,\

D gradient 2.0 " -33°C/kin

-

~O

O

2.5 O I----4 •

• \O

CorrectedB H T m ~ e n t Palenlcmperamre data Range ~timate Point estimate Minimum estimate

', Maximum estimate

Fig. 2. Method of estimation of the amount of uplift and erosion using palaeotemperature estimates from mixed data. This construction is only appropriate when the relevant unconformity is at the present ground surface.

330

I.R. DUDDY E T A L .

that the only source of heat is basal heat-flow. The latter assumption is critical in regions where fluid flow might be expected and may be the prime reason for grossly inaccurate erosion estimates as discussed further below. The slope of the trend of Ro% with depth (Ro%/100 m) is not equivalent to the palaeogeothermal gradient in the section at the time of maximum palaeotemperatures. Extrapolation of a log R o % / l O O m gradient to estimate the amount of section missing through uplift and erosion was suggested by Dow (1977). However, this can only be done when the kinetic relationship between Ro% and temperature and time is taken into account (as, for example, in extrapolated palaeotemperature profiles). Thus, the simple projection of a log Ro%/100 m gradient to a depositional Ro% value (usually somewhere

Thrust-Fold Belt

Foreland Platform

RECHARGE

• Well R

................ "":'-.:'.--'---.:"""...~-"-.'-.'-~"

isotherrn'~

DISCHARGE

n~i.s

~

STEADY-STATE FLUID FLOW Fig. 3. Schematic illustration of fluid movement in an active foreland basin. Ingress of cold fluid in the elevated fold/thrust belt results in gravity-driven flow discharging heated fluids on the foreland platform (after Majorowicz et al. 1985).

Heated water discharge area

Cold water recharge area

Temperature

Temperature

Q., O.

,.C; O.

c~

E)

Heat flow

Heat flow

e-

121

o

Fig. 4. Theoretical temperature and heat flow profiles for two wells located in the recharge and discharge areas on Fig. 3 (after Majorowicz et al. 1985).

RECOGNIZING HOT FLUID FLOW IN SEDIMENTS between 0.15 and 0.25%) will not give an accurate estimate of the amount of missing section, even in the case where the relevant unconformity is at the present-day ground surface, as the slope of log Ro% with depth is not constant with increasing Ro%, even for a constant geothermal gradient. The conversion of Ro% values to palaeotemperatures using the Burnham & Sweeney (1989) kinetic relationship overcomes many of the difficulties associated with projection of log Ro% for erosion estimates and is the approach adopted here (a full discussion of the estimation of uplift and erosion from VR data will be presented elsewhere, but see also discussion in Katz et al. 1988). Once a palaeogeothermal gradient has been determined, it is then possible to go on to constrain palaeo-heat flow at the time of maximum temperatures. However, calculation of heat flow, while very commonly attempted, requires a detailed knowledge of both the thermal conductivity distribution in the stratigraphic section and the porosity reduction history. These latter two parameters are almost never available in an active exploration setting and depend to a large degree on the magnitude of the eroded section which is to be determined! Therefore, the palaeogeothermal gradient data is preferred as a more direct measure for thermal history assessment.

Fluid flow in foreland basins The expected variation in surface heat flow and temperature profiles across an active foreland basin derived by Hitchon (1984) for the Alberta Basin, Canada, and illustrated by Majorowicz et al. (1985), is illustrated in modified form in Fig. 3. The effect of gravity-driven regional cold water recharge at the advancing thrust front is to lower the temperature in the upper part of the sediment pile, with the water transporting this deeper into the pile and then laterally and upwards as the water moves onto the foreland platform and discharges. Overall, this depresses the isotherms in the recharge area and raises them over the foreland platform (Fig. 3). Thus, at locations in the former recharge area and where sedimentation is greatest, the geothermal gradient in the upper part of the section is lowered. Meanwhile, on the foreland platform where fluids are moving laterally and upwards, the geothermal gradient is raised (Fig. 4). This perturbation of the temperature profile by the lateral transfer of heat by fluid gives low measured surface heat flow in the recharge area and a high surface heat flow in the discharge

331

Temperature

.

1 km

j

Geothermal gradient higher than

I

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

Temperatureincrea

\ \ \

\ \ \ \ \ \ Geothermal gradient equal to background value but shifted to higher temperatures

Fig. 5. Expected stead-state temperature profile developed in response to hot fluid flow in a shallow confined aquifer. area, even though basal heat flow across the region may be uniform (Fig. 4). In detail, complex vertical variation in palaeogeothermal gradient may develop in the region where the heated fluids are confined in distinct aquifer horizons on the passage to discharge. The expected steady-state temperature associated with a single aquifer horizon (Ziagos & Blackwell 1986) is illustrated in Fig. 5, with an elevated geothermal gradient above the aquifer and a background geothermal gradient below the aquifer but shifted to higher temperatures. Confinement of flows of similar temperature fluids to distinct aquifer horizons separated by a considerable vertical thickness of shale may result in a zero thermal gradient through the shale section (Fig. 5b). Where local cold recharge interacts with regional hot flow, it may be possible to have hotter fluids in a shallow aquifer than a deeper one, resulting in negative geothermal gradients through the intervening aquitard section (Fig. 7).

Preservation potential of fluid flow thermal signatures A thermal signature will be observed only in a preserved stratigraphic section if the present temperature at which a section resides is lower than the maximum palaeotemperature reached during the fluid flow regime.

332

I.R. DUDDY E T A L .

Temperature

Therefore, once the active development of a foreland basin ceases, the ability to observe the thermal signature of the regional fluid flow regime at the present day varies considerably depending on the location within the former basin. Recharge area. The probability of observing the low palaeogeothermal gradients due to fluid flow in the upper part of the sediment pile in the recharge area (Fig. 3) is low for two reasons. Firstly, uplift and erosion in the part of the basin close to the advancing thrust complex may be high, removing the section which experienced the low geothermal gradient. Secondly, even without uplift and erosion, cessation of cold water ingress will result in a rapid rise in geothermal gradient to the 'normal' background value, as the cold water will no longer be suppressing the transmission of the basal heat flux to the surface. Thus, present-day observations of palaeotemperature profiles in wells at such locations will reveal a normal background geothermal gradient corresponding to the present thermal regime, and the previous influence of fluid flow on the thermal gradient will have been erased. To observe palaeo-fluid flow thermal signatures in a recharge area requires rapid cooling of the section while the fluid flow system is still active, without subsequent reheating. This will require the magnitude of cooling due to uplift and erosion to be greater than the magnitude of rapid heating which must follow cessation of fluid movement. The low fluid flow-induced palaeogeothermal gradient itself makes this situation difficult to achieve. For example, for a palaeogeothermal gradient of 10°C km -~, 3 km of uplift and erosion will only cool the section by 30°C, while return to a typical background gradient of, for example, 30°C km -I will result in preservation of the evidence of maximum palaeotemperatures associated with fluid flow in only the upper 1.5 km of section. However, in practice, the maximum temperatures in this upper 1.5 km will have been between only 30 and 45°C, a temperature range where thermal indicators are least sensitive (vitrinite reflectance would be less than 0.3% over the entire interval). Thus, the true thermal history is unlikely to be resolved in such a case and a far less complicated history would be interpreted from measured data (probably in terms of maximum temperatures at the present day throughout the section and therefore no uplift and erosion). Discharge area. Observation of the thermal

Geothermalgradient higher than background value

~'~~~~f \ lkm

J= C}

\ [i!i!i~i:upper ,!i aquifer horizon: fluid at n*c~i::ilii~i:i!~iii!!!ii::!.!i !i!!:il;i~i~i::~.~:i il \ \ Geothermal gradient is zero

\

--~-

\

\ \

Geothermalgradient equal to background value bu1 shifted to higher temperatures Fig. 6. Temperature profile around two aquifers carrying fluid at the same temperature separated by a thick aquitard.

signature of fluid flow in the discharge area is potentially more straightforward. In this region, cessation of fluid flow will result in the cooling of the section as the isotherms rapidly decline through the section. Without uplift and erosion, this will preserve a VR palaeotemperature profile in the section characteristic of the fluid flow regime; in the simple case illustrated in Fig. 5, a high geothermal gradient above the aquifer horizon, and normal geothermal gradient below. Application of A F F A in the section would additionally reveal the time of cooling from maximum palaeotemperatures, which would be the time of cessation of fluid flow. Uplift and erosion following or coincident with the cessation of fluid flow would enhance the observed magnitude of cooling, but would make the elucidation of the fluid flow thermal signature more difficult, as in the case illustrated, where the characteristic high-interval gradient shallow in the section would be completely removed as a result of only 1 km of uplift and erosion. The remaining section would then only preserve a record of the normal geothermal gradient below the aquifer. Reconstruction of the burial history from such thermal data will result in the over-estimation of the amount of removed section. The more complex palaeotemperature profiles which may develop in regions of discharge, with zero or negative geothermal gradients (Figs

RECOGNIZING HOT FLUID FLOW IN SEDIMENTS

333

D i s c h a r g e area

Temperature

Maximum burial . j . Geothermal gradient higher than

Temperature (*C) 10

50

k\

1 km

100

I

150

l

\~

I

Interval gradient

.;:::.Upper aquifer h0riz0ni f'iuidatn*C:~ :~i: ~. i::, :. :.: : Geothermal gradien

-

NN maintained at 80"C

a

\ \

i Lower aquifer horizon: fluid at < n * C

: :

\

~

Interval

\

gradient

\ ~*C/km

: :

~2. r~

r~=,o

-

no.

Geothermal gradient equal to background value bul shifted to higher temperatures

n.i~ 300c~,.

3-

\

-

\

100"C

Fig. 7. Temperature profile around two aquifers with hotter fluid flowing in the shallower aquifer resulting a negative palaeogeothermai gradient in the intervening aquitard.

6 and 7), may provide the best clue to a fluid flow origin, as such profiles will not be produced by variation in thermal conductivities in typical classic lithologies. In all of these cases, the measurement of Ro(max) and AFTA data allows accurate determination of palaeotemperatures and time of cooling in the preserved section, which will allow the time of source rock maturation to be quantitatively assessed. However, reconstruction of the burial history in fluid-controlled regimes, particularly the amount of section eroded at unconformities, is difficult and may be impossible, as discussed further below.

\

-

140"C

Fig. 8. Steady-state temperature profile in the discharge area developed at maximum burial and controlled by a confined aquifer at I km depth. to a palaeo-surface temperature may result in unrealistically large estimates of removed section. This is illustrated in Figs 8 to 11. Figure 8 shows a theoretical palaeotemperature profile developed in a discharge area well at the time of maximum burial with an aquifer at 1 km depth carrying fluid supplied from deeper in basin at

D i s c h a r g e area Post-uplift

Paleotemperature 10 I

50 I

100 I

(*C) 150 I

Estimates of uplift and erosion in fluid flow regimes Estimating the magnitude of uplift and erosion using thermal methods (i.e., VR, AFTA, Tmax, biomarkers, CAI etc.) in fluid-flow dominated regimes is difficult and dangerous, as the palaeotemperature estimates made in the preserved section, no matter how accurate, cannot be unequivocally transferred to the removed section assuming a constant heat flow. Where low palaeogeothermal gradients (i.e. low apparent palaeo-heat flow) are observed at the present day, simple projection of these gradients

Q.

Pm~nt-day 2 ~ ge.othcrmal ~ gradient , e 30*C/kin

# % Palengenthermal gradient 30*C/kin

Fig. 9. Palaeotemperature profile observed in a discharge area well following uplift and erosion of 1 km of section.

334

I.R. DUDDY E T A L . Time Me)

"Reconstructed" paleotemperature profile

r~

o

Eroded section estimate = 2.3 km

,

l I

~

Exu'apoladonof the

/ I \ ,./,.==,,=dp.l.og-,otho,~ | J \,I" gradient through the / I \ ".,i,.~g.=.o,,',ol~c 1.0

Present-day

groundsurface

| I /~10 ,~_/y ~

~ 2-

Temperature(°C) 5.0 1 \ : ) 0 l ~1

3 150 I

4 A

Time (Ma) J= .,... t-s t~

..-.. 1

J

\\ \ ~

'

] ldlomema= of sectionuplifted :derodcd

Paleogcothe'a~al gradicm 30°C.Jkm

Fig. 10. Reconstruction of eroded section in the discharge area assuming that the geothermal gradient measured in the preserved section extended through the missing section. The calculated value eroded section value overestimates the actual value by a factor of around 2.3, with consequent apparent increase in burial rate in the period prior to cooling.

80 ° C. The palaeotemperature profile recorded in this well at the present day and following removal of I km of section is shown in Fig. 9. If it is assumed that the observed palaeotemperature result from simple deeper burial, implying a constant basal heat flow as the only source of heat, then c. 2.3 km of eroded section would be estimated (Fig. 10). This overestimation of removed section has profound effects on the reconstruction of the burial history, not only in terms of overestimation of the magnitude removed section but also in the overestimation of the burial and heating rates (Fig. 11). Therefore, it is essential that thermal data be used in conjunction with seismic sections or other information which can give a more direct estimate of depth of burial, such as shale density or sonic response (e.g., Magara 1976), chalk porosity (e.g. Hillis 1991). Even so, these techniques may also involve considerable uncertainties, particularly in recognition of the absolute baseline datum. Stratigraphic reconstructions of removed thickness should also be

=.

~2 Actual burial history

3 Fig. 11. Comparison of burial history in the discharge area reconstructed with 2.3 km of missing section assuming no lateral heating due to fluid flow, and the actual burial history with only 1 km of missing section. treated with caution, as they will commonly underestimate the true magnitude, particularly in foreland basins where the region of maximum erosion may correspond with the region of maximum deposition (e.g. Beaumont 1981). In the Alberta Basin, for instance, a number of workers have made estimates of uplift and erosion using thermal data (e.g. Hacquebard 1977; Hitchon 1984; Nurkowski 1984; Beaumont et el. 1985; England & Bustin 1986; Kalkreuth & McMechan 1988; Majorowicz et el. 1990), but no real account has been taken of either the likely non-linearity of palaeogeothermal gradients through the removed section, or the time at which the measured palaeogeothermal gradients are applicable. The estimates of Hacquebard (1977) and Nurkowski (1984) based on equilibrium moisture content of near surface coals were claimed to be non-thermal.

RECOGNIZING HOT FLUID FLOW IN SEDIMENTS However, the good correlation between moisture content, rank and vitrinite reflectance that they report suggests that moisture content may predominantly reflect maximum palaeotemperature rather than simple burial alone. England & Bustin (1986) used vitrinite reflectance data to estimate palaeogeothermal gradients in the Alberta Basin, obtaining values in the range 8-15°C km -1, and from these, estimated amounts of uplift and erosion between c. 5 and 9 km. Majorowicz et al. (1990) have commented on problems with these geothermal gradients on other grounds, but it is clear that extrapolation of low palaeogeothermal gradients from a preserved section, however determined, will result in erosion estimates which are much too high, when lateral heat flow resulting from fluid movement such as expected in the Alberta Basin (e.g. Hitchon 1984; Garven 1989) are not considered. Magara (1976) used shale compaction data to estimate uplift and erosion across the Alberta Basin, and his values are systematically lower than those estimates from VR data, ranging between zero and c. 1400m. He suggested essentially no uplift and erosion in the east, while subsequent studies using thermal data in the same general region suggest minimum values near 1 km (e.g. Nurkowski 1984). It is possible that Magara's (1976) sonic velocity estimates suffer from a baseline datum problem as documented elsewhere (e.g. Bray et al. 1992), although no published estimates of uplift and erosion in the Alberta Basin appear firmly based. AFTA provides a method for addressing part of the problem, by constraining the time at which the palaeogeothermal gradient was operating. This provides a fixed point in time where the thermal and possible burial histories can be evaluated against regional stratigraphic data, time of thrusting, shale density compaction data etc. Limited fission track data have been collected in the Alberta Basin, but these data have not been used either to estimate independently the amounts of eroded section, nor the time of cooling from maximum palaeotemperatures (Issler et al. 1990; Willet & Issler 1992). A practical example of the difficulties associated with reconstruction of the burial history in a foreland basin is illustrated in the next section with data from a well in the Papuan Fold Belt.

A n example o f heated palaeo-fluid flow in a foreland basin, the Papuan Fold-belt The Papuan Fold-belt, a region of Jurassic to Tertiary sedimentation, was in a foreland basin

335

setting from the Mid-Miocene (Davies 1990; Home et al. 1990; Osborne 1990). Collision of the Australian plate with the Melanesian arc at around 10Ma culminated in formation of the Papuan fold and thrust belt and foreland basin on the northern margin of the Australia plate (Smith 1990). The thrust belt includes large basement-involved ramp anticlines, as well as what appears to be a thin-skinned type of deformation in the overlying sedimentary section, with a large number of thrusts and folds in what is now the Papuan Highlands apparently formed by re-activation of earlier extensional structures (Hill 1990). Major hydrocarbon discoveries have recently been made in this region, with most discoveries concentrated along the southwestern margin of the fold-belt trend (e.g. Hedinia, Iagifu) (Fig. 12). The general fold-belt stratigraphy is of Palaeozoic to Permian basement on which was deposited a mid- to late Triassic volcanic and greywacke sequence. Sedimentation from the Jurassic to early Cretaceous was dominated by fine-grained deposits derived from contemporaneous volcanism, with the reservoir facies of the late Jurassic to early Cretaceous Toro sandstone accumulating in a period of volcanic quiescence. Seal to the Toro sandstone is provided by thick shales and silts of the Ieru Formation produced by a return to contemporaneous volcanism through to the late Cretaceous. A major unconformity from the Cenomanian to the Late Oligocene between the Ieru Formation and the overlying Darai Limestone marks regional uplift and erosion associated with the opening of the Coral Sea (Smith 1990). Remnants of a syntectonic clastic depositional phase, the Late Miocene to early Pliocene Orubadi Formation overlies the Darai Limestone in some parts of the fold belt. The outcropping Darai Limestone is deeply karstifled in the fold belt, virtually eliminating the acquisition of seismic data (Lamerson 1990), with hydrocarbon exploration therefore relying largely on outcrop mapping. Published fission track studies (Hill & Gleadow 1989, 1990) have shown that the cooling and uplift and erosion which accompanied thrusting in the fold belt region occurred dominantly in the late Miocene to early Pliocene, between c. 6 and 4 Ma. In some places, the entire Mesozoic and Tertiary section was removed, exposing Permian granite-cored anticlines as at Muller and Kubor (Hill & Gleadow 1989). In other areas of the fold belt, only a portion of the Miocene Darai Limestone and perhaps the overlying early Miocene Orubadi clastics have been removed by Pliocene erosion.

336

I.R. DUDDY ET AL. N

8 8HIELD

FORELAND II

PAPUAN

FLY PLATFORM

I

FLY-ITRICKLAND DEPOGENTflE

FOLD-THflUBT

BELT

li HEOINA IAGIFU

(proJeolsd)

0

- *-*'i~

~"'"

. . .•. ,. • ,, ,i,, ~ A +i,

• 0"*

"

:" .......

*"

--

' " " "1"! . . . . . .

":'d,i

p

.........

o

N

N

-pLEr, IITO

II'E'II'T! _~.,,. .'I.+. .l~. .,,+. . *. . * * * , + + , , ~ * l+** * * * * , i"+ * * , ++!1+ 'Ii *~ '!I+ ++

::::::~+

I

' " "" ~" ":"

:/.++++i+++

~--6000

w Q 10000-,

"

ION MAP

0 !

**:** '" "1 -10000

lOOkm !

~

Llmoelone

~

MRIn reservoir (Toro Osndotono)

Fig. 12. Schematic cross-section showing the present-day structural relationship between the Fly platform and the Papuan fold thrust belt (modified from Osborne 1990). Note the thick shale seal of the Ieru Formation above the confined Toro Formation main reservoir.

Structural and stratigraphic studies indicate around a maximum of 1 to 2 km of missing late Miocene to early Pliocene section in those areas where about 1 km of the Darai Limestone is still preserved. In contrast, more than 3kin of Jurassic to Miocene section has been removed from the Muller and Kubor anticlines (Hill & Gleadow 1990). Thermal history reconstruction in a fold belt well. Figure 13 shows a palaeotemperature-depth plot derived from A F T A in four samples and VR data in three samples from Well A and a VR sample from a nearby outcrop of the Late Miocene to early Pliocene Orubadi Formation. The well penetrates a typical thickness of Miocene Darai Limestone unconformably overlying the Cretaceous Ieru Formation, a dominantly mudstone unit with a drilled thickness of about 1.3 km at this location. The Toro Sandstone reservoir target, typically ! 00 m thick, and the Jurassic Imburu Formation source rock sequence underlie the Ieru Formation, but were not reached in this well. The present-day geothermal gradient, derived from corrected BHT data, is 16°C km -1 and the surface temperature is 15° C. The data is Fig. 13 show that the maximum palaeotemperatures in the well determined from the A F T A and VR data are consistent, and comparison with the present-day temperatures

demonstrate that considerable cooling has occurred. The A F T A results further showed that cooling from maximum palaeotemperatures commenced within the last c. 5 Ma (unpublished A F T A interpretation). Palaeotemperatures over c. 3.0 km of section vary from c. 70°C to c. 100°C, giving a palaeogeothermal gradient before cooling between c. 5 and 10° C km -] . At face value, projection of this range of gradients to a palaeo-surface temperature of 15°C implies between c. 6 and 11 km of uplift and erosion of post-Miocene section. This amount of post-Miocene section is unquestionably unrealistic in this region, with structural (Hill 1990) and stratigraphic (e.g. Lamerson 1990) constraints suggesting a maximum of 1-2 km of section removed since that time. The thermal and geological data can be reconciled if the palaeogeothermal gradient through the section at the time of maximum palaeotemperatures was non-linear, with a high geothermal gradient in the shallow eroded section. In this situation, there is no independent method of constraining both the palaeogeothermal gradient and the amount of eroded section using the palaeotemperature data alone. A viable model for the depositional and uplift history of well A is shown in Fig. 14. This model assumes c. 1200 m of uplift and erosion in the Pliocene combined with the estimated geothermal gradient of 5°C km -1 below the

RECOGNIZING HOT FLUID FLOW IN SEDIMENTS

337

Temperature (°C) Surface 15°C 4 0 0 _. .

60 •

80 100 120 140 tl~

•

|

I; °

;.0-

~

?Orubadi

I

Well A

I t I I

l ~

0.5-

I

3t

•

J L~t

Darai Lst Miocene

r±l

I t

J

t ].5-

1

l i

4~ z

I

-

L

l

I

| ii

° . .

E v

t-

2.0

".'.'2

\

"

r.'-:

gradient -5 to 9°C/kTn

° . ° °°°

(

a 2.5

° . .

[]

. ° . ,

"

ieru Fm Cretaceous

°°.

Present-day ~'~ gradient

-16 C/kin

3.0 =

f

4 D

°°° °°. °°. °°o .°. °.°

"

~

~"

41

PNG Fold belt

r

i

TC -3000m

O

C o r r e c t e d B H T measurement

•

Maximum palcotemperaturc from V R Maximum paleotemperaturc from AFTA at - 5 Ma

Fig. 13. Actual palaeotemperature profile constructed using AFTA and VR data from a Papuan fold belt well, well A. The data allows calculation of a palaeogeothermal gradient throughout the preserved Late Miocene to Early Cretaceous section, between c. - 5 and 9°C km -I.

Orubadi Formation. These parameters allow calculation of a geologically reasonable upper interval geothermal gradient of c. 45°C km -1 when projected to a palaeo-surface temperature of 15°C. For a range of 1-2km of removed section, the corresponding upper interval geothermal gradients would be c. 55 to 28°C km -1, and are equally plausible. Thus, combination of stratigraphic and structural data with the thermal data tends to support a model with migration of fluids heated to c. 70°C through a confined aquifer assumed to be in the vicinity of the boundary between the Orubadi and Darai Formations. The low palaeogeothermal gradient measured through the Darai and Ieru Formations suggests that fluid

movement also occurred in the Toro Formation aquifer just below TD in the well. Fluid temperatures in the Toro Formation would be expected to be a maximum of c. 100-120°C, given the palaeotemperatures derived from the Ieru Formation near total depth. Transient heated fluid flow in an upper aquifer alone is an alternative explanation for the data, as this could also give a near vertical vitrinite reflectance profile as discussed in the next section. This is thought unlikely, however, as evidence from the Toro Formation throughout the fold belt shows that it is a major fluid conduit for both water and oil, and sufficient time was available prior to cooling for steady-state conditions to be established.

338

I.R. DUDDY E T AL.

IPI

o

E Icl

M II

.~_ 1.0

-J ° m

D

E ~" 2.0

;~:..ii{

2 90°C " ~

1

100oo

- - -

Reconstru c ted history WellA 140

120

100

80

60

40

20

0

Time (Ma) Fig. 14. Probable reconstructed history in Papuan fold belt well, recognizing a likely fluid-flow effect on the observed palaeotemperature profile, and assuming only 1 km of section eroded since 5 Ma rather than c. 6-11 km calculated from the palaeotemperature data assuming that the only source of heat was basal heat flow. Finally, the present-day geothermal gradient in this well is itself low when compared with the regional values, and this may reflect the influence of present-day cold water flow through the Darai Limestone karst system. Therefore any assumption that the present-day thermal conditions can be applied to the past to estimate the magnitude of uplift and erosion is also extremely dangerous.

Thermal effects of intrusions transmitted by fluid flow Transient contact thermal effects surrounding igneous intrusive bodies in sediments are well documented. As a general guide, conductive heating effects in the country rocks are felt for a distance of about 1.5-3 times the thickness of the intrusive body (Carslaw & Jaeger 1952). Evidence for the transient nature of such conductive

heating effects is provided by vitrinite reflectance data from the country rock, which shows a rapid decrease in values away from the contact with the intrusive (e.g. Dow 1977; Raymond & Murchison 1988, 1991). These vitrinite reflectance profiles can be interpreted in terms of a positive palaeogeothermal gradient in the sediments above the intrusive and high negative palaeogeothermal gradient below the intrusion giving a bell-shaped profile as illustrated in Fig. 15. The bell-shape of the thermal profile associated with fluid flow in active geothermal systems has been modelled by Ziagos & Blackwell (1986). They have shown the time taken to move from transient to steady-state thermal profiles around a shallow aquifer carrying hot fluid is geologically short (less than 100000 years), so a negative palaeogeothermal gradient below an aquifer could be diagnostic of transient thermal effects (Fig. 16). Note that these transient

RECOGNIZING HOT FLUID FLOW IN SEDIMENTS

VR-derived palaeotemperatures from the underlying finer-grained section may allow short-term transient effects to be recognized. In this context, fluid inclusion-trapping temperatures from diagenetic phases in the aquifer sequence may be useful in demonstrating higher temperature in the aquifer than in the underlying sequence. However, in general, it is probable that only comparison of these perturbed gradients with regional gradients due solely to conduction of basal heat flow will allow a distinction between steady state and transient effects to be made.

Vitrinite reflectance

" ~ ~ i

High +ve paleogeothermal

e~ 121

:,'SJLL.'-:-2..2

~

I

aa<:kgro,.,,:l / VR prof,le

339

'1 "",gg._'eP~.'e°'l '~,_~.eadTeert al J

Fig. 15. Typical transient VR profile developed in lithified, low porosity sediments by conductive heating near an intrusive sill. The width of the thermal aureole (A) is typically 2-3 times the width of the intrusion. temperature profiles around aquifers are essentially the same as those produced around intrusive dykes, where heating is solely by conduction (Fig. 15). Comparison of AFI'A-derived palaeotemperatures from within the aquifer section with

00.

20

An example of heated palaeo-fluid flow associated with intrusion, the Canning Basin, Northwestern Australia The Canning Basin of Northwestern Australia (Fig. 17) records a history of sedimentation from the early Palaeozoic to the present day. A number of major tectonic events resulting in significant unconformities occur through the Palaeozoic and Mesozoic. Principal breaks occur in the Carboniferous, 'the pre-Grant Formation unconformity', and in the Late Triassic-Early Jurassic, 'the Jurassic unconformity'.

TEMPERATURE (°C) 40 60 80

100

EA 100~

,,E

°

=,o - .......

\t(""

,'\\t

///

//\.

/ ,ooo/sooo -s

i : +

-

300 _Z Fig. 16. Transitionfrom backgroundto a steady-statetemperatureprofile around a shallow geothermalaquifer (after Ziagos & Blackwell 1986).Numbersagainstthe transientprofiles indicate yearsafter initiation of flow at 60° C fluid in the aquifer. Note that the backgroundgeothermalgradient (dashed)and the steady-state geothermal gradient are equivalent.

340

I.R. DUDDY ET AL. I

I

122"

123"

I

Northwestern (3anning. %. o o o

l~r~c, i n .

,

1~0.8,

":.-'~ " - - ~

I

Perindi-1 $

~.~'~urringa-l~/

¢~,1 ~

-17"

0I -18"

I

50I

'

/ Brl°°rne~

I

"1

Fig. 17. Canning Basin, Northwestern Australia, showing location of wells studied.

Permian intrusive activity is widespread in the Fitzroy Trough-Lennard Shelf area with a number of sills, dykes and laccoliths of dolerite identified on gravity and magnetics and by drilling (Reeckmann & Mebberson 1984). On stratigraphic grounds, these bodies are Late Permian in age and are further constrained by fission track dating of apatite to have been emplaced at 280-270Ma (Gleadow & Duddy 1984). Thermal history reconstruction. VR values associated with a 150m thick dolerite intrusion intersected in the Perindi-1 were raised to between 1 and 1.3% (Ro max) over a vertical distance c. 550m above the dolerite (Reeckmann & Mebberson 1984), and c. 300 m below the dolerite (Fig. 18). Due to the dominantly sandy section, no VR results were obtained close to the intrusion so the presence of a conductive profile is not revealed. Nevertheless, the available VR data indicate little variation in palaeotemperature over a c. l k m depth range, indicative of a palaeogeothermai gradient in this section close to zero. This pattern of high palaeotemperatures was interpreted to result from the circulation of fluids in the sandstone section heated by interaction with the intrusive body (Reeckmann & Mebberson 1984). VR values in the overlying Jurassic and

Cretaceous section are only 0.3-0.4%, consistent with relatively little additional post-intrusion burial at the well site. Maximum palaeotemperatures of at least 160°C are required to give these VR values if temperatures were maintained for a few million years. In addition, the fission track results in detrital apatites derived from the sandstones over this broad interval show that it cooled from palaeotemperatures in excess of 110°C (for the same heating time) at around 250-270Ma (Gleadow & Duddy 1984). Comparison of the fission track ages and the Late Permian stratigraphic constraints on the age of the intrusion indicates that cooling throughout the section occurred rapidly, from maximum palaeotemperatures in excess of 160°C to temperatures similar to the present temperatures of between c. 50 and 60° C. Intrusion of these large dolerite bodies in the form of sills and laccoliths at shallow levels into porous sandstone of the Permian Grant Formation has resulted in distinct thermal effects observed kilometres from the closest intrusives (Reeckmann & Mebberson 1984). Using apatite fission track analysis, Gleadow & Duddy (1984) observed that the Late Permian time of cooling in the Grant Formation section in the Kambara1 well was indistinguishable from the time of cooling of dolerites intruding the same Grant

RECOGNIZING HOT FLUID FLOW IN SEDIMENTS

Tappers Inlet-1

Perindi-1

0O-

341

JurassicCretaceous - 500

Jurassic-

Cretaceous

"~

500 -

-1000

1000-

1500

:D"

e-

~" 1500-

a

2000Ordovician

O~ 0

~

2500-

="~evonlan....... ~..i 2000

erindi-1

-2500

Tappers .5

1

Inlet-1 2

-3000 3

Ro max Fig. 18. VR (crosses are TAI values converted to equivalent VR) depth plot for Perindi-I and Tappers Inlet-1 wells, Canning Basin, illustrating the lower, near vertical, palaeogeothermal gradient associated with the dolerite sill in Perindi-I (after Reeckmann & Mebberson 1984).

Formation section in the Perindi-1 well 13km away, although Kambara-1 is actually only 3 km from the nearest intrusion believed to be of the same age (Fig. 19). The maximum palaeotemperatures experienced by the Grant and Poole Formations in Kambara-I are lower than those in Perindi-1, but were still in the range of c. 90-110°C (Gleadow & Duddy 1984). However, AFTA palaeotemperature estimates do not show any simple relationship with increasing depth, implying that better aquifer zones within the formations may have experienced slightly higher temperatures than less porous intervals. In addition, limited VR data from shales in the sequence did not show any pronounced increase in reflectance which might be attributable to intrusion-related heating (Reeckmann & Mebberson 1984). Such a situation is consistent with transient heating by fluids distal from the site of intrusive heating. Raymond & Murchison (1988) and Murchison & Raymond (1989) have documented a similar variation in thermal effects associated with two types of sills intruding the Carboniferous sediments in the Midland Valley area of Scotland. Sills of quartz dolerite composition show a distinct conductive aureole defined by vitrinite reflectance data, while no apparent change in the vitrinite reflectance profile is observed near teschenite bodies. These differing effects were

interpreted as the result of shallow-level intrusion by the teschenite sills into water-saturated sediments with the heat rapidly removed from the vicinity of the intrusive by fluid circulation, while the quartz dolerite sills were intruded after deeper burial of the section, with heat transmission largely by conduction. Similar widespread thermal effects attributed to large-scale expulsion of heated pore fluids were described by Einsele et al. (1980) from sediments in the Gulf of California. Thus, in the Canning Basin, A F T A and VR data have been used to demonstrate widespread thermal effects due to intrusion of dolerite bodies at shallow levels into an aquifer sequence (Reeckmann & Mebberson 1984; Gleadow & Duddy 1984). Temperature effects in the vicinity of the intrusions suggest that fluid movement operated for a sufficient time to achieve a steady state, while those some kilometres from the nearest intrusion may have been transient.

Discussion: distinguishing transient and steady-state thermal effects It is not possible to evaluate independently the time scale of palaeo-heating using AFTA and VR palaeotemperature data, as both methodologies have similar underlying kinetic processes, and show effectively parallel responses

342

I.R. DUDDY E T A L . I

I

121045'

122000 .

~

I

I

122°15'

122030.

~0

_

~

Sill

intruding Permian s

? I

Kambara-1

.~.

16 4s

'

1pk~ I

I

I

Fig. 19. Distribution of shallow dolerite intrusions around the Kambara-1 well, in which the Grant formation section shows transient heating effects (after Reeckmann & Mebberson 1984). for the typical range of geological heating times (Duddy et al. 1991). Only in the case of transient heating occurring at the present day can the kinetics of A F F A and VR be used to estimate how long the elevated temperatures have been operating. Thus, palaeotemperature estimates will be lower than expected for the thermal history estimated from the preserved stratigraphic section and the present temperatures, showing that the present temperatures can only have been operating for a short time. A bell-shaped palaeotemperature profile with a negative palaeogeothermal gradient beneath an igneous sill or aquifer horizon is good evidence for transient heating. On the other hand, the lack of expected conduction profiles around an igneous intrusion is an initial clue that intrusion has occurred into porous, water-saturated sediments. In this case the thermal effects will be of lower magnitude but will be more widespread due to the transport of heat by large volumes of fluid. Such fluidfacilitated heating may also be transient, but heating may also persist for more than 10-~years, in which case steady-state thermal profiles would be developed. Unfortunately, the shape of the palaeotemperature profiles set up by long-term fluid movement and shorter-term fluid circulation at higher temperatures may be similar. The same difficulty is present in gravity-driven fluid flow regimes, where zero or negative palaeogeothermal gradients maintained over kilometres of preserved section may be indicative of transient flow in the case of a single

shallow aquifer horizon, or steady-state flow if the low gradient section is confined between two aquifer horizons. Detailed data on lithological distribution are required to aid distinction between transient and steady-state flow regimes. Whether transient or steady-state fluid flow is involved, the perturbation of palaeotemperature profiles resulting from hot fluid flow are distinctive and must be identified if meaningful burial histories are to be constructed from palaeotemperature data.

Conclusions (1)

(2)

(3)

Identification of the thermal effects of fluid flow is possible by interpretation of palaeogeothermal gradient from a plot of maximum palaeotemperature with depth obtained from a borehole section. Transient fluid flow at some time in the past may be recognized by bell-shaped palaeotemperature profiles, with negative palaeogeothermal gradients below an aquifer or intrusive igneous body; variation in AFTA-derived palaeotemperature estimates over a narrow range of depths in sandstone sequences; and fluid inclusion-trapping temperatures from diagenetic phases in an aquifer which are higher than estimated from AFTA or VR data estimated assuming steady-state heating. Present-day transient fluid flow may be recognized by bell-shaped temperature

RECOGNIZING HOT FLUID FLOW IN SEDIMENTS

(4)

profiles around aquifer horizons and evidence from AFTA or VR that temperatures must have increased very recently. Steady-state fluid flow in the past may be recognized by very low to negative paiReDgeothermal gradients in normal clastic sequences over kilometres of vertical section; and very high palaeogeothermal gradients which are unsustainable to moderate depths without implying melting within the sedimentary pile.

References ARNE, D.C. 1992. Evidence from apatite fission track analysis for regional Cretaceous cooling in the Ouachita Mountain fold belt and Arkoma Basin of Arkansas. American Association of Petroleum Geologists Bulletin, 76,392-402. ~, GREEN, P.F. & DUDDY, I.R. 1991. Regional thermal history of the Pine Point area, NorthWest Territories, Canada, from apatite fission track analysis. Economic Geology, 86,428-435. BEAUMONT, C. 1981. Foreland basins. Geophysical Journal of the Royal Astronomical Society, 65, 291-329. , BOUTILIER, R., MACKENZIE, A.S. & RULLKOTTER,J. t985. lsomerization and aromatization of hydrocarbons and the paleothermometry and burial history of the Alberta Foreland Basin. American Association of Petroleum Geologists Bulletin, 69,546-566. BETHKE, C.M. 1986. Hydrological constraints on the genesis upper Mississippi Valley mineral District from Illinois Basin brines. Economic Geology, 78,233-249. BRAY, R.J., GREEN, P.F. & DUDDY, I.R. 1992. Thermal history reconstruction using apatite fission track analysis and vitrinite reflectance: A case study from the UK East Midlands and Southern North Sea. In: HARDMAN,R.F.P. (ed.) Exploration Britain: Geological Insights for the Next Decade. Geological Society, London, Special Publications, 67, 3-25. BURNHAM, A.K. & SWEENEY,J.J. 1989. A chemical kinetic model of vitrinite maturation and reflectance. Geochimica et Cosmochimica Acta, 53, 2649-2657. CARMEN,G.J. 1990. Occurrence and nature of eocene strata in the Eastern Papuan Basin. In: CARMEN, G.Z. & CARMEN,C. (eds) Petroleum Exploration in Papua New Guinea. Proceedings of the First PNG Petroleum Convention, Port Moresby, 169-183. CARSLAW, H.S. (~ JAEGER, J.C. 1952. Conduction of Heat in Solids, Second Edition. Oxford University Press, Oxford. COOK, A.C. 1989. The use of vitrinite reflectance as an indicator of the level of organic maturation. In: New Directions on Studies of Maturation in Thermal History of Sedimentary Basins. Australian Mineral Foundation seminar, Adelaide.

343

DAVIES, H.L. 1990. Structure and evolution of the border region of New Guinea. In: CARMEN,G.Z. CARMEN, C. (eds) Petroleum Exploration in Papua New Guinea. Proceedings of the First PNG Petroleum Convention, Port Moresby, 245-269. Dow, W.G. 1977. Kerogen studies and geological interpretations. Journal of Geochemical Exploration, 7, 79-99. DUDDY, I.R., GREEN, P.F., HEGARTY, K.A. & BRAY, R.J. 1991. Reconstruction of thermal history in basin modelling using apatite fission track analysis: What is Really Possible? Offshore Australia Conference Proceedings, 1, III-49-III-61. & LASLETT,G.M. 1988. Thermal annealing of fission tracks in apatite: 3. Variable temperature behaviour. Chemical Geology (Isotope Geoscience Section), 73, 25-38. EINSELE, G., GIESKES, J.M. & 17 OTHERS. 1980. Intrusion of basaltic sills into highly porous sediments, and resulting hydrothermal activity. Nature, 283,441-445. ENGLAND, T.D.J. & BUSTIN, R.M. 1986. Thermal maturation of the western Canadian Sedimentary Basin south of the Red Deer River: 1. Alberta Plains. Bulletin of Canadian Petroleum Geology, 34, 71-90. GARVEN, G. 1989. A hydrogeologic model for the formation of the giant oil sands deposits of the Western Canada Sedimentary Basin. American Journal of Science, 289, 105-166. GLEADOW, A.J.W. & DUDDY, I.R. 1984. Fission track dating and thermal history analysis of apatites from wells in the North-West Canning Basin. ln: PURCELL, P.G. (ed.) The Canning Basin, WA. Proceedings of the Geological Society of Australia/Petroleum Exploration Society of Australia Symposium, 377-387. GREEN, P.F., DUDDY, I.R., GLEADOW, A.J.W. & LOVERING,J.F. 1989a. Apatite fission track analysis as a paleotemperature indicator for hydrocarbon exploration. In: NAESER, N.D. & MCCULLOCH, T. (eds) Thermal History of Sedimentary Basins - Methods and Case Histories. Springer-Verlag, New York, 181-195. --, TINGATE, P.R. & LASLEWr, G.M. 1986. 'Thermal annealing of fission tracks in apatite: 1. - A qualitative description. Chemical Geology (Isotope Geoscience Section), 59, 237253. --, LASLErr, G.M., HEGARXV, K.A., GLEADOW, A.J.W. & LOVERmG, J.F. 1989b. Thermal annealing of fission tracks in apatite: 4. Quantitative modelling techniques and extension to geological timescales. Chemical Geology (Isotope Geoscience Section), 79, 155-182. HACOUEBARD,P.A. 1977. Rank of coal as an index of organic metamorphism for oil and gas in Alberta. In: DEROO, G., POWELL, T.G., TtSSOT, B. & MCCROSSAN, R.G. (eds) The Origin and Migration of Petroleum in the Western Canadian Sedimentary Basin: A Geochemical and Thermal Maturation Study. Geological Survey of Canadian Bulletin, 262, 11-22. HILL, K.C. 1990. Structural styles and hydrocarbons in

344

-

-

I.R. D U D D Y E T A L .

the Papuan Fold Belt: A review. In: CARMEN, G.Z. & CARMEN,C. (eds) Petroleum Exploration in Papua New Guinea. Proceedings of the First PNG Petroleum Convention, Port Moresby, 301-310. & GLEADOW, A.J.W. 1989. Uplift and thermal history of the Papuan Fold Belt, Papua New Guinea: fission track analysis. Australian Journal of Earth Sciences, 36, 515-539. & 1990. Apatite fission track analysis of the Papuan Basin. In: CARMEN,G.Z. & CARMEN, C. (eds) Petroleum Exploration in Papua New Guinea. Proceedings of the First PNG Petroleum Convention, Port Moresby, 119-136. HILLIS, R.R. 1991. Chalk porosity and Tertiary uplift, Western Approaches Trough, SW UK and NW French continental shelves. Journal of the Geological Society, London, 148,669-679. HITCt4ON, B. 1984. Geothermal gradients, hydrodynamics and hydrocarbon occurrences, Alberta, Canada. American Association of Petroleum Geologists Bulletin, 68,713-743. HOME, P.C., DALTON, D.G. & BRANNAN, J. 1990. Geological evolution of the Western Papuan Basin. In: CARMEN, G.Z. & CARMEN, C. (eds) Petroleum Exploration in Papua New Guinea. Proceedings of the First PNG Petroleum Convention, Port Moresby, 107-117. ISSLER, D.R., BEAUMONT, C., WILLETT, S.D., DONEHCK, R.A., MOOERS, J. & GRIST, A. 1990. Preliminary evidence from apatite fission track data concerning the thermal history of the Peace River Arch Region, Western Canada Sedimentary Basin. Bulletin of Canadian Petroleum Geology, 38A, 250-269. KALKREUTn, W. & MCMECHAN, M. 1988. Burial history and thermal maturity, Rocky Mountain Front Ranges, Foothills, and Foreland, EastCentral British Columbia and adjacent Alberta, Canada. American Association of Petroleum Geologists Bulletin, 72, 1395-1410. KATZ, B.J., PHEIFER, R.N. & SCHUNK, D.J. 1988. Interpretation of discontinuous vitrinite reflectance profiles. American Association of Petroleum Geologists Bulletin, 72,926-931. LAMERSON, P.R. 1990. Evolution of structural interpretations in Iagifu-Hedinia field, Papua New Guinea. In: CARMEN, G.Z. & CARMEN, C. (eds) Petroleum Exploration in Papua New Guinea. Proceedings of the First PNG Petroleum Convention, Port Moresby, 283-300. LASLEaq', G.M., GALBRArrH, R.F. & GREEN, P.F. 1994. The analysis of projected fission tracks. Nuclear Tracks, in press. , GREEN, P.F., DUDDV,I.R. & GLEADOW,A.J.W. 1987. Thermal annealing of fission tracks in apatite: 2. A Quantitative Analysis. Chemical Geology (Isotope Geoscience Section), 65, 1-13. LAW, B.E., N u c o o , V.F. & BARKER, C.E. 1989. Kinky vitrinite reflectance well profiles: Evidence of paleopore pressure in low-permeability, gasbearing sequences in Rocky Mountain Foreland Basins. American Association of Petroleum Geologists Bulletin, 73,999-1010. -

-

MAGARA, K. 1976. Compaction and fluid migration. Developments in Petroleum Science, 9, Elsevier, Amsterdam. MAJOROWICZ, J.A., JONES, F.W. & ERTMAN, M.E. 1990. Relationship between thermal maturation gradients, geothermal gradients and estimates of the thickness of the eroded foreland section, Southern Alberta Plains, Canada. Marine and Petroleum Geology, 7, 138-152. ~, RAHMAN, M., JONES, F.W. & MCMILLEN, N.J. 1985. The paleogeothermal and present thermal regimes of the Alberta Basin and their significance for petroleum occurrences. Bulletin of Canadian Petroleum Geology, 33, 12-21. MILLER, D.S. & DUDDY, I.R. 1989. Early Cretaceous uplift and erosion of the Northern Appalachian Basin, New York, based on apatite fission track analysis. Earth and Planetary Science Letters, 93, 35--49. MURCmSON, D.G. & RAYMOND,A.C. 1989. Igneous activity and organic maturation in the Midland Valley of Scotland. International Journal of Coal Geology, 14, 47-82. NURKOWSKI, J.R. 1984. Coal quality, coal rank Variation and its relation to reconstructed overburden, Upper Cretaceous and Tertiary Plains Coals, Alberta, Canada. American Association of Petroleum Geologists Bulletin, 68,285-295. OSBORNE, D.G. 1990. The hydrocarbon potential of the Western Papuan Basin Foreland - with reference to worldwide analogues. In: CARMEN, G.Z. & CARMEN,C. (eds) Petroleum Exploration in Papua New Guinea. Proceedings of the First PNG Petroleum Convention, Port Moresby, 197-213. PERSON, M. & GARVEN, G. 1992. Hydrologic constraints on petroleum generation within continental rift basins: Theory and application to the Rhine Graben. American Association of Petroleum Geologists Bulletin, 76,468-488. RAYMOND, A.C. & MURCmSON, D.G. 1988. Development of organic maturation in the thermal aureoles of sills and its relation to sediment compaction. Fuel, 67, 1599-1608. -& -1991. Short Paper: The relationship between organic maturation, the widths of thermal aureoles and the thicknesses of sills in the Midland Valley of Scotland and Northern England. Journal of the Geological Society, London, 148,215-218. REECKMANN,S.A. & MESBERSON, A.J. 1984. Igneous intrusions in the North-West Canning Basin and their impact on oil exploration. In: PURCELL,P.G. (ed.) The Canning Basin, WA. Proceedings from the Geological Society of Australia/Petroleum Exploration Society of Australia Canning Basin Symposium, Perth, 389-399. RUSSELL, N.J. 1990. Application of vitrinite reflectivity to paleogeothermometry studies: some examples from Papua New Guinea basins. In: CARMEN, G.Z. & CARMEN,C. (eds) Petroleum Exploration in Papua New Guinea. Proceedings of the First PNG Petroleum Convention, Port Moresby, 403-420.

R E C O G N I Z I N G HOT FLUID FLOW IN SEDIMENTS SMrrn, L. & CHAPMAN, D.S. 1983. On the thermal effects of groundwater flow, 1 - Regional scale systems. Journal of Geophysical Research, 88, B1, 539--608. SMITH, R.I. 1990. Tertiary plate tectonic setting and evolution of Papua New Guinea. In: CARMEN, G.Z. • CARMEN,C. (eds) Petroleum Exploration in Papua New Guinea. Proceedings of the First PNG Petroleum Convention, Port Moresby, 229-244. SUMMER, N.S. & VEROSUB, K.L. 1987. Maturation anomalies in sediments underlying thick volcanic strata, Oregon: Evidence for a thermal event. Geology, 15, 30-33. SUMMER,N.S. & VEROSUB,K.L. 1989. A low temperature hydrothermal maturation mechanism for sedimentary basins associated with volcanic rocks. In: PRICE, R.A. (ed.) Origin and Evolution

of Sedimentary Basins and their Energy and Mineral Resources. IUGG, 3, Washington, 129-

-

136. &~ 1992. Diagenesis and organic maturation of sedimentary rocks under volcanic strata, Oregon. American Association of Petroleum Geologists Bulletin, 76, 1190-1199.

345

SWEENE¥,J.J. 8¢. BURNHAM,A.K. 1990. Evaluation of a simple model of vitrinite reflectance based on chemical kinetics. American Association of Petroleum Geologists Bulletin, 74, 1559-1570. WAPLES, D.W. 1990. Kinky vitrinite reflectance well profiles: Evidence of paleopore pressure in low-permeability, gas-bearing sequences in Rocky Mountain Foreland Basins: Discussion.

American Association of Petroleum Geologists Bulletin, 74,946-947. WILLETT,S.D. & CHAPMAN,D.S. 1987. Temperatures, fluid flow and thermal history of the Uinta Basin. In: DOLICEZ,B. (ed.) Migration of Hydrocarbons in Sedimentary Basins, Editions Technip, Paris, 533-551. & ISSLER, D.R. 1992. Apatite fission track thermochronometry applied to the Western Canada Sedimentary Basin - A case history from the Peace River Arch region. In: Low Temperature Thermochronology, Mineralogical Association of Canada Short Courses, 20, 157-185. ZIAGOS,J.P. & BLACKWELL,D.D. 1986. A Model for the transient temperature geothermal systems.

Journal of Volcanology and Geothermal Research, 27,371-397.

The use of natural He, Ne and Ar isotopes to study hydrocarbonrelated fluid provenance, migration and mass balance in sedimentary basins C.J. B A L L E N T I N E 1 & R . K . O ' N I O N S 2

I Paul Scherrer Institute, Wiirenlingen und Villigen CH-5232 Villigen PSI, Switzerland 2 University o f Cambridge, Department o f Earth Sciences, Downing Street, Cambridge CB2 3EQ, UK Abstract: Hydrocarbon accumulations contain rare gases derived from the atmosphere, the

crust and, in some cases, the mantle. The distinctive isotopic structure of these different rare gas components allows them to be resolved. The relative abundances of the He, Ne and Ar in the crustal, mantle and atmosphere-derived components provides information on the physical processes which have operated in the subsurface. When combined with mass balance considerations, which place constraints on the scale of fluid movements in sedimentary basins, the rare gases provide powerful constraints on fluid provenance and transport. Results of case studies from two extensional basin systems, the Pannonian and Vienna basin, and the sub-Alpine loading basin of the Po, illustrate how the rare gases provide information about the extent of hydrocarbon-groundwater interaction, constrain the mechanism of hydrocarbon gas transport, and provide an insight into the behaviour of the fluid regimes in different types of sedimentary basin.

The importance of the role of fluids within geological systems cannot be over-emphasized. Whether we are considering fundamental geological problems such as the formation of lithophile-rich continental rocks from primitive magmas, the role of fluids in metamorphic terrains, or attempting to predict the occurrence of ore deposits, hydrocarbon reserves, hydrothermal reservoirs or, at a more basic level, the availability of drinking water, all require a full understanding of the fluids involved. This not only means an understanding of the fluid provenance but also of the mechanisms of fluid migration, of the extent of interactions with differently sourced fluids and, equally as important, just how much fluid is involved in these different processes. Rare gas isotopic compositions may be used to study the behaviour of crustal fluid systems (Zartman et al. 1961). This is because any interaction between differently sourced fluids results in the equilibration of isotopically distinct rare gases between the different fluid systems. For example, hydrocarbons equilibrate to a variable extent with the groundwater system containing dissolved atmosphere-derived rare gases. This results in the transfer of atmospherederived rare gases into the hydrocarbon phase (Bosch & Mazor 1988; Ballentine et al. 1991).

In many crustal environments mantle-derived volatiles, particularly 3He, are also present (Mamyrin & Tolstikhin 1984; Poreda et al. 1986; Oxburgh et al. 1986) and are usually associated with lithospheric extension and melting (Oxburgh & O'Nions 1987; O'Nions & Oxburgh 1988). Furthermore, decay of the radioelements U, Th and K, either within the fluid reservoir volume or deeper within the crust, will produce radiogenic and nucleogenic rare gas isotopes which may also be present in hydrocarbon reservoirs. With precise isotopic analyses it is now possible in most crustal fluid samples to resolve rare gases from these different sources, quantify their relative contributions, and place constraints on the processes responsible for their input (Fig. 1). Since the rare gases are chemically inert, changes in their elemental ratios, which are well defined in the different sources (e.g. Ozima & Podosek 1983), occur only by physical processes and therefore must reflect the physical behaviour of the bulk phase in which the rare gases are carried. For example, rare gas fractionation caused by partitioning between two different fluid phases is often distinct and resolvable from that caused by a diffusive/kinetic fractionation process. Furthermore, as there are no chemical sinks for the different rare gas species, they form

From PARNELL,J. (ed.), 1994, Geofluids: Origin, Migrationand Evolution of Fluidsin Sedimentary Basins, Geological Society Special Publication No. 78, 347-361.

347

348

C.J. B A L L E N T I N E & R.K. O'NIONS \

/

"",,IAtTsp"i'cc° p°ne i Jifer recharge I

JJ

ormation Water Fig. 1. Schematic diagram of a hydrocarbon reservoir, illustrating the different rare gas components which may occur within hydrocarbon fluids. Atmosphere-derived rare gases are input into the hydrocarbon fluid by

interactions between the hydrocarbon phase and the groundwater system which contains dissolved atmosphere-derived rare gases. Radiogenic rare gases produced by the natural decay of the radioelements U, Th and K are also incorporated into hydrocarbon fluids. Within areas of recent continental extension, rare gases derived from the mantle may also be associated with hydrocarbon fluids. The distinct isotopic composition of each of these rare gas components allows the extent of their contribution to the hydrocarbon phase to be resolved. conservative tracers within the subsurface, and enable quantitative constraints to be placed upon the volume of the original fluids involved in any such interactions. In many instances rare gases are closely associated with major gas species. For example, at mid ocean ridges 3He and CO2 are degassed at a near constant ratio (Marty & Jambon 1987). Within the continental environment the ratio of 4He/N2 within any one region is often constant, such as in the H u g o t o n - P a n h a n d l e gas fields, Texas (Pierce et al. 1964). A methane gas phase which separates from a CH4-saturated groundwater has a characteristic CH4/36Ar ratio (Ballentine et al. 1991). Where these relationships are preserved, the rare gases provide information about the provenance of the major gas species. However, surprisingly little use has been made of rare gases, other than He, in the study of crustal fluid systems. This work updates a review by Ballentine & O'Nions (1993) to include recent progress that has been made in the application of He, Ne and Ar isotopes to

25°C

Temperature

4

i

2.8

I

3.0

,

I

3.2

i

I

3.4

~o(

,

I

3.6

,

I

3.8

,

4.0

Xe/Ar (10 -5 ) Fig. 2. The relative solubilities of the rare gases are temperature-dependent. The dissolved atmospheric Ne/Ar and Xe/Ar ratios in fresh water equilibrated with the atmosphere are shown as a function of the water temperature. This distinct rare gas pattern enables palaeo-recharge temperatures of groundwaters to be determined to within +0.5 ° C.

He, Ne & Ar IN HYDROCARBON FLUIDS

(a)

i

i

i

349

i

i

i

25°~,....~OIL

.g6 < Z

0oc

Z 4

I

2

i

3

I

n

4

I

i

,

4 6

5

i

2

~

i

3

,

i

4

Ar cc (STP) / g (10

Ar cc (STP) / g (10 -4 )

,

5

6

-4 )

Fig. 3. (a) The Ne/Ar ratio of dissolved atmosphere rare gases in groundwater is plotted against its Ar concentration, for seawater and fresh water. An increase in salinity does not significantly change the relative solubilities of Ne and Ar, but does decrease the concentration of the atmosphere-derived rare gases in solution. This relationship can be used to determine the salinity of groundwater at recharge. (b) Interaction between groundwater and another phase will result in partitioning of rare gases between the two phases. The predicted effect of oil/water and gas/water interactions on the dissolved atmospheric Ne/Ar ratio and Ar concentration in the water are shown.

(a)

8

- ~

i

[

f

(b)

I

1

I[--II

II

Mantle

6

_

;~

"1" 2 _ 0 0

Nr--

I

•11

•

o#~:.-~, • t,~,a)

-5

I

Mantle

-10

:C:o.r0, -15

-20

-25

~13C(CO2-gas)(%o)

2-

-30

(b

b;4;" I . L ;•

o,...', ~

.

-2

-4

o

.,~(a)

%td/•

,or~;.~.l.

-6

~13C(C02

-8 -10

I

-12 -14

-gas)(%o)

Fig. 4. Relationship between the isotopic composition of carbon, given a s ~13Cco2 and the isotope composition of He, presented as R/Ra values (From Griesshaber et al. 1992). Samples of water and gas from NW Europe define two distinct mixing trends. (a) To a first approximation the He-C isotope relationship of these fluids can be described as a simple two component mix of a magmatic component with R/Ra > 6.0 and -7%0 < ~13Cco2< -4%0 and a crustal component characterized by R/R, < 0.1 and ~13Cco2< - 15%orepresenting a fluid enriched in radiogenic 4He and principally organic-derived carbon (labelled a) (b) Superimposed upon this data, a small number of samples from the Ardennes and the Hunsriick-Taunus fault (labelled b) is significantly enriched in ]3C and shows the presence of a third component of crustal carbon in these groundwaters, such as might be liberated by the decarbonation of limestone.

problems of hydrocarbon migration and transport. Case studies from hydrocarbon gas fields in the N e o g e n e Pannonian Basin, Hungary, the Po Basin, Italy, and the Vienna Basin, Austria, provide the first systematic application of rare gases to the study of a hydrocarbon rich environment and are reviewed here (Ballentine 1991; Ballentine et al. 1991; Ballentine & O'Nions 1992; Elliot et al. 1993). Readers are also referred to the work of Hiyagon & Kennedy

(1992) which discusses the rare gas systematics of CH4-rich gas fields in the Alberta Basin, Canada. These studies illustrate how differently sourced rare gases are resolved using their isotopic composition, how they provide information about the extent of hydrocarbon/ groundwater interaction, constrain the mechanism of hydrocarbon transport, and provide an insight into the overall behaviour of the fluid regime in sedimentary basins.

350

C.J. BALLENTINE & R.K. O'NIONS

Vienna

( N~_~._~///~~?~..///~-"~;;~-I ~_v.~ zY//'/

v__/_~Budapest ~/~ ~

~3-----J~HAJD0SZOBOSZL6

I / , .--.,/

/~ ~--~

°

, ~

"~ .~"~'_~°~)erecske ~" E R l : : ~ - ~ S u b - B a s i n

)

,'

\ J

~-

basementisopach (kin

•

,,

S~c,e ''

%

~1~

~,~////~//////////////~/

(a)

R. Danube l

Tisza River

~::"

E

~

"---------~ - : : - ~

1000

v v v v V

•

~

o

1

(b)

o

o

2

C.-_-'~-_--

~

~

,

,

3

~

v

:,~,~,v I

1

I

0

50 km

100

4

5

V

v

~ ~ ~ ,, v ~

6

7

He, Ne & Ar IN HYDROCARBON FLUIDS A

B HAJDOSZOBOSZL6

1

-

~

"~ 3-

~

5-

/ /

/

Top Pre-Pannonian basement

(c)

__

GAS POOLS - t ' - - - -

/

_ ~xx/x~Tx

GAS POOL

~-

~"

DERECSKE / SUB-BASIN--

: : 12500km 3

FILL "~

:

~

,,o~m

:

:

._

Groundwater provenance and rare gas solubility Atmosphere-derived rare gases are present in many subsurface fluids. They are probably introduced into the subsurface dissolved in groundwater which has equilibrated with the atmosphere. The solubility of rare gases in groundwater and the information this provides about the provenance of groundwater is briefly reviewed. The relative solubility of rare gases in water is strongly temperature dependent (Crovetto et al. 1982; Smith 1985). The relative proportions of rare gases dissolved in water which has equilibrated with the atmosphere, containing a fixed rare gas composition, is therefore determined by the temperature of the water at equilibration. This is shown for atmosphere-derived Ne/Ar and Xe/Ar ratios in fresh water over the temperature range 0-25°C (Fig. 2). Waters which subsequently enter the subsurface, either buried as formation water or as part of aquifer recharge, retain this characteristic rare gas elemental abundance pattern. This property of rare gases has been used successfully in determining the palaeo-recharge temperature of old groundwater systems (e.g. Mazor 1972). Current analytical techniques enable the palaeorecharge temperature to be determined with a precision of + 0.5°C (e.g. Stute & Deak 1989). The solubility of rare gases in water also depends on salinity. While an increase in salinity does not greatly affect the relative solubility of the rare gases, the absolute solubility of the rare gases decreases with an increase in salinity (Smith & Kennedy 1983). This is illustrated in Fig. 3a, where the Ne/Ar ratio of dissolved atmosphere-derived rare gases is plotted against the Ar concentration in fresh water and sea water over a temperature range of 0-25 ° C. At any given temperature there is a marked decrease in the concentration of dissolved Ar with increasing salinity, but little change in the Ne/Ar ratio. This relationship has been used, for

351

Fig. 5. (a) The location of the Pannonian Basin is shown in relation to the areas of alpine uplift and associated basins. The expanded map shows the location of the Hajduszoboszlo gas field in relation to the basement isopachs, other oil and gas fields and the exposed pre-Pannonian basement (after Ballentine et al. 1991). (b) The SW-NE cross section shows the general features of the present day flow regimes. The shallower water is topographically driven and supported by aquifer recharge. The deeper regimes, which connect locally to the shallow flow, is saline and possibly connate. 1, Principle groundwater flow direction; 2, Quaternary; 3, Upper Pliocene; 4, Pliocene with saline pore water; 5, Upper to Middle Miocene; 6, Miocene volcanic; 7, Mesozoic or older (after Martel et al. 1989). (c) Schematic section across the Hajduszoboszlo gas field along the section A-B. This diagram emphasizes the efficiency of fluid focusing and the relative scales of transport, rather than identifies any particular fluid provenance (after Ballentine et al. 1991). example, to distinguish between formation waters originating as evaporational brines and those originating as seawater in the Palo Duro Basin, Texas (Zaikowski et al. 1987). Both groundwater palaeo-temperature and mass balance investigations require that the aqueous rare gas system has remained closed since it last equilibrated with the atmosphere. In some instances it is found that the groundwater contains more atmosphere-derived rare gases than can be accounted for by equilibrium solubility, and this is ascribed to the addition of excess air dissolved in the groundwater system (Heaton & Vogel 1981). On the other hand, if the groundwater equilibrates with another phase, such as oil or gas, after entering the subsurface, rare gases dissolved in the groundwater will partition into the oil or gas phase (e.g. Bosch & Mazor 1988). The rare gases remaining in the water will retain a characteristic rare gas abundance pattern (e.g. Zaikowski & Spangler 1990), depending on the phase equilibrated with the water, the extent of interaction between the two phases, and the temperature and salinity of the water during this equilibrium (Fig. 3b). In practice, however, it is often only the hydrocarbon phase, containing the atmospherederived rare gases, which has been sampled and is used to deduce the relationship between the hydrocarbon and any groundwater phase.

Mantle and deep crustal-derived gases within basin systems In addition to the atmosphere-derived rare gases dissolved in groundwater, it is also possible in many cases to resolve rare gas contributions from both mantle sources and radiogenic rare

352

C.J. BALLENTINE & R.K. O'NIONS

,.ofen

exposedbasementIWildendurnbach t..,,,,~.,., ~__~Joil & gas fields f-&

/

] ~;~"

t ~'"

I@"X

basement isopach L ~

i

I"

,,¢ (km)F_..=. ,,(#,/~" major faults 1 ~ ' , i'(

..,

.-,.

/

"~

v /

~,

~

~ / ~ . ~ [ [ /-'"~( F 4 / i / ~

//t

"
""?

li,A,,dej:klaa Breit/enlee

/

t

VIENNA BASIN

t •~:=

:. •

~, (.)

0

km

,..

o.

'~~

Basin ~

•

.

"

I

""

". ! "Transylvanian

400 '

//~.Basin

0

I

5

I

10 km

1

15

Fig. 6. (a) The location of the Vienna Basin shown in relation to the areas of alpine uplift and associated basins. The expanded map shows the location of the sampled gas fields in relation to the basement isopachs, other oil and gas fields and the exposed pre-Neogene basement. (b) Schematic cross-section through the Austrian section of the Vienna Basin showing sample locations in relation to the different tectonic units. Within the Central Vienna Basin the temperature gradient is c. 31 ° C km -~ and the oil generation window currently lies between 4000-6000 m (Ladwein 1988). The maturity of the Neogene sediments, and the organic content of the calcareous Alpine-Carpathian nappes are both too low for these formations to be significant hydrocarbon sources. It appears that most gas, condensate and oil in the Vienna basin were probably sourced in the Late Jurassic basinal marl, which matured only during the Miocene subsidence of the Vienna Basin (Ladwein 1988). (After Ballentine & O'Nions 1992). gases p r o d u c e d from the decay of the radioe l e m e n t s K, U and Th within the crust. U n d e r certain circumstances it is possible to estimate the age of g r o u n d w a t e r systems from the

predicted ingrowth of the radiogenic rare gases within an aquifer system (e.g. A n d r e w s 1985). H o w e v e r , it has b e e n shown that in m a n y aquifers the a m o u n t of radiogenic rare gases is

He, Ne & Ar IN HYDROCARBON FLUIDS WNW

ESE Sch6nkirchen H6flein Pottenhofen ~ s e

~

Aderklaa Breitenlee

I 4~ \ ;

|l::l

" ~ ~'/'/yr~.~

o ,s

Km

~o

Orth I

"4~ ,

"

~ A

NAPPES

~

'w~SSIF

- -

~-

353

al. 1992). Where these relationships are preserved, they are diagnostic of the source of the major gas species and of mixing between different systems (e.g. Fig. 4). Furthermore, due to the constant rate of radiogenic production of rare gases within any source rock, they also have the potential to provide age constraints on the ime elapsed since fluid was last expelled from the source system.

L I

(b)

Methane gas fields in the Po, Pannonian and Vienna basins

far in excess of that which can be reasonably produced within a closed system (Torgersen & Clarke 1985; Torgersen & Ivey 1985; Torgersen et al. 1988), while the addition of a mantle component often identifies communication with much deeper regions of the crust (Martel et al. 1989; Marty et al. 1992). The injection of deeply sourced rare gases into aquifers provides a powerful tool in the study of the hydrodynamic behaviour and mixing of different aquifer systems within any basin complex (e.g. Stute et al. 1992; Pinti & Marty 1993; Castro & Marty 1993). The rare gases must have been transported from these deep regions as a trace component in a carrier fluid phase, as the length scale of a diffusive process cannot account for the required transport distances over any reasonable time period. It is possible that the carrier fluid may be formed by the rocks themselves during any chemical changes resulting in fluid expulsion, from very low grade metamorphism producing hydrocarbons, through dehydration reactions and eventual decarbonation (e. g. Frey et ai. 1080; Walther & Orville 1982). It is also possible that externally sourced fluids, such as those derived from magmatic systems, may also provide a medium in which to transport the rare gases produced in the deep crust to shallower regions (Oxburgh et al. 1986). The radiogenically produced gases must also be available for transport away from their site of production. He is more mobile than Ar due to its higher diffusivity, and in shallow low temperature regimes is often decoupled from radiogenic Ar (Elliot et al. 1993; Ballentine etal. 1993). It would appear, however, that even with low temperature prograde metamorphism, indicated by vitrinite reflectance maturities of Ro-> 2.0, both radiogenic He and Ar are quantitatively released into any carrier phase present (Ballentine et al. 1993). It is not surprising, therefore, to find a genetic link within continental systems between He, Ar and CH4 production (Ballentine et al. 1993), He and N2 production (Tongish 1980; Jenden et al. 1988) and He and CO2 sources (Griesshaber et

The three case studies reviewed in this paper are based upon hydrocarbon gas fields from the Po (Elliot et al. 1993), Pannonian (Ballentine et al. 1991) and Vienna basins (Ballentine 1991; Ballentine & O'Nions 1992) (Figs 5, 6 & 7). The Pannonian and Vienna basin systems appear to have developed by extension in the early to mid-Miocene, creating a series of deep basins separated by shallower basement blocks (Royden 1988; Wessely 1988). Within the Pannonian Basin this was accompanied by surface vulcanism. While there is no evidence of this in the Vienna Basin, mantle 3He, which is associated with recent melt emplacement, is ubiquitous within fluids from both basins (Martel etal. 1989; Marty et al. 1992). The basins are filled by Neogene-Quaternary sediments which reach maximum depths of 7000m and 5000m in the Pannonian and Vienna Basins respectively (Berczi 1988; Jiricek & Tomek 1981). Groundwaters in the Miocene of both basins are highly saline and communicate only locally with more shallow systems (Wessely 1983; Erdelyi 1976; Ottlik et al. 1981). In distinct contrast, the Po basin is a subalpine loading basin with the sedimentary fill varying between 4 and 5 km at its northern border to more than 12km near the Appennines. Furthermore, typical of rapidly subsiding basins, geothermal gradients are generally in the range 12-20°C (Mattavelli & Novelli 1988). Consequently the organic maturity in the basin is similarly low, and below 0.5% on the vitrinite reflectance scale (Ro) to depths > 5 km (Mattavelli & Novelli et al. 1988; Pieri & Matavelli 1986). The Hajduszoboszlo gas field in the Pannonian Basin (Fig. 5) and the Dosso degli Angeli field in the Po Basin (Fig. 6) were used to investigate the detailed systematics of rare gases in single gas fields within very different geothermal and tectonic basin environments. Both are composed of a series of discrete gas pools at different depths within Neogene interbedded clays, shales and sandstones. Samples from six gas fields in the Vienna Basin, located in different stratigraphic

354

C.J. BALLENTINE & R.K. O'NIONS

Dosso Degli Angeli

J-"r

...... ..- ..... ..- .. .........

B

Pliocene-Quaternary

IT] . . . . . . Pre-Pliocene !$@Mesozoic

Fig. 7. (a) The location of the Po Basin shown in relation to the areas of alpine uplift and associated basins. he exp%ded map shows the location of the Dosso degli Angeli gas field in-relatibn to the basement isopachs, other oil and gas fields and the exposed pre-Neogene basement. (b) Schematic cross section from SW to NE across the Po Basin. The Dosso delgi Angeli gas field lies within the Pliocene-Quaternary sediments located above the faulted Mesozoic-Miocene basement. The gas field lies between 3200 and 3500 m depth. The thermal gradient in this region is particularly low c. 18°C km-' (after Elliot et al. 1993).

He, Ne & Ar IN HYDROCARBON FLUIDS c

i

=

Ii

i

,

i

i,

m

lOOO

355

l

"-" r: 2000

_o

¢" 3000

I

uO Re

8.000

b

•

iil~

•

•

•

me

;i1~

nile

i

I

II ..mml~ NII-~

__.i .,

-<,,

Radiocjenic Ne,._ r

Mantle He

5000 6000 0.0

(a)

(~.2

().4 R I Ra

i

0.6

i

0.8 0.02

(b)

,

0'.03

0.04' '

21 N e 122 N e

0105 '0.06 0

(c)

Radio,qenic Ar

1600

2b00

'

3000

4o A r 1 3 6 A r

Fig. 8. He, Ne and Ar isotope ratios from the central Vienna Basin (Ballentine 1991), plotted as a function of depth, illustrate how they can be used to resolve rare gases of different provenance. (a) R/Ra ratios are the measured 3He/4He ratio (R) normalized to the atmospheric ratio (R,). Within the Vienna basin, atmosphere-derived He contributions are negligibly small and R/R~ ratios reflect a two component mixture of mantle and crustal-radiogenic He, with end member values of R/R, = 8.0 and 0.02 respectively. Measured R/R, values of between 0.2 and 0.7 therefore represent between 2 and 8% contribution of mantle 4He tO the total 4He content. (b) 21Ne/EENeratios are close to the atmospheric value of 0.0290 in the shallowest samples. An increase in the 21Ne/22Neratio with depth is due to the addition of radiogenic 2tNe. In the deepest sample, with an 21Ne/E2Neratio of 0.05311,45% of the 21Ne is radiogenic. (c) 4°Ar/a6Ar ratios show similar behaviour to the nNe/E2Ne ratios. In the shallowest samples 4°Ar/36Ar ratios are close to the atmospheric value of 295.5. An increase in the 4°Ar/a6Ar ratio with depth is due to the addition of radiogenic 4°At. In the deepest sample, with a 4°Ar/36Ar ratio of 2460, 88% of the *JAr is radiogenic. and structural units between 0.5 and 5.5km depth (Fig. 7), provide an impression of rare gas behaviour on the basin scale.

Resolving different rare gas components The abundance and isotopic composition of He, Ne and Ar have been determined for some 50 samples of natural gas (Ballentine 1991 ; Elliot et al. 1993). Both the Pannonian and Vienna basin samples display similar rare gas isotope systematics. The Vienna Basin samples are used here to illustrate the type of information obtainable from the rare gas isotopes. In the central Vienna Basin, the measured 3He/aHe ratios (R), normalized to the atmospheric value (Ra = 1.4 X 10-6), vary between R/Ra = 0.2 and 0.7 (Fig. 8a). 2°Ne within these samples is predominantly atmosphere-derived, as subsurface radiogenic sources are negligible, and mantle contributions are small. The 4He/Z°Ne ratio in these samples varies between 290 and 32700, far greater than the atmosphere value of 0.288, which implies that the He in these samples can be considered to be a simple two-component mixture of mantlederived and radiogenic He, with negligible He contribution from atmospheric sources. Radiogenic H e has R/Ra = 0.02 and mantle He has R/Ra=8.0, therefore these samples contain between 2% and 8% mantle-derived 4He (Fig. 8a). In contrast, R/Ra values from the Dosso degli Angeli field in the Po Basin vary between 0.027 and 0.046 and can be considered to be entirely radiogenic in origin.

Similarly, the 21Ne/22Ne and 4°Ar/36Ar isotopic ratios may be used to resolve the contribution of Ne and Ar from different sources. Figure 8b and c shows the 2~Ne/22Ne and 4°Ar/a6Ar ratios in the Vienna Basin plotted as a function of depth. In the shallowest samples, both ratios are close to the atmospheric values of 21Nef12Ne = 0.0290 and 4°Ar/36Ar = 295.5. There is a general if not systematic increase in these ratios with depth to values of 0.05311 and 2460 respectively in the deepest sample. As with Z°Ne, 22Ne and 36Ar are predominantly atmosphere-derived, and because the mantle Ne and Ar contributions are small (Ballentine et al. 1991; Ballentine & O'Nions 1992), the increase in the 21Ne/22Ne and 4°Ar/36Ar ratios must be due to additions of radiogenic 21Ne and 4°Ar to an atmospherederived component. Within the shallowest samples, the contribution of radiogenic 21Ne and ~ A r is 0% and 6% increasing to 45% and 88% respectively in the deepest sample. The Po Basin samples once more contrast with these results, with 2'Nefl2Ne and 4°Ar/36Ar ratios in all samples indistinguishable from atmospheric values, indicating that there is no resolvable radiogenic 21Ne or 4°Ar within these samples.

Groundwater hydrocarbon relationships In all gas samples, the 2°Ne and 3~Ar are almost entirely derived from the atmosphere. Other contributions, either from nuclear reactions in the crust or contributions from the mantle, are

356

C.J. BALLENTINE & R.K. O'NIONS

(a) PANNONIAN AND VIENNA BASINS ~il

i

I

i

I

I

T

-600 ""-.,..o

i~

"'".

Di~

e"

CI

-1200

• Pannonian Basin0 Vienna Basin

"'-....

I

A

-1000

I

,o

-800

E

I

"""''" " % . .

g,; !

I •

%"

%-.... ""..A .... I V-.,...........• ...,.

l i

'o~ -1400 . ~

brine i~pure

-1600

1

~

"°'"'-...•••. o.

water I

0

I

10

I

I

20

I

I

3O

40

50

CH4/36Ar (10 6)

(b) PO BASIN u

-2600

p

-2800

I

1

I

~

1

brine

ll~ql/-~-i-~

..~ -3000 ¢" -3200 Q. a -3400 -3600

Ii

!

LI ~ I 0 I 1

-3800-

I

-4000 0

10

I 20

t

.... I 30

I

40

50

CH 4 / 3 6 A r (10 6) Fig. 9. (a) Variation in measured C H J 3 6 A r ratios in the Hajduszoboszlo and Vienna Basin gas fields as a function of depth. These values are compared with the CH4/36Ar ratios predicted for CH4 degassed from air-equilibrated groundwater at the reservoir pressure and temperature, for both pure water and a 5M NaCI brine (after Ballentine et al. 1991). The CH4/36Ar ratios from the Hajduszoboszlo gas field fall close to the theoretical values, suggesting that gases from this field have been transported in solution. Values from the Vienna Basin are much larger and indicate that the extent of interaction between these gases and air-equilibrated groundwater is much less. The Vienna basin gases are unlikely to have been transported in an aqueous phase. (b) Variation in measured CHj36Ar ratios in the Po Basin gas field as a function of depth. The CH4/36Ar ratios from the Dosso delgi Angeli gas field fall close to the value predicted for a gas originating from air-equilibrated groundwater at the reservoir pressure and temperature, suggesting that gases from this field have also been transported in solution. (After Elliot et al. 1993.) insignificantly small• H y d r o c a r b o n gas incorporates ZONe a n d 36Ar by interaction b e t w e e n the h y d r o c a r b o n phase a n d g r o u n d w a t e r containing dissolved a t m o s p h e r e - d e r i v e d rare gases. Concentrations of Z°Ne and 36Ar within the

h y d r o c a r b o n p h a s e t h e r e f o r e reflect the extent of equilibration b e t w e e n the gas and the g r o u n d w a t e r system. T h e ratio of CH4/36Ar can be p r e d i c t e d for a gas p h a s e which has d e g a s s e d from an

He, Ne & Ar IN HYDROCARBON FLUIDS air-equilibrated water, and is determined by the concentration of atmosphere-derived 36Ar dissolved in the water at recharge and the volume of CH4 which can be dissolved in the water under reservoir conditions (Ballentine et al. 1991). CH4 solubility in water depends on pressure, temperature (i.e. depth), and salinity. Assuming an appropriate temperature gradient and hydrostatic pressure, the predicted CH,d36Ar ratios for pure water and a 5M NaCI brine saturated with CH4 at reservoir pressure and temperature may be compared with the measured values in the Vienna and Pannonian basin gas samples (Fig. 9a) and those from the Po Basin (Fig. 9b). The CH4/36Ar ratios measured in both the Po and the Pannonian basin gas samples are similar to the range of values predicted for a gas phase which has separated from a CHa-saturated groundwater which derived its 36Ar through equilibration with the atmosphere. This is consistent with transport of CH4 in a saturated water phase, from which a gas phase separates upon cooling or decompression (e.g. Andrews & Wilson 1987). This might occur for example, at topographic highs such as occupied by the Hajduszoboszlo gas field (Fig. 5c). In contrast, samples from the Vienna Basin have CH4/36Ar ratios much higher than the range predicted for a CH4-saturated water (Fig. 5a). These values require relatively less interaction between the hydrocarbons and air-equilibrated groundwaters, and suggest that gas transport in the Vienna Basin has not occurred primarily in a CH4-saturated groundwater.

Constraints on transport The relative abundances of rare gases in an atmosphere-equilibrated groundwater, and those produced by nuclear reactions in the crust, are well defined. Any exceptional variation in these patterns must reflect the physical processes which have occurred in the subsurface, and therefore may be used to identify and quantify these processes. The patterns of fractionation of rare gases in all three Basin gas samples are very similar. Data from the Vienna Basin are used to illustrate the results obtained (Fig. 10). The radiogenic 4He/4°Ar and 4HefllNe ratios have been normalised to their crustal end-member production ratios in order to obtain a fractionation value F, for each sample. A value of F = 1 indicates that no elemental fractionation from the end-member production ratio has occurred. The normalized radiogenic ratios are compared with the atmosphere-derived 2°Ne/ 36Ar ratios in the gas samples, which is normalized to the expected ratio in groundwater.

357

Where fractionation is observed in the radiogenic 4He/4°Ar ratio, then fractionation of similar magnitude is also observed in the atmospherederived 2°Ne/36Ar ratios (Fig. 10a). Fractionation of the radiogenic 21Ne/4°Ar ratio (not shown) also occurs coherently with that of the atmospherederived 2°Ne/36Ar ratio. This coherent He/Ar, Ne/Ar fractionation implies that both crustalradiogenic and atmosphere-derived rare gases were intimately mixed before the fractionation occurred. In marked contrast, the radiogenic 4He/21Ne ratio does not show any resolvable fractionation from the predicted end member value, regardless of the extent of fractionation observed in the atmosphere-derived 2°Ne/36Ar (Fig. 10b). If a kinetic/diffusive process was responsible for the fractionation observed in the He/Ar and Ne/Ar ratios, then fractionation would also be expected in the He/Ne ratio. Diffusive processes cannot therefore be resolved in the rare gases. Diffusion must therefore be assumed to have played only a minor role in the transport of the hydrocarbon gases, and other processes must be considered to account for the observed pattern of rare gas fractionation. The partitioning of rare gases between liquid and gas phases at equilibrium depends upon the relative solubilities of the rare gases and the gas/liquid volume ratios. For example if FNe/Ar = (Ne/Ar)gafl(Ne/Ar)tiquia, FNe/Ar ~ KNflKAr as the gas/liquid volume ratio ~ 0, where FNe/Ar, KN~ and KA~ are the fractionation value of Ne/Ar in the gas phase and the Henry's solubility constants of Ne and Ar respectively. The solubilities of He and Ne in both water (Crovetto et al. 1982; Smith 1985) and oil (Kharaka & Specht 1988) are very similar, whereas the solubility of Ar is quite different. For example in a 5M NaC1 brine at 31 OK, KH~ = 476 000 atm, KNe = 472 000 atm, and KAr = 233000atm. In this case as the gas/brine volume ratio --~ 0, then FH~/N~~ 1.01, FHe/A~---> 2.04 and FNe/A~---> 2.03. Thus rare gas partitioning between an oil or water with a gas phase fractionates the He/Ar and Ne/Ar ratios by the same magnitude but would not be resolvable in the He/Ne ratios. Observed rare gas fractionation patterns in Fig. 10 are therefore consistent with liquid/gas phase partitioning. They are also consistent with the observation that both the Po and Pannonian basin gases are degassed from a water phase (Fig. 9). The Vienna Basin gases, however, do not have such a straightforward relationship with the groundwater system (Fig. 9a). It is entirely possible that they have been subject to interactions with an oil phase, either within closely associated oil deposits or with the source rocks, from which they are now separated.

358 (a)

C.J. BALLENTINE & R.K. O'NIONS 6

,

1

'

I

'

I

'

1

i

A ///

/I

(b)

i

2

A///I / / //T

T A

~'~ 4

////

,it

1

Z

/

O4

a}

=' T

"1,+ 2 LL

T T -T-

=

TT

1

2

3

4

5

6

i

L

v

I

ii iI

iI

0

-r

-r"

I:) I

0

4----

---

7

F (2°Ne/36Ar)atm

0

L

l

I 2

I 3

I 4

F (2°Ne/36Ar)atm

Fig. 10. Comparison of (4He/4°mr)rad (4He/21Ne),aaand (2°Ne/36Ar)atm, normalized to their predicted end-member ratios, and expressed as fractionation values (F). (a) Comparison of F(4He/4°Ar),adand F(2°Ne/36Ar)atm.All data fall on or close to the line of equal-fractionation, labelled A, suggesting that the radiogenic and atmosphere-derived rare gases must have been intimately mixed before rare gas fractionation. (b) Comparison of F(4He/mNe)raa,and F(Z°Ne/36Ar)atm. All data are within error of the line F(4Hef21Ne)~ad = 1, labelled B, indicating that no fractionation in the radiogenic 4HeP1Ne ratio is resolvable, regardless of the extent of fractionation in the atmosphere-derived component. This indicates that diffusive/kinetic fractionation processes have been relatively unimportant. The pattern of rare gas fractionation is typical of that predicted by a gas/liquid phase partitioning model.

Scale of fluid flow Some limits may be placed on the total volume of groundwater which has interacted with the natural gas accumulation by considering the volume of groundwater required to introduce the atmosphere-derived rare gases present in the gas fields. In addition, a crustal-radiogenic rare gas mass balance may place some limits on the crustal degassing history of the region. Such mass balance calculations have been made from the Hajduszoboszlo field (Ballentine et al. 1991). The volume of gas in the Hajduszoboszlo gas field is estimated at 33 km 3 gas STP. Given the average concentration of atmosphere-derived 2°Ne and 36Ar in the gas samples, it contains 7.3 × 103 m3(STP) 2°Ne and 1.3 × 104 m3(STP) 36Ar. The corresponding volume of water that must degas completely to account for the total reservoir volume of 2°Ne and 36Ar is 50 km 3 and 17km 3 respectively. This assumes that the groundwater originated as seawater (Erdelyi 1976) and that the concentrations of 36mr and 2°Ne in seawater at 25°C are 7.65 × 10 -7 cm3(STP)/g(H20) and 1.47 × 10-7 cm3(STP)/ g(HzO) respectively (Ozima & Podosek 1983). The difference between the water volumes calculated from the Ne and Ar concentrations reflects the fact that 2°Ne/36Ar ratio in the reservoir gas is not the same as that in atmosphere-equilibrated groundwater. In reality therefore, the volume of water must be much larger, as the fractionation observed

between 2°Ne and 36Ar in the reservoir gas requires that the gas/water volume ratio was very small, which in turn results in only partial transfer of the atmosphere-derived rare gases from the water to the gas phase (Ballentine 1991). The estimated reservoir pore volume is 0.367km3; these water volume estimates are larger than this by a factor of between 50 and 140. Assuming an average pore space of 5% at a depth of 3.5 km (Dovenyi & Horvath 1988), the m i n i m u m volume of water required to have interacted with the gas phase would occupy a volume of between 350 and 1000km 3 of the sedimentary rock. This can be placed in perspective by considering the Derescke sub-basin to the southeast of the Hajduszoboszlo field, which has a sedimentary fill of 3000 km 3 (Fig. 5c). Rare gases produced radiogenically within the crust also form a significant proportion of the total rare gas inventory in the Hajduszoboszlo gas field. Given the total volume of gas in the Hajduszoboszlo field and the average concentration of the radiogenic 4He, the gas field contains 23 × 106 m3(STP) 4He,ad. This amount of 4He is large compared to the volume which could have been produced by a-decay of U and Th within the reservoir itself, and even exceeds the volume expected to have been produced within the sedimentary fill of the Derescke sub-basin. The possibility that deeper regions of the crust have contributed to the 4He mass balance must be considered. For an average U,

He, Ne & Ar IN HYDROCARBON FLUIDS Th content of the upper crust of 2.8ppm and 10.7ppm (Taylor & McLennan 1985), and an average rock density of 2.7 cm 3 g-l, the volume of rock required to produce the measured 4nerad over 1, 10 and 50Ma is estimated at 1 x 105, 1 x 104 and 2 × 103 km 3 respectively. Over the time of basin formation, the equivalent of the entire crustal rare gas production down to a depth of 10km over the area of the Derescke sub-basin (Fig. 5c) is therefore required to produce the volume of 4Herad in the Hajduszoboszlo gas field.

Regional fluid flow overview The information now available from the rare gases associated with hydrocarbon reservoirs starts to provide a picture of the fluid behaviour within these different Neogene sedimentary basin systems. The large volume of radiogenic 4He observed in just the one Pannonian Basin gas field at Hajduszoboszlo, requires a mechanism of storage of these radiogenic rare gases within the deep crust. The release of radiogenic rare gases from the extensional Pannonian and Vienna basins is presumably related to increased heat flow during basin formation (O'Nions & Ballentine 1993) and is accompanied by mantlederived fluids. In addition, the radiogenic rare gas ratios of (4He/E1Ne)rad, (4He/'*°Ar)r~d and (21Ne/4°Ar)rad are close to the theoretical production ratios except where the atmospherederived (2°Ne/36Ar)atm is also fractionated by the same extent from its predicted end member value. The radiogenic rare gases have not, therefore, undergone fractionation from the end member values until after mixing with the atmosphere-derived component. This observation requires that the radiogenic rare gases are stored within the crust at close to their production ratios. Furthermore, both the release and transport of the radiogenic rare gases also occurs without any relative fractionation. Therefore, the mechanism of transport of the radiogenic rare gases to shallow regions within these extensional basins must be principally by advection and most probably as part of a single phase system. In distinct contrast, within the Po basin there is no observable radiogenic 4°Ar or 21Ne, only radiogenic 4He, while the presence of mantle derived volatiles is also unresolvable. This supports the observation by O'Nions & Oxburgh (1988) that loading of the continental lithosphere should not be accompanied by an influx of mantle fluids from depth. The presence of only radiogenic 4He with no accompanying radiogenic 4°Ar or 21Ne suggests that within this

359

environment preferential release of 4He over 4°Ar from the site of production on a local scale occurs and is not accompanied by a flux of fluid derived from any great depth. An obvious difference between the Po and the Vienna/ Pannonian basins is that the former has a geothermal gradient less than half of the others. The volume of water required to introduce the atmosphere-derived rare gases into the Pannonian Basin gas field is significantly larger than the reservoir volume. The radiogenic rare gases, mostly derived from the deeper crust, mix with air-equilibrated groundwater on a regional scale, after which the coherent fractionation of both radiogenic and atmosphere-rare gas abundance ratios occurs through liquid-gas phase partitioning. Clear evidence for kinetic fractionation of rare gases has not been found and diffusional transport of hydrocarbons must be assumed to be relatively unimportant in both the Vienna and Pannonian Basin gas fields. In the case of both the Dosso degli Angeli field in the Po basin and in the Hajduszoboszlo gas field within the Pannonian basin, the concentration of natural gas has occurred, at least in part, by dissolution and transport in groundwater. The Authors are very grateful for reviews by B. Sherwood-Lollar and B. Marty. This work has been supported by the Royal Society and EEC Contract No. JOUG-0006-UK.

References ANDREWS, J.N. 1985. The isotopic composition of radiogenic helium and its use to study groundwater movements in confined aquifers. Chemical Geology, 49,339-351. -& WmSON, G.B. 1987. The composition of dissolved gases in deep groundwaters and groundwater degassing. In: FRITZ,P. & FRAPE,S.K. (eds) Saline water and gases in crystalline rocks. Geological Society of Canada Special Papers, 33, 245-252. BALLENTINE, C.J. 1991. He, Ne and Ar isotopes as tracers in crustalfluids. PhD Thesis, University of Cambridge. & O'N[oNS, R.K. 1992. Mantle neon in Vienna Basin hydrocarbon reservoirs. Earth and Planetary Science Letters, 113,553-567. - - & - - 1993. The Use of natural He, Ne and Ar isotopes as constraints on hydrocarbon transport. In: PARKER,J.R. (ed.) Petroleum Geology of NW Europe: Proceedings of the 4th Conference. The Geological Society, London, 1339-1345. - - , OXBURGH,E.R., HORVATH,F. & DEAK,J. 1991. Rare gas constraints on hydrocarbon accumulation, crustal degassing and groundwater flow in the Pannonian basin. Earth and Planetary Science Letters, 105,229-246.

360 ~,

C.J. BALLENTINE & R.K. O'NIONS

MAZUREK, M. & GAUTSCHI, A. 1993. Rare gas constraints on fluid provenance and migration: the Palfris formation, Wellenberg (Switzerland). In: PARNELL,J., RUFEELL,A.H. & MOLES, N.R. (eds) Geofluidss '93. Torquay 4-7 May 1993, 293-298. BERCZI, I. 1988. Preliminary sedimentological investigation of a neogene depression in the Great Hungarian Plain. In: ROYDON,L.H. & HORVATH, F.H. (eds) The Pannonian Basin: a Study in Basin Evolution. Memoirs of the American Association of Petroleum Geologists, 45,107-116. BOSCH, A. & MAZOR,E. 1988. Natural gas association with water and oil as depicted by atmospheric noble gases: case studies from the south eastern Mediterranean coastal plain. Earth and Planetary Science Letters, 87,338-346. CASTRO, M.C. & MARTY, B. 1993. Argon isotope measurements and radiogenic 4He/4°Ar ratios in groundwaters of the Paris Basin, France. Hydrodynamic implications. In PARNELL,J., RUFFELL, A.H. & MOLES,N.R. (eds) Geofluid '93. Torquay 4-7 May 1993,290--292. CROVE1"ro, R., FERNANDEZ-PRINI,R. & JAPAS, M.L. 1982. Solubilities of inert gases and methane in HzO and in D20 in the temperature range of 300 to 600K. Journal of Chemical Physics, 76, 1077-1086. DOVENYI, P. & HORVATH, F. 1988. A review of temperature, thermal conductivity and heat flow data for the Pannonian Basin. In: ROYDON,L.H. & HORVATH,F.H. (eds) The Pannonian Basin: a Study in Basin Evolution, Memoirs of the American Association of Petroleum Geologists, 45, 195-233. ELLIOT, T., BALLENTINE, C.J., O'NIONS, R.K. & RIccmuTo, T. 1993. Carbon, Helium, Neon and Argon isotopes in a Po basin natural gas field. Chemical Geology, 106,429-440. ERDELYi,M. 1976. Outlines of the hydrodynamics and hydrochemistry of the Pannonian Basin. Acta Geologica Academiae Scientiarium Hungarieae, 20,287-309. FREY, M., TEICHMULLER, M., TEICHMULLER, R., MULLIS, J., KUNZl, B., BREITSCHMID, A., GRUNER, U. & SCHWIZER, B. 1980. Very low grade metamorphism in external parts of the central Alps: Illite crystallinity, coal rank and fluid inclusion data. Eclogae Geologicae Helveticiae, 73,173-203. GRIESSHABER, E., O'NIONS, R.K. & OXBURGH,E.R. 1992. Helium and carbon isotope systematics in crustal fluids from the Eifel, the Rhine Graben and Black Forest, F.R.G. Chemical Geology, 99, 213-235. HEreON, T.H.E. & VOGEL,J.C. 1981. 'Excess air' in groundwater, Journal of Hydrology, 50, 201216. Hn'AGON, H. & KENNEDY, B.M. 1992. Noble gas in CH4-rich gas fields, Alberta, Canada. Geochimica et Cosmochimica Acta, 56, 1569-1589. JENDEN, P.D., KAPLAN,I.R., POREDA,R.J. & CRAIG, H. 1988. Origin of nitrogen rich gases in the Californian Great Valley: Evidence from helium,

carbon and nitrogen isotope ratios. Geochimica et Cosmochimica Acta, 52,851-861. JIRICEK, R. & TOMEK, C. 1981. Sedimentary and structural evolution of the Vienna Basin. Earth Evolution Sciences, 3--4, 195-203. KHARAKA,Y.K. & SeECHT,D.J. 1988. The solubility of noble gases in crude oil at 25-100°C. Applied Geochemistry, 3,137-144. LADWEIN, H.W. 1988. Organic geochemistry of the Vienna Basin: model for hydrocarbon generation in overthrust belts. American Association of Petroleum Geologists Bulletin, 72,586--599. MAMYRIN, B.A. & TOLSTIKHIN, I.N. 1984. Helium isotopes in nature. Elsevier Science Publications, Developments in Geochemistry, 3. MARTEL, D.J., DEAK, J., DOVENYI,P., HORVATH,F., O'NIoNS, R.K., OXBURGH,E.R., STEGNA,L. & STUTE, M. 1989. Leakage of helium from the Pannonian Basin. Nature, 432,908--912. MARTY,B. & JAMBON,A. 1987. C/3He in volatile fluxes from the solid earth; implication for carbon geodynamics. Earth and Planetary Science Letters, 83, 16-26. ~, O'NIONS, R.K., OXBURGH,E.R., MARTEL,D. & LOMBARDI, S. 1992. Helium isotopes in Alpine regions. Tectonophysics, 206, 71-78. - - , TORGERSEN,Z., MEYNIER,V., O'NIONS, R.K. & DE MARSILY,G. 1992. Helium isotope fluxes and groundwater ages in the Dogger aquifer, Paris Basin. Water Resources Research, in press. MATrAVELLI, L. & NOVELLI, L. 1988. Geochemistry and habitat of natural gases in Italy. In: MATrAVELL1, L. & NOVELLI, L. (eds) Advances in Organic Geochemistry. Organic Geochemistry, 13, 1-13. MATrARELLI, L., RICCHINTO, T., GRIGRARI, D. & SCHOELL, M. 1983. Geochemistry and habitat of natural gases in the Po Basin, Northern Italy. American Association of Petroleum Geologists Bulletin, 67, 2239-2254. MAZOR, E. 1972. Paleotemperatures and other hydrological parameters deduced from noble gases in groundwaters; Jordon Rift Valley Israel. Geochimica et Cosmochimica Acta, 36, 1321-1336. O'NIONS, R.K. & BALLENTINE, C.J. 1993. Rare gas studies of basin scale fluid movement. Philosophical transactions of the Royal Society of London, A344, 141-156. & OXBURGH,E.R. 1988. Helium, volatile fluxes and the development of continental crust. Earth and Planetary Science Letters, 90,331-347. OrrLlK, P., GALFI, J. & HORVATH,F. 1981. The low enthalpy geothermal resource of the Pannonian Basin, Hungary. In: RYBACH, L. & MUFFLER, L.J.P. (eds) Geothermal systems: principles and case histories. Wiley, New York, 221-245. OXBURGH,E.R. & O'NIONS, R.K. 1987. Helium loss, tectonics and terrestrial heat budget. Science, 237, 1583-1588. & HILL, R.I. 1986. Helium isotopes in sedimentary basins. Nature, 324,632-635. OZIMA, M. & PODOSEK,F. 1983. Noble gas geochemistry. Cambridge University Press. PIERCE, A.P., Gorr, G.B. & MYrrON, J.W. 1964. -

-

He, Ne & Ar IN H Y D R O C A R B O N FLUIDS Uranium and Helium in the Panhandle gas field, Texas, and adjacent areas. United States Geological Survey Professional Papers, 454-G. PmRI, M. & MArrAVELLI, L. 1986. Geological framework of Italian petroleum resources. American Association o f Petroleum Geologists Bulletin, 70, 103-130. PrNxl, D.L. & MARre, B. 1993. Hydrocarbon transfer in the Paris Basin (France): a Helium isotope study. In: PARNELL,J., RUFFELE,A.H. & MOLES, N.R. (eds)Geofluids '93. Torquay 4-7 May 1993, 293-298. POREDA, R., JENDEN,P.D., KAPLAN,I.R. & CRAIG, H. 1986. Mantle helium in Sacramento Basin natural gas wells. Geochimica et Cosmochimica Acta, 50, 9-33. ROYDEN, L.H. 1988. Late cenogenic tectonics of the Pannonian basin system. In: ROYDON, L.H. & HORVA'rn, F. (eds) The Pannonian Basin: a Study in Basin Evolution. Memoirs of the American Association of Petroleum Geologists, 45, 27-48. SMnn, S.P. 1985. Noble gas solubilities in water at high temperature. LOS, Transactions o f the American Geophysical Union, 66,397. -& KENNEDY,B.M. 1983. The solubility of noble gases in water and in NaCI brine. Geochimica et Cosmochimica Acta, 47,503-515. STUTE, M. & DEAK, J. 1989. Environmental isotope study (14C, ~3C, 180, D, noble gases) on deep groundwater circulation systems in Hungary with reference to paleoclimate. Radiocarbon, 31, 902-918. --, SONNTAG, C., DEAK, J. & SCHLOSSER, P. 1992. Helium in deep circulating groundwater in the Great Hungarian plain: Flow dynamics and crustal and mantle helium fluxes. Geochimica et Cosmochimica Acta, 56, 2051-2067. TAYLOR,S.R. & MCLENNAN,S.M. 1985. Thecontinental crust." in composition and evolution. Blackwell Scientific, London. TONGISn, C.A. 1980. Helium - its relation to geologic systems and its occurrence with the natural gases, nitrogen, carbondioxide and Argon. Bureau of

361

Mines Report of Investigations. US Department of the Interior, 8444. TORGERSEN, T. & CLARKE, W.B. 1985. Helium accumulation in groundwater I: and evaluation of sources and the continental flux of crustal 4He in the Great Artesian Basin, Australia. Geochimica et Cosmochimica Acta, 49, 1211-1218. - & Iwv, G.N. 1985. Helium accumulation in groundwater II: a model for the accumulation of the crustal 4He degassing flux. Geochimica et Cosmochimica Acta, 49, 2445-2452. --, KENNEDY, B.M., HrVAGON, H., Cmov, K.Y., REYNOLDS, J.H. & CLARKE, W.B. 1988. Argon accumulation and the crustal degassing flux of 4°Ar in the Great Artesian Basin, Australia. Earth and Planetary Science Letters', 92, 43-56. WALTnER, J.V. & ORVILLE, P.M. 1982. Rates of metamorphism and volatile production and transport in regional metamorphism. Contributions to Mineralogy and Petrology, 79,252-257. WESSELY, G. 1983. Zur Geologie und Hydrodynamik im siidlichen Wiener-becken und seiner Randzone. Mitteitung Osterreich Geologisehe Gesellschaft, 76, 2748. - 1988. Structure and development of the Vienna Basin in Austria. In: ROYDON, L.H. & HORVATn, F. (eds) The Pannonian Basin: a Stud)' in Basin Evolution. Memoirs of the American Association of Petroleum Geologists, 45,333-346. ZAInOWSKI,A. & SeANGLER,R.R. 1990. Noble gas and methane partitioning from groundwater. An aid to natural gas exploration and reservoir evaluation. Geology, 18, 72-74. - - , KOSAt~KE,B.J. & HUBBARD,N. 1987. Noble gas composition of deep brines from the Palo Duro Basin, Texas. Geochimica et Cosmochimica Acta, 51, 73-84. ZARTMAN, R . E . , WASSERBURG, G.J. & REYNOLDS, J.H. 1961. Helium, Argon and Carbon in some natural gases. Journal of Geophysical Research, 66, 277-306.

Index

accretionary prisms dewatering 114 geological setting 113--14 acetate ions 155,157,164 concentration 182,193 stability 187, 188, 189, 190 acid-sulphates 225 advection 31 aerosols, marine 160 Ag see silver Alaska gold province 58 see also North Slope Basin Alberta Basin crude oil metal content 204 formation water salinity 154 rare gas systematics 349 uplift and erosion history 334-5 albite 134, 162,317 ailanite 308 Alleghany gold province 59 Alpine Fault 5 aluminium ion complexes 184 ion systematics 164-5 ore 178, 194 alunite 226 amino acids 178,180 amphiboles 160,308 analcime 302 andesite 3, 4 anhydrite 157,226,302 anions in formation water 154-5 see also n a m e d species

Antarctica, Bransfield Strait study 262,266 antimony in crude oils 205 apatite in red beds 308 apatite fission track analysis 326,335 use in cooling time estimation 327-8 use in palaeogeothermal gradient estimation 329 use in palaeotemperature estimation 327 aquathermal pressuring 236-7 Aquitaine Basin 47, 49 Ar/Ar isotope analysis 278 aragonite dissolution 136 architecture, modelling behaviour of 146 argon ratios 348,351,355,356-7,359 transport effects 357-8 arsenic in crude oils 205,209 asphaltic oils 203-4,208 atacamite 309, 310 Atlantis Deep 262,266 Australia sedimentary basin studies Canning Basin 339-41 Great Artesian Basin 34, 36 Murray Basin 160 Perth Basin 283 gold province studies

Cloncurry 61-6 Kalgoorlie 59 Victoria 58 Austria see Vienna Basin azurite 309 backscattered scanning electron microscopy 100 bacteria role in ore formation 281,282,309 role in reduction 157 Baffin Bay 49 Bakken Shale 242 Barbados prism 114,115,116 barite 280,293,317 barium ion distribution 171 base metal ions in ore fluids 193 basin modelling 9-12 applications method 19 results 20-1 results discussed 21-3 fluid flow 14-15 formation and evolution 12 intraplate stress 15-17 loading 13-14 Bentheim 276 bicarbonate ions 155,157, 164,225-6 Big Horn Basin 34 biodegradation 213 biomarkers 203,214 biotite in red beds 308 bismuth 280 bisulphide ions 193 bitumen biodegradation 213-14 composition 203-5 dating 285-6 defined 203,234 formation 209-10 maturation effects 213 metal interactions 205-9,281-5 migration effects 213 in MVT ore 293 occurrence authigenic mineral data 280-1 ore deposits 278-80 orogenic setting 280 reservoir 276-8 veins 275-6 role in ore exploration 286-8 bornite 306,309 Bransfield Strait 262,266 brine behaviour in basins 28, 38 defined 152 role in metal ion transport 310-11 brittle behaviour faulting 71 plastic processes compared 101 363

364 brochantite 310 bromide ions 156-7,164,205 buoyancy, role in fluid flow of 389 burkeite 303 butyrate ions 182 C isotope studies radio 268 stable 240 calcite 276,280,293 cement 131,136-7 diagenesis 239-40 formation 157 solubility 128,129 veins 240 calcium ions 155,161,194 calcrete 302 California Alleghany gold province 59 San Andreas Fault 80 San Joaquin Basin 154, 185,188,239 Canada Alberta Basin 154,204,334-5,349 Baffin Bay 49 Western Canada Basin 47-9,215,235 Canadian Shield 28, 38, 160 Canning Basin setting 339-40 thermal history 340-1 carbohydrates 180 carbon dioxide formation 175,180, 184 helium ratio 348 role in hydrothermal systems 269 carbonate anions 157 cement 131,136--7 diagenesis 239-40 eodiagenesis 303 mesodiagenesis 305 telodiagenesis 306 solubility 128,129 carboxylic acid anions 179-80 di- 184-5 mono- 182-4 Cascadia prism 115, 116 cataclastic flow 104 cathodoluminescence 100 cations in formation water 155 see also named species

cementation processes 130, 131-2,136-7 red beds 304 sources 128-9 cerussite 313 chalcocite 306 chalcopyrite 205,306,309 chelate compounds 177 see also ligands

Chile prism 114,116 chloride ions in crude oils 205 role in diagenesis 166 role in transport 309-10

INDEX systematics 155,156, 157, 164,224 enthalpy plot 229 chlorinity, role in dissolution of 127, 128 chlorite 302 chromium in crude oils 205 cinnabar 279-80 clay minerals dehydration 236--7 effect on metalloporphyrins 211 clinoptilolite 302 Cloncurry gold province metasomatism 61-2 gold provinces compared 65-6 reactions 62-5 cobalt 179,205,306 collision tectonics 4-5 Colorado 33--4 Colorado Plateau 306 compaction and fluid flow 37-8 concentration gradients 127 conduction, thermal 128 connate water 152,224 continental crust equations 28-31 fluid flow controls 33-9 heat transport 31-3 permeability 27 convection, thermal 1,2, 131,269 cooling time analysis fission track method 327-8 vitrinite reflectance method 328 coordination compounds 177 copper in metalloporphyrins 206 ores 178,179 deposition 278,282 Lubin deposits 309 red beds 301-2,306,307 redox controls 309-10 sources 169-70 Corella Formation 62, 63 Cornish Granite 160 coseismic strain 87-8 cross bedding 146 crude oil biodegradation 213-14 composition 203-5 formation 209-10 maturation 213 metal ion species 205-9 migration 213 crystobalite 226 D isotope studies 165 Darai Limestone 335,336,337 Darcy flow 245 decarbonation 55, 57, 58 decarboxylation reactions 187-8 decollement 114 deformation environmental effects metamorphic conditions 107-8 sedimentary basins 107 factors affecting 101

INDEX deformation---cont'd

mechanisms cataclastic flow 104 diffusive mass transfer 105 dislocation creep 105-7 fracture 101-4 independent particulate flow 104 methods of analysis 100 permeability effects 113, 115-17 Darcyan v. dynamic 117-19 measurements 117 relation to microstructure 119-20 relation to volume change 120-2 degassing reactions 3 dehydration 55, 57-8,236-7 Denver Basin 33-4 desulphidation 55, 58 detachment zone 114 devolatilization 55, 57-8 dewatering 4 diagenesis carbonate rocks 136-7 effect on permeability 118 effect on petroleum 239-41 migration 277-8 pore water behaviour 127-8 effect on calcite 129 effect on quartz 129-30 flux 128 low temperature 133-5 rates of reaction 165 red beds eodiagenesis 302-4 mesodiagenesis 304-5 pH controls 313 redox controls 313 telodiagenesis 305-6 thermodynamic modelling 313-16 role of fractures 131-2 role of organic matter 191-3 thermal effects 131 diagenetic/metamorphic boundary 55-6 dickite 226 diffusion equations 30-1 role in petroleum migration 127,234,243 diffusive mass transfer (DMT) 105 dilatancy and stress cycle 74, 76--7 effect of fluid pressure 76 effect of mean stress 75-6 effect of shear stress 75 dilatancy-diffusion hypothesis 74 dislocation creep 105-7 dismigration 234 disulphides 178,179 Dogger Formation 45 Doherty Formation 63 dolomite 136-7,280,293,317 dynamic fluid viscosity 29 dynamic permeability s e e permeability as a dynamic concept earthquake stress cycle and fluid flow 74, 86--7 dilatancy 74-7

365

fault valves 78-80 post-seismic redistribution 77 East African Rift 49 hydrothermal system 262,266,267,268 East Pacific Rise 262,266 eduction 4 Eh s e e redox potential electron microprobes 100 elevation, significance of 5-6 eodiagenesis 302-4 epidote 5,308 erionite 302 erosion rates 5 estimation of 329-31,333-5 Escanaba Trough hydrothermal system 262,263,264-5 petroleum migration history 267,268 evaporation index (EI) 155 evaporation processes 155-7 evaporite dissolution 157-8 explosive migration 237 expulsion efficiency 250-1 numerical modelling 251-3 expulsion fractures 236,242 extension fractures 71 fatty acids 180 fault gouge 104 fault-valves 78--80 fault/fracture mesh 72-4 faults dilatancy effects 74 fluid pressure effects 70 modelling of 80 hydrology effects 85, 86-7, 91 permeability effects 70-1 modelling of 146-7 Fe s e e iron feldspar dissolution 133-5,302,304 in red beds 308 s e e a l s o albite a l s o plagioclase ferromagnesian minerals 302 ferrous ions 193 Fischerschiefer 243 fission track analysis s e e apatite fission track analysis flood basalt 2-3 flow studies deformation effects 101 mechanisms cataclastic flow 104 diffusive mass transfer 105 dislocation creep 105-7 fracture 101-4 independent particulate flow 104 methods of analysis 100 environmental effects metamorphic 107-8 sedimentary 43,107,131 foreland basin studies 331-5 Papuan Fold Belt 335-8 mathematical model 14-15 measurement 358-9 permeability effects 113,115-17

366

INDEX

flow s t u d i e s - - - c o n t ' d measurements 117 Darcyan v. dynamic 117-22 stress cycle effects 74 mechanisms 74-80 upper crustal studies descriptive equations 28-31 mechanisms 33-9 fluid inclusions 293,309 fluids flow s e e flow studies mixing effects 316-18 pressure effects 70, 313-16 sources 55-7, 69 fluidized bed injection 4 fluorite 293 foreland basins fluid flow 331-5 heat transfer properties 46-9 Papuan fold belt 335-8 formate ions 182,188 formation water 224,231 fractionation, role in petroleum migration of 249-51 fractures cementation 131-2 effect on fluid flow 101-4,131,236,241-2 permeability effects 71 France 45-6, 47, 49, 154 Frio Formation 242,298 Fulmar Sandstone 134 fulvic acid 177-8,180 galena 179,280,306 gallate ions 188 gangue minerals 317 garnet in red beds 308 gas formation 175,184 Gavarnie Thrust mylonite 102,104,106 Germany 243,276 geochemistry of ocean water 2 geoporphyrins 206 geopressure zones 295-6 role in ore formation 296-8 geothermal energy 302 geothermal fluids 222-3 classification 223-4 bicarbonates 225-6 chlorides 224 sulphates 225 hydrological controls 226-7 effect on gold ore 227-9 effect on oilfield water 229-31 steam generation 226 zone recognition 227 water origin 224-5 geothermal gradient analysis 328-31 glutarate ions 184 goethite 226 gold 282 ions in ore fluids 193 ore formation 178,227-9 provinces fluid behaviour 58--61

regional study 62-5 granite and aquifers 5 granodiorite 5 Great Artesian Basin 34, 36 Great Plains Province 34 Green River Shale 204,206 greenstone gold province 58 greywackes 58 groundwater flow rates 358 interaction with rare gases 355-7 Guayamas Basin hydrothermal system 262-4,270 petroleum expulsion 267,268 gypsum 157,303 halite in red beds 302,303,305 halloysite 226 heat transport 1 conduction 128 convection 131 descriptive equations 31-3 s e e also thermal characteristics Hebgen Lake 86--7 helium cycling 2 gas field study 353,355 ratios in natural gases 348,359 transport 357-8 hematite 226,280 red beds 306,317-18 Hg s e e mercury Hill fault/fracture mesh 72-4 hot spots 2 Huldra Field 278 humic acid 177--8,180 Hungarian Basin 52 s e e also Pannonian Basin hydraulic conductivity 29-30 hydraulic diffusivity 31 hydraulic fracturing 235-6 hydraulic head causes 69 measurement 28-9 hydrocarbon solubility 167 hydrocarbons s e e gas also oil a l s o petroleum hydrofracturing 27-8 hydrology, effect of faults on 85, 86-7 hydrosphere buffering 1 hydrothermal eruption breccia 228 hydrothermal systems classification 261,262 effect of organic matter continental 266-7 submarine 262-6 expulsion 267-9 fluid interactions 269 hydrothermal veins 69 hydrous pyrolysis 237 hydroxides, role in sorption of 318 hypersaline waters see salinity and saline waters hyposaline waters 160 Ieru Formation 335,336,337

INDEX Illinois Basin 154,158 ores 295 illite effect on reservoir properties 278,281 formation 134 illitization 236,278,281,302,304 Imburu Formation 336 immiscible flow s e e oil-water flow independent particulate flow 104 inert gases s e e rare gases intracratonic basins 45--6 iodide ions 205 ion microprobe 100 Ireland, ore exploration in 286,288 Irish Sea Basin 278 iron ions 184, 193,204,205,206 minerals 307-8,309 ores 178,179, 194,221 isotope studies radio dating 166,278,285,286,294 stable composition 59,165-6,216,240 Italy s e e Po Basin j arosite 226 Jean d'Arc Basin 50 Juan de Fuca plate 3 juvenile water 224 K/Ar dating 278 Kaiko Project 3 Kalgoorlie gold province 59 kaolinite 225,280,306 dissolution 240 formation 133 134 in red beds 308 Kebrit Deep 262,266 Kennedy Basin 33 kerogen catagenesis 205,209 degradation 186-7 role in petroleum migration 245,247 Keuper Sandstone 278 Kimmeridge Clay 240 Kupferscheifer 309 La Luna Shale 215,242 laser ICPMS 100 laser microprobe stable isotope analysis 100 lead ores 169-70,179,221 characteristics 293 Mississippi Valley Type history of study 293-5 source model 295-7 summary 297-8 in red beds 302,306,307 transportation 310-12 ligands action 177-80 defined 177 destruction 187-8 distribution 180--6 origin 186-7 reaction thermodynamics 188-91

role in diagenesis 191-3 role in ore formation 193-4 lithofacies analysis 239 Louisiana Basin 188 formation water study 154, 159,161,164 Magellan Basin 243 magmatic water 224-5 magnesium ions 155,156, 157,161 complexes 194 magnetite in red beds 308 role in decarboxylation 187-8 malachite 309-10 maleate ions 185 maleic acid 184 malonate ions 185,188, 189,190, 193,194 malonic acid 184 manganese 205,221 mantle convection 2, 3--4 water 2 Manville Formation 48 Maracaibo Basin 215 marcasite 205 Maronan Supergroup 63 mathematical modellings e e modelling maturation 213 Melones fault zone 69 membrane filtration 158-9 mercaptans 178,179 Mercia Mudstone Group 278,307 mercury in crude oils 205,209 ions in ore fluids 193 ore deposits 279-80,288 mesodiagenesis304-5 metal-bearing minerals 307-8 metal ion complexes 175 characteristics 205-8 in crude oil 203 evolution initial 209-13 secondary 213-14 organic complexes 170-1,177,281 metal fulvate 177-8 metal humate 177-8 modelling behaviour diagenesis 191-3 ore formation 193-4 s e e a l s o metalloporphyrins transport 308-13 metal ores 169-70 exploration 286-8 metalloporphyrins 206-8 as biomarkers 214-16 stability 210-13 metamorphic fluids 3 defined 55,224 role in gold ores 65--6 Cloncurry Province 62-5 sources 58--61 devolatilization 57-8 metamorphic/diageneticboundary 55-6

367

368

INDEX

metasomatism and gold ores 65-6 Cloncurry province 62-5 meteoric water defined 224 flux 133-5 role in carbonate diagenesis 136-7 methane formation 3 rare gas mixtures 353-5 ratios 348 sources 355-7 role in hydrothermal systems 269 Mexico, Gulf of 296 mica dehydration 57 hydrolysis 160 see also biotite also muscovite Michigan Basin formation water 10,154, 164 heat transfer 45, 46 micro-organisms 1 microcracks 71 microfracturing 235-6 microstructures 119-20 Mid Atlantic Ridge 262,266 Middle Valley hydrothermal system 262,265-6 petroleum expulsion 267,268 migration 213 see also primary migration also secondary migration mineralization 93-4 red beds 313 fluid flow modelling 316-18 sorption 318--20 thermodynamic modelling 313-16 Mississippi Valley Type (MVT) ores 193,221,279,280 characteristics 293 origin history of study 293-5 source model 295-7 summary 297-8 temperature controls 309 setting 293 Missouri groundwater study 33 modelling basin applications method 19 results 20-1 results discussed 21-3 fluid flow 14-15 formation and evolution 12 intraplate stress 15-17 loading 13-14 rift shoulder erosion 12-13 diagenesis 191-3 fault zones 80 flow 252 fracturing 252-3 MVT ores 293-7 organic ligands 191--4 petroleum pathways 253,254 expulsion 251-2 oilfield water 229 red beds

fluid mixing 316-18 mineral-fluid equilibria 313-16 strain 87-8 thermal history 326 two-phase flow 144-9 Molasse Basin 243 monazite 283,308 Monterey Shales 73,239 montmorillonite 226 role in decarboxylation 187-8 mordinite 226 Mount Isa 61,306 mud volcanoes 4 Murray Basin 160 muscovite 308,317 mylonite 102--4 Na see sodium Nankai prism 114,115,116,117 143Nd/144Ndratios 216 neon ratios 351,355,356-7,359 transport effects 357-8 Neuquen Basin 276 New Albany Shale 204,206 New Zealand 5,204,216,262,266-7 nickel 179,276 effect of biodegradation 213-14 effect of maturation 213 effect of migration 213 occurrences 204-5,206,208-9,210 use as biomarker 214-16 nitrogen : helium ratios 348 North Sea Basin 9, 49,278 North Slope Basin 34-6, 37, 50-2 petroleum geochemistry 215 Nova Scotia Shelf 51 O isotope studies 165-6,240 ocean floor flux 2 ocean water geochemistry 2 volume 3 oil-water flow 141-2 equations governing 142-3 experimental study 143-4 modelling 144-6 anisotropy effects 148 pervasive faulting effects 146--7 results discussed 148-9 sediment architecture effect 146 oilfield water composition 180-1 modelling behaviour 229-31 opal phases 133 optical microscopy 100 Oregon 326 Oregon prism 116 ores deposition 278-80 exploration 286-8 formation modelling 193--4 role of organic matter 175-6 mineralogy

INDEX ores-----cont'd

relation to formation water 169-71 sedimentary 132-3 see also n a m e d m e t a l s

organic acids 167--8, 180 destruction 187-8 method of measuring 181-2 origin 186-7 see also carboxylic acid organic matter 175-6 accumulation continental 266-7 submarine 262-6 chemical behaviour compound distribution 180-6 compound origins 186-7 compound types 177-80 ligand reaction thermodynamics 188-91 destruction 187-8 effects of hydrothermal system 261,262,270 expulsion 267-9 fluid interactions 269 modelling behaviour diagenesis 191-3 ore formation 193-4 role in diagenesis 209-10 role in formation water salinity 167-9 role in metal ion complexes 170-1,281 organometallic compounds 177 organosulphur ligands 178-9 Orubadi Formation 335,336,337 osmotic flow 158 Ouachita Mts 276,280 overburden effect on migration 236 overpressure 18-19 oxalate ions 184, 185,188,189,190,193,194 oxides, role in sorption of 318 oxygen fugacity see redox potential Ozark Mt ores 295 palaeogeothermal gradient analysis 328-31 palaeosurfaces, significance of 227,228 palaeotemperature analysis fission track method 327 vitrinite reflectance 326-7 palladium ores 178,179 Pannonian Basin 9, 11 fluid flow 18-19 hydrocarbons 18-19 modelling methods 19 results 20-1 results discussed 21-3 origin 17,353 rare gas occurrence 353,358 subsidence 18 Papuan fold belt setting 335 stratigraphy 335-6 thermal history 336-8 Paris Basin formation water salinity 154 heat transfer properties 45-6 passive margin, heat flow 49, 50

Pattani Basin 154 Pb ore see lead 2°7pb/2°6pb dating 285 pelite metamorphism 57 peptides 180 peridotite deserpentinization 5 serpentinization 2 permeability deformation-created changes 101,104-5 as a dynamic concept Darcyan v. dynamic 117-19 effect of microstructure 119-20 effect of volume 120-2 introduction 113, 115-17 measurement 117 effect of diffusive mass transfer 105 effect on faults 70-1 effect on hydraulic conductivity 29-30, 37 experimental study 143-4 mathematical modelling 144--6,148--9 anisotropy effect 148 pervasive faulting effect 146-7 sediment architecture effect 146 measurements 242-5 continental crust 27 sandstone 131 shale 131 relation to stress field 71 role in immiscible fluid flow 141-2 equations governing 142-3 Perth Basin 283 Peru prism 114, 116 petroleum hydrothermal generation 261,262,270 continental 266-7 submarine 262-6 migration see primary petroleum migration pH organic matter effects 175,180 pore water 161 role in red bed diagenesis 313,316 role in sorption 320 phenols 180 Phosphoria Shale 215 plagioclase dissolution 134 in red beds 308 plastic behaviour, brittle processes compared 101 platinum 178,179,282 Po Basin 353-5 pore size classification 243 pore water characteristics 127-8 composition 151-2 flux calculation 128 low temperature reactions 133-5 mixing effects 137 porosity deformation-created changes 101 effect on cementation 130 effect of depth 19 effect on diagenesis 56 effect of diffusive mass transfer 105

369

370

INDEX

porosity----cont' d

effect of dislocation creep 105 effect of dissolution 240 porphyrins 179 as biomarkers 214-16 metallic ion links 206-7 stability 210-13 Posidonia Shale 215 post-seismic redistribution 77 potassium cations 155,161 pressure solution, role in petroleum migration of 241 pressure/temperature controls hydrothermal systems 269 Mississippi Valley Type deposits 309 role in diagenesis 314 primary petroleum migration controls diagenesis 239-41 fractures 241-2 hydrothermal system 267-9 lithology 238-9 defined 233--4 driving forces 236-7 expulsion efficiency 250-1 fractionation effects 249-50 laboratory simulation 237-8 modelling 251-3 modes 234-6 pathway analysis geochemical 246-9 petrophysica1242-6 prism taper angle 114 propionate ions 182, 189,193,194 pyrite 205,226,280,293,306 from hematite 305 as reductant 309 role in decarboxylation 187-8 pyrolysis 186-7,245 pyroxene 308 quartz 293 cement mineral 129--30, 131,134,304-5 formation water content 165 gangue mineral 317 role in decarboxlyation 187-8 solubility 128 radio-isotopes see under isotopes Rannoch Formation 148 rare gases groundwater relationships 355-7 levels in gas field study 353-5 sources 351-3 transport methods 357-8 use in fluid flow studies 358-9 Rayleigh convection 131 Rb/Sr dating 294 red beds diagenesis 313 eodiagenesis 302-4 mesodiagenesis 304-5 telodiagenesis 305-6 thermodynamic modelling 313-16 formation 281-2

mineralization 306 fluid flow 316-18 metal distribution 307-8 metal transport 308-13 sorption 318-20 origin 301 related base metal (RBRBM) deposits 193 Red Sea 262,266 Red Sea Rift 49 redox potential effect on metal ion complexes 203,215 effect on metal precipitates 281,309 effect of organic matter 175,180 formation waters 168-9 role in diagenesis 313,316 eodiagenesis 303-4 mesodiagenesis305 role in sorption 320 reduction reactions 304, 305,309 reverse osmosis 158--9 Rheingraben 34, 49,325 rift basins and heat transfer properties 49-52 Rikuu earthquake 86-7 river flow and earthquakes 86-7 Rough Rock Group 285 rutile 205 safflorite 209 Salina Formation 154, 160 salinity and saline waters composition 154, 160-6 defined 152 effect of organic matter 167-9 history of study 152-3 origin 155--60 pH effects 161 role in dissolution 127,128 role in gas solution 351 San Andreas Fault 80 San Joaquin Basin 185,188 formation water 154 porosity study 239 sandstone diagenesis 130 permeability modelling 144-6, 148-9 anisotropy effects 148 pervasive faulting effects 146-7 sediment architecture effects 146 saturation, role in petroleum migration of 245--6 Saxony Basin, Lower 243 Sb see antimony sea water evaporation 155-7 secondary electron scanning electron microscopy 100 secondary migration 234 sediments sedimentary basin studies characterization 44-5, 52 foreland 46-9 intracratonic 45--6 rift 49--52 fluid flow 43 thermal characters equations 44 time factors 43-4

INDEX seismic pumping 74 selenium in crude oils 205,209 serpentinization 2, 5 Shaban Deep 262,266 shear stress and dilatancy 74, 75 shear zones and microstructural analysis 119,120 Sherwood Sandstone Group 278,302 diagenesis 302,306 silica sinter 226,228 silver 193,282,306 Silvermines ores 288 skutterudite 209 ~47Smgeochemistry 216 smectite dehydration 236 formation 134 replacement reactions 302,304,307 sodium in crude oils 205 ions and complexes 155,161,194 solubility, factors affecting 128 sorption 318-20 role in petroleum migration 245-6 South Africa 58 South Caspian Basin 37 SPEX method 237 sphalerite 179,280,306 spilite 3 Sr dating 59,166 steam zone 226,227 stibnite 280 strain cycle 92-3 modelling 87-8 stress cycle and dilatancy 74-7 map of Europe 10 static field effects 71 strike-slip faulting 4-5 strontium dating 59, 166 ion concentrations 161 stylolites relation to stress field 71 role in petroleum migration 241 subduction 3-4 erosion 114 succinate ions 184, 191,194 suction pumps 77 sulphate anions 155,157,164,225 sulphides ions 193 ores 221,282 role of organic matter 179 role in red bed diagenesis 317 sulphur 205,226 Swiss Molasse Basin 47, 49 sylvite 302 Tanganyika Lake hydrothermal system 262, 266,267, 268 Taranaki Basin 204,216 tectonic erosion 3 tectonic fractures 242

telodiagenesis 305-6 temperature role in diagenesis 314 role in dissolution 127,128 temperature/pressure controls hydrothermal systems 269 Mississippi Valley Type deposits 309 role in diagenesis 314 tenorite 310,313 tetrapyrrole 179 thermal characteristics descriptive equations 44 sedimentary basins 52 foreland 46-9 intracratonic 45-6 rift 49-52 time effects 43--4 thermal conduction 128 thermal convection 1, 2, 131,269 thermal diffusivity 44 thermal history modelling 326 cooling time estimation fission track analysis 327-8 vitrinite reflectance 328 fluid flow analysis foreland basins 331-5 Papuan fold belt 335-8 intrusions 338-9 Canning Basin 339-41 maximum palaeotemperature estimation fission track analysis 327 vitrinite reflectance 326-7 palaeogeothermal gradient 328-31 transient v. steady state effects 341-2 thermodynamic buffering 162 thermodynamic modelling 313-16 thiols 178,179 thiophene 178, 179 thorium ores 282-4 thrusts 3, 5 titanite in red beds 308 tonalite 5 topography and fluid flow 33-7 Toro Sandstone Formation 335,336,337 total dissolved solids 152,157 total organic carbon 245 trace elements crude oil content 203-4 in porphyrins 206-8,210-13,213-14 Trans-Atlantic Geotraverse 262,266 travertine 226 trona 303 two-phase fluid flow 141-2 equations 142-3 experimental study 143-4 mathematical modelling 144-6 anisotropy effects 148 pervasive faulting effects 146-7 results discussed 148--9 sediment architecture effects 146 Tynagh ores 286 U/Pb dating 286 Uinta Basin 34, 47, 49,326

371

372 ultrafiltration 158-9 underplating 114 underthrusting 4--5 uplift estimation 32%31,333-5 uranium ores 178 effect of bitumen 278,279,282,284 in red beds 302,306, 307,309, 310 use in dating 285,286 USA basin studies Big Horn 34 Denver 33-4 Illinois 154, 158 Kennedy 33 Louisiana 154, 159, 161,164, 18 Michigan ] 0, 45, 46, 154, 164 Missouri 33 Uinta 34, 47, 49,326 Williston 34, 36, 46,243 gold provinces Alaska 58 Alleghany 59

INDEX volatiles inventory 2 volume change effects in permeability 120-2 Waiotapu 262,266-7 waste disposal problems 1 water inventory 1-2, 69 role in global cycle 2 role in hydrothermal systems 269 rote in ocean recycling 2 sources 224--5 water table 33-7 Waterstones 278 well logs, use of 238 Wessex Basin 302,305 West Siberia Basin 276 Western Canada Basin heat transfer properties 47-9 petroleum analyses 215,235 wettability, role in petroleum migration of 246 Williston Basin 34, 36, 46,243 Witwatersrand gold province 58 xenon ratios 351

valerate ions 182 vanadium ions 179,276 effect of biodegradation 213-14 effect of maturation 213 effect of migration 213 occurrence 204-5,206,210 use as biomarker 214-16 ores 278,282 Vancouver prism 115, 116 Venezuela ores 287 Victoria gold province 58 Vienna Basin 353-5 vitrinite reflectance (VR) 293,326 measurements 335 use in cooling time analysis 328 use in palaeogeothermal gradient analysis 329 use in palaeotemperature analysis 326-7 volatile fatty acid (VFA) 180

Yellowstone National Park 262,266 zeolites 302 zinc ion behaviour 178, 179,193,194,205 ores Mississippi Valley Type ore 221,293 history of study 293-5 source model 295-7 summary 297-8 red beds mineralization 302,306,307 transportation 310-t2 sources 169-70 zincite 313 zircon 205,283 in red beds 308

Geofluids: Origin, Migration and Evolution of Fluids in Sedimentary Basins edited by John Parnell (Queen's University of Belfast, UK)

Geological fluids are a central theme linking the petrography and chemistry of all rock types, deformation processes on the microscopic to the continental scale, and the concentration of economic resources. The fundamental importance of fluid migration and evolution to rock composition and structure is reflected in a growing interest in fluid processes, including a series of successful conferences on water-rock interaction. The papers in this volume are intended to give a review of the whole spectrum of current geofluids research. The papers include international case studies and are written by leading experts in the field. • • • • • •

First overview of fluids research Examples of commercial applications of fluids research Reviews research for the non-specialist Up-to-date extensive bibliographies Promotes methodological transfer between oil and minerals industries Explains theory behind and consequences of fluid flow processes

This volume will be of interest to geologists in the oil and minerals industries and in academia, and to hydrogeologists and geochemists.

• 168illustrations • 374 pages • 23 chapters • index

Cover illustration: Multi-phase hydrocarbon bearing fluid inclusion within a quartz crystal, Dolyhyr, Wales.