Water Resources of Arid Areas

WATER RESOURCES OF ARID AREAS

PROCEEDINGS OF THE INTERNATIONAL CONFERENCE ON WATER RESOURCES OF ARID AND SEMI ARID REGIONS OF AFRICA (WRASRA), AUGUST 3–6TH 2004, GABORONE, BOTSWANA

Water Resources of Arid Areas Edited by

D.Stephenson Civil Engineering Department, University of Botswana, Gaborone, Botswana E.M.Shemang & T.R.Chaoka Department of Geology, University of Botswana, Gaborone, Botswana

A.A.BALKEMA PUBLISHERS LEIDEN/LONDON/NEW YORK/PHILADELPHIA/SINGAPORE

Copyright © 2004 Taylor & Francis Group plc, London, UK All rights reserved. No part of this publication or the information contained herein may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, by photocopying, recording or otherwise, without written prior permission from the publisher. Although all care is taken to ensure the integrity and quality of this publication and the information herein, no responsibility is assumed by the publishers nor the author for any damage to property or persons as a result of operation or use of this publication and/or the information contained herein. Published by: A.A.Balkema Publishers, a member of Taylor & Francis Group plc http://www.balkema.nl/ and http://www.tandf.co.uk/ This edition published in the Taylor & Francis e-Library, 2005. “To purchase your own copy of this or any of Taylor & Francis or Routledge’s collection of thousands of eBooks please go to http://www.ebookstore.tandf.co.uk/.” ISBN 0-203-02340-4 Master e-book ISBN

ISBN 04 1535 913 9 (Print Edition)

Table of Contents Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

Preface

xii

Keynote address Sustainable water management in arid and semi-arid environments W.Kinzelbach, P.Bauer, P.Brunner & T.Siegfried

3

Theme A: Problems in obtaining hydrological and geo-hydrological data Slug tests in fractured rock formations: value, pitfalls and misinterpretations P.D.Vermeulen & G.J.van Tonder Flow simulation model performance assessment using entropy approach A.M.Ilunga & D.Stephenson Data collection experiences in water level monitoring, borehole archive and research projects in semi arid Botswana M.Magowe, T.Obakeng & P.Makobo Rainfall characteristics in semi-arid Kitui district of Kenya A.O.Opere, V.O.Awuor, S.O.Kooke & W.O.Omoto Quantification of the impact of irrigation on the aquifer under the Vaalharts Irrigation Scheme R.G.Ellington, B.H.Usher & G.J.van Tonder

21

29

36

43 60

Theme B: Groundwater recharge: natural and artificial Groundwater development—identification of artificial recharge areas in Alla, Eritrea K.S.Viswanatham, F.Tesfaslasie, M.Asmellash, A.Kumar & S.A.Drury

75

Subterraneous injection of nutrient rich groundwater to the coastal waters K.K.Balachandran & J.S.Paimpillil A new method for the estimation of episodic recharge J.Bean, G.van Tonder & I.Dennis Prioritisation of the impacts of pollutants on groundwater flow systems in South Africa I.Dennis, B.Usher & J.Pretorius Understanding problems of low recharge and low yield in boreholes: an example from Ghana A.J.E.Cobbing & J.Davies Spatial variation of groundwater recharge in semi-arid environment—Serowe, Botswana L.M.Magombedze, B.Frengstad & M.W.Lubczynski Quantification of artificial ground water recharge G.C.Mishra The architecture and application of the South African Groundwater Decision Tool I.Dennis & G.J.van Tonder The development of a groundwater management tool for the Schoonspruit dolomitic compartment B.H.Usher & S.Veltman Effects of mining and urban expansion on groundwater quality in Francistown, Botswana B.Mafa & H.Vogel In situ remediation potential for Southern African groundwater resources S.Clarke, G.Tredoux & P.Engelbrecht Coastal aquifers intrusion at semi-arid region of Turkey L.Yilmaz Evaluation of groundwater recharge rates in the Kizinga catchment in Dar es Salaam region Y.B.Mkwizu & H.H.Nkotagu

85

92 99

109

122

133 145

156

168

181

191 198

Theme C: Socio-economic aspects

KNUST experiences in capacity building in the water and sanitation sector S.N.Odai, F.O.K.Anyemedu, S.Oduro-Kwarteng & K.B.Nyarko

213

Strategic partnerships for sustainable water education and research in developing countries S.N.Odai, K.A.Andam & N.Trifunovic Assessing demand for clean and safe domestic water in eastern Zimbabwe E.Manzungu, M.Machingambi & R.Machiridza The role of supplementary irrigation for food production in a semiarid country—Palestine M.Y.Sbeih Conversion of priority water rights to proportional water permits and conflict management in the Mupfure river catchment, Zimbabwe T.Mpala Impacts of water development in arid lands of Southern Africa: socio-economic issues J.P.Msangi Institutional challenges for small towns’ water supply delivery in Ghana K.B.Nyarko Socio-economic performance of Sepeteri irrigation project in Nigeria O.O.Olubode-Awosola & E.O.Idowu

221

227

240

254

262

275

287

Theme D: Application of geophysical, GIS, and remote sensing techniques

Mapping vegetation for upscaling transpiration using highresolution optical satellite and aircraft images in Serowe, Botswana Y.A.Hussin, D.C.Chavarro, M.Lubczynski & O.Obakeng Gravity study on groundwater structure in Central Butana (Sudan) K.M.Kheiralla & A.E.Ibrahim Remote sensing and electrical resistivity studies on groundwater structure zones in Central Butana (Sudan) K.M.Kheiralla & A.E.Ibrahim Monitoring and modeling of fluxes on Kalahari—setup and strategy of the Kalahari Monitoring project. Serowe study case, Botswana M.W.Lubczynski & O.Obakeng Geoelectrical investigation for aquifer delineation in the semi-arid Chad Basin, Nigeria A.Iliya & E.M.Shemang

302

313 330

346

357

Monitoring of evapotranspiration on Kalahari, Serowe case study, Botswana O.Obakeng & M.W.Lubczynski Electro-seismic survey system S.R.Dennis, M.du Preez & G.J.van Tonder Borehole site investigations in volcanic rocks of Lolmolok area, Samburu district, Kenya J.K.Mulwa Groundwater evaluation in a complex hydrogeological environment—a GIS based approach B.Mudzingwa, J.L.Farr, R.Gumiremhete & T.Kellner Application of 2-D resistivity imaging combined with time domain electromagnetic survey to map shallower aquifers in Kunyere valley, northwest Botswana E.M.Shemang, H.Kumar & J.Ntsatsi

364

381 388

406

420

Theme E: Climate change and its impact Hydraulic studies in the design of sand dams A.S.Nzaba, H.O.Farah, T.C.Sharma & C.W.M.Sitters Designing and implementing an aircraft survey mission using highresolution digital multi-spectral camera for vegetation mapping for upscaling transpiration of Serowe, Botswana Y.A.Hussin, M.W.Lubczynski, O.Obakeng & D.C.Chavarro Relevance of groundwater interaction with surface water to the eco-hydrology of semi-arid regions J.Y.Diiwu Impacts of climate change in water resources planning and management A.Opere Turning a liability into an asset: the case for South African coalmine waters B.H.Usher & F.D.I.Hodgson Environmental hydrogeology of the dolomite aquifer in Ramotswa, Botswana M.Staudt & H.Vogel Investigation of natural enrichment processes of nitrate in soil and groundwater of semi-arid regions: case study—Botswana S.Stadler, M.von Hoyer, W.H.M.Duijnisveld, T.Himmelsbach, M.Schwiede, J.Böttcher & H.Hötzl Hydroclimatological approach to sustainable water resources management in semi arid regions of Africa U.T.Umoh

430 442

450

461

467

479

489

505

Impact of cultivation practices on multiple uses of water in the Alemaya catchment, eastern Ethiopia Y.E.Woyessa & A.T.P.Bennie Geochemical evidence and origin of salinity in the shallow basinal brine from the Makgadikgadi Pans Complexes, northeastern Botswana L.N.Molwalefhe

514

528

Theme F: Vulnerability and risk Decision support for optimal water system planning: a Wadessy case study A.A.Ilemobade & D.Stephenson The importance of constructing a correct conceptual model for an aquifer G.van Tonder, I.Dennis & D.Vermeulen Water resources development and risk assessment in mountain regions of Africa H.Scheuerlein Reliability, resilience and vulnerability for reservoir sizing and operation J.G.Ndiritu Hydrological impact of dam construction in an arid area D.Stephenson & Z.Chengeta The geochemistry of fresh water supplies in Botswana L.Molwalefhe & S.Vriend Groundwater modelling with limited data: a case study of Yobe River Basin, North East Nigeria M.Hassan, R.C.Carter & K.R.Rushton

541

551

559

572

580 589 600

Theme G: Water resources management Apple and grape vinegar application as c-source in water denitrification Ş.Aslan & A.Türkman

613

Water resources management in the National Park, central Australia E.R.Rooke Integrated water resources management and agriculture in southern Africa M.McCartney, H.Sally & A.Senzanje

623

634

Challenges for managing water resources in semi-arid areas: a case study from two rural communities in Zimbabwe F.T.Mugabe & A.Senzanje An Integrated Water Resources Management tool for Southern Africa allowing low flow estimation at ungauged sites M.J.Fry, S.S.Folwell, H.A.Houghton-Carr & Z.B.Uka Organization of water services in Malawi and strengths and weaknesses in implementing Integrated Water Resources Management (IWRM) M.Selemani Towards best water resources management practice in small town water supply system in Tanzania A.Mvungi & M.Makuya Water management in the Mauritian textile wet processing industry N.Kistamah & S.Roseunee Analysis of the microbiological situation of the quality of domestic water sources and identification of the microorganisms in them, located in the semi-arid regions of the Eastern Cape, South Africa M.Zamxaka, G.Pironcheva & N.Y.O.Muyima Dry season Kalahari sap flow measurements for tree transpiration mapping—Serowe study case, Botswana M.W.Lubczynski, A.Fregoso, W.Mapanda, C.Ziwa, M.Keeletsang, D.C.Chavarro & O.Obakeng Heavy metals and radioactivity in the groundwater of Khartoum State, Sudan A.M.Ahmed Impediments to the effective implementation of a groundwater quality Protection strategy in Botswana T.R.Chaoka, E.M.Shemang, B.F.Alemaw & O.Totolo Spatial assessment of groundwater pollution vulnerability of the Kanye wellfield in SE Botswana B.F.Alemaw, E.M.Shemang & T.R.Chaoka The effect of socio-economic activities on watershed management: the case study of Gaborone Dam catchment in Botswana G.S.Thabeng & D.B.Kemiso

Author index

643

650

661

666

678 685

693

702

707

718

725

737

Preface Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

Africa’s water resources are threatened by population growth with the resultant increase in water demand, the stresses of water use for various activities, desertification, global warming and climate change, and other interventions in the water cycle by man. These effects are more pronounced in the Arid and Semi-Arid regions of Africa in particular and the world in general. It is therefore vitally important that the water resources in Arid and Semi-Arid regions are developed and managed in a sustainable and integrated manner. Integrated management of water resources in the arid and semi-arid regions of Africa requires a spectrum of efforts from local and community stakeholders to national and transboundary river basin management. This conference aims at sharing the best practices of water use and conservation around the globe. The main objective of this conference was therefore to bring together educators, researchers, practitioners, managers, policy makers and NGO’s from Africa in particular and the world in general involved in various aspects of water resources in arid and semi arid regions. The more specific objectives of the conference were to (i) Assess the current state of the art of water resources management in arid and semi arid regions with particular emphasis on African regions. (ii) Address the future water stress due to limited water resources, population growth, increasing demand and pollution and other related risks, resulting in insufficient water supply. (iii) Promote dialog and interaction between different disciplines and professions. (iv) Forster insights into issues of global sustainable development and set concrete targets to meet the need for drinking water and water borne sanitation in arid and semi arid countries of Africa and the world in general. We received an overwhelming response to our call for papers. We received over 120 abstracts and each abstract was reviewed and more than two thirds of the abstracts were accepted. Authors were then requested to submit full text of papers. The full texts of the papers were reviewed by the conference organizing committee and 68 papers were finally accepted for conference. The papers in this book “Water Resources of Arid and Semi-Arid Regions of Africa” constitute the conference proceedings. This book is subdivided into seven sections. Section 1 deals with problems in obtaining data. Section 2 deals with groundwater recharge: natural and artificial; Section 3 deals with Socio economic aspects of water demand management; Section 4 deals with geophysical, GIS and remote sensing

techniques for groundwater exploration; Section 5 deals with climate change and its impact on water resources; Section 6 deals with vulnerability and risk assessment and Section 7 water management. This book will be of interest to researchers and practitioners in the field of surface water hydrology, groundwater hydrology, environmental engineering, agricultural engineering and earth sciences, as well as those engaged in water resources planning, development and management in arid and semi arid areas. Graduate students and those wishing to conduct research in hydrology, environmental science and engineering and water resources will find the book to be of value. Dr A.R.Tombale Permanent Secretary Ministry of Minerals, Energy and Water Resources, Botswana

Keynote address

Sustainable water management in arid and semi-arid environments W.Kinzelbach, P.Bauer, P.Brunner & T.Siegfried Institute for Hydromechanics and Water Resources Management, Swiss Federal Institute of Technology, Zurich, Switzerland Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Scarcity of water often leads to its non-sustainable use. The globally most widely spread non-sustainable practices are overpumping of aquifers, drying up of wetlands and soil salination on irrigated land. Only with much more careful management of scarce resources sustainability in the long term can be reached. Modeling is a valuable tool in the analysis of management options and scenarios. New types of data from remote sensing, airborne geophysics, and environmental tracers to name a few allow reaching a new quality of prediction. Three field studies illustrate the points.

1 INTRODUCTION Fresh water is a scarce resource on a worldwide basis. This becomes apparent when looking at the basic items of the global freshwater balance (Postel et al., 1996). Of the 110,000km3/a of precipitation on the landmass of the earth, 50,000 are returned to the atmosphere via evapotranspiration by the planet’s natural plant cover. Another 21,000 are used by man-made ecosystems (18,000km3/a by rain fed agriculture and 3,000km3/a by irrigated agriculture). This shows that agriculture and natural vegetation are already fierce competitors for the available freshwater. Of the accessible runoff of 13,000km3/a about 4,000 are appropriated by mankind. 70% of those go into irrigated agriculture. This means that a global water crisis would above all be a global crisis in food production. Compared to the 13,000km3/a available the abstracted 4,000 appear small. One should, however, not forget that these figures are averaged in time and space and therefore hide the real problem, e.g. droughts and floods. One can still use the ratio a of withdrawals and available renewable resources as an indicator. Due to the variability of the quantities involved, it is the experience that a value of a>0.4 already reflects severe scarcity. On a global scale a=0.31 is found (Alcamo et al., 2003). This indicates that scarcity on a global level is a reality today, with the arid world already experiencing very severe scarcity problems.

Water resources of arid areas

4

2 SUSTAINABILITY Sustainable water management is a practice, which avoids irreversible or quasiirreversible damage to the resource water and other natural resources linked to it, such as soil and ecosystems. Such a practice conserves the ability of the resource water to provide its services including ecological services. Water scarcity and poverty are often the causes of non-sustainable behavior as they lead to overexploitation and depletion of stocks. What are the relevant issues for global sustainability in the water sector? To identify big and possibly existential problems for whole regions we have to look for ubiquitous negative global trends. In that sense there are a number of non-sustainable practices, which are of global importance. Above all, these are – the overpumping of aquifers, – the destruction of wetlands, – the salination of soils, and – the pollution of aquifers with persistent pollutants. Globally about 800km3/a of freshwater are abstracted from aquifers. About one quarter of this abstraction is non-sustainable in the sense that it is not replaced by recharge, i.e. it is taken out of the available stock. On the Arabian Peninsula, in North Africa, China and the arid Western United States for example, abstractions for large-scale irrigation have withdrawn large quantities of fossil water, which under present climatic conditions are not replenished any more. The global area of wetlands has diminished by 50% since the year 1900. This has a dramatic impact on species diversity. It is a consequence of the competition between natural and man-made ecosystems for land and water resources. The tendency is unbroken. Of the 260 million ha of irrigated farmland in the world about 80 million are affected by soil salination. Salination is a common phenomenon in hot climates. It occurs if in a soil more salt is deposited by evapotranspiration than is removed by drainage. In irrigated agriculture, the most common mechanism leading to salination is the groundwater table rise due to seepage of irrigation water. Once a groundwater table is closer than 2m to ground level, capillary rise leads to direct evaporation from groundwater and to fast salination of the topsoil. Finally, there is the deterioration of groundwater quality by persistent pollutants. One might expect that among those chlorinated hydrocarbons are the most important. This is only true for industrialized countries while globally the most prominent pollutant is salt, especially in arid regions and coastal areas, where seawater intrusion occurs. In principle, all these violations of sustainability are reversible. But the required timescales are on the order of several generations. For all practical purposes these damages are irreversible. In the following, three examples from projects of the Institute of Hydromechanics and Water Resources Management are shown, illustrating the first three globally important sustainability problems. The common features of these examples are that – there is water scarcity (all three areas are in arid or semi-arid climate zones), – a model is developed to analyze and understand the system,

Sustainable water management in arid and semi-arid environments

5

– the model is used for the testing of management strategies and/or for optimization, and finally – the connection is made to the field of socio-economics.

3 OVERPUMPING OF THE NORTH-WEST SAHARA AQUIFER SYSTEM The North Western Sahara is underlain by one of the world’s largest aquifer systems, which covers approximately 1,000,000km2 and consists of two major aquifers, the deeper Continental Intercalaire (CI) and the shallower Complexe Terminal (CT) (Fig. 1). Their water resources are being utilized by the three countries Algeria, Tunisia and Libya mainly for irrigation purposes. The system nowadays hardly receives any recharge. At most 30m3/s are estimated as recharge

Figure 1. Overview of the NorthWestern Sahara aquifer system and its water balance. along the southern flank of the Saharan Atlas where the aquifers strike out (ERESS, 1972). Compared to the size of the system, this recharge flux is—if at all—only of importance locally. The system discharges mainly via the sink of the Chotts or salt lakes, which are the topographic lows of the endorrheic basin. Here approximately 10m3/s evaporate. A very small portion of no more than 5m3/s is thought to discharge to the sea in Libya. Until 1950 abstractions were small. Since then the population has tripled and with it the amount of water pumped for irrigation. An estimated rate of 180m3/s is abstracted today. As a consequence the large springs in the vicinity of the Chotts have run dry (Fig. 2). Artesianism has vanished over large areas and the water, which before flowed at no energy cost, now has to be pumped.

Water resources of arid areas

6

The present situation is characterized by an abstraction, which is 6 times as large as the recharge rate. This brings up the question whether a non-renewable resource should be used at all. Looking at the size of the system and its storage coefficient, an enormous amount of about 100000km3 of water is stored. About one tenth of that amount can be accessed with an economically feasible drawdown of less than 250m. With a projected future withdrawal rate of 500m3/s the total resource can still last for about 600 years. But this water comes at a price. First, energy is necessary for its pumping and distribution and investments in pipes and boreholes have to be made. Second, pumping can lead to deterioration of water quality. Sources of pollution are various. Near the Chotts for example, large drawdowns will reverse the hydraulic gradient, which under natural conditions is always directed from the oases to the Chotts. A reversed gradient mobilizes brine, which finally leads to degradation of the water quality pumped and contributes to the die-off of oases. A similar phenomenon is observed along the coast, where overpumping leads

Figure 2. Development of discharges from springs in Southern Tunisia (1887–1985) (Source: Mamou, 1990)

Sustainable water management in arid and semi-arid environments

Figure 3. Modeled head distribution in CI, 1950 (heads from 530 mamsl to 70 in steps of 35m).

Figure 4. Predicted head distribution in CI, 2050 for planned pumping (heads from 530 mamsl to –250 mamsl in steps of 55m).

7

Water resources of arid areas

8

to seawater intrusion. Salt water can also be mobilized from lower saline aquifers such as the Turonian. With a reduced pressure overburden this confined aquifer can infiltrate at a larger rate into the CT from below. A numerical model of the system has been built which demonstrates that with the required total pumping rate by 2050 large areas of the presently strongly pumped regions will face a piezometric decline with economically infeasible drawdowns of more than 250m below ground level (area with sawtooth pattern in Fig. 4). At the same time the constraints for water quality locally can no longer be fulfilled. The groundwater model was then coupled with optimization algorithms to find allocation patterns that conform to demand, drawdown and quality constraints in time while minimizing overall provision costs. The wells in an optimal scheme spread out over the area to equilibrate distribution cost with pumping cost, which depend on drawdown. They further spread to the CI from the CT. Two variants were analyzed. In the first, the existing pumping locations were used and the pumping rates at those constituted the set of decision variables. At the Chott cells, gradient constraints were introduced to prevent gradient reversal and thus preserve the productivity of the oases. On the whole the costs are exploding over time, with the running cost of water increasing by a factor of about 30 in 50 years (Siegfried, 2003). In a second variant, pumping at any location was allowed with the costs being the only criterion for choosing a specific cell. The results show that compared to the first variant much better abstraction schemes are possible with considerably lower running cost (and total costs) over the next 50 years. However interesting such a scenario is, it would require a complete renewal of infrastructure. Realistically, only a gradual transition from today’s pumping well distribution to a more favorable one in the future will be feasible. The model demonstrates that it is possible to minimize pumping cost to reasonable levels and provide water for the next 50 years. This time however must be used to develop alternatives. All optimization runs were carried out ignoring national borders in order to assess benefits from cooperative management. As the results demonstrate, cooperation between the three countries involved brings considerable advantage in the exploitation of the resource. Nevertheless, in the long run the conservation of the oasis culture requires heavy subsidies as the substitution between the production factors water and capital progresses. 4 MANAGEMENT OF THE OKAVANGO DELTA, BOTSWANA The Okavango River flows from the Benguela plateau of Southern Angola in southeastern direction through the northern tip of Namibia and then into Botswana, where it forms an inland

Sustainable water management in arid and semi-arid environments

9

Figure 5. Satellite image of the Okavango Delta (length from right to left about 550km). Delta in which it is consumed completely by evapotranspiration (Fig. 5). The Delta is one of the largest wildlife areas in Africa and is an attraction for numerous tourists. The yearly floods of the river turn a large area of the Delta into a seasonal swamp (Hutchins et al., 1976; Scholz et al., 1976; McCarthy et al., 1986; Thomas and Shaw, 1991; Ellery et al., 1993; McCarthy et al., 1993; McCarthy and Ellery, 1994; Modisi et al., 2000; Gumbricht et al., 2001). As the flood takes 3 months to propagate from the inflow at Mohembo to the distal side at Maun, it is out of phase with the local rainy season and thus increases the water availability over the year. The upstream countries are discussing plans to abstract water from the river and/or build dams for electricity production or agricultural purposes. In Botswana itself, various sectors of the economy have also proposed to make use of the Okavango water, be it for agriculture or for mining purposes. All measures proposed threaten the existence of the Delta as the unique ecosystems it is. Both abstraction of water in the upstream and acceleration of the through-flow by dredging of channels etc. will cause a decrease in the size of the seasonal swamp. In order to assess the impact of hydraulic measures on the size and distribution of the flooded area a numerical model was constructed which contains the surface water and the groundwater in two coupled layers. In an innovative approach satellite data on the time-varying size of the Delta were used to calibrate the model (McCarthy et al., 2003; Bauer et al., 2004). Further data used in this approach are a high-resolution digital terrain model obtained from the flooding patterns and the related vegetation patterns (Gumbricht et al., 2001, 2003), the inflow at Mohembo, the precipitation from METEOSAT data (Herman et al., 1997), the

Water resources of arid areas

10

evapotranspiration from multi-spectral satellite data (Bastiaanssen et al., 1998a, 1998b), and last not least local measurements which are routinely performed by the Botswana Department of Water Affairs. The model is able to reproduce satisfactorily the seasonal dynamics of the flooded areas both in total extent and in distribution over a period of 20 years for which data are available (Fig. 6). The sink of all water is essentially the evapotranspiration by the plant cover. This process also governs the distribution of salts in the Delta. Pronounced salt crusts indicate areas, which are natural disposal sites of salts. Their continued functioning is of considerable importance to the conservation of the Delta. This process will be incorporated in a future version of the model. One example of measures with potentially serious impacts on the Delta is the abstraction of water upstream of the inflow (Fig. 7). It is seen that an abstraction is amplified i.e. the relative reduction in area is considerably larger than the relative reduction in inflow. Dams have an effect on both inflow reduction and temporal inflow distribution. Model calculations showed that the change in input distribution not necessarily is detrimental to the size of the flooded area. A more stretched out flood will bring water further downstream. Morphological changes such as dredging of channels and removal of blockages by papyrus have also a pronounced effect, not so much on the total flooded area as on the distribution of flooded areas within the Delta (Bauer et al., 2004).

Figure 6. Observed and modeled flooding frequency (%).

Sustainable water management in arid and semi-arid environments

11

Figure 7. Flooded area for different abstraction scenarios in comparison with the modeled development of the last 20 years. The local abstraction for household consumption, be it directly or from the aquifer, is so small that it will at no stage be of relevance for the Delta. The tentative ranking of different interventions according to their severity is as follows: – abstractions larger than 2m3/s in the upstream, – building of large dams in the upstream, – change to a drier climate, – morphological changes (dredging, cutting of vegetation, tectonics), – local drinking water supply. The model can provide a quantitative basis for the political debate between the three riparian nations. It is clear that the conservation of the Delta must bring some revenue to the upstream in exchange for the guaranteed inflow. The key parameter for an administrable negotiated solution will be the minimum inflow at Mohembo and its seasonal variability. 5 SALINATION OF SOILS AND WATER IN YANQI BASIN, XINJIANG, CHINA The third example studies a region in China’s arid west. The Yanqi basin is formed by the lowlands of the Kaidu River and Lake Bostan (Fig. 8). The area has been used intensively for agriculture over the past 50 years. The main products are grapes, cotton and red peppers.

Water resources of arid areas

12

Figure 8. Satellite map of the Yanqi basin showing the irrigation areas along the Kaidu River, Lake Bostan and the Kongque River. As precipitation is only 70mm/a and thus negligible compared to the potential evaporation of 1800mm/a, no agriculture is possible without irrigation. The last 50 years have seen a tremendous growth of the population. This has led to a strong increase in agricultural production. The intensive irrigation with river water caused a water table rise, followed by serious soil salination. To maintain production, over-irrigation is required to push salt from the surface beneath the root zone. This practice increases the amount of water used per unit crop and contributes again to water table rise. A vicious circle is triggered, leading to higher and higher salinity in the water flowing off the irrigation area both in the subsurface and in the drains. One could argue that the applied irrigation techniques and efficiencies in the Yanqi Basin are sustainable, as a steady state has been reached (the amount of salt transported out of the Yanqi Basin is equal to the amount of salt imported by the Kaidu) and production stabilized on a level still profitable. This of course cannot be called sustainable because only the needs of the farmers in this particular irrigation system are satisfied. With the rising groundwater table and the increased non-productive evaporation of water the salinity in the lake has increased and the lake level has fallen. The amount of water available for the downstream of lake Bostan, carried by the Kongque River, has decreased thus limiting natural vegetation growth and agriculture in the so called Green Corridor. The Green Corridor is a landscape, which extends down to Lop Nor and is characterized by the riverine desert poplar forests. Today, no water reaches Lop Nor due to the high consumptive use in the upstream irrigation systems. In order to improve the situation of the system as a whole, a number of measures in the upstream have been proposed (Dong, 2001). They include

Sustainable water management in arid and semi-arid environments

13

– the reduction of irrigation area, – substitution of irrigation water from the river by groundwater thus guaranteeing that the groundwater table stays below critical levels, – changes in the crop mix and irrigation techniques (e.g. drip irrigation for grapes), – the transfer of water directly from Kaidu river to the Kongque river bypassing the lake, – the lowering of the lake level in order to reduce non-productive evaporation of the lake and others. In an integrated modeling approach all these options will be investigated. Again, some relevant data can be obtained using remote sensing techniques. In this case we constructed a digital terrain model from stereo images of radar satellites based on methods described by Zebker and Goldstein, 1986. The absolute elevations were obtained from single point DGPS measurements (Fig. 9). The ground surface elevation is of particular interest in salination problems as evaporation from groundwater is a function of the distance to groundwater table. Hence salinity observed at the

Figure 9. Digital terrain model of the Yanqi basin.

Water resources of arid areas

14

Figure 10. Correlation between measurements at ground control points (GCP) of soil conductivity and spectral match to salt pixel. ground level is a data type, which allows the regional verification of the groundwater model. The distribution of surface salinity was obtained from multispectral ASTER data and measurements in the field. Based on the spectral response of a completely salinized pixel, the closeness of any pixel to this reference is determined yielding an uncalibrated salinity map. To convert these values into salinity or its physical measure of electrical conductivity, a calibration with ground truth is required. The ground truth was obtained both by single core samples and less time-consuming geophysical measurements. A good correlation between ground truth and the uncalibrated salinity map was found (Fig. 10). Of course, this correlation only holds for the non-irrigated areas. The salinity map (Fig. 11) clearly shows the salt accumulation in the paths between fields while in the irrigated fields themselves no increased salinity is visible due to overirrigation. While a coupled groundwater-surface water model is still under development, preliminary estimates are already available on the basis of a multi-box approach with the irrigation area, the aquifer and the lake being the respective boxes. Despite the fact that the box approach is a major simplification, it demonstrates that steady states for groundwater tables as well as salt concentration exist. Depending on how water in a steady state is exported from the system, reaching a steady state salt concentration can take a very long time compared to reaching a steady state in groundwater tables (Fig. 12). The steady state salt concentration is directly determined by

Sustainable water management in arid and semi-arid environments

15

Figure 11. Salinity map obtained from a multispectral satellite image.

Figure 12. Steady state for groundwater tables (hss, m below surface) and salinity (css, g/l) in the Yanqi basin aquifer as functions of applied irrigation water (in 107m3/a). the ratio of the flux of water draining from the aquifer into the lake and the groundwater recharge. Furthermore, the box approach shows that the rate of accumulation of salt increases rapidly as soon as direct evaporation from the aquifer occurs. Pumping groundwater for irrigation purposes would not only reduce the need for overirrigation, but also directly contribute to the water availability downstream. This solution is more expensive as groundwater comes at about 10 times the price of surface water. However, if the water table can be kept low by pumping groundwater, the conservation of soil and the increased availability of surface water in the downstream might strike the balance with a higher price of water.

Water resources of arid areas

16

6 CONCLUSIONS In arid countries the problems of sustainability in the water sector are prominent. On a worldwide basis the three subjects discussed are the most widespread. They show several common features. Water management in the arid and semi-arid environment must include salt management. Modern tools such as remote sensing, geophysics and modeling hydrological science help even in regions with weak infrastructure to quantify the implications of human interaction and to give advice to decision makers on the sustainability of water management practices. Models summarize the state of affairs and are the only means to make predictions. They are bound to be crude and simulations will always be idealized. Still, they can serve as points of reference. A further common feature is that sustainable solutions require the system boundary to be taken sufficiently large, often transgressing political boundaries. While science can give some decision support, the decisions for or against sustainability are made in the political arena. REFERENCES Alcamo, J., Doll, P., Henrichs, T., Kaspar, F., Lehner, B., Rosch, T. & Siebert, S. 2003. Development and testing of the WaterGAP 2 global model of water use and availability. Hydrological Sciences Journal-Journal Des Sciences Hydrologiques, 48(3):317–337. Bastiaanssen, W.G.M., Menenti, M., Feddes, R.A. & Holtslag, A.A.M. 1998a. A remote sensing surface energy balance algorithm for land (SEBAL). 1. Formulation. Jnl. of Hydrology, 212– 213:198–212. Bastiaanssen, W.G.M., Pelgrum, H., Wang, J., Ma, Y., Moreno, J.F., Roerink, G.J. & van der Wal, T. 1998b. A remote sensing surface energy balance algorithm for land (SEBAL). 2. Validation. Jnl. of Hydrology, 212–213:213–229. Bauer, P., Gumbricht, T. & Kinzelbach, W. 2004. A large-scale coupled surface water/ground water model of the Okavango Delta, Botswana. Water Resources Research, submitted. Dong, X., Jiang, T. & Jiang, H. 2001. Study on the pattern of water resources utlilsation and environmental conservation of Yanqi Basin. In: G.Li (Ed.), Development, Planning and Management of Surface and Groundwater Resources. IAHR congress proceedings. Tsinghua University Press, Beijing, China: 333–340. Ellery, W.N., Ellery, K., Rogers, K.H., McCarthy, T.S. & Walker, B.H. 1993. Vegetation, hydrology and sedimentation processes as determinants of channel form and dynamics in northeastern Okavango Delta, Botswana. African Jnl of Ecology, 31:10–25. ERESS 1972. Etude des Ressources en Eau du Sahara Septentrional. Rapport sur les Résultats du Projet, Conclusions et Recomm endations, UNESCO, Paris. Gumbricht, T., McCarthy, T.S. & Bauer, P. 2003. Microtopography of the Okavango Delta using correlation between land cover and elevation. Earth Surface Processes and Landforms, in press. Gumbricht, T., McCarthy, T.S. & Merry, C.L. 2001. The topography of the Okavango Delta, Botswana, and its tectonic and sedimentological implications. South African Jnl. of Geology, 104:243–264. Herman, A., Kumar, V.B., Arkin, P.A. & Kousky, J.V. 1997. Objectively Determined 10 Day African Rainfall Estimates Created for Famine Early Earning Systems. International Journal of Remote Sensing, 18(10):2147–2159. Hutchins, D.G., Hutton, L.G., Hutton, S.M., Jones, C.R. & Loenhert, E.P. 1976. A summary of the geology, seismicity, geomorphology and hydrogeology of the Okavango Delta, Geological Survey Botswana, Gaborone.

Sustainable water management in arid and semi-arid environments

17

Mamou, A. 1990. Charactéristiques et evaluation des resources en eau du sud Tunisien. Dissertation, Université de Paris-Sud, Centre d’Orsay. McCarthy, J., Gumbricht, T., McCarthy, T.S., Frost, P.E., Wessels, K. & Seidel, F. 2003. Flooding Patterns of the Okavango Wetland in Botswana between 1972 and 2000. Ambio, 32(7):453–457. McCarthy, T.S. and Ellery, W.N. 1994. The effect of vegetation on soil and ground water chemistry and hydrology of islands in the seasonal swamps of the Okavango fan, Botswana. Journal of Hydrology, 154: 169–193. McCarthy, T.S., Ellery, W.N., Rogers, K.H., Cairncross, B. & Ellery, K. 1986. The roles of sedimentation and plant growth in changing flow patterns in the Okavango Delta. South African Journal of Science, 82: 579–584. McCarthy, T.S., Green, R.W. & Franey, N.J. 1993. The influence of neo-tectonics on water dispersal in the north-eastern regions of the Okavango swamps, Botswana. Journal of African Earth Sciences, 17(1): 23–32. Modisi, M.P., Atekwana, E.A., Kampunzu, A.B. & Ngwisanyi, T.H. 2000. Rift kinematics during the incipient stages of continental extension: Evidence from the nascent Okavango rift basin, northwest Botswana. Geology, 28(10):939–942. Postel, S.L., Daily, G.C. & Ehrlich, P.R. 1996. Human appropriation of renewable fresh water. Science, 271(5250):785–788. Scholz, C.H., Koczynski, T.A. & Hutchins, D.G. 1976. Evidence of incipient rifting in Southern Africa. Geophysical Journal of the Royal Astronomical Society, 44:135–144. Siegfried T. 2003. Management of internationally shared groundwater resources in semiarid and arid region s: the Northern African Aquifer System. In E.Servat et al. (eds), Hydrology of Mediterranean and Semiarid Regions, IAHS Publ. No. 278, 2003. Thomas, D.S.G. & Shaw, P.A. 1991. The Kalahari Environment. Cambridge University Press, Cambridge. Zebker, H.A. & Goldstein, R.M. 1986. Topographic Mapping from Interferometric Synthetic Aperture Radar Observations. Journal of Geophysical Research-Solid Earth and Planets, 91(B5):4993–4999.

Theme A: Problems in obtaining hydrological and geohydrological data

Slug tests in fractured rock formations: value, pitfalls and misinterpretations P.D.Vermeulen & G.J.van Tonder Institute for Groundwater Studies, University of the Free State, Bloemfontein Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Currently slug tests in South Africa are used with two objectives in mind: (i) to get a first estimate of the yield of a borehole (relationship obtained by Viviers et al., (1995) and (ii) to estimate the Kvalue (or T-value) of the aquifer in the vicinity of the borehole. The paper shows that the use of currently available slug test interpretation methods to analyse slug tests in fractured rock aquifers to estimate a T or K-value is problematic. The estimated value is dependent on the flow thickness (thickness of the part of the aquifer in which flow occurs due to the slug input). If this thickness of flow is known, the estimated K-value is more representative of that of the fracture zone. By using the total thickness of the formation for the estimation of the K-value in slug test analysis, the estimated K-value (and thus KD-value) does not represent the T-value of the formation.

1 INTRODUCTION In performing a slug test, the static water level in a borehole is suddenly lowered or raised. This is usually done by lowering a closed cylinder into a borehole. The cylinder replaces its own volume of water within the borehole, thus increasing the pressure in the borehole. As the equilibrium in the water level is changed, it will recover or stabilise to its initial level. If the rate of recovery or recession of the water level is measured, the transmissivity or hydraulic conductivity of the borehole can be determined (Kruseman and De Ridder, 1994). In South Africa slug tests are conducted for the following two reasons: ● To estimate the hydraulic conductivity (K) of the aquifer in the vicinity of the borehole and ● To get a first estimate of the yield of a borehole (Vivier et al., 1995).

Water resources of arid areas

22

Vivier et al. (1995) performed slug tests on 32 boreholes, of which the maximum yield was known and they then derived empirically the following formula (there is a 93% correlation between the actual yield and the yield estimated with the formula): Q=117155.08t−0.83 (1) where: Q=yield of the borehole in l/h and t=recession time of the slug test in seconds (90% recovery). Usually the Cooper method (Cooper et al., 1967) or the Bouwer and Rice method (1976) is used to estimate the K-value (or T-value in the case of the Cooper method). In the following section slug test results, as well as pumping and tracer test results for borehole UO5 on the well-known Campus Site of the University of the Free State, South Africa (Figure 1) will be discussed to illustrate the problems associated with the interpretation of slug tests in a borehole drilled in a fractured aquifer.

Figure 1. Map of the RSA.

Slug tests in fractured rock formations

23

Figure 2. Diagram of the geological formation at the Campus Test Site (relative thickness of the aquifers in brackets). 2 BOREHOLE UO5 ON THE CAMPUS SITE The Campus Test Site is underlain by a series of mudstones and sandstones from the Adelaide Subgroup of the Beaufort Group of formations in the Karoo Supergroup (Figure 2). There are three aquifers present on the Site. The first, a phreatic aquifer, occurs within the upper mudstone layers on the Site. This aquifer is separated from the second and main aquifer, which occurs in a sandstone layer of between 8 and 10m thick, by a layer of carbonaceous shale with a thickness of 0.5 to 4m. The third aquifer occurs in the mudstone layers (more than 100m thick) that underlie the sandstone unit.

Water resources of arid areas

24

Figure 3. Acoustic scan of borehole UO5 at a depth of 20m to 25m below the surface.

Slug tests in fractured rock formations

25

Figure 4. Borehole video image of the fracture zone in borehole UO5 showing a fracture-zone thickness of about 200mm.

Figure 5. Constant rate pumping test data of UO5. Table 1. Hydraulic parameters estimated for UO5.

Water resources of arid areas

Parameter

26

Value

T of formation* (m2/d) 19 K of fracture zone (m/d) 3600 T of fracture zone (m2/d) 576 K of matrix (m/d) 0.17 T of matrix** (m2/d) 3 *Average for fracture+matrix, obtained from Cooper-Jacob fit to late drawdown values. **For 20m thickness.

A major characteristic of the main aquifer is the presence of a horizontal fracture that coincides approximately with the centre plane of the sandstone layer, and which intersects all 11 boreholes with significant yields on the Site, of which UO5 is one. The remaining 14 boreholes all have very insignificant yields. The fracture zone thickness is approximately 10mm, but the adjacent 200mm of sandstone is also highly permeable. Figure 5 shows a graph of the data from a constant rate test conducted on UO5 at a rate of 1.25L/s. Measurements were also taken in the observation borehole UO6. These pumping test data were analysed with a numerical 3D model (Van Tonder et al., 2001), and the following parameters were estimated in Table 1. The thickness of fracture zone (referred to in Table 1) was obtained from tracer tests and the borehole video, and is 0.16m. The hydraulic parameters given in Table 1 are regarded to be accurate (Van Tonder et al., 2001). It would now be interesting to analyse the data of a slug test (Figure 6) conducted on borehole UO5 and compare the estimated values with the values given in Table 1. The 90% recovery occurred after about 9 seconds, and using Equation (1) the yield of borehole UO5 is estimated as 5.3L/s. The tested blow yield of borehole UO5 was 6L/s during drilling. The Bouwer and Rice method (1976) was applied to the data in Figure 6. The Bouwer and Rice equation reads: (2) where: rc=radius of the unscreened part of the borehole where the head is rising rw=horizontal distance from the borehole centre to the undisturbed aquifer Re=Radial distance over which the difference in head h0 is dissipated in the flow system of the aquifer

Slug tests in fractured rock formations

27

Figure 6. Data collected during a slug test conducted on UO5. Table 2. Estimated K-values with the Bouwer and Rice method (1976) for different values of the flow thickness. Thickness open to flow (m)

K (m/d)

30 20 10 1 0.16 0.001

12 17 32 231 541 3600

T (m2/d) 360 340 320 231 86 3.6

d=length of the borehole screen or open section of the borehole h0=head in the borehole at time=0 ht=head in the borehole at time t The estimated K-value of Bouwer and Rice is dependent on the thickness open to flow, d, and Table 2 shows the different K-value estimates for different flow thicknesses. Note that a flow thickness of 30m will indicate the depth from the water level to the end of the borehole and that a thickness of 0.16m is the thickness of the fracture zone in borehole UO5. 3 DISCUSSION Comparison of Table 1 and Table 2 shows the following important issues: ● An incorrect K-value is obtained from the slug test if the thickness of the aquifer (total formation) is used as the flow thickness. For a thickness of 30m, a K-value of 12m/d

Water resources of arid areas

28

(or T=360m2/d) is estimated from the slug test, which is neither the T-value of the fracture zone nor the T-value of the matrix. ● For a flow thickness of 0.16m (i.e. the thickness of the fracture zone), a K-value of 541m/d is estimated with the Bouwer and Rice (1976) slug test method. This estimated K-value is more representative of the K-value of the fracture zone. ● The average T-value of the formation, which is important for management purposes, was estimated as 19m2/d from the constant rate pump test. It is impossible to estimate the T- or K-value of the aquifer formation via a slug test.

4 CONCLUSIONS The use of the current available slug test interpretation methods to analyse a slug test in a fractured rock aquifer to estimate a T- or K-value is problematic. The estimated value is dependent on the flow thickness (thickness of the part of the aquifer in which flow occurs due to the slug input). If this thickness of flow is known, the estimated K-value is more representative of that of the fracture zone. By using the total thickness of the formation for the estimation of the K-value in slug test analysis, the estimated K-value (and thus KD-value) does not represent the T-value of the formation. REFERENCES Bouwer, H. & Rice, R.C. 1976. A slug test for determining hydraulic conductivity of unconfined aquifers with completely pr partially penetrating wells. Water Resources Research, 12:423–428. Cooper, H.H, Bredehoeft, J.D., & D Papadopulos, I.S. 1967. Response of a finite-diameter well to an instantaneous charge of water. Water Resources Research, 3:263–269. Kruseman, G.P. & de Ridder, N.A. 1994. Analysis and Evaluation of Pumping Test Data. 2nd ed. International Institute for Land Reclamation and Improvement. Publication 47. Wageningen, the Netherlands: 237–247. Vivier, J.J.P., Van Tonder, G.J. & Botha, J.F. 1995. The use of slug tests to predict borehole yields: correlation between the recession time of slug tests and borehole yields. In Conference Proceedings: Groundwater’95: Groundwater Recharge and Rural Water Supply, Midrand, South Africa. Van Tonder, G.J., Botha, J.F., Chiang, W.H., Kunstmann, H. & Xu, Y. 2001. Estimation of the sustainable yields of boreholes in fractured rock formations, Special issue of Journal of Hydrology: No 241.

Flow simulation model performance assessment using entropy approach A.M.Ilunga Civil Engineering, University of the Witwatersand, South Africa D.Stephenson Civil Engineering, University of Botswana, Gaberone, Botswana Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Hydrological data (e.g. rainfall, river flow, etc) are used in water resources planning and management for planning reservoir and operation. However, it happens sometime that the appropriate site where a hydraulic structure (e.g. reservoir) should be built has no available data due for example to inaccessibility to erect a flow gauging station, etc. This is particularly a problem in arid areas. Very often hydrologists make use of simulation models to estimate the flows data series at the very site from the nearby stations and using some physical characteristics of the catchment area. In this paper, a merely methodology is proposed to evaluate the performance of simulation models in terms of entropy (e.g. reduction of the uncertainty of flows) before and after applying a model to the site. This reduction should be above a certain threshold value for the model to be retained as performing well. An example is illustrated through RAFLER model, which is used to simulate yearly flows at Braamhoek.

1 INTRODUCTION For planning, management and effective control of water resources systems, a considerable amount of data on hydrologic variables such as rainfall, streamflow, etc. are required. It sometime happens that the appropriate site where a reservoir should be built has no available data due for example to inaccessibility to erect a flow gauging station. This is particularly a problem in arid areas. Physical models, semi-distributed models, statistical models, conceptual model, embracing probabilistic, fitting curve, black box, etc are often used to simulate/estimate flows. In this paper, a merely methodology is proposed to evaluate the performance of simulation models in terms of entropy (e.g. reduction of the uncertainty of flows) before and after applying a model to the site in a similar way of Panu (1992). These authors used entropy approach for infilling hydrological data problems (e.g. reduction of uncertainty before and after infilling the

Water resources of arid areas

30

data series), while in this paper the same approach is applied to cases where no data series is available at all at the target site. In very recent paper (see, Ilunga and Stephenson, 2003b) the methodology for evaluating the model performance was roughly used but it was not explained systematically as done in this paper. It should be noted that the reduction evoked above should be above a certain threshold value for the model at hand to be retained as performing well. An example is illustrated through RAFLER model, which is used to simulate yearly flows at Braamhoek. 2 INFORMATION CONTENT OF HYDROLOGICAL VARIABLES Traditionally, the information content of a hydrological variable can be measured through variance, which shows the variability of the hydrological variable with respect to its mean. However this approach was criticized for cases where information available about the hydrological variable is little (Singh, 1998). Since 70’s hydrologists tried to find another way of measuring information by theoretic entropy (a term borrowed from communication, see Shannon & Weaver, 1949). Thus the concept has been applied in water resources (Singh & Florentino, 1992; Amorocho & Espildora, 1973) and water related fields. The entropy is considered as a measure of the amount of chaos or lack of information about a system. The entropy can be viewed as a measure of ignorance about the system described in classical sense by a probability distribution. Indirectly, it measures the information about the system. Mathematically entropy of a system {xi} is defined in its discrete form by the following expression (1) where K: is a function of the base used or the scale factor (bits for base 2, napiers for base e, decibels for base 10), i=1, 2,…, n and pi is the probability of occurrence of the event i. It can be shown that the value of H(X) reaches its maximum when all variate values xi are equally likely, that is, when the outcome has maximum uncertainty (Amorocho & Espildora, 1973). In this case the entropy becomes Hmax (X)=log n (2) The theoretic entropy definition was extended to hydrology. Hence entropy is considered as a measure of the degree of uncertainty of random variable hydrologic processes (Amorocho & Espildora, 1973). Since the reduction of the uncertainty by means of making observations is equal to the amount of information gained, the entropy criterion indirectly measures the information content of a given series of data (Harmancioglu et al., 1994). It arises that the distribution of the variable can be unknown a prior although some of its properties may be known, e.g. mean, variance, normality condition. These proprieties (information) enable to determine the distribution of the variable, which maximizes the entropy function. In this way the distribution is consistent with the available information,

Flow simulation model performance assessment using entropy approach

31

but retains maximum uncertainty within the feasible domain and thus ensures the least bias; that is the principle of maximum entropy (POME) introduced by Jaynes in 1968. This principle has been applied intensively in hydrology in the last two decades. 3 ENTROPY APPROACH AS HYDROLOGICAL MODEL PERFORMANCE EVALUATION Amorocho and Espildora (1973) suggested that the mutual information (between the observed values and the simulated ones) could be used as a criterion in the selection of hydrological models; e.g. rainfall-runoff prediction. Note that the mutual information concept is derived from entropy notion and for more details; the reader is referred for example to the above-mentioned paper. Later the directional information transfer index (DIT) appeared as a generalization of the mutual information Yang and Burn (1994) and was used for dependency evaluation between streamflow gauging station pairs. Recently, it is argued that since mutual information is used for model performance assessment, its generalization i.e. DIT can be extended to model performance evaluation (Ilunga & Stephenson, 2003a). The above considerations are valid when the estimated values have to be compared to the observed ones. In that respect statistical criteria such as root mean square error, etc can be also used to crosscheck the results (Ilunga and Stephenson, 2003a). However it becomes difficult to use these considerations when missing values are encountered in the data series. Thus Panu (1992) introduced the notion of reduction of uncertainty of the hydrological variable before and after infilling the data series. The reduction of uncertainty Re d(%) at a given site as defined by Panu (1992) can be given as follows: (3) where Hcc and Hcomp are entropy values before and after infilling the data series respectively. It should be noted that this concept was applied to cases of consecutive missing data values, e.g. hydrological data exist before and after the missing values. 4 PROPOSED METHODOLOGY Panu (1992) used expression (3) for infilling data problems, in other words some data exist before infilling process. In this paper the same expression (3) is proposed for cases where no available flow data exist at all at the site. It is more natural to say that a case where no data is available, the uncertainty is higher than a case where data exist. Thus it is assumed that the uncertainty should be maximum (e.g. if all hydrological events would occur equally likely) at a site where no data is known. Thus, in this case expression (3) can be re-written as (Ilunga & Stephenson, 2003b): (4)

Water resources of arid areas

32

Where Hmax is defined by expression (2) The following are the different steps for evaluating the performance of a flow simulation model for cases where no data exist at all at given site. 1. Having the physical parameters of the catchment area and information (e.g. rainfall) from the nearby sites, compute the simulated flows. 2. Compute the frequency (probability) distribution of the flows. 3. Compute the marginal entropy of the site using formula (1) and set the entropy before simulating flows to its possible maximum value, e.g. see formula (2). 4. Compute the reduction of entropy at the site using expression (4) and set a threshold reduction of entropy to an arbitrary value. If the computed value for the reduction is greater or equal than the threshold value, the model is considered as performing well. Otherwise, the model performs poorly. Terminate.

5 SHORT NOTE ON RAFLER RAFLER is an acronym for Rainfall Flow Erosion. A model (RAFLER) is a deterministic model based on the physics of runoff, soil infiltration and soil transport and which converts rainfall data to runoff over a length of time, e.g. years. The model uses monthly rainfall figures to reproduce monthly stream flow series and soil erosion. Some simplification is made to enable the model to be run with a minimum of data. And the rainfall period each month is estimated from the number of rain days to enable true flow rates to be calculated. This model requires a number of modules including catchment, channels and reservoirs. The general theoretical background of the model can be traced in Stephenson (2002). 6 STUDY AREA AND DATA AVAILABILITY Braamhoek is situated in the Free State, in South Africa. The catchment area is about 62km2. Neither rainfall data nor stream flow data is available at this particular site. Thus it was possible to simulate flows at Braamhoek using rainfall data from the nearby sites; viz at Van Reenen (MAP=1002mm/month); at Moorside (MAP=839mm/month) and at Baldergow

Flow simulation model performance assessment using entropy approach

33

Figure 1. Simulated yearly frequency (probability) distribution at Braamhoek. Table 1. Model performance evaluation at Braamhoek.

Before applying RAFLER Applying RAFLER

Marginal entropy (Napiers) at Braamhoek

Reduction of uncertainty (%) at Braamhoek

4.47

−

1.65

63.06

(MAP=887mm/month). The monthly rainfall data (1916–2002) were obtained from the Weather Bureau, South Africa. 6.1 Application of the methodology to Braamhoek The application of the model, i.e. RAFLER to simulate the total annual flows (from 1916–2002, e.g. 87 data points) at Braamhoek site gave the following results. Figure 1 from which the entropy calculations were possible shows the probability (frequency) distribution estimated from the model. The threshold value of the reduction of uncertainty was set to a value of 50% napiers. Table 1 shows the results of entropy calculations before and after applying RAFLER models. It is therefore concluded that the reduction in uncertainty of the yearly flows at

Water resources of arid areas

34

Braamhoek was 66.06% by this model. This value is the equivalent of information inferred about the site using RAFLER model. This model could be thought to perform well. Thus RAFLER model could be used for flow prediction at Braamhoek with regard to the total yearly flows. Nonetheless the model needs to be tested on other flow regimes for that specific site. 7 CONCLUSION The focus of this paper was to give a methodology for evaluating the performance of simulation models using entropy approach. The methodology has been tested with RAFLER model on Braamhoek site where records were simulated. Recall that this methodology was roughly used in Ilunga and Stephenson (2003a), but without presenting systematically the steps involved as been done in this paper. The computations from the entropy criterion showed that RAFLER model could be used for simulating the total yearly flows at Braamhoek when a threshold value of 50% is considered for the reduction of uncertainty before and after simulation. Investigation should also be done on other flow regimes. REFERENCES Amorocho, J. & Espildora, B. 1973. Entropy in the assessment of uncertainty in hydrologic systems and models. Water Resources Research, 9(6):1511–1522. Harmancioglu, N.B., Alpaslan & Singh, V.P. 1994. Assessment of the entropy principle as applied to, water quality monitoring network design. Stochastic and Statistical Methods in Hydrology and Environmental Engineering., 3:135–148. Ilunga, M. & Stephenson, D. 2003a. Performance of hydrological data infilling techniques using entropy approach: Expectation maximization algorithms. 11th South African National Hydrology Symposium, Port Elizabeth, South Africa: 6. Ilunga, M. & Stephenson, D. 2003b. Entropic measures for comparing flow simulation models at Bedford site. Paper submitted to the J. Hydrology, Elsevier. Panu, U.S. 1992. Application of some entropic measures in hydrologic data infilling procedures. In: Singh, V.P. & Fiorentino, M. (Eds) Entropy and energy dissipation in water resources, Kluwer Academic Publishers, The Netherlands: 175–192. Shannon, C.E. & Weaver, W. 1949. The Mathematical Theory of Communication. University of Illinois Press Urbana, Chicago, London. Singh, V.P. 1998. Entropy as a decision tool in environmental and water resources. J. Hydrology , Indian Association of Hydrologists, 21(1–4):1–12. Singh, V.P. & Florentino, M. 1992. A historical perspective of entropy applications in water resources. In: Singh, V.P. & Fiorentino, M. (Eds) Entropy and energy dissipation in water resources, Kluwer Academic Publishers, The Netherlands: 21–61. Singh, V.P. & Krstanovic, P.F. 1987. A stochastic model for sediment yield using the principle of maximum entropy. Water Resources Research, 23(5):781–793. Stephenson, D. 2002. “Modular kinematic model for runoff simulation”. In: V.P.Singh, & D.K.Frevert (Eds). Mathematical models of small watershed hydrology and applications. Water Resources Publications, LLC, pp. 183–218, Chapter 7. Yang, Y. & Burn, H. 1994. “Entropy approach to data collection network design”. J. Hydrology, 94:307–324.

Flow simulation model performance assessment using entropy approach

35

Yevjevich, V. (1972). Probability in hydrology. Water Resources Publications, Colorado, U.S.A.: 331.

Data collection experiences in water level monitoring, borehole archive and research projects in semi arid Botswana Magowe Magowe, Thothi Obakeng & Paul Makobo Department of Geological Survey, Hydrogeology Division, Lobatse, Botswana Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: The Department of Geological Survey (DGS) of Botswana has been carrying out water level monitoring of non-operational wellfields since 1983 and various research projects throughout the country. This involved intense collection of data on water levels and rainfall manually and later on automatic data collection instruments were introduced. This data was stored as hardcopies without quality assurance. Problems experienced with this data collection encompass logistical, equipment and human input. The logistical problems include poor accessibility due to the country’s hostile environment such as dust and extensive ponding during heavy showers. This hostile environment leads to reduction of the lifespan and the poor performance of these instruments. Poor handling of data, equipment failure, lack of the right set of equipment and local operational knowledge also poses problems. Therefore, as a consequence, valuable data is normally lost. In 2003, a quality assurance process was resumed for the 1999–2003 water level monitoring and rainfall records. Common problems that were encountered are data gaps which could be explained by the above causes. During the quality assurance of 1999–2002 data for the ten (10) monitoring network areas, data gaps or unavailable data constituted 80% of all recorded problems (DGS, 2003). This paper discusses the experience of the DGS in collecting hydrogeological data in Botswana semi arid environment.

1 INTRODUCTION The DGS Hydrogeology division has been collecting data in the areas of monitoring, borehole archive and research projects. Data collection in monitoring started in 1983 and since then boreholes from various groundwater projects have been added to the

Data collection experiences in water level monitoring, borehole archive and research projects

37

monitoring network. The data serves various purposes such as establishment of benchmarks and parameters of various systems. Monitoring is comprised of water level and rainfall measurement for establishing the natural piezometric surface. Knowledge of the natural piezometric surface is also needed as an input in water resources modelling and groundwater recharge estimation efforts. National Borehole Archive acts as storage for all borehole data whereas Research projects collects data on various hydrogeological parameters. This data collection is wrought with problems that are of logistical, equipment and human nature. 2 MONITORING Water level measurement by manual dipping and change of rainfall charts are done on monthly basis in selected boreholes. The problems recorded in the data collection sheets of ten monitoring networks are depicted in Figure 1.

Figure 1. Distribution of problems as recorded in water level records for 1999–2002 period. 3 MONITORING LOGISTICAL ISSUES However, this schedule is not followed due to several problems such as poor accessibility due to the country’s hostile environment and its vastness. Since the present monitoring network is determined by the groundwater potential projects carried out, most of the monitored areas are in remote areas far from the main office and are laden with heavy

Water resources of arid areas

38

sands and muddy soils from heavy thunderstorms. This means that a four wheel drive vehicle is a necessity. However, in some cases these vehicles are not available and this results in data gaps for some months. The other problem is late availability of transport which results in the late changing of water level/rainfall charts and water sampling. This renders the data useless as it results in the superposition of more than one line on the charts making it difficult to read. In some cases, other stakeholder dealing with water connect these monitoring boreholes to water supply out of emergency to supply people with water and this creates data gaps in the records as the borehole will not be accessible and cannot be used. In other cases, other water authorities drill production boreholes adjacent to these monitoring boreholes and the pumping creates interference and as such the borehole will be taken out of the network. Some of the boreholes are in private property and in some cases there is no access as the gates are locked. In other situations the dipper is reported to be stuck for the whole year such as in borehole 6736 in Lethakeng/Bothapatlou monitoring network. 4 MONITORING DATA COLLECTION ISSUES 4.1 Recording One of the major problems is the recording system. There are several problems associated with data recording and these include the following: – Unclear or no comments, this include things such as reporting same borehole dry and blocked/ collapsed for different months or reporting “no hole for dipper” – Water level taken from a borehole with unknown number – Incorrect entry of measured values – Measured values being different from the chart recorded value. This unclear recording resulted in data being discarded hence creating data gaps. 4.2 Quality check All the collected data between 1999 and 2002 was just filed without being checked and the lack of quality check was evident by a lot of problems encountered during the quality assurance of this 1999–2002 data. This lack of timely quality checking resulted in some data points being thrown out as it was difficult to know the exact reasons for these discrepancies/anomalies. This included problems such as recording problems. 4.3 Data storage Until 2003, all the collected hard copy data was not being digitised and it was not filed properly. The records/charts were either misfiled in different binder or thrown in drawers. This resulted in some of the records missing and these added to the issue of unavailable data.

Data collection experiences in water level monitoring, borehole archive and research projects

39

4.4 Equipment Lack of proper preparation for the field also creates data collection problems. In some cases, it is reported that dipper or its light was not working and this results in partial or no data collection. Due to lack of timely quality check it becomes difficult in ascertaining the source of the mal/non-functioning of the equipment, whether it is the sensor or the batteries. The other problem was from the mechanical water level recorders being used, these were perceived as a better replacement of manual dipping, and however, they came with their own shortcomings. In most cases, the monitoring equipment used was designed and manufactured in Europe where the environment is completely different from the semi arid conditions of Botswana. This has lead to tremendous reduction of the lifespan and the poor performance of these instruments. Some of the problems experienced include – Stuck pens rendering the data – No marking on the chart resulting in blank chart.

5 MONITORING HUMAN RESOURCES ISSUES 5.1 Knowledge In most cases, the personnel operating some of the equipment such as water level recorders, dippers, sampling pumps and rain gauges lack the technical know-how necessary to implement first line maintenance. This result in late or no acquisition of data and hence data gaps develop while the equipment is sent for maintenance or replacement. The lack of knowledge sometimes results in the equipment not being calibrated or set up properly and this indicated off scale water level and rainfall curves. 5.2 Availability The personnel used for data collection are at technician and artisan level. In most cases these personnel are shared among the various on going research projects and the normal monthly monitoring. This results in one of these activities suffering due to nonavailability of the personnel for a certain period and as such data gap will be inevitable. 6 NATIONAL BOREHOLE ARCHIVE (NBA) NBA records borehole data on daily basis. This includes registering privately drilled boreholes; entering borehole data in the database including plotting boreholes on hardcopy maps and storing rock chip samples. 6.1 NBA Logistical issues There are several logistical problems associated with NBA. One of these problems is the running out space for rock chip samples storage. The lone core shed is full and this has resulted in the new and reliable samples being piled without proper storage. The other

Water resources of arid areas

40

problem is the use of outdated topographical maps for plotting boreholes which results in difficulty in borehole location verification. The lack of physical verification of registered boreholes is also one of the problems. This is due to the fact that boreholes are drilled almost daily all over the country and due to its vastness; it is difficult to cover all the drilled boreholes. 6.2 NBA Quality check Since the NBA is the storage for all drilled boreholes in the country, the issue of data quality check is very important. However, until recently it has been neglected especially on the privately owned boreholes. As an authority responsible for registration of private boreholes, one of the main tasks is to verify the location of the boreholes as provided by the owner/driller so that the borehole could be plotted correctly in the map. However, this has not been done adequately and as such a lot of boreholes have uncertain location. 6.3 National Borehole Archive Human Resources issues 6.3.1 Availability The personnel used for data collection are at technician and artisan level. In some cases these personnel is shared between NBA and the normal monthly monitoring. This results in one of these activities being suspended for a certain period and creating backlog. 7 RESEARCH PROJECTS (RP) The Hydrogeology division has been running various projects ranging from groundwater potential survey to hardcore research projects such as Groundwater Recharge Evaluation Studies (GRES) and the Kalahari Research Project. These projects are multi-disciplinary and use different equipment and collect different data sets. In most cases these projects are carried out jointly with external partners and therefore timely bound. These projects, especially those run in-house experience a lot of problems. 7.1 Research Projects logistical issues In most cases these projects are carried out in remote areas and several problems are experienced in the field. This includes transport problems such as vehicle breakdown which takes a long time to fix due to the long process to be followed. In some cases there is a need to seek permission from other stakeholders and these requests can result in extension of the program while waiting for approval. This includes funds approval and changing the scope of the study. For example, the approval of funds and changing of project scope may take three (3) to four (4) months of valuable field activity time. This negatively affects projects that require time based data.

Data collection experiences in water level monitoring, borehole archive and research projects

41

8 RESEARCH PROJECTS HUMAN RESOURCES ISSUES 8.1 Knowledge In most cases these projects use specialised equipment that requires special operating level. However, most of the personnel have never been exposed to this equipment and this can result in loss of data or collection of unreliable data. For, example various software used to operate Skye data loggers are still unknown to a good number of hydrogeological technical personnel within Hydrogeology Division. 8.2 Availability The personnel used for data collection are at technician and artisan level. In most cases these personnel is shared among the various on going research projects and the normal monthly monitoring. This lack of technical level staff impacts negatively on the running of these projects. This results in the project using unqualified staff to fill the gap; however, that has serious implications on the quality of the collected data. In other situations, there is a need to have specialised personnel such as a welder to develop a specialised piece of equipment on site. This can result in delays especially if that person is unavailable or is occupied with other departmental work. 8.3 Equipment Some of this necessary specialised research equipment needs special care and due to harsh conditions prevailing in these remote areas, a lot of time is lost when the equipment breaks down since it must be sent overseas to be fixed. This also results in loss of data especially temporally dependent data. 9 CONCLUSION On the basis of this experience, we conclude that the following aspects are vital for a successful and reliable hydrogeological data collection effort in the semi-arid Botswana environment. – Routine analysis of the archived data should be a must rather than an option, in order to eliminate useless data before it accumulates in large amounts within records. – Routine training programmes for technicians on field equipment should be designed to enable technicians to keep abreast with the changing technology that is specific for hydrogeological applications. – Increasing manpower capacity by recruiting personnel with basic hydrogeological monitoring and database knowledge in order to facilitate data collection and reduce data losses arising from lack of knowledge. This will increase the reliability of the collected data. – The general public and other stakeholders need to be informed about the importance and relevance of hydrogeological research, borehole archiving and monitoring

Water resources of arid areas

42

activities, so that they can allow such activities in their private properties such as farms. – Manufacturers must be encouraged to design field equipment suited to the hostile semi arid and saline conditions of Botswana, so that the durability of the field equipment can be guaranteed. – Periodical inspection of water level monitoring boreholes should be a must in order to curtail issues of “dry” or “blocked” boreholes, hence maintain a continuous and an accurate water level record. – A comprehensive process map of water level monitoring program which include recording of the environmental status or changes in the vicinity of the monitoring borehole such as new pumping borehole. – Conduct a routine water sampling of observation boreholes. Currently the Hydrogeology Division is engaged in improvement of data collection and archiving through implementation of the following programs. – Acquisition of digital water level and rainfall recorders and accessories to replace mechanical ones and manual dipping. This will reduce human errors and improve data quality. – Development of proper databases and checking data immediately from the field to ensure that issue of unclear comments and data anomalies are reduced hence maintaining good quality data. – Development of process maps to improve the quality of the data being collected and being entered into the databases. This will ensure that all factors are considered before a inexplicable conclusions such as “dry” boreholes are reached. – Field programs are being carried out to review borehole location maps. This is to ensure that borehole locations are correct and indeed the plotted boreholes do exist. Regular data collection even if it is not part of a specific study, helps to build a picture of the general behavior of the system. The data collected provides valuable comparisons and context when the system is studied in more detail. However, all this will not be possible if the data collected is wreaked with a lot of problems. REFERENCES Department of Geological Survey. 2003a. Groundwater Monitoring of Non-Operational Wellfields and Other areas of Development Interest, compiled by T.Kellner, vol. 1b–1d. Department of Geological Survey (2003), Groundwater Monitoring of Non-Operational Wellfields and Other areas of Development Interest, compiled by T.Kellner, vol. 2b, 2d–2f, 2h–2k.

Rainfall characteristics in semi-arid Kitui district of Kenya A.O.Opere, V.O.Awuor, S.O.Kooke & W.O.Omoto Department of Meteorology, Nairobi, Kenya Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Variable semi-arid climate characterized by precipitation patterns unfavorably distributed in space and time and high evaporation rates reaching up to 100% of the incoming monthly precipitation is a challenge facing water resources management in arid and semi-arid lands (ASALs). Kitui district in Kenya is an example of an ASAL environment where water resources management issues are particularly important and sensitive. Sources of water are nearer to the people in the wet season, but as the seasonal rivers dry up, distance to water points get as far away as 25–30km. There is, however, great potential for rainwater harvesting. This is dependent on proper understanding of the patterns of precipitation both in space and time. This would be useful in understanding drought characteristics in order to develop strategies to capture, store and redistribute the available water. The spatial characteristics were determined through principal component analysis (PCA). Season lengths, drought severity and frequency were determined. The results indicated that, on average, the onset for the long rains (March–May) was centred on day 82.36 while cessation was on day 126.3. The longest season was 107 days during the long rains while the shortest lasted 7 days during the short rains (October–November). Severe droughts in the district were experienced in 1980, 1985, 1990 and 1995. The largest seasonal total for the long rains was 768mm with a return period of 25.9 years while the smallest total was 81.1mm with a return period of 1.0 years. For the short rains, the largest total was 1022.0mm with a return period of 15.7 years and the smallest total was 205.4mm with a return period of 1.0 years.

1 INTRODUCTION Efficient and sustainable use of available water resources is paramount for a peaceful, sustainable and equitable development of any region. Some of the challenges to integrated and sustainable water resource management include:

Water resources of arid areas

44

Changing land use, land degradation by erosion, deteriorating water quality and competing water demands by stakeholders. There exists, therefore, a strong demand for an integrated water allocation and decision support system. The backbone of such a system must be scientifically sound to be accepted and trusted by stakeholders. Efforts to improve catchment management and to impose a sustainable water resources management are of economic and political importance for any country. Variable semi-arid climate characterized by precipitation patterns unfavorably distributed in space and time and high evaporation rates reaching up to 100% of the incoming monthly precipitation is a challenge facing water resources management in arid and semi-arid lands (ASALs). Kitui district in Kenya is an example of an ASAL environment where water resources management issues are particularly important and sensitive. Sources of water are nearer to the people in the wet season, but as the seasonal rivers dry up, distance to water points get as far away as 25–30km. The main problems in water development and management in this district include: ● unreliable rainfall and inadequate supply to meet the demand, ● the available water resources are unevenly distributed and inaccessible to all, ● traditional farming methods lack water conservation principles, ● most of the water projects have since been abandoned, and ● High rates of potential evaporation on the available water resources. There is great potential for rainwater harvesting. This is dependent on proper understanding of the patterns of precipitation both in space and time. This would be useful in understanding drought characteristics in order to develop strategies to capture, store and redistribute the available water. Droughts have been the phenomena of great concern throughout the continent of Africa, because of the devastating effects they have inflicted on the economies of some of the countries in the continent. Kitui District located in Eastern Kenya is no exception and is an example of one of the most vulnerable areas to the effects of drought. Droughts are usually classified as meteorological, hydrologic or agricultural depending on the variable under investigation. Definition of droughts has also been given on the basis of theory of runs and stationary structure of time series, Yevjevich (1967). The most important variable in meteorological drought is rainfall, in hydrological drought is availability of water in rivers, lakes, reservoirs, and ground water resources; and in agricultural drought is the soil moisture content to sustain the crop growth. Drought analysis involves investigation of duration, magnitude or severity, frequency and regional spread of the event. There have been limited investigations on meteorological droughts in Kenya. However, substantial work exists on the drought characteristics for the various agro-climatic regions of South Africa; Dyer and Tyson (1977), Zucchini and Adamson (1984), Dent et al. (1987). The study investigates the duration, magnitude or severity as well as frequency aspects of drought within Kitui District.

Rainfall characteristics in semi-arid Kitui district of Kenya

45

2 DATA AND METHODS 2.1 Data: types, sources and problems 2.1.1 Historical data Historical data from thirteen rainfall stations within and around the project area were analyzed. The first part of this analysis was to determine the onset and cessation dates of rainfall in each year for the short and long rains. The duration for each season was then derived. 2.1.1.1 Filling in missing records Many rain gage records are incomplete. It is necessary to estimate the missing data in order to utilize partial records, especially in data sparse areas. The methods used include: ● Arithmetic average method ● Normal ratio method ● Correlation method ● Inverse distance method The problem of filling data at un-gauged location involves transmitting data at the nearby index gages to the un-gauged location. The missing data can be formulated as (1) ai is the weighting factor of the ith gage with record Pi and N is the number of index gages while Px is the rainfall to be estimated at x. The different methods differ in their methods of estimating ai’s where i=1, 2,…………………N 2.1.2 Data quality control One of the sources of error in rainfall measurement is the location of the gage in relation to obstructing objects such as trees and buildings. In the progress of time, trees grow and buildings come up. This means raingages must be moved. This may affect the consistency of the records from the raingage i.e. the records before and after the movement might be different. In addition, change in observational procedure might also affect the consistency.

Water resources of arid areas

46

Inconsistency in a rainfall record may be detected by graphical or statistical methods such as double mass analysis, the Vonn Neumman ratio test, cumulative deviations, likelihood ratio test and runs test. 2.2 Methods 2.2.1 Principal component analysis (PCA) The method of PCA involve the transformation of a greater number of unorthogonal (manifest) variables into smaller number of orthogonal variables, which present common causes of manifest variable changes. It can therefore reduce the dimensionality of a problem by replacing the measured variables and the inter-correlated variables by using a smaller number of uncorrelated variables. This can be useful in reducing the amount of basic data to be processed. Depending on the data, it is possible to interpret the orthogonal functions in terms of some underlying physical processes. Castell (1966) proposed a method of retaining significant factors in PCA solutions. Similar methods have been used by Ogallo (1988a,1989) and Basalirwa et al. (1995) for East Africa and Tanzania respectively. This method was used to group rainfall records from the study catchment into homogeneous zones. Mathematically, a variable Z may be transformed in terms of m common empirical orthogonal function (factors) and n unique factors as below: (2) where, Zi is variable i in the standardized form Fi represents the common orthogonal vector (factor) ui is the unique factor for variable i ai1=standardized multiple regression coefficient of the variable i on the common factor 1 (factor loading). The unique term diui=0 since principal component analysis does not consider the unique component of the variance. Details of this method are available in many referencesincluding Drosdowsky (1993), Ogallo (1989) and Basalirwa (1991) among others. 2.2.1.1 Identification of representative rainfall station Principal component analysis (PCA) solutions were used to identify the most representative station in the area of study. This formed the basis for further analyses including onset/cessation of rainfall, dry spell lengths and frequency. Two stations from the thirteen stations were chosen for detailed analyses. These were Kitui Secondary School (identified as station with highest communality from PCA results); thus is a representative station based on communality concept. Kitui Water Office was also used for comparison of results.

Rainfall characteristics in semi-arid Kitui district of Kenya

47

2.2.2 Drought indices Drought duration is a crucial component particularly if one has to plan for storage that can last certain duration of drought for given water demand. The duration portrays the season lengths hence the potential success or failure of a water supply scheme that can be put up. At each of the stations, the onset and cessation dates were determined for each year and for each season i.e. long and short rains. The actual, earliest and latest onset and cessation dates was the basis of deriving average, shortest and longest duration at each of the stations. Mhita (1990) has made similar attempt in Tanzania. Two methods used were the water balance technique and pentad method. Both methods are based on preset hydrological conditions, to determine the onset and cessation of rainfall. Details of these methods are briefly discussed in the following sections. 2.2.3 Pentad method Definition of the start of the rains that is used is based on preset hydrological conditions. The first occasion after March 1st and October 1st that the running 5 day total exceeded 25mm and there being no dry spell exceeding 7 days in the next 21 days (Successful start, threshold of 1mm). In a nutshell, the pentad method involves computing a 5-day total rainfall for each year. The cumulative values of the 5-day total are divided by the annual total for each year and expressed as a percentage, that is, ΣPi*100/Annual total. These are plotted against pentad numbers. The onset and cessation dates are then determined from the plots. 2.2.4 Water balance technique The first occasion after March 1st and October 1st for the long and short rains respectively when the water balance goes to zero (capacity 100mm, daily evaporation 6mm). A water balance approach was used with a threshold of 1mm of rainfall. Evaporation rate for the area is taken to be 6mm per day on average. The soil moisture capacity was taken to be 100mm (i.e. average soil moisture during dry days). No runoff is generated since the rainfall amounts cannot even satisfy the evaporation demand. 2.2.5 Determination of season duration The maximum duration for each season was obtained using the earliest onset and latest cessation for the period of study in each of the two stations. The longest and shortest duration for each season was also determined for the two stations. 3 RESULTS AND DISCUSSIONS The results of the PCA, onset/cessation of rainfall, dry spell lengths and frequency are presented in the following sections.

Water resources of arid areas

48

3.1 Results of the principal component analysis From the results, three Eigen vectors were retained based on the Scree method (see Fig.1).

Figure 1. Scree test of Castell. Kitui Sec. School had the highest communality and was thus picked as the most representative station for further analysis. Results from PCA can be seen in Table 1 and Figure 2. The spatial map was obtained by mapping the factor loadings at the station locations. Three homogeneous regions were delineated from these results indicating complex rainfall variability within the study area.

Table 1. Rotated loading matrix. Variable (rainfall stations) 9137012 9137010 9137003 9138000 9137020 9137028 9137045 9137058 9137073

Factor 1

2

3

0.212 0.0373 0.164 0.458 0.191 0.385 0.757 0.274 −0.027

0.288 0.119 0.154 0.122 0.786 0.293 0.179 0.531 0.141

0.325 0.656 0.643 0.549 0.171 0.396 0.331 0.251 0.770

Rainfall characteristics in semi-arid Kitui district of Kenya

9137076 9137094 9138003 9138013 Variance explained % of total Variance

49

0.796 0.196 0.056 0.063 0.792 0.082 0.848 0.129 0.106 0.546 0.147 0.256 2.908 1.876 2.290 22.37 14.429 17.612

Figure 2. Homogeneous rainfall zones of Kitui. Table 2. Examples of onset/cessation dates. Average Earliest/shortest Latest/longest Std. (Day (Day no.) (Day no.) dev. no.) (Day no.) 9137012 LR onset LR cessation LR duration SR onset SR cessation

96.9 131.4

76 115

123 11.84 141 8.19

35.45

7

56 11.37

307 342.2

280 333

335 12.89 366 8.483

Water resources of arid areas

SR duration 9137010 LR onset LR cessation LR duration SR onset SR cessation SR duration 9137003 LR onset LR cessation LR duration SR onset SR cessation SR duration 9138000 LR onset LR cessation LR duration SR onset SR cessation SR duration

50

36.18

11

69 17.31

82.35 143.1

0 124

122 25.21 163 10.04

62

18

149 30.68

306.9 349.4

287 328

319 9.462 366 12.45

46.17

14

76 19.04

93.7 138.4

64 122

129 16.35 152 6.905

46.05

11

83 20.02

307.8 342.3

289 336

324 10.26 361 8.42

37

16

71 15.02

97.92 124.5

78 99

144 14.44 172 16.38

27.63

8

47 12.15

316 353.8

296 336

341 12.51 366 11

41.13

6

60 15.66

Rainfall characteristics in semi-arid Kitui district of Kenya

Figure 3. Water balance for Kitui Water Office.

Figure 4. Pentad for Kitui Water Office 1989.

51

Water resources of arid areas

52

Figure 5. Water balance for Kitui Sec. School—1984. 3.2 Onset and cessation dates of rainfall Some results for onset/cessation dates and season duration are given Table 2. A few examples of the results from pentad method for the two stations are given in Figure 4 and Figure 6. The annual water balance plots from the water balance technique is also given in Figure 3 and Figure 5.

Figure 6. Pentad for Kitui Water Office 1988.

Rainfall characteristics in semi-arid Kitui district of Kenya

53

Table 3. Dry spell lengths for Kitui Sec. School and Kitui Water Office. Kitui Sec. School Year L S 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998

22 31 27 24 35 27 84 18 36 66 32 81 69 70 33 8 55 61 37 12 14 23 7 16

Kitui Water Office Year L S

18 1981 23 1982 39 1983 155 1984 18 1985 79 1986 23 1987 14 1988 169 1989 9 1990 17 1991 149 1992 83 1993 41 1994 18 1995 12 1996 15 1997 27 140 11 13 114 107 26

11 10 23 8 13 56 31 43 18 22 70 46 20 5 7 114 7

12 17 25 67 30 13 72 28 47 17 72 64 28 26 31 23 107

The two methods were found to be comparable. For example, Kitui Water Office: Average onset of long rains is on day 82.36 and cessation on day 126.3 using the two methods. On the other hand onset is on day 84.37 and cessation on day 137.5 using the pentad method. These results are also presented in Table 2.

Water resources of arid areas

54

Figure 7. Dry spell lengths during long rains—Kitui Sec. School.

Figure 8. Dry spell lengths during short rains—Kitui Water Office. 3.3 Season lengths/duration Some of the results of the season lengths for Kitui Water Office and Kitui Sec. School are given in Table 4 and presented in Figure 7 to Figure 10. From the results in Table 4, for instance, Kitui Sec. School shows the longest long rains duration in 1997 (107 days). The same year also shows the shortest short rains duration (7 days). The result is confirmed in Kitui Water Office for the year 1997.

Rainfall characteristics in semi-arid Kitui district of Kenya

55

3.4 Drought severity This was based on the anomalies of the seasonal totals for each year for the two stations. A normal expectation was taken to be ±0.5 s.d. For Kitui Water Office, 1985, 1990 and 1995 were severe in terms of the long rains totals. The same years are also severe for the short rains. In Kitui Sec. School, 1980, 1985, 1990, 1995 were severe during the long rains as well as during the short rains. The seasonal totals for the very dry years can be used as a basis for planning for a water storage facility. Examples of the anomalies are given in Figure 11 and Figure 12.

Figure 9. Dry spell lengths during long rains—Kitui Water Office.

Water resources of arid areas

56

Figure 10. Dry spell lengths during short rains—Kitui Sec. School. 3.5 Drought frequency Gamma distribution was used to fit the seasonal totals for the period of study. The results are shown in Figures 11–12 and Table 4, while Figure 13 indicates the gamma distribution. In Kitui Water Office (KWO), the largest seasonal total for the long rains is 768.3mm and has a return period of 25.9 years. The smallest seasonal total is 81.1mm with a return period of 1.0 years. For the short rains, the largest total is 848.6mm with a return period of 20.1 years and smallest total is 98.76mm with a return period of 1.0 years. In Kitui Sec. School (KSS), the largest seasonal total for the short rains is 1022.0mm with a return period of 15.7 years while the smallest seasonal total n is 205.4mm with a return period of

Table 4. Values expected frequency MAM(KWO) OND(KWO) Gamma dist. Gamma dist. Mean 344.1 & Mean 441.2 & k of 2.929 k of 4.133 <=81.1 0.71 81.1 to 106.60.64 106.6 to 122.30.49

<=98.760.23 98.76 to 167.10.98 167.1 to 175.80.19 175.8 to 198.60.57

OND(KSS) Gamma dist. Mean 578.1 & k of 4.799 <=205.40.73 205.4 to 244.50.57 244.5 to 261.20.30 261.2 to 319.31.30

Rainfall characteristics in semi-arid Kitui district of Kenya

122.3 to 198.6 to 178.52.15 244.11.40 178.5 to 244.1 to 183.80.23 257.40.46 183.8 to 257.4 to 192.60.38 306.31.82 192.6 to 306.3 to 197.10.20 424.90.72 324.9 to 197.1 to 20 1.70.20 325.40.02 201.7 to 325.4 to 286.63.72 342.30.66 286.6 to 333 342.3 to 1.86 405.32.40 333 to 344.50.43 405.3 to 417.70.45 344.5 to 417.7 to 391.81.61 427.20.34 391.8 to 427.2 to 477.2 425.91.00 1.66 425.9 to 477.2 to 459.70.87 507.60.91 459.7 to 507.6 to 51 465.10.13 10.10 465.1 to 511 to 546.81.59 540.70.80 546.8 to 540.7 to 594.80.68 570.40.72 594.8 to 570.4 to 658.20.67 632.61.25 658.2 to 632.6 to 768.30.72 672.10.64 >768.30.73 672.1 to 690.70.26 Maximum 690.7 to T=25.9 723.80.41 Minimum T=1.0 723.8 to 724.20.00 724.2 to 848.61.04 >848.60.95 Maximum T=20.1 Min T=1.0

319.3 to 333.30.37 333.3 to 338.30.14 338.3 to 402 1.85 402 to 537.14.21 537.1 to 571.41.01 571.4 to 583.50.34 583.5 to 592.30.24 592.3 to 626 0.90 626 to 69 1.9 1.56 691.9 to 727.30.73 727.3 to 791 1.13 791 to 848.50.82 848.5 to 909.50.70 909.5 to 979.30.61 979.3 to 10220.29 <10221.21 Maximum T=15.7 Minimum T=1.0

57

Water resources of arid areas

58

Figure 11. OND seasonal anomalies— Kitui Water Office. 1.0 years. For the long rains, the totals are 757.1mm with a return period of 29.8 years and smallest total is 73.22mm with a return period of 1.0 years. 4 CONCLUSION Three rainfall regions were delineated. The longest spell is in 1997 (107 days) during the long rains. The same year also showed the shortest spell (7 days) during the short rains Average onset for long rains is on day 82, cessation was on day 126. The years 1980, 1985, 1990, 1995 were severe during long rains as well as during short rains. Average onset for long rains is on day 82, cessation was on day 126 using the two methods. The largest seasonal totals had return periods ranging between 15–25 years while the smallest seasonal totals had return period on average of 1 year, and indication of frequent drought. This information is crucial if water management and planning is to meet a particular demand in a specified duration of water stress. REFERENCES Basalirwa, C.P.K. 1991. Raingauge network designs for Uganda, Ph.D Thesis, Univ. of Nairobi, Kenya. Basalirwa, C.P.K. et al. 1995. The climatological zones of Tanzania based on rainfall characteristics. Water Resources Engineering Research Report, University of Dar-es-Salaam. Castell, R.B. 1966. The Scree test for the number of factors. Multivar. Behav. Res., 1:245–276.

Rainfall characteristics in semi-arid Kitui district of Kenya

59

Dyer, T.G.J. and Tyson, P.D. 1977. Estimating above and below normal rainfall periods over South Africa. Journal of Applied Meteorology 16(2):145–147. Dent, M.C., Schulze, R.C., Wills, H.M. & Lynch, S.D. 1987. Spatial and temporal analysis of the recent drought in the summer rainfall region of Southern Africa, Water SA, 13(1):37–42. Drosdowsky,W.1993. An analysis of Australian seasonal rainfall anomalies 1950–1987I: J climat. 13:1–30. Mhita, M.S 1990. The onset and cessation of rains andr importance for cropping strategies in Tanzania. Ogallo, L.A. 1988a. The spatial and temporal clusters of the East African Seasonal Rainfall anomalies derived from Principal component analysis. J.Climatol. 6:1–23. Ogallo, L.A. 1989. The spatial and temporal patterns of East African Seasonal Rainfall derived from Principal component analysis. J.Climatol. 9:145–167. Yevjevich, V. 1967. An Objective Approach to Definitions and Investigations of Continental Hydrologic Droughts. Hydrol. Paper 23, Colorado State University, Fort Collins, Colorado. Zucchini, W. and Adamson, P.T. 1984. The occurrence and severity of droughts in South Africa. WRC Report No.91/1/84, Department of Civil Engineering, University of Stellenbosch, South Africa.

Quantification of the impact of irrigation on the aquifer under the Vaalharts Irrigation Scheme R.G.Ellington, B.H.Usher & G.J.van Tonder Institute for Groundwater Studies, University of the Free State, Bloemfontein, South Africa Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: The Vaalharts is the largest irrigation scheme in the country, with water imported from the Vaal River to supplement to rainfall in the area. Approximately 30000 hectares of land is currently being irrigated. The salinity of the irrigated water has steadily increased over time. Several previous research projects have been undertaken to determine the fate of the added salts. The conclusion in these reports is that a very large proportion of the salts added to the subsurface due to irrigation are not returned to the surface water. The underlying aquifer was postulated as sink for these salts, with limited storage capacity. Once this capacity has been exceeded, a flow reversal was postulated to occur. This process is likely to add a tremendous salt load (estimated to be approximately 100000t/year) to the Harts river system. The adverse effects of such an addition would be catastrophic to the irrigation scheme, and all downstream irrigation schemes and water users. Investigations into the hydrogeology and hydrochemistry were conducted to quantify the impact of irrigation on the groundwater resources. This included drilling, aquifer testing, groundwater monitoring and empirical and numerical modeling. Findings included a water and salt balance for the area and the understanding of the underlying lithology as fractured rock aquifer.

1 INTRODUCTION The Vaalharts Irrigation Scheme was initiated approximately 55 years ago. It is the largest irrigation scheme in South Africa at approximately 32000ha. Vaal River water is transferred via an extensive canal system from Warrenton into two subsequent canals, namely the North Canal and West Canal. This research entailed a detailed groundwater investigation including drilling of additional boreholes, aquifer parameter determination from slug, pump and tracer testing,

Quantification of the impact of irrigation on the aquifer

61

groundwater monitoring of water levels and water quality for more than a year, and the construction of water and salt balances using empirical and numerical modelling techniques. 1.1 Overview of the study area The geology within the Vaalharts valley is largely sedimentary of Karoo age, although the pre-Cambrian basement geology appears igneous. The Vaalharts valley is largely overlain by aeolian Kalahari sands. Also of Quaternary age are calcretes and alluvial gravels. Below these Quaternary sediments lie shales, tillites and mudstones. The preCambrian igneous lithologies form the lower boundary of the system. Natural drainage has been found to be poor. This is attributable to the flat topographical gradient, and typical soil profiles found in the area. The upper, generally impermeable calcretes are found at depths varying between 0m and 5m (Gombar & Erasmus, 1976). According to Streutker (1977) the water table was found to be lying at approximately 24m below ground level (mbgl) for the period between 1935 and 1940, although it seems that no comprehensive borehole drilling to determine the water levels in the irrigation area was undertaken across the entire scheme (Herold & Bailey, 1996). No extensive measurement of the water levels seems to have been undertaken during the period of 1940’s to 1970’s. To combat waterlogging, a comprehensive network of 240 subsurface drains was installed between the years 1976 and 1979 at an approximate depth of 1.8mbgl. The drains were found to successfully control the water table, and in so doing, improve the crop yields. In 1976, prior to the drains’ installation, approximately 3000ha of soils were saline or saline-sodic to a depth of 0.3mbgl. The end of 1977 had reduced this reduced to approximately 1500ha, while in 1980 there remained approximately 1000ha of saltaffected soils (Herold & Bailey, 1996). 1.2 Previous investigations Research has been conducted in the Vaalharts since the 1960’s addressing increasing water levels and salinisation within the irrigation area. Most recent was a report by Herold & Bailey (1996) discussing the long-term salt balance for the Vaalharts Irrigation Scheme. This report stated an annual loss of 100000t of salts to groundwater, and predicted that, as these salts were not being measured in the Harts River, that they would be seen in the form of a sudden salt reversal to the Harts River, thereby adding a massive strain to an already stressed river system. The basic hypothesis for this to occur was the existence of a “perched” aquifer below the irrigation scheme, which in turn is underlain by a deep aquifer. This deep aquifer would be the sink of these excess salts. Once the postulated deep aquifer’s storage was exceeded, these salts would be added to the Harts River, causing a water quality deterioration in the downstream Spitskop Dam. The irrigation schemes in the lower Orange River such as those at Douglas would be negatively impacted on. In another study Gombar & Erasmus (1976) sampled various boreholes in the North Canal area. The average TDS at the time was determined to be 1005mg/l.

Water resources of arid areas

62

2 METHODOLOGY 2.1 Hydrocensus An initial field recognisance study was undertaken. Literature indicated that 41 diamondprospecting boreholes had been drilled during the 1970’s. It was hoped that these boreholes could provide access to the aquifer. During a hydrocensus inspection to locate these boreholes, all were found to either be destroyed by farming practices, or have been blocked by stones. A second hydrocensus across the Vaalharts Irrigation Scheme to obtain an indication of the groundwater response was conducted. The boreholes located during this hydrocensus were equipped with mono-pumps, and had been encased in concrete. In total, 22 boreholes were discovered. Only pumped samples were possible from these boreholes. 2.2 Field investigation The need for drilling was made evident by the relative lack of accessibility to open boreholes, as the majority of boreholes are covered by cement blocks fitted with monopumps. The drilling method used was air percussion, with the boreholes drilled to a diameter of 0.165m, using a drill rig provided by the Department of Water Affairs and Forestry. A total of 17 boreholes drilled across the Vaalharts Irrigation Scheme, of which three were located on the riverbanks of the Harts River, and the remaining 14 were drilled on the plots. In order to determine the hydrogeologic variation over the extent of the lithology, the first seven boreholes were drilled to random depths until they reached the lava bedrock, while the remaining 10 boreholes were drilled to a depth of approximately 20m. The boreholes were predominantly cased with a 4mm steel casing, although three of the boreholes drilled on the banks of the Harts River were cased with Johnson screens. Furthermore, three of the seventeen boreholes drilled during this research were equipped with piezometers. In all boreholes, casing and screens used for the borehole construction were slotted to allow free flow of groundwater through the boreholes and accurate groundwater investigation. The boreholes were slotted from a metre below the depth of the Kalahari sands to prevent clogging of the borehole from these sands. Piezometers were installed in three boreholes across the North Canal area. The boreholes used are believed to present an accurate representation of the general geology in the Vaalharts. The piezometers were installed to test the conceptual model of Herold & Bailey (1996), where they assumed there to be two aquifers in the Vaalharts—an upper, perched aquifer relating to the calcretes, and a deeper aquifer. 2.3 Aquifer testing Three types of aquifer tests were undertaken to obtain the hydraulic parameters of the aquifer underlying the irrigation scheme.

Quantification of the impact of irrigation on the aquifer

63

2.3.1 Slug tests A slug test is a method used to measure the hydraulic conductivity or transmissivity of a borehole. This is done by measuring the rate of recovery or recession in the borehole, following a sudden addition into the borehole or extraction of water from the borehole of a known volume. The Bouwer and Rice (1976) equations were used to analyse the slug tests. 2.3.2 Pumping tests Multi-rate and constant discharge rate pump tests were undertaken on selected boreholes. Use was made of traditional analysis methods such as the Bisroy-Summers (1980) method for multi-rate tests, the Cooper-Jacob method for constant rate tests and the more recent suite of methods based on the Flow-Characteristics methodology (van Tonder et al., 2002). 2.3.3 Tracer tests Tracers are identifiable substances that, from the examination of their behaviour in a flowing medium, may be used to infer the general behaviour of the medium (Riemann, 2002). Leap and Kaplan first described the single-well Injection Withdrawal Tracer test for the estimation of groundwater flow velocities in 1988. The single-well Injection Withdrawal Tracer test is conducted by injecting a known volume of tracer solution into the test borehole, allowing the tracer to drift under the influence of the natural hydraulic gradient for a period, and then removing the tracer by pumping the test borehole to recover the tracer. Adopted methodologies as described by Riemann (2002) were used, with NaCl and NaBr as artificial tracer. 2.4 Groundwater monitoring Groundwater monitoring has been ongoing since inception of the drilling program. Six groundwater-monitoring runs have been undertaken in the period April 2003 to March 2004. Groundwater levels have been measured using electronic contact dip meters. Sampling of the boreholes has been undertaken with depth-specific sampling equipment. Hydrochemical profiling using YSI- Sonde 6000 multi-parameter probe which measures pH, specific conductance, dissolved oxygen and redox potential with depth, was done at three occasions. 2.5 Numerical and analytical methods Empirical and numerical methods were employed. 2.5.1 Analytical methods Darcy’s law was used to determine an expected flux to the river. Salt loads were calculated using the Ogata equation. The salt load estimation in the GW reserve program uses the Ogata equation (from Freeze and Cherry, 1998):

Water resources of arid areas

64

where: l=distance along the flow path, v=the average linear water velocity, t=time, C=concentration at time t, C0=initial concentration and D=the coefficient of molecular diffusion, for the solute in the porous medium. The expected salt leaching from the irrigation scheme was obtained from du Preez et al., (2000), who used the Aragues model and the Soil Water Balance to determine salt leaching from the scheme. These values were used as input in the empirical and numerical models for salt balance calculations. 2.5.2 Numerical models The numerical model used to simulate the aquifer system in the Vaalharts irrigation area was Modflow. Modflow is a modular two- or three-dimensional finite difference groundwater flow model that was developed by McDonald and Harbaugh of the United States Geological Survey for the purpose of computation of hydraulic heads in saturated porous medium with uniform water temperature and density (Harbaugh & McDonald, 1996). The Modflow mass transport program used during the Vaalharts transport modelling was MT3D. The Vaalharts numerical groundwater model is, as all groundwater models, a representation of the naturally occurring conditions. Certain assumptions therefore had to be made, while certain limitations also persisted in representing natural conditions. The following assumptions were made: ● The rivers in the area were treated as fixed heads. ● As there is no significant groundwater extraction in reality due to the water allocation from the Vaal River, no discharge was included. ● As there are large volumes of water being applied by irrigation, a higher volume for recharge was applied. ● The basic lavas in the stratigraphy were accepted as being the lower boundary within the stratigraphy due to their relatively impermeable nature. The model was assigned 320 rows and 152 columns with a cell size of 250m×250m. This equates to a model area of 3040km2. The model area’s co-ordinates are −3120000, −2000 (lower right corner) to −3040000, −40000 (upper left corner). The Vaalharts model constructed during this project made use of a two-layer model. Confined conditions were applied to these layers. The layer depths were based upon geology encountered during literature reviews of Vaalharts specific data and drilling that took place during the course of this project. The upper layer was assigned values for the sands, according to geological logs drilled during this and other projects, averaged at approximately 6m. The lower layer was assigned average values for the calcretes, clays, gravels and shales due to their relatively similar range of depths, and depths of the geological strata from borehole logs. The various pre-determined areas therefore each had

Quantification of the impact of irrigation on the aquifer

65

separate hydraulic conductivities applied to them, based upon knowledge of the geology and tested aquifer parameters. For instance, areas with a higher degree of gravels were assigned a higher hydraulic conductivity for that area. In other areas, where fractures with significant yields were encountered, an increased hydraulic conductivity was assigned. The drain package in MODFLOW was applied to the North Canal and West Canal areas during the simulations. The drains applied in the model accurately represent the processes occurring naturally in the Vaalharts system. The subsurface drains were positioned 2mbgl to simulate natural conditions. For the mass transport simulations, an initial concentration of 500mg/l TDS was assigned, while the input concentration from irrigation was obtained from the salt leaching models described by du Preez et al., (2000). 3 RESULTS 3.1 Geology From the drilling a consolidated geological model was constructed using the Rockworks (Version 3.4.1.6, Rockware Incorporated) program. The geology model shows thicker shales to the northern side of the Vaalharts Irrigation Scheme, with thinner shales in the south of the scheme. In addition, while the model indicates calcretes throughout the scheme, the calcretes are more pronounced in the southern half of the scheme. The southern half of the scheme’s geology is represented more by gravels and clays. This may be due to possible erosion and deposition as the Harts River meandered during the Vaalharts’ geological history. 3.2 Aquifer parameters The aquifer parameters obtained through pump and tracer testing show the underlying strata to be heterogeneous, with high transmissivity values where fracture zones are encountered. The tracer test results indicate a similar range of values, with high velocities occurring wherever fracture zones exist. 3.3 Water levels The water level variation over the time of measurement was not significant, with minor summer and winter levels observed. Apart from a few exceptions, the water levels in the area are very similar, varying between 1.6–2.0m below surface. This level coincides with the depth of the installed drainage systems prevalent in the area. Water levels in piezometers installed within the same borehole exhibited less than 1cm variation between piezometers installed to different depths.

Water resources of arid areas

66

3.4 Water quality The groundwater quality in the area varied spatially and is dependent on the geology and position relative to the irrigation. The majority of the electrical conductivity values are between 100mS/m and 270mS/m. The average TDS in the irrigation area was determined to be 1350mg/l. The two boreholes illustrating the highest and lowest electrical conductivities respectively, are interestingly enough both present on the same plot, and are within a 50m distance of each other. Borehole 6L16-1, illustrating the lowest conductivity of the samples obtained, is drilled within 10m of a canal, which seems to be leaking water into the groundwater system via cracks in the concrete. Borehole 6L16-2 is however located within 10m of irrigated land. Several of the groundwater samples have relatively high nitrate values, but considering this is a heavily

Table 1. Transmissivity values for tested boreholes. Borehole

Transmissivity (m2/d)

1G14-1 1K10-1 6L16-1 6L16-2 1D3-1 1D7-1 2J14_RIV-3 8H14-1 2J5-1

43.3 31.6 4.2 4.5 0.21 194.0 58.0 1.2 123.0

Table 2. Tracer test results. Borehole number 1B10-1 1D3-1 1D7-1 2J5-1 8H14-1

Darcy velocity (m/d) 2 1.5 3 1 22

Seepage velocity (point dilution) (m/d) 21 15 29 9 217

Quantification of the impact of irrigation on the aquifer

67

Figure 1. Electrical conductivity values of the samples taken from the boreholes drilled during this project. Table 3. Values used for Vaalharts Water Balance using empirical values, and model values where possible empirical values calculated for the Vaalharts Water Balance (Mm3/annum). Name North Canal West Canal Rainfall Groundwater going to river Canal Tailends Recharge Drainage Runoff Evapotranspiration Totals Difference (Inflow—Outflow)

Incoming water (Mm3/a)

Outgoing water (Mm3/a)

272.01 42.97 309.60 14.34

624.58 −0.41%

23.35 28.38 23.63 4.10 533.30 627.16

cultivated area the 50th percentile value for the groundwater of 2.2mg/l N is considerably lower than expected. As nitrate is often used as a tracer to highlight the effect of cultivation on water quality (e.g Pulido-Bosch et al., (1999)) such a low value would seem to indicate that vertical migration of salts from the cultivated lands to the groundwater is less pronounced than previously expected. This is confirmed by the low

Water resources of arid areas

68

potassium values, another key constituent in fertilizers, where 95% of the measured values fall under 15mg/l. Consideration of the major parameters using interpretive diagrams such as Piper plots, showed no dominant anion or cation, although some of the cations do tend towards the Na+K field. Comparison with surface waters and geology indicated a Mg- enrichment in the Dwyka shales, and sulphate largely from incoming Vaal River water used for irrigation. The in situ water quality, as determined through hydrochemical logging, exhibited only minor variations with depth. No significant evidence of stratification of poorer water quality was observed in any of the boreholes. 3.5 Water and salt balance Water balances were determined using the numerical model and empirical calculations for various scenarios.

Table 4. Salt balance permutations Vaalharts salt balance (tons/year). Components

Option 1 Option 2

North Canal 112884 112884 West Canal 17832 17832 Groundwater going to 15873 15873 river Canal Tailends 17977 17977 Drainage 17979 17979 Recharge1,2 84287 111758 25962 25962 Salts taken up by crops1 Fertilizer addition1 48302 48302 1900 1500 Salts in soils1 Incoming salts 179018 179018 Outgoing salts 163979 191050 Incoming less outgoing 15039 −12032 Percentage difference 8.401% −6.721% 1 Based on values obtained from du Preez et al., 2000 and the numerical model. 2 Upper and lower values of calculated salt leaching used.

For the purposes of the salt balance water qualities of various water types were obtained from DWAF, previous reports and measured in this project. Combination of the water balance, these concentration values and the output of salt leaching models reported by du Preez et al., (2000), allowed salt loads to be calculated. The following was used to for the salt balance. There are several permutations of these options but these give similar results regarding the overall salt balance. Of importance is the recharge salt addition to the groundwater system. Using the median value of approximately 98000t/year of salts added to the

Quantification of the impact of irrigation on the aquifer

69

groundwater and the assumption that the net storage in aquifer remains relatively constant over time, the expected net increase in TDS should be in the order of 14mg/l. 4 DISCUSSION AND CONCLUSIONS The main aim of the study was to ascertain the impact of many years of irrigation on the groundwater resource. More specifically, emphasis was placed on the assessment of previous hypotheses related to the aquifer system and salt migration within this system. 4.1 Previous hypotheses The regular water level measurements over a period of a year in the seventeen boreholes have shown that the subsurface drainage installed by the farmers is effectively controlling groundwater levels. All the water levels in close proximity to such drainage exhibit limited variation over time, and lie at a depth consistent with the installed drains. The installed piezometers in several boreholes have shown that the water levels in deeper and shallower systems are within a few centimetres of one another. The water quality in these piezometers is also very similar for each borehole. Hydrochemical profiling has indicated that no significant stratification of water occurs. All these factors point to the conclusion that the deeper lying aquifer, thought to be the salt sink, does not behave independently. The system is dynamic enough, and the shallow impermeable layers too localised, to have a system of cascading groundwater finding its way to a deeper system. This finding has positive consequences, in that no sudden catastrophic event of salt reversal is likely to occur. However, the indications are also that the groundwater quality is showing a steady deterioration over time. 4.2 Groundwater quality The ongoing monitoring has shown the groundwater quality to be fairly poor, but not as saline as was suspected. The average TDS of 1350mg/l compared to 1050mg/l in 1976 is cause for concern. The numerical model has shown that groundwater is expected to leave the scheme and contribute to water quality deterioration downstream in the Harts River system. The observed water quality change is approximately 13mg/l year. Based on the salt balance calculations, which takes the incoming and expected outflowing volumes into account, the increase was expected to be around 14mg/l. These two results are therefore in good agreement. 4.3 Water and salt balances The empirical water and salt balances showed good agreement with one another. These balances showed that irrigation is the most important driver on water quality and volumes in the system. The salt balance also highlighted the fact that the greatest contribution to the incoming salt load is the irrigation water sourced from the Vaal River. The salts added in this way are more than double those from fertilizer addition and management of this incoming water is therefore the key to the salt accumulation in the irrigation scheme.

Water resources of arid areas

70

Short of addressing the upstream Vaal River practices and means by which to ensure cleaner water entering the Vaalharts system, the Vaalharts itself needs to be addressed. The logical approach to ensure lower salinity water enters the groundwater in the Vaalharts system is to make use of less water. Since agriculture accounts for nearly 70% of all water withdrawn from rivers, lakes, and underground aquifers for human use, the greatest potential for conservation lies with increasing irrigation efficiencies (Clarke, 1991). What is needed in the Vaalharts is a more efficient manner of irrigation, where less water is applied per unit area, and therefore fewer salts enter the groundwater via leaching. A more efficient means of irrigating would be drip irrigation, with a field application efficiency of 95%, which is 40% to 60% more efficient than gravity systems (Postel, 1997). The installation of drip irrigation in the Vaalharts would increase the efficiency of irrigation, thereby reducing the volumes of water needed. This would simultaneously decrease the mass of salts applied to the Vaalharts Irrigation Scheme, and reduce the tonnage leaching to groundwater and, eventually, entering the Harts River. REFERENCES Annandale, J.G., Benadi, N., Jovanovic, N.Z. & Du Sautoy, N. 1998. SWB: A user friendly irrigation scheduling model. Soils and Crops towards 2000 Congress, South African Society of Crop production, Alpine Health, Kwazulu-Natal. Aragues, R.M. 1996. Conceptual irrigation return flow hydrosalinity model. In K.K.Tanji (ed.). Agricultural Salinity Assessment and Management. Am. Soc. Of Civ. Eng., New York. Clarke, R. 1991. Water: The international crisis. London, Earthscan:. 193. Birsoy, Y.K. & Summers, W.K. 1980. Determination of aquifer parameters from step tests and pumping data. Groundwater, 18:137–146. Chiang, W.-H. & Kinzelbach, W. 2000. Processing Modflow (PMWIN), Version 5.1. The Institute for Groundwater Studies, University of the Free State, Bloemfontein, South Africa, 9300. Du Preez, C.C., Strydom, M.G., Le Roux, P.A.L., Pretorius, J.P., Van Rensburg, L.D. & Bennie, A.T.P. 2000. Effect of Water Quality on Irrigation Farming along the Lower Vaal River: the Influence on Soils and Crops. WRC Report No. 740/1/00. Water Research Commission. Harbaugh, A.W. & McDonald, M.G. 1996. User’s documentation for MODFLOW-96, an update to the U.S. Geological Survey modular finite-difference ground-water flow model: U.S. Geological Survey Open-File Report 96–485, 56p. Herold, C.E. & Bailey, A.K. 1996. Long Term Salt Balance of the Vaalharts Irrigation Scheme. Water Research Commission. Gombar, O. & Erasmus, C.J.H. 1976. Vaalharts Ontwateringsprojek, Technical Report GH2897. Department of Water Affairs. Leap, D.I. & Kaplan, P.G. 1980. A single-well tracing method for estimating regional advective velocity in a confined aquifer theory and preliminary laboratory verification. Water Resources Research, 23(7): 993–998. Postel, S. 1993. Water and Agriculture. Water in Crisis. New York, Oxford University Press. pp. 55–66. Pulido-Bosch, A., Bensi, S., Molina, L., Vallejos, A., Calaforra, J.M. & Pullido-Leboeuf, P. 1999. Department of Hydrogeology, University of Almeria. Canada, Spain. Riemann, K. 2002. Aquifer parameter Estimation in Fractured Rock Aquifers using a combination of Hydraulic and Tracer Tests. PhD thesis. Institute for groundwater studies. South African Weather Service. 2002. Climate data. http://www.weathersa.co.za/

Quantification of the impact of irrigation on the aquifer

71

van Tonder, G., Bardenhagen, I., Riemann, K., van Bosch, J., Dzanga, P., Xu, Y. 2002. Manual on Pumping Test Analysis in Fractured-Rock Aquifers. WRC Report No. 1116/1/02. Water Res. Comm.

Theme B: Groundwater recharge: natural and artificial

Groundwater development—identification of artificial recharge areas in Alla, Eritrea K.S.Viswanatham, Filmon Tesfaslasie & Michael Asmellash Water Resources Department, Government of Eritrea, Asmara, Eritrea, N.E. Africa Arun Kumar Department of Earth Sciences, University of Asmara, Asmara, Eritrea, N.E. Africa S.A.Drury Department of Earth Sciences, The Open University, Milton Keynes, UK Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Remotely sensed data acquired by the EO satellite have been analyzed to demarcate potential target zones for the development of future irrigation projects in a part of Eritrea. The study area is one of the important centers of horticulture activities and is currently facing water problem. Various lithologic units, tectonic signatures, land use and geomorphology related to groundwater assessment have been identified and interpreted from the imagery. The northern part of the area, covered by alluvial fans and presence of NW-SE dykes indicates favorable potential areas for groundwater development. Based on present investigation, construction of check dam in the Ghadien River and subsurface dam near Bazit Village is recommended in order to improve groundwater potential in the area.

1 INTRODUCTION The advent of polar-orbiting satellite remote sensing has provided hydrogeologists with a sophisticated and reliable tool for rapidly assessing natural resources of an area with reasonable accuracy. The focus of this paper is on the analysis of remotely sensed data combined with ground truth to delineate geological and geomorphologic patterns and their effect on groundwater occurrence and movement. The area selected for the study lies south east of Asmara, the capital city of Eritrea, in Sub-Sahelian Africa (Fig. 1). Alla-Ghadien and its vicinity are well known for horticultural activities. Irrigation is from open dug wells and in a limited way from bore

Water resources of arid areas

76

wells. The continual exploitation of water resources has resulted in gradual decline of water levels and also a decrease in well yields (Habteab 2000). The area lies within a depression at the edge of the Red Sea escarpment, and is an intermontane basin enclosed by several outstanding ridges. Ghadien is the administrative center of five villages, Alla, Belesto, Adi-Asambo, Adi-Moya and Ghadien itself. It is located 16km northeast of Dekemhare town. The elevation of the area ranges from 1680m to 1936m above mean sea level. The horticultural activities in the area commenced during the Italian colonization. In the early times, sufficient groundwater was found at shallow depths (five meter); however at present it is difficult to get water in some areas, even at depths of 30 meters (Habteab, 2000). The total population of the Ghadien administrative center is about 2773, with a total cultivated area of 2888.24 hectares. The average annual rainfall is 463mm. Based on the information of the Ghadien administrator the number of wells is about 600, of which 200 are dry (Habteab, 2000). Some of the existing problems in the area are—acute shortage of groundwater, lowering of water table, high density of wells with close spacing, over exploitation of groundwater, mismanagement of the existing water, lack of assessment of groundwater potential, extensions of irrigated areas,

Figure 1. Location map. construction of wells and small dams without knowing the groundwater potential. A reconnaissance survey in April 2001 indicated a lowering of water levels and decreasing

Groundwater development—identification of artificial recharge areas in Alla, Eritrea

77

well yields. The main objective of the present work is to identify the existing hydrogeological problems, to test the water quality at random sites and to evaluate water use and management. With the help of image interpretation and field checks, areas for future groundwater development and suitable site for check dam and subsurface dam have been suggested. The remotely sensed data employed are from the NASA EO-1 satellite’s Advanced Land Imager (ALI) instrument (NASA, 2001). The deployment aims at testing the suitability of 9 spectral bands in the visible to short wave infrared, with 30m resolution, and a single panchromatic band with 10m resolution, for the eventual replacement of the Landsat-7 Thematic Mapper. The chosen bands include 6 similar to that of the TM instrument, with an additional 3 that augment possibilities for vegetation and mineral discrimination. Sharpened by the 10m panchromatic band, various combinations of the spectral bands offer unprecedented (for Eritrea) opportunities for resource assessment. 2 PHYSICAL SETTING Eritrea is located on the western flank of the Red Sea at about 12.5°–18° North, 36.5°– 43.5° East. Its border length in the north and west with Sudan is about 624km, that with Ethiopia to the south is about 917Km, and that in the southeast with Republic of Djibouti is about 104km. With a total area of about 124000km2, it has more than 350 islands and a coastline of more than 1200km. The population of Eritrea is approximately 3.5 million. Eritrea has five major river basins, namely Mereb-Gash, Setit, Barka-Anseba, Red Sea and Danakil depression. All the rivers (except the Setit) and their tributaries are mostly seasonal and intermittent. It has four physiographic regions namely the Central Highlands, Western Lowlands, Eastern Lowlands and Coastal Lowlands. The average temperature ranges from 3°C–28°C in the highlands and 20°C–48°C in the lowlands. Eritrea being an arid and semi-arid country is not endowed with rich water resources. It has a vulnerable environment due to recurrent and devastating droughts, being part of Sahelian Africa. Rain-fed agriculture is the main occupation of most rural people. The majority of the population depends on groundwater as the main water-supply source. Rainfall is torrential in nature, i.e. high intensity and short duration, and is monsoonal. Annual precipitation ranges from 300mm to greater than 800mm in the Central Highlands and Southwestern Lowlands, 200mm to 300mm in the Northwestern Lowlands, and 100 to 200mm in the Eastern and Southeastern Lowlands. Situated close to the Red Sea escarpment, the Alla-Ghadien area receives intermittent orographic rainfall, as well as that during the two monsoonal periods in March–April and June–September. 3 PREVIOUS WORK Drury and Berhe (1993) reported significant regional geological controls for groundwater in Eritrea and their expression on satellite images. According to them, the main potential for groundwater developments occur as fracture systems, carbonates with enhanced permeability, granites (which develop deep, coarse and porous soils and extensive joint

Water resources of arid areas

78

systems), fissile rocks in shear zones, deep basins of unconsolidated sediments in the mountains and coastal plains, igneous intrusion which act as natural barriers, and outcrops of lava and laterite. Many such features have been identified on images of Eritrea. Some of them were considered to be targets for locating bore wells in difficult terrains of northwestern Eritrea, and drilling them met with a success rate of 85%. Drury et al. (2001), focusing on the hydrogeological potential of fracture systems, indicated that NNW-SSE Precambrian shear zones, normal faults roughly parallel to those earlier structures and prominent East-South-East-West-North-West dilatational fractures offer considerable scope for groundwater development. Prospect geophysical profiling across several of these structures in both lowland and highland terrains revealed conductive features believed to relate to saturated zones in large, regionally extensive fractures. Zerai and Solomon (1993) identified five main hydrogeological units: (i) Unconsolidated sediments with variable inter granular permeability. (ii) Volcanic rocks (basalts) with fracture and fissure permeability. (iii) Fissured and karstic carbonate aquifers. (iv) Metamorphic and intrusive rocks with localized low to moderate permeability along fractured and weathered zones. (v) Aquitards and aquicludes and groundwater barriers (acid to intermediate volcanic).

4 GEOLOGY Knowledge and understanding of the geological events of an area are important for groundwater investigations. The geology of Eritrea is made up of a Precambrian basement complex comprising high-to low-grade metamorphic rocks and associated intrusives, which are overlain by predominantly Mesozoic sedimentary rocks and Tertiary to Quaternary volcanic and sedimentary rocks (JICA Report 1998 Drury and Berhe 1993). Precambrian granites are exposed in the eastern, western and northern parts of Alla-Ghadien. The southern part of the study area is a wide and flat plain between actively rising ridges of metamorphic rocks. The geological events in Eritrea are summarized as follows: (1) Precambrian: Formation of the crystalline basement complex and its associated intrusive rocks; (2) Paleozoic: Peneplanation of the basement complex and deposition of sparse sedimentary rocks; (3) Jurassic: Transgression-regression of the Mesozoic sea, which deposited lower sandstone, Adigrat sandstone and Antalo limestone during subsidence towards the Indian Ocean; (4) Upper Eocene-Miocene: Uplift forming domes, extensional fault systems, basaltic flood volcanism and opening of the Red Sea rift system; (5) Miocene Period: Formation of upper sandstone; and (6) Quaternary Period: Formation of alluvial, eluvial and colluvial unconsolidated sediments.

Groundwater development—identification of artificial recharge areas in Alla, Eritrea

79

5 HYDROGEOLOGY The reconnaissance survey done in Ghadien and Alla areas shows that the plain is covered by Quaternary alluvial and colluvial sediments and with subordinate outcrops of granitic intrusive rocks. There are two types of aquifers in the study area: alluvial sediments and weathered and fractured acidic granitoids. Moreover major and minor lineaments exist in the area, many being fracture zones or igneous dykes, which naturally serve as conduits for or barriers to groundwater flow respectively. That the NNW-SSE basaltic dikes serve as groundwater flow barriers is confirmed by high yields of wells around the dikes. Minor structures trending roughly NE-SW are found to serve as a conduit for the water flow, confirmed by the high yield of the wells in the Ghadien area, where the structures are dominantly observed. In 1994 and 2000 the Water Resources Department (WRD) did a well inventory in Alla-Ghadien area, encompassing location, static water levels and well design. Based on the well inventory data in 1994 there were 144 wells, the average well depth was 15m+ with static water level (SWL) 14.4m. In the year 2000 the average well depth was 17 meters & SWL 16.51. Therefore the lowering of water level in 4 years is 2 meters. In the Alla-Ghadien area, hand dug wells are very closely located, from approximately 20m to 100m apart. An ad-hoc assessment of the catchments of 1465km2 area based on rainfall for Alla has shown that the total groundwater recharge is 643Ha.m, while the total annual groundwater draft is 1062Ha.m leaving a negative balance of −419Ha.m (K.S.Viswanatham 2002). The above figure indicates “mining” of water. The Alla-Ghadien area could therefore be classified as an overexploited area. Areas where groundwater resource assessment shows stage of groundwater development more than 100% and both pre and post monsoon groundwater levels show a significant longterm decline are classified as over exploited areas. 6 METHODOLOGY An integrated approach that involves interpretation of remotely sensed data and groundbased ancillary investigations has been implemented for the Alla-Ghadien study. Groundwater zonation was prepared using various thematic maps at 1:50,000 scales, which include geology, geomorphology, lineament trends and land use/landcover. Field survey was done to correlate the image characteristic to ground feature to confirm the interpretation. Specific field data, such as well inventories of SWL, yield, depth, diameter, quality of water and drilling logs were collected. Images (bounded by UTM 7500000–7516000 E 1666000–1682000 N) of the Alla-Dekamhare area were from an EO-1 overpass in April 2001 and combine ALI bands 5, 4 and 3 as red, green and blue components, sharpened by the use of the ALI panchromatic band to modulate intensity. This combination is optimum for expressing vegetation cover, but does discriminate some lithologies and soil types, as well as revealing small-scale topographic features.

Water resources of arid areas

80

7 SIGNIFICANT PARAMETERS OF GROUNDWATER OCCURRENCE AS REVEALED BY IMAGE INTERPRETATION The various water-bearing and movement properties, and the controlling parameters in the study area based on image interpretation and ground check are as follows: Fluvial/alluvial sediments and alluvial fans, Tertiary laterite, Granite, Tertiary basaltic rocks, as lavas and dykes.

Geologically, the area is dominated by granitic terrain. The alluvial areas of Alla and Ghadien are eastward sloping plains surrounded by hills with mostly steep slopes. The alluvial sediments have a yellowish colour and define a roughly triangular area. Tertiary laterites are indicated by grayish white color in irregular shapes along the streams and on the plains. Areas of bedrock comprise granites, which are traversed by NNE to SSW basaltic dikes. In the imagery, the granites are bluish to dark blue in color and occupy topographical ridges. The granitic exposures form circular and semi-circular shapes, for example Bazit Hill near Bazit village. There is a possibility of marble being present as roof pendants in the granitic masses at the northern flank of the area (represented by light to dark gray tones with signs of bedding), which have to be checked and confirmed in the field. Topographically the area is surrounded by hills with undulating slopes, and almost plain on the center. These features could be observed north west of Ghadien. Dikes show as linear features mostly concentrated north of Ghadien and roughly north of the Alla plains. The major land use of the area is classified based on color, shape and texture as fallow land, barren rocky terrain, dense vegetation (mainly horticulture gardens), river sediments and sparse vegetation. The Ghadien, Sesah and Bazit Rivers drain the area. The drainage pattern is distributed and dendritic. All the rivers flow eastwards. Some flow features follow a structurally controlled direction. Low drainage density in the alluvial plains, which probably indicates high rates of infiltration, suggests good groundwater prospects in parts of them. 8 FIELD CHECKS From the fieldobservations, the area can be divided into high- and low-potential zones. In the upper Ghadien river successful hand dug wells are being pumped for 4–7 hours per day and irrigating 6–10ha of horticulture gardens, particularly citrus bushes. In addition, this high-potential area has been investigated by the geophysicist from the Ministry of Agriculture, and recommended to be a potential area for development. In the satellite image alluvial fans and fills clearly represent the upper Ghadien zone.

Groundwater development—identification of artificial recharge areas in Alla, Eritrea

81

The low-potential zones, as defined by the well inventory data, are represented in the satellite image by granitoid and granitic ridge with adjoining alluvial cover. The farmers are pumping water for 2–3 hours per day from wells there, some of which have been abandoned due to poor yields or non-availability of water. The geophysical surveys have also confirmed the absence of suitable aquifers at deeper levels in such low potential zones. A preliminary interpretation on the regional geology was attempted, based on the imagery. In order to corroborate this interpretation, a quick field trip was made. The field observations revealed additional information on the geology, which necessitated some modification to the preliminary assumptions about the geological set-up. The alluvial fans as interpreted from the imagery are observed in the field to be a thin cover of alluvium on granitic rocks. The possibility of carbonate rocks/Marble in the North West part of the area is ruled out as granites varying in color, composition and texture represent these outcrops. Only few small patches of marble, which are not mapable, have been reported, which could be the extension of carbonates that supply abundant spring water in the Maihabar area 15km to the north. Beside this, cherts and conglomerates are observed on the Northwest part (assigned as basement metasediments), diorite on the Northeast, and metabasalt on the Southeast part of the study area (Fig. 2). In the Alla area, four bore wells drilled in the range of 47 to 50 meters are reported to yield 1.4 to 2.2 liters per second. The depths of the alluvial cover at those sites are from 13 to 21 meters, followed by granites. In the Bazit area, there are 3 boreholes, each having depths of 50 meters, and yields of 1.4lps, 2.5lps and 5Lps. In the Ghadien area,

Figure 2. Geological map of AllaGhadien.

Water resources of arid areas

82

there is only borehole with a depth of 50 meters and yield of 2.5Lps. The alluvial cover is 21 meters. In Chuhot area 3 boreholes, with depths of 36 meters, 48 meters and 50 meters respectively proved to be highly successful, with yields of 4Lps, 5Lps and 5.5Lps. 9 TARGET AREAS FOR GROUNDWATER DEVELOPMENT (1) The areas around Alla (UTM 504000–508000 E and 1673000–168000) are favourable for groundwater exploitation. (2) The NNE-SSW dikes on the North flank of Alla plains, which can be, extrapolated beneath the alluvial cover show a positive indication for potential groundwater development as they form natural sub-surface barriers. (3) The source of water is mainly from the alluvial formation whose origin is by weathering of granites.

10 ARTIFICIAL RECHARGE SITES The alluvial formations in Alla plain provide suitable sites for sub-surface dams, infiltration galleries and check dams. This is due to fine sediments predominantly quartz gravels and sand derived from granitic terrain. The natural barriers namely the NNESSW dikes have to be taken into consideration while constructing the sub-surface dams and check dams for the suitability of the structures. 11 RECHARGE AND DISCHARGE AREA Areas can be delineated into recharge and discharge areas depending on whether water is added to or abstracted from the zone of saturation. In the case of the water table aquifer, usually the areas occupying higher elevations with deeper water tables constitute the recharge areas while the

Groundwater development—identification of artificial recharge areas in Alla, Eritrea

83

Figure 3. Water table contour map of Alla-Ghadien area. topographic lows with shallow water tables comprises natural discharge areas (Karanth, 1994). Therefore in the case of Alla-Ghadien area, Ghadien is a useful recharge area due to its higher elevation and deeper water table than Alla. This is shown by a water table contour map (Fig. 3) where the ground water flow is NNW to SSE of the area. To confirm this, further investigations are necessary. Implementing a program of artificial recharge and abstraction requires the construction of check dams and sub-surface dams at finalized sites. 12 CONCLUSION & RECOMMENDATIONS (1) Construction of check dams specifically in the Ghadien River because of high slopes with fractured granitic rocks and catchment area is recommended. (2) The sub-surface dam near Bazit village needs to be further probed by integrated geophysical investigations. (3) Bore wells drilled in the Chuhot area gave reasonably good yields (5ls−1), which if used through drip irrigation could cover large areas. Integrated groundwater studies are to be taken up for further bore well locations.

Water resources of arid areas

84

ACKNOWLEDGEMENTS The authors acknowledge with thanks Mr. Ghebremichael Hagos Director General for giving an opportunity for the investigations and permitting to publish this paper. The authors also thank Mr. Ghebremichael Temenewo and Mr. Michael Negash for discussions from time to time, which helped to improve the paper. Ms. Meron Teshome is acknowledged for her support in preparing the maps. REFERENCES Drury, S.A and Berhe, S.M. 1993. Remote Sensing and Water Exploration in Eritrea WRD Eritrea/EIAC/ GREADCO and Open University UK. (Unpubl) 6 pp. Drury, S.A and Berhe S.M. 1993. Accretion Tectonics in Northern Eritrea revealed by remotely sensed imagery. Geol Mag, 130(2):177–190. Drury, S.A, R.J.Peart M.E. & Andrews Deller. 2001. Hydro geological potentials of major fractures in Eritrea. Journal of African Earth Sciences, 2(2):163–177. Habteab.T. 2000. Groundwater Depletion in Alla Commercial Farm. Department of Environment, Ministry of Land Water and Environment (Unpubl). Karanth, K.R. 1994. Groundwater Assessment, Development and Management, New Delhi, Tata McgrawHill Publishing Company Limited: 720 pp. JICA Report. 1998. Study on Groundwater Development and Water Supply for Seven Towns in Southern region of Eritrea.Water Resources Department and Sanyu Consultants Inc., Japan. NASA, 2001. EO-1Science Validation Team Home Page, http://eol.gsfc.nasa.gov/science/SVTAuth.cfm Viswanatham, K.S. 2002. Water Resources Development Management of Critical Areas in Eritrea. Journal of Applied Hydrology, XV(4), Oct:21–25. Zerai Habteab. 1996. Groundwater and Geothermal Resources of Eritrea with the emphasis on their chemical quality: Journal of African Earth Sciences, 22:415–421.

Subterraneous injection of nutrient rich groundwater to the coastal waters K.K.Balachandran National Institute of Oceanography, Regional Center, Cochin Joseph Sebasgtian Paimpillil Center for Earth Research & Environment Management, Cochin, India Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Existence of a subterranean flow connecting a tropical backwater to coastal waters through submerged porous lime shell is inferred from nutrient distribution patterns in the coastal waters. The activation of ‘mud banks’ triggers high productivity in oligotrophic coastal waters of southeastern Arabian Sea. The current results represented a period when the mud banks were not activated but fertilization at certain compartments of the coastal zone by injection of nutrients by hitherto unknown processes was evident. The higher dissolved concentrations of ammonia, nitrate and silicate originating from shallow depths and extending to offshore have indicated a clear groundwater based nutrient source. The enriched particulate organic carbon and Chlorophyll a were also notable features of the nutrient injection region. It is difficult to point out a definite source for the high nutrient introduction as fresh water discharge was at the minimum during the nutrient injection duration. A band of N/P>15 funneling out during non-mud bank period gave a clear indication of an ‘external ground water source’ of nitrogenous compounds to the coastal water which deserve identification as it is traced to a region far away from any river mouth and the injection of nutrients was observed during non-monsoon months when mud banks were passive. The existence of subterraneous channels as the artifacts of porous nature of the lime shell base of the region transporting the nitrogenous compounds cannot be ruled out in the region.

1 INTRODUCTION The west coast of India is environmentally more sensitive than the east coast primarily because it is bordering one of the most sensitive ecosystems in the world, the Arabian Sea. The environmental property of the northern Arabian Sea is unique which manifests in rich biological production throughout the year through different processes and thus,

Water resources of arid areas

86

explain for the Arabian Sea ‘Paradox’ Mathupratap et al (1996). The symptoms are there to show considerable impact of deterioration of estuarine waters on the coastal ecosystem Nair et al (1991), Naqvi et al (2000), Jayakumar et al (2001). The emerging industrial establishments and human settlements along the west coast of India, thus necessitates a critical evaluation of the nature and quantum of inputs to the Arabian sea as well as their regional assimilative capacities. If there is a possible threat to the well being of the living resources of EEZ of India, then the coastal waters of southwest coast of India, and in particular, Cochin region is the prime location prone to trigger it. The booming city of Cochin has population of nearly 1.5 million Anonymous (1998) and 60% of the chemical industries of Kerala are situated in this area Cochin backwaters are the largest of its kind on the west coast of India with an area of 256Km2. The 16 major and several minor industries situated in the upstream region of the backwaters discharge nearly 0.105Mm3d−1 of effluents Anonymous (1996). The fertilizer consumption in Kuttanad region (the main agricultural field draining to Cochin backwater) alone is reported to be 20,239ty−1 Anonymous (1996). The backwater

Figure 1. Map showing the study area (A) and study region with location of stations and bathymetry (B). receives organic wastes ~260td−1 Anonymous (1998) and an annual dredge spoil from the harbor area to the tune of 107m3. Conventional understanding of coastal waters of southeastern Arabian Sea is that activation of mud banks by monsoon forcing triggers intense geochemical processes leading to high productivity. Mud banks, as they appear only during monsoon and disappear with its retrieval, are unique in their formation and functions, and have turned out to be economically important for its rich biological resources. As far as the chemical features are concerned, the general picture so far emerged out is that except during the

Subterraneous injection of nutrient rich groundwater to the coastal waters

87

monsoon periods, the southwest coastal waters remained oligotrophic and surface chlorophyll a typically ranges from 0.1 to 5.3mgm−3, while primary productivity ranges from 100 to 360mgCm−2d−1. Recent studies as the one discussed here contradict these findings and show that even after the monsoon period, fresh injection of nutrients by hitherto unknown processes fertilize the coastal waters that are either permanent or quasipermanent in nature. One of the major mudbank regions (Fig. 1 A, B) of southwest coast of India was selected for observation that indicates episodic introduction of nutrients into the coastal waters during periods when mud banks are passive. 2 RESULTS AND DISCUSSION During the typical pre-monsoon (February) months, the nitrogenous nutrients in water remained low except for the southern transects centered on Chethi and Alleppey. The phosphate concentrations did not show any spatial or vertical variation in the water column, but higher concentrations of ammonia, nitrate and silicate were observed at selected regions starting in the near shore regions and extending offshore (Fig. 2 A–D). The Nitrate-N concentrations point towards a clear source between Chethi and Pazhayangadi, where it peaked up to >8µM and decreased towards offshore. A similar trend was observed for ammonia-N with the source centered on Chethi (at about 15m depth). It may be assumed that the ammonia released were either rapidly utilized by phytoplankton or oxidized within the system itself where the waters were saturated with dissolved oxygen. Distribution of silicate-Si was similar to that of nitrate (4–10µM), higher than the corresponding values reported for the waters of Southeastern Arabian Sea. The input of these nutrients supported high primary production up to 14mg/m3 of chlorophyll a (peak column production of

Water resources of arid areas

88

Figure 2. Distribution of ammonia-N (A), nitrite-N (B), nitrate-N (C) phosphate-P (D) at the surface and bottom during October (a, b), February (c, d) & November (e, f). 1529mgCm2d−1), approximately 3 times greater than the peak values reported so far from these waters Qasim et al (1978). The peaks in chlorophyll a and ammonia showed a preference of ammonia among the nutrients for primary production. It is difficult to point

Subterraneous injection of nutrient rich groundwater to the coastal waters

89

out a definite source to these high nutrients during this period, as the fresh water discharge was at the minimum.

Figure 3. N/P peak values funneling out from mud bank region.

Water resources of arid areas

90

During post monsoon (November), homogenous mixed layer prevailed in the entire region. While the physical characteristics were more or less stable, there was considerable variability in the nutrients and in chlorophyll a concentration (Fig.2 A–D). A marked decrease in sub-surface dissolved oxygen (2.8–4.8ml/l) was the characteristic feature of this period, which was concomitant with enriched nitrite (0.5–2.0µM), phosphate (0.4–2.8µM) and silicate (0.5–14µM). The ammonia (1–7µM) and nitrate (1– 6µM) were also elevated at some regions along southern transects. The enriched particulate organic carbon (>3.5mg/l) and Chlorophyll a (14.8mg/m3) were also the notable features of this period. It is likely that chlorophyll a values were proportionate to carbon production indicating a strong positive relationship binding it with nutrient related factors rather than seasonal or diurnal fluctuation. The elevated nitrite and phosphate levels around Cochin may be due to the input from the backwaters. Higher values of nitrite, POC and chlorophyll a towards the southern offshore waters off Pallana were conspicuous and the regions with high nitrite had nitrate levels up to 6µM and the low levels of ammonia had ruled out the nitrification as a significant process responsible for nitrite accumulation. The remarkable co-existence of nitrite with nitrate strongly suggested that the nitrite production should mostly be due to assimilatory reduction. This was further substantiated by the high concentration of chlorophyll a (4–9.8mg/m3) on these transects. The N/P ratio in the coastal waters was below 15 during November (Fig. 3), possibly due to the disproportionate release of P from mudbank sediment. However, a band of N/P>15 funneling out from Alleppey region was indicative of an ‘external source’ of nitrogenous compounds into the coastal waters. A comparison of long-term (decadal) trend in the chlorophyll data of this region showed “greening” of near shore waters Devassy (1983). This suggests that phytoplankton standing crops had increased historically, possibly in response to watershed nutrient inputs. These sources of nutrients deserve identification as it was traced to a region, far away from any river mouths. The current observations in general indicated the presence of a nutrient source between Chethi and Pallana. This region has mud banks but the release of nitrogenous compounds cannot be accounted from sediments. The injection of nutrients was in nonmonsoon months when mud banks were passive and a new influence of Vembanad Lake on the coastal waters is very clear. One of the recent estimate shows that in spite of receiving 42.4×103mold−1 of inorganic phosphate and 37.6×103mold−1of inorganic nitrate from Periyar side of the estuary, the export to the coastal waters is only 28.2×103mold−1of inorganic phosphate and 24×103mold−1 of inorganic nitrate Naik (2000) and the lake acts as a sink for the nutrients, flushing out only a portion of the pollution load that it receives. Increased human population along the coastal belt has also resulted in concomitant increases in widespread use of septic tanks and nutrient inputs to coastal waters, particularly from regions occupying limestone beds. It has been found that domestic wastewater from septic tanks provide more nitrogen than that due to precipitation or use of fertilizers. The situation is exacerbated in the present study region, as more than 70% of households in these coastal belt and adjacent areas of Vembanad Lake do not have proper sanitation facilities. Significant amounts of nutrients from fertilizer applied in agricultural fields (approx. 94kg/ha) leach out into waterways, groundwater and to the coastal bays inducing coastal fertilization due to direct discharge into coastal ocean and through ground water seepage.

Subterraneous injection of nutrient rich groundwater to the coastal waters

91

3 CONCLUSIONS The nutrient fluxes into coastal region were influenced by fluxes from Cochin backwater and by the mud bank formation. The present study isolates a possible link between Vembanad Lake that supplies primary nutrients to the adjacent coastal waters and precondition it for rich primary production during non-monsoon months. The causative factors discussed are indicative of existence of a subterranean flow connecting Vembanad Lake to the adjacent coastal waters through the submerged porous lime shell beds. Continuous nutrient entry through such process is bound to upset coastal water productivity pattern. If the existence of the subterraneous channels linking Vembanad Lake to the adjacent coast is proved, it might even re-construct the historical evidence that the subterraneous flow plays a decisive role in the formation of mud banks along this region. A sub aqueous injection of nutrients into the coastal waters through this region is possible even after the rainy season. This assumption need further study to establish cause and affect mechanisms and quantify actual trends created by increased nutrient loading. REFERENCES Anonymous, 1996. Pollution potential of industries in coastal areas of India. Coastal Pollution Control Series: central Pollution Control Board Report. COPOCS/9/1995–96. Anonymous. 1998. NEERI- carrying capacity based developmental planning of Greater Kochi Region. Phase I Report. Devassy, V.P. 1983. Mahasagar, Bull Bull Nat. Inst. Oceanogr.7:101–105. Hema Naik, 2000. Budgets for Periyar estuary, Kerala. Presented at Regional Training Workshop on Biogeochemical Budgeting and Socio-Economic modeling for Coastal Scientist. APN/SASCOM/LOICZ, 18–22 September, Colombo. Jayakumar D.A., Naqvi S.W.A., Narvekar P.V. & George M.D. 2001. Methane in coastal and offshore waters of the Arabian Sea. Mar. Chem. 74:1–13. Mathupratap N.M., Prasanakumar S., Bhattathri P.M.A, Dileepkumar M., Reghukumar S., Nair K.K.C. & Ramaiah N. 1996. Mechanism of the biological response to winter cooling in the north eastern Arabian Sea. Nature, 384:549–551. Nair C.K., Balchand A.N. & Nambisan N.P.K. 1991. Heavy metal speciation in sediments of Cochin estuary determined using chemical extraction techniques. Sci.Total Environ. 102:113– 128. Naqvi S.W.A., Jayakumar D.A., Narvekar P.V., Naik H., Sarma V.V.S., D’Souza W., Joseph S. & George M.D. 2000. Increased marine production of N2O due to intensifying anoxia on the Indian continental shelf. Nature, 408:346–349. Qasim, S.Z., Wafar, M.V.M., Sumithra Vijayaraghavan, Joseph P., Royan. & Krishna Kumari, L. 1978. Ind. J. Mar. Sci.,7:84–93.

A new method for the estimation of episodic recharge J.Bean, G.van Tonder & I.Dennis Institute for Groundwater Studies, University of the Free State, Bloemfontein, South Africa Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: A new stable isotope-based technique, the Modified Amount Effect (MAE) Method, was developed during this study. This technique provides insight into episodic recharge processes by estimating the proportion of preferential pathway-to-matrix-derived flow entering an aquifer, and the amount of rainfall required to initiate recharge via the respective flow paths. Significantly, the proportion of bypass flow can be determined without undertaking expensive and time-consuming unsaturated zone studies, both factors often of primary concern when undertaking recharge investigations in developing countries.

1 INTRODUCTION There can be no doubt that the South African water industry has been profoundly transformed over the last 10 years, with millions of rands invested in water infrastructure aimed at ensuring that every South African has access to fresh drinking water. In drier, more isolated, inland areas of the country, this has often meant that available groundwater resources must be exploited. As such, government and non-government organisations have invested in research associated with developing new assessment techniques so that these resources can be managed sustainably. In common with all these strategies is the need for recharge processes to be understood, and if possible, quantified. An understanding of site recharge behaviour is far more important than many geohydrologists realise, and goes beyond estimating the average proportion of rainfall entering a given aquifer. For example, from a planning viewpoint, groundwater ingress into a mine is seldom problematic to mine management, providing it is constant; problems occur when unpredicted increases occur, such as those associated with the sudden entry of recharge water into surrounding aquifers. Thus, through understanding the episodic nature of recharge in semi-arid and arid areas, and therefore the thresholds that must be exceeded before recharge occurs, geohydrologists are better able to provide predictive advice for their clients. This paper discusses a new stable isotope-based

A new method for the estimation of episodic recharge

93

technique, the Modified Amount Effect (MAE) Method. This was developed during the study, which provides insight into episodic recharge processes. 2 THE MODIFIED AMOUNT EFFECT METHOD 2.1 General This technique provides insight into episodic recharge processes by estimating the proportion of preferential pathway-to-matrix-derived flow entering an aquifer, and the amount of rainfall required to initiate recharge via the respective flow paths. Significantly, the proportion of bypass flow can be determined without undertaking expensive and time-consuming unsaturated zone studies, both factors often of primary concern when undertaking recharge investigations in developing countries. 2.2 The Methodology Recharge water is progressively enriched in δ18O due to evaporation, which occurs with movement through the unsaturated zone. The slope of the evaporated water line for matrix water in the unsaturated zone (EWL-U) is generally around 5, but sometimes as low as 2. The variable d was corrected to represent evaporation within the unsaturated zone by constructing a line through the 2H and 18O average for those samples representing background recharge (i.e. δ18O<−4.3‰), resulting in an EWL-U of δ2 H=2.5δ18O+ −21.05. Laboratory studies undertaken by Allison et al. (1984) confirm that if site recharge is constant, evaporation-induced enrichment within the unsaturated zone should also be constant, resulting in a line parallel to the local meteoric water line (LMWL), herein referred to as the Matrix Water Line (MWL-U). Thus, if exchange processes between aquifer materials and groundwater are ignored, preferential recharge areas can be inferred in cases where the LMWL and MWL-U are not parallel (i.e. the LMWL has a steeper slope and greater d-excess than the MWL-U). A similar assumption can be made if the line-of-best-fit through site groundwater data (GWL) is not parallel to the LMWL, as the isotopic characteristics of water stored in the aquifer represent a long-term average of recharge processes. An approximate estimate of the contribution of preferred pathwaysderived recharge to aquifer storage can therefore be determined by constructing lines parallel to MWL-U, GWL, and EWL-U, through the average 2H and 18O composition of groundwater derived from direct recharge i.e.:

Where PPflow=the proportion of recharge derived from preferential flow, and dGWL, dEWL2 U, and dMWL-U represent the d excess in δ H (‰) for GWL, EWL-U, and MWL-U, respectively. It should be appreciated that the calculated value PPflow is sensitive to: ● Variations in the orientation of the LMWL and EWL-U.

Water resources of arid areas

94

● The recharge threshold. At low recharge thresholds (i.e. recharge occurs rapidly in most years), particularly in more temperate areas, evaporation effects may not be represented in the stable isotopic composition of groundwater data. In these areas, transpiration, and not evaporation, probably has a greater potential to reduce the recharge flux to site aquifers. ● The source of recharge water. The suggested method assumes that recharge water is derived solely from precipitation, with no contribution from an adjacent surface water body where pre-recharge evaporation has occurred. 2.3 Adapting the MAE Method 2.3.1 Background information Groundwater samples were taken from the vicinity of Liebenberg’s Pan near Petrusburg, Free State Province, South Africa. Pans occur throughout the Western Free State, generally following the strike of the Ecca Series, which forms part of the Carboniferous to early Jurassic aged Karoo Basin sediments. The Ecca Series at the pan is comprised of mudstone, sandstone, and shale interbeds. Cretaceous-aged dolerite dykes and sills have intruded these sediments, with high yielding aquifers (>10L/s) often occurring at the structure/sediment interface. These have been locally overlain by calcretes to a maximum depth of about 15m. Petrusburg has a semi-arid climate, with an evaporation excess of 1920mm (2380– 460mm MAP) annually. Given that the water table is generally less than a metre below the pan surface, groundwater here is exposed to continuous evaporation in most years, the exception being those years where sufficient rainfall occurs to flood the entire pan for a few months of the wet season. Thus, the slope of EWL-U in these areas should be parallel to GWL because, while preferential pathway-derived water may not be as evaporated as matrix-derived during recharge events, it will eventually be evaporated to the same degree after entering the aquifer. Liebenberg’s Pan-derived brine with chloride concentrations in excess of 100000mg/L is further concentrated in evaporation ponds that have been constructed on site as part of a commercial salt-extraction enterprise operated by a local farmer. The pan itself is the lowest topographical feature in the landscape. Land use varies with soil type, topography, and access to irrigation water, with grazing and dairy farming predominant to the north and west of the pan in the steeper dolerite hills that occur there, and irrigated cropland located to the south and east on deeper soiled, gently sloping ground. 2.3.2 Calculating recharge using MAE The orientation of the EWL determined from brine samples was determined to be δ2H= 3.65 δ18O+−4.71 (refer to Figure 1). Another characteristic of Petrusburg data of interest is that evaporation-induced enrichment has not been excessive as would be expected in this type of environment, with all groundwater samples having a δ18O concentration 3.3‰ or less. This suggests that brine has mixed with isotopically depleted water from another source, the most likely being groundwater from upslope areas, a finding

A new method for the estimation of episodic recharge

95

supported by the occurrence of freshwater springs at various locations around the perimeter of the pan, and the observed decrease in brine concentrations in production bores over time. PPflow estimated using the MAE Method is between 33 and 25% assuming an EWL-U slope of 3.65.δ18O and 2.5.δ18O, respectively (refer to Figure 2). However, given the potential for brine/ fresh groundwater mixing, the lower figure would be more acceptable in this instance. Further insight into site recharge processes can be obtained when variations observed in recharge threshold estimates obtained using the cumulative rainfall departure method (CRD) and MAE techniques are considered. On the basis of 98 years of rainfall data for Petrusburg, the average

Figure 1. Stable isotope characteristics of groundwater samples taken in the vicinity of Liebenberg’s Pan, Petrusburg. “GW”, “PW”, and “Ave” denote groundwater samples taken from boreholes surrounding the pan, in the pan, and the background isotopic average, respectively.

Water resources of arid areas

96

Figure 2. Line characteristics used to determine PPflow at Liebenberg’s Pan. monthly rainfall is 35.7mm; this value also representing the long-term recharge threshold for an aquifer in equilibrium if seasonal conditions are ignored. In theory therefore, there would be no change in water levels if 35.7mm of rain fell at the site every month. Under field conditions however, this does not occur; prolonged periods of below average rainfall are evident throughout the Petrusburg dataset. Thus, in order to restore equilibrium conditions such that the average recharge threshold again decreases to 35.7mm/month, a given catchment must receive above-average rainfall. This observation is significant because it indicates that, for a given aquifer in a semi-arid and arid area, multiple recharge thresholds will be represented in site water level data. Multiple recharge thresholds that are likely to be of importance include those necessary to induce recharge via: 1. Preferential pathways after a period of below-average rainfall; 2. The matrix after a period of below-average rainfall; 3. Preferential pathways once aquifer equilibrium has been restored; 4. The matrix once aquifer equilibrium has been restored. Each of these recharge thresholds can be approximated using available site stable isotope data by applying the mass balance equation: RTave.δ18ORT−ave={RTlow.(X.δ18ORT−low)}+{RThigh.(X−1)δ18ORT−high} Where, RT=Average recharge threshold expressed as an equivalent rainfall depth (mm); RTlow=Average recharge threshold to be exceeded if recharge via preferential pathways is to occur (mm); RThigh=Average recharge threshold to be exceeded if recharge via the matrix is to occur (mm) and δ18ORT-low=Average δ18O concentration of preferential pathway-derived recharge water (‰); δ18Orw-high=Average δ18O concentration of matrixderived recharge water (‰); and, X=Preferential pathway to matrix proportioning factor. The average thresholds to be exceeded before recharge occurs via the matrix, and preferential pathways. Note that these values represent long-term averages, and not the upper and lower limits of recharge thresholds. These limit thresholds can be calculated,

A new method for the estimation of episodic recharge

97

however, by considering CRD and long-term average values together. For example, the CRD Method indicates that, for an aquifer under equilibrium conditions, the recharge threshold is approximately 35mm/month. Since, on average, the recharge threshold cannot be lower than this amount, it must represent the average lower recharge threshold. Thus, the respective average lower recharge thresholds can be calculated once the isotopic composition of rainfall for an equivalent depth of 35mm has been estimated from amount effect data. Once the lower and average long-term thresholds for both preferential pathway (RTlowand RTave-pp) and matrix-medium recharge (RTlow-uzm and RTave-uzm), the upper recharge pp thresholds RThigh-pp and RThigh-uzm can also be calculated, i.e. RThigh=2.RTave−RTlow Only 25% of recharge at Petrusburg occurs via preferential pathways. On average, recharge occurs via these pathways in more than 50% of all rainfall events (RTave−pp=56.4%). Therefore in episodic recharge environments, resource managers must ensure that allocated water can be used for the entire period between major recharge events, which where recharge via the matrix predominates, can be significant. Indeed, in many instances it may be more realistic to base groundwater allocations on the proportion of bypass flow-derived recharge entering site aquifers initially, the allocations increasing once aquifer storage, recharge threshold, and recharge event return period characteristics are better understood. 3 CONCLUSIONS Four recharge thresholds can be identified using the MAE Method; the low and high recharge thresholds that must be exceeded before recharge occurs via preferential pathways or the matrix, respectively. These represent threshold limits, the low value only of importance following successive months of wet weather, the high value representing the rainfall that must be received to restore an aquifer system to equilibrium after prolonged dry spells. Once these thresholds are known, the recharge history of a site can be modelled using available rainfall data by adapting the CRD Method. An important finding of modelling undertaken during this investigation is that in those semi-arid to arid areas where most recharge water enters, the aquifer via the matrix, the period of time that elapses between successive rainfall events that exceed the matrix recharge threshold often extends to scores of years. This has significant resource management implications for much of the region, as it indicates that the current approach of basing allocations on average recharge estimates is only justified if sufficient groundwater is available for use over the entire period between recharge events. The MAE Method was found to be sensitive to the recharge history of the site, the returned recharge estimate significantly higher when calculated immediately after recharge via the matrix had occurred. This is not to say that these estimates were incorrect (indeed they were representative of site recharge processes at the time of sampling), but that rainfall in the preceding months should be considered prior to sampling. In general however, sampling should be undertaken near the end of the dry season, which in the summer-dominant rainfall areas of Southern Africa is between

Water resources of arid areas

98

September and November (allowing for a 30 to 60 days lag time between rainfall and subsequent recharge). REFERENCES Alison, G.B., Barnes, C.J., Hughes, M.W. & Leaney, F.W.J. 1984. Effect of climate and vegetation on oxygen-18 and deuterium profiles in soils. Isotope Hydrology 1983. IAEA Symposium 270, September 1983, Vienna.

Prioritisation of the impacts of pollutants on groundwater flow systems in South Africa I.Dennis, B.Usher & J.Pretorius Institute for Groundwater Studies, University of the Free State, Bloemfontein, South Africa Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Groundwater pollution can occur, as a result of various activities of man. With increased human settlement and economic development, a range of undesirable waste products are produced which can end up in the environment. If these waste products are not well handled, they can cause pollution of groundwater. The threat caused by undesirable substances on groundwater is recognized in South Africa and measures have been put in place through legislation to protect groundwater from pollution. Although groundwater pollution incidences have been reported countrywide, we do not have an indication of the extent of the problem. The results of the investigation discussed in this paper are therefore geared towards filling the gap in the understanding of groundwater pollution in South Africa’s urban environments. By doing so the principal pollutants can be identified and based on their risk prioritised. This will facilitate better management of groundwater quality through the country.

1 INTRODUCTION Groundwater pollution can occur, as a result of various activities of man. With increased human settlement and economic development, a range of undesirable waste products are produced which can end up in the environment. According to the National Water Act (Act No. 36, 1998), pollution is defined as the direct or indirect alteration of the physical, chemical or biological properties of a water resource so as to make it— 1. Less fit for any beneficial purpose for which it may reasonably be expected to be used; or 2. Harmful or potentially harmful— ● to the welfare, health or safety of human beings;

Water resources of arid areas

100

● to any aquatic or non-aquatic organisms; ● to the resource quality; or ● to property. The main aim of the investigation discussed in this can therefore be summarized as the prioritization of the type of pollutants and their associated sources which present a threat to groundwater, the environment and health in South Africa’s urban catchments. 2 METHODOLOGY 2.1 Factors taken into account The methodology followed evaluated the sources and contaminants separately. The results of the evaluation were then combined to determine a final risk based prioritization. 2.1.1 Sources Sources, in this context, refer to the origin of the substances (inorganic species, organic compounds or microbial agents) that are causing, or may potentially cause, the pollution. The term

Table 1. Systems for classification of groundwater contamination sources. Classification Examples system based on Way of release Loading history Location Degree of localization Origin Likelihood of occurrence

Discharge sources, transport sources Spill or continuous Above ground surface, below surface Point (or line) and non-point sources Industrial sources, mining sources For example petrol service stations found more often than chemical manufacturing plants

is used very broadly over a range of scales and may describe physical entities (e.g. a pond, a tank, a pipeline); human activities (e.g. mining, irrigation, wastewater treatment); the site at which potential pollutants are stored, used or disposed (e.g. wastewater treatment works, cemeteries, fuel filling stations) or even large scale phenomena (e.g. atmospheric deposition).

Prioritisation of the impacts of pollutants on groundwater flow systems in South Africa

101

Source of pollutant plays a large role in whether the pollutant will reach the groundwater table and if it does the rate at which the pollution will enter the groundwater system. There are also several existing methods for classifying the sources of groundwater pollution. A simplified classification based on that of Nonner (2002) was used to classify South African sources (see Table 1). 2.1.2 Pollutants Pollution refers to levels of hazardous substances in the environment over and above what would ordinarily be found in the absence of local activities. Groundwater pollution therefore refers to the occurrence of substances (inorganic species, organic compounds or microbial agents) in concentrations above those that would naturally be found in an aquifer. The substances themselves, both chemical and microbial, are called pollutants. There are various ways in which to group or classify groundwater pollutants. Each of these has major classes which can then be broken down into smaller categories. The choice of system and level of detail of the classification is dictated by the purpose of the classification for the sake of this investigation pollutants were classified according to: ● Fate in the environment – Degradable pollutants, which can be rendered harmless by natural processes and need therefore cause no permanent harm if adequately dispersed or treated; and – Persistent pollutants, which eventually accumulate in the environment and may be concentrated in food chains. – Pollutants may also be divided by their behaviour in water into: (a) Soluble pollutants, which includes most inorganic species and some organics. (b) Insoluble substances, which are small enough to be carried through the aquifer matrix, including microbial pollutants and colloidal inorganic pollutants. (c) Non-aqueous phase liquids (NAPLs), which are organic compounds that do not dissolve readily in water and remain as a separate liquid phase. These are further subdivided into Light Non Aqueous Phase Liquids (LNAPLs) and Dense Non Aqueous Phase Liquids (DNAPLs). ● Human health impacts – Non-harmful substances, which have no observed effects on human health. – Toxic substances, which cause various effects on the body from short-term exposure or long term accumulation, ranging in severity depending on the dose e.g. nausea, rashes, kidney failure or neurotoxic effects. – Carcinogenic substances, which are known to cause cancer. – Pathogenic substances, which are known to cause diseases in humans. ● Other aspects that are taken into account: – Duration of pollution—if the pollution results from a single (once-off) spill, the impact will probably be smaller than that resulting from continuous pollution. – The vulnerability of the aquifer represents the intrinsic characteristics that determine the sensitivity of an aquifer to the adverse effects resulting from the imposed

Water resources of arid areas

102

pollutant (Lynch et al., 1994). Factors taken into account include depth to groundwater, recharge, aquifer media, soil media, topography and impact of the vadose zone. 2.2 The risk-based methodology Rating occurs when contaminant sources are given a quantitative or qualitative measure of the potential hazard they pose to groundwater. Prioritisation methods focus on aspects such as contaminant loading, mobility, persistence and hazardousness while risk assessment develops these further into potential human health impacts. A risk analysis estimates the probability and consequences of a contaminant event and usually considers both the properties of the contamination source and the hydrogeological environment. Conventional set theory (Boolean) states that an element is either a member of a set or not. Fuzzy logic is an extension of conventional set theory enabling an element to belong to a set to a degree. The degree of membership is a function that defines the membership of an element to a set according to the value of the element. Membership is expressed as a value between 0 and 1. Zero implies 0% membership and 1 implies 100% membership. Linear membership functions are seldom used in practice in contradiction to sinusoidal functions which are very popular. In most cases risk analysis will involve more than one input to be considered in the analysis. Fuzzy logic makes it possible to generate a set of decision rules according to the number of inputs and these rules must then be evaluated by an expert in the field of study. The number of rules generated is given by the following equation is: n=2inputs where n represents the number of rules generated. The rules consist of all possible binary combinations of the respective inputs with a weight assigned to each rule representing the risk. The risk is then calculated using the following formula:

where n=number of rules, DOM=degree of membership and Wn=weight of rule n. 2.3 A tiered approach Based on the amount of data available a tiered approach is followed when considering risk assessments. The first tier (LEVEL 0) is a rapid assessment of sources in which minimal data are required and it produces low confidence results. This assessment should be completed within a few minutes and is based on a rating system. LEVEL 1 is the second tier which is a rapid assessment of contaminants on a local scale. It is intended to

Prioritisation of the impacts of pollutants on groundwater flow systems in South Africa

103

give the assessor a guideline of the risks. The assessment should take a couple of hours to complete. The next tier (LEVEL 2) is an intermediate assessment. The first step in the intermediate assessment is to collect all relevant data. Data requirements include aquifer and contaminant parameters, as well as health information. General information will be obtained from databases, but it is sometimes necessary to have site-specific data. The confidence attached to this assessment should be medium to high. Both the second and third tiers include risk assessments based on a fuzzy logic methodology. Figure 1 is a schematic representation of the tiers and the function performed on each level of assessment.

Figure 1. Tiered approach to South African prioritization methodology. In order to protect boreholes wellhead protection areas (WHPAs) need to be delineated. A WHPA can be defined as the surface and subsurface area surrounding a

Water resources of arid areas

104

borehole or wellfield, through which contaminants are reasonably likely to move and reach such a borehole or well field. In many cases it is difficult to protect the whole area, therefore various zones are established within the area. 3 CASE STUDY: CAPE FLATS WASTEWATER TREATMENT WORKS 3.1 The wastewater treatment works within the City of Cape Town prioritization of sources and contaminants on a regional scale The City of Cape Town (CCT) is located in the Western Cape Province on the southeastern corner of South Africa. A major portion of the CCT consists of the area known as the Cape Flats, which has an elevation of between 20 and 45m above sea level. CCT has a mean annual rainfall of 515mm/annum and an average temperature of 16.7°C. It is a winter rainfall area. The current population of the CCT is estimated at 3.2 million with the highest population density occurring on

Table 2. Source prioritization for CCT (incomplete list). Source prioritisation (from highest to lowest risk) On-site sanitation Petrol service stations (underground storage tanks) Cemeteries Stormwater/sewer systems Agriculture (general and crop cultivation) Feedlot/poultry farms Wastewater treatment

Table 3. Contaminant prioritization for CCT. Contaminant prioritisation (from highest to lowest risk) Nitrate Chloride Phosphate Potassium Ammonia & sulphates

Prioritisation of the impacts of pollutants on groundwater flow systems in South Africa

105

Table 4. Information used in site-specific risk assessment. Parameter

Assigned value

Vulnerability –Recharge 65mm/yr –Soil media Sa-LmSa –Aquifer media Intergranular –Vadose zone Beach sand –Groundwater depth 8m –Topography 1% Duration Continuous Contaminant* Nitrate Level of management Low *Once the contaminant is entered the software automatically pulls in the health risk information and physio-chemical behaviour from a database.

the Cape Flats and there are approximately 90000 consumers on informal sites. There are 21 wastewater treatment plants within the CCT. According to TIER 0 the wastewater treatment works are rated as the 7th highest pollutant source within the CCT. Due to the length of the complete list only the 7 highest potential polluters have been documented in Table 2. Typical contaminants found at wastewater treatment works include ammonium, nitrate, potassium, phosphate, chloride, sulphate and faecal pathogens. Micro-organisms were not included in the investigations and will therefore not be included in the prioritization list. The prioritization of the above-mentioned chemicals is listed in Table 3. 3.2 The Cape Flats wastewater treatment works risk assessment The wastewater treatment works has unlined sewage sludge drying ponds. The wastewater treatment works are situated on an unconfined primary sand aquifer. The information used to determine a site-specific risk is listed in Table 4. For the sake of demonstration only the risks for nitrates will be determined in this paper. The results of the assessment are summarized in Table 5.

Table 5. Results of risk assessment. Assessment

Risk (%)*

Source 58 Vulnerability 52 Health 99 Physio-chemical 75 Total 68 * Higher the risk higher the negative impacts.

Water resources of arid areas

106

Table 6. Data used to calculate protection zones. Parameter

Assigned value

Abstraction rate Transmissivity Effective porosity Hydraulic gradient Saturated thickness

10l/s 100m2/d 0.1 0.01 20m

Table 7. Calculated protection zone. Definition

Radius (m)

Zone 1: Highly protected area around the borehole. Its purpose is to protect the borehole from the direct introduction of pollutants into the borehole and its immediate area from spills, surface runoff, or leakage from storage facilities or containers. Potential pollutant sources in Zone 1 should be strictly monitored. 25 Zone 2: Is established to protect a borehole from contact with pathogenic micro-organisms which can emanate from a source located close to the borehole, as well as to provide emergency response time to begin active cleanup and/or implementation of contingency plans should a chemical contaminant be introduced into the aquifer near the borehole. 470 Zone 3: Is designed to protect the borehole from chemical contaminants that may migrate to the borehole; it typically includes a major portion of the 750 recharge area or the capture zone.

The results of the risk assessment for nitrates indicate there is a 68% chance that there are going to be negative impacts on the environment (including human health) as a result of groundwater becoming polluted with nitrates as a result of the wastewater treatment works. 3.3 Protection of boreholes The distance between a pollution source and a protected borehole can be calculated to ensure the borehole is not polluted. The zone of protection can then be delineated around the borehole. These wellhead protection zones can also be used to plan new boreholes. If all pollution sources are known then the ‘safe’ distance from a source can be calculated.

Prioritisation of the impacts of pollutants on groundwater flow systems in South Africa

107

Wellhead protection zones were calculated for boreholes in the Cape Flats. Table 6 contains the information needed for the calculations and Table 7 defines the protection zones and gives the radius of protection zones. 4 CONCLUSIONS AND RECOMMENDATIONS Pollution of South Africa’s urban aquifers presents a threat to the sustainability of this water resource. Man’s activities, use of chemicals and generation of wastes tend to concentrate potential sources of pollution in the urban areas. The threat caused by undesirable substances is recognized in this country, but the understanding of the extent of the problem in South Africa’s urban catchments is poor. This paper therefore briefly outlined a risk-based methodology to prioritise and determine the impacts of pollutant sources and pollutants. The methodology takes the following into account: ● Characteristics of pollutant sources ● Characteristics of pollutants ● Human health impacts of pollutants ● Vulnerability of South African aquifers ● Duration of pollution The methodology was then applied to determine: ● National list of priority chemicals and sources ● Regional list of priority chemicals and sources for the large South African urban areas ● Local risk assessments to determine the risks of certain pollutants ● Delineation of protection zones The results are intended to help groundwater practitioners and water authorities in assessing the likely transport, fate an impact of pollutants in the subsurface in an urban environment. It is recommended that the following aspects receive more attention in future research investigations: ● Based on the paucity of groundwater-related microbial data encountered in this project, the inclusion of these aspects in urban groundwater management must be regarded as a priority. ● Petroleum products, industrial thinners and mineral oils and other non-aqueous phase liquids represent a category of potential pollutants that have been largely overlooked by regulatory agencies and legislature, despite their harmful effects at small concentrations. ● A general lack of data on groundwater pollution from pesticides is evident. This is due to: (i) surface waters are the main source of water supply in the country; (ii) cost and difficulty to measure organic contaminants; (iii) private companies are often sensitive to make public data related to pollution problems. Therefore there is a need to investigate pesticides in groundwater.

Water resources of arid areas

108

REFERENCES Lynch, S.D., Reynders, A.G. & Schulze, R.E. 1994. Preparing input data for a national-scale groundwater vulnerability map of Southern Africa. Water SA, 20(3):239–246. National Environmental Management Act. Act 107 of 1998, Pretoria, South Africa. National Water Act. Act 36 of 1998, Pretoria, South Africa. Nonner, J.C. 2002. Chapter 3: Sources of groundwater contamination. In: A. Zaporozec (ed.) Groundwater contamination inventory: A Methodological Guide.UNESCO, IHP-VI, Series on Groundwater No. 2. 23–38.

Understanding problems of low recharge and low yield in boreholes: an example from Ghana A.J.E.Cobbing & J.Davies British Geological Survey, Wallingford, Oxon, UK Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: The Afram Plains region of Ghana experiences acute seasonal water shortages during the four to five month long dry season. The long-term development of the limited groundwater resources of the region has proved to be difficult as the hydrogeology is poorly understood. Failure of boreholes is common, and there is little or no monitoring of groundwater levels. A two-year study led by the British Geological Survey, including the monitoring of borehole-drilling programmes, has led to a better understanding of the groundwater resources in the area and has provided guidelines for data collection.

1 INTRODUCTION The Afram Plains area is located in the Eastern Region of Ghana, in the Volta River basin between latitude 6°30′ and 7°30′N and longitude 1°00′W and 0°15′E (Figure 1). The area is about 4285km2 in extent, and lies between lake water level at 76m and 300m above mean sea level. The topography is subdued, with the main feature being a low northeast to southwest trending ridge 200–300m high. The Afram Plains supports savannah vegetation that is being progressively cleared for agricultural use. Coarse tussock-grass with a few stunted trees covers the low-lying lakeside plain and dense bush with large trees covers the better-drained ridge area. Since initial settlement in 1930, the rural population of the Afram Plains has increased rapidly following the construction of the Akosombo Dam in the 1960s. Between 1970 and 1984 census data show a 250% increase in the farming population, attracted by fertile soils and improving infrastructure. There are now more than 140 villages on the Afram Plains. Access to the area is poor, with the principal route by ferry across Lake Volta. The main town is Donkorkrom, which has a hospital, secondary school and post office.

Water resources of arid areas

110

Figure 1. Map of Ghana showing the Afram Plains study area. 2 THE WATER SUPPLY PROBLEM IN THE AFRAM PLAINS Before construction of the Akosombo Dam, village water supplies were obtained from the perennial Afram and Volta Rivers, seasonal flows and pools along ephemeral tributary streams and shallow water-filled dugouts. Rainfall on the Afram Plains is seasonal, with an average of about 1200mm/year falling almost entirely between April and October. Surface drainage is mainly ephemeral, storm water draining by sheet flow as short-lived floods. The seasonal rainfall and limited surface water storage result in acute water shortages during the November to March dry season. Reliance on unprotected pools and dugouts for water supply results in water-washed and diarrhoeal disease, and much time and effort in water collection. Guinea worm infections occasionally occur in the Afram Plains. 2.1 Regional geology The Afram Plains are located at the southern end of the large (>100,000km2) Voltaian Sedimentary Basin formed during the Precambrian to early Palaeozoic Pan-African Orogeny of 730–550Ma. (Black and Liegeois, 1993, and Shackleton, 1976). The Voltaian Basin is interpreted as a foreland basin; with sediments of marine and terrestrial origin filling a flexural depression at the margin of the West African Craton (Ako and Wellman, 1985). Kesse (1988) and Anani (1999) describe the Voltaian basin sediments as fairly flat bedded sandstones, shales, pebble beds, mudstones, limestones and siltstones deposited unconformably upon older Precambrian rocks. The molasse type sediment pile, that is estimated to be more than 4km thick, resulted from erosion of mountain chain fold belts that occurred along the present Ghana-Togo border to the east. The Voltaian Formation

Understanding problems of low recharge and low yield in boreholes

111

Obosum Beds that underlie much of the southern Afram Plains have yet to be studied in detail. Present geological knowledge has been derived from rapid reconnaissance surveys, several deep exploration boreholes and a number of shallow groundwater boreholes. 2.2 Previous groundwater development Development of the groundwater resources of the Voltaian sediments of the area began in 1963–65 when the Geological Survey of Ghana and the Volta River Authority (VRA) drilled a series of test and production boreholes in response to populations displaced by the rising lake waters. During the late 1960s and early 1970s the Catholic Church funded the construction of 28 hand-dug wells to supply small villages. These were mainly located in valley sites to replace unprotected shallow dugout sources. Additional boreholes were installed by the VRA at Donkorkrom and Kaklakoklope in 1983/84. UNICEF provided a borehole for the secondary school at Donkorkrom in 1983. During 1984, Prakla Seismos drilled 47 village boreholes for the German NGO Misereor. Of these, 19 boreholes were dry and 17 had yields greater than 301min−1. Although Lake Volta forms the eastern boundary of the area, the underlying low permeability rocks are the main source of water especially in the more remote western area. During the 1990s, more than 300 boreholes were drilled to meet the water supply needs of the expanding population.

Table 1. Summary of borehole drilling on the Afram Plains, 1963 to 2001. Organisation

Period No. Wet Dry of Bhs Bhs Bhs

Volta River Authority Prakla Seismos for Misereor World Vision International World Vision International WaterAid/Afram Plains Dev. Org. DANIDA exploration boreholes Totals

1963– 1965 1984

10

6

4

47

28

19

1990– 1995 1999– 2000 1996– 2001 2001

152

92

60

66

?

?

101

67

34

5

5

0

381

198 117

Many of the boreholes drilled were dry whilst other nominally successful boreholes showed a progressive decline in yield to fail after two to three years of use, especially in the west of the area. Due to their short period of land tenure, communities have yet to develop coping strategies to manage the limited water available during dry periods.

Water resources of arid areas

112

Populations attracted to the area by the groundwater supply have no effective alternative water source if borehole yields fail after several years of use. 3 RECENT WORK ON THE AFRAM PLAINS The British Geological Survey (BGS) first worked on the hydrogeology of the Afram Plains in 1985–86, in a project examining shallow wells and boreholes in the Donkorkrom area (Buckley, 1986). Beginning with a visit to the area in February 2000, the BGS collaborated with the Afram Plains Development Organisation (APDO), WaterAid, DANIDA, Legon University (Accra) and other partners in a two-year project specifically aimed at investigating the hydrogeological problems of the area. The BGS work was funded by the British Department for International Development (DfID). The project was timed to coincide with the drilling of 36 village water supply boreholes on the Afram Plains, funded by WaterAid working with the APDO. In addition, DANIDA funded the drilling of a further 5 deep (>100m) exploration boreholes, the first four of which were sited and geologically logged by BGS hydrogeologists. Studies carried out by the BGS in collaboration with local partners included: ● A reconnaissance geological and hydrogeological survey, and the creation of a GIS base map of the area. ● The geophysical survey of four of the five deep exploration borehole sites using frequency domain electromagnetic induction (EM34). ● The geological logging of rock chip samples produced during drilling, and the recording of penetration rates and drill stem yields. ● The test pumping of boreholes, including the demonstration of bailer tests and low yield “whale” pumps. ● The sampling of groundwaters for hydrochemical analysis of major and minor ions, and isotopes. ● The geophysical logging of the deep exploration boreholes was carried out by DANIDA.

4 SURFACE GEOPHYSICS ON THE AFRAM PLAINS Electrical resistivity and EM geophysical exploration surveys have been undertaken in the Afram Plains during other development projects. As in other hydrogeologically “difficult” areas in Africa, these methods have fallen out of favour in the Afram Plains, being seen as relatively expensive and time consuming for little benefit. This is due to a combination of: ● the mode of groundwater occurrence in the area, such as deep fractures, often thick weathered zones, that cannot be defined using geophysical surveys, ● the lack of experienced personnel capable of interpreting geophysical results for sedimentary environments.

Understanding problems of low recharge and low yield in boreholes

113

BGS undertook 11km of EM34 surveys at the first four exploration borehole sites, using 10m, 20m and 40m inter-coil separations. Readings were made in both vertical and horizontal orientations. The survey results were correlated with the geological logs from the exploration boreholes. Geophysicists from the University of Ghana, Legon, undertook max-min EM and electrical resistivity geophysical traverses along the main road in the eastern Afram Plains (Banoeng-Yakubo and Armah, 2001). The result of these studies demonstrate that geophysical surveys can be used in the eastern Afram Plains to differentiate between near surface shale, siltstone, sandstone and conglomerate bands, as well as delineate possible fault zones. In the west of the area, re-cemented sandstones up to 60m thick form a low permeability homogeneous layer below an ancient weathered surface. Thin water bearing fracture or weathered zones beneath this layer cannot be detected using EM34 or VES equipment. 5 DATA GATHERED DURING BOREHOLE DRILLING Useful geological and hydrogeological data that can be gathered during the drilling of a borehole includes: ● Geological data ● Penetration rate data ● Flow data ● Hydrogeological data Rock chip samples produced during drilling were collected at 1m intervals. Weathered zones (colour changes) and fracture zones (calcite and quartz mineralisation) enabled identification of water bearing zones. The chip samples were placed in a marked half pipe and photographed to produce pseudo-core logs. This procedure allowed zones of water inflow to be correlated with changes in lithology, and deductions regarding the nature of groundwater occurrence to be made. The rate of drill penetration and flow rate, determined at water strike zones and at the end of each drilling rod can be correlated with changes in lithology and weathered zones. Photo logs can show the nature of the weathered zones. The results obtained from exploration borehole showed that the rock types present are generally tight and fine-grained, with water being produced from horizontal weathered zones and along lithological boundaries rather than near-vertical fractures. In the western half of the area, the presence of a thick duricrust weathered zone, stopping recharge to underlying aquifer systems, was recognised. 6 BOREHOLE GEOPHYSICS A suite of geophysical logs was obtained from six boreholes in the study area, i.e. the five deep exploration boreholes together with a water supply borehole located at the APDO office in Tease. The calliper logs show the fracture zones, which can be correlated with the drillers report and the chip sample logs. The fracture zones are also indicated by lower resistivity measurements.

Water resources of arid areas

114

Pumped fluid logging of the boreholes clearly shows that fluid inflows occur at discrete fractured or weathered horizons, and that most of the water obtained from the boreholes is derived from these features. The discontinuous nature of the fracture systems that supply water is illustrated by drilling at the APDO office in Tease: in 2000, a 70m deep borehole (“hole no. 28”) was drilled in an attempt to provide a water supply for the office. This borehole proved to be dry and was backfilled. In 2001 a further two boreholes were drilled within 20m of this hole, one to 54m and a deep exploration borehole to 152.8m. Both of these boreholes yielded water. 7 TEST PUMPING OF BOREHOLES Pumping test data in fractured aquifers is more difficult to interpret compared with intergranular systems. There is often a distinct change between early and late time drawdown rates, due to the effect of fracture dewatering. This can allow erroneous interpretations to be made, particularly if pumping tests are carried out over only short periods of time. Pumping test interpretation requires specific training, and pumping tests have sometimes been done on the Afram Plains merely as required by the contract, without the pumping test information being used to inform the borehole completion. BGS developed a simple bail test, which allows field personnel on the Afram Plains to decide in a general way whether or not to equip a borehole, without going through the lengthier and more complex process of a pumping test (Davies and Cobbing, 2002). There are cases however where the bail test is inconclusive and the borehole requires a pumping test. Bail tests are

Understanding problems of low recharge and low yield in boreholes

115

Figure 2. Geophysical logs of deep exploration boreholes at Gazeri Camp (left) and Samanhyia, near Tease. Fractured and weathered zones can be seen on the calliper and induction resistivity logs, and the pumped flowmeter logs show that most flow into the boreholes is derived from these horizons. recommended as a rapid and simple field procedure to be used by staff not trained in pumping test interpretation to decide whether or not to equip a borehole with a pump. Simple pumping tests give indications of the productiveness of the systems but the results obtained are from “fractured” aquifer systems with high secondary permeability zones are difficult to reconcile. Such systems can initially give high yields but when they are dewatered during extended periods of over-pumping these systems can suddenly fail. 8 HYDROCHEMISTRY RESULTS Water samples for hydrochemical analysis were obtained from 29 boreholes and wells during the 2001 visit. Samples were taken from sources after several minutes of pumping where possible. Measurements of pH, specific electrical conductance (SEC), temperature and bicarbonate were taken at each site. Filtered acidified and non-acidified samples were obtained from each source for laboratory analysis. A GPS was used to locate the areal co-

Water resources of arid areas

116

ordinates of each sample site. Stable isotope analysis (δ2H and δ18O) was carried out on twelve samples by mass spectrometry. The results of these determinations plot close to the world meteoric water line. There is some evidence for the possible mixing of lakederived waters with aquifer waters in some areas. The major and minor ion analyses show that most determinants are within World Health Organisation (WHO) Guide Values, with the exceptions of boron and sodium that are a problem in the unfractured shale and sandstone area. Nitrate and ammonium levels in a few boreholes were evidence for anthropogenic pollution, which can occur because water is able to move relatively rapidly through fractures. The fluoride concentration in one sample exceeded WHO Guide Values. 9 HYDROGEOLOGY OF THE AFRAM PLAINS A five-fold hydrogeological division of the rocks of the Afram Plains can be produced, based on the conclusions of Bannerman (1990) and Acheampong (1996), and taking the current study into

Figure 3. Five hydrogeological divisions on the Afram Plains. account (Figure 3). The hydrogeology of each of these units is summarised in Table 2. 1. Massive conglomerate and sandstone. 2. Fractured shale and grey sandstone. 3. Quartzitic sandstone and conglomerate.

Understanding problems of low recharge and low yield in boreholes

117

4. Feldspathic sandstone, arkose, siltstone and mudstone. 5. Unfractured shale and sandstone.

10 DISCUSSION In regions of seasonal or low rainfall with ephemeral drainage patterns rural settlements may be totally dependent upon groundwater supply during the dry part of the year. Such is the present and future shortage of land in many areas that communities once settled in water poor areas are difficult to move. Therefore understanding of groundwater resources is a vital factor for long-term development plans of such marginal areas. Groundwater development in the Afram Plains has followed a pattern that is typical of areas underlain by low permeability rocks in sub-Saharan Africa. Reconnaissance level geological and hydrogeological surveys were first undertaken with limited drilling more than thirty years ago. Some borehole drilling by the VRA was undertaken at the time of population resettlement following the building of the Akosombo Dam and consequent flooding in the 1960s, but these boreholes have fallen into disuse following lack of maintenance. NGO-led water supply programmes, undertaken by World Vision International, a Catholic Church Group and WaterAid, funded the drilling of some 370 boreholes on the Afram Plains during 1984–2001. During these programmes the economic design and construction of boreholes, and borehole drilling “success rates” were emphasised. A borehole was judged a success if “wet” at the completion of drilling. The hand pump equipped boreholes were expected to supply 250 people with at least 20 litres per capita of water per day. In the Afram Plains the acceptable yield minimum is about 121min−1, due to the low borehole yields obtained. The high borehole “failure rate” (40%) has led to further study of the distribution of fracture and near surface weathered zones, these being perceived as the best groundwater bearing targets. Although many boreholes have been drilled, the geology of the area, groundwater occurrence, and the nature of the water resource remain poorly understood. This problem is exacerbated by the failure of apparently successful boreholes after 3–4 years of use.

Table 2. Summary of the hydrogeology of the five hydrogeological units. Description Ground of rock/ water hydrogeology targets unit Obosum Massive Beds— conglomerate Upper and sandstone Voltaian System

Ground Ground Field Technology Comments water water techniques potential quality

Weathered ** zones and fracture zones. Success rate ~66% wet 38%≥30l/min

Good. Presence of NO3N and NH4 indicates pollution in heavily used

Weathered Boreholes conglomerate 60–100m gravel often visible at surface: EM34—used to locate fractures and sandstones/ conglomerate

Good recharge, best sites located in valleys. Boreholes should be drilled to below present day

Water resources of arid areas

118

boreholes near surface. in village VES— centre. indicates depth of weathering

Quartzitic Weathered ** sandstone and zones and conglomerate fracture zones. Success rate ~67% wet 40%≥30l/min

Good.

Quartzitic Boreholes sands often 100–150m visible at surface. EM34—used to locate fractures and sandstones/ conglomerate near surface. VES— indicates depth of weathering

Feldspathic sandstone, arkose, siltstone and mudstone

Good.

Weathered Boreholes purple brown 100–150m sandstone platform surface beneath thin ferrecrete. Difficult to identify fractures with EM34, sandstones have been recemented to 60 m. VES—may indicate

Weathered */** zones and fracture zones. Success rate ~66% wet 39%≥30l/min

lake level. May be able to induce flow from the lake along fracture zones, Problems with pollution in villages. Moderate recharge, best sites located in valleys. Boreholes should be drilled to below present day lake level. May be able to induce flow from the lake along fracture zones. Problems with pollution in villages. Very poor recharge potential due to recemented layer down to ~60m. Deep holes may intercept weathered zones, Remoteness precludes direct recharge from lake

Understanding problems of low recharge and low yield in boreholes

119

depth of weathering

along fractures, Fractures poorly defined.

Description Groundwater Groundwater Groundwater Field of rock/ targets potential quality techniques hydrogeology unit

Technology Comments

Unfractured Weathered * shale and grey zones and sandstone fracture zones. Success rate ~50% wet 14%≥30l/min

Poor to saline. Low lying Boreholes— low altitude 50–100m lake side areas. EM34— moderate to high conductivities, used to locate fracture zones VES— indicates depth of weathering

Fractured shale Weathered *? and sandstone zones and fracture zones. Success rate Unknown due to lack of data

Poor to saline?

Poor to moderate recharge to tight formation except where conglomeratic bands area present. Boreholes should be drilled to below present day lake level. Boreholes— Unknown 50–100m

Low lying low altitude lakeside areas. EM34— moderate to high conductivities, used to locate fracture zones VES— indicates depth of weathering

KEY: Groundwater potential: *Low; **Moderate; ***High. Note: Groundwater Potential is an overall function of groundwater storage, groundwater yield and groundwater residence time (length of time groundwater remains in the unit, i.e. rate of groundwater throughflow). It indicates both the available yields and the length of time these are available for: i.e. high, moderate or low yields, available only during the wet season and immediately afterwards, or yearround. See below for more detail. EM34 conductivity response: High>50mmhos/m; Moderate 20– 50mmhos/m; Low <20mmhos/m. Yield: High >1l/s; Moderate ~0.5l/s; Low <0.2l/s. Note: Where groundwater residence times are long, groundwater availability is likely to be less vulnerable to variations in seasonal rainfall—e.g. one year of drought. Where few data are available locally, the interpretations given here are preliminary, and should be updated as new data are provided by continuing groundwater development work.

Water resources of arid areas

120

11 CURRENT CONCEPTUAL MODEL OF THE AQUIFER: A SUMMARY The main features of the aquifer model for the Afram Plains as a whole are as follows: ● Groundwater is thought to occur in discrete fracture systems or zones of weathering. ● The geological units have different hydrogeological characteristics but all are relatively low yielding. ● In the west of the Afram Plains in particular, the aquifer units may not be adequately recharged during successive wet seasons, leading to the progressive mining of groundwaters that leads to the failure of boreholes with time. Old water is often present in the fracture systems. ● In areas where recharge of surface water occurs, rapid movement through near surface weathered zones and fracture systems can lead to rapid transport of contaminants into boreholes below sanitary seal zones, as indicated by high ammonium and nitrate levels discerned in the central village borehole water sources on the Afram Plains. ● Water bearing weathered zones may be too deep and discrete to be determined using geophysical survey methods. ● Drilling deep boreholes to below the present day lake level may allow interception of fracture and weathered systems that can potentially be recharged by lake water. This process of recharge from the lake remains to be proven. ● The collection of accurate geological and hydrogeological data is vital for better understanding of the aquifer systems present. The use of currently available data is hindered by a lack of accurate site locations. The interaction of geological factors such as lithology, diagenesis, recent weathering, ancient weathering, tectonism with ancient and modern water level changes needs to be understood.

12 CONCLUSIONS The water supply problems on the Afram Plains cannot be solved by borehole drilling and groundwater development alone. The failure of boreholes after two or three years of use is particularly serious since in that time communities come to rely on the groundwater resource. Conjunctive use with rooftop rainwater catchment systems and small dams may need to be considered as well as artificial recharge to aquifers. There is a need to understand recharge mechanisms before borehole drilling commences. This project has demonstrated the types of data that can be easily collected at little additional cost during borehole drilling, and the uses to which such data can be put to the benefit of subsequent water supply programmes. A regional summary of groundwater occurrence in this “difficult” hydrogeological area has been built up, and presented in a format that can be used in subsequent groundwater development. The general shift from centralised groundwater development towards demand-driven, private organisation or NGO led work in Africa has had some benefits in terms of sustainability, community involvement and ownership issues, and the targeting of resources at the poorest communities. However, the negative effect has been the non-collection, storage and sharing of basic groundwater data, which leads to a lack of understanding in those areas where the groundwater resources are limited or difficult to access. At present, data collection is frequently seen

Understanding problems of low recharge and low yield in boreholes

121

as an unaffordable “optional extra”, adding mainly difficulty and expense to a project Such data as are collected often become difficult to access, since no effective central repository for data is currently in operation. The project aimed to overcome this by depositing the data collected in easily retrievable Word and Excel based packages with the Afram Plains Development Organisation staff. REFERENCES Acheampong, S.Y.1996. Geochemical evolution of the shallow groundwater system in the Southern Voltaian Sedimentary Basin of Ghan. A dissertation submitted in partial fulfilment of the requirements for the degree of Doctor of Philosophy in Hydrology/Hydrogeology, Univ. of Nevada, Reno. Acheampong, S.Y. & Hess, J.W. 1998. “Hydrogeologic and hydrochemical framework of the shallow groundwater system in the southern Voltaian Sedimentary Basin”. Hydrogeology Journal, 6:527–537. Ako, J.A. & Wellman, P. 1985. The margin of the West African craton: the Voltaian Basin. Journal of the Geological Society of London, 142:625–632. Anani, C. 1999. Sandstone petrology and provenance of the Neoproterozoic Voltaian Group in the southeastern Voltaian Basin, Ghana. Sedimentary Geology, 128:83–98. Bannerman, R.R. 1990. Afram Plains borehole drilling programme, hydrogeological survey for WaterAid Ghana. Banoneng-Yakubo, B. & Armah, T. 2001. Hydrogeological and geophysical test investigations in the Afram Plains, Ghana. Department of Geology, Univ. of Ghana, Legon, for DANIDACWSA Project, Eastern Region, Ghana. Black, R. & Liegeois, J.-P. 1993. “Cratons, mobile belts, alkaline rocks and continental lithospheric mantle: the Pan-African testimony”. Journal of the Geological Society, London, 150:89–98. Buckley, D.K. 1986. Report on advisory visit to WaterAid projects in Ghana. British Geological Survey Technical Report. Davies, J. & Cobbing J. 2002. An assessment of the hydrogeology of the Afram Plains, Eastern Region, Ghana. British Geological Survey. Technical Report CR/02/137N. Grant, N.K. 1967. Complete Late Precambrian to Early Palaeozoic orogenic cycle in Ghana, Togo and Dahomey. Nature, 215:609–610. Kesse, A.O. 1988. The Mineral and Rock Resources of Ghana. A A Balkhema. Shackleton, R.M. 1976. Pan-African structures. Philosophic Transactions of the Royal Society, London. 280: 491–497. World Vision 1995. The Conrad N Hilton Foundation Funded World Vision Ghana Rural Water Project, Hydrogeological Report, Second Phase.

Spatial variation of groundwater recharge in a semi-arid environment—Serowe, Botswana L.M.Magombedze & B.Frengstad Geological Survey of Norway, Trondheim, Norway M.W.Lubczynski International Institute of Geoinformation Science and Earth Observation, Enschede, Netherlands Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: The estimation of groundwater recharge by conventional direct methods of subtracting evapotranspiration (ET) from rainfall is practically limited in semi-arid regions by the difficulty in determining ET with sufficient accuracy for calculation of very low recharge fluxes in the order of a few mm/yr. In the Serowe study area two alternative methods are used to estimate spatial distribution of groundwater recharge. Based on 127 point measurements of chloride concentration in groundwater, the chloride mass balance method gave recharge rates ranging from 2mm/yr to ~30mm/yr with a harmonic mean of 12mm/yr. In order to present recharge spatially the data were interpolated by kriging of the logarithms of the net recharge values. GIS map modelling technique was used to integrate the influence of various recharge attributes in a single semiquantitative recharge potential map. The relative influence of factors such as soil type, vegetation, lineament density and slope among others was weighted and subsequently validated by site specific recharge rates. This method can give a valuable first estimate of the recharge rate based on surface properties identified from maps and remote sensing in areas where detailed hydrogeological information is limited.

1 INTRODUCTION Serowe area, like the whole of Botswana, is characterised by a low rainfall pattern and lack of surface water resources. People in this area depend mainly on groundwater. Recently, the population of Serowe has increased tremendously, making it the largest village in the country, thus widening the gap between demand and availability of water. In this regards, it is important for sustainability and management purposes to determine the renewable groundwater resources and how net recharge varies in space and time. This

Spatial variation of groundwater recharge

123

study focuses on estimation of net recharge rates using the chloride mass balance method and the assesment of spatial variability of both net recharge and potential recharge. Serowe study area is situated in the Central District of Botswana, at the eastern fringe of the Kalahari Basin and is about 275km northeast of the capital, Gaborone (Fig. 1). It is characterized by semi-arid climate with cool dry winters (May to September) and hot moist summers (October to April). It receives an average annual rainfall of 447mm/year. The most prominent feature in the area is the 90–150m high escarpment, which extends NNE-SSW splitting the area into two hydrologically contrasting areas: eastern and western. The western part slopes gently to the west and is covered by thick Kalahari sands. The eastern part slopes steeply to the east and Kalahari sand cover is thin or absent where the Ntane Sandstone or the basalt outcrops. The study area boundaries were assigned after the boundaries of the numerical model set up by Wellfield Consulting Services (WCS 1998, Lubczynski 2000). The northern boundary is a regionally stretching impermeable graben. The eastern and south-eastern boundary is delineated along the

Figure 1. Serowe study area. eastern wedging limit of the Ntane sandstone. The southern and south-western boundary is an impermeable dolerite dyke. The western boundary is an artificial one with no physical meaning and it stretches approximately from UTM coordinates (400000, 7517249) to (420150, 7544376). The study area is located entirely in the Karoo strata, where the most important aquifer is the ~100m thick Ntane Sandstone layer (Lubczynski 2000). The Ntane aquifer is underlain by the impermeable Mosolotsane Formation, mainly siltstone and mudstone and overlain by locally fractured Stormberg Basalt with thickness ranging from 0m to 143m. Groundwater recharge is expected where the basalt is absent so the Ntane aquifer lies directly under the Kalahari and where basalt is thin and fractured. The Kalahari sand cover, which overlies the Stormberg Basalt, is thick in the western part of the study area 50–75m and thin <10min the eastern part. The Kalahari sands have a high recharge potential where they do not contain calcretes and silcretes. The structurally dominant feature in the study area is the WNW-ESE trending fault system that is intruded by dolerite dykes of 10−40m width (SGC 1988, Lubczynski 2000). These features have divided the area into a series of horsts and grabens (Fig. 1), which enforces structurally controlled groundwater flow outward to the west and to the east of the groundwater divide located on the western side of the escarpment (Fig. 1).

Water resources of arid areas

124

2 GROUNDWATER RECHARGE ASSESSMENT The net groundwater recharge is defined by Lubczynski (2000) as the amount of water that reaches the groundwater minus groundwater evapotranspiration representing discharge of groundwater by tree transpiration and evaporation from groundwater in the form of upward convective water flux. Groundwater recharge is the most important factor in evaluating the renewability of groundwater resources of regional aquifer systems in arid and semi-arid environments and it is unfortunately the most difficult to quantify (Allison 1988). Several methods of estimating recharge have been developed. These can be divided into physically based, chemical and isotopic methods (Simmers 1997, Lerner 1990). Recently, also numerical models have been used to estimate groundwater recharge. While making site-specific recharge measurements, one of the most important problems to overcome is the spatial data presentation (Allison 1988, Lerner 1990). In this study, two techniques i.e. kriging interpolation (Ahmed et al. 1995, Gieske 1999) and spatial extrapolation with intergrated GIS recharge modelling technique were applied to assess spatial variability of recharge. 2.1 Chloride mass balance measurements The chloride mass balance method, developed by Eriksson & Khunakasem (1969), allows calculation of the average recharge rate in groundwater, RT in mm/yr provided the Cl ion behaves

Figure 2. Interpolated recharge map with net recharge estimates from the chloride mass balance method. conservatively and mass is conserved (Beekman et al. 1996): (1)

Spatial variation of groundwater recharge

125

where RT=total recharge rate (mm/yr), TCl=average annual total chloride deposition at the surface (mgm−2/yr); P=rainfall (mm/y); ClP=chloride content in rainfall (mg/l); Clgw=chloride content in groundwater (mg/l); D=dry deposition of chloride measured during dry season (mgm−2/yr). Due to lack of total dry deposition data during this investigation, estimation of recharge was based on the 1986/93 dry deposition of chloride (D) for Serowe of 442±124mgm−2/yr determined from rain gauge measurements (Selaolo 1998). Also Gieske (1992) recommended a similar D of 400–500mgm−2/yr for Serowe area. Based on 127 measurements of chloride in groundwater collected in September 2001, site-specific net recharge rates ranging from 2mm/yr to 74mm/yr were calculated according to equation 1. The results are shown in Figure 2. Two high values were regarded as outliers and have been omitted from further calculations. The method gave a harmonic mean recharge rate of ~12mm/yr, which is within range of estimates from previous studies. For example, SGS (1988) calculated net recharge rate of 11.7mm/yr from the chloride mass balance method. 2.2 Recharge interpolation Groundwater recharge is a function of space and time and is often highly variable. With regard to spatial variability, usually geostatistical analysis is carried out using only quantitative point data and qualitative geological information is often neglected (Ahmed 1991 in Ahmed et al. 1995). The key to geostatistics is the semi-variogram, which is a graphical presentation of spatial correlation in a given data set (Cohen & Spies 1990). The method assumes normally distributed spatial data. In this study the recharge data set was tested for normality using the Anderson-Darling normal probability test applying Minitab software. The test showed that the data is not normally distributed. The same data was successfully normalised by taking the logarithm of the net recharge values. These values were then further analysed geostatistically. The calculated semi-variogram and the spherical semi-variogram model are shown in Figure 4. A value of R2=0.71, indicating a good fit between the experimental data and the semivariogram model was found. Based on the parameters obtained from the semi-variogram model, the interpolation of the logarithms of the net recharge values was done using the ordinary kriging method. Finally, the antilogarithm of the interpolated values was taken resulting in the net recharge

Water resources of arid areas

126

Figure 3. Groundwater recharge potential map with point net recharge estimates.

Figure 4. Semi-variogram and semivariogram model for recharge. map (Fig. 2). The map of the net recharge distribution shows a general decrease in net recharge from the south-east and east towards the north and the west. Bulbs in the map

Spatial variation of groundwater recharge

127

indicate places of locally high and low net recharge. Low net recharge in the north could be explained by confinement by very thick basalt layer hampering net recharge. High evapotranspiration resulting in high chloride concentration in groundwater is the likely cause of locally low net recharge in the east. High net recharge in the northeast could be due to the absence of a confining basalt layer. Generally, recharge is lower in the western part than in the eastern part. This is due to the thick sand and basalt cover, deeper groundwater table and also lower rainfall in the western part of the study area. 2.3 Recharge attributes for GIS modelling Spatial variation of recharge is influenced by the spatial variability of factors specific to the study area like rainfall, vegetation type and density, soil type and texture, geology, landuse, topography,

Table 1. Kalahari thickness scores. Kalahari thickness (m)

Score

0–10 10–20 20–30 30–40 >40

5 4 3 2 1

Table 2. Depth to water table scores. Depth (m)

Score

<50 50–60 60–70 70–80

5 4 3 2

landform and depth to water table (Lerner et al. 1990). However, these factors influence recharge with different weights. Therefore, GIS modelling, which involves combining maps of different recharge attributes was used to come up with a recharge potential map, qualitatively displaying the spatial distribution of groundwater recharge potential. The modified index overlay method described by Bonham-Carter (1997) was applied in this study. Each attribute map was subdivided into six classes, with most recharge suitability assigned score 5 and 0 representing the least recharge suitability. Each attribute map was then assigned a weight according to its significance in controlling recharge. The average score of each pixel is defined as: (2)

Water resources of arid areas

128

where S=the weighed score for each pixel, Wi=the weight of the ith map and Sij=the score of the jth class of the ith map. The recharge potential map of the study area was derived considering the following factors influencing recharge potential: – Basalt cover: Basalt lacks primary porosity and impedes vertical flow of water except where it is fractured. Therefore high recharge suitability is expected where basalt is absent (score 5) and least recharge where basalt is present (score 0). – Kalahari thickness and depth to groundwater table: The thinner the Kalahari layer the faster water passes the unsaturated zone and reaches the aquifer and the less chances of water encountering duricrusts. The shallower the water table the faster water reaches the saturated zone. Assigned scores are shown in Tables 1 and 2. A correlation matrix showed that Kalahari thickness and depth to groundwater table are positively correlated. Therefore, the two scored maps were crossed and combined in a new map which was used for assigning recharge suitability scores. – Soils: Porous and coarse textured soils have high infiltration capacity and low field capacity and therefore enhance recharge processes. Soils with high silt and clay content do not release water to the lower zone fast enough to allow recharge processes to take place. Scores were assigned according to soil type and average infiltration rates as described by De Wit & Nachtergaele (1990), see Table 3. – Vegetation density: Areas with high vegetation density have also large density of root network and are therefore less suitable for recharge. High vegetation density may also facilitate infiltration by reducing runoff, but this is less important in case of Serowe due to the generally high infiltration capacity of the Kalahari sands. A NDVI map constructed from the Landsat TM 7 image of 24 April 1998 was classified into three recharge suitability classes (Table 4). – Slope: Slope of the landscape influences recharge rate. Generally, the steeper the slope the more runoff and the less the amount of water infiltrating the soil. A slope percentage map was assigned scores according to the FAO slope classification (Allen et al. 1998) see Table 5. – Lineament density: The influence of lineaments on recharge is greatest when the fractures and faults are deep, continuous over some distance and are not filled with secondary material. Not

Table 3. Scores for soils. Soil type Average infiltration rate (cm/hr)

Score

Arenosols 25–33 Regosols 22–30 Luvisols 0.05–0.8 Lixisols 0.05–0.8 Vertisols <0.05

5 4 3 3 1

Table 4. Scores for vegetation density. NDVI

Vegetation density class

Score

0–130

Low

5

Spatial variation of groundwater recharge

130–140 Medium 140–250 High

129

3 1

Table 5. Scores for slope. Slope (%)

Description

Score

0–2 2–8 8–16 16–30 >30

Flat Gently sloping Undulating Rolling Hilly

5 4 2 1 0

Table 6. Lineament density scores. Density

Class

Score

<20 20–40 >40

Low Moderate High

1 3 5

Table 7. Scores for vegetation cover. Vegetation cover

Score

Bare Open savanna Open savanna shrub Open savanna tree Pans Escarpment dunes Escarpment woodlands Outcrop area Hardveld shrub Hardveld woodland Agricultural area Riverine woodland Grassland Hardveld savanna

3 5 4 3 2 3 2 4 3 2 3 2 5 4

Table 8. Weight assigned to recharge attributes. Attribute

Model Model Model Model 1 2 3 4

Soil type Vegetation density Basalt cover Lineament density Slope

4 6

5 6

5 6

4 6

4 3

4 3

4 3

7 3

2

2

2

2

Water resources of arid areas

Kalahari 7 thickness/ depth to water table Vegetation cover 7

130

7

8

7

6

7

6

every lineament, but the majority of them can provide paths for infiltrating water. The density of lineaments was assessed from a combination of a lineament map obtained from WCS (1998) and lineaments derived from satellite TM5 data (Table 6). – Vegetation cover: Evapotranspiration and interception vary with vegetation cover. More evapotranspiration is expected from woodlands than from shrubs and grass. Therefore high recharge is associated with grasslands and low recharge with woodlands. In this assessment the vegetation map prepared by Ecosurve (1998) was used (Table 7). 2.4 GIS map modelling for mapping of recharge potential The recharge attributes considered for GIS map modelling do not bear equal importance to recharge potential mapping. Each map was therefore assigned a weight according to its perceived importance (Table 8). Most importance was attached to Kalahari thickness, depth to groundwater table, and vegetation cover, followed by soil type and vegetation density since they are

Table 9. Classification of recharge zonation maps. Weighted score

Recharge class

0–2.8 2.8–3.0 3.0–3.5 3.5–4.2 4.2–5

Very low Low Moderate High Very high

fundamental determinates of amount of water available for recharge followed by basalt cover. Lineament density and slope were considered the least important. There is not much variation in slope in the study area and the faults and fractures on the Kalahari are masked by the thick Kalahari sand cover. In the eastern part of the study area higher weights were applied to the lineament occurence. GIS map modelling refers to adjustments of attribute scores and attribute weights in order to provide the most realistic spatial distribution of recharge potential. Four different models based on weight sets given in Table 8 were used in this process. Four recharge potential map models were then produced following equation 2. In order to come up with recharge zonation models, each of the four recharge potential maps was subdivided into five recharge potential classes namely, very low, low, moderate, high and very high according to Table 9. Conceptually, all four recharge potential models are realistic when visually assessed. The four recharge potential models were then overlaid with recharge rates from the chloride mass balance method for comparison and validation of the GIS recharge potential models. In principle GIS recharge modelling should be manipulated by

Spatial variation of groundwater recharge

131

changing scores and weights to fit the point distribution of recharge (Lubczynski & Gurwin 2004). However, due to time constraints, this could not be done resulting in deterministic recharge potential rather than recharge potential modelling. Though verification data is missing for the far eastern and northern part of the study area, Model 3 is in best agreement with most of the point chloride mass balance results and was therefore considered the most realistic. Figure 3 shows the spatial variation of recharge potential based on Model 3 overlain with point net recharge values. In this map there are places in the central part of the study area with inconsistencies, where the moderate recharge values of the chloride mass balance fall in the very low recharge category of the recharge potential map. This could be explained as inaccuracy of the lineament assessment in terms of the size i.e. depth, width and openness which could not be incorporated in the GIS map modelling. Also, it is difficult to assess which lineaments are important for recharge and which ones are not. Other inconsistencies such as those in the northeast, north and center where low chloride mass balance recharge rates fall in the moderate to high recharge potential zones are most likely attributed to the subjective way of scoring and weighting in map modelling. It could also be attributed to the underestimation of recharge due to the presence of groundwater evapotranspiration in that area. 3 CONCLUSIONS – The chloride mass balance technique gives an insight into the spatial variation of net recharge. Net groundwater recharge in the Serowe study area is spatially variable and it generally ranges from 2mm/yr to ~30mm/yr with a harmonic mean of 12mm/yr. – The recharge potential map obtained by GIS recharge map modelling provides a semiquantitative distribution of recharge. This technique is a very useful tool in reconnaissance studies and in numerical model calibration, particularly when quantitative data is limited, because it is able to integrate spatial data from various sources. The zones do not necessarily show that recharge occurs or how much recharge occurs but give an indication of where recharge is most likely to occur. Such an assessment can therefore also be used as a planning tool taking into account local hydrogeological knowledge and constraints in the development, management and use of groundwater resources. – The process of GIS modelling as presented is largely based on expert knowledge of the modeller as well as knowledge of the area. More objective solutions can be obtained by optimization of the scores and weights to best fit the recharge attribute zones with the point measurements (Lubczynski & Gurwin 2004). – The pattern of net recharge interpolated by kriging is more or less similar to the extrapolated recharge pattern obtained by intergated GIS modelling of recharge potential. The differences can be attributed to inaccuracies in both methods and also to the fact that chloride mass balance gives net recharge while GIS map modelling shows potential recharge.

Water resources of arid areas

132

REFERENCES Ahmed, S., Sankaran, S. & Gupta, C.P. 1995. Variographic analysis of some hydrogeological parameters: Use of Geological soft data. Journal of Environmental Hydrology 3(2). http://www.hydroweb.com/ Allen, R.G., Pereira, L.S., Raes, D. & Smith, M. 1998. Crop evapotranspiration, guidelines for computing crop water requirements, Food and Agriculture Organisation of the United Nations, FAO Irrigation and Drainage paper 56, Rome, Italy. Allison, G.B. 1988. A review of the physical, chemical and isotopic techniques available for estimating groundwater recharge. In I.Simmers (ed.) Estimation of Natural groundwater recharge. NATO ASI series, C222, Reidel, Dordrecht: 49–72. Beekman, H.E., Selaolo, E.T. & Nijsten, G.J. 1996. Groundwater Recharge at the Fringe of the Botswana Kalahari-The Letlhkeng-Botlhapatlou Area. Botswana Journal of Earth Sciences 3. Cohen, W.B., Spies, T.A. & Bradshaw, G.A. 1990. Semi variograms of Digital Imagery for analysis of conifer canopy structure. New York: Elsevier Inc. Ecosurve. 1998. Vegetation mapping and ground truthing for Radar Imagery (Vegetation report). Serowe. Ecosurve project. Eriksson, E. & Khunakasem, V. 1969. Chloride concentration in groundwater, recharge rate and rate of deposition of chloride in the Israel coastal plain, Journal of Hydrology 7:178–197. Hendrickx, J.M.H. & Walker, G.R. 1997. Recharge from precipitation. In I.Simmers (ed.) Recharge of Phreatic Aquifers in (Semi-) Arid Areas. Rotterdam: Balkema. Gieske, A. 1992. Dynamics of groundwater recharge: A case study in the semi-arid eastern Botswana. PhD thesis, Vrije Unversiteit Amsterdam, The Netherlands. Gieske, A. 1999. Geostatistics for hydrologists, Principles and applications, Lecture notes, Adapted from de Marsily, 1986, ITC, Enschede, The Netherlands. Lerner, D.N., Issar, A.S. & Simmers, I. 1990. Groundwater Recharge: A guide to understanding and estimating natural recharge. International contributions to Hydrogeology 8. Lubczynski, M.W. 2000. Groundwater evapotranspiration—Underestimated component of groundwater balance in a semi-arid environment—Serowe case, Botswana. In Oliver Sililo et al. (eds), Groundwater: Past achievements and future challenges: 199–204. Rotterdam: Balkema. Lubczynski, M.W., Gurwin, J. 2004. Integration of various data sources for transient groundwater modelling—Sardon study case, Spain. Journal of Hydrology—in revision. Selaolo, E.T. 1998. Tracer studies and groundwater recharge assessment in the eastern fringe of the Botswana Kalahari, The Letlhakeng-Botlhapatlou Area. PhD thesis, Vrije Universiteit Amsterdam, The Netherlands. Swedish Geological Survey (SGS) 1988. Serowe Groundwater Resources Evaluation Project, Final Report, Ministry of Mineral Resources and Water Affairs, Department of Geological Survey, Lobatse, Botswana. Wellfield Consulting Services (WCS) 1998. Serowe wellfield 2 extension project (TB10/3/10/95– 96), Main report, DWA, Gaborone, Botswana. Wellfield Consulting Services (WCS) 2000. Serowe wellfield extension project, Groundwater Modelling report, DWA, Gaborone, Botswana.

Quantification of artificial ground water recharge G.C.Mishra Water Resources Development Training Centre, Indian Institute of Technology, Roorkee, India Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Sedimentary groundwater basins are mostly comprised of alternate layers of sand and clay. Occurrence of a clay layer at the top surface prevents direct recharge from rainfall. Natural and man-made surface drains in such a region are likely to carry away most part of the rainfall as direct runoff. These drains, while conveying the runoff, can be considered as a source of water for artificial recharge. Vertical shafts (recharge well) or a pit may be constructed in the bed of the surface water body through the intervening clay layer to facilitate recharge to the underlying confined aquifer. If the piezometric surface of the aquifer stands below the water level in the drain, the recharge would take place under the action of gravity. The recharge rate is governed by (i) the difference in the hydraulic heads at the water body and in the confined aquifer under the shaft, (ii) diameter and length of the shaft, (iii) transmissivity and storage coefficient of the aquifer being recharged, and (iv) hydraulic conductivity of the coarse material the shaft may be filled with.

1 INTRODUCTION Sedimentary groundwater basins are mostly comprised of alternate layers of sand and clay. Occurrence of a clay layer at the top surface prevents direct recharge from rainfall. Natural and man-made surface drains in such a region are likely to carry away most part of the rainfall as direct runoff. These drains, while conveying the runoff, can be considered as a source of water for artificial recharge. Vertical shafts (recharge well) or a pit may be constructed in the bed of the surface water body through the intervening clay layer to facilitate recharge to the underlying confined aquifer (Sandford, 1938). If the piezometric head in the aquifer stands below the water level in the drain, the recharge would take place under the action of gravity. The recharge rate is governed by (i) the difference in the hydraulic heads at the water body and in the confined aquifer under the shaft, (ii) diameter and length of the shaft, (iii) transmissivity and storage coefficient of

Water resources of arid areas

134

the aquifer being recharged, and (iv) hydraulic conductivity of the coarse material the shaft may be filled with. Appropriate methods of artificial recharge for different geohydrological conditions have been described in detail (Todd, 1985; Oaksford, 1985). In the present study, an analytical method is described to quantify time variant recharge from a surface water body to a confined aquifer through a vertical shaft or a recharge well under the action of gravity. In the paper, analytical solutions have been obtained applying unit response function coefficients known as discrete kernels and convolution technique to quantify time variant recharge from a surface water body to a confined aquifer through a vertical shaft and a recharge well. The following cases have been dealt: (i) a vertical shaft marginally penetrating into an aquifer and filled with coarse sand; (ii) a vertical shaft marginally penetrating into an aquifer; (iii) a recharge well fully penetrating an aquifer. A shaft with radius ranging from 1 to 2m filled with coarse sand can recharge at a significant rate between 250 to 700m3/day. If the shaft is not filled with sand, the rate of recharge at large time is twice that of when the shaft is filled with coarse sand. Recharge through fully penetrating well is more than 10 times that of the recharge through a vertical shaft filled with coarse sand. 2 STATEMENT OF THE PROBLEM A sedimentary groundwater basin consists of a confined aquifer overlain by an aquiclude and underlain by an impervious stratum. The aquifer is homogeneous, isotropic, and of infinite areal extent. The water level in the surface water body is at a height h1 above the bottom impervious base chosen as the datum. The thickness of the upper clay layer beneath the surface water body is L. Prior to onset of recharge, the piezometric surface in the aquifer stands at a height h2 above the datum. The height h2 is lower than h1. The aquifer can be recharged by constructing a vertical shaft in the bed of the surface water body through the intervening clay layer. The shaft may be filled with a filter material such as coarse sand to restrict contamination of groundwater. The aquifer can be recharged through a fully or partially penetrating recharge well. Quantification of recharge rate is sought for the following structures: (i) a vertical shaft marginally penetrating into the upper aquifer and filled with a coarse material; (ii) a vertical shaft marginally penetrating into the confined aquifer; (iii) a recharge well fully penetrating into the aquifer. 3 ASSUMPTIONS The assumptions made to quantify the recharge rate are:(i) the time span is discretised by time steps of uniform size ∆t (day); within each time step the recharge rate is uniform; the varying recharge is a train of pulses, (ii) an unsteady state is a succession of steady states, (iii) within a time step, Bernoulli’s equation is applicable.

Quantification of artificial ground water recharge

135

4 ANALYSIS 4.1 Case 1: A vertical shaft penetrating marginally into the aquifer and filled with a coarse material A vertical shaft penetrating marginally into an aquifer can be treated as a recharge well of zero penetration. Hantush (1961) has derived an analytical expression for evolution of piezometric surfaces in response to continuous uniform pumping from a well with zero penetration. The corresponding Hantush’s well function can be used to compute the evolution of rise in piezometric surface due to a unit pulse recharge. The response of a linear system to a unit pulse perturbation has been designated as discrete kernel coefficient (Morel-Seytoux, 1975). Morel-Seytoux and Daly (1975) have demonstrated the use of kernel coefficients in solving complex ground water flow problems. The rise in piezometric surface is expressed in terms of varying recharge and kernel coefficients derived from Hantush’s well function using a convolution technique. Let the time span be discretised by time steps of uniform size ∆t and let the time varying recharge through the shaft be treated as a train of pulses. Let R(γ) be the recharge during γth time step. Let δp(m, ∆t) be the rise in piezometric surface at the well face at time m∆t due to unit recharge (unit pulse input) that occurs during the first time step only. The expression for δp (m, ∆t), the kernel coefficient, is given in Appendix-1. The rise in piezometric surface, s(rw, m∆t), at the recharge well face at time m∆t due to variable recharge, R(γ), γ=1, 2, …, m, is given by: (1)

The hydraulic head at time m∆t at the bottom of the shaft is summation of the initial height h2 and the rise in piezometric surface, s(rw, m∆t). Applying Darcy’s law, the recharge during mth time step is given by: (2)

in which, kf=hydraulic conductivity of the coarse material the shaft is filled with. The term within the bracket is the hydraulic head difference dissipated in length L of the shaft. Solving for the recharge during the mth time step from (2)

Water resources of arid areas

136

(3)

R(m), m=1,2,…n, can be found in succession starting from m=1 to n.

Figure 1. A vertical shaft penetrating marginally into the aquifer and filled with a coarse material.

Quantification of artificial ground water recharge

137

4.2 Case 2: A vertical shaft penetrating marginally into the aquifer, with no filling material An analytical expression for recharge is derived applying Bernoulli’s equation (vide Streeter and Wylie, 1981). Accounting for the entry, exit, and friction losses and applying Bernoulli’s equation between points 1 and 2 for mth time step: (4)

(5) in which, ce=coefficient of entry loss, f=friction loss factor, L=length of the shaft, γw=unit weight of water, g=acceleration due to gravity (m/sec2), v=velocity of water in the shaft (m/sec) during mth time step, R(m) recharge volume (m3) during mth time step, ∆t=time step size (day). The first term in the right hand side of equation (4) accounts for entry loss, the second term accounts for friction loss in the shaft and the third term is the expansion loss at the exit of the shaft. s(rw, m∆t) is the rise in piezometric surface at the recharge well consequent to the recharge taken place. Incorporating (1) and (5) in (4) the following quadratic equation in R(m) is obtained: (6)

Considering the positive root of the equation (7)

in which,

For m=1, c=h2−h1. R(m), m=1, 2,…n, can be found in succession starting from m=1. The recharge rate during mth time step is equal to R(m)/∆t (m3/day).

Water resources of arid areas

138

4.3 Case 3: A recharge well fully penetrating into the aquifer The procedure for finding recharge through a fully penetrating well is same as that for the partially penetrating well described above except that the rise in piezometric surface at the well face is to be computed using discrete kernel coefficients pertaining to the fully penetrating well. The unit pulse response function coefficients δ1(m, ∆t) are obtained from the unit step response function derived by Hantush (1964) for a fully penetrating well of finite radius (Appendix II). 5 RESULTS The kernel coefficients are generated assigning values to the aquifer parameters. The thickness of the intervening clay layer is taken as 10m; the hydraulic conductivity of the packed porous medium is 10 times the hydraulic conductivity of the aquifer medium and is assumed to be 380 m/day. The friction factor f=0.02; entry loss coefficient ce=0.05; bl/rw. The corresponding non-dimensional recharge rate, R(m)/∆t T(h1−h2), with dimensionless time factor, , are presented in Figure 2. Since, the recharge rate is governed by the difference in the hydraulic heads at the entry and exit points of the shaft, and the head difference decreases with time, the recharge,

Figure 2. Variation of dimensionless recharge rate with time factor for

Quantification of artificial ground water recharge

139

Table 1. Average recharge rate for various radii; ce=0.05; f=0.02; kf=380m/day. Radius of the shaft (m) 0.1 0.2 0.4 0.5 1 2

Rate of recharge (case 1) (m3/day) 5.4213 19.8253 67.4339 97.8430 283.9961 700.755

Rate of recharge (case 2) (m3/day) 59.0824 116.8527 229.4277 284.1505 541.7252 991.7594

Rate of recharge (case 3) (m3/day) 1979.721 2078.622

therefore, decreases with time. When the shaft is filled with the coarse material, the rate of recharge at large time is half of the recharge rate that would occur without filling material in the shaft. Recharge through fully penetrating well is more than 10 times that of the recharge through a vertical shaft with a filler material. Numerical results are presented for the following aquifer parameters for various radii of the recharging structure: transmissivity, T=655.5m2/day; storativity, ; thickness of clay layer, L=10m; initial hydraulic head difference, h1−h2=5m. The average recharge rates during 120 days for different well radii are presented in Table 1. A vertical shaft with 2m radius, 10m length filled with a filter material having hydraulic conductivity of 380m/day, can recharge at an average rate of 700m3/day under an initial hydraulic head difference of 5m. 6 CONCLUSIONS Analytical methods are presented to estimate unsteady recharge, that can occur under the action of gravity, through (i) a vertical shaft filled with coarse sand, (ii) a well penetrating marginally into an aquifer, and (iii) a fully penetrating well. Application of unit response function coefficient is illustrated while quantifying the recharge rate. A vertical shaft with radius ranging from 1 to 2m filled with coarse sand can recharge at a significant rate between 250 to 700m3/day. 7 APPENDICES 7.1 Appendix I: Discrete kernel, δP (m, ∆t) Let the unit step response function for piezometric rise at the well face of a marginally penetrating recharge well and a confined aquifer system be designated as U(rw, t). According to Hantush(1961)

Water resources of arid areas

140

(1)

in which, T=transmissivity (m2/day), ф=storativity, and b1=thickness of the upper aquifer(m); rw=radius of the well or shaft(m),

and

Let the time domain be discretised by time steps of uniform size ∆t. The unit pulse response function of the system, δp(m, ∆t), is given by: (2)

W(u) and Wn(u, nπrw/b) are improper integrals as the upper limit of integration is infinite. W(u) is Theis’ Well function and can be computed using the polynomial and rational approximation (Abromwitz and Stegun, 1970) Wn(u,n πrw/b) is evaluated using Gaussian quadrature after converting the improper integral into proper integral and changing the limit. The procedure is as follows.

Quantification of artificial ground water recharge

141

As x→ −1, the value of the integrand in the second integration is found as follows:

These integration can be performed numerically using Gauss quadrature. 7.2 Appendix II: Discrete kernel, δ1 (m, ∆t) Hantush(1964) has derived the well function for computation of drawdown in an artesian aquifer due to pumping from a fully penetrating well of finite radius starting from the basic solution given by Carslaw and Jaeger (1959) for an analogous heat conduction problem. Let the unit step response function for piezometric rise at the well face of a fully penetrating recharge well and a confined aquifer system be designated as U1(rw, t). According to Hantush(1964) it is given by: (1)

in which,

Water resources of arid areas

142

functions of first kind of zero and first order respectively; Y0(x) Y1(x)=Bessel functions of second kind of zero and first order respectively; T=transmissivity (m2/day), and ø=storativity of the upper aquifer; rw=radius of the well or shaft(m). The integral in (1) is an improper integral as the upper limit of integration is infinite. The improper integral is reduced to a proper integral as described below.

Expanding the exponential term, and applying L’ Hospital’s rule, it can be shown that as v tends to −1, the integrand tends to 0. The integral I1 is a proper integral and can be evaluated numerically using Gauss quadrature.

Limit of the integrand at the lower is found as described below.

Quantification of artificial ground water recharge

Similarly,

Thus I2 can be evaluated using Gauss quadrature.

143

Water resources of arid areas

144

REFERENCES Abramowitz, M. & Stegun, I.A. 1970. Handbook of Mathematical Functions. Dover Publications, Inc, New York, 231pp. Carslaw, H.S. & Jaeger, J.C. 1959. Conduction of Heat in Solids. New York, Oxford Univ. Press: 338pp. Hantush, M.S. 1961. Drawdown around a partially penetrating well. J. Hydr. Div., ASCE, 87(HY4):83–98. Hantush, M.S. 1964. Hydraulics of wells. Advances in Hydroscience, Ed. Ven Te Chow, Vol. 1, 340pp. Morel-Seytoux, H.J. 1975. Optimal legal conjunctive operation of surface and ground water. Proc. Second World Congress. Intl. Water Resour. Assoc., New Delhi, Vol. IV:119–129. Model-Seytoux, H.J. & Daly, C.J. 1975. A discrete kernel generator for stream-aquifer studies. WaterResour. Res., 11 (2):253–260. Oaksford, E.T. 1985. Artificial Recharge: Methods, Hydraulics, and Monitoring. Artificial Recharge of Groundwater. Ed. Takashi A. Butterworth Publisher: 69–127. Sandford, H.J. 1938. Diffusing pits for recharging water into underground formation: chemical well cleaning methods. American Water Works Association Journal, 30(11):1755–1766. Todd, D.K. 1985. Groundwater Hydrology. New York, John Wiley & Sons: 458–493.

The architecture and application of the South African Groundwater Decision Tool I.Dennis & G.J.van Tonder Institute for Groundwater Studies, University of the Free State, South Africa Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Groundwater forms part of an integrated water resource and needs to be managed accordingly. Currently, however, there are limited tools available for groundwater professionals as well as water resource regulators to make informed decisions concerning groundwater use and management as part of Integrated Water Resource Management (IWRM). The South African Groundwater Decision Tool (SAGDT) was developed to assist the Department of Water Affairs and Forestry (DWAF) and Catchment Management Agencies (CMA) in decision-making with regard to aquifer protection and management in South Africa. This paper discusses the legal requirements and policies that the SAGDT complies with together with the application architecture to highlight the methodology used in the design. Finally a simple case study is given to demonstrate the application of the tool.

1 INTRODUCTION Groundwater forms part of an integrated water resource and needs to be managed accordingly. Currently, however, there are limited tools available for groundwater professionals as well as water resource regulators to make informed decisions concerning groundwater use and management as part of Integrated Water Resource Management (IWRM). Traditional water resource planners and engineers find it difficult to conceptualise groundwater hydraulics and to come to terms with the estimated impact of groundwater utilisation on surface water sources. Groundwater professionals, on the other hand, need to know at what level they have to do their investigations to satisfy the requirements of the regulator.

Water resources of arid areas

146

The purpose of the SAGDT is to assist the Department of Water Affairs and Forestry (DWAF) and Catchment Management Agencies (CMA) in decision making with regard to aquifer protection and management. The following applies to the SAGDT: ● Consists of a standard system of consistent methods/rules to guide planning and decision making about water resources. ● Allows transparency, accountability and long-term goal-setting to be incorporated into water resource management. ● Calculates the level of confidence of results obtained. This paper discusses the legal requirements and policies that the SAGDT complies with together with the application architecture to highlight the methodology used in the design. Finally a simple case study is given to demonstrate the application of the tool. 2 LEGAL REQUIREMENTS AND POLICIES The SAGDT is aligned with existing legal requirements, policies and DWAF activities which are discussed in the sections below. 2.1 Legal requirements There are 3 Acts that are taken into account, namely; Constitution of the Republic of South Africa (Act No 108 of 1996), National Water Act (Act No 36 of 1998) and Water Services Act (Act No 108 of 1997). The Water Acts are aligned with the Constitution, which states: Everybody has the right to an environment not harmful to their health and well-being, to have an environment protected for the benefit of present and future generations, and to have access to sufficient food and water. The underlying principles of the Water Acts are therefore the sustainability and protection of water resources balanced by the use thereof. The SAGDT is based on the same underlying principles namely sustainability and protection of groundwater resources (quantity and quality). Warning systems are included when the sustainability of the system is at risk. For more information concerning the warning systems, refer to Section 3.3. 2.2 Resource Directed Measures (DWAF, 1999, 2003 & 2004) To implement the National Water Act, the DWAF has initiated resource directed measures (RDM). The steps in the RDM process includes delineating the area under investigation, classifying the resource, quantifying the reserve, setting resource quality objectives, and implementing monitoring. The first four of the five can be done within the SAGDT. In addition there are different levels of investigation, namely: desktop, rapid, intermediate and comprehensive. The SAGDT provides information, methodologies and guidance to perform the various levels of assessment. A DWAF initiative Framework program for Education and Training in Water (FETWATER) focuses on training and capacity building in the water sector in South Africa. A groundwater RDM network has been developed under this initiative. One of the

The architecture and application of the South African groundwater decision tool

147

objectives of the groundwater network is to develop training material and workshop the groundwater RDM process. The SAGDT has been included in this training and the relevant components will be presented at the workshops. 2.3 National Resource Water Strategy (DWAF(a), 2003) A National Water Resource Strategy (NWRS) is currently being developed as a framework for managing water resources in this country. The NWRS aims to provide a framework balancing the sustainability and protection of water resources and the use thereof. This is the focus of the SAGDT. In addition the 19 Water Management Areas (WMAs) defined, the NWRS will be included in the SAGDT. According to the NWRS a water allocation plan must be developed for each of these WMAs, the SAGDT can assist the water planner in developing the groundwater component of these water allocation plans. 3 FUZZY LOGIC IN RISK ANALYSIS Conventional set theory (Boolean) states that an element is either a member of a set or not. Consider the following real-life problem: A person is said to be young when they are under the age of 25 and a person is said to be old when they are over 40. In which group would we place a person of the age of 30? Fuzzy logic is an extension of conventional set theory enabling an element to belong to a set to a degree. The degree of membership is a function that defines the membership of an element to a set according to the value of the element see the Figure 1. Membership is expressed as a value between 0 and 1. Zero implies 0% membership and 1 implies 100% membership. The solid line describes the membership function for the set of people older than 40 and the dotted line describes the membership function for the set of people younger than 25. Note that in most cases the membership functions of the two sets will be inverses. To answer the question of where will a person of the age of 30 fit in can be as follows: That person belongs 75% to the set of young people and 25% to the set of old people.

Water resources of arid areas

148

Figure 1. Age membership function. Table 1. Fuzzy logic rule set for 3 inputs. Rule Weight Input 1 no. 1 2 3 4 5 6 7 8

0.0 ? ? ? ? ? ? 1.0

Input 2

Input 3

Favourable Favourable Favourable Favourable Favourable Unfavourable Favourable Unfavourable Favourable Favourable Unfavourable Unfavourable Unfavourable Favourable Favourable Unfavourable Favourable Unfavourable Unfavourable Unfavourable Favourable Unfavourable Unfavourable Unfavourable

Selection of the membership function is done by an expert on the field of study. Linear membership functions are seldom used in practice in contradiction to sinusoidal functions, which are very popular. In most cases risk analysis will involve more than one input to be considered in the analysis. Fuzzy logic makes it possible to generate a set of decision rules according to the number of inputs, and these rules must then be evaluated by an expert in the field of study. The number of rules generated is given by Equation 1. n=2inputs (1) where n represents the number of rules generated. The rules consist of all possible binary combinations of the respective inputs with a weight assigned to each rule representing the risk. Table 1 shows the decision rules generated for 3 inputs. Instead of true and false the terms favourable and unfavourable are used to make the rules easier to read.

The architecture and application of the South African groundwater decision tool

149

Rule number 1 is read as: If input 1 is favourable and input 2 is favourable and input 3 is favourable then the risk is 0%. All of the other rules are read in the same fashion and an expert must evaluate each rule individually to assign the appropriate risk. For each input a membership function must be defined with a favourable and unfavourable limit defining the two sets. One function will represent the favourable set and the other the unfavourable set. Thus, for each input, a favourable and an unfavourable value can be read from the membership functions. For each input the table of decision rules is then populated with the respected favourable and unfavourable degree of membership and the risk is calculated using Equation 2: (2)

where n=number of rules DOM=degree of membership Wn=Weight of rule n Note that the minimum function must return the minimum value of all inputs for each rule. 4 SYSTEM ARCHITECTURE The sub-systems comprising the SAGDT are discussed in the sections that follow. Refer to Figure 2 as reference for the sections to follow. 4.1 3rd Party Software The 3rd party software included in the SAGDT is to provide a one-stop application to the user. The SAGDT makes provision for most user requirements without having to search for additional resources. This software will not interface automatically with the SAGDT environment, but should be seen as additional utilities provided to the user.

Water resources of arid areas

150

Figure 2. High level architecture of the South African Groundwater Decision Tool. 4.2 GIS Tool The GIS database will consist of a core set of shape files that ship with the application. Users are able to extend this database by importing custom generated shape files. New shape files added to the database will not be used in the risk assessment unless the user maps the shape file over an existing core shape file (certain criteria must be met for this operation). The GIS Tool comes with a query builder that interfaces with the whole GIS database, and users are encouraged to extend their personal database to take advantage of the query builder functionality. Using the GIS Tool provides some of the functionality of a full-blown GIS application, with a fraction of the complexity of commercial GIS systems. 4.3 Risk tool The risk tool is a fuzzy logic-based multi-criteria risk assessment tool. Two spatial environments exists in the SAGDT, i.e. the external and internal worlds as shown in Figure 2. The external world represents the whole of South Africa and the user selects a single coordinate as starting point of the scenario. The internal world then uses the specified coordinate as reference to query the GIS database. The user builds a scenario in the internal world through the use of objects available in the object repository. A simplified version of the object model is shown in Figure 3.

The architecture and application of the South African groundwater decision tool

151

In some instances more than one object performing the same function exists, and they differ only in the detail their attributes require as well as their respective confidence levels. In building a scenario, an object tree will result due to the parent child relationships of the objects used. The base object will always be a GIS object representing the area of the internal world. This object tree is then passed to the fuzzy logic engine to determine the risk assessment result, and further analysed to determine the confidence levels associated with it. A warning system exists so that the user will be notified when to do a more detailed scenario. The warning system uses the confidence level, risk assessment result and policy as inputs to

Figure 3. Simplified SAGDT object model for illustration purposes. determine if a warning should be issued. The SAGDT supports the following risk assessment categories: Sustainability, Health, and Ecological. After the completion of a risk analysis the SAGDT produces a risk profile report containing the following information: ● Description of area (area object name) ● Picture of area (internal world snapshot) ● Summary of object properties and calculated values ● Risk assessment per applicable category together with the confidence level ● Warning system response

Water resources of arid areas

152

4.4 Scenario wizard Tutorials are available in the form of a scenario wizard. The wizard gives users step by step instructions to build predefined scenarios. The wizard also guides the inexperienced user in not only learning the software, but also the methodology used in the design of the software. 4.5 Application help file The application help file is a Windows® based help file providing the user with help on the Graphical User Interface (GUI), operation and functionality of the application. 4.6 Groundwater Dictionary The Groundwater Dictionary contains terms and definitions related to groundwater illustrated by graphics, photos and animations where possible, to ensure that concepts are correctly understood. 4.7 Object help A scenario is built using available objects in the repository or library. An extensive help database is available that specifically describe each object and its functionality. This help file contains the mathematical description of each object and all associated properties, as well as interfacing with the other objects. By making this available to the user, the methodology applied in each object can be understood and an object is not just a “black box” to the user, but allows him to make educated decisions when populating the object attributes. 5 CASE STUDY 5.1 Geology The region consists of mudstones, shales and sandstones from the Adelaide Subgroup of the Beaufort Group within the greater Karoo Super Group. Post-Karoo dolerite sill intrusions are present, which have to a large extend been eroded, exposing the underlying sedimentary rocks. The surface dolerite on the Campus are highly fractured with little ground cover and it is assumed that recharge over these areas is probably high. 5.2 Study area The Campus test site is located on the grounds of the University of the Free State, South Africa, and covers an area of approximately 180×192m2. The aquifer is intersected by thirty percussion and seven core-boreholes as shown in Figure 4. Core samples indicate parallel horizontal fractures, the most significant of which is at a depth of 21m In more

The architecture and application of the South African groundwater decision tool

153

weathered sections of the aquifer, diagonal fractures intersect the bedding plane fractures. The sandstone containing the most horizontal fractures also forms the main watercarrying formation.

Figure 4. Borehole positions on the Campus test site. 5.3 Scenario The scenario that will be evaluated for this case study is the determination of the sustainable yield of UO5 when pumped continuously for 2 years. The assumptions used in the case study are as follows: ● Only UO5 will be pumped at variable rates over the assessment period. ● The assessment period is 2 years. ● Three levels of evaluation will be done, that is each successive analysis uses objects with higher confidence levels than the previous set of objects. This implies that more detailed data are needed for objects used with higher confidence levels. 5.4 Results Figure 5 shows the results obtained from the SAGDT for the specified scenario. From the graphs it is clear that the higher the confidence level of the scenario, the more accurately the risk of failing can be determined. There exists a good correlation between the 99%

Water resources of arid areas

154

risk of failure for each of the confidence levels evaluated, but the higher confidence scenarios give opportunity for better management. As a recommendation one could propose a 0.4L/s abstraction rate from the 89% confidence scenario, which indicates a 20% risk of failure. From extensive field investigations it has been proved that UO5 can be pumped for 6 months at 0.33L/s without failing, which correlates well with the proposed recommendation. It is important to note that the tool will produce an overestimate for the risk when the confidence is low. This is why a warning system was implemented to make sure the user is aware of the fact that the risk is too high according to policy and that a more detailed analysis is required to confirm the high risk situation. This will prevent users from making management decisions based on high-risk results with low confidence.

Figure 5. Risk of UO5 failing for various scenarios. 6 CONCLUSIONS The SAGDT has proven to be a powerful groundwater management tool. The tool provides a common framework for all groundwater practitioners in South Africa in which they can perform groundwater risk assessments that relate to policy. By employing fuzzy logic to do the risk analysis the user has the knowledge of an expert captured in the application to assists in the decision-making process. The SAGDT also acts as a groundwater educational environment, due to the extensive groundwater dictionary and object help files available.

The architecture and application of the South African groundwater decision tool

155

REFERENCES Constitution of the Republic of South Africa. Act No 108 of 1996, Pretoria, South Africa. DWAF 1999. Water Resources Protection Policy Implementation—Resource Directed Measures for the Protection of Water Resources Version 1.0 Volumes 2–6; Department of Water Affairs and Forestry, Pretoria. DWAF 2003. Resource Directed Measures—Module 1—Introductory module; draft edition August 2003, Department of Water Affairs and Forestry, Pretoria. DWAF 2003a. National Water Resource Strategy; current draft edition, Department of Water Affairs and Forestry, Pretoria. DWAF 2004. Groundwater resource directed measures training manual. Sponsored by Fetwater, Department of Water Affairs, Pretoria. National Water Act. Act 36 of 1998, Pretoria, South Africa. Water Services Act. Act No 108 of 1997, Pretoria, South Africa.

The development of a groundwater management tool for the Schoonspruit dolomitic compartment B.H.Usher Institute for Groundwater Studies, University of the Free State, Bloemfontein, South Africa S.Veltman Department of Water Affairs and Forestry, Geohydrology Division, Free State Region, Bloemfontein, South Africa Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: The Schoonspruit dolomitic compartment is situated in the relatively arid Northwest province of South Africa. The compartment is a critical management area since it feeds the Schoonspruit “eye”, a perennial spring feeding into a stream along which several communities and irrigation farmers abstract water. Due to large-scale groundwater abstractions and declining water levels, the area was declared as a groundwater protectorate in 1995. Water management has been devolved to local level, and in this instance a user-friendly practical management tool was needed for the Water User Association. The compartment was investigated to identify the compartment boundaries, groundwater flow directions, recharge relationships and the eye flow response over time. Use was made of long-term monitoring data, isotopic and hydrochemical data to identify recharge zones. Two zones were identified as groundwater management units in the compartment and groundwater balances for the two zones were defined. The cumulative rainfall departures method and a method of moving averages were among techniques employed to define calibrated recharge relationships for the eye. Different threshold rainfall values and recharge factors have been determined for two zones. These equations were incorporated into a simplified electronic management tool. Time-dependent inputs into this tool include rainfall, hydrochemical and water level data, while factors such as the basic human needs, reserve requirements and allocated volumes are built in. This is translated into allocable volumes for irrigation into the compartment and predicted spring flow volumes. With this tool, groundwater management is facilitated and the sustainable use of the water resources in this area can be more accurately considered.

The development of a groundwater management tool

157

1 INTRODUCTION 1.1 Overview of the study area The Schoonspruit dolomitic compartment is a dolomitic aquifer situated to the North and Northwest of the town Ventersdorp in the Northwest province. The compartment has been named after the Schoonspruit Eye, which is dependent on the compartment for flow. The Schoonspruit Eye, in turn, is the sole reason why the Schoonspruit has a constant flow and provides a municipality and two surface water irrigation boards with surface water all year. The Schoonspruit dolomitic compartment is situated north and northwest of Ventersdorp in the Northwest province. With the proclamation of The National Water Act, Act 36 of 1998, a new responsibility towards groundwater and groundwater management developed. Regional Offices were given the responsibility of managing these resources as acting Catchment Management Agencies. Groundwater is a resource that needs management and decisions based on sound scientific principles, regarding allocable volumes from the groundwater resources must be made. With the above principles in mind, the aim of the study was the development of a technical methodology and a first-order technical groundwater management tool to manage groundwater in the Schoonspruit compartment according to geohydrological principles, within a Groundwater User Association. This management tool had to be able to, on a year-to-year basis, determine the volumes available in the aquifer for allocations. 2 THE SCHOONSPRUIT COMPARTMENT 2.1 Overview The setting can be described in more detail as the compartment is categorised as Transvaal Highlands with elevation changes of more than 100m over a 40-km distance. The topography slopes downward from the northeast to the southwest. The Pretoria Formation in the north forms the water divides, in the north of the compartment, between the Vaal and Limpopo rivers. The Schoonspruit compartment falls within the surface water drainage area C24, drained by the Schoonspruit, and circular depressions can be found in the area that show elements of karstic evolution. Most of the rainfall occurs from November to February. The average rainfall for the area is 606mm and the average evaporation is in the vicinity of 1900mm. 2.2 Geology The geology of the area can best be described by differentiating between the main geological systems. In general the geology is known as dolomites of the Malmani Subgroup that plunge regionally northward and are overlain by the Pretoria Group. Outcrops of the Witwatersrand Supergroup appear along the southern boundary of the dolomites (Fleisher, 1981). The Malmani Subgroup is described as dolomite, banded iron formation, chert and shale. This series consists mostly of layered strata of calcium

Water resources of arid areas

158

magnesium carbonates (CaMgCO3), some layers massive and some with chert bands. Secondary limestone also occurs in the dolomites and is widely mined for the manufacturing of cement. Dolomites in this area are generally easily weathered and form undulating landscapes. (Kok, 1972) The Subgroup is further described as representing the dolomitic sequence and is largely concealed by overburden throughout the study area and therefore difficult to trace (Polivka, 1987). The majority of outcrops and aquifers in the area are associated with the Malmani Subgroup (Kotze, 1994). 2.3 Geohydrology The dolomitic aquifer of the Schoonspruit dolomitic compartment consists of four different formations (Polivka, 1987). Of these the chert-rich formations, Monte Christo and Eccles, are better aquifers compared to the chert-poor formations, Oaktree and Lyttleton, and boreholes drilled on fault intersections also gave high yields (>10l/s) (Kotze, 1994). The strata dip northward and are overlain by the Pretoria Group. Average borehole yields differ for the different formations and range from 11l/s in the Eccles to 3l/s in the Lyttletone formation (Polivka, 1987). Agriculture has the most important influence on the compartment’s water quality. As such, factors such as nitrate pollution are of particular importance. 2.4 Springs and water users The area is solely dependent on groundwater either directly, as abstraction from boreholes, or indirectly, as abstraction out of the Schoonspruit. The Ventersdorp Local Municipality indicates an average daily consumption of 16.3m3, including the four communities within the Ventersdorp Municipal District and on the dolomitic compartment: Ga-Mogopa, Ga-Motlatla, Goedgevonden and Tsêtsê. The largest water use within the compartment is abstraction for agricultural irrigation

Figure 1. Monthly abstraction values in the Schoonspruit dolomitic compartment.

The development of a groundwater management tool

159

Figure 2. Annual and monthly flow volumes of the Schoonspruit Eye. purposes. In the Schoonspruit river below the eye, there are four water users, claiming water rights from the Schoonspruit Eye. They are Ventersdorp municipality, the Klerksdorp Irrigation Board (IB), the Schoonspruit Governmental Water Scheme and the vlei down stream from the weir. (Darcy Consultants, 2002). Figure 1 shows the monthly abstraction volumes as used in the water balance methods. Note that these abstraction volumes include abstraction from all boreholes, spring flow volumes from the system, leakages and water consumed by evapotranspiration. There are several springs of which the Schoonspruit eye is by far the most productive, although the flow of the eye has decreased over time. This decrease is most likely associated with increased groundwater abstraction, as well as decrease in rainfall. Figure 2 shows the annual and monthly flows of the Schoonspruit Eye up to September 2002. Annual flows were calculated for hydrological years, starting October and ending in September of the following year, e.g. 1-Sep-81 would be flows for October 1980 to September 1981. It is clear from the flows that an average flow volume would not be an indication of the true situation in the compartment and it is necessary to do an analysis of the effect of rainfall and the lag-time effect of recharge through the compartment on the flow of the eye.

Water resources of arid areas

160

3 GEOHYDROLOGICAL DESCRIPTION The most important aspect for the geohydrological description was the determination of an appropriate water balance methodology for the compartment. Details regarding the verification of the compartment boundaries, current water quality and aquifer parameter determination are contained in Veltman (2003). 3.1 Recharge and water balance methods The most important input into the water balance of the area is incoming water. While some water does migrate into the compartment across the boundaries of the compartment, the vast majority of the water in this dolomitic terrain is derived from recharge from rainfall. Several recharge methods, based on methods described by van Tonder et al., 1999, were used. Recharge in the area is high due to soils that are transmissive and areas of karstification, which allow rapid recharge. Several recharge methods were used, a level of certainty assigned to each method and a weighted recharge estimate obtained. By using the weighted values, a value that is reliable for recharge can be obtained. Recharge was estimated as 6.0% of annual rainfall, amounting to the average volume of 70.68Mm3/a. 3.2 Springflow simulation Simulations of the groundwater monitoring borehole levels were done using the CRD and MA methods and the Schoonspruit Eye flow were done using MA springflow software currently being developed by Bredenkamp, 2003. 3.2.1 Groundwater simulations Groundwater levels and rainfall data are a critical input into most simulations and evaluations, of which the Saturated Volume Fluctuation (SVF), Cumulative Rainfall Departure (CRD) and Moving Average (MA) methods give the best simulations for determining natural water levels and various aquifer parameters. Where spring flows are linearly related to the CRD and MA of a rainfall series, and thus also linearly related to the groundwater system from which it is recharged, both the natural flows and the effect of abstraction can be simulated (Bredenkamp, 2000). The simulations of the groundwater monitoring borehole levels were done using the CRD and MA methods and the Schoonspruit Eye flow using MA springflow equation. 3.2.1.1 Moving average method This method mimics the groundwater level of a specific month to the average rainfall over a number of preceding months (Bredenkamp, et al., 1995) and is described by Equation 1 (Bredenkamp, 2000):

The development of a groundwater management tool

161

hi=(b/s)/n.Rfj+F (1) Where: hi=the groundwater level for month I, b=coefficient of rainfall representing recharge, s=aquifer storativity, n=number of months, Rf=average rainfall for preceding months J, F=inferred depth of aquifer below surface. 3.2.1.2 CRD method The groundwater balance can be explained by the concept that equilibrium in an aquifer is established over time between recharge (average gains) and drainage (average losses) and is expressed with Equation 2 (Bredenkamp, et al., 1995): Rfave=ROave+REave+EVTave (2)

Table 1. Aquifer parameters determined with the MA & CRD methods. Parameter

Zone A

Zone B

Aquifer thickness (m) Recharge (%) Threshold monthly rainfall (mm) High rainfall recharge factor (%) S

8 8 26

13 7.63 24.33

30

30

0.027

0.023

where: Rfave=average rainfall, ROave=average runoff, REave=average recharge, and EVTave=average evapotranspiration. The CRD method corresponds to the concept that equilibrium is established in an aquifer over time, therefore matching the groundwater level fluctuations to the cumulative rainfall departure from the average rainfall, can mimic the hydrological balance of an aquifer (Bredenkamp, et al., 1995). Defining the CRD relationship for different time intervals yields Equation 3 CRDi=CRDi−1+Rfi−k.Rfave (3) Where: CRDi=CRD at month I, CRDi−1=CRD at the month preceding month I, Rfi=rainfall at month I, Rfave=average rainfall and k=coefficient representing abstraction. 3.2.1.3 Springflow simulations When doing the CRD_MA simulations one has to incorporate inflows and outflows (abstractions) and adjust storativity and recharge values to attain the best possible curve to fit the actual groundwater level measurements that was taken. The simulation therefore

Water resources of arid areas

162

incorporates the groundwater balance to a degree, although levels and not volumes are of concern. The storativity and recharge values, attributed to the specific borehole’s reaction to averaged rainfall over an area, are obtained. In the dolomitic aquifer these values are not as related to fracture flow, since the dolomites characteristics cause variations in groundwater levels to be smoothed over an area. The values are therefore indicative of the aquifer characteristics. Further information gained from these simulations is the effective depth of the aquifer and the threshold values of rainfall before recharge will take place. Table 1 summarise the information as acquired with these simulations. The simulations provide valuable information for use in regional modelling of the aquifer, and for determining the groundwater balance. However, for proper management of the relationship between the system’s response and the flow of the Schoonspruit Eye needs refinement. 3.2.1.4 Spring flow When relating the CRD relationship to spring flow, Equation 4 can be used to simulate this relationship (Bredenkamp, 2000): QI (4) spring=J/S.ρ.CRDI+CFLOW Where: QI spring=spring flow at month I; J=hydraulic coefficient+flow cross section width constant; S=aquifer storativity; ρ=coefficient of rainfall representing recharge; CRDI=CRD at month I and CFLOW=long-term average spring flow around which the flow fluctuates. Various factors can be introduced to simulate different situations, e.g. aerial extent, abstraction or different lag time effects. When the moving average of rainfall is used, lag time effects of rainfall events can be included and its integration over the aquifer (Bredenkamp, et al., 1995). When simulating the spring flow all known parameters are incorporated and the unknown parameters are calibrated to attain the best fit. Spring flow parameters have been incorporated together with the different moving averages of rainfall. The Schoonspruit Eye simulation is shown in Figure 3 and simulated with the following equation: Schoonspruit Flow (Mm3/m)=ReN+ReF−AbsGW

The development of a groundwater management tool

163

Figure 3. Schoonspruit Eye simulated flow with a 96 month moving average. where: ReN=recharge under normal rainfall events, ReF=recharge under flood rainfall events and AbsGW=groundwater abstractions from the drainage area (Mm3/m) And ReN=ReN%/100*Rf24MMA/Rf120MMA*(Rf96MMA−ThN)*A/1000 ReF=ReF%/100*(RfFLOOD)*A/1000 RfFLOOD=IF((Rf120MMA−ThF)>0, Rf120MMA−ThF) The calibrated parameters for the system have been determined as ReN (7%), ReF%(44), ThN(26mm) which the recharge threshold and ThF(43mm) which is the flood recharge threshold. The Rf values all refer to the month lag included. The equation therefore amounts to Equation 5: Schoonspruit Flow (Mm3/m)=(0.07*Rf24MMA/Rf120MMA*(Rf96MMA−26)*0.842) (5) +(0.44*(IF((Rf120MMA−43)>0, Rf120MMA−43))*0.842)−AbsGW This equation is used in the groundwater management tool of the Schoonspruit dolomitic compartment. The biggest advantage of this method is that abstractions can now be incorporated into the simulation and predictions can be made with long-term predicted rainfall. The effective recharge for the Schoonspruit Eye was determined as 13% for 2002.

Water resources of arid areas

164

4 MANAGEMENT TOOL 4.1 A first order groundwater management tool The basic principle of a first order tool is to include the essential mechanisms in an understandable format, which will be used by the most basic groundwater manager. Inputs into the tool must be simple and outputs easily usable, while a layman should not change the driving equations. When new information becomes available, the tool should be easily modified by professionals, to include refined parameters or simulations. Developing a groundwater management tool is dependent upon the geotechnical controls essential to the management of the dolomitic compartment and which are only beneficial. Therefore the tool cannot be developed before the geohydrological evaluation has been complete and all essential controls have been defined and determined. Essential outputs from the tool include groundwater balances for various zones in the compartment, annual recharge volumes and therefore allocable volumes in the compartment, spring flow simulations for predictions from rainfall, including allocable volumes for both groundwater and surface water users and a classification of the groundwater quality based on the standards for the use of the water on the compartment. Beneficial outputs from the tool include a warning system if the Resource Quality Objectives are not met and the management class of the aquifer incorporated into the allocable volumes. 4.2 Users The users of such a tool range from the groundwater user’s association to the regulators and also groundwater consultants operating in the area. The tool needs to be versatile and contain all the necessary geohydrological equations, yet at the same time to be userfriendly. Equations for inclusion in this tool included the Schoonspruit Eye simulation equation and incorporation of domestic and ecological requirements. 4.3 Input The tool was constructed in such a way that only the latest rainfall and water quality need to be included as time-variant data. Aspects such as the reserve requirements and rainfall/recharge equations are built in. 4.4 User-friendly tool The management tool was programmed with macros which guide the user, in MS Excel. The Title page of the Tool is an information page only and a navigational button move to the next sheet, the Menu sheet. The “Enter Data” button navigates to the Data sheet, the “Compartment Map” button to the Map sheet and the “Assign/check Volumes” button to the Volume sheet. This is a navigational sheet only and the pathways are inserted here is up to the developer. The Data sheet of the Tool, allows input of the groundwater

The development of a groundwater management tool

165

quantities and qualities. Data input includes compulsory data inputs and optional data inputs. The optional data are helpful if available, but the simulations are not dependent on these cells to run. Navigational buttons to other sheets and data entry points are also included. The Drinking Water Quality Classes for the different parameters are included as fixed parameters. On the Prediction sheet of the Tool all the calculations for the simulation of the Schoonspruit Eye flow, therefore allocable volumes are done. The simulated flow (Mm3/m) is then determined using the spring flow equation as only rainfall values, and not equations, are now incorporated. Input data to this sheet is obtained from the Data sheet. Allocable volumes are determined with the simple equation of subtracting surface water demand from the simulated flow, as this has already taken into account current groundwater use. Figure 4 shows the spring flow and allocable volume graph. 5 DISCUSSION The aim of the groundwater management tool was to provide a first order technical tool, which is a practical and workable tool, for use by the WUA in determining allocable volumes. The following conclusions are made with regard to the groundwater management tool: ● Input and output parameters as outlined in this paper were used and proved to be sufficient for defining quantity and quality concerns in the Schoonspruit dolomitic compartment. ● Allocable volumes can be determined for the two zones using predicative rainfall data. ● The Schoonspruit Eye can be simulated using predicative rainfall data with the following equation:

Schoonspruit Flow (Mm /m)=(0.07*Rf24MMA/Rf120MMA*(Rf96MMA−26)*0.842) +(0.44*(IF((Rf120MMA−43)>0, Rf120MMA−43))*0.842)−AbsGW 3

Water resources of arid areas

166

Figure 4. Spring flow and allocable volume graph of the SGM tool. ● The drinking water quality classes were introduced, as a useful parameter, as part of an early warning system where drinking water quality is of concern. ● The tool is sufficient to continue with groundwater management in the dolomitic compartment. ● Verification of the lawful users is of utmost importance for groundwater management to be successful. ● The tool is a practical and useable tool for all groundwater managers and planners. The following recommendations are made with regard to the groundwater management tool: ● Groundwater management should commence at once and the tool tested against annual data. ● Verification of lawful water uses should continue and be completed as soon as possible. ● The tool should be tested and applied to other dolomitic areas.

REFERENCES Bredenkamp, D.B., Botha, L.J., Van Tonder, G.J. & Van Rensburg, H.J. 1995. Manual on quantitative estimation of groundwater recharge and aquifer storativity. Report no. TT 73/95. Water Research Commission, Pretoria. Bredenkamp, D.B. & Swartz, A. 1987. Reconstruction of the flow of springs by means of annual recharge estimates. Technical report no. GH 3525. Department of Water Affairs, Directorate Hydrology, Pretoria. DARCY Groundwater Scientists and Consultants. 2002. A catchment management plan for the Schoonspruit and Koekemoer Spruit catchments: A groundwater situation analysis. Department of Water Affairs & Forestry, Bloemfontein.

The development of a groundwater management tool

167

Fleisher, J.N.E. 1981. The geohydrology of the dolomite aquifers of the Malmani Subgroup in the SouthWestern Transvaal, Republic of South Africa. Technical report no. GH 3169. Department Water Affairs & Forestry, Directorate Hydrology, Pretoria. Kok, T.S. 1972. Wes-Transvaal en Noord-Kaap waterbeplanningstreek—geologie, fonteine en myne in opvanggebied. Technical report no. GH 1758. Department of Mines, Geological Survey, Pretoria. Kotze, J.C. 1994. Summary of the Geology, Geohydrology, and Boundaries of the proposed SGWCA, District Ventersdorp, Drainage Area C24. Technical report no. 3833. Department of Water Affairs & Forestry, Directorate Hydrology, Pretoria. National Water Act, Act No. 36 of 1998. Polivka, J. 1987. Geohydrological investigation of the Schoonspruit compartment in the dolomitic area of Ventersdorp. Technical report no. GH 3524. Department of Water Affairs, Directorate Hydrology, Pretoria. Schoeman & Vennote. 1996. Ventersdorp Oog Ondergrondse Staatswaterbeheergebied. Report no. B0307/2. Department of Water Affairs & Forestry, Sub directorate Water Allocation, Pretoria. Selaolo, E.T. 1998. Tracer Studies and Groundwater Recharge Assessment in the Eastern Fringe of the Botswana Kalahari. Ph.D. thesis, Free University of Amsterdam. GRES Project Publication. Van Tonder, G. & Xu, Y. 2001. A guide for the estimation of groundwater recharge in South Africa. The Institute of Groundwater Studies, Bloemfontein. Vegter, J.R. 2001. Groundwater development in South Africa and an introduction to the Hydrogeology of groundwater regions. Report no. TT 134/00. The Water Research Commission, Pretoria. Veltman, S. 2003. A Methodology for Groundwater Management in Dolomitic Terrains with the Schoonspruit Compartment as Pilot Area. Unpublished M.Sc thesis. University of the Free State, Bloemfontein, South Africa.

Effects of mining and urban expansion on groundwater quality in Francistown, Botswana Benjamin Mafa Department of Water Affairs, Gaborone, Botswana Horst Vogel Department of Geological Survey, Lobatse, Botswana Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: This study was carried out as part of a technical cooperation project between the Department of Geological Survey in Lobatse and the German Federal Institute for Geosciences and Natural Resources. The aim was to determine if groundwater pollution had taken place in Francistown (NE Botswana) due to urban expansion and/or historic gold mining activities, and to delineate affected areas as well as potential groundwater hazards on thematic maps which were designed in a digital and easily readable form for future development planning by urban planners. The results revealed that groundwater in Francistown had indeed become polluted through pit latrines, gold mine tailings dumps, and waste disposal sites (landfills). Different pollutants were associated with specific pollution zones. Groundwater from boreholes located within these zones was not suitable for human consumption because it exceeded certain World Health Organization and Botswana Bureau of Standards recommendations for drinking water. The study revealed that groundwater pollution due to nitrates constitutes a real health hazard and environmental and health hazards emanating from abandoned mines jeopardized human safety and environmental protection and was obvious from observed chemical “cocktail” conditions of tailings dumps and trace element concentrations in some boreholes.

1 INTRODUCTION Francistown is the oldest established town in Botswana. Born during the late 19th century as a gold mining town at the confluence of the ephemeral Tati and Ntshe sand rivers,

Effects of mining and urban expansion on groundwater quality in Francistown, Botswana

169

Francistown is the commercial hub in the NE of Botswana. The city’s rapid economic development, in particular since the 1970s, has caused its population to triple over the last three decades to approximately 100,000 inhabitants. Today Francistown is the second largest city in Botswana. In the not too distant past, water demands were entirely met by groundwater locally available from shallow alluvial and fractured volcanic rock aquifers. However, in the 1970s it was found that groundwater produced from the city’s public wells contained elevated concentrations of nitrate. In addition, the available limited groundwater resources could no longer meet the steadily rising demand for water. For these reasons public water supply was shifted in 1982 to surface water from the Shashe dam, which is located at a distance of approximately 30km to the SW of Francistown. The Shashe dam was built during the 1970s to supply the copper-nickel mine in Selebi-Phikwe. 2 GEOLOGY The prevailing dendritic drainage pattern consists of a system of irregularly branching tributaries and forms junctions at various acute angles. This is a manifestation of the complex folded and contorted metamorphosed rocks where lithological variations (in terms of weathering and erosion) are insufficient to modify this pattern. A significant portion of the Francistown study area consists of rocks of the basement complex including meta-volcanics of the so-called Tati schist group. The basement complex is divided into various granitic formations and two non-granitic lithostratigraphic units (Gibb & Partners, 1987). These are subdivided into three formations, the first of which is correlated with the Lady Mary volcanic formation. This formation consists of a homogeneous succession of dark coloured, fine-grained amphibolitic schists. The Lady Mary formation is overlain by the Penhalonga formation, which includes both metasediments and meta-volcanics. The latter are predominantly meta-andesite (greenstone) lavas, tuffs and agglomerates with amphibolite and meta-tuff beds (Key, 1976). The Selkirk formation at the top of the schist relic is laterally more restricted than the other two formations and consists of mainly felsic meta-volcanic extrusives with minor intercalations of meta-sedimentary schists. Gold mineralization in the Francistown area is mainly from quartz reefs and fissure veins of the Tati schist relic. Indeed the Tati schist relic has also been recognized for its base metal potential. Copper and nickel deposits have been identified and are now mined at the Selkirk and Phoenix mines near Matsiloje, 40km further to the SE of Francistown. Copper-zinc anomalies have also been reported near the contact between the Penhalonga and Lady Mary formations as well as in several ironstones in these formations. 3 HYDROGEOLOGY Very little detailed groundwater monitoring of the Francistown aquifers was undertaken since the first abstractions in the early 1950s and since the recommendations made by

Water resources of arid areas

170

consultants in 1974 (Colquhoun et al., 1974) and in 1979 respectively (Gibb & Partners, 1987). Groundwater consultants identified the major aquifer in Francistown as the Penhalonga mixed formation about 1.5km wide extending for at least 7km downstream of the Tati and Ntshe river confluence (Colquhoun et al., 1974). The most productive aquifers were recognized as relatively shallow discontinuous zones of fracturing. These fracture zones have a high transmissivity and draw from storage in the overlying weathered rock and alluvium. They may be up to 4m thick and are usually semi-confined by alluvial sediments and clayey weathered rock. Confining layers composed of sandy horizons contain water and contribute leakage into the underlying aquifer thereby acting as perched aquifers. Weathering appears to be confined to certain horizons within the Penhalonga mixed formation where it appears to be restricted to the easily weathered acid meta-volcanics. Indeed the river Tati is an excellent outward expression of this feature since it also follows the geological strike of this formation within these acid metavolcanics. The river tends to change its course where it traverses more competent members of the Penhalonga mixed formation. Groundwater also occurs in the sandy channels of the rivers Tati and Ntshe and this perennial baseflow component may also be regarded as an aquifer. Upstream of the confluence, the river Tati is 35 to 40m wide with the average thickness of the sand bed being 1.7m. However, sand pockets of up to 3m deep exist and increase the saturated storage of this aquifer. Downstream of this confluence, larger volumes of water can be stored since the river becomes wider with widths ranging from 20 to 100m and deeper sand beds of more than 2m in parts. 4 METHODS AND MATERIALS The study commenced with a census of all existing wells so as to establish their distribution, usage, and availability for sampling. A Garmin 40 hand-held GPS (http://www.garmin.com/) was used for coordinate acquisition. Similarly, all industries and other sites that may have a negative impact on groundwater quality were mapped. Boreholes that were found to be accessible in terms of water level measurement were used together with the topographical elevation to infer groundwater flow directions. An electrical dipper was used for water level measurements and a Trimble high-precision GPS for ground elevation measurements as well as more accurate Cartesian coordinates. The sampling of accessible boreholes involved the use of a Grundfos MP1 submersible pump (http://www.grundfos.com/) equipped with riser pipes of up to 90m. All discharge water generated while pumping was released at least 30m away from the borehole down gradient of the prevailing land slope. The method of sampling was such that electrical conductivity and groundwater reaction were measured continuously until both parameters had stabilized. Once they had stabilized a groundwater sample was taken from the particular borehole. The sample bottles were all made of plastic. Upon sampling, water reaction (pH), electrical conductivity (EC), and dissolved oxygen (DO) were measured using hand-held meters (http://www.wtw.com/). Bicarbonate determined through titration.

and carbon dioxide (CO2) were

Effects of mining and urban expansion on groundwater quality in Francistown, Botswana

171

The data obtained from the various fieldwork exercises and the hydro-chemical laboratory analyses were used to produce several environmental geology maps. For this to materialize, all data were transferred to the ArcView GIS software (Version 3.2) environment (http://www.esri.com/) where the various data layers were put together to produce the thematic maps. Data obtained from the chemical analyses were also used to deduce redox conditions, to delineate redox zones, and to determine the predominant redox processes. 5 RESULTS Out of the total of 202 boreholes that were identified during the well census, only 48 could be sampled for groundwater. All the others were inaccessible because of collapse, vandalism, or else, they had fallen dry. The vast majority of the accessible boreholes were concentrated along the two rivers Ntshe and Tati. However, groundwater yields were generally low. Several of the few known borehole yields were below 2m3/h, hence their proximity to the rivers. Only very few such as the monitoring boreholes at the abandoned and the new waste disposal site were beyond the rivers. The chemical analyses revealed that there was not much variation in groundwater reaction (pH). Most of the samples had neutral pH levels around 7, which is normal for groundwater. No groundwater sample showed acid conditions. Magnesium (Mg2+), calcium (Ca2+), and bicarbonate were the most important ions. Hence, Mg-Ca-HCO3 type of water was dominant. In some places, NaMg-Ca-HCO3 type of waters were prevalent that also featured elevated concentrations of , chlorine (Cl¯), and sulfate The concentration of total dissolved nitrate solids (TDS) was less than 1000mg/L in these particular boreholes. Over most of the built-up city area the groundwater was strongly influenced by anthropogenic activities. This was evident from TDS levels greater than 1000mg/L, and and constituted the dominant anions. In order to identify and delineate Cl¯, distinct groundwater pollution zones, all chemical groundwater parameters were used as environmental indicators and mapped individually. The concentration of oxygen allowed to identify zones with different aeration status, namely zones with aerobic (oxic) and those with anaerobic (probably reduced) groundwater conditions. This was the starting point towards defining likely pollution zones and also towards predicting redox states. In order to allow for a sound investigation it was necessary to examine the main indicator species for redox state, namely sulphate

ferrous iron (Fe2+),

nitrite and ammonium . manganous manganese (Mn2+), nitrate As was to be expected, a comparison between these species revealed that areas with high and . Equally, levels of Fe2+ and Mn2+ had at the same time low levels of areas rich in sulphates and nitrates coincided with zones high in dissolved oxygen (O2), indicating oxidizing (aerobic) conditions, while zones high in ferrous iron and

Water resources of arid areas

172

manganous manganese overlapped with zones very low in dissolved oxygen, thus indicating reduced (anaerobic) environments. The change from one zone to another was gradual. Important information in order to identify buffering systems was the presence of carbon dioxide (CO2) and bicarbonate . The distribution of these two species did not show a significant relationship to the redox state of the water. It rather was related to the calcium (Ca) and magnesium (Mg) distribution. The next parameter under consideration was the distribution of chloride (Cl¯). Due to its significant mobility, Cl¯ was meant to point to possible pollution sources. Yet, only few zones of high concentration could be identified. The next step was to seek out possible pollutants, that is heavy metals and other trace elements. All heavy metals that were detected in the study area showed distributions quite different from each other and were possibly related to mine waste sites. Zinc (Zn2+) however was not connected to mine dumps only; very strong concentrations of reduced Zn were much wider spread. The spatial distribution of the different pollutants revealed that they formed zones. Thus, once the areas with oxic (aerobic) and anoxic (anaerobic) groundwater environments and the spatial distribution of pollutants were identified, the study area could be divided into different pollution zones (Fig. 1):

Effects of mining and urban expansion on groundwater quality in Francistown, Botswana

173

Figure 1. Groundwater pollution zones in Francistown in the year 2000. Zone I—Oxidizing conditions, lots of dissolved oxygen, lots of nitrate, sulphate present, very low concentrations of ferrous iron and manganous manganese; mostly organic pollution. Zone Ia—A local disturbance within Zone I showing presence of trace elements. A small mine dump was next to this site. Zone II—High concentrations of ferrous iron and manganous manganese, enhanced zinc, no sulphate, no nitrate, ammonia present indicating very reduced conditions possibly due to organic pollution but also danger from precipitation of sulphides of different metals.

Water resources of arid areas

174

Zone III—Typical mining waste point. Elevated heavy metal and trace element concentrations, very high arsenic concentration, but also high nitrate and sulphate levels. Ferrous iron and manganous manganese strongly decreased. Zone IV—Reduced zone with high Zn2+ and Cr concentrations; Fe2+ and Mn2+ enhanced. Some trace elements present. Zone IVa—Disturbance within Zone IV with high oxygen concentration and some elevated heavy metals and trace elements; Cl¯ also highly concentrated. Zone V—Oxidised, very high sulphate and chloride concentrations; no nitrate; trace elements in significant concentrations; some heavy metals present; controlled landfill. Zone VI—Reduced, but no Fe2+ and Mn2+; Cl¯ highly concentrated; Mg2+ and Ca2+ ; some trace elements present; old and abandoned enhanced as well as CO2 and landfill. Zone VII—Oxidising conditions; no organic pollution but very diverse trace elements present (some of them concentrated); no potential source of pollution could be identified from the groundwater hazards map. 6 DISCUSSION The results of this study revealed that groundwater in Francistown had become polluted through three major sources, namely pit latrines (Zones I and II), mine tailings dumps (Zones Ia, III, IV, IVa, and VII), and waste disposal sites (Zones V and VI). 6.1 Pit latrines—Zones I and II Chemical analyses showed that nitrate concentrations (Fig. 2) well above the Botswana drinking water standard of 45mg/L (BOS, 2000) were frequent within the city area, that is zones I and II. The areas where the boreholes revealed elevated nitrate levels matched well with the areas where pit latrines where to be found. Pit latrines were located all along the river Tati throughout the built-up area. They constitute a constant source of organic pollution in the form of human excrements. This problem is made worse by the fact that pit latrines are also being used to discharge household wastewater. It is likely that a considerable amount of pollution may have been transferred into zone I from the reduced zones IV and VII upstream to the north. Zone I was characterised by a high concentration of dissolved oxygen. Nitrogen originating upstream as well as from zone I itself was probably oxidized to nitrate, which showed an extremely high concentration in the centre of the zone. Downstream the concentration of dissolved oxygen decreased. At the same time the nitrite concentration increased, probably due to denitrification. The high nitrate concentrations gradually decreased towards zone II. In contrast, the concentrations of Fe2+ and Mn2+ were very low in the centre of zone I and gradually increased towards the reducing zone II. This was typical of a redox state controlled by bacterial activity. In zone I there was a lot of organic matter input, which may have been used by bacteria as a source of carbon for the oxidation of Fe2+ (cf. Christensen et al., 1995). In the southern part of Francistown, the river Tati flows in south-easterly direction and then bends back towards the west in the middle of zone II (cf. Fig. 1). Between the two

Effects of mining and urban expansion on groundwater quality in Francistown, Botswana

175

bends ephemeral river flows are slowed down and the observed depth to the groundwater was shallower here than elsewhere in Francistown. In the crest of the second river bend there were big alluvial

Figure 2. Nitrate concentrations (mg/L) in Francistown in July/August 2000. deposits, which probably resulted in the accumulation of organic pollution and reduced groundwater conditions. At this point was hardly measurable but showed strong concentration. Fe2+ and Mn2+ were also strongly concentrated along with Zn2+. This was indicative of strong anaerobic bacterial reduction processes. On the edge of zone II towards zone III there was an old sewage pond. This location could be picked up in the form of a prolonged reduced zone characterized by lower and very high Mn2+ levels. Surprisingly though, the concentration of Fe2+ was low. From this it appeared that Mn reduction was somehow favoured over Fe reduction, which may have been controlled by the redox state of the pond (Mn needs less energy for oxidation than Fe). 6.2 Mine waste dumps—Zones Ia, III, IV, IVa, and VII Several groundwater zones were indicative of pollution due to historic gold mining activities. The strongest evidence came from the wider surroundings of the Lady Mary mine, which is located in the SE corner of the study area (zone III). Two boreholes located close to this abandoned mine site (strongly) violated international and Botswana drinking water arsenic standards, which allow for a maximum of 10 ppb (µg L−1). Yet, the groundwater in the two boreholes featured levels of 26 and 244µg (ppb) As L−1

Water resources of arid areas

176

respectively. Arsenic is very problematic in the environment because of its relative mobility over a wide range of redox conditions (Smedley & Kinniburgh, 2001). Zinc (Zn), copper (Cu), cadmium (Cd) and nickel (Ni) were also present in very high concentrations. In addition, other compounds such as cobalt, titanium, scandium, antimony, mercury, tellurium, rubidium, and thallium also showed elevated concentrations in zone III. Since zone III is situated in the most downstream spot of the study area, it is likely that all sorts of organic groundwater pollutants and products of anaerobic processes originating from zones I and II were also transferred into this area. Because the concentration of dissolved oxygen was rather high, the redox processes obviously went towards oxidation. The nitrite and nitrate levels were also rather high, which was indicative of active oxidation of ammonia that must have originated from reduced zone II. Sulphate was also very high and possibly originated from the oxidation of FeS or MnS. At the same time the concentrations of Fe2+ and Mn2+ were strongly decreased, which was probably the result of oxidation and the formation of insoluble Fe3+ or Mn4+ compounds. All this suggested very strong bacterial processes. In addition, the Cl distribution in this zone was also indicative of a site characterized by pollution input. All in all, the broad range of organic and highly toxic inorganic pollutants in zone III calls for urgent attention. Another menacing mine site is Monarch (Vogel & Kasper, 2002), which is located north of the confluence of the rivers Tati and Ntshe. Surprisingly though, zone IV did not indicate elevated levels of heavy metals or other trace elements. The only irregularities compared to the surroundings were very high O2 and elevated cobalt and silver concentrations but low nitrate and very low Fe2+ and Mn2+ levels. Possibly this was due to a combination of factors such as limited rainfall in this semi-arid environment, the fine grain-size distribution of the tailings material, and the huge size of the Monarch tailings, which may not easily provide for acid mine drainage (leaching). Rather most of the pollutants may remain in the oxidised crystal form. In contrast, a couple of areas (Ia, IV, IVa, VII), which at first had not appeared conspicuous, revealed strange irregularities in their groundwater composition. Zone IV was very reduced with a medium concentration but high Fe2+ levels. Surprisingly, it also showed a high zinc and a very high chromium concentration. Thallium, rubidium, tellurium, and cadmium were also present in increased concentrations. The distribution of Cl¯ indicated a strong pollution input upstream from this zone. The data obtained from this zone suggested that somewhere there must have been a very strong but unrecognized source of pollution, or else, the natural geological environment may have caused the formation of reduced groundwater conditions and the release of metal ions into the water. The latter was however unlikely given the granitic nature of the resident rock. The situation was similar in zone IVa. A low dissolved O2 level and therefore a low concentration, increased Fe2+ and Mn2+ but also increased arsenic, copper, selenium, beryllium, tin, caesium, yttrium and tungsten concentrations. Data from this site also revealed strong inorganic pollution even though no obvious inorganic waste source was detected. Zone VII was located north of zone IVa. Again, data showed enhanced concentrations of heavy metals but not of the elements identified in zone IVa. Because no pollution

Effects of mining and urban expansion on groundwater quality in Francistown, Botswana

177

source could be detected in these two zones it is suggested that remnants of old mine deposits may still exist in these two areas. A small mine dump within zone I (cf. Lehmann, 2001) caused raised concentrations of zircon, tantalum, hafnium, cerium, niobium, bismuth scandium and titanium and it was therefore separated out as mine waste zone Ia. Increased concentrations of Cl¯ and clearly pointed to anthropogenic pollution. The oxygen and nitrate concentrations at this site were strongly reduced but nitrite was increased. This indicated a change in bacterial populations from nitrifying to denitrifying bacteria. Given the obvious similarities in groundwater pollution between the above sites, they were put in the same pollution risk group. They may be even more hazardous than zone III since they are situated upstream from the built-up areas. Clearly, more investigations need to be carried out and immediate attention must be given. 6.3 Waste disposal sites—Zones V and VI Waste disposal sites pose an environmental hazard if they give rise to the formation of leachate plumes. The two most important factors governing the biogeochemical processes within a leachate plume are (1) the redox state, and (2) the content of the leachate. Determining the redox state of polluted groundwater is not easy. It is based on the identification of redox-sensitive species. The primary redox-sensitive species in groundwater are the dissolved ions of Fe2+, Mn2+, , , , , HS¯, the dissolved gasses CH4, N2O and O2, and also some organic substances (Christensen et al., 2001). Most of these processes are driven by bacteria and therefore slow. Bacterial populations are differentiated according to the presence (aerobs) or absence (anaerobs) of oxygen. Hence, a crucial step for this part of the study was to determine the presence of dissolved oxygen (O2) in the groundwater samples. It was obvious that the two waste disposal sites in Francistown were quite different in terms of aeration. The old landfill site (zone VI) was very poor in dissolved oxygen (O2). It is assumed that the long-lasting deposition of waste had formed the reduced environment and that anaerobic processes had probably taken place. In contrast, the new landfill site (zone V) had not yet developed a reduced zone of influence. There the concentration of O2 was quite high. In both zones, Fe2+ and Mn2+ were only present in very low concentrations. Similarly, was also only present in a very low concentration, and and were probably absent. Since no significant increase in Fe2+ and Mn2+ levels and no decrease in could be observed, and given the fact that there was only very little groundwater in both areas (in fact, during pumping one of the sampled boreholes dried up), it is assumed that the geochemistry and the redox states were not governed biologically. Bacteria need water in order to thrive. The very enhanced concentration of sulphate in zone V was probably the result of the presence of oxygenated water and the deposition of ash and building material at this site. Spreading out in a radial manner, sulphate looked like a serious problem. Very

Water resources of arid areas

178

similar pictures emanated from the spatial concentrations of rubidium, thallium, silver, uranium, molybdenum, lanthanum, zircon, titanium, sodium, bromide, and boron. may have caused The observed slightly enhanced concentrations of CO2 and the dissolution of Ca2+ and Mg2+ out of the carbonates. Probably as a result of this, the concentrations of these two cations were slightly raised (cf. Christensen et al., 2001). This could have influenced the buffering system of the sediments. Both waste disposal sites also featured high Cl¯ concentrations, though the concentration was much wider at the new (zone V) as compared to the old landfill site (zone VI). This supported the assumption that there was no new input of pollution at the old landfill site. So far, the new landfill is only used to deposit inorganic waste. Once it will be used for other kinds of waste, different processes may set in. Considering the semi-arid environment in Francistown it may be assumed that pollution at both landfill sites is localized, will not move readily from place to place, and is probably confined to the soil only. From this it would follow that the two landfills had no significant adverse effect on groundwater quality in the study area. On the other hand, natural remediation in the form of transporting pollutants to other places or through bacterial degradation is also not likely to take place. Thus pollution would probably stay as a hazard for a long time. 7 CONCLUSIONS The study highlighted that groundwater quality in Francistown had deteriorated drastically due to the influence of urban expansion and historical mining. The three dominant sources of pollution were identified as pit latrines, mine waste dumps, and waste disposal sites (landfills). However, pollution from these sources was spatially confined to those zones within which pit latrines, mine waste dumps, and landfills were located. Groundwater from boreholes located within these zones was not suitable for human consumption because it exceeded certain World Health Organization (WHO, 1998) and Botswana Bureau of Standards (BOS, 2000) recommendations for drinking water. Amongst the three pollution sources, pit latrines were found to have had the worst impact on groundwater quality. The chemical analyses of groundwater samples from a total of 48 public and private wells sampled within and around Francistown showed that nitrate concentrations were frequently well above the maximum allowable level of nitrate in drinking water. Groundwater sampled from boreholes situated in remote areas outside the city featured considerably less nitrate. In most cases the nitrate levels in remote areas outside the city were below 40mg/L, which supported the assumption that the cause of nitrate contamination was anthropogenic. Finally, the addition of nitrate through faecal waste had in turn triggered complex redox processes that had raised the ferrous iron (Fe2+) and sulphate concentrations of the groundwater. Mine dumps and/or tailings also contributed to the deterioration of groundwater quality through the addition of heavy metals, and by raising the sulphate concentration in certain zones. However, since the vast majority of the sampled boreholes were located along the rivers Tati and Ntshe and thus far away from the tailings, the real groundwater

Effects of mining and urban expansion on groundwater quality in Francistown, Botswana

179

hazards emanating from the tailings may have gone unnoticed. Clearly, further investigations are required. Amongst the three major pollutants, landfills had the least impact on groundwater quality. They are sited away from the main aquifer and within rock formations that yield little groundwater. Because of the limited rainfall in the study area, pollutants within these zones are likely to stay contained within the area. Only occasionally will they be flushed out during the rainy season and become diluted. 8 RECOMMENDATIONS Groundwater from a substantial number of boreholes was found to be not suitable for human consumption. It is therefore necessary to determine which boreholes are used for humans so as to discontinue their use. As a rule, the Francistown city council ought to adopt a development strategy that places more emphasis on an environmental approach to planning taking into account the existing water resources. For example, all new infrastructures should be placed as far away as possible from the rivers because the aquifers in the area are dependent on rainfall and river recharge. Activities such as the recent aligning of the sewage pipelines along the riverbanks must in future be avoided by all means. Such activities not only destroy a natural flood barrier but they may in fact lead to serious water pollution. Similarly, any new development must not include pit latrines. Since a sewage reticulation system has been put in place throughout the city, it is necessary to educate the residents on the need to connect to the sewerage and put an end to the use of pit latrines. So far, connection to the sewerage system is on a voluntary basis and pit latrines (and septic tanks) are currently still the main means of wastewater discharge in the newly connected areas. The study also confirmed that environmental and health hazards emanating from abandoned mine tailings must be dealt with in a way that guarantees human safety and environmental protection. The reported chemical “cocktail” conditions of tailings and the observed trace element concentrations in some boreholes make this obvious. The waste disposal (landfill) sites appeared to have been well sited in areas of low groundwater yields. But continuous monitoring is necessary in order to determine the dynamics of possible plume development so as to act upon possible groundwater pollution. Further investigations are also necessary to determine the source of heavy metals and other pollutants at the new landfill site. REFERENCES Colquhoun, B., O’Donnel, H. & Partners 1974. Redevelopment of the Francistown groundwater studies report. Phases I, II and III. Australian Groundwater Consultants. BOS 2000. Water quality—Drinking water—Specification. BOS 32, Botswana Bureau of Standards, Gaborone, Botswana. Christensen, T.H., Kjelsden, P., Bjerk, P.L., Jensen, D.L., Christensen, J.B., Baun, A. Albrechtsen, H.J. & Heron, G. 2001. Biogeochemistry of landfill leachate plumes. Applied Geochemistry: 659–718.

Water resources of arid areas

180

Gibb, A. Sir & Partners 1987. Francistown Water Development. Pre-Investment Study. Appendices B1 and B2. Water Resources. Water Utilities Corporation, Botswana. Key, R. 1976. The geology of the area around Francistown and Phikwe, Northeast and Central Districts, Botswana. District Memoir 3, 121p. plus maps, Dept. Geological Survey (DGS), Lobatse, Botswana. Lehmann, A. 2001. Conceptual map of the urban soils of Francistown. Draft map and explanations with special reference to town planning and environmental quality. Report by the Environmental Geology Division, Dept. of Geological Survey (DGS), 48p, Lobatse, Botswana. Smedley, P.K. & Kinniburgh, D.G. (2001). Source and behaviour of arsenic in natural waters. In: United Nations Synthesis Report on Arsenic in Drinking Water. Vogel, H. & Kasper, B. 2002. Mine soils on abandoned gold mine tailings in Francistown. Report by the Environmental Geology Division, Dept. of Geological Survey (DGS), 43p., Lobatse, Botswana. WHO (1998). Guidelines for drinking water quality. World Health Organization, 2nd ed., Volumes 1 and 2, Geneva, Switzerland.

In situ remediation potential for Southern African groundwater resources Sumaya Clarke, Gideon Tredoux & Pannie Engelbrecht Water Programme, Environmentek, CSIR, Stellenbosh Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: In situ groundwater remediation is practised in Europe, the United States, New Zealand and Canada. Widely accepted treatment methods include, permeable reactive barriers, redox manipulation, the Vyredox® method and biological denitrification. The permeable reactive barrier is widely used for contaminant removal. It consists of a constructed trench filled with a contaminant specific reagent, such as sawdust or wood chips for the promotion of biological denitrification. Other techniques include in situ redox manipulation which requires dithionite injection and in situ biological denitrification, which uses substrates such as ethanol or methanol. The Nitredox® and Vyredox® methods use a sophisticated arrangement of injection and aeration boreholes to manipulate oxidation and reduction to control nitrate, iron and manganese concentrations in the aquifer. Concern about nitrate as a chemical constituent of groundwater is increasing, especially in the arid and semi-arid regions of Southern Africa. Livestock losses, and “blue baby syndrome” in humans, result from high nitrate concentrations in drinking water. Hence, there is a need to remediate groundwater with nitrate concentrations above the required standard. Low cost, robust and simple treatment technologies are needed for rural water supply in Southern Africa. This paper gives an overview of the performance of full scale and pilot scale treatment plants. An estimate was made of the costs of applying selected in situ treatment options for a South African town. The order of difference in cost between in situ and ex situ treatment plants is calculated. The geological and hydrogeological parameters required for successful operation of most in situ treatment systems are described. Advantages and disadvantages of in situ treatment are also mentioned.

Water resources of arid areas

182

1 INTRODUCTION Research into groundwater remediation methods has been intensified and various techniques have been tested and applied. “Pump and treat” technologies have been widely used in the USA. The success of this method has been questioned, considering its excessive costs, (Simon et al., 2001). As an alternative, in situ technologies are being developed and applied for removing contaminants in the aquifer. Literature references to more than 100 successfully operating sites confirm that permeable reactive barriers can remove a variety of contaminants including halogenated organic compounds, metals, nitrates, acid mine drainage, phosphorous, chromium and gasoline/petrol derivatives using various reactive materials in the barrier. The currently operating sites range from household scale permeable reactive barriers to industrial sites, to mining and wastewater treatment plants and municipal well fields (Robertson and Cherry, 2003). Slowly degradable carbon sources are placed in barriers perpendicular to the flow and such treatment occurs with a high success rate. The Nitredox® plant in Vienna, Austria to treat nitrate, iron and manganese; has been operated successfully for more than a decade. Various low cost, robust treatment techniques like permeable reactive barriers and biological denitrification have proven to be successful in Canada, New Zealand, Austria, France, and the USA. Cost implication of implementing any in situ technology is important and need to be taken

Figure 1. Map showing the distribution of nitrate (as NO3) in Southern Africa.

In situ remediation potential for Southern African groundwater resources

183

into consideration. A cost estimation, performed for a town in the Northern Cape Province, South Africa demonstrates the cost difference between in situ and ex situ treatment methods. Internationally a nitrate concentration of 10mg/L as N (45mg/L as NO3) is accepted as guidelines for health risk. The maximum allowable level is set as 20mg/L as N (90mg/L as NO3) in South Africa. Nitrate concentrations in groundwater are alarmingly high in some parts of Southern Africa as shown in Figure 1. The northern provinces of South Africa all have many groundwater sources with nitrate concentrations ranging from 251– 500mg/L. In some areas concentrations up to 1000mg/L occur, while the Southern Kalahari has concentrations of up to 2000mg/L, particularly in the more saline areas, (Marais, 1999, Tredoux et al., 2000). In view of the prevalence of nitrate in groundwater, this paper focuses on in situ denitrification as a viable treatment option for town and rural applications. It is crucial that groundwater pollution be taken seriously, and that remediation and protection of the groundwater resources available be considered as a priority in countries affected by pollution. Surface water resources are limited, more particularly in arid and semi arid regions, and with groundwater being unfit for use by inhabitants of these regions, a serious threat is posed to the survival and growth of communities affected. 2 AVAILABLE DENITRIFICATION TECHNOLOGIES Many methods are successfully being employed to denitrify groundwater. Permeable Reactive Barriers and Biological Denitrification methods are the most widely used of the many methods identified, hence these methods will be discussed in further detail. 3 PERMEABLE REACTIVE BARRIERS Permeable reactive barriers (PRB) are constructed across the flow path of the migrating plume of contaminated groundwater. These systems are typically designed as a continuous trench, filled

Water resources of arid areas

184

Figure 2. Permeable reactive barrier in situ denitrification system, shown here on a rural/local scale. with permeable, reactive material. Alternatively, a funnel and gate configuration is used, which includes impermeable sections, directing the groundwater flow through the permeable treatment “gates” (Robertson and Cherry, 1995, 2000, Blowes et al., 2000, Schipper & Vojvodic-Vukovic, 2000). These treatment systems may be applied for the removal of various anions, cations, organic compounds and inorganic compounds. Configurations and system design is generally site and contaminant specific, e.g. for mitigation of nitrate at on site sanitation (see Fig 2). Requirements for the denitrification barriers (“walls”) include the following: ● The site should have a shallow water table; ● Aquifer parameters should be well understood; ● The aquifer thickness and composition should allow for constructing the wall i.e. not more than 10m deep. ● Boreholes should be placed on either side of the wall to sample groundwater to monitor chemical and microbiological changes. ● Analysis of groundwater and soil should be done prior to installation of the PRB to estimate the amount of carbon substrate required.

4 BIOLOGICAL DENITRIFICATION Biologically enhanced denitrification requires injection of readily available carbon substrates such as ethanol, methanol, sucrose and glucose to serve as a source of energy for promoting microbiological activity. Various configurations of the method may be

In situ remediation potential for Southern African groundwater resources

185

used to suit site-specific requirements. The most successful configurations are those used in Vienna and Nebraska. A list of operational systems is presented in Table 1. The daisy configuration is shown in Figure 3. Most of the biological denitrification systems use variations of this basic configuration. The area in the sketch labelled (IV) represents nitrate polluted water. The “daisy” represents the area (in plan view) progressively affected by the denitrification due to substrate (carbon source) injection. The reaction takes place in zone I, followed by filtration of any by-products in zone II while the nitrate free water (or water with a lowered nitrate concentration) is found in zone III. The Nitredox® system consists of one pumping borehole located at the centre of two concentric circles of injection boreholes. It involves injection of an organic substrate (outer ring) to enhance denitrification, but includes an additional phase of aerated water injection for the oxidation and removal of iron (inner ring) once the nitrogen is removed. The groundwater recovered from the

Table 1. Some pilot and field operational denitrification sites and their experiences. Method and Period NO3location Nmg/L

Aquifer

Carbon substrate

Injection/ barrier

Nitrate removed

PRB, Canada 5 yrs+ 5–57 Primary Sawdust/ Emplaced 58–91% (1) woodchips barrier PRB, New 5 yrs+ 5–15 Unconfined, Sawdust Emplaced 95+% Zealand (2) sandy barrier Electrokinetics/ Test Controlled Primary/ None: Abiotic Emplacement 84–87% Fe-wall, period amounts secondary of wall and USA(3) electrodes NitrEI system, Many Up to Primary/ None: Electrodes Reduced Canada (4) Currently 1000 unsaturated Electrochemical down to operating zone electrodes 0.1mgN/L sites Daisy wheel, 40 Sand and Ethanol C and P 35%-cISBD, gravel injection injection; Nebraska (5) 90–100% pinjection Nitredox, 15 years 14 Primary Ethanol P injection 75% ISBD, Vienna aquifer (6) ISBD, line of 226–565 Chalk Ethanol 80% injection boreholes, France (7) C-Continuous injection, P-pulse injection. (1) Robertson and Cherry (1995, 2000), Blowes et al., (1999), (2) Schipper & Vojvodic-Vukovic, 2001, (3) Chew and Zhang, 1998 and Loo, 2000, (4), (5) Khan & Spalding (1998), (6) Braester and Martinell (1988), Jechlinger et al., (1991), (7) Chevron et al., (1998).

Water resources of arid areas

186

Figure 3. Biological denitrification treatment system with “daisy” configuration (after Mercado, 1988). Table 2. Cost estimation of implementation of in situ denitrification compared to ex situ treatment (amounts in S A Rand). Method

PRB ISBD Ex situ Method

Capital investment Operation and maintenance per m3 Projected running expenses over 5 yrs Projected total cost over 5 yrs

61332 100289 350000 0.1 0.3 2 71144 21343 2845740 132476 313720 3195740

central production borehole is partly free of nitrate but completely free of iron, manganese or other by-products (Braester and Martinell, 1988). This method has been applied to coastal aquifers and primary aquifers. Where biological denitrification is implemented, it is important to know and monitor the permeability and porosity. Clogging may result when the carbon substrate injection exceeds the amount required for denitrification. The method has been applied mainly to primary aquifers where flow dynamics are well understood.

In situ remediation potential for Southern African groundwater resources

187

5 OPERATIONAL SITES In Southern Africa, denitrification per se is not applied. Treatment methods use expensive ex situ pump and treat systems such as desalinisation by ion exchange. These do not specifically treat nitrate and does not obtain optimum results with respect to nitrate concentrations. There is a number of test and full-scale in situ denitrification plants all over the world. Table 1 lists some of these sites, their experiences and shows the variety of configurations, carbon sources and aquifer types to which in situ denitrification has been applied. These methods are mainly applied in sand; gravel and other primary aquifer type settings, although it has been used in chalk aquifers as well. Implementation in secondary aquifer settings is said to be possible and has been modelled for sites in the UK (Cartmell et al., 1999). 6 OPERATIONAL EXPERIENCES The PRB systems in Canada have been operational for more than 5 years and are used on various scales including household, municipal, and huge water treatment plants. In New Zealand, the reactive barrier had to be replaced after 5 years of operation as reactive material had clogged parts of the aquifer. In the USA, the electrokinetic methods worked better when combined with iron walls. The biological denitrification used in Nebraska used both continuous and pulse injection regimes. The continuous carbon source injection was more efficient in denitrification but led to complete biofouling after 10 days. The system used inner oxidation ring to remove possible nitrite, iron and manganese. In Vienna, where the Nitredox® method is currently operational, clogging was experienced. Pulse injection and reduction of the amount of ethanol (carbon source) prevented clogging of the system. In France, natural in situ denitrification was carbon limited. Remediation by carbon source addition was selected to accelerate denitrification. Denitrification was achieved in long time operation (450 days). Rates of denitrification were improved when trace metals were supplied in conjunction with the carbon substrate. 7 ECONOMIC FEASIBILITY OF TREATMENT TECHNOLOGIES The costs of implementing denitrification were estimated for Marydale in the Northern Cape Province of South Africa. The results are shown in Table 2. In situ application was compared with a conventional ex situ method. In Marydale, a well field containing 10 boreholes is used as the town water supply. Half of these boreholes, produce groundwater of nitrate concentration above the maximum allowable 20mg/L, (Hofmann, 1997). Microbiological sampling showed that coliform counts of 15/100ml in some boreholes were three times as much as the SABS specification (Hofmann, 1997). No faecal

Water resources of arid areas

188

coliforms were detected. This gives an indication that no human or animal waste reaches the boreholes. Exploration boreholes drilled in the area revealed that a primary and a secondary aquifer are present in the area. The secondary fracture system is not well understood, but it is believed that the bed rock is not very permeable. The main water bearing unit is the alluvial cover of more or less 12m thick. The aquifer material consists of sedimentary layers containing primarily sand and silt. The Projected water demand of the town for 2005 is 142287m3/a (Shand and VSA, 1997). Capital expenses for permeable reactive barriers (PRB) include excavation costs, wall emplacement costs and dewatering prior to wall emplacement. Woodchips or sawdust was considered as suitable permeable reactive barrier material as these are cheap and slowly degradable carbon sources. The barrier size is based on the size of the well field and the depth to bedrock. The largest contribution to capital costs for in situ biological denitrification (ISBD) methods include borehole construction costs, purchasing of injections pumps among other costs. Erecting infrastructure is a major expense for conventional treatment plants. Running costs were based on the projected annual water demand and estimated chemical costs. The PRB method requires limited maintenance. In the calculation, operation and running costs are included; however, they may not occur frequently for methods like PRB. Operational and maintenance costs are relevant especially when clogging or partial clogging of wells occurs. Pump and treat methods and other ex situ methods (in this case ion-exchange) generally cost an order of magnitude more than in situ methods. It is clear from this information that rural communities for which funding is not always in surplus may capitalize on this advantage as well as the ease of use of some of these methods. Proper management and monitoring of sites are essential to detect potential clogging cases early and to put remedial measures in place. 8 DISCUSSION Field scale plants have proven in situ technologies to be successful. The nitrate removal rate at most currently operating sites are high. It is evident from Table 2 that the permeable reactive barrier method is the most cost effective method. Capital costs are relatively low and it requires little or no additional treatment of groundwater after passing through the system. Installation and running costs of ex situ treatment exceed that of in situ methods. Operational sites in the US, Canada and New Zealand showed that barrier material replacement was only required after 5 years, while ex situ methods have set running expenses per cubic meter of water. The largest full-scale in situ denitrification plant uses the Nitredox® principle. This plant is located at Bisamberg, Vienna (Austria) and has been operating successfully for more than a decade (Jechlinger et al., 1991). It uses ethanol as the carbon substrate and the process is regulated to ensure that the raw water nitrate, which exceeds 65mg/L, is reduced to approximately 35mg/L in the product water. There are advantages as well as some disadvantages of in situ treatment technologies. Some advantages of implementing such a treatment system include minimal exposure to

In situ remediation potential for Southern African groundwater resources

189

dangerous chemicals, job creation in rural villages, little or no treatment required at the surface, possible treatment of other contaminants due to redox changes, costs savings in comparison to conventional ex situ treatment plants, low maintenance costs, simple to use technologies, and no need for electricity (PRB- method). Disadvantages include possibilities of clogging of boreholes. This occurs when the carbon dosage is in excess of the required amount. Sulphate reduction may occur when carbon substrate dosage is too high and result in acetate production as a by-product of microbial activity (Israel, 2004, unpublished data). Loss of hydraulic permeability of the aquifer is possible if carbon addition is not effectively managed. Preferential flow of groundwater can occur, where a great contrast develops between the treatment zone and the rest of the aquifer and the path of least resistance is taken by the groundwater. Hence monitoring of the above mentioned parameters is very important. Cautionary measures include proper estimation of the required amount of carbon substrate, and monitoring the effective porosity and permeability of the aquifer before, after and during treatment. Management of implementation and monitoring are essential for success. It is important to note that no microbes are added to initiate the process, as this would affect the ecosystems that are already established at any specific site. There are many strains of bacteria that occur naturally under the various environmental conditions, which are capable of denitrification. Although some scientists may prefer to add appropriate bacteria to initiate the process but the addition of a carbon source is sufficient to activate resident bacteria. 9 CONCLUSION In situ groundwater treatment methods are widely used and accepted in the US, Canada, Europe and New Zealand. Literature shows that various in situ methods for a range of heavy metals, organic compounds and other constituents have been successfully implemented at field scale in these countries. In situ denitrification methods are also viable treatment methods which are successfully implemented and currently operating. The cost analysis performed in South Africa, showed that there is an order of magnitude difference between the costs of ex situ and in situ treatment plants. Optimal conditions for most in situ treatment methods include the following: ● Primary aquifer systems, or well understood secondary aquifers (with respect to flow characteristics and porosity/permeability). ● Aquifer material can include sand, gravel, and chalk material. ● A known concentration of nitrate-nitrogen is important for estimation of the appropriate quantity of carbon substrate. ● A maximum aquifer thickness of 20m for injection type methods (e.g. ISBD) and 10m for emplacement methods (e.g. PRB). ● Monitoring of aquifer parameters (permeability, hydraulic conductivity etc.), chemical changes (pH, Eh, etc.) and microbiological changes with time. It is important to note that no foreign microbes are added, as this would affect the ecosystems that are already established at any specific site. Although some scientists may

Water resources of arid areas

190

prefer to add appropriate bacteria, addition of a carbon source is sufficient to activate resident bacteria. REFERENCES Blowes, D.W., Ptacek, C.J., Benner, S.G., McRae, C.W.T., Bennett, T.A. & Puls, R.W. 2000. Treatment of inorganic contaminants using permeable reactive barriers. Jnl. of Contaminant Hydrology, 45:123–137. Braester, C. & Martinell, R. 1988. The Vyredox and Nitredox method in situ treatment of groundwater, Wat. Sci. Tech., 20(3):149–163. Cartmell, E., Clark, L., Oakes, D., Smith, S. & Tomkins, J. 1999. Feasibility of In situ Bioremedition of Nitrate in Aquifer systems, R & D Technical Report P277, WRC report no. EA 4683. Chevron, F, Lecomte, P., Darmendrail, D. & Charbonnier, P. 1998. Rehabilitation de qualitè physicochimique d’un aquifere contaminepar des nitrates d’origine industrielle- un example en region Nord-Pas de Calais. L’Eau, L’Industrie, Les Nuisances, 208(31–35) (In French). Chew, C.F. & Zhang, T.C. 1998. In situ remediation of nitrate contaminated ground water by electrokinetics/ iron wall process. Water Science and Technology, 38(7):135–142. EPA, 1995. In situ remediation technology status report: Treatment walls. Report No. EPA/540/K94/004. Office of Solid Waste and Emergency Response, US Environmental Protection Agency. Israel, S., 2004, Subsurface Manipulation of the Nitrogen Cycle: In-Situ denitrification and its potential for remediation of contaminated soil and ground water resources: Case Study: Marydale, Northern Cape, MSc research, unpublished data, University of Stellenbosch. Jechlinger, G., Schöller, F., Seidelberger, F., & Zibuschka, F. 1991. Denitrification In Situ. In: Proc. of I.W.S.A workshop: Inorganic nitrogen compounds and water supply. Hamburg, 27–29 Nov:113–122. Khan, I.A., & Spalding, R.F. 1998. Denitrification using a daisy well system. Presentation to National Sanitation Foundation International Symposium, Safe Drinking Water in Small Systems: Technology, Operations, and Econimics. Washington D.C., May 10–13. Kruithof, J.C., Van Paasen, J.A.M., Hijnen, W.A.M., Dierx, H.A.L. & Van Bennekom, C.A. 1985. Experiences with nitrate removal in the eastern Netherlands. Proc. Nitrates Dans les Eaux, Paris 22–24 October. Loo, W.W. 2000. Electrokinetic treatment of hazardous wastes. Standard Encyclopedia of Environmental Science and Technology, New York, McGraw Hill,: 14.69–14.84. Mercado, A., Libhaber, M. & Soares, M.I.M. 1988. In situ biological groundwater. denitrification: Concepts and preliminary field test. Wat. Sci. Tech., 20(3):197–209. Ninham Shand and VSA Consulting, 1997, Geohidrologiese Ondersoek van die Groundwaterbronne by Marydale, Noord-Kaap Provinsie, VSA Conculting pty. Ltd. Robertson, W.D. & Cherry, J.A. 1995. In situ denitrification of septic system nitrate using reactive porous media barriers: Field trials. Ground Water, 33(1):99–111. Robertson, W.D., Ford, G. & Lombardo, P. 2003. Wood-Based Filter for Nitrogen Removal in Septic Systems, (Submitted to: Journal of Environmental Quality), (unpublished). Schipper, L.A. & Vojvodic-Vukovic, M. 2000. Nitrate removal from groundwater and denitrification rates in a porous treatment wall amended with sawdust. Ecol Engineering, 14:269–278. Schipper, L.A. & Vojvodic-Vukovic, M. 2001. Five years of nitrate removal, denitrification and carbon dynamics in a denitrification wall. Wat. Res. Research, 35(14):3473–3477. Tredoux, G, Talma, A.S. & Engelbrecht, J.F.P. 2000. The increasing nitrate hazard in groundwater in the rural areas. Paper presented at WISA 2000, Sun City, RSA, May 2000.

Coastal aquifers intrusion at semi-arid region of Turkey L.Yilmaz Technical University of Istanbul, Civil Engineering, Hydraulic Division, Maslak, Istanbul, Turkey Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Coastal aquifers are important sources of water for domestic, agricultural and industrial uses. Under natural conditions the hydraulic gradient is towards the sea so that there is a natural outflow of fresh groundwater. Frequently the hydraulic gradient is small. Therefore very little extraneous activity is required to disturb the natural system and cause the fresh water to become. This situation poses a difficult management problem which is best addressed by means of mathematical models. This research describes the use of such models together with the difficulties likely to be encountered.

1 INTRODUCTION 1.1 Relationship between the level of the water table and the depth to the saline wedge A relation between the level of the water table and the depth to the saline wedge in an unconfined aquifer under steady conditions of flow (Badon Ghijben, 1889; Herzberg, 1901, Davis, 1978) points out that Joseph DuCommun (1828) made similar observations. Prior to the work of these pioneers it was thought that saline water occurred at a depth close to sea level. The saline water close to the sea shore is defined by Badon GhijbenHerzberg equation, which was derived by a simple application of hydrostatics. The weight of a column of fresh water of height hf+z is equal to the weight of a column of saline water of height z. If rof and ros are the densities of fresh and saline water respectively, it is given in equilibrium conditions rosgz=rofg(hf+z) (1) or (2) If the relative densities of fresh and saline water are taken as 1.0 and 1.025 respectively, then

Water resources of arid areas

192

z=40 hf (3) This expression is a good approximation in steady state conditions when the zone of dispersion is only a small fraction of the saturated thickness of the aquifer. Since fresh water is flowing along the interface some mixing will occur due principally to microscopic and macroscopic dispersion. When the saline and fresh water mix in the zone of dispersion then the diluted saline water becomes less dense and will rise along a seaward path. The resulting mechanism is similar to thermal convection, the only difference being that the gradients are caused by changes in density due to changes in salinity instead of temperature (Cooper, 1964). This flow will advect some saline water towards the sea. Therefore, in order to preserve the saline mass balance, a small flow of saline water must occur in the landward direction. This flow creates a head loss, thus a reduction in pressure at the interface and a reduction in the level of the interface. This application gives the position and movement of a saline front in a coastal aquifer. 1.2 Groundwater flow equation Using the Darcy’s law and the law of conservation of mass to a control volume, (Pinder and Bredehoeft, 1968; Konikow and Bredehoeft, 1978) gives (4) where Tij=transmissivity tensor (L2T−1) h=hydraulic head (L) above a reference point S=storage coefficient (−) W=source or sink volume flow term (LT−1), positive for outflow (=W (xi, t), i=1, 2. This is usually the recharge, pumping and evapotranspiration). xi, Xj=Cartesian coordinates (L) t=time (T) The advection-dispersion equation is given using the notation of Konikow and Bredehoeft (1978) as; ∂(Cb)/∂t=∂/∂xi(bDij∂C/∂xj)−∂/∂xi(bCVi)−C′W/ε (5) where Dij=coefficient of hydrodynamic dispersion (L2T−1) Vi=seepage velocity in the direction xi (LT−1) C=concentration of the pollutant (ML−3) C′=concentration of the pollutant in the source or sink fluid (ML−3) b=saturated thickness of the aquifer (L) E=effective porosity of the porous medium (−) This equation gives the change in chemical concentration due to kinematic dispersion and diffusion, the effect of advective transport and the removal of pollutant due to fluid sources and sinks.

Coastal aquifers intrusion at semi-arid region of Turkey

193

2 MAIN OBJECTIVE 2.1 Solutions of the groundwater equations Therefore the interface is not sharp and a mixing zone exists, the thickness of which depends upon the hydrodynamics of the aquifer. If this transition zone is only a small fraction of the saturated thickness of the aquifer then the assumption of a sharp interface is reasonable and a good mathematical description of the shape of the saline wedge can be obtained. The thickness of the fresh water wedge decreases in the seaward direction and the slope of the water table steepens towards the coast. Therefore the shape of the interface is concave upwards. If the more realistic view is taken that the fresh and saline water are miscible, then the interface cannot be sharp and the mathematical description of the problem becomes more complicated. The assumption of a sharp interface cannot be considered reasonable if the flow situation varies with time since the hydrostatic pressure distribution will vary and the assumed interface will move either landwards or seawards. This results in the sharp interface being replaced by a zone of dispersion in which the salinity of the water varies from fresh to very saline. Clearly the simplifying assumption of a sharp interface makes for a mathematically simpler but less accurate model. 2.2 Sharp interface models The relationship between groundwater levels and the depth to the saline wedge is given by Badon Ghijben and Herzberg who, working independently, developed a relationship between the level of the water table and the depth to the saline wedge in an unconfined aquifer under steady conditions of flow (Badon Ghijben, 1889; Herzberg, 1901). Davis (1978) points out that Joseph DuCommun (1828) made similar observations. Prior to the work of these pioneers it was thought that saline water occurred at a depth close to sea level. In contemporary practice the result is always referred to as the Badon GhijbenHerzberg equation. This equation is derived by a simple application of hydrostatics. Since the interface is stationary then the weight of a fresh water above the interface is exactly balanced by the pressure of the saline water below the interface. By consideration of the Figure it can be seen that the weight of a column of fresh water of height hf+z is equal to the weight of a column of saline water of height z. If rof and ros are the densities of fresh and saline water respectively, then for equilibrium rosgz=rosg(hf+z) (5) If the relative densities of fresh and saline water are taken as 1.0 and 1.025 respectively, then z=40 hf (6) This simple expression gives a remarkably good first approximation to the depth below sea level of the interface under steady state conditions when the zone of dispersion is only a small fraction of the saturated thickness of the aquifer.

Water resources of arid areas

194

This concept can be further developed to determine the extent of the penetration of the saline wedge inland. Many analyses can readily be developed, for example the determination of the shape of the interface when the seepage surface is submerged beneath the sea (Glover, 1964) and the shape of the saline upcone beneath a pumping well in a coastal aquifer (Schmorak and Mercado, 1969; Sahni, 1972). 2.3 A sharp interface with some mixing When a sharp interface is assumed, then this interface is a flowline in the same way as the water table is a flowline. Hence it is a boundary condition for the problem. Since fresh water is flowing along the interface some mixing will occur due principally to microscopic and macroscopic dispersion. When the saline and fresh water mix in the zone of dispersion then the diluted saline water becomes less dense and will rise along a seaward path. The resulting mechanism is similar to thermal convection, the only difference being that the gradients are caused by changes in density due to changes in salinity instead of temperature (Cooper, 1964). This flow will advect some saline water towards the sea. Therefore, in order to preserve the saline mass balance, a small flow of saline water must occur in the landward direction. This flow creates a head loss, thus a reduction in pressure at the interface and a reduction in the level of the interface as predicted from the Badon Ghijben-Herzberg equation. The mechanism is shown in Figure. It is possible to extend these concepts to determine the solutions to various moving interface problems. However, except for some very restrictive cases, analytical solutions do not exist. Hence numerical methods are required to solve the resulting equations, which usually means that it is more convenient to use commercially available groundwater quality models. Approximate solutions for moving interface problems, including numerical ones, are discussed by Bear (1979). 2.3.1 Equations of groundwater flow and advection—dispersion The above section dealt with some very simple first approximations for determining the position and movement of a saline front in a coastal aquifer. Whilst these are useful in the early stages of a study they do not permit a full solution to the majority of aquifer problems. For example they cannot deal with spatial variations of geology or aquifer parameters nor with multi-layered aquifers. In order to achieve this two equations are required, one to describe the groundwater flow and one to describe the movement of the salt. These will be considered in turn. 2.4 Solutions of the groundwater equations The finite difference method is used for solving this type of differential equations. The first step is to give the area of the model in mathematical terms, which is called the solution domain. This solution domain is covered by a rectangular grid which can be either regular or irregular. The differential equation is replaced by a set of difference equations, one for each grid point. This results in n×m simultaneous equations which have to be solved, where n is the number of rows and m is the number of columns of the

Coastal aquifers intrusion at semi-arid region of Turkey

195

grid. The finite difference method is, perhaps, the most frequently used technique for solving the flow equation. However, it is not often used to solve the advection-dispersion equation because of a phenomenon known as numerical dispersion. The numerical solution usually appears to advance the solute at a rate which is greater than is physically possible. Finite difference schemes can be developed to minimize dispersion. However they are liable to cause either overshooting or undershooting which appear in the solution as oscillations. Van Genuchten (1976) analyzed and gave as a result that the finite element schemes will usually yield more accurate solutions than finite difference ones. There are some rules which can be helpful in minimizing the effects of dispersion. These use a form of the Peclet number, Pe, and the Courant number, C. The grid should be designed such that Pe(=dx/De)<4, where dx is some characteristic grid size and De is some characteristic dispersivity. The finite element method was first developed in the solid mechanics (1950). Then it was used to solve the groundwater flow equation. The finite element method is an integral (as opposed to differential) approach in which the regular grid of the standard finite difference method is replaced by an irregular polygonal mesh which allows the modeller to describe natural shapes more precisely. In groundwater the polygonal shape is, almost always triangular. The finite element mesh can be adapted to describe the irregular shape of the boundary and obtained an accurate description of rapidly varying phenomena. In this approach the piezometric surface is approximated by a series of small triangular surfaces which can be flat or curved. If the chosen basic functions are linear then the surfaces will be flat and the variation of head within each element will be linear. The point of intersection of the triangles is called a node and each triangle is called an element. The equation is solved by a weighted residual technique of which the most popular one is the Galerkin method. In this method the weighting functions are made equal to the basic functions and the integration is then performed over each element and summed to yield the contribution from all the elements that make up the solution domain. The finite element method is a powerful and mathematically elegant technique but it is difficult to program. The method of characteristics was developed to solve hyperbolic partial differential equations (Courant and Friedrichs, 1948) and was first used for flow through porous media by Gardner et al. (1964). They proposed the method because they argued that when flow velocities become large the dispersion equation is, essentially, hyperbolic. The method has been extensively applied for solving the advection-dispersion equation and is now the basis of one of the standard solute transport models (Konikow and Bredehoeft, 1978). The solutions are x=x(t), y=y(t) and C=C(t), where x and y are the coordinates in a Cartesian system, C is the pollutant concentration and t is time. These are called the characteristic curves of, in this case, the advection-dispersion equation. Once these solutions are available then a solution of the advection-dispersion equation can be obtained by following the characteristic curves. Gardner et al. (1964) state that “Each point corresponds to one characteristic curve and values of x, y and C are obtained as functions of t for each characteristic”. Essentially this is the Lagrangian approach of classical hydrodynamics. It is particularly useful for making cross-sectional models of saline intrusion. There are many other methods for solving the groundwater flow and advectiondispersion equations, as integrated finite differences (Tyson and Weber, 1963; Goodwill,

Water resources of arid areas

196

1980), boundary element methods (Liggett and Liu, 1983) and analytic elements (Strack, 1989). 3 RESULTS The model requires substantial amounts of field data, the collection of which is both time consuming and expensive. The hydrological and geological data are used for the area to be modeled. These data are: ● Surface and subsurface geology. ● Piezometric levels for all the aquifers contained in the system. ● Aquifer characteristics and likely boundaries, soils, land use and vegetations. Since aquifers are subject to recharge and pumping, data on the quantities and timing of these will be required. Such data will include precipitation, evapotranspiration and pumping. If irrigation is undertaken, then also rates of application and return flows and river flows are required, including flows to and from the rivers to the aquifers if they are not in direct hydraulic contact. Since the concern here is with saline intrusion then data on salinity, both areally and vertically, will be required. If any of these data do not exist or are too scanty, then a field programme will be required to collect them. In order to collect and plot all these data an accurate topological map is essential, the scale of which will depend on the size of the aquifer and the scale of the problem being studied. This map should show the surface contours, surface water bodies, streams and man-made watercourses such as irrigation canals and drainage ditches. REFERENCES Bear, J. 1972. Hydraulics of Groundwater. New York, McGraw-Hill Book Co., 567 p. Bras, R.L. & Rodriguez-Iturbe, I. 1976. Evaluation of mean square error involved in approximating the areal average of a rainfall event by a discrete summation, Water Resources Research, 12(2), 181–184, a. Hubbert, M.K. 1940. The Theory of Ground-Water Motion, The Journal of Geology, 48(8), Part-I, Nov.–Dec. Pinder, G.F. 1982. Finite Element Simulation in Surface and Subsurface Hydrology, Gallagher, Vol. 4. Pinder, G.F. & Abriola, L.M. 1982. Calculation of Velocity in three space dimensions from hydraulic head measurements, Groundwater, 20, 205–213. Pinder, G.F. & Gray, W. 1982. Finite Elements in Water Resources, edited by P.Holz, V.Meissner & C.A.Brebbia (eds), Berlin, Springer Verlag: 4. Rushton, K.R. & Redshaw, S.C. 1979. Seepage and Groundwater Flow, Wiley, Winchester, UK, 339 pp. Sarma, S.V.K. & Silva, T.C. 1987. Hydraulic response to pumping in free aquifers, ABAS, 11, 26– 32. Sarma, K.V.S. & Antonio, A.P. 1997. Decontamination of pollutants from aquifers using the concept of induced flow from adjacent rivers. Intl Conf. on Large Scale Water Resources Projects, Oct. 20–23. Kathmandu, Nepal, EI 17–24.

Coastal aquifers intrusion at semi-arid region of Turkey

197

Sumer, B. 1980. The Determination of Water Quality at the Sapanca Lake, TUBITAK Project No. QA6–4, Sakarya. U.S. Environmental Protection Agency, 1988. Model Assessment for Delineating Wellhead Protection Areas, Office of Groundwater Protection, Washington DC, 210 pp. Yilmaz, L., Agiralioglu, N. & Saltabas, L. 1999. The determination of the water-use capacity of the Sapanca Lake in Turkey. Proc. Intl. Conf. on Water, Environment, Ecology, Socio-economics and Health Engineering (WEESHE), Oct. 18–21, Seoul National University, Seoul, Korea, Water Resources Pubs., LLC, 162–166.

Evaluation of groundwater recharge rates in the Kizinga catchment in Dar es Salaam region Y.B.Mkwizu Lawyers’ Environmental Action Team (LEAT), Dar es Salaam, Tanzania H.H.Nkotagu Department of Geology, University of Dar es Salaam, Dar es Salaam, Tanzania Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: This paper focuses on the need to manager groundwater exploitation by comparing its abstraction rates with natural recharge rates and a case study was done in the Kizinga catchment. Despite the increase in groundwater exploitation, there is no comprehensive legal mechanism to ensure proper management of groundwater resources and no specific provision for groundwater abstraction under The Water Utilization Act. Instead groundwater exploitation management has been provided for in the same way as surface water. The Act has not demanded operators of wells or boreholes to submit data and records to water authorities. It is therefore possible that substantial commercial drilling of groundwater has been conducted without adequate monitoring and controls. It is suggested the proper management of groundwater utilization be established to balance recharge and discharge. This will ensure that groundwater abstraction is done sustainably and thus avoid negative consequences of groundwater depletion.

1 INTRODUCTION The high demand for freshwater in the Dar es Salaam City, suggests clearly that, surface water can no longer meet the projected total demand. The second best alternative remains on groundwater. Even with the presence of surface water, groundwater can still be preferred on the basis of easier protection from pollution, better dependability during drought periods, and on the supplying costs. It is necessary however that groundwater resource is used with proper management focusing on both its quality and quantity.

Evaluation of groundwater recharge rates in the Kizinga catchment in Dar es Salaam region

199

Evaluation of groundwater recharge rates and areas is an important step towards thorough understanding of its quality and quantity. 1.1 Location and climate of the study area The study area is located to the south of Dar es Salaam region bearing geographical coordinates 39°02E and 39˚18E and 6°50S and 7°00S. It is built on a low lying coastal plain with an altitude varying between 20m to 240m above sea level in Kurasini area and in Pugu and Kisarawe hills respectively with a total surface area of 191km2 (Service Plan 1997). The study area is drained by Kizinga river having its upper reaches in Pugu and Kisarawe hills. The river flows in a NE direction towards Indian Ocean (Fig. 1). Climatically the daily temperature ranges from 18.1°C to 32.2°C with a mean value of 25.75°C, while actual evaporation has a mean monthly value of 160mm ranging from 128mm to 181mm. The mean annual precipitation is 1124mm.

Figure 1. Location of the study area. 1.2 Objectives The purpose of the study was to evaluate groundwater recharge rates in the area. Specifically, the research aimed at: Determining the mean annual groundwater recharge rates in the study area.

To determine the source of groundwater recharge in the study area.

Water resources of arid areas

200

1.3 Groundwater evaluation in the study area Very few studies on the behavior of groundwater resource have been documented in the study area. The study on Coast/Dar es Salaam Water Master Plan by the Ministry of Energy and Minerals (MEM) (1979) for example, found that the coastal sedimentary deposits of Coast and Dar es Salaam regions did not include aquifers that provide major groundwater supplies. The project revealed further that within the area; almost seventy per cent of the area was found to be underlain by material that could yield water of insufficient quantity or inadequate quality. On another study, Matondo (1978), found that the Kizinga basin is a potential source of groundwater. He also noted that groundwater from drilling is very close to the ground surface. He further found that, only 8% of annual rainfall appears as total runoff while the major part of the rainfall volume was stored within the aquifer. Both studies didn’t make evaluation on groundwater recharge rates. 2 METHODOLOGY Data that incorporate various hydrogeological units and taking into account all flow components such as discharge, infiltration, subsurface inflow to and outflow from the basin’s aquifer and abstractions through pumped wells were collected. The main data/and data sources were: – Boreholes and wells and their hydrologic information drilled in the study area up to 1999 from the Ministry of Water, borehole drilling unit-Ubungo – Monthly rainfall data, maximum and minimum temperature and evaporation for Dar es Salaam from Tanzania Meteorological Agency, Dar es Salaam office – Runoff measurements from rivers Kizinga from 1967 to 1980, From Ministry of Water, RBM-Ubungo

Table 1. Chloride concentration of rainwater in the study area in (mg/l). Location Ukonga Kiwalani Yombo Temeke Mbagala Pugu Gongolamboto Total Mean

Mar Apr May Mean 3.1 2.8 3.0 1.9 3.2 2.4 3.7 2.6 3.6 2.8 3.2 1.8 3.0 2.8 22.8 17.1 3.26 2.44

2.6 3.9 2.8 3.4 3.0 2.4 3.2 21.0 3.04

2.83 2.93 2.80 3.23 3.13 2.47 3.00 20.39 2.91

Evaluation of groundwater recharge rates in the Kizinga catchment in Dar es Salaam region

201

Figure 2. The groundwater flow net in the study area. Abbreviation Site Pg Gt Uk Airport Kw Tr Ym Ch Tm Mb Mt Kr

Mean hydraulic head (m)

Pugu Gongolamboto Ukonga Kiwalani Tazara Yombo Chang’ombe Temeke Mbagala Mtoni Kurasini

224 63.5 43.2 37.4 22.2 20.1 14.1 12.6 10.9 11.2 9.0 6.3

The fieldwork involved rainwater sampling in different sites within the study area for the rainfall period of March, April and May in the year 2000. Another fieldwork activity was to collect water from boreholes in some selected locations in the study area. Chloride determination from water samples was undertaken following standard method as reported by (APHA 1985). The results obtained from rainwater samples are given in Table 1. 3 RESULTS 3.1 Groundwater resource accumulation and flow Hydraulic heads from water level measurements at various points and elevation at various borehole points indicate the presence of a close correspondence to topographic heights.

Water resources of arid areas

202

The comparison between hydraulic head and topographical height shows that the direction of groundwater flow is roughly to the drainage pattern which follows the gradient of the land surface. Groundwater movement in the study area is therefore approximately to the north east which is generally the direction of Kizinga river (Figs 1 & 2). The above observations on the groundwater movement in the study area suggest that groundwater starts flowing from Pugu and Kisarawe hills towards low lying plains of Yombo; Changombe; Temeke; Mbagala; Mtoni and Kurasini. This shows that precipitation within Pugu and Kisarawe hills is the major source of groundwater recharge in the study area. 3.2 Aquifers properties in the study area Layers of sand are important for the hydrogeology of the study area as they are mostly accompanied with good water bearing capacity and they allow significant quantities of water to be drawn from them (Fig. 3(d)). Generally the texture of sands are medium to coarse with gravels and pebbles existing in clay matrix. Clay layers tend to hold water which can not be withdrawn easily and therefore they don’t have good water bearing capacity (Figs 3(a & c)). Significant amount of groundwater has been found basically to occur in two types of aquifers namely sands and gravels and limestone. Clay and clay bound sands are poor aquifers and their importance is mainly on the formation of confining layers for most of sand and gravels aquifers (Fig. 3(b)). 4 ESTIMATION OF GROUNDWATER RECHARGE RATES Evaluation of groundwater recharge has significant implications for not only the study of groundwater quantity, but also water quality. Infiltrating water can carry contaminants from

Evaluation of groundwater recharge rates in the Kizinga catchment in Dar es Salaam region

203

Figure 3(a & b). Lithology of aquifers with their scales in (m) in some sites of the study area. ground surface to the aquifer. Understanding the rate and mechanism out of which such infiltration takes place can therefore lay a foundation in setting down strategies for the prevention of groundwater contamination. Three methods were used to estimate groundwater recharge. These are: the water balance method, hydraulic method and chloride profile method.

Water resources of arid areas

204

Figure 3(c & d). Lithology of aquifers with their scales in (m) in some sites of the study area.

Evaluation of groundwater recharge rates in the Kizinga catchment in Dar es Salaam region

205

Figure 4. A comparison between annual groundwater recharge, rainfall and actual evaporation. 4.1 Water balance method The method was used under the following assumptions: – Precipitation is the only inflow into the basin i.e., there is no leakage or underground channel to the basin – River discharge and evapotranspiration is the ultimate water outflow – Basin storage is steady and hence the in storage is considered to be zero for long term period. The water balance method requires the use of a combination of actual evaporation, surface runoff and precipitation data in order to estimate annual groundwater recharge. Groundwater recharge in this method is calculated as a remainder when losses, identified in the form of runoff and evaporation have been deducted from precipitation. This can be presented in the following equation: P=E+R+∆S Where: P—Precipitation (mm), E—Actual evaporation (mm), R—Runoff over the catchment (mm) and ∆S—Change in Storage (mm). The value of groundwater recharge obtained using water balance method gave the average value of 81.3mm/year. The comparison between annual groundwater recharge, annual rainfall and annual actual evaporation shows that, the values of recharge in most cases increase with increase in rainfall and decrease with increase in actual evaporation (Fig. 4) and that, annual groundwater recharge and annual rainfall are highly correlated.

Water resources of arid areas

206

The annual actual evaporation in the study area always exceeds annual rainfall. However a plot of mean monthly actual evaporation and rainfall for the one year period (Fig. 5), shows that there are few months when rainfall exceeds actual evaporation. These months are mainly, March; April; and November and it is expected that during these months, groundwater recharge takes place. 4.2 Hydraulic method This method was used to estimate the amount of annual groundwater in flow (Q) in to the Kizinga river catchment area. An average hydraulic gradient (I) of 8.063×10−3, hydraulic conductivity (k)

Figure 5. Mean monthly variation of rainfall with actual evaporation for the year 1968 in the study area. of 3.14×10−5m/s along a cross section at the center of the basin having the maximum aquifer thickness of 28m and a surface width of 27km, were used. Darcy’s law, Q=KIA were used Where: Q is the quantity of water (m3/s), K is the hydraulic conductivity (m/s), I is the hydraulic gradient and, A is the area (m2). The value obtained was 6036076.153×109mm/year and when extrapolated to the entire basin’area of 191km2, this amount of annual groundwater in flow in mm per year was calculated such that, Recharge rate=(6036076.153×109mm/year)/191×1012mm=31.6mm/year.

Evaluation of groundwater recharge rates in the Kizinga catchment in Dar es Salaam region

207

4.3 Chloride profile method The technique regards chloride as an inert element, and compared with other inorganic ions, it is not added or removed by water rock interaction. The element is considered as an inert in the hydrological cycle having its source from the atmosphere. It has the advantage over tracers involving water molecule in that atmospheric inputs are conserved during recharge processes allowing a mass balance approach to be used (Nkotagu 1996). On using the chloride profile method, it is assumed that the amount of water and chloride added at the surface should equal the amount of water and chloride percolated down. This however is not always true and therefore the following assumptions were considered when using this method. (i) Recharge is only that derived from precipitation (ii) Recharge is largely by piston flow mechanism (iii) Chloride in soil water is from precipitation and dust only (iv) The precipitation amount used in the recharge and soil water age equations is reasonable for the time represented by samples (v) The total chloride input value used in the recharge and soil water age equations is reasonable for the time represented by the samples (vi) Dispersive mixing of water and chloride is small (vii) The chloride uptake by plants is negligible Knowing that there is a possibility of groundwater being affected by marine intrusion or marine connate source as the study area is situated in the coast, the use of the method involved boreholes with water whose chloride concentration values fall within acceptable ranges of freshwater and

Table 2. Summary of the groundwater yield in the study area. Total No. of boreholes Maximum yield Minimum yield Total annual yield

130 60.923m3/h 0.220m3/h 9.4×106m3/year

out of seawater. This was achieved by including in the determination of groundwater recharge, only boreholes with chloride content whose ratio is less than 1. Groundwater recharge using this method was estimated as follows. The mean chloride content of precipitation was found to be 2.9mg/l and the mean chloride concentration of groundwater was found to be 71.8mg/l. The ratio of chloride content of precipitation to groundwater was determined to be 0.04. A long term mean annual precipitation of 1124mm was used in the calculation. Estimation resulted in groundwater recharge rate of 45.4mm/year.

Water resources of arid areas

208

5 DISCUSSION OF THE RESULTS The three methods gave a mean value of 52.8mm/year(Table 2), equivalent to 10.1×106m3/year. This indicates that, 4.7% of the long term mean annual precipitation, which is 1124mm, ends up as annual groundwater recharge. It can then be concluded that at present, annual groundwater production rates which is 9.4×106m3/year is approximately equal the annual natural groundwater recharge rates from the study area. The value obtained using water balance method is much larger as compared to the values of the other two methods. The big value in this method is likely to have been contributed by errors in the process of data taking. The system that was used in measuring runoff which based on water level recording two times a day is likely to have introduced some errors. Such errors may be missing of flood peaks especially that occurring at night and therefore registering less storm runoff than real values happening in nature. According to Matondo 1978, the personnel engaged in data collection by that time in most basins including Kizinga, were unskilled or only semi skilled and therefore less accurate and efficient in collecting data. There was no even distribution of rain gauge stations in the basin. Mostly only Dar es Salaam airport had a recording gauge (Matondo 1978). The existing raingauge therefore, did not facilitate studying the rainfall distribution in respect to time and space. This restricts further the accuracy of investigations. It can therefore be seen that the rainfall and runoff records might have errors, which then could possibly contribute to more errors in the computation of effective rainfall and average precipitation, and therefore the final results. In additional to shortcomings mentioned above, the data used in this method were recorded about 20 years ago while the other two methods used the data recorded within last two years. It is possible therefore that, the soil condition 20 years back supported more groundwater recharge. This is supported by the fact that the land in the study area has been disturbed through construction and cultivation. 6 CONCLUSION AND RECOMMENDATION The major contribution in recharging the study area has been found to be the faults on the slopes of Pugu and Kisarawe hills, which are quite permeable and that the direct infiltration of rainwater is the main source of groundwater recharge in the study area. The average value of annual groundwater recharge rates has been found to be 52.8mm/year after combining all the three methods used. This value is approximately 10.1×106m3/year and it is 4.7% of the long term mean annual precipitation of 1124mm. A projection of maximum groundwater abstraction showed an increase of more than 100% after every two years. However the natural groundwater recharge is expected to remain constant, and therefore the maximum production rate will be more than natural recharge rate within a period of few years to come. This relation when projected over a long period may result in negative consequences such as depletion of groundwater supply, reduced

Evaluation of groundwater recharge rates in the Kizinga catchment in Dar es Salaam region

209

stream flow, deterioration of water quality and more importantly land subsidence may be expected in the near future. On the other hand there is no comprehensive legal mechanism to ensure proper management of groundwater resources. The Water Utilization Act has not provided a holistic approach to the management of water resources especially with regard to the management of groundwater resources. There is no specific provision for groundwater abstraction under the Act and instead groundwater exploitation has been provided for its management in the same way as surface water. There is no separate provision under the Act to regulate groundwater-drilling operations. However, abstraction of groundwater of more than 22,700 litres per day requires a water right issued under the Act. The Act has not demanded for operators of wells or boreholes to submit data and records to the water authorities. It is therefore possible that, commercial drilling of groundwater of substantial scales has been conducted without adequate monitoring and controls. Clear understanding on the aquifer parameters and recharge rates for other parts of Dar es Salaam and the country at large is fundamental before embarking on further exploitation of the resources. Studies on the subject however are very limited to academic purposes and have not been able to comprehensively provide a clear understanding on the aquifers that provide the resource. In general, the studies that have been conducted so far indicate a negative trend in the status of groundwater in different parts of the country. The absence of adequate data and legislation has been impinging on effective management of groundwater resources. The Kazimzumbwi forest which is in the Pugu and Kisarawe hills is suffering a massive deforestation from illegal harvesting of forest products. The capacity of the faults to serve as groundwater recharge will consequently be affected. It is possible therefore that, in the near future the rate of natural recharge will decrease while that of abstraction will keep on increasing. It is recommended that, the faults on the slopes of Pugu and Kisarawe hills, which serve in recharging the area be conserved and protected from disturbance to ensure continuation of safe and enough groundwater supply. Possible alternatives for freshwater supply should be used. Measures that reduce water wastage must be introduced and encouraged. More studies on groundwater recharge and aquifer performance need to be conducted in the study area and others around Dar es Salaam using other methods so as to have a complete understanding on the aquifer system in Dar es Salaam city. ACKNOWLEDGEMENTS We would like to thank all those people who contributed to the preparation and completion of this work. Our special thanks go to the Department of Geology of the University of Dar es Salaam for the facilities they have provided to us during our research. Many thanks go to MHO program through the Faculty of Science University of Dar es Salaam for the financial support in undertaking the research. We wish to acknowledge the assistance of the Staff of the institutions in that time; Ministry of Water, Ubungo Maji; DAWASA headquarters and Tanzania Meteorological Agency Dar es Salaam office for supporting us with necessary data. We are grateful to the Management

Water resources of arid areas

210

of Lawyers’ Environmental Action Team (LEAT), for their support on access to the organizational facilities and information during the preparation of this paper. REFERENCES APHA, 1985. Standard Method for the Examination of Water and Wastewater. 16th Edition. American public Health Association, Washington, D.C. Matondo, J.I. 1978. A kinematic conceptual model for Kizinga basin for Estimation of Hydrological variables. M.Sc. Thesis, University of Dar es Salaam. Ministry of Energy and Minerals (MEM) 1979. Coast/Dar es Salaam regions water master plan. Dar es Salaam. Nkotagu H.H. 1996. Hydrological and Isotopic characterization of a fractured basement groundwater flow system in Semiarid Area of Dodoma, Tanzania. Znge: Berlin, Techn. Univ. Diss. Serviceplan, 1997. Report on the evaluation of groundwater sources of Dar es Salaam. Supporting Report B. Dar es Salaam.

Theme C: Socio-economic aspects

KNUST experiences in capacity building in the water and sanitation sector S.N.Odai, F.O.K.Anyemedu, S.Oduro-Kwarteng & K.B.Nyarko Department of Civil Engineering, KNUST, Kumasi, Ghana Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Capacity building for the water and sanitation sector in Ghana has been in existence in the Civil Engineering Department (DCE) of Kwame Nkrumah University of Science and Technology (KNUST) since the university was established. The activities of the department in the area of capacity building for the water and sanitation sector have become very popular after the establishment of the water supply and environmental sanitation programme (WSESP) in 1996. In fact, this recent development has brought the department to the forefront of capacity building and applied research in the water and sanitation sector of Ghana. The project, aimed at capacity building for sustainable development and growth in the water supply and sanitation sector in Ghana and the West African sub-region has so far produced several professionals from Ghana, from three West African states and one from an East African State. Capacity building through short courses and tailormade programmes has helped in training several institutions and communities in Ghana leading to sustainable development. The project has impacted positively on both institutional and human capacities in the country. The paper looks at the experiences of KNUST in providing highlevel and low-level capacity building for Ghana.

1 INTRODUCTION The human resources requirements of the water and sanitation sector are largely similar to those of other professions. Factors militating against capacity building in the sector have been mainly due to lack of funds and lack of understanding of the urgency of the need to improve the sector. In many developing countries of Africa, improvement in the water supply and environmental sanitation sector has lately been recognised to have a direct positive impact on public health (Ghana Government, 2003). The development of the sector is seen as crucial to the successful control and eradication of communicable diseases in general. As a result, governments, policy makers, non-governmental agencies,

Water resources of arid areas

214

and external support agencies have begun directing attention to issues related to these sector in an effort to improve the overall health of citizens by facilitating easy accessibility to water and sanitation. Until the last decade, the progress and growth in the water supply and environmental sanitation sector has been little and very slow. One of the major factors hampering the desired rate of progress and growth in the sector has been the lack of adequate personnel and professionals with the requisite skills, expertise and experience to lead and manage the sector. In fact, during the 1991 UNDP symposium at IHE-Delft, on A Strategy for Water Sector Capacity, it was acknowledged that capacity building in the water supply and environmental sanitation sector was essential for the development, growth and sustenance of the sector at the local, national and even sub-regional levels (KNUST, 2002a). In the wake of this symposium, it was felt that the training and re-training of professionals for and within the water supply and environmental sanitation industry must become part of the central focus of academic institutions that have long traditions of providing quality leadership training as well as professional expertise. The UNESCOIHE Institute for Water Education in Delft and the Department of Civil Engineering of the Kwame Nkrumah University of Science and Technology (KNUST) in Kumasi quickly took the initiative to develop a programme for supporting the restructuring and strengthening of the water supply and sanitation sector in Ghana and the sub-region. The programme was preceded with a needs assessment of the situation in several neighbouring countries, to assess the percentage coverage of sanitation and water supply in urban and rural communities, and the statistics were stunning. Following this needs assessment, the DCE of KNUST in collaboration with UNESCO-IHE in 1996 initiated a programme for developing human resource capacity in the water supply and environmental sanitation for a wide range of beneficiaries, spanning from sector professionals to low-level operators and sometime even uneducated water board members. The high-quality training programme in the water supply and environmental sanitation sector was conceived under the project name “Water and Environmental Sector Capacity Building and Sustainable Development in Ghana and the Region” designed by the KNUST-Kumasi and UNESCO-IHE. The project aims at providing the sector within the West African sub-region with the necessary skills, knowledge and expertise to meet the demands, challenges and opportunities anticipated with the projected growth in the region. This paper presents the experiences of KNUST in capacity building in Ghana and the sub-region. 2 HISTORY OF CAPACITY BUILDING IN THE WATER SECTOR AT KNUST The Kwame Nkrumah University of Science and Technology (KNUST) was established in 1951 to train Scientists and Technologists for both Ghana and other African countries. Academic programmes are run at the undergraduate and post-graduate levels by different faculties, schools and institutes. Training in water-related disciplines is offered in several departments in the university.

KNUST experiences in capacity building in the water and sanitation sector

215

Presently, the DCE, the Chemistry Department, and the Geological Engineering Department collaborate in delivery of education and training in water quantity and quality. The programmes of the Department of Civil Engineering cover both surface and groundwater resources and water quality analyses, while the Chemistry Department is involved mainly in water quality analyses, and the Geological Engineering Department involved in groundwater resource development. In addition, the Agricultural Engineering Department looks at irrigation and water for food, the Biological sciences Department works on environmental science, while the Physics Department masters in groundwater and limnology. Since 1996, the DCE and UNESCO-IHE have been developing human resource capacity in water supply and environmental sanitation, targeted at all levels of sector professionals (sometimes including persons with low-level education). Since the establishment of WSESP which has the primary focus of postgraduate training for the sector, several, short courses such as Public Private Participation, Water Treatment, Wastewater Treatment, Solid Waste Management, Urban Water Transportation and Distribution, etc., have been offered annually. 3 INSTITUTIONAL CAPABILITY OF KNUST The institutional capacity of KNUST to act as a capacity building centre in the water and sanitation sector is discussed under the following sub-headings. 3.1 Human resources The departmental academic strength comprises of 22 highly qualified lecturers, mostly PhD graduates from prominent universities in addition to experienced technicians who support the execution of educational programmes and consulting services; eleven out of the 22 staff members in the department specialise in water and sanitation issues. Specifically there are 5-PhD, and 6-Msc holders in the two sections. Three out of the six MSc’s are currently pursuing sandwich PhD programmes with UNESCO-IHE; one in Wastewater Treatment, one in Utility Management, and the third in Water Treatment. 3.2 Facilities The facilities at KNUST for training purposes include refurbished classrooms, a computer laboratory that gives access to each participant, a refurbished laboratory that allows for water and wastewater quality analyses for training and research purposes. Students and lecturers in the DCE, together with other departments in the School of Engineering, have access to a well-equipped library, apart from the main library of the University. In addition, a collection of specialised books is at the disposal of the staff and participants in WSES project. A 30-room hostel has been built purposely for the MSc participants. These rooms are also available for use by short course participants when these are organised during the vacation periods. The university now has Internet connectivity for research, and a website has been created for the project.

Water resources of arid areas

216

Lecture notes, mostly from UNESCO-IHE, in addition to modern ICT equipment facilitate our objective of capacity building. 3.3 Networking and outreaching capabilities At the national level, the Department has strong links with sector organisations in Ghana, such as the Ghana Water Company Limited, Community Water and Sanitation Agency, Ministry of Local Government and Rural Development, The Environmental Protection Agency and the Water Resources Commission. This link is maintained through quarterly meetings held together with these agencies. The main objective of this networking is to enable the university identify the problems of industry at first hand and provide the necessary advice. The department also depends on its international links to facilitate information acquisition. Presently, the university has active links with UNESCO-IHE in Delft, the University of Newcastle in UK, the University of Bristol in UK, and the National University of Rwanda. In fact, the department’s links with industry and the sector professionals have led to the active participation of these professionals in part-time lectures and the running of some short courses. The strategy is such that use is made of these professionals for training and lecturing where there is a gap. The departmental strategy is to maintain this alliance. 4 EXPERIENCES OF CAPACITY BUILDING UNDER THE WSES PROJECT The Water Supply and Environmental Sanitation Project is aimed at capacity building for sustainable development and growth in the water supply and sanitation sector in Ghana and the West African sub-region. It seeks to strengthen the sector through training of high-level personnel for institutions and organisations that have a stake in the water supply and environmental sanitation industry and professionals with active careers in the sector, in addition to training of low-level personnel. In Ghana, the programme targets the following sector organisations and their professionals: ● Ghana Water Company Limited (GWCL), ● Community Water and Sanitation Agency (CWSA), ● Environmental Protection Agency (EPA), ● Ministry of Works and Housing MWH), ● Ministry of Local Governments and Rural Development (MLGRD), ● Ministry of Health (MoH), ● Ministry of Environment and Science (MES), ● Water Research Institute (WRI), ● Consulting firms and contractors, ● Water-related industries, ● Small Towns Water Operators and Managers.

KNUST experiences in capacity building in the water and sanitation sector

217

4.1 Postgraduate studies The training is a formal type that aims at providing advanced level training at MSc degree level for people with interest in the water supply and sanitation sector and related fields. This group of trainees takes up top positions in agencies such as Community Water and Sanitation Agency and Ghana Water Company Limited. Besides young BSc holders, the programme is also designed to cater for professionals with industrial experience who want to improve their analytical skills and enhance their career skills and professional expertise. The table below gives the statistics of students who have graduated from the official MSc programme since its commencement in 1997. To our satisfaction, most of the graduates are working in responsible positions to implement the techniques they have acquired from their study. The major challenges facing potential participants of the programme include: – Employers not willing to release employees for two years – Employers not willing to grant study leave with pay – Duration of two years seems too long for some participants – Discontinuation of scholarship for foreign students. To help overcome some of the challenges above we consider using distance and electronic learning and modularised programmes. Scholarships for foreign students may be difficult to obtain but through our partnerships schemes, these may be possible (Odai et al, 2004). 4.2 Short/refresher courses Besides the MSc programme, the project develops short and refresher courses to enrich and refresh the knowledge of professionals already working or engaged in projects in the sector. Several topics have been treated since the establishment of the project. The topics are selected based on needs assessment in the sector through interviews and informal discussions. The target group includes staff from the organizations mentioned above in addition to consulting firms and contractors. Topics of the short courses ran include: – Operation and maintenance of water distribution systems (with computer applications) – Solid waste management – Water treatment technologies – Wastewater handling, management and disposal – Water quality analysis

Table 1. Statistics of graduates. Year Total Ghanaians 1999 2000 2001 2002

7 6 15 15

Foreigners 7 6 13 12

0 0 2 3

Water resources of arid areas

2003 Total

10 53

218

9 47

1 6

Table 2. Statistics of short courses (since 1999). Year Number of short Total number of courses participants 1999 2 2000 6 2001 3

17 50 27

– Groundwater flow pollution modelling – Public private participation in the water sector. It has been realized that most organizations that benefit from our training have more and interconnected needs, hence the most likely way forward is to develop tailor-made programmes for them, which will meet their needs holistically, rather than individual courses. The other reason being, marketing of the courses are becoming more demanding. 4.3 Tailor-made training The tailor-made courses are unique because they are usually developed through interactive and feedback discussions, in order to meet the requirements of the client institution. Target groups include technicians and professionals from the district assemblies, civil servants from the ministries, NGOs and rural water systems operators. Professionals from these agencies are given these specialised trainings to improve their performance. The WSESP is presently running two tailor-made programmes for sector agencies. They are municipal engineering course and small towns’ water supply operators’ course. Municipal engineering course: This course is organized with the primary aim of upgrading the technical and managerial capacity of the District Assemblies within the Ministry of Local Government and Rural Development (KNUST, 2002b). The personnel upon completion of the course are required to be able to do the following in addition to other assignments. – Plan their needs in infrastructure and resources, – Operate and maintain infrastructure, – Estimates materials and manpower needs for the construction of infrastructure, etc. The course is organised for duration of four weeks and the courses offered are – Planning and Management – Refuse management – Excreta disposal – Drainage – Infrastructure – Water supply – Roads and Highway – Electricity Service

KNUST experiences in capacity building in the water and sanitation sector

219

– Telecommunication – Structural Building works

Table 3. Statistics of municipal engineering trainings. Year

Total no. of participants

February 2003 August 2003 January 2004 Total so far

10 23 15 48

Table 4. Small towns’ water systems trainings. Year

Total

August 2003 February 2004 Total so far

145 38 183

Small towns’ water supply operators’ course: This course is organised with the primary aim of upgrading the technical and managerial capacity of water board members and the system operators to enable them achieve sustainability. Needs assessment showed that there are more than 2000 people in the 210 districts in Ghana needing the first round training. The objectives of the training are (KNUST, 2003) – To update the knowledge, technical and management skill, and attitude of operating staff of small town water systems for effective and efficient operation and maintenance, and management of the water systems. – To train water and sanitation board members to effectively oversee and manage the water system and deal with consumer complaints and requests satisfactorily. The global objective is to attain sustainability, which will subsequently lead to improved health conditions and productivity. The attendance statistics is shown in the table below, and it is impressing to note that the number of participants has increased since we started the tailor-made courses. The course is organised for duration of one week and the courses offered are – Water supply system operation and maintenance – Borehole pumping system operation and maintenance – Managing system information – Water and health – Roles of the water board – Budgeting and tariff setting.

Water resources of arid areas

220

5 CONCLUSION In Ghana, statistics show that there is low capacity in the water and sanitation sector. Our experience in capacity building shows that organizations are aware of their need but sometimes they need help to articulate such needs. The capacity building programme of sector professionals and low-level personnel will be strengthened, and sustainability enhanced. It is anticipated that the 210 districts in Ghana will eventually benefit from our programmes. The department is gradually shifting to tailor-made programmes since the short courses do not proof to be financially sustainable and looking for clients who will patronise the course places extra demands on us. The recent thinking of tailor-made programmes is catching on since there is usually money available for particular organizations to build capacity. We therefore develop programmes, which we discuss with donors and such organizations; upon approval of the courses we then prepare teaching materials and the cost estimates. This approach seems to becoming popular with us because of the high response of participants and the monetary value. Thus we can ensure sustainability in this process of capacity building in the water and sanitation sector in Ghana. REFERENCES Ghana Government, 2003. Ghana’s poverty reduction strategy. National Development Planning Commission, Accra, Ghana. 112–113. KNUST, 2002a. Brochure for MSc programme in Water Supply and Environmental Sanitation. KNUST, Kumasi, Ghana. KNUST, 2003. Evaluation report on operation and maintenance. Department of Civil Engineering, KNUST, Kumasi, Ghana. KNUST, 2002b. Proposal for municipal engineering and infrastructure management course. Department of Civil Engineering, KNUST, Kumasi, Ghana. Odai, S.N., Andam, K.A., & Trifunovic, N. 2004. Strategic partnerships for sustainable water education and research in developing countries. Proc. Int. Conf. on Water Resources of Arid and Semi Arid Regions of Africa, Gaborone Botswana, 3–6 August 2004.

Strategic partnerships for sustainable water education and research in developing countries S.N.Odai Department of Civil Engineering, KNUST, Kumasi, Ghana K.A.Andam Vice-Chancellor, KNUST, Kumasi, Ghana N.Trifunovic UNESCO-IHE, Delft, The Netherlands Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: The past decade has seen tremendous investment in water and sanitation for both the urban and the rural communities in developing countries. For effective and sustainable implementation of these projects, local capacity is required. The training of such professionals is becoming increasingly expensive for governments of developing countries, in addition to the fact that there are few institutions offering such programmes. The Kwame Nkrumah University of Science and Technology (KNUST) realising the modern trend in water education and research has entered into strategic partnership with some twenty other institutions around the world to jointly train professionals in the sector; and other external support agencies for financial support. This strategic partnership for water education and research is seen as one of the major emerging options for water education and research around the world. This paper looks at the benefits and challenges of such global partnerships for developing countries.

1 INTRODUCTION The development of the water and sanitation sector is seen as a crucial process to the successful control and eradication of communicable diseases in general. Increasingly, more governments are realising that improvement in the water supply and environmental sanitation sector has a direct positive impact on public health. There are stories about communities where charging for the provision of potable water caused the inhabitants to go back to other sources, e.g. nearby streams, which are not potable. This practice resulted in some persons suffering from diarrhoea, while others suffered from guinea

Water resources of arid areas

222

worm infestations. These are some of the challenges encountered by governments of developing countries. As a result of these attitudes, the government, external support agencies and policy makers have begun directing attention to issues related to the sector in an effort to improve the overall health of the people (Ghana Government, 2003). To achieve this noble objective, water education and research is being enhanced at the tertiary institutions. Water education and research is becoming increasingly diverse and subsequently, more expensive for most developing countries around the world. Thus providing the needed education and research in the sector requires more input which is not accessible locally in one country, hence the need to enter into strategic partnership with other institutions to complement each other and to share experiences. In addition, it is necessary to collaborate with external support agencies working in the sector for financial assistance. In general, the financial base for service provision is quite weak for most institutions. For example, the Department of Civil Engineering (DCE) of the Kwame Nkrumah University of Science and Technology (KNUST) depends on subventions from central government in the provision of its services. However, this is not adequate for postgraduate education in water and sanitation, hence the need to collaborate with agencies who can support us financially. The department organises short courses for sector professionals and offers consultancy services to make additional income. However, as mentioned above these efforts must be augmented for sustainable delivery of the services of the department. Current trends encourage the need to enter into strategic partnerships with institutions and agencies to enhance the sustainable delivery of our services. Generally, in a large number of developing countries there are still only a few people and institutions that have sufficient knowledge to solve the complex technological as well as institutional sector problems of concern. As indicated by different groups of professionals on many occasions, the major causes are (UNESCO-IHE, 2002): ● Lack of sufficient capacity provided by qualified professional staff; ● Lack of highly developed knowledge centres; ● Lack of sufficient institutional and governance qualities and capabilities to ensure an integrated approach to sector problems; ● Lack of ‘communities of practice’ through which professionals and institutions can effectively exchange information, knowledge, experience and good practices.

2 STRATEGIC PARTNERSHIPS To overcome some of these challenges, the DCE has entered into partnerships with several institutions and external support agencies for water education and research. This is in recognition of the modern trend of education, which takes advantage of virtual classrooms and uses expertise around the world as guest lecturers.

Strategic partnerships for sustainable water education and research in developing countries

223

2.1 Partnership with WANet The West Africa Network for Capacity Building in IWRM (WANet) was established in June 2002 for the purpose of training Policy/Decision Makers, IWRM Professionals, and Technicians. Three institutions, namely, the KNUST of Ghana, the National Water Research Institute of Nigeria, and EIER/ETSHER of Burkina Faso are involved. The various training activities in IWRM proposed for these identified target groups are presented below. 2.1.1 Policy/decision makers This group is made up of professionals, executives and, bureaucrats who are involved in policy-making and implementation of IWRM programmes. Specifically, the group includes government policy makers (political leaders, ministers of water, agriculture, environment etc, Governors, Mayors), and senior managers of IWRM-related institutions. 2.1.2 IWRM professionals This group is also quite broad, made up of Water Experts/Consultants, Future Managers in the Water/IWRM domain, and Trainers/Educationalists. IWRM consultants include professionals with diverse specialisations in water management and its use. These include Planners, Engineers, Economists, Social Scientists, Agricultural Workers and Ecologists (Donkor and Nyarko, 2002). The Future Managers sub-group will concentrate on students at the postgraduate level specialising in IWRM-related subjects. Trainers/Educationalists include academic staff of universities in the field of IWRM and trainers and professionals at training institutes. 2.1.3 Technicians This group is made up of those charged with the operation and maintenance of facilities used in the direct provision of IWRM services such as water treatment plants, maintenance of water supply system, etc. Cap-net helped establish this partnership and they continue to support WANet financially for its activities in getting the institutions involved to build the capacity and the materials they need to kick-start their trainings. WANet is however faced with the challenge of offering training in English and French. 2.2 Partnership with UK institutions through British Council British Council (BC) offers opportunities for linkage between universities in UK and other countries. The British Council wholly provides the budget for these links. The activities under such links include staff exchange between the two universities, equipment purchase, research and sometimes students’ exchange. Currently, the DCE has two of such links with two universities in the UK, under different themes. There is one with the University of Newcastle upon Tyne under the

Water resources of arid areas

224

theme “shelter and sanitation for the homeless” (1999–2005), and the other with the University of Bristol under the theme “sustainable water delivery for the poor” (2003– 2006). These links are established based on the mutual consent and interest of the two institutions involved since they will have to make their expertise and facilities available for use by the other institution. Under the link arrangements, staff from UK institutions may serve as lecturers in Ghana, with the cost borne by BC. On the other hand, staff from Ghana benefit from sitting in some courses, having discussions with experts, having access to their library materials and electronic journals. The staff exchange grants staff from both institutions the mutual benefit of learning of new areas of research. Ghanaian students on exchange to a UK institution benefit from having access to facilities for research, while students from UK benefit from best practices in Ghana. Over 50 Ghanaian students and 40 British students have benefited from this partnership exchanges. This students exchange programme has been at the BSc level, and recently extended to PhD students. MSc students are yet to be included in the programme. 2.3 Partnership with DANIDA DANIDA is a bilateral agency, very active in the water and sanitation sector in Ghana. The agency supports Ghana in physical and institutional developments in the water and sanitation sector. They have already invested millions of euros in the sector. They have been working with the Water Resources Commission, Water Research Institute, Hydrological Service Department, and the Meteorological Service Department for data collection for water resources management. In December 2002, the DCE was invited to attend a stakeholders’ workshop at which it presented its strategic plan for education and training in the sector. The department’s strategic plan included networking with sector agencies for education and research, undertaking problem-oriented research to help develop the country. The result of that workshop led to the signing of agreement between the DCE and DANIDA. DANIDA has since been supporting the DCE with scholarships to train MSc students in water resources engineering and management, and providing funds for external examiners from any part of the world. This partnership is making financial resources available for the department to enhance research and education. The department also now has a better relationship with the above-mentioned agencies for easier access to data. These agencies can also now fall on the department for consultation in research or capacity building. 2.4 Partnership for Water Education and Research (PoWER) The Partnership for Water Education and Research (POWER) was founded in November 2002 by 17 institutions around the world (from countries in Africa, Asia, Europe, and South America). UNESCO-IHE has been in the center of all these activities. The main objectives of the partnership are: ● To develop a sustainable and mutually beneficial global partnership in water education and research between UNESCO-IE and regional collaboration centres that promotes life-long learning through generation and sharing of knowledge in integrated and sustainable water and environmental systems relevant to the developing world.

Strategic partnerships for sustainable water education and research in developing countries

225

● To combine the strengths of all partners and enhance the capacity of each partner in order to produce joint products, such as, deliver capable professionals in the sector, find innovative solutions for sector challenges, and build institutional capacity for better efficiency. In the process of combining strengths and levelling the capabilities of the individual partners, joint products in the field of education, training, communities of practice, staff exchange, and collaborative research will be developed in a multidisciplinary manner. These shall be demand-responsive and duly accredited. 3 BENEFITS OF SUCH PARTNERSHIPS The list of the benefits of such partnerships is endless. Some of them are ● Combining expertise from various institutions to do research and publish papers ● Combining strengths to prepare lecture materials ● For very expensive experiments, if one institute has the equipment the other institutions can have access to them for their work ● There is leadership in research, since one institute may have all strength in a particular area; and the other institution may depend on such an institution for direction ● Mutual benefit of learning of new areas of research ● Encourages distance and electronic learning ● Financial assistance usually available for research and training ● Knowledge from the north is made available to the south ● Best practices developed in the south are made available to the north.

4 CHALLENGES OF SUCH PARTNERSHIPS The challenges that come with such partnerships are numerous but not destructive. Some of them are mentioned below. ● Each institution must strive to attain excellence and international recognition/accreditation ● Strive to stay modern by improving ICT equipment and providing or having access to video-and tele-conferencing facilities ● Make your strengths available for other institution to benefit from ● Source funding for the partnership ● Difference in languages of present and potential partners. The present partnership is expected to grow into an international collaboration within which partners will complement each other’s effort as done in the aviation industry, e.g., the alliance between KLM, Northwest and Kenya Airways.

Water resources of arid areas

226

5 CONCLUSIONS Strategic partnerships for sustainable water education and research in developing countries has come to stay, because modern technologies are available for easy communication and sharing of knowledge. It is also getting more expensive to maintain all the facilities and the internationally renowned experts in every institution, therefore partners will share their facilities and expertise to help reduce cost of education in water. Partner institutions are also continually challenged to stay abreast with modern trends in water education. REFERENCES Donkor, E. & Nyarko, K.B. 2002. Establishing nodal resource center in West Africa for capacity building in integrated water resources management. Department of Civil Engineering, KNUST, Kumasi, Ghana. Ghana Government 2003. Ghana poverty reduction strategy. National Development Planning Commission, Accra, Ghana: 112–113. KNUST 2002. Brochure for MSc programme in Water Supply and Environmental Sanitation. KNUST, Kumasi, Ghana. Odai, S.N., Anyemedu, F.O.K., Oduro-Kwarteng, S. & Nyarko, K.B. 2004. KNUST experiences in capacity building in the water and sanitation sector. Proc. Int. Conf. on Water Resources of Arid and Semi Arid Regions of Africa, Gaborone Botswana, 3–6 August 2004. UNESCO-IHE 2002. PoWER and knowledge for sustainable development. http://www.ihe.nl/power/%20knowledge.htm

Assessing demand for clean and safe domestic water in eastern Zimbabwe E.Manzungu, M.Machingambi & R.Machiridza Department of Soil Science and Agricultural Engineering, University of Zimbabwe, Harare, Zimbabwe Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: This paper assesses rural people’s demand for clean and safe domestic water in two districts in eastern Zimbabwe. It explores the role played by physical, socio-economic and cultural factors in influencing rural people’s willingness to pay for domestic water. A semistructured questionnaire was administered in January 2002 to representatives of some 239 randomly selected households in Chitakatira and Nyanyadzi wards. Willingness to pay for clean and safe domestic water was found to be influenced by the availability of other water sources, rainfall received in the catchment area, perceived safety of a water source, age and occupation of respondents. Season and gender did not significantly affect respondents’ willingness to pay. Demographic characteristics influenced willingness to pay for clean and safe water, which should be taken into account when implementing cost recovery policies in the domestic water sector. For completeness, effective demand, best illustrated by ability to pay, should be determined.

1 INTRODUCTION In 1996 it was estimated that 2.5 million Zimbabweans had no access to safe water (Chenje and Johnson, 1996). The situation was worse in the rural areas where only 64% of the population had access to safe water compared to 99% in the urban areas. A recent survey confirmed that water in the rural areas was largely unsuitable for human consumption due to bacterial contamination (Moyo and Mtetwa, 2000). The situation has deteriorated in the last 5 years because of severe economic problems, worsened by the withdrawal of donor support. Since independence in 1980 the donor community has heavily financed Zimbabwe’s Rural Water and Sanitation Programmes. Reduction in funding has resulted in poor maintenance of water supply facilities forcing rural communities to revert back to unsafe water sources (NAC, 1997).

Water resources of arid areas

228

Local communities are increasingly being called upon to contribute in cash and kind to the operation and maintenance of the domestic water sources, a development linked to the Economic Structural Adjustment Programme that the Government of Zimbabwe started in 1991. This World Bank/International Monetary Fund-supported programme advocated for cost recovery in many areas including social services. Provision of domestic water was not spared. For example, the enactment of both the Water Act and the Zimbabwe National Water Authority Act in 1998 incorporated policies like cost recovery and economic water pricing. This signified a policy shift towards the concept of treating water as an economic good (Manzungu, 2001). However the concept, borrowed from the international community, as captured in the World Water Vision (Cosgrove and Rijsberman, 2000), is characterised by a number of inconsistencies (Savenije and van der Zaag, 2002). In a country like Zimbabwe, where 75% of the rural population is regarded as poor, (GOZ, 1995) there are legitimate grounds to ask whether such a policy best serves the population. A recent survey found that, at the local level, there were mixed signals regarding people’s willingness to pay for water (Machingambi and Manzungu, 2003). Respondents wanted the cost of water point establishment and repairs shared between the community (69%), the government (11%) and the donors (5%). Sixty-three percent of the respondents wanted the government to take the responsibility of establishing water points. Close to half (43.9%) indicated that they had individually contributed towards the establishment of the water points they were currently using. There was also a willingness to participate in the maintenance of most domestic water sources except in the piped water scheme apparently because of the high costs involved. The question is: Does this willingness to participate in operation and maintenance of domestic water facilities translate to a demand for clean and safe water by rural communities in Zimbabwe? This study sought to determine whether there was a demand for clean and safe water among the rural people by assessing their willingness to pay for domestic water in the Lower Odzi subcatchment in Chimanimani and Mutare districts in eastern Zimbabwe. In many respects this area typifies most rural areas in the country. In the study the demand for safe and clean water was assessed using the contingent value method (Pearce, Markandya and Barbier, 1989). The method is based on eliciting, from respondents, valuations/bids, which to some extent reflect the strength/ depth of feeling i.e. degree of concern about access to clean and safe water on the basis of a hypothetical market. The hypothetical market is taken to include, not just the good itself, but also the institutional context in which it would be provided, and the way in which it would be financed. The respondent is asked to indicate whether or not they would be willing to pay (WTP) a “starting-point bid/price (SPP)”. An iterative procedure then follows: the SPP is increased to determine whether or not the respondent would still be willing to pay the increment in the price. The last accepted bid, then, is the “maximum willingness to pay (MWTP)”. Besides WTP, respondents were asked about their ability to pay for clean and safe water.

Assessing demand for clean and safe domestic water in eastern Zimbabwe

229

2 MATERIALS AND METHODS 2.1 The study area Chitakatira ward in Mutare district, and Nyanyadzi ward in Chimanimani district in eastern Zimbabwe, were selected as the study areas because of the different rainfall amounts received by the two areas. Water availability was hypothesised to influence demand for water. Nyanyadzi is located in natural region V that receives less than 650mm per annum. Chitakatira falls in natural region III where annual rainfall amounts of 680–800mm are received (Vincent and Thomas, 1960). The amount and distribution of rainfall received in region V is less reliable than that received in region II. The ward was used as a sampling unit since it is the government’s planning and administrative unit. Six villages, each made up of 100 homes, normally constitute a ward. A ward is therefore made up of approximately 600 homes (Makumbe, 1996). Nyanyadzi ward had a total of 832 households (CSO, 1992). The Nyanyadzi Rural Service Centre, which falls within the ward, is supplied with piped water as well as some of the surrounding villages. An irrigation scheme near the service centre also constituted a domestic water source for some respondents. According to the 1992 census there were 1224 households in Chitakatira (CSO, 1992). It also has a rural service centre, and villages, supplied with piped water. The water is drawn from Zimunya dam. It is treated by the national water utility, Zimbabwe National Water Authority (ZINWA). 2.2 Data collection and analysis Respondents were drawn from villages with and without piped water, and from the nearby rural service centres. A semi-structured questionnaire was administered in January 2002 to a total of 239 randomly selected households from the two districts. The questionnaire was administered to one respondent per household who was either the head of the household or a representative. A total of 118 people (82 males and 36 females) and 121 people (70 males and 51 females), were interviewed in Nyanyadzi and Chitakatira respectively. In Nyanyadzi a total of 83 respondents with access to piped water were interviewed compared to 100 in Chitakatira. The questionnaire sought to obtain answers to the role played by physical, socio-economic and cultural factors in influencing respondents’ willingness to pay for water from boreholes, deep and shallow wells, rivers, dams, canals and water taps. Informal interviews were also conducted with ZINWA, Department of Agricultural Technical and Extension Services (AGRITEX) now split into the Department of Agricultural Research and Extension (AREX) and the Department of Agricultural Engineering (DAE) and local government officials. Data was analysed using the Statistical Package for Social Sciences (SPSS) for windows version 10. Descriptive statistics were run for different variables so as to obtain the frequencies, means and cross tabulations. Further analysis was done to determine the effect of different variables on the SPP and MWTP using analysis of variance (ANOVA). The Levene test was used to assess whether the used ANOVA model had a good fit. To validate the ANOVA results, non-parametric tests were done. The Mann-Whitney test

Water resources of arid areas

230

was used to determine which of the factors; season, age, region, gender, occupation and access to piped water affected respondents’ WTP regarding the establishment and repairs of different water points. The Moses test was used to determine whether the observed variation between variables was due to the influence of some of the aforementioned factors. The Kruskal-Wallis test was used to assess if the means of the variables in each category were the same at 5% significance level. 3 RESULTS In some cases, there were no significant differences in the observations between the two wards that were studied. For this reason some data is presented in a collated form. Where there were differences the data is presented separately for the two wards. 3.1 Profile of respondents The profile of the respondents was hypothesised to have an effect on the demand for clean and safe water. Table 1 presents the age distribution of the respondents. At least 70% of the respondents from both wards were communal smallholder farmers while 5% of the respondents from each of the wards were government officials. The proportion of student respondents in Chitakatira and Nyanyadzi was 2% and 5% respectively. Nyanyadzi also had traditional healers among its respondents. 3.2 Conceptions of water Table 2 shows the various uses to which water was put as well as whether it was regarded as primary or commercial water use. Water was classified as primary when it was regarded as a basic need i.e. when used for daily subsistence requirements for most households. Commercial water use, for which people were supposed to pay, according to the respondents, was any use that was not regarded as primary, especially if the use generated a financial income. Most of the water uses were classified as primary.

Table 1. Age distribution of respondents. Age (years) 15–30 31–45 46–60 61–75 >75

Number of respondents Chitakatira Nyanyadzi (n=121) (n=181) 29 46 34 10 1

22 42 32 17 3

Respondents did not make a distinction between clean and safe water although the general perception was that clean and safe water was free of bacteria. Table 3 gives perceived characteristics of clean and safe water.

Assessing demand for clean and safe domestic water in eastern Zimbabwe

231

The perception of whether water was clean or safe was influenced by the source of the water and the season in which that water was used as shown in Table 4. In the wet season, across all water sources, water was perceived to be unsafe. 3.3 Willingness and ability to pay Eighty-two percent of the respondents felt that primary water use should be accessed free of charge, and 79.5% of the respondents felt that people had to pay for commercial use of water. Some 2.1% of the respondents felt that water had to be paid for to enable the maintenance of

Table 2. Patterns and classification of water uses. Water use Drinking Cooking Bathing Laundry Irrigating gardens Livestock watering Brick making Irrigating plots

Classification (%) % citing Primary Commercial water use 99.2 99.2 99.2 99.2 80.8

98.3 54.4 98.7 98 70.3

0.8 0.4 0.4 0.8 10.9

31

16.7

14.3

23.4 25.9

9.2 7.9

14.6 18

Table 3. Characteristics of clean and safe water. Characteristic % attributing % attributing it it safe clean Clear Bacteria free Chlorinated Protected Tasty Not rusty Not dirty Piped Treated

43.1 34.3 3.8 9.6 1.3 0.4 0.4 0.4 0.8

38.1 36.4 3.8 9.6 1.3 0.4 0 0 2.9

Water resources of arid areas

232

Table 4. Perception of water quality of various sources across seasons. Water quality in Water quality in summer (%) winter (%) Water Safe Not Clean Not Safe Not Clean Not source safe clean safe clean Borehole Deep well Shallow well River Dam Canals and water taps

66.1 1.3 6.7 2.1

67 6.7

0.4 53.1 1.7 2.1 3.3 1.3

53.6 3.3

1.3 1.3

0.4 2.1

0.4

2.1 0.4 1.3

0.4

1.3

4.2 31.8 3.3 5 25.5 1.7

5.4 30.5 3.3 12.6 2.5 5.9 2.9 3.3 25.3 1.7 16.7 1.7

4.6 11.3 2.1 4.2 16.7 1.7

infrastructure. About 80% of the respondents said it was the responsibility of the government to ensure that the communities had enough water. Domestic water provision was said to involve some cost by 28.9% of the respondents. Of these, 4.8% put the costs as ranging between Z$50 and Z$400 (Z$4000=US$1). This was however misleading since the figures coincided with the monthly water bills for respondents, especially those staying at rural service centres. The ability to pay different sums of money for water for the month is as shown in Figure 1. Ability to pay, it should be noted, is a function of affordability. It, however served, as a good indicator of respondents’ demand since it showed the price respondents would want their water supplied at. It was observed that generally the number of respondents decreased with the increase in the amount to be paid. However Z$10 was the most common WTP figure. Ability to pay was linked to the SPP and MWTP, which was affected by a number of factors such as the source of the water, the season and the treatment water was subjected to. 3.4 Factors affecting SPP and MWTP A regression model to find the effect of rainfall received (β1), season (β2), gender (β3), age (β4), occupation (β5) and access to piped water (β6) on the SPP and MWTP was run at 5% significance level. The general model was defined as: SPP/MWTP=a+β1+β2+β3+β4+β5+β6+ (1) The following hypotheses were then formulated to test the assumption of normality on equal variance using the Levene’s test. H0: δ12=δ22=δ32=δ42=δ52=δ62 (2)

Assessing demand for clean and safe domestic water in eastern Zimbabwe

(homogeneous variance) H 1: δ12≠δ22≠δ32≠δ42≠δ52≠δ62 (heterogeneous variance)

233

(3)

Accept H0 if p>0.05 Further analysis was undertaken using non-parametric tests, namely the MannWhitney, the Moses and the Kruskal-Wallis.

Figure 1. Respondents’ ability to pay for domestic water per month. Table 5. Comparison of respondents’ MWTP in different seasons. MWTP range (Z$/bucket—25 litre container) 0 1–20 21–40 41–100 >100 Source S W D S W D S W D S W D S W D of water Shallow 8.8 7.9 7.1 33 32.9 33.9 0.4 0.8 1.3 0.4 0.8 1.3 0 0 0 well

Water resources of arid areas

234

Deep 6.7 5.9 5.4 12.6 22.6 23.4 0.4 2.1 0.8 0 0 0.8 0.4 0 0.4 well Borehole 4.6 5 4.2 34.8 31.8 33.1 1 7 4.3 1.3 2 1 1.7 2.5 0.8 1.3 1 7 River 9.6 8.8 7.9 17.5 18 19.3 0.4 0.4 0.4 0.8 1.3 1.7 0 0 0 Dam 7.9 7.9 6.3 18.7 17.9 18.9 0 0.4 0.4 0.8 1.3 2.1 0.4 0.4 0.4 Piped 2.5 2.5 2.5 48.9 47.6 45 2.5 2.9 4.2 8.2 8.7 9.5 2.1 2.5 2.8 scheme Canals 0.4 0.4 0.4 2.1 2.1 2.1 0 0 0 0.4 0.4 0.4 0 0 0 & water taps Key: S—Summer; W—Winter; D—Drought; MWTP— Maximum willingness to pay.

Rainfall: There was a significant effect of rainfall received in a particular area on the demand for clean and safe water. Chitakatira generally had lower SPP and MWTP values compared to Nyanyadzi. Water was therefore perceived to be more valuable in the generally drier Nyanyadzi than in the wetter Chitakatira. The ANOVA test on the means showed that except for the SPP and MWTP to repair broken down pipes, all other variables were significantly affected by the amount of rainfall received. The MannWhitney test confirmed that the amount of rainfall had an effect on all variables except the SPP for repairing a broken down pipe. The Moses test showed that all variables, except the SPP and MWTP for establishing a new borehole, had major differences within them because of the rainfall factor. The amount of rainfall received in a particular location can therefore be said to influence respondents’ WTP. Source of water: Water from boreholes, piped water schemes, canals and water taps, shallow and deep wells had a modal MWTP figure of Z$10 whilst that of dam and river water was Z$0. Piped water was the most popular water source followed by borehole, shallow well, deep well, river, dam, canals and water taps. Season: The measures of association between season and the different variables were also very small (ranging between 0.015–0.098) confirming minimal association between season and the SPP and MWTP values. However, drought had the highest mean followed by winter and summer. Table 5 shows the distribution of respondents’ MWTP across the seasons. Water treatment: Treatment of water marginally changed the proportion of respondents willing to pay more than Z$40 across all the water sources, especially for piped, river and dam water. When treatment was factored in, the modal MWTP figures remained the same for the respective water points although the MWTP figures rose to Z$999. The proportion of respondents willing to pay Z$40 or less actually increased for water sources such as the piped scheme and river, whilst it decreased in the case of water from boreholes, shallow and deep wells. The modal MWTP figures for different water sources did not change. The distribution of respondents’ MWTP with and without treatment is shown in Tables 6 and 7. Gender: The ANOVA, Mann-Whitney and Moses tests revealed that gender did not significantly affect the SPP and MWTP values of the respondents. Except for the SPP and MWTP of establishing a new borehole and repairing it, all other variables showed that males had higher mean WTP figures than their female counterparts.

Assessing demand for clean and safe domestic water in eastern Zimbabwe

235

Age: The means for the different variables were found to have small variances with the 30–45 and 60–75 year age groups having the highest SPP and MWTP figures whilst the dependent (15–30 and 75+ year) age groups had the lowest values. ANOVA tests showed that the SPP and MWTP for a new deep well were affected by age. The KruskalWallis test showed that the SPP for a new water source, as well as the SPP and MWTP for the repair of a borehole, were different for

Table 6. Comparison of respondents’ MWTP for treated water. Treated water—MWTP range (Z$/bucket—25 litre container) Source of 0 1–20 21–40 41–100 >100 water Shallow well Deep well Borehole River Dam Piped scheme Canals & water taps

6.7 32.9

2.5

1.3

0.4

5.4 4.6 7.1 5.9 2.9

2.1 3.7 0.8 0.8 5.1

1.7 2.5 2.1 2.1 9.2

0.4 1.3 0.4 0.4 2.4

0

0.4

0

22.1 32.5 20.6 19.8 44.7

0.4 2.5

Table 7. Comparison of respondents’ MWTP for untreated water. Untreated water—MWTP range (Z$/bucket—25 litre container) Source of 0 1–20 21–40 41–100 >100 water Shallow 9.2 32.5 well Deep well 6.7 23.4 Borehole 5 34.2 River 8.8 18.4 Dam 7.1 19.2 Piped 5.4 45.5 scheme Canals & 0.4 2.1 water taps

0.8

0.4

0

1.7 1.7 0.8 0 3.7

2.1 0.8 0.4 0.8 7.8

0.4 1.7 0 0.4 1.2

0

0.4

0

respondents with the different ages. It can therefore be concluded that age affected these variables particularly the 15–30 and 75+ year age groups. Occupation: Traditional leaders’ MWTP was Z$1 for borehole water, Z$5 for piped water and Z$0 for all the other water points as well as for all repairs. Communal farmers,

Water resources of arid areas

236

local government officials and students had higher SPP/MWTP figures for most variables in decreasing intensity than traditional leaders, Agritex and ZINWA officials. ANOVA tests revealed that the SPP and MWTP for establishing a new deep well, SPP and MWTP for a new borehole, SPP and MWTP for repairing a borehole, and SPP and MWTP for repairing a broken down piped scheme, had significant variances due to occupation of the respondent. There was minimal negative association between occupation and the SPP and MWTP for a new water source, SPP and MWTP for a new borehole, SPP and MWTP for a new deep well, and the SPP and MWTP for borehole repairs. The SPP and MWTP for establishing a new deep well, SPP and MWTP for a new borehole, SPP and MWTP for repairing a borehole and SPP and MWTP for repairing a broken down piped scheme were shown to be significantly affected by occupation according to the Kruskal-Wallis test. Occupation had a weak association with these variables. This could be explained by the fact that more respondents did not have a stable income hence their responses masked those of respondents with stable sources of income. Access to piped water: The Levene test for equality of variances on the impact of access to piped water on respondents’ SPP/MWTP showed that there was homogenous variance, which implied that access to piped water did not affect differences in the SPP/MWTP. However, the Mann-Whitney test showed that access to piped water affected WTP for all water sources although respondents with piped water had lower SPP/MWTP figures compared to those without piped water except for borehole water. In order to establish whether demand for a better water service delivery existed in the communities, respondents were further asked whether they were willing to contribute towards the establishment of a new water source that would save women time compared to an old source, new borehole and deep well. Investment in a new water source for women’s needs: Respondents’ SPP and MWTP were not much different for a new water source that would save women time compared to an old source across the seasons. However, during drought the SPP/MWTP figures were higher than in winter and summer. The Levene test showed that variances in the SPP and MWTP values observed were due to the effect of the amount of rainfall received. Parameter estimates showed that Chitakatira had lower SPP and MWTP values than Nyanyadzi. Agritex/ZINWA officials, traditional leaders and local government officials had a decreasing effect that is lower SPP and MWTP values compared to communal farmers and students. Gender did not affect the respondents’ SPP and MWTP values although males had higher SPP and MWTP figures than females. The Mann-Whitney test showed that access to piped water affected the SPP/MWTP. Respondents with piped water had higher SPP/MWTP than their counterparts without. For the establishment of a new borehole, Chitakatira had lower WTP figures compared to Nyanyadzi, which decreased the SPP and MWTP values. Traditional leaders had lowest WTP figures whilst the Agritex/ZINWA officials had the highest SPP and MWTP figures. Drought had higher SPP and MWTP values followed by winter then summer. In this case females were found to have higher SPP and MWTP values than males. Access to piped water was shown not to affect respondents’ SPP/MWTP for establishing a new borehole. Occupation, age and region had significant effects on the SPP and MWTP for the establishment of a new deep well. Traditional leaders still had a decreasing effect on the SPP and MWTP whilst students had the highest SPP and MWTP values. The 15–30 year

Assessing demand for clean and safe domestic water in eastern Zimbabwe

237

age group had the lowest SPP and MWTP values. Chitakatira had lower WTP values than Nyanyadzi, which lowered the SPP and MWTP values. Gender did not significantly affect the SPP/MWTP values although males had higher SPP and MWTP values than their female counterparts. Season did not have a significant effect on the SPP and MWTP values for the establishment of a new deep well. 4 DISCUSSION The evidence gathered in this study showed interesting perceptions held by rural people in relation to WTP for clean and safe domestic water. At a general level it can be said that there is no substance in the assertion that poor people do not want to pay for water. In Zimbabwe poor people have already begun to meet operational and maintenance costs in domestic water sources (Machingambi and Manzungu, 2003), and in publicly owned irrigation schemes, contrary to claims that the government maintained these schemes (Manzungu, 1999). Worldwide it has been documented that poor people tend to pay the highest amounts for domestic water (Cosgrove and Rijsberman, 2000). Where payments are not forthcoming the problem may be a lack of money rather than willingness to pay (Machingambi and Manzungu, 2003). Poor community mobilisation methodologies may also be another reason (Global Water Partnership, 2000). The study has also provided insights into specific issues concerning the supply of domestic water in rural areas, which may be of interest to policymakers and practitioners. It was clear that there was a high awareness of the potential danger caused by consumption of water containing bacteria. Respondents characterised clean and safe water as being free of bacteria. There was also a realisation of the likely causes of the contamination. This was shown by the fact that water was perceived to be generally unsafe in the wet season (hence lower WTP figures than for the dry season). Piped water had the highest WTP figures as it was rated the safest. River water was rated the most unsafe; it had the largest proportion of respondents not willing to pay anything for it in summer. Treating water had the effect of increasing MWTP figures. In some cases physical scarcity of water also affected SPP and MWTP. This explains why Nyanyadzi, the drier of the two regions, had respondents who were willing to pay higher amounts of money for their water than their counterparts in Chitakatira. The impact of physical scarcity of water on shaping the management of water resources is increasingly being acknowledged internationally. It is not absolute scarcity of water that is a problem but an economic scarcity regarding the availability of finances for the development and management of water resources (IWMI, 2000). This explains the paradox of a country like Zambia with more water resources than South Africa, but has a greater percentage of the population suffering from water scarcity more than the latter. Some commentators have also argued that water scarcity can lead to better adaptive capacities, which may mean the adoption of more intensive water uses (Turton and Ohlsson, 1999). This underlines the importance of analysing the role of social, cultural and economic factors in influencing willingness to pay for water. Socio-economic and cultural factors affected respondents’ WTP for water. The effect of the respondents’ economic circumstances on WTP for water was illustrated by the fact that economically dependent individuals’ (15–30 and the 75+ year age groups) were not

Water resources of arid areas

238

interested in contributing towards their water use. Their WTP was not affected by whether the water was safe or unsafe since they showed no interest in paying for water. Generally traditional leaders were not willing to pay for water because they considered themselves as owners of the water, underlining the role of cultural factors in influencing the demand for water. Perceptions about who owns water also affected WTP (Machingambi and Manzungu, 2003). However safety was a fundamental factor in influencing respondents WTP as even the traditional leaders who were unwilling to pay for water from any other water source wanted to pay for the “safe” piped water. Females were found to have lower WTP than males probably due to the fact that they normally do not handle finances in the home. They therefore tended to be more conservative regarding money issues than the males. However females had low WTP figures in relation to investing in new water sources to reduce labour upon women. The influence of physical, socio-economic and cultural factors on the WTP provided a basis for respondents to portray their degree of concern about access to reliable, safe and clean water, the ideal institutional context in which water could be provided and the way in which it would be financed. Addressing such issues constitutes a more holistic intervention in water issues affecting respondents, rather than merely focusing on cost recovery. 5 CONCLUSION Demographic characteristics of respondents are important in influencing WTP for clean and safe domestic water. Cost recovery policies should therefore be related to demographic characteristics of the intended beneficiaries. While demand for reliable, accessible, clean and safe water was shown to exist in rural areas, success of cost recovery policies depends on the ability to pay. ACKNOWLEDGEMENTS The authors wish to thank the Water Research Fund for Southern Africa (WARFSA) for providing the grant that made the study possible. Mr Chimedza of the University of Zimbabwe’s Department of Statistics is greatly acknowledged for the assistance with data entry and analysis. REFERENCES CSO, 1992. Central Statistics Office, Census 1992, Provincial Profile: Manicaland. Government Printers. Chenje, M. & Johnson, P. (eds.) 1996. Water in Southern Africa, SADC/IUCN/SARDC. Harare, Print Holdings. Cosgrove, W.J. & Rijsberman, F.R. 2000. World Water Vision: Making Water Everybody’s Business. London. Earthscan Publications Ltd. Global Water Partnership, 2000. Towards Water Security: A framework for Action. GWP, Stockholm.

Assessing demand for clean and safe domestic water in eastern Zimbabwe

239

Government of Zimbabwe, 1995. Poverty Assessment Study, Harare, Government Printers. International Water Management Institute, 2000. World Water Supply and Demand: 1995 to 2025. DTP Unit, IWMI—January 2000, Colombo. Machingambi, M. & Manzungu, E. 2003. An evaluation of rural communities’ water use patterns and preparedness to manage domestic water sources in Zimbabwe. Physics and Chemistry of the Earth: Water Demand Management for Sustainable Use of Water Resources, 28(20–27):1039– 1046. Pergamon Press. Makumbe, J.Mw, 1996. Participatory development: the case of Zimbabwe. Harare, University of Zimbabwe Publications. Manzungu, E. 1999. Strategies for smallholder irrigation management in Zimbabwe. PhD Thesis, Wageningen University, The Netherlands. Manzungu, E. 2001. A lost opportunity: The case of the water reform debate in the fourth parliament of Zimbabwe. Zambezia, XXVIII (i):97–119. Moyo, N.A.G. & Mtetwa, S. 2000. Water Quality Management Strategy for Zimbabwe. A paper prepared for the Ministry of Environment and Tourism, Harare. National Action Committee, 1997. Sustainability strategy for the National Rural Water Supply and Sanitation Programme, Government of Zimbabwe, Harare. Pearce, D., Markandya, A. & Barbier, E.B. 1989. Blueprint for a green economy. London, Earthscan Publications. Savenije, H. & van der Zaag, P. 2002. Water as an economic good and demand management: paradigms with pitfalls. Water International, 27(1):98–104. Turton, A.R. & Ohlsson, L. 1999. Water scarcity and social adaptive capacity: Towards an understanding of the social dynamics of managing water scarcity in developing countries. Paper presented at the Workshop on Water and Social Stability at the 9th Stockholm Water Symposium on Urban Stability through Integrated Water-Related Management, hosted Stockholm Water Institute (SIWI), Stockholm, Sweden, 9–12 August. www.soas.ac.uk/Geography/WaterIssues/OccasionalPapers/home.html Vincent, V. & Thomas, R.G. 1960. An agricultural survey of Southern Rhodesia: Part I—agroecological survey. Salisbury, Government Printer.

The role of supplementary irrigation for food production in a semi-arid country—Palestine Mohammed Yousef Sbeih Irrigation Project Coordinator, American Near East Refugee Aid (ANERA), Ramallah, West Bank, Palestine Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Palestine consists of the West Bank and the Gaza Strip. The proclaimed state of Palestine has a land area of 6657km2. Water is considered an essential factor of life and needs to be developed in arid countries. Reuse of treated wastewater for irrigation as supplementary irrigation will increase the irrigated area in Palestine and replace fresh water.

1 INTRODUCTION Palestine consists of the West Bank and the Gaza Strip. The proclaimed state of Palestine has a land area of 6657km2. Water is always considered as an essential factor of life and development in arid and semi-arid countries. In Palestine the total per capita water consumption is 139m3. The total available water for Irrigation is 239 M.C.M. which is responsible for irrigating only 330000 dunums out of 2314.000 dunums cultivated that can be irrigated if water is available i.e. 5% of the total cultivated land. The average rainfall is 450mm and unfortunately there isn’t any water harvesting structures i.e. dams, most of this rainwater flowing towards the Dead Sea or the Mediterranean Sea as waste. So harvesting this water in individual farmer land and using this water for supplementary irrigation to irrigate olive trees, almonds, grapes and cereals will be of a great impact on the Palestinian land for feed production. It should be noted that there are few farmers who practice supplementary irrigation for production of vegetables that are planted in summer as individual initiative. The quantity and quality of production that they have is extremely tangible. Since most of the land in Palestine is planted by olive, grape, and cereals, supplementary irrigation should be introduced and practiced where the production of wheat via irrigation by treated wastewater was three times that under rain fed planting project implemented in a pilot project.

The role of supplementary irrigation for food production in a semi-arid country-Palestine

241

Reuse of treated wastewater for irrigation as supplementary irrigation will increase the irrigated area in Palestine and will replace the fresh water that can be used for domestic purposes. 2 THE NEED FOR SUPPLEMENTARY IRRIGATION IN PALESTINE As it was mentioned before, Palestine is a semi-arid country, where the average rainfall is 450mm. The availability of water is questionable. Furthermore, the availability of water for agriculture is reducing in a tangible way due to the following: 1. The normal increase in growth rate, the population of the country is increasing, so the demand for domestic water is also increasing. This will affect the availability of water for agriculture. 2. Since rainwater is the only source of water, the quantity of rainwater (rainfall) has been decreasing in the recent years. 3. There is a huge conflict on water issues at this stage between the Palestinians and the Israelis since Israel occupied Palestine. It should be mentioned here that during early negotiations in the peace process, four main issues have been delayed since 1992; they are Jerusalem, refugees, water and borders. Still after 8 years of negotiations, there hasn’t been any significant movement on these issues. So the quantity of water that can be available for the Palestinians will probably not be increased. 4. The quality of ground water wells especially in Gaza and Jericho becomes saline and shortly it cannot be safely available for agriculture. From the above, it seems that extra availability of water for additional irrigated area or even to sustain the irrigated area is not an easy task. Total cultivated area in the West Bank is 2100.00 dunums, but the irrigated area is 110000 dunums. From the small experience (pilot project) for this field as well as other country experience i.e. Syria. It has been proven that the production of crops under supplementary irrigation is 3 times higher than under rain fed crop, in addition to the increase in the quality of the product. So if supplementary irrigation has been practiced we can easily increase the production of rained crops to three times or twice. This will play a major role in providing food for the people and even exports can take place and the net income of the country will be increased. 3 BACKGROUND It is foreseen that the world’s food production has to be doubled in the next 25 years, and thus, the agriculture continues to be an important sector in the 21st century. Meanwhile, the agriculture sector remains the largest user of the water resources, and it is evident that there is a decline of agricultural water due to increasing demands from cities, industries, and hydropower utilities in the developing countries such as Asia. Much of the water has to come from irrigation water savings.

Water resources of arid areas

242

Population and economic growth in many developing countries of Asia have created serious problems, such as the shortage of food, the scarcity of water, and the deterioration of the environment. Some of the irrigation and drainage projects have been seriously criticized due to their high-cost and low-efficiency for the construction and maintenance. The concept of maximum yield is now changing to optimum yield for creating an efficient irrigation schedule. The water saving is the most sustainable conservation, because it reduces the new construction needs to meet the increased water demand. The major issues of agricultural water are how to increase withdrawals about 15–20% by water saving, how to increase storages 10–15% by new irrigation facilities, and how to conserve the water quality of irrigation. 4 SUPPLEMENTAL IRRIGATION 4.1 Definition ICARDA defines supplemental irrigation (SI) as; the addition of essentially rain fed crops of small amounts of water during times when rainfall fails to provide sufficient moisture for normal plant growth, in order to improve and stabilize yields. Accordingly, the concept of SI in areas having limited water resources is built on three bases: First: water is applied to rain fed crops, that would normally produce some yield without irrigation; Second: since precipitation is the principal source of moisture for rain fed crops, SI is only applied when precipitation fails to provide essential moisture for improved and stabilized production and; Third: the amount and timing of SI are not meant to provide moisture stress-free conditions rather to provide minimum water during the critical stages of crop growth to ensure optimal instead of maximum yield. The management of supplemental irrigation is seen as a reverse case of full or conventional irrigation (FI). In the latter the principal source of moisture is the fully controlled irrigation water, while the highly variable limited precipitation is only supplementary. Unlike FI the management of SI is dependent on the precipitation as a basic source of water for crops grown. Water resources for supplemental irrigation are mainly surface, but shallow ground water aquifers are being increasingly used lately. Non-conventional water resources are of a potential for the future, but an important one emerging is water harvesting (Dwas 2001). 4.2 Improving production with SI Research results from ICARDA and other institutions in the dry areas as well as harvest from farmers showed substantial increases in crop yields in response to the application of relatively small amounts of supplemental irrigation. This increase covers cases with low as well as high rainfall. Average increases in wheat grain yield under low, medium and high annual rainfall in Tel Hadya reached about 400%, 150% and 30% using amounts of

The role of supplementary irrigation for food production in a semi-arid country-Palestine

243

SI of about 180, 125 and 75mm respectively. Generally, optimal SI amounts range from 75mm to 250mm in areas with annual rainfall between 500 to 250mm, respectively. Determining the optimal amount under various conditions will be discussed later (Oweis 2001). When rainfall is low, more water is needed but the response is greater, but increases in yield are remarkable even when rainfall is as high as 500mm. The response was found to be higher when rain distribution over the season is poor. However, in all rain fed areas of the region it was found that some time in the spring there is usually a period of stress, which threatens, yield levels. This soil moisture stress usually starts in March, April of May, if total annual rainfall received is low, average or high respectively (Oweis 2001). In Syria average wheat yields under rain fed conditions are only 1.25t/ha and this is one of the highest in the region. With SI the average grain yield was up to 3t/ha. In 1996 over 40% of rain fed areas were under SI and over half of the 4 mil tons national production was attributed to this practice. Supplemental irrigation does not only increase yield but also stabilizes farmer’s production. The coefficient of variation in rain fed production in Syria was reduced from 100% to 10$ when SI was practices. This is of special socio-economic importance since it affects farmer’s income (Oweis 2001). 4.2.1 Introduction Historical Palestine is located between the Mediterranean Sea and the Jordan River, as well as to the Red Sea from the south. The present proposed Palestinian state consists of West Bank and Gaza Strip. The other part of Palestine is occupied by Israel in 1948. This study focuses on the West Bank and the Gaza Strip. The proclaimed state of Palestine has a land area of 6657 square kilometres (Kateeb 1993). Population senses has been taken place recently by the Bureau of Statistics early 1998. It is reported that the population of the West Bank is 1571571 and Gaza Strip is 963026 where the total population of the Palestinian people is 2534598 people. Ground water is the main water source in the country. It is recharged by rainfall. Rainfall varies from 100mm in the south east to 800mm in the north. The average rainfall is 550mm (Sbeih 1995). Where the average rainfall in Jordan Valley is from 100mm to 270mm/year (Zaru 1992), and in Gaza is 200–400mm/year (Abu Safieh 1991). Not all the rainwater is available to the Palestinian due to Israeli Military orders. Water is abstracted from the ground water through 340 wells in the West Bank and 1781 wells in Gaza. In addition to that springs contribute a lot, where half of the irrigation water in the West Bank is due to springs. The quality of the available water varies from almost rain water to brackish water. In the Jordan Valley where it is the lowest point in the elevation in the world where temperature is very high in this area especially in summer. As example, the chloride content is reaching 68mg/l and the SAR reaches 11.7 where the TDS reaches 5000PPM. Still the utilization of this saline water is not as efficient and environmentally safe as it should be where further utilization of this water could play a major role in developing the area where still the irrigated area consists of not more than 6% of the cultivated area in the West Bank.

Water resources of arid areas

244

It should be mentioned that not only saline water does already exist and utilized unproperly, but it also seems to be that the additional water that can be allocated for irrigation is also saline water which is going to be from: 1. The Eastern aquifer to be used in Jordan Valley. 2. From the treated waste water from different cities and villages in the West Bank. 4.2.2 Water sources in occupied Palestine 4.2.2.1 West Bank Two main water sources are available for Palestinian in the occupied Palestine (West Bank and Gaza Strip) for agricultural, domestic and industrial use. These are rainfalls and ground water sources—Palestinians consume water mainly through ground water wells and springs (where rainfall is considered the main recharge). The total annual water springs discharge varies according to the rainfall. The total annual flow of the 113 fresh water springs in the West Bank ranges between 24 M.C.M. (as in the year 1978/79) to 119.9 M.C.M. (as in the year 199/92) and with an average of 52.9 M.C.M. as calculated from the annual flow in the past 24 years. Around 86% of the total annual flow of these 113 springs is within the eastern drainage (in/or toward the Jordan Valley), while the other 14% is within the western and south-west (Nusseibeh 1995) where the total estimated annual water discharge from ground wells is 60 M.C.M. (Awartani 1992). So that the total annual water available to Palestinian is 113 M.C.M. In addition to that there is another 2.5 M.C.M. is collected directly from the rainfall in cisterns in Palestinian houses. So that the total available water is 116 M.C.M./year, for more information see Table 1. 4.2.2.2 Gaza Strip Water situation in Gaza Strip is very critical. The Gaza Strip lies on top of two water strata. The upper is fresh water, the lower carries saline water. The annual consumption of water is at present in the vicinity of 100 M.C.M. These aquifers get replenishment of some 60% leaving a deficit of 40 M.C.M. of water (Shawwa 1991). Even the Gaza water is lower in quality than West Bank, but due to the complication of the situation there and due to the geographic location where my work is more in the West Bank. This paper will address West Bank issues more clearly.

Table 1. Basic land and water indicators for Israel and the occupied Palestinian and other Arab territories. West Bank

Gaza Strip

Israel

Total area 5573000 360000 20000000 (dunums) Population (1988) 900000 600000 4300000

The role of supplementary irrigation for food production in a semi-arid country-Palestine

245

Area of land 2100000 214000 4250000 cultivated (dunums) Area of land 110000 120000 850000 irrigated (dunums) Percentage of total 5 56 44 irrigated land (%) Percentage of total 38 59 21 land cultivated (%) 95 80 1320 Annual water consumption for irrigation (million m3) 27 21 325 Annual water consumption for households (million m3) 3 2 125 Annual water consumption for industry (million m3) Total annual water 125 103 1770 consumption (million m3) Total per capita 139 172 411 water consumption (m3) 35 75 Per capita water 30 consumption per household (m3) Per capita water 3.3 3.3 29 consumption for industry (m3) Per capita water 106 133 307 consumption for irrigation (m3) 1 dunum=1000m2. Source: Israeli land and water policies and practices in the occupied Palestinian and Arab territories, unpublished study in Arabic (Economic and Social Commission for Western Asia, Baghdad, 1990), p. 8.

4.2.3 Irrigated areas in the occupied Palestine In Palestine, being a semi arid country, we are confronted by a demographic growth, and agricultural development as well livestock and industrial development. Thus in essential growing water requirement makes the rational management of water resources supremely important in order for development to be lasting and for environment to be served.

Water resources of arid areas

246

On a global basis at least 60% of all water abstracted at present is used for agricultural production. In Palestine 70% of all water consumed is due to agriculture. Here in Palestine, agriculture is considered to be one of the main national income. Agricultural production contributes 47.61% of the total national income in 1970. The potential for irrigation to raise both agricultural productivity and the living standards of the rural poor has long been recognized. Irrigated agriculture occupies approximately 17% of the world’s total available land but the production from this land comprises about 34% of the world total. In Palestine, irrigation is considered to be the spinal chord of plant production for the following reasons: 1. Palestine is considered as a semi arid region where some of the crops cannot be grown without irrigation (example, citrus). 2. In the Jordan Valley, which constitutes the main agricultural production for the country, irrigation is a must due to low rainfall and high temperature. 3. With irrigation the same plot of land can be planted up to three times per year while it cannot be planted more than two times with dry farming. 4. Different varieties and crops can be planted in any region due to the availability of water i.e. more flexibility of planting several crops at different regions in different times of the year. 5. Job creation: Since the labour requirement per irrigated dunum is more than double that of job required per dry farming per one season. This has now become more vital due to continuous of closures of the West Bank and Gaza Strip where the number of labourers that are working in the Palestinian part that occupied in 1948 is sharply reduced. 6. Agricultural production is much higher for irrigated farming than for dry farming per dunum per season. As example average tomato production per dunum is as follows: – Dry farming: 2–3 ton per dunum per season. – Irrigated (open land) 6–8 ton per dunum per season – Irrigated (greenhouse) 12–16 ton per dunum per season 7. Net income per dunum of dry farming does not exceed $150 while from irrigated area the net income can exceed $1500 per dunum. 8. Especially in Palestine, where the horizontal expansion in agriculture by increasing the total cultivated area due to the Israeli occupation, and shortage of water. The vertical expansion could be the main parameter to play with. Irrigation will be the main element in this formula. So that providing extra water for irrigation to irrigate as much as possible of the cultivated area is a must. This implies that Palestinian should use any drop of water. Regardless the quality of that water practically and efficiently: Table 2 shows the irrigated area in each district in Palestine where the total irrigated area in 1993–94 was 217000 dunum (PSBS 1996).

The role of supplementary irrigation for food production in a semi-arid country-Palestine

247

Table 2. Distribution of area that could be irrigated in the West Bank (Source: Awartani, 1991). Location

Dunum

Plains in Jenin and Tulkarem Highland Eastern slopes Jordan Valley Total

99600 27740 64.6 93.5 535.1

4.2.4 Available area that is ready for irrigation Where in Gaza Strip the irrigated area could be doubled or tripled in terms of topographical situation but due to the limitation of the water both quality and quantity it is very difficult to increase the irrigated area while in the West Bank the area that could be irrigated in terms of topographical conditions estimated to be 535 thousand dunums (Awartani 1991) as in Table 2. Where in the study conducted by PWA in 1992 in order to develop a plan for the western Ghore the following locations could be the most suitable area to be ready for irrigation. 4.2.4.1 Northern Ghore The areas suitable for irrigated agriculture in this region include: 18000 dunums in Ein Al Beida, Bardalla villages 5300 dunums in the Ghore 3500 dunums in the Ghore But the Ghore and Zhor areas are mostly closed by the Israeli Military orders. 4.2.4.2 El-Bique Valley This is a large flat area to the west of the hills of northern Ghore. This area includes about 18500 dunums of fertile smooth deep soil. The Palestinian farmers as rainfed excluding 5500 dunums where the two settlements their (Baquat and Roi) are occupying cultivate all this area. 4.2.4.3 Upper El Fara’ Valley area (Semi-Ghore) In this area, there are 13100 dunums that are suitable for irrigation and can be easily irrigated as follows: Sahl Tubas Sahel Tayassear Sahel Tammun Sahel El Fara’ El Nassarieh (additional)

3600 dunums 900 dunums 1900 dunums 5000 dunums 1700 dunums

Water resources of arid areas

248

Where there are another 7000 dunums, which are already irrigated. 4.2.4.4 The middle and south Ghore This region extends from approximately grid north 180 (northern of Marj Najeh) in the north to the Dead Sea in the south and from the Jordan River in the east to the feet of the west-bank mountains. The total area that could be ready for irrigation in this area is 145500 dunums. In summary, the total area that can be used in irrigated agriculture in the western Ghore will be: Northern Ghore Biquia Valley Semi-Ghore Southern Ghore Total

26800 dunums 18500 dunums 201000 dunums 145500 dunums 210900 dunums

Where about 44000 (PCBC 1991) dunums of this area is currently irrigated. So the total additional area that could be irrigated in the West Bank is (210900−44 000+(535100−935000)= 608500 dunums. It should be mentioned that the Jordan Valley produces more than 59% of the vegetables produced in the West Bank. It also produces 100% of the bananas produced in Palestine. 5 PALESTINIAN EXPERIENCE OF SUPPLEMENTARY IRRIGATION Still the term supplementary irrigation is not even used formally and officially in Palestine. Until this time there is not any plan of implementing any project of supplementary irrigation. This is mainly due to the lack of qualified staff at the Ministry of Agriculture as well as to the lack of great interest to agriculture from M.O.A. due to the following reasons: 1. The lack of responsibility of the Palestinian Authority on most of the agricultural land due to the occupation. 2. The lack of finance and funding to development projects. Nevertheless, there are individuals who attempt to use supplementary irrigation, an example of that are few farmers in Sinjel town in the Ramallah area. 6 DESCRIPTION OF AGRICULTURAL AREA IN SINJEL This village is located just between Ramallah and Nablus cities, situated 20km to the north of Ramallah. The total agricultural area in the village excess 4000 dunums, out of these areas. About 1000 dunums are plain and flat.

The role of supplementary irrigation for food production in a semi-arid country-Palestine

249

This 1000 dunums is planted with vegetables in summer and cereals in winter. All of this area is rain fed, there are no source of water for irrigation since this area is located close to the village (houses), it is easy for the farmers to bring water by mobile tanks. Usually the farmers in summer, bring some water and store them in a container (barrel) of 200 liter capacity each, since the ownership of land is between 3–5dunums, the number of barrels used are 6–8. In summer farmers used to plant vegetables, at the time of planting the seedlings, farmers used to irrigate the seedling by a bucket. Farmers used to mix the fertilizer water and irrigation at the time of planting the seedlings. Later on, after 20 days the second irrigation with fertilizer is applied. The third one and the last one are provided with fertilizer before flowering. The total amount of water applied per each plant is not more than 1 liter, for a dunum of 1000 plants, 1000 liter is applied 1 cubic meter of water applied for the whole season per one dunum. While for the irrigated area the minimum irrigation water requirement is 70m3/dunum per the season. In this village, Sinjel, and through my investigation, in the year 2000 I found 3 farmers who are using this approach technology, when I asked one of them what is the result that you will expect, he broadly replied: 1. The quality of agricultural product that I used to obtain for the last two years where I used to use supplementary irrigation is much better than the product of my neighbor in the same plot of land in the village, so the price per 1kg. That I got is much higher also. 2. The total production is much higher than that of my neighbor, i.e. I got 4 tons each per dunum, my neighbor got 2 tons of squash per dunum. 3. The period of production that I have is much bigger than that of my neighbor has, this means that total income that I gained is much higher. I used the produce vegetables for 2 months, while my neighbor only one month, i.e. the harvesting period is much higher when supplementary irrigation used. I informed this farmer that I am working on an irrigation project coordinator for an NGO that provides funds for farmers. Since this farmer believes that he was happy from his production since he has only 3dunums and all of his family working in this plot of land, he did not ask what service that since that we offered, this totally indicated that he is happy, and he did not need any further assistance. At that time there was visiting irrigation professor from Canada. This professor told me that we should use him as a model to encourage people using appropriate technology. Another example of using supplementary irrigation is found in Hebron where a farmer from Al Tamimi family, who has a grape field and luckily a pipe water pass through his field and used to get some water from this pipe and provide some water for his grape. In winter since the rainfall in Hebron is not exceeding 300mm, as well as in July.

Water resources of arid areas

250

Table 3. Results of El Bireh wastewater treatment pilot plant using treated wastewater. Treatment

Production of wheat (anber variety), all the plants (kg/dunum)

Irrigation with treated wastewater with fertilizer Irrigation with treated wastewater without fertlizer Without irrigation, with fertilizer Without irrigation, without fertilizer

2520

20036

1600 572

It is well known in Hebron, that the quality of grape of that man is the best in Hebron, since Hebron is of the biggest producing city (country) in Palestine. Since the municipality constructed a pilot treatment plant, it thought of planting crops using the treated effluent. This was funded by American Near East Refugee Aid (ANERA). Three crops were selected by the Agriculture Department to be planted for the first time in Palestine using treated wastewater: ● Artichokes on 150m2—planted on October 31, 1993. ● Onion frozen production on 500m2—planted on November 6, 1993. ● Wheat on 1000m2—planted on November 22, 1993. ● Drip irrigation as well as sprayers were used. Several treatments were made as follows: ● Irrigation with wastewater used, fertilization was used. ● Same as above, but without application of fertilization. ● Irrigation not used but fertilization was used. ● No irrigation and no fertilization (dry land farming). All the agricultural practices were used (pesticides, ploughing, seed control, etc. Table 3 shows the production of each kind of treatment. The impact of using treated wastewater appears clear. Notes: 1. Time of planting was October 1993; all the crops received rainfall during the growing period. 2. Time of harvesting was June 2, 1994. 3. Production with irrigation with treated wastewater with fertilization was five times without irrigation and fertilization. 4. Production increased the soil when irrigated with treated wastewater where fertilization was applied on both cases (irrigated and non-irrigated).

The role of supplementary irrigation for food production in a semi-arid country-Palestine

251

7 METHODOLOGY OF PRACTISING SUPPLEMENTARY IRRIGATION IN PALESTINE Since the ownership of land is very small in size i.e. from 5–10 dunums, supplementary irrigation can be easily implemented for vegetables, trees and to cereals to some extent constructing of small ponds of 40–50m3 capacity, i.e. this pond can be located on a 14–18 meter square area. This pond can be located on the lowest point in elevation of the individual land. This land serves two farmers if agreed upon where it can be sited on the border of each farmer land. Distributing of water to the plant can be done manually by lifting the water and distributing it to the plants by a bucket. Another way of distributing this water that this water can be lifted manually from the pond and poured into a barrel that can be located on the dip of the pond with 1/2meter raised over the surface so water can be distributed to the plant by gravity through pipe line. The farmer can distribute the water pipe from the plant to another. These methods can be implemented

Table 4. Results of Al Beireh Pilot Wastewater Treatment, 1994.

Crop Wheat 870 type

Kind of treatment

Production (kg/dunum) Seed Hay

Irrigation with 687.5 fertilizer Irrigation 656.70 without fertilizer Rainfed with Rainfed with 537.5 fertilizer fertilizer Rainfed without 500 fertilizer Wheat anber Irrigation with 864 type fertilizer Irrigation 824 without fertilizer Fertilizer Rainfed with 600 fertilizer Rainfed without 236 fertilizer

1375 1373

1187.5 1531.25 1656 1212

1000 336

easily with zero operation cost. Since only the farmer himself can conduct this job easily, another method of distributing water is by using a small pumped electricity is available since the head required is very small. In the case of cereals water can be distributed easily by establishing ponds, so water can be discharged into the farm then water can flow by gravity. In order to reduce the

Water resources of arid areas

252

cost of pumping farms can cooperate between themselves when each farmer can construct his pond on the highest point in elevation on his land. His pond can receive water from his neighbour’s field and so on. 8 THE ECONOMY OF SUPPLEMENTARY To construct a pond of 50m3 the following is needed with estimated costs: 1. 2. 3. 4.

Excavation of 50m3 Construction works Plastering Parallel, pipes, buckets Total estimatated cost:

=$3900 =$2000 =$500 =$120 =$3100

9 REVENUES Assume a plot of land of 5 dunums planted with vegetables. The production of vegetables of rainfed per dunum is 3 tons/dunum, the production of dunum with supplementary irrigation is 4.1 ton. The price per ton is $200 for rainfed crops.

The income per supplementary irrigation is 4×250=$1000.

The price per ton for supplementary irrigation is $250.

The net income due to supplementary irrigation will be 1000?=400 per dunum.

S the income per rainfed dunum=3×200=$600.

5 dunums×400=2000 per session per 5 dunums.

10 CONCLUSION AND RECOMMENDATIONS 1. In Palestine the total cultivated area is 2314000 dunums, while the irrigated area is 230000 dunums, so any efforts for increasing the productivity of the cultivated area should be considered due to the large area, while the production of the irrigated area is on its maximum. 2. Providing of extra water or even to sustain the existing water for both irrigation and domestic purposes is questionable due to the increase demand for domestic purposes first and due to the Palestinian-Israeli water conflict. 3. Practicing supplementary irrigation is not costly and did not need that much complicated technology. 4. The irrigated area only represents 6% of the cultivated area, where the land that can be easily irrigated is estimated to be 608600 dunums. In the West Bank only, which is 6 times the land that is already irrigated but water is needed. 5. The salinity of the ground water is deteriorated by time due to over pumping, sea intrusion and the low rainfall especially in the Jordan Valley and in the Gaza Strip, so providing fresh water for irrigation is questionable.

The role of supplementary irrigation for food production in a semi-arid country-Palestine

253

6. The additional water that will be available for the Palestinians will be either from (a) Eastern aquifer, (b) Jordan River, or (c) Treated wastewater. Where all of this water is saline water, where there are another source such as the mountain aquifers, but this seems to be difficult to be secured soon. 7. The early possible of expansion in irrigation will be in Jordan Valley where the existing water wells and the future water that might be available is saline. 8. Since the treated water is in the full control of the Palestinians, more attention and care should be paid in order to better and safe utilize of this water for developing the agricultural sector in Palestine, and this water can be used for supplementary irrigation. 9. The productivity of one cubic meter of water with supplementary irrigation is much higher than that of irrigated land since the water prepared by irrigated dunum is 7 times more than the required for supplementary irrigation. 10. The existing irrigated area is already exhausted since this land used to be planted two or three times a year where the other land used to be cultivated once a year even it kept fallow on some years. 11. Palestinian Agricultural Ministry and Palestinian Water Authority should recognize the situation and consider supplementary irrigation as a major element for food supply.

REFERENCES Abdul Jabar, Q. 1996. Chemical analysis of Jericho wells. PhD, Jerusalem. Abu Arafeh, A. 1894. Jordan Valley, Jerusalem, published by Arab Studies Society. Al Khateeb, N. Palestinian water supplies and demands. A proposal for the development of a required water master plan, IPC, Jerusalem. Awartani, H. 1991. Irrigated Agriculture in the Occupied Palestinian Territories, Al Najah National Univseristy, Nablus. Awartani, H. 1992. Groundwater wells in the Occupied Palestinian Territories, PhD, Jerusalem. ICID, 2001. General Report to the First Asian Regional Conference of International Commission on Irrigation and drainage. Agricultural, Water and Environment, Seoul, Korea. Nusseibeh, M. & Nasser Eddin, T. 1995. Palestinian Fresh Water Springs, Jerusalem, Palestine. Palestinian Water Authority, 1992, Water Development Plan, Jerusalem, Palestine. Sbeih, M. 1995. Recycling of treated wastewater in Palestine: Urgency, obstacles, and experience to date, Elsevier. Oweis, T. 2001. Supplemental Irrigation for field crops, water saving and increasing water productivity: challenges and options. University of Jordan, Amman, Jordan.

Conversion of priority water rights to proportional water permits and conflict management in the Mupfure river catchment, Zimbabwe Tamsanqa Mpala Scientific & Industrial Research & Development Center (SIRDC), Harare, Zimbabwe Water Resources of Arid Areas–Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: The Zimbabwe Water Act of 1976 was repealed and replaced by the Water Act of 1998. The Water Act of 1998 (Chapter 20:24) came into force on the 1st January 2000 and with it came the abolishment of the prior date system of water allocation. In the promulgation of the new Water Act, existing water rights used under the priority date system were to be converted to proportional water permits. The objective was to promote equitable access to water resources for all Zimbabweans and to encourage the sustainable utilization of the resource. The challenge for the newly established catchment councils was therefore to determine all existing water rights in their respective catchments and conduct the conversion exercise based on water generated in the catchment and water use by water right holders. This study analyzed the existing water rights and water generation for the Mswenzi River catchment and conducted the conversion to water permits with the Sanyati Catchment Manager.

1 INTRODUCTION Any resource, such as water, when used by more than one user in a single catchment or river basin, tends to attract conflicts about how it is shared and distributed. This paper focuses to a large extent on a highly committed small river catchment; namely, the Mswenzi River Catchment that forms part of a greater basin, which eventually drains off into the Zambezi River and into the Indian Ocean. The Mswenzi is a tributary of the Shuru Shuru River in the northwestern region of Zimbabwe. The Shuru Shuru River in turn is a tributary of the Mupfure River that drains northwesterly and into the Zambezi River. The study catchment has a total surface area of approximately

Conversion of priority water rights to proportional water permits and conflict management

255

158km2 with the Mswenzi River stretching a distance of 22km. The study catchment generates on average an estimated 70mm of surface water per year, or 11.0*106m3/year. Hence, this paper looks into the operationalization of water rights in the Mswenzi river catchment and gives an analytical recommendation of converting the old water rights used under the Zimbabwe Water Act (1976) to new water permits as recommended under the new Zimbabwe Water Act of 1998. A water right, for the purpose of this study, has been defined as a right to use beneficially a certain volume of water expressed in absolute volumetric units per time unit, whilst a water permit is a permit or allowance for the use of water, which specifies and restricts the use of water allocated. At the same time the paper attempts to bring out the conflicts involved in the Mswenzi catchment between upstream and downstream commercial water users. The paper discusses the legislative policies by looking at the old and new Water Acts and highlights the important principles that will govern conflict management and help spells out recommendations for water authorities and catchment management institutions. 2 THE WATER ACTS In Zimbabwe, the Water Act of 1976, which was repealed by the new Water Act of 1998, vested all public water in the President and private water was water belonging to the owner of land on which it was found. The right to use water was dependent on the type of use. For primary use no right was required. Access of water for non-primary use was based on the prior appropriation doctrine, where an appropriated right was based on the application of the appropriated public water to some beneficial use. The granting of any water right was the exclusive function of the Administrative Court sitting as the Water Court. The right would only be granted if public water was available and if it could be ascertained that the water would be put to beneficial use. The right granted was dependent on the date on which the application for the right was made. This date determined the applicant’s priority in the use of water applied for. The new Zimbabwe Water Act of 1998, which replaces the old Water Act (1976) sought to bring about equal and fair distribution of the available water resources in the national interest for the development of the rural, urban, industrial, mining and agricultural sectors. The major principles of the new Water Act were that all water would now be owned by the State and any use of it other than for purposes of primary use should be approved by the State. All stakeholders should be involved in decision-making processes and contribute to sustainable management of water resources. Water resources would be managed at catchment and sub-catchment levels, and the environment would also be considered a legitimate user of water. One of the important changes of the Water Act that is brought out in this study is that the priority date system of first come first served of water allocation was abolished to enable the principle of equitable access to water and sharing of water at all times. A fractional allocation system is now the recommended allocation system for non-primary water use.

Water resources of arid areas

256

3 THE CONVERSION EXERCISE The Mswenzi River Catchment has an area of approximately 160km2, and on average generates 11.0*106m3/a of blue water. Water generated in the catchment is based on the annual unit runoff and the catchment area, where the water generated (m3/a), is a product of the two. Mean unit runoff for the catchment is 70mm/year and the calculated catchment area is 158.15km2. This gives an amount of 11.0*106m3/a as water generated in the catchment. The Mswenzi River catchment has a total of 17 existing water rights owned by mainly commercial farmers, who have built reservoirs with a total capacity of 7.3*106m3. Among the 17 water rights, the total commitment level of the catchment is 65%. This is water used, (m3/a) as a fraction of water generated (m3/a). The major dams in the catchment are Balwearie and Tawstock dams, which have a combined capacity of 4.9*106m3, and serve a total of 8 water rights to various farmers in the catchment. The other 9 water rights are served from smaller dams along the Mswenzi River. Table 1 shows existing water rights and current users in the catchment as well as the priority date for each property. The procedures for the conversion of water rights to water permits were that, (i) the priority date attached to each water right be removed, (ii) volumes of flow and storage rights as allocated in the old system remain the same and shall be used as permits until such a time as the water authorities see fit to amend or revise the permit, and that (iii) water rights be converted according to the applicants ability to beneficially use the water. In the conversion exercise a simple formula has been recommended for the purposes of the study catchment, which follows: Permit (m3/a), P=[S1+S2+S3…]+F1+F2… Where, S=storage (m3/a) abstraction from stored water. F=flow (m3/a) abstraction from river flow. The storage permit is the volume of water permitted to be stored and used while a flow permit is the volume of water, which can be abstracted and used directly from the river. The validity of the permits, is for a limited period of 20 years, thereafter they are reviewed.

Table 1. Review of existing water rights and current users in the Mswenzi catchment. Water Property right number 7573 8881

Kasama

River Priority Abstraction Abstraction Dam right from right from flow storage (103m3/a) (103m3/a)

Mswenzi 19/10/66 tributary Dodington Mswenzi 06/03/70 tributary

0

45 Weir

0

136 Farm dam

Conversion of priority water rights to proportional water permits and conflict management

10626

12007

Rem. of Mswenzi 31/10/73 Luton Balwearie Mswenzi 31/05/73 Rem. of Mswenzi 05/10/83 Luton Balwearie Mswenzi 01/06/81 Balwearie Mswenzi 14/11/73 Strathspey Mswenzi 28/05/49 Strathspey Mswenzi 23/09/70 Strathspey Mswenzi 24/10/73 Handley Mswenzi 16/01/80 cross Cornucopia Mswenzi 24/10/73 Handley Mswenzi 24/10/73 cross Merchiston Mswenzi 28/01/81

14322

Merchiston Mswenzi 20/10/88

0

14321

Merchiston Mswenzi 29/10/90

0

12014

Merchiston Mswenzi 20/02/81

0

10156 10626 10156 10156 2276 2276 2276 12398 10364 9101

Total Source: Field notes, 2002.

0

910 Balwearie

0 0

1040 Balwearie 520 Balwearie

0 450 0 0 0 0

130 Balwearie Balwearie 400 Tawstock 182 Tawstock 518 Tawstock 1530 Suri Suri canal 523 Tawstock 714 Tawstock

0 0 0

450

257

611 Farm dam 20 Farm dam 5 Farm dam 20 Farm dam 7304

Balwearie Farm, which is served by Balwearie Dam, utilizes an estimated 1.59*106m3/a of water for its various crops under irrigation, hence the permit using the above formula will be: P=[1040*103m3/a+130*103m3/a] +450*103m3/a =1620*103m3/a for a validity of 20 years. Conversion for other properties in the catchment include: Kasama, P=[45*103m3/a] for a validity of 20 years. Dodington, P=[136*103m3/a] for a validity of 20 years. Rem. Of Luton, P=[910*103m3/a+520*103m3/a] =1430*103m3/a for a validity of 20 years. Strathspey, P=[(400+182+518)*103m3/a] =1100*103m3/a for a validity of 20

Water resources of arid areas

258

years. Handley Cross, P=[2244*103m3/a] for a validity of 20 years. Cornucopia, P=[523*103m3/a] for a validity of 20 years. Merchiston, P=[(611+20+5+20)*103m3] =656*103m3 for a validity of 20 years. The new permits as recommended for the study catchment after the conversion exercise is shown in the Table 2. 4 CONFLICT MANAGEMENT IN THE UPSTREAM, DOWNSTREAM CASE The case singles out the dispute between upstream and downstream commercial farmers holding water rights in the Balwearie and Tawstock Dams where the latter is downstream. Downstream

Table 2. Water permits for Mswenzi river. Water Property permit number 7573 8881 10626 10156 2276 12398 10364 12007 Total

Kasama

River Abstraction permit (103m3/a)

Mswenzi tributary Dodington Mswenzi tributary Rem. of Mswenzi Luton Balwearie Mswenzi Strathspey Mswenzi Handley Mswenzi cross Cornucopia Mswenzi Merchiston Mswenzi

Dam

45 Weir

Validity 20 years

136 Farm dam

20 years

1430 Balwearie

20 years

1620 Balwearie 20 years 1100 Tawstock 20 years 2244 Suri 20 years Suri/Tawstock 523 Tawstock 20 years 656 Farm dam 7754

farmers who held earlier water right priorities in Tawstock Dam were outraged that they were not receiving sufficient water from upstream Balwearie Dam and as a result jeopardized their operations. As earlier applicants, Tawstock farmers were entitled to water first, which meant that Balwearie farmers had to open the outlet gates at Balwearie Dam and release water for Tawstock farmers before the former could store and use any water. In 1982 An investigation came about as a result of a submission for a decision made to the Administrative Court by Tawstock farmers who possessed water rights no, 2276,

Conversion of priority water rights to proportional water permits and conflict management

259

10364 and 9101 of Tawstock Dam and who are referred to as the applicants. The applicants were concerned that the holders of water rights no, 10156, 10626 and 10659 (Balwearie Farmers) of Balwearie Dam who are referred to as the respondents, were unable to pass sufficient water from their storage works to satisfy the downstream priorities. The applicants maintained that the reason for this was that the outlet pipe of Balwearie Dam was of insufficient internal diameter for this purpose and therefore the operation of their prior rights was jeopardized. In this case the two dams have a similar capacity and are separated by a mere 2km where each dam has three participants in the utilization of the stored water. Two parts of water right no. 2276 have the earliest priority, after which water right no. 10156 has its turn. Then the remaining part of water right no. 2276 followed by the other two participants is satisfied. The other two participants in water right no. 10156 then follow each with separate priorities. This rather complex situation involves 6 separate water rights and 9 priorities. An agreement was reached at the Administrative Court between the applicants and the respondents. It was agreed that: ● A siphon of 12 inches diameter be constructed and installed together with an outlet pipe of not less than 12 inches diameter in Balwearie Dam. Both devices were to be used to pass water that flows into Balwearie Dam and down to Tawstock Dam to meet the entitlements of holders of water rights no, 2276, 9101 and 10364 together with the primary requirement of 85 liters per second. ● Not less than 425l/s will be released from Balwearie Dam, and the construction and installation of the siphon and the gauging weir immediately upstream of the headwaters of Balwearie Dam shall be carried out by Balwearie farmers so as to be in full operation.

5 SUMMARY The case has described the way in which the operation of priorities of this complex situation worked under the old system and how the issue of satisfying earlier priorities was resolved. In arbitration (a conflict management tool), the Tawstock farmers submitted their argument before the Administrative Court who acted as the judge and a solution was reached whereby both the Tawstock and Balwearie farmers signed an agreement. 6 RECOMMENDATIONS It is recommended that measures be taken to establish actual water use rather than assume values of water rights so that true commitment levels are achieved. It is also recommended that conflict management play a more important role in water resource management so as to empower local water authorities to handle such situations involving disputes over water allocation. It is hoped these recommendations will lay the platform for increased participation, negotiation and dialogue for better basin management.

Water resources of arid areas

260

7 CONCLUSION The main objective of this study was to establish how existing water rights in a small catchment were operated under the old system based on the 1976 Water Act and to describe the conversion process with the requirements of the permit system. The results from the study show that most storage rights were operated with staff gauges installed in the basin that enabled the stored volume and abstraction to be determined for any reservoir level. The study also showed that senior water rights consumed water impetuously without much consideration for downstream users and therefore new users found it difficult to receive a full entitlement of water allocated. The new system now allows new users the opportunity to be given an abstraction permit for their beneficial use therefore disregarding priority. The study revealed through a questionnaire that there was little cooperation and communication over data between upstream and downstream users that often resulted in disputes over water allocation. Of the seven farmers interviewed, six of them said they were not aware of the water reform. It is important therefore that water authorities and catchment agencies seriously consider the issue of enhancing dialogue and cooperation between different users and assist in the issuing of water permits to improve the management of water at catchment and basin level. The results of the study showed that the Mswenzi is a highly committed catchment with a total use of flow and storage water rights totaling 6900*103m3/a for all the water right holders in the catchment. The Mswenzi generates on average per year, 11070*103m3/a of water, bringing the water commitment of the catchment to 62.3%. The study showed that in the conversion process from the old water rights into new water permits, the permits would have to use the same volumes as previously granted for their water rights and discard the priority date. Therein catchment councils have the obligation of amending or revising the water permit according to beneficial use of the permits and accommodation of new entries among other criteria. Perceptions of the commercial farmers in the catchment have shown that the majority, almost 80%, of the big stakeholders in the catchment are unaware of the principles of the water reform and how their new water permits will be operated. Conflicts over water allocation have emerged under the old system due to misperceptions and lack of adequate data and it is anticipated that the new permit system will mitigate the grounds for conflicts in the future. REFERENCES Ashton, P. 1999. Southern African water conflicts: Are they inevitable or preventable? In: Water for Peace in the Middle East and Southern Africa, Geneva, Green Cross International, 2000. Biswas, A.K. 1993. Management of International Waters: Problems and Perspectives, Water Resources Development, 9. Heun, J. 1998. Water resources planning and analysis, Lecture Note, Dept. of Civil Engineering, University of Zimbabwe, Harare. Huffaker et al. 2000. The Role of Prior Appropriation in Allocating Water Resources into the 21st Century, Water Resource Development, 16.

Conversion of priority water rights to proportional water permits and conflict management

261

Jaspers et al. 1999. An external review of the Mupfure Catchment Integrated Water Management Project, Prepared for the Royal Netherlands Embassy, Harare. Lang, H. 1997. Options for a New Water Rights System. Draft Unpublished Paper, October. Manzungu, E. 1999a. Conflict management in the Umvumvumvu Catchment, In: Water for Agriculture in Zimbabwe: Policy and Management Options for the Smallholder Sector, University of Zimbabwe, Harare. Manzungu, E. 1999b. Strategies of smallholder irrigation management in Zimbabwe, PhD Thesis, Wageningen University, Netherlands. Manzungu, E, Senzanje, A & Van der Zaag, P. 1999. Water for Agriculture in Zimbabwe: Policy and Management Options for the Smallholder Sector. University of Zimbabwe Publications, Harare. Mazvimavi, D. 1998. Water Resources Management in the Water Catchment Board Pilot Areas, Phase I: Data Collection, CASS Technical Paper, NRM Series; CPN 95/98, University of Zimbabwe, Harare. Natsa, T.F. 1999. From priority date to fractional allocation: Towards equitable distribution of surface water resources in Zimbabwe. MSc Thesis, University of Zimbabwe, Harare. Resolve Inc. et al. 2000. Participation, Negotiation and Conflict Management in Large Dams Projects, Final version, Cape Town, Republic of South Africa. Van der Zaag, P & Nyagwambo, L. 1998. Water Allocation Criteria for the Mupfure Catchment, Final Document, December 1998, Harare. Van der Zaag, P. 2001. Water Law Lecture Notes, Department of Civil Engineering, University of Zimbabwe, Harare. Wallensteen, P & Swain, A. 1997. Comprehensive assessment of the freshwater resources of the world. International fresh water resources: Conflict or cooperation, Stockholm, Sweden. Wolf, A.T. 2000. Indigenous Approaches to Water Conflict Negotiations and Implications for International Waters, Published in: International Negotiation: A Journal of Theory and Practice, December 2000. Zimbabwe 1996. Zimbabwe Water Act 1976, (Chapter 20:22), Government Printers, Harare. Zimbabwe 1999. Zimbabwe Water Act 1998, (Chapter 20:24), Government Printers, Harare.

Impacts of water development in arid lands of Southern Africa: socio-economic issues J.P.Msangi University of Namibia, Windhoek, Namibia Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Aridity characterizes an expansive area of Southern Africa. By manipulating their environment and available resources, the inhabitants of this area have devised mechanisms that enable them eke a living. Although unfavorable climatic and environmental conditions contribute to precarious living conditions for the inhabitants of the arid lands all over the world, those in Southern Africa are particularly vulnerable due to low technology and high dependence on natural resources exploitation. Many water development projects have been undertaken causing varying environmental impacts that have reduced the performance of the economy and undermined sustainability of projects meant to off set the difficult situations prevalent in the arid lands in Southern Africa. Impacts of such undertakings must be identified prior and after project implementation and mitigative measures taken into consideration during the planning stage. The multisectoral nature of water resource projects should be taken into consideration during the planning and implementation phases. Partnerships and indigenous knowledge are vital in ensuring success.

1 INTRODUCTION Southern Africa has an expansive area that is characterized by aridity, a condition of perpetual moisture scarcity. The inhabitants of such areas have devised mechanisms which enable them eke a living through manipulations of the environment and available resources, water being key to all the activities. The activities of these areas are predominantly changing due to adjustments that must be made in response to the prevailing climatic conditions. Although recurrent droughts, climatic variability and uncertainty heavily contribute to precarious living conditions for the inhabitants of the arid lands all over the world, those in Southern Africa are particularly vulnerable due to low technology and high dependence on natural resource exploitation. Due to the prevailing climatic conditions and high sense of uncertainty, many water development projects have been undertaken or are proposed to provide more dependable water

Impacts of water development in arid lands of Southern Africa

263

supplies to meet both animal and human requirements in Southern Africa. In some areas, irrigation water has been provided to overcome inadequate and/or unreliable rainfall while in others domestic water supply schemes have been constructed including dam building and borehole construction. The main justifying reason for such undertakings is always that water in its natural state is seldom in a position to satisfy the requirements which include public water supply for domestic use and/or livestock production; regulated flow for hydro-electric power production; adequate supplies for industrial processing and irrigated agriculture. Water development projects have been recorded to cause adverse impacts to the environment the world over, Southern Africa included. Such impacts are known to reduce the performance of the economy and undermine the sustainability of projects implemented to off set the difficult situations found in the arid lands. Such impacts need to be identified and/or predicted during the project planning stage so that appropriate mitigative measures can be taken into consideration before and after the project is implemented. Environmental impact issues in Southern Africa include high population concentrations (both human and animal) attracted by the putting up of a reliable water source; soil erosion; agricultural and chemical pollution from irrigated fields as well as over exploitation of groundwater aquifers which may lead to collapse and eventual destruction of the aquifer. Others include denying down stream populations and habitats fresh water supply through damming or excessive abstractions to meet upstream water demands. Through environmental impact studies, the multisectoral nature of water resource development can be taken into consideration during the planning and implementation phases of water resource development projects. Strong partnerships and indigenous knowledge considerations are necessary to make sure all aspects of the resource are included in such studies. 2 INTRODUCTION 2.1 Environmental impact An impact is described as a strong impression or effect on something or even somebody so that a lasting impression is observable and/or measurable (Oxford advanced Learners’ Dictionary). An impact or repercussion of something can also be anticipated or expected or predicted before it takes place. On the other hand, an EIA has been described as a formal study that is used in achieving successful development of major projects through incorporation of environmental considerations in project planning and management (Westman, 1985). An EIA has been identified as both a planning and management tool for sustainable utilization of natural resources. It seeks to ensure that development options are environmentally sound and that any environmental consequences are recognized early and taken into account in project planning, design and implementation. EIA has its origins in USA where during the 1970s initial developments focused on impacts to the biophysical environment and subsequently moved on to encompass and integrate social, health, economic, improved public participation, risk and uncertainty. During mid and late 1980s emphasis included cumulative effects, the integration of project level environmental impact assessment with policy, planning, legislation, monitoring and auditing. During recent times, EIA has been described as “a planning and

Water resources of arid areas

264

assessment process that involves forecasting the environmental consequences of a proposed development process” (Mubvami, 2000). It involves “identifying, predicting, evaluating and mitigating the biophysical, social and other relevant effects of proposed projects and physical activities prior to major decisions and commitments are made” (Mutter, Topfer & Wichterich, 2002). EIA has evolved into a flexible planning tool that allows governments, donor agencies and project developers to evaluate the environmental implications of project proposals during the planning stage. Since mid 1980s, many investors and funding agencies including the World Bank and other multilateral banks require that their borrowers carry out EIA for proposed projects and programmes. During this time, many lending institutions and international environmental agencies like UNEP issued guidelines to assist such assessments and ensure that projects are designed and implemented in an environmentally sound manner. EIA is now widely accepted in both developed and developing countries as an important tool for project planning. The role, fully acknowledged at the 1992 Earth Summit, has led to several countries putting in place legislation that requires that an EIA be conducted before projects are implemented. The purpose of EIA focuses on providing a systematic, holistic and multidisciplinary view of the impacts of a proposed project or undertaking such as the impacts of constructing a dam across a river valley. These impacts include those affecting the natural environment (both living and non-living) and the people who inhabit and use the specified natural environment. In short, EIA examines the environmental and socioeconomic consequences of a proposed undertaking such as a river development project. It emphasizes prevention or minimization of adverse impacts of the project on the environment and the people. It also looks at the effects of the environmental factors on the proposed project as well as the impacts of the people’s activities on the proposed undertaking. Further, an EIA ensures that the ability of the biosphere to absorb effects of proposed activities is not diminished. It is undertaken in order to identify, analyze and assess potential environmental effects of a proposed project and where possible mitigate against negative effects. An EIA exercise can have varied consequences on a proposed undertaking. It can be used to modify and improve the design of a proposal, it can ensure efficient use of resources, it can enhance the social aspects of a proposal and it can be used to identify measures for monitoring and managing impacts and to provide justification for a proposed activity. The effectiveness of the EIA process will have a direct bearing on how many of these results will be achieved. This is more so in highly fragile and vulnerable ecosystems found over much of the arid lands of the world including those in Africa. Since an EIA is conducted before an undertaking, its ultimate goal should be to ensure that current development meets the needs of the present generations without comprising the ability of future generations to meet their own needs. Thus EIA would contribute towards the attainment of sustainable development. 3 PURPOSE OF WATER DEVELOPMENT Most always, the purpose of water development is to provide adequate supplies of water to meet various water demands. This is necessitated by the fact that in most instances,

Impacts of water development in arid lands of Southern Africa

265

water in its natural state is not in a position to satisfy the numerous demands placed upon it, which may include: – Public water supply for domestic use; – Hydro-electric power production; – Irrigated agriculture development; – Water for livestock production; – Industrial water for processing or for cooling machines; – Water for sewage treatment; – Fishing; – Navigation; – Recreation. Water development projects are known to cause both positive and adverse environmental impacts to the environment that need to be identified and/or predicted prior to project implementation. Examples range widely depending on the nature and scale of the project, its location and the type and level of technology required for its sustainability. A good example is where a project involving groundwater recharge using wastewater must take into consideration the danger of pollution from the wastewater from industrial and residential sources. The cost of treatment of the water before using it to recharge ground water must be considered included in the assessment. Similarly, the impacts of other land use management activities taking place in the project area are known to affect the water projects. Activities of unprotected catchment areas including unplanned deforestation and/or overgrazing would produce sediments which would reduce the capacity and adversely affect the life span and operation of a down stream reservoir greatly undermining the project performance and disrupting its sustainability such as happened in Kisongo dam in Arusha and Imagi dam in Dodoma, Tanzania (Msangi, 1987; Kitheka, 1993; Christiansson, 1981). On the other hand a well-managed cultivation system (including terracing and aforestation) of a catchment area will prolong the life span of the reservoir making it possible to meet multiple demands from its waters. All the occurring impacts need to be identified and/or predicted during the project planning stage so that appropriate mitigative measures can be considered before a project is implemented. Needed also is post planning project monitoring to assess the impacts resulting from the undertaking. 4 ASSESSMENT OF ENVIRONMENTAL IMPACT OF WATER DEVELOPMENT PROJECTS In the process of considering whether to carry out a major water project or not, the first main concern is the extent of environmental change that will result from the construction and particularly after the project is implemented. This change will be both to the physical environment as well as to the people and other living organisms inhabiting that environment. The changes to the physical environment will influence the socialeconomic environment just like the people’s response to the physical change will affect the physical environment. Thus both types of change must be considered when impacts are predicted and later on monitored. Most of the changes will be permanent and in some

Water resources of arid areas

266

cases cumulative. Alternatives form part of the pre-implementation phase when impacts are weighed and compared to provide careful and sufficient consideration of all possible impacts, both negative and positive ones. The assessment process for water development projects mainly addresses environmental impacts resulting from the project itself and land use management practices occurring in the river catchment area. These impacts could include high surface runoff; soil erosion and increased sediment flow and raised concentrations that are a result of induced land use changes in a river catchment where a project is proposed. Moreover, lack of knowledge or lack of apprehension of the consequences of overgrazing the catchment areas contributes to the shortening of the life span of a project (Msangi, 1996). Changed ecological conditions, such as the creation of an ideal habitat for disease carrying flies, snails and mosquitoes, have also made areas around a reservoir created behind a dam undesirable and unhealthy (Kaduma, 1977; 1972). Other impacts may include changes to groundwater levels, changes to river flow and flow peaks, flooding or drying up of a river (Christelis & Struckmeier, 2001). Others could be introduction of agricultural chemicals and fertilizer residues by surface runoff from farmlands where irrigated agriculture is part of the project or where the project will induce such undertakings. An indirect impact on an undertaking emanates from the attitudes of the population towards the proposed project as well as attitudes held on water resource use and management. In some parts of the dry lands in Southern Africa, attitudes centering on cattle numbers as wealth are most likely to override environmental conservation so that once water becomes available, the number and intensity of grazing increases without due regard to carrying capacity of the range (Msangi, 1992; 1996; Darkoh, 1989; Ellis & Swift, 1998). Such attitudes and other related social practices should form part of the assessment during and after implementation of a water development project. The actual impacts of a water resource project depend on the purpose, scale and location of the project. For a small water supply project for example, the positive impacts will include the expected socio-economic benefits such as drinking water and water for other domestic uses. Raised health and sanitation standards and the general well being of the people will also be included. An indirect impact will include elevated economic standards, as people enjoying good health will be able to work to produce more and thus generate some economic returns. On the other hand negative impacts will include over grazing of the land around watering points and beyond, along animal tracks and adjacent land etc (Darkoh, 1989; Msangi, 1991, 1996; Stone, 1991). For a large water supply project with pipelines and house connections and staggered animal watering points, the adverse impacts on the physical environment would be minimized or controlled completely. Thus alternatives ought to be considered carefully before implementation. A water development project involving the construction of large structures such as dams and canals will produce a varying range of impacts on both the physical and socioeconomic environments. Possible adverse impacts could include the displacement of people and animal populations as happened when Lake Kariba and other large dams in Southern Africa (Kaduma, 1997) and others including loss of flora and fauna; changes in groundwater conditions, triggering seismic activity due to the presence of a large body of water; deterioration of the health status of the environment through creation of ideal habitats for disease vectors; lowered water quality from rotting inundated vegetation and altered river flow characteristics (Kaduma, 1997). Conversely, the advantages of a large

Impacts of water development in arid lands of Southern Africa

267

scale water development would include creation of new habitats such as wetlands; the production of much needed electricity for irrigation water to support agricultural production; provision of hydro electric power for homes and for industrial establishments; job creation from undertakings utilizing the water and generated power; support improved economic conditions and the well being of the people; regulated river flow and improved utilization of a river including the establishment of a fishery. The list of indirect impacts is long and varied. The creation of sub-surface dams that are more environmentally feasible than surface dams are faced with various limitations including water recovery. High investments required during construction and maintenance and operation of pumps (be it petrol/diesel, solar energy or windmills) are often not economically justified given the low land productivity inherent in most parts of the dry lands of Southern Africa. Hand pumps are only feasible if recovery is from shallow wells (Msangi, 1996). 5 DRY LANDS IN SOUTHERN AFRICA Southern Africa has an expansive area that is characterized by aridity, aridity being a condition of perpetual moisture scarcity. The inhabitants of such areas have devised mechanisms which enable them eke a living through manipulations of the environment and available resources. The activities of these areas are predominantly changing due to adjustments that must be made in response to the prevailing climatic conditions. An exceptionally wet year may see the cultivation and harvesting of quick maturing crops that dry years will not. More often than not the survival techniques include livestock rearing, mostly keeping of small stock such as goats and sheep. Cattle are kept for milk, export beef and as a source of wealth in the areas that enjoy relatively humid conditions as opposed to those that are very dry (Msangi, 1996). Due to excessively high temperatures, the little moisture that may be received in the form of rain gets evaporated very quickly soon after a downpour (Msangi, 1996). Climatic variability and uncertainty has led to precarious living conditions for the inhabitants of the arid lands all over the world, those in Southern Africa high in the list. Due to the prevailing climatic conditions and high sense of uncertainty, many water development projects have been undertaken or are proposed to provide more dependable water supplies for both animal and human requirements as well as industrial water needs. In others, irrigation water has been provided to overcome inadequate and/or unreliable rainfall (Chenje & Johnson, 1996). The main justifying reason for such undertakings is always that water in its natural state is seldom in a position to satisfy the requirements which include public water supply for domestic use; regulated flow for hydro-electric power production; adequate supplies for irrigated agriculture development and for livestock production. Water development projects have been recorded to cause negative impacts to the receiving environment (Kaduma, 1977). While these have been investigated and documented worldwide, in Southern Africa, the need still exists to identify and/or predict them before the proposed projects are implemented and follow up monitoring after they are in operation (Chenje & Johnson, 1996; Wood, Stedman & Mang, 2000). Such projects need to be monitored as they may be affected by environmental factors caused by other land use management activities taking place in the

Water resources of arid areas

268

project areas. Such activities are known to reduce the performance of projects and do undermine the sustainability of projects implemented to off set the difficult situations found in the arid lands (Biswas, 1978; Msangi, 1996). All these impacts need to be identified and/or predicted during the project planning stage so that appropriate mitigative measures can be considered before the project is implemented. The inhabitants of the dry lands in Southern Africa have a rich heritage of managing and living with their environment including water. They have been irrigating their lands for centuries. The communities inhabiting the dry lands have lived and adapted to the environmental conditions arising from many years of experience and folklore handed down generations. This harmonious existence with nature was interrupted and interfered with during the last a hundred years or so through the introduction of western cultures and new ways of viewing the environment. The introduction of improved health and nutritional facilities as well as monetary economy together with the institutional requirements that go with it, has disrupted and partially changed the lifestyles of these communities (Msangi, 1996, 1992; Stone, 1991; Ellis & Swift, 1988). The population of both people and animals has increased rapidly as food aid and western medicine have increased survival chances and increased fertility rates. Before this interruption, land, water and vegetation successfully supported the life styles and economic activities of the dry lands inhabitants. The forces of nature had adequately checked imbalances between man and nature so that simple social and economic patterns had developed and had been harmoniously maintained. The installation of schools, hospitals, central governments and all their branches imposed new requirements on the communities and therefore the environmental resources. Water being the central and most scarce resource in these lands has been subjected to various manipulations and new development approaches geared towards meeting both the communities’ traditional and new institutional demands. Due to increasing populations of both people and animals, water demand far exceeds supply, thus the need to practice wise use, management and conservation of water resources in the dry lands of Southern Africa. This requires that social attitudes be reoriented so that communities appreciate the implications of limited supply as opposed to the ever increasing demands on the scarce water resources, limited groundwater recharge rates and the need to conserve the resources such as controlling pollution and recycling, all new concepts to most of the dry land communities in Southern Africa. Many cases of efforts to conserve, develop and manage the water resources in the dry lands have been made and are documented in numerous plans and consultant reports (Msangi, 1992). Few successful cases have been recorded and many failures have been experienced. The reasons for failure are mainly due to the inappropriateness of the technology adopted to the existing environmental conditions or, most often, to the wants and wishes of the local communities. Many times the wrong sector of the community has been targeted for training. Women and their children who are the ones mainly responsible for collecting and managing water for domestic use and sometimes tending small stock, tend to be side lined for the men who are users rather than managers of the resource. Women should be at the center of any training aimed at improving existing management technologies or introducing new ones. Sustainability rests on clear understanding of the people’s social organizations and gender roles and economic patterns in a given community.

Impacts of water development in arid lands of Southern Africa

269

Economic and other development activities intended to be introduced into the dry lands should be focused more on the needs and wants of the people bearing in mind environmental constraints. For example instead of introducing irrigated agriculture, dry land farming based on indigenous crop varieties should be employed instead of sprinkler irrigation to grow exotic crops with high water demand. Flood irrigation and other high water requiring methods of crop growing have rendered useless large tracts of land through salinization. The high evaporation rates inherent in the dry lands of Southern Africa do not favor these methods. 6 LAND USE MANAGEMENT SYSTEMS AND THEIR IMPACTS ON WATER DEVELOPMENT PROJECTS Water development projects are affected by the land use in many different ways. Unregulated land use system such as indiscriminate clearing of tree cover from a catchment area can lead to reduced water yield and cut short the lifespan of a project. Cultivation and/or heavy grazing of such a catchment area can lead to soil erosion and subsequent sedimentation and siltation of a reservoir, intakes and irrigation channels. Such a system can also lead to increased surface runoff, flooding in the lower reaches and lowered groundwater in the upper reaches thus jeopardizing a water development project. Other impacts include reduced water quality from suspended sediment and agricultural chemicals and residues from farmlands. This will lead to increased costs for water purification or adverse health conditions for those depending on the water source either for domestic use or industrial processing. Habitats for fish and other aquatic animals will be damaged and the economic standing of those dependent on them will be adversely affected. Furthermore, maintenance costs for structures will increase dramatically if sediment has to be cleared regularly, unless ofcourse this was foreseen and budgeted for right from the beginning. Closely connected to decreased infiltration and reduced water yield due to compaction is the loss of water sources such as wells and springs. Reduced infiltration leads to reduction in levels of ground water table that may cause ground subsidence (ground surface collapse and curving in) due to over pumping; or if close to coastal areas lead to intrusion of coastal salt water and soil salinization that may reduce crop production through increased accumulation of harmful salts in soil particularly where irrigated agriculture depend on wells or boreholes (Christelis & Struckmeier, 2001). Other types of land use such as urban land use may cause water pollution due to inadequate water and waste management from dwellings and industries. Pathogens as well as organic and chemical pollution can lower the water quality necessitating expensive water treatment to meet set water quality standards. Alternatively, high concentrations of discharged organic compounds may create excessive demand on oxygen resources of a body of water during the conversion process to the extent that the oxygen concentration in the water is reduced and eventually depleted resulting in death of living organisms including fish. High concentration of organic matter may also raise the fertility of the water body to the extent that eutrophication occurs leading to life decimation in the water body (Wood, Stedman & Mang, 2000).

Water resources of arid areas

270

7 CONCLUSIONS AND RECOMMENDATIONS It becomes apparent that water resource development projects produce serious and definable impacts on a community both socially and economically while other activities in the vicinity of the water projects affect and influence the performance of the projects. Feasibility studies for all impacts must be carried out prior to water projects implementation. Successful water development and resource conservation should always strive to incorporate environmental considerations during project planning and project implementation stages. Similarly, integrated catchment management should encompass the various resource components and associated management practices to achieve stable systems. Environmental legislation should make EIA mandatory in all water development projects in order to ensure sustainability and high quality water supply for industrial, agricultural and domestic usage. People centered planning should be adopted where social, economic and environmental consequences of an undertaking are given deserving emphasis. Therefore social-economic as well as environmental impacts should be considered alongside the often-emphasized physical and technical impacts. 8 APPENDIX 1 (AFTER WESTMAN, W.E. 1985) Questions Useful in Planning the Pre-Impact Phases of an Impact Assessment: 8.1 Phase I: Defining study goals What information is needed and how precise must it be for: – The proponent to minimize environmental impact. – The government agency to reach a decision on approving the project. – Concerned groups to know how they will be affected What resources are needed for the study? What resources are available? – Needed expertise; available expertise – Needed time for baseline and experimental studies – Remaining time before the project is supposed to begin – Required funds to conduct the proposed study; available funds. 8.2 Phase II: Identifying potential impact What are the boundaries of potential impacts? – Area affected – Organisms or ecological functions affected – Duration of the project

Impacts of water development in arid lands of Southern Africa

271

– Interval before effects occur – Duration of effects with and without mitigation. What is the range of potential impacts? – Major direct actions – Major ecological components (air, land, biota, structure) affected – Major ecological processes affected – Secondary or higher-order interactions – Indirect effects triggered at a future time or different place – Other actions (past, present, reasonably foreseeable future) that may add to the present action, causing cumulative effects. Which potential impacts are most significant? Which effects will: – Violate existing laws, plans or policies. – Cause major disruption to ecosystem processes, processes, affecting species significantly. – Cause major adverse effects on species numbers. – Cause health risks, economic losses or significant social disruption to people. 8.3 Phase III: Measuring baseline conditions and predicting significant impacts Baseline conditions: What are the significant features of the ecosystem presently? – What is the current pattern of fluctuation in popular sizes for important species (measured over sufficient time to characterize the range of variations)? – Which species are playing a dominant or critical role in maintaining ecosystem processes? What is their abundance, distribution and function of behaviour? – What is the condition (Quality, quantity, dynamics) of physical resources of the ecosystem? – What are the major pathways of interaction between ecological components? – What sources of stress from natural or human-induced sources already exist (fire, air pollution, grazing etc)? With what intensity and periodicity do these stresses occur? Predictions: What will be the major effects of the proposed action? What is known from each of the following? – Case studies: Extrapolation of effects from similar instances of disruption to the same or similar ecosystems elsewhere. – Modelling: Predictions from conceptual or quantitative models of ecosystem interaction. – Bioassay and Microcosm Studies: The effects of simulated disturbances on ecosystem components under controlled conditions. – Field Perturbation Studies: Response of a portion of the proposed project area to experimental disturbance.

Water resources of arid areas

272

– Theoretical Considerations: Predictions of effect from the current ecological theory. Estimation of likelihood: – What is the probability of occurrence of the predicted events? – How precisely can the magnitude and likelihood of impacts be estimated? Summarizing and analyzing findings: – How can findings be summarized in table, graphs or indexes so that key findings emerge? – What is the ecological interpretation of the findings? 8.4 Phase IV: Evaluating significance of findings How are the effects distributed among the affected groups? – What is the nature and magnitude of impact on each affected group? – What weight shall be given to the concerns of each group? – What weight does each group give to the significance of predicted effects? How well are goals achieved by the proposal? – Proponent’s goals? – Goals of affected groups? What is the overall social significance of the predicted ecological effects? – How can effects be expressed in terms that allow meaningful comparison with other social goods, services and values? – If monetary values are placed on normally unpriced goods and services, what features are inadequately evaluated by this procedure? 8.5 Phase V: Considering alternatives to the proposed action – What would be the effect of not proceeding with the project? – What would be the effect of achieving ultimate project goals by an entirely different means? (e.g. maintaining electrical service to a growing population by conserving energy rather than building a new power plant) – What alternative designs could achieve project goals? What steps could be taken to mitigate adverse environmental effects of the proposed project? – Could parts of the proposal be reduced or eliminated? – Could expected damage be repaired or rehabilitated? – Could ongoing management procedures be instituted to reduce damage?

Impacts of water development in arid lands of Southern Africa

273

– Could affected components be replaced or owners compensated? – Could project design be modified to reduce effects? – Could effects be monitored, and provision made for future mitigation of project effects when the exact nature and extent of effects are better known?

REFERENCES Chenje, M. & Johnson P. (eds.) 1996. Water in Southern Africa. SADC, IUCN SARDC, 238 Biswas, A.K. 1978. Environmental Implications of Water Development for Developing Countries, Water Supply and Management Journal, 2, 283–297 Christelis, G. & Struckmeier, W. (eds) 2001. Groundwater in Namibia: An explanation to the Hydrological Map, 128 Christiansson, C. 1981. Soil Erosion and Sedimentation in Semi-arid Tanzania: Studies on Environmental Change and Ecological Imbalance. Uppsala: Scandinavian Institute of African Studies and Department of Physical Geography, University of Stockholm. Darkoh, M.B.K. 1989. Combating Desertification in the Southern African Region: An Updated Regional Assessment. Nairobi: UNEP Ellis J.E. & Swift, D.M. 1988. Stability of African Pastoral Ecosystems: Alternative Paradigms and Implications for Development. Journal of Range Management, 41(6), 450–211. Kaduma, J.D. 1972. Some Development and Economic Aspects of the Mindu Dam Project, Morogoro: A Background Analysis for Decision Making, M.A thesis, University of Dar es Salaam, 140 Kaduma, J.D. 1977. Man-made Lakes: Their Social, Economic and Ecological Impacts—The Case in Tanzania, PhD thesis, University of Dar es Salaam, 400 Kitheka, J.U. 1993. Soil Erosion and Its Impacts on Surface Water Reservoirs: A Case study of Nguu Tatu Catchment, NE Mombasa District, Kenya, Proc. 4th Land and Water Management Workshop, Nairobi, Kenya, 309–329 Mubvami, T. 2000. Environmental Impact Assessment as a Policy Tool for Environmental Management. IUCN-ROSA A Handbook on Approaches to the Environmental Policy Analysis in Southern Africa. IUCN—The World Conservation Union. Mutter, T., Topfer, J. & Wichterich, C. 2002. A Comprehensive Study of the Heinrich Boell Foundation’s projects abroad in Political Ecology and Sustainability. 1st Ed. Heinrich Boell Foundation. Msangi, J.P. 1987. Conservation of Water Resources in the Semi-arid Areas of Tanzania. Journal of Eastern Africa Research and Development. Vol 17, 63–73. Msangi, J.P. 1991. Sustainability in Exploitation, Development and Management of Hydrological Resources of Turkana District. Journal of Eastern African Research and Development, 21, 21– 39. Msangi, J.P. 1992. Social-Cultural and Demographic Factors in Desertification Control in Kenya’s Arid and Semi-arid Lands. Proc. of the Workshop on Desertification Monitoring, Assessment and Control. Nairobi: National Environment Secretariat, 21–32. Msangi, J.P. 1996. Social-Cultural Factors Affecting Non-Adoption of New Water Harvesting Technology Among the Dryland Communities in East Africa. In Yue-man Yeung (ed) 1996 Global Change and the Commonwealth. Hong Kong Institute of Asia-Pacific Studies, Chinese University, Hong Kong, 233–253. Stone, J.C. (ed) 1991. Pastoral Economies in Africa and Longterm Responses to Drought. Aberdeen: Aberdeen University African Studies Group. Wood, A. Stedman-Edwards P. & Mang, J. 2000. The Root Causes of Biodiversity Loss. Earthscan, 399.

Water resources of arid areas

274

Westman, W.E. 1985. Ecology, Impact Assessment, and Environmental Planning. John Wiley & Sons.

Institutional challenges for small towns’ water supply delivery in Ghana Kwabena Biritwum Nyarko Kwame Nkrumah University of Science and Technology, Kumasi, Ghana Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: The small towns’ water supply sector emerged in Ghana after 1994 when water supply delivery was separated into urban water supply (served by a public utility) and Community Water Supply (CWS) decentralised under community ownership and management. A small towns’ water system, which falls under CWS, is defined as a piped system serving a community with inhabitants between 2,000 and 50,000 that is willing to own and manage its water supply system. A national programme for community water delivery has been in place since 1994. Ensuring the sustainability of the small towns’ systems are fundamental concerns, which makes the study of institutional issues that affect the sector timely. The paper describes how the institutional arrangement to support the delivery of small towns’ water services in Ghana has evolved, and also discusses the experiences, lessons and the challenges. The paper also makes recommendations to improve service delivery.

1 INTRODUCTION About 32% of the Ghanaian population do not have access to safe water (WDI, 2002) making access to safe drinking water a challenge. The situation is even worse in the rural an small communities where the majority of the population lives. In 1994, water supply delivery in Ghana has been separated into urban water supply (served by a public utility, GWCL) and Community Water Supply (CWS) under community ownership and management (Nyarko, 2000 & CWSA, 2003a) to improve the supply of water to the people of Ghana in a sustainable manner. Community water supply consists of rural and small towns’ water supply. The small towns’ water supply is a piped system, serving communities with inhabitants between 2,000 and 50,000 who are willing to own and manage the water system. As at the end of 2001 there were 254 small towns’ water supply systems under community ownership and management (CWSA, 2003a). These systems are decentralised and do not enjoy any cross subsidies and other benefits of economies of scale as the urban water supply.

Water resources of arid areas

276

Consequently, interest in the delivery of water services in small towns has grown rapidly in recent years due to the peculiar characteristics of the small towns and the number of inhabitants it serves. To ensure the sustainability of the small towns’ systems are fundamental concerns, which makes the study of institutional issues that affect the sector timely. This paper examines the institutional framework of the small towns’ water supply component of community water supply with the aim of enhancing the sustainability of the service delivery.

2 BACKGROUND TO THE SMALL TOWNS’ WATER SECTOR 2.1 Historical development of community water supply sub-sector The Ghana Water and Sewerage Corporation (GWSC), the predecessor of Ghana Water Company Limited (GWCL), was established in 1965 by Act 310 for the provision, distribution and conservation of both the urban and rural water supply in Ghana, for public, domestic and industrial purposes. In 1986, the first attempt to enhance service delivery in the rural areas (defined as communities with less than 5000 inhabitants) led to the establishment of the rural water department within the GWSC (Asamoah, 1998). At that time, the approach of providing water services to customers was a supply driven one. With more promising revenue from the urban areas (with higher income levels) as well as technically challenging “engineering” of providing urban water services, GWSC focused more on the urban areas. In 1991, the sector ministry for water, Ministry of Works and Housing (MWH) organised a workshop to discuss the provision and sustainability of rural water supply and sanitation. The outcome of the workshop was the National Community Water and Sanitation Programme (NCWSP), which aims at accomplishing the following objectives: ● To provide basic water and sanitation services to communities that will contribute towards the capital cost and pay the normal operations, maintenance and repair cost of their facilities ● To ensure sustainability of these facilities through community ownership and management, community decision making in their design, active involvement of women at all stages in the project, private sector provision of goods and services, and public sector promotion and support ● To maximise health benefits by integrating water sanitation and hygiene education In line with the NCWSP, the urban and rural water supply systems were separated in 1994. The rural water division of GWSC was transformed into a semi-autonomous department known as the Community Water and Sanitation Department (CWSD), with the responsibility of implementing the NCWSP. CWSD was further transformed by Act 564 of 1998 into an agency, the Community Water and Sanitation Agency (CWSA) (GOG, 1998) with the responsibility of facilitating community water and sanitation services under community ownership and management.

Institutional challenges for small towns’ water supply delivery in Ghana

277

2.2 NCWSP and GOG decentralisation reforms The implementation strategy of the National Community Water and Sanitation Programme (NCWSP) is consistent with the Ghana decentralisation policy, which transfers authority, responsibility and capacity from the Central Government, Ministries and Departments to the District Assemblies. The decentralisation policy is backed by the Local Government Act, 1993, Act 462, which aims at devolving central administrative authority and divesting implementation responsibility to district levels (GOG, 1993). It re-assigns functions making Central Government Ministries/Departments undertake policy planning, monitoring, evaluation and promotion; and makes regions, (through the Regional Co-ordinating Councils and their respective Regional Planning Co-ordinating Units), play the role of co-ordination, whilst, the District Assemblies become responsible for implementing development programmes (CWSA, 2000). 2.3 The small towns water supply sub-sector A small town is defined in the CWSA Act as “a community that is not rural but is a small urban community that has decided to manage its own water and sanitation systems”(GOG, 1998). A small town water system is also defined as a piped system serving communities of between 2,000 and 50,000 inhabitants who are prepared to manage their water supply systems in an efficient and sustainable manner (CWSA, 2003b). The Act further defines a rural community to be those with less than 5000 inhabitants. The MWH’s Comprehensive Development Framework 1999 for the water sector also defined a small town based on a population range of between 5–15,000 (MWH, 1999). The implementation of each Small Towns’ Project follows the following cycle (CWSA 2003b): ● Project Promotion—for the prospective Community to be familiar with the project cycle and procurement procedures. ● Community Selection and Approval—by the District Assembly in collaboration with the CWSA. ● Community Mobilisation—An extension team is engaged to provide relevant community mobilisation and extension services in each beneficiary community. ● Hygiene Education and Sanitation ● Participatory Planning—to ensure that beneficiary communities are adequately informed and are responsible for decisions made on the system ● Design-Water supply systems shall be adequately designed to provide reliable and good quality water in sufficient quantity over the design period. ● Construction, Operation and Maintenance of the facility ● Post Project—The CWSA shall provide relevant post project support (up to one year) to beneficiary communities to promote achievement of system sustainability. A typical small towns’ water system consists of the following (Jonah. E, 2003): ● a source (usually a mechanised borehole),

Water resources of arid areas

278

● pump house (a submersible pump powered by a 3-phase voltage transformer), ● source of power (AC power from the national grid, local diesel Power generator or Solar panels (only few cases in the northern region) ● Pipelines (transmission and distribution pipes made of uPVC and HDPE), ● An elevated reservoir, standpipes and appurtenances.

3 POLICY, LEGAL AND INSTITUTIONAL FRAMEWORK 3.1 Institutional arrangements The institutional framework is shown in Fig 1. CWSA is under the oversight of the Ministry of Works and Housing (MWH), the sector ministry responsible for water. The District Assemblies (DA) is the highest political and administrative authority in the district, with responsibility for development and management of basic infrastructure, municipal works and services (GOG,

Figure 1. Institutional arrangement for small towns’ water supply delivery. 1993). The Regional Co-ordinating Councils (RCC) and their respective Regional Planning Co-ordinating Units, play the role of co-ordination, whilst the District Assemblies are responsible for implementing development programmes.

Institutional challenges for small towns’ water supply delivery in Ghana

279

The Environmental Protection Agency (EPA), under the Ministry of Environment Science and Technology (MEST) is charged with environmental regulation. The Water Resources Commission (WRC) has the obligation to allocate and grant water rights. The Ministry of Finance (MF) is responsible for negotiation and approval of credit facilities (loans) in the water supply and sanitation industry. The CWSA regional officers constitute the Regional Water and Sanitation Team (RWST), which is composed of a Hydrogeologist, Water and Sanitation Engineer, an Extension Services person and a Financial Specialist. The District Water and Sanitation Team (DWST) is the focal point in the District Assembly (DA) for water service delivery. It’s a three-member team with members seconded from the Public Works Department, the Department of Community Development and the Department of Environmental Health. Their role includes the identification of interested communities and providing support to the Water and Sanitation Development Board (WSDB). The External Support Agencies (ESAs) provide technical and funding support to the subsector. The WSDB is responsible for the management of the small towns’ water supply system. It is composed of elected community (small towns) members. The WSDB is also responsible for appointing the operational staff, promoting and disseminating information within the community, ensuring that all community members participate in decision making, setting tariff and ensuring proper financial management. 3.2 Policy framework Based on the January 2001 draft policy for small towns’ water supply (CWSA, 2001) the key policy statements were: ● Beneficiary communities would pay a part of the capital cost and take up all operations and maintenance costs. The community contribution depends on the levels of service selected by the community. It is 5% of the capital cost for Basic Water Supply Services, which is the supply of 20l/c/d (standpipes) for 80% of population and 60l/c/d (house connection) for 20% population. For higher levels of service the community contribution shall be 50% of the capital cost. ● District Assembly shall contribute 5% of the capital cost. ● Water produced shall meet WHO International Drinking Water Quality guidelines. ● Delivery of water should be in a cost effective manner (not exceeding the cedi equivalent of $1.0/m3) CWSA has reviewed the 2001 draft policy in attempt to improve service delivery. The main changes and additions based on the 2003 draft policy for small towns’ water supply (CWSA, 2003b) are: ● Community contribution for capital expenditure would be 2.5% for Basic Water Supply Services and 50% of the Additional Cost for Levels of Service Higher than Basic Water Supply Services; ● The membership of WSDBs shall exclude Traditional Authorities and DAs. Where necessary, they may participate in WSDB meetings as observers. ● Tariffs shall be set by the WSDBs in accordance with CWSA approved tariff setting guidelines. DAs shall review and approve all tariffs. Any reduction in expected tariff

Water resources of arid areas

280

revenue as a result of action by the DA, e.g., reduced tariff, etc., shall require that the DA pay the difference in revenue into the WSDB account. The CWSA in collaboration with the Regional Co-ordinating Council (RCC) shall ensure compliance. ● The implementation of small towns’ water supply and sanitation projects shall be in accordance with the regulations of the Environmental Protection Agency (EPA) and Water Resource Commission (WRC). The CWSA Guidelines for operations and maintenance (CWSA, 2003c) also stipulates that: ● communities through tariffs shall undertake all major repairs and replacements. But for total rehabilitation, cost sharing arrangement and procurement procedures shall be the same as for new systems. ● Water produced shall meet Ghana Standards Board Drinking Water Quality standards. 3.3 Legal authority of the WSDB The concept of community management is achieved by having community representatives, the Water and Sanitation Development Board (WSDB) in charge of the water supply management in the community. As part of the decentralization policy of the Government and in accordance with section 15 of Act 462 the District Assembly has the power to delegate its functions other than its legislative functions to an individual or group (GOG, 1993). The Water and Sanitation Development Board (WSDB) takes its legal authority from the District Assembly through a byelaw. CWSA has developed generic bye-laws for WSDBs to adapt for their local circumstances.

4 INSTITUTIONAL ANALYSIS OF THE SMALL TOWNS’ SUBSECTOR The research approach utilised both quantitative and qualitative methods to gain insight into the institutional issues of the sector. A literature review of small towns’ water sector project documents was first conducted to get a thorough understanding of the sector. Literature on institutional issues such as what the institutional framework should offer was also reviewed. Based on the literature review, the research instruments (interview guides and questionnaires) were developed for the various stakeholders, in the sector to identify the main institutional issues and challenges. Specific institutional analysis tools used were a combination of the Activity Responsibility Matrix (ARM) and the Strength Weakness Opportunity Treats (SWOT) as well as National Macro-environment Analysis. The data was collected from field visits conducted in 20 small towns’ water supply systems, five district assemblies and three regional CWSA offices. In addition discussions with 20 WSDB Chairmen and 22 Technical Managers who attended a short course at Kwame Nkrumah University of Science and Technology, Kumasi, Ghana in August 2003 were used to validate the study.

Institutional challenges for small towns’ water supply delivery in Ghana

281

4.1 SWOT Analysis of the STWSS sub-sector See Table 1. 4.2 Activity responsibility matrix See Figure 2.

5 DISCUSSION OF KEY ISSUES 5.1 Institutional arrangement and framework The District Assemblies (DAs) are the managers of the NCWSP at the district level. The DAs full time field team for water and sanitation is the DWST. As mentioned already the DWSTs have been seconded from different departments and are ultimately answerable to their respective regional directors, who can effect transferred without consulting the district assemblies. There are cases of such transfers to different district on a totally different assignment. It was revealed that the time taken for the vacancy to be filled normally spans 3–12 months, after which re-training would have to be organised. It was also realised that the monitoring and supervision of the WSDB activities by the DAs is weak. For example, there are no proper monitoring mechanisms to ensue that records (operational,

Table 1. SWOT analysis. Strengths • •

•

•

•

Weakness

• The DWST is not well anchored in High sense of ownership among the DA structure since the communities Almost universal acceptance of the need individual members have been seconded to the DA making DWST to pay for water services even at rates staff subjected to indiscriminate higher than what prevails in the urban transfers by their mother water sector organisations. Users’ perception of the water service delivery is high, since in most cases the • The technical/administrative water situation was poor before the capacity at the local level (DAs, boards took over. WSDB) is weak. • Data collection and record-keeping In a number of situations, the responsiveness of WSDBs/Communities have been poor, regular water quality monitoring has not yet in servicing breakdowns is high, started; compared to the previous situation of centrally-managed systems under • Even though the various boards GWCL. have been taught the guidelines for Accountability to the users in the tariff-setting, in a majority of the community via public fora. systems visited, the tariff was not based on a rational analysis of the cost components

Water resources of arid areas

Opportunities • Government of Ghana support

•

Good will and support from External Support Agencies

•

Government decentralisation policies

282

• Regulation and monitoring from DAs is poor and in most cases no reports are sent by the WSDBs. Even where these are sent no analysis or follow-up is done by the DAs; • Lack of appropriate institutions at the local level to manage the water resource • Revenues accruing from water sales are sometimes mis-appropriated for other purposes; • Membership of some boards is dwindling due to a lack of interest, presumably because of the poor remuneration. This leaves a few who take decisions that may not always be in the interest of the whole. Threats • Political interferences in community management (WSDB affairs). • Relatively high levels of tariff in small towns’ compared with the urban water supply • ESAs/Donor fatigue • Inadequate attention to ensure water resource management • Inadequate attention to ensure financial sustainability

technical and financial) are well kept and that reports are submitted to DAs and CWSA. In addition, when reports are prepared and sent to the DAs it hardly get comments from the DAs. Majority of small towns does not perform routine water quality tests as stipulated by CWSA. This has been attributed to low capacity at the DA level especially the DWST to perform their function and inadequate resources at the DWSTs disposal for their duties. The DAs is also expected to play the role of the Water Resources Commission (WRC) at the district level in the areas of water abstraction rights and permitting. This aspect is not yet operational at the DAs level. The Ministry of Works and Housing (MWH) is responsible for policy making in the water sector. The provision of infrastructure is the responsibility of the DAs, which is under the Ministry of Local Government and Rural Development (MLGRD). CWSA as a facilitating agency cannot force the DAs or the community (WSDB) to execute its water related activities (eg. ensuring the submission of periodic reports, water quality monitoring and using appropriate water tariffs). This is attributed to the following: lack of effective accountability mechanisms between MLGRD/DA and CWSA; the location of

Institutional challenges for small towns’ water supply delivery in Ghana

283

CWSA and DAs in different ministries and the fact that the DAs is an authority on its own.

Figure 2. Scored activity responsibility matrix. To improve on the situation the following option are considered important. First, establish a works department under the DAs structure with responsibility for infrastructure (including) delivery so as to deal with CWSA. This has been accepted in principle but its implementation is yet to start. Secondly, establish clear and explicit accountability mechanism with benchmarks between the RWST and the DAs within the region would help. The introduction of yard stick competition with incentives for the DAs in a particular region or even nationwide would be useful to provide a check on the DAs to enhance performance. It is important for the CWSA regional team to have good collaboration with the RCC the appropriate institution to supervise the DAs activities to enhance CWSA monitoring role of the NCWSP. The roles and responsibilities of the WSDB demand certain skills, such as technical, financial, managerial etc. The selection criteria initially specified gender and interest groups representation, without mention of the skills required. The new draft policy dated July 2003 adds that the membership of WSDBs shall exclude Traditional Authorities and DAs, but where necessary, they may participate in WSDB meetings as observers (CWSA, 2003b). This is laudable since there were interference from traditional authorities. For example in one small town the WSDB chairman was the chief and his nephew was also the treasurer. In a number of communities political agitation and social tensions have affected the membership, tenure and therefore the effectiveness of WSDBs. Some of the examples are:

Water resources of arid areas

284

● In a small town (Bimbilla), after a meeting with all stakeholders to increase the tariffs “a youth movement” managed to convince the District Chief executive (DCE) to reverse the decision. ● In a small town (Bekwai), the District Security Council dissolved the WSDB in response to a proposed demonstration threat by some community members. However, the WSDB were re-instated after about six months. ● In a small town (Juaso) the chief requested for money for farming and was granted by the WSDB. The DA got to know of it and demanded the money back and the WSDB was dissolved. As at now (over 6 months) the new WSDB do not have access to their Bank accounts because signatories have not changed. At a recent training course organised at KNUST for WSDB treasures, majority of the participants confirmed having illiterates on the WSDB and indicated that it affects performance. They attribute that to the community sensitisation during project preparation, which made them understand that the WSDB is the community representative and that any one elected by the community could do it. As a result some of the WSDBs members do not understand the issues, and this reflects in the system’s performance. In cases where members have the required skills, performance has been exceptional (Arthur, 2002). The field visits also revealed that, WSDB perceives themselves as owners of the water supply, which seems to explain why periodic operational reports are not sent regularly to the DAs and CWSA. The DAs also do not have incentive mechanisms in place to enhance the WSDB performance. Most or some of the WSDB do not have approved byelaws and hence do not have legal recognition. 5.2 Cost recovery Initially CWSA policy for the small towns made it clear that water tariffs would have to cover all the operations and maintenance cost, but was not explicit on the recovery of capital expenditure (CWSA, 2001). The new policy indicates that, after the initial community contribution, water tariffs should cover operations and maintenance, major repairs, replacements, and extension to new areas (CWSA, 2003b). However, the operations and maintenance guidelines also states that for total rehabilitation of existing system components cost sharing arrangement and procurement procedures for new projects would be followed (CWSA, 2003c). The lack of definitions to differentiate between major repairs and the total rehabilitation makes the policy unclear. Furthermore, a blanket policy, which does not consider special cases such as a small town with a rather small population but requiring a complex technology, could worsen the plight of some communities with respect to the achievement of public health benefit. 5.3 Interface between “CWSA” and “GWCL/urban” small towns’ water supply In addition to small towns’ water supply being facilitated by CWSA and there are small communities in the urban areas that receive service from the urban public utility (Ghana Water Company Limited, GWCL) as part of urban water supply. The potential transfer from “CWSA small town” to GWCL small town” and vice versa, raises the following

Institutional challenges for small towns’ water supply delivery in Ghana

285

questions (Sarpong Manu, 2001): Whose prerogative is it to make this decision—MWH, DAs or the WSDB?, What will be the criteria for any such transfer? At the moment a small town (Ejisu) with a population of about 15,000 but under GWCL supply service areas wants to come under CWSA and benefit from the small towns facility. They claim that for the past 15 years water supply from the urban water supply utility has been basically non existent.

6 CONCLUSION AND RECOMMENDATIONS At the national level the institutional linkages and the accountability mechanisms between MLGRD/DAs, CWSA, MWH and the Water Resources Commission (WRC) are weak. This results in ineffective monitoring and management of the small towns’ water system as well as the water resources. At the district level the DWSTs members seconded from the other department does not make the DWSTs permanent in the DAs structure affecting delivery of water services. The policy on cost recovery is not clear especially on the recovery of capital expenditure such as rehabilitation and major replacement. Based on the conclusions, the following recommendations are made: ● At the national level there is the need for a closer collaboration between CWSA, MWH, WRCand MLGRD through the inter-ministerial coordination and at the regional level DAs, RCC and CWSA (regional office). In addition there is the need to include accountability mechanisms in the Memorandum of Understanding (MOU) between the DAs and CWSA. ● The Government of Ghana should speed up the process of establishing the Works Department within the DA to strengthen the DWST position in the DA. ● CWSA should clarify the policy on cost recovery. Where there is room for subsidies, the source of the subsidy and criteria for eligibility should be made explicit.

REFERENCES Arthur, E. 2002. Manpower survey at the district assembly and community levels for small towns’ water supply.. BSc Thesis, Kwame Nkrumah Univ. of Science and Technology, Kumasi, Ghana. Asamoah, K. 1998. Ghana: The Community Water and Sanitation Project. Paper presented at the Community Water Supply and Sanitation Conference at the World Bank, Washington, DC. CWSA, 2000. The Project Operational Manual (POM) of CWSP-2, CWSA: 10–11. CWSA, 2001. Small Towns Water supply and Sanitation Policy, CWSA. CWSA, 2003a. Investment Opportunities in the Community Water and Sanitation sub-sector. A presentation to an Americo-German Investors in Ghana, CWSA, pp 2, 8. CWSA, 2003b. Small Towns Water and Sanitation Policy. Community Water and Sanitation Agency, Ministry of Works and Housing, Government of Ghana. CWSA, 2003c. Small Towns Water and Sanitation Policy. Operation and Maintenance Guidelines. Community Water and Sanitation Agency, Ministry of Works and Housing, Government of Ghana. GOG, 1993. Act 462, Local Government Act, Ministry of Local Government and Rural development. Government Printer, Assembly press, Accra.

Water resources of arid areas

286

GOG, 1998. Act 564, Community Water and Sanitation Agency Act, 1998. Government Printer, Assembly press, Accra. Jonah, E. 2003. Performance Assessment of Small Towns Water Supply System: The role of management models an institutional structure, MSc Thesis, Kwame Nkrumah University of Science and Technology, Kumasi, Ghana. MWH, 1999. WATER. Comprehensive Development Framework. Ministry of Works and Housing, Ghana. Nyarko, K.B. 2000. Ghana Water and sanitation sector: Drivers for water performance. PhD Proposal, IHE Delft, The Netherlands. Sarpong Manu, K. 2001. PPIAF/CWSA PSP in Small Towns Water Study. CWSA, Sept 2001. WDI, 2002. World Development Indicators, http://www.worldbank.org/data/dataquery.html

Socio-economic performance of Sepeteri irrigation project in Nigeria O.O.Olubode-Awosola & E.O.Idowu Department of Agriculture Economics, Obafemi Awolowo University, IleIfe, Osun State, Nigeria Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: This study tried to assess the economic, social and financial viability of irrigation the project and examine the efficiency of resource use among the project farmers. Data from all the Sepeteri Project’s farmers during the 2001/2002 seasons were used. Factor share approach was used to examine the resource use efficiency among the farmers. Records of the project’s activities from 1995/96 to 2001/2002 seasons were summarised to some socio-economic performance indices. All the farmers perceived irrigation fee cheap however, 77% attributed their low demand to lack of credits. The irrigation service was acceptable to the farmers with ease of collection. About 67% of farmers do not accept responsibility of making the project a success. Farmers were not efficient in resource use. It was concluded that while the irrigation fee is far below its economic value, it is high enough for the farmers and this prompts them to prefer rain-fed to irrigated cropping. The project was not financially viable due partly to insufficient funding and low level of demand from farmers.

1 INTRODUCTION Up to 1960s, Nigeria was almost self-sufficient in staple food crops from the relatively abundant rainfall. However, from the 1970s, the long drought; the resulting recurrent desert encroachment and the substantial rate of increase in population brought set back Nigeria’s agriculture. The emergence of these three phenomena necessitated public investment in formal irrigation. Irrigation involves development of water resources, conveyance and distribution of water supply at the field coupled with necessary water management exercises (Chukwuma 1993). River Basin and Rural Development Authorities (RBRDAs)’ Irrigation projects were established between 1973 and 1979 to cover every part of the country. They were to carry out a number of functions particularly the development of irrigation infrastructure in their respective areas of operation. Consequently, thousands of tons of crops such as tomatoes, groundnut, wheat, cotton,

Water resources of arid areas

288

millet, maize, etc, were grown by small-holder and commercial farmers. However, the overall performance of the existing irrigation facilities had been on a decline owing to a combination of technical, socio-economic and institutional factors (Nwa 1993). The projects are saddled with inadequacy of untimely funding (Akinkoye 2001) hence, not self-sustained. Consequently, the 1995–97 Corporate Plan mandated RBRDA to generate funds internally to cushion the dwindling funds from the budgetary allocations in order to meet substantial portions of their recurrent costs. Ordinarily, a guided increase in agricultural commercialization leads to increase in purchase of farm inputs, scale (farm size) and specialization in farm enterprises and changes in the role and nature of farm labour inputs. However, this move, if not guided, has the tendency to expose the RBRDAs to been more concerned with the activities that return highest internally generated revenue and possibly less concerned with meeting the irrigation needs of the intended beneficiaries is vital. To this end, while most research efforts on improving the performance of public irrigation projects have focused on the structure, technology and environmental issues, this research focused on social and economic performance of the system as well as the resulting effect on its sustainability and achievements of its statutory functions in the face of commercialization and eventual privatization. The broad objective of this study is the assessment of the socio-economic performance of Sepeteri Irrigation Project. The specific objectives are to: i. Examine the socio-economic characteristics of the farmers; ii. Examine the factors that affect demand for and supply of irrigation services in the project area; iii. Assess the economic, social and financial viability of the irrigated cropping and; iv. Determine the efficiency of resource use among the irrigation project farmers.

2 CONCEPTUAL FRAMEWORK Farmers are primarily concerned with the profitability of their enterprises at individual farm levels. Hence, economic performance of irrigation project farmers is based on a production function and viewed in terms of the efficiency with which farmers combined irrigated plot with other resources such as labour, fertilizers, etc., in the context of institutional framework and management practices of irrigation projects. Farmers are expected to meet economic optimum criteria by adjusting inputs and outputs to relative prices. According to the concept of marginal productivity, a rational producer allocates each variable input according to its market price. This concept is supported by the theory of equilibrium in factor markets under profit maximization, which requires that a factor input be paid its value of marginal physical product (VMP). If a factor is paid higher than its VMP, it is over-utilized and if otherwise it is under-utilized (Henderson & Quandt 1980). However, public irrigation projects in developing nations like Nigeria usually do not meet the conditions for competitive market analysis as explained above because its outputs are natural resources (i.e. land and water), which are developed for national economic efficiency and development. According to Schreiner et al. (1989), such a project is characterized by concepts like natural monopoly, derived demand, etc.

Socio-economic performance of Sepeteri irrigation project in Nigeria

289

Therefore, it is quite possible for project to record negative returns to the agency management because of high cost of capital, yet farmers are expected to make profit consistently. Therefore, rational irrigation policy should ensure that the only sound reason for fixing irrigation fee is the net additional benefits it offers. These benefits accrue to the region and the society as a whole. The major impact of these benefits is to be found in land use, employment, cropping pattern, farm inputs, etc. So, priorities are always given to these benefits above financial returns accruing to the government from irrigation fees. 3 LITERATURE REVIEW 3.1 The state and roles of irrigation projects in Nigeria’s agriculture The practice of irrigation in Nigeria dates back to 700AD. Formal irrigation scheme started in 1926 in Kware, Sokoto State. Subsequently, an irrigation policy for the Northern Nigeria was promulgated in 1963 to construct village-level irrigation schemes. Studies were then conducted to examine the water resources and irrigation development potential in Nigeria. The productivities of Sokoto-Rima and Chad Basin Development Authorities established in 1973 were huge and impressive (Adegbola & Akinbode, 1986). The then Federal Military Government established nine (9) more RBRDAs by 1976 to promote irrigated agriculture in order to enhance food self-sufficiency programmes (OORBRDA 1998). Sepeteri Irrigation Project is a farmer-based irrigation project under the Ogun-Oshun River Basin and Rural Development Authority (O-ORBDA), a parastatal of the Federal Ministry of Water Resources and Rural Development. Consequent to the commercialization programme and the addition of rural water supply function, the River Basin Development Authorities (RBDA) became River Basin and Rural Development Authorities (RBRDAs) since January, 1995. Through the RBRDAs, a number of hectares of lands were available under irrigation. However, the sustainability and efficiency of these formal irrigation projects have started to decline. To arrest this situation and to further explore irrigation potentials, corporate farmers, State’s Agricultural Development Projects (ADPs), private organizations and local governments started investments in small-, medium-scale irrigation projects. Despite these efforts, the various irrigation systems developed so far have not regained the initial performance especially production of import-substitute and export crops. Research findings showed that the productivity of the existing irrigation schemes is on the decline as a result of a combination of technical, social, economic, institutional and political factors (Kolawole 1988). 3.2 Empirical irrigation project performance evaluation The conscientious definition of irrigation performance was reiterated by Bos (1997) as a measure of the degree at which irrigation agent responds to the irrigation needs of its farmers and the efficiency with which the farmers use the resources. Omezzine & Zaibet (1998) used both allocating and irrigation efficiencies as indices of modern irrigation performance in Batinah Region of Oman to examine the economic and technical

Water resources of arid areas

290

efficiencies respectively. Both revealed inefficient water use. The report is that size of irrigated farms and unit cost of water are factors to be considered in the studies of water use and management. Mandal et al. (1995) in an attempt to examine resource use efficiency with respect to farm size of Irrigated HYV Boro Rice Cultivation in Mymensingh District of Bangladesh used factor share approach by estimating a CobDouglas production function. The study revealed that no farm size group allocated resources efficiently. 3.3 Water pricing and irrigation project performance Moore et al. (1994) in a study of four regions of the Western United States discovered that farmers respond to increase in water price by shifting to crops that require less volume of water, hence reduction of acreage of crops requiring high volume of water. This implies that levying too high a charge results in under utilization of facilities such as has occurred on the Sarda Canal in India (NCAER 1959). However, Krishna (1963) found that increase in general water rates would no doubt increase the technical efficiency with which water is applied. It is observed that water abstracted and lifted to field level from wells by human or animal power or by pumps at high private cost, is utilized with much greater efficiency than cheap government canal supplies. Kwanashie et al. (2000) investigated the extent to which poor pricing, poor planning, lack of good management and poor project monitoring and evaluation have affected water resource use in Nigeria. They studied Bakalori Irrigation Project in Nigeria and concluded that these factors undermined water resource management in Nigeria. They then recommended market-based strategies for allocating water between competing users for efficient and cost effectiveness. 4 METHODOLOGY Sepeteri Project is one of the O-ORBRDA’s Irrigation projects supplying irrigation services up to 2001/2002 cropping periods. O-ORBRDA has seventeen (17) farmer-based irrigation projects. However, only two (2) Sepeteri and Itoikin still supplied irrigation services as at 2001/2002 cropping periods. The Sepeteri Project is located in the SakiEast Local Government Area of Oyo State. This is a typical agrarian community. The project was then purposively selected for the study. The project was planned to irrigate 2000ha with sprinkler system. It has 2 dams—Sepeteri I and Sepeteri II of 2.1mcm and 1.3mcm storage capacities respectively for the production of dry season vegetables and Okro. Primary and secondary data were used. One set of structured questionnaire was used to purposively collect primary data from all the forty-four (44) project farmers. They all cultivated a total of 22.35ha. Cross-sectional data of the farming activities during the March 2001 to October 2001 rain-fed cropping period and the November 2001 to March 2002 irrigated cropping period were collected. Secondary data were also obtained from the project records for the periods of 1995/1996 to 2001/2002 cropping seasons. Descriptive and inferential statistics were used to summarize the distribution of data on respondents’ socioeconomic characteristics and factors affecting the supply of

Socio-economic performance of Sepeteri irrigation project in Nigeria

291

irrigation service. Index numbers were used to summarize series of figures over years. The indexes show how much one-year figures differ from another. Usually, a fairly typical year’s figure is taken as a base year figure and others are compared to the base year’s figure. The commonest application of index numbers is in comparison of a series of annual figures. In this study, 1995/96 irrigation season was taken as a base year such that subsequent figures were compared with 1995/96 figures. This year was chosen as a base year to examine how fare have the projects been performing in irrigation services since commercialization move in 1995. This year was assumed to be a logical base year for evaluating performance of a previously public irrigation project. To observe the average annual percentage change over the years, the average annual percentage change was computed as follows: (1) where Indexi=base year Index; indexl=last year index and n=number of years over which the trend is studied (Harper, 1991). To examine the performance of the irrigation projects, performance indices according to Bos (1997) were used. The indices used included Fee Collection Performance, Relative Water Cost, Users Stake in Irrigation System, Financial Self-Sufficiency and Relative Cropping Profit Indices were used to assess the operational and strategic performances of the project management agency. The indexes are specified below: (2) where Irrigation fees collected=total revenue collected on irrigation service during an irrigation season

and Irrigation Fees Due=total revenue collectible on irrigation

service during an irrigation season . The Fee Collection Index reveals the level of acceptance of irrigation delivery as a public service to the farmers i.e. the ease of enforcement of irrigation fee or how affordable the fee is among the intended beneficiaries. (3) where Number of Active project farmers=number of farmers in attendance and Total number of project members=number of project farmers informed and expected to be in attendance. Users Stake in Irrigation Project Index reveals the social capacity of intended beneficiaries and organization in managing and sustaining the project i.e. the level of acceptability of responsibility in making the project a success. The “activeness” of members were quantified using acquired data on the attendance of farmers during the last five consecutive regular meetings called by the management for an agreed upon task such as water distribution, conflict resolution, plot maintenance, etc. (4)

Water resources of arid areas

292

where Actual Income=Total internally generated revenue from irrigation related services and Total MO+M Expenditure=total expenditure on irrigation related services. Financial Self-sufficiency Index reveals the financial viability of the project. (5) where Irrigation Cost per ha=cost of irrigation service per ha (

) and Total

production cost per ha=average total cost of irrigated cropping ( ). Relative Water Cost Index reveals the tendency of farmers abandoning or continuing with irrigated cropping. It is computed on the average. However, The Relative Water Cost Index is perceived to be inadequate to measure the tendency of the farmers abandoning or continuing with irrigated cropping since for some farmers in the developing nations, ends justify the means i.e. (Alimi, pers. comm.) they consider profit far more than the cost. It will then be modified to incorporate the ends, profits from irrigated and rain-fed croppings as specified below: (6) To determine the efficiency of resource use by the respondents, Ordinary Least Squares (OLS) technique was used to estimate parameters of explanatory variables in the postulated Cobb-Douglass production function. The marginal values of inputs used were computed indicating the proportion by which value of crop output changed with one per cent change in the quantity of each input when the quantities of other inputs were kept constant. The production elasticities of the inputs were added together to obtain the returns to scale indicating the proportion by which value of crop output changed with one percent change in the quantities of all the inputs. It is assumed that the value of output depends on level(s) of input(s) such as land, labour, capital and management used and that the production function is a one-equation model (Ogunrowora et al. 1979; Omotesho et al. 1993; Ayanwale 1995). Thus the production function for project farmers was specified as follows: Y=f(X1, X2, X3, X4, u) (7) where Y=Total value of crop output

; X1=size of irrigated farm plot (ha);

X2=expenses on fertilizer and other agro-chemicals ; X3=number of farm household members that assisted in farming activities (man-day); X4=amount spent on hired labour and u=error term. Crop output (Y) was measured in monetary term because two crops—Vegetable (Amaranth sp.) and Okro (Abelmoscus esculentus) were grown together without measuring for sale in standard unit like kilogram. The log-log stochastic production function was fitted for the respondents’ values of crop outputs as follows: ln Y=ln α0+α1 ln X1+α2 ln X2+α3 ln (8)

Socio-economic performance of Sepeteri irrigation project in Nigeria

293

X3+ln α4X4 where ln α0 is the regression constant and ln αi is regression coefficient of Xi. The condition of optimum use of inputs as postulated by the theory of equilibrium in factor market under profit maximization is given by the equation: VMPi=MPPi×Py (9) where VMPi=value of marginal physical product from using additional unit of input Xi; MPPi=marginal physical product from using additional unit of input Xi and Py=market price of the output. So the Allocating/Pricing Efficiency Index is given as (10) where MICi=marginal input cost of input i. (i=1, 24). a priori, the expected signs of the explanatory variables are positive. 5 RESULTS AND DISCUSSION 5.1 Respondents’ socio-economic characteristics Table 1 reveals that about 57% of the farmers are indigenes of the project village. The rest 43% are of distant origin. Almost all the respondents, (95%) resides within the project village. This

Table 1. Summary of socio-economic characteristics of respondents. Characteristic Village of origin: Project/Neighbouring villages *Distant villages Village of residence Project/neighbouring villages *Distant villages Age: 21–30 31–40 41–50 Above 50 Sex: Male Female Level of formal education:

Frequency

% Frequency

25 19

56.82 43.18

42 02

95.45 04.55

04 19 13 08

09.10 43.20 29.50 18.20

40 04

90.90 09.10

Water resources of arid areas

294

No formal education 07 15.91 Primary education 16 36.36 Secondary education 18 40.91 Vocational studies 03 06.82 Years of farming experience: 5–10 10 22.73 11–20 18 40.91 above 20 16 36.36 Year of project participation: 3–6 16 36.40 6–10 23 52.20 above 10 05 11.40 Nature of farming occupation: Full-time 05 11.36 Part-time 39 88.64 Factors affecting size of irrigated plot demand: Credit availability 34 77.27 Cost of irrigation plot in high 05 11.36 Other non-farm engagement 04 09.09 Others 01 02.27 *These are villages farther than 50Km away to Sepeteri, the project village.

revealed that the indigenes within the Project area participated well in the project. About 73% of respondents were within age bracket of 31 and 50 years while about 18% are above 50 years. Also, women scarcely participated in the projects. About 91% of farmers were male. The reason may be that the community is a typical agrarian community where men are predominantly engaged in farming and women engage in other economic activities or assist the male household heads in farming operations. About 16% had no formal education while none had tertiary education, while 77% had between primary school and secondary school education. All the Project farmers had above 5 years of farming experience. In fact, about 37% had above 20 years of farming experience. This result justifies locating the project in the area to help agricultural development and also support the tenet that land should be allocated to farmers with proven commitment to farming as a career. In the same vein, the project farmers had long years of participation. Above half had been with the project for over 6 years. However, few (about 11%) of farmers were full-time farmers. Majority, 88.6% were involved in other economic activities. About 77% acknowledged credit availability as

Socio-economic performance of Sepeteri irrigation project in Nigeria

295

*Table 2. Farm specific characteristics of the project and factors affecting supply of irrigation services. Characteristic

Description

Irrigated plot cost (irrigation fee)

–

3300 per ha and collected after marketing of produce

Upland cost

–

Operation problems

– – – – –

800 per ha and collected after marketing of produce Insufficient resource provision Deterioration of physical structures Inadequate finance Lack of tractor and equipment Occasional invasion by Fulani cattle nomads Harsh harmattan between January and February Monkey pest invasion by February Fuel shortage for water pumping Break down of vehicle to transport produce to market Lack of ready markets for produce Electricity failure for pumping irrigation water 4 months (usually between January and April) Implementation committee comprising of project agents and farmers

Management constraints

Other risks and problems peculiar to the project

– – – – – –

Irrigation period

–

Who is responsible for water – allocation * Response from the project staff.

limiting factors of irrigated plot size while 11% perceived irrigation service fee as high while others expressed engagement in other activities as a constraint. 5.2 Factors affecting delivery and supply of irrigation services Table 2 shows some of the farm characteristics and factors affecting supply of and demand for irrigation services in the project. There is crop restriction to vegetable and Okro. The irrigated and upland costs were 3300 and 800 per ha per cropping season respectively and were constant over years. There was no coordinated Water Users Association but the project manager and farmers met occasionally as matter arose especially to allocate land. The project often witnessed inadequate finance, breakdown of tractor and occasional invasion by Fulani cattle nomads. The physical structures are deteriorated. The irrigation period was usually between 4 months of January and April each year. Some risks and peculiar problems that usually discouraged project participation included harsh harmattan between January and February each year, monkey pest invasion by February each year and lack of ready markets for the produce.

Water resources of arid areas

296

5.3 Socio-economic and financial performance of the project Table 3 reveals that the average irrigation fee collection performance index is 96%. This implies ease of enforcement of irrigation service charge i.e. despite the payment is allowed until after marketing of produce. The User’s stake performance index is 67%. This may be interpreted to mean that about 67% of the farmers were actively involved in the last five (5) obligations required of them for running the project. However, the performance declined at an average rate of 8%. Average Financial self-sufficiency index is 29%. This implies recovering about 29% of the expenditure on the irrigation services rendered to farmers. This implies decline in costs recovery at the annual rate of 25%. Cost of irrigated plot as a percentage of total cost is 20%. This is not much different from 18% reported by Mandal et al. (1995) among irrigation farmers in Mymensingh area of Bangladesh. However it should be noted that the percentage of irrigated plot cost to total production cost is high enough to make farmers abandon irrigated for rain-fed cropping because most mentioned lack of credit facility as a limiting factor to demand for irrigated plot. Besides, the ratio of profit from irrigated cropping to profit from rain-fed cropping is 1.08. This implies that there is no statistically significant difference between profits to irrigated and rain-fed croppings.

Table 3. Socio-economic performance indices of the project. Index Fee collection Users’ stake Financial selfsufficiency Relative irrigated plot cost (as a % of total cost) Relative cropping profit

Average index (%)

Average annual % change in index

96 67 29

4.3 −8 −25

20

1.08

Table 4. Estimates showing efficiency of resource use from the 2001/2002 irrigated cropping. Variable

Average Regression VMP coefficient

MIC

Dependent variable: Y 41,569.91 Explanatory variables: Intercept – 10.67 – – – 0.51 0.718*(4.55) 58,686.93 3300 17.78 X1 1900 0.069 (0.89) 3.16 1900 0.00167 X2

Elasticity of production

– 0.72 −0.069

Socio-economic performance of Sepeteri irrigation project in Nigeria

X3

1.0

X4

500

−0.078 −648.49 500 −1.30 (0.44) −0.094 −1.90 500 0.0038 (0.51)

297

−0.078 −0.094

Other statistics: N=44 R2=0.39 F=5.79* Return to scale=0.62 *Significant at 5%.

5.4 Efficiency of resource use among the respondents From Table 4, 39% of variability in the value of output was explained by the set of explanatory variables captured in the model. The 5.79 F-statistic is statistically significant at 5% level indicating that joint effect of these explanatory variables is significant. The coefficient of irrigated plot size is positive and statistically significant at 5% level. The coefficient of amount spent on chemical has positive sign but not significant. This implies negligible increase in output value results from additional unit increase in the amount spent on chemicals. The coefficient of family labour is negative and not significant. This implies additional use of family labour brings about decrease in output value. This is contrary to expectation and may result from cultivating too small a plot. Coefficient for hired labour is negative and insignificant. This implies that additional use of hired labour results in decrease in output value. The regression constant is 10.67. This is positive and implies that on the average farmers are technically efficient in realizing as much as 11 times in value of input used. This agrees with result reported by Ogunfowora et al. (1979), Omotesho et al. (1993) and Ayanwale (1995). However, the farmers were not efficient in resource allocation. The values of marginal physical products are far different from corresponding marginal input costs (MIC). VMP of irrigated plot is much higher than MIC of irrigated plot. This implies there is scope to increase irrigated plot size to generate higher income. The small and or negative ratios of VMP to MIC of other inputs imply they were over utilized in combination with irrigated plot. Also, the return to scale is 0.62. This indicates decreasing return to scale that the farmers operating under irrational zone of production. 6 CONCLUSION In conclusion, there is higher level of participation from the neighbouring villages to the project. They were mostly resident farmers. Also, they were mostly male with low level of formal education but they had long years of farming experience and project participation. Majority was aged between 31 and 50 years and above. Most of them were part-time farmers and attributed their small level of irrigation participation to lack of

Water resources of arid areas

298

credit facility. Inadequate funding and deteriorating structures hindered the level of irrigation supply of the project. The performance indices revealed that the irrigation service is acceptable to the intended beneficiaries. Higher cost per hectare of irrigated cropping connotes tendency among the farmers to abandon irrigated cropping for rain-fed cropping. However, the farmers do not accept responsibility of making the project a success. The project is not financially self-sufficient partly for insufficient funding for operations and partly low level of demand from farmers. The farmers were inefficient in resource use. The results call for reform in management of irrigation system such that privatization efforts should not tie down peoples’ land unused; Specifically farmers should be encouraged into coordinated and recognized WUA that incorporate credit lending and efficient marking. Further research should look at ways to rationalize family labour. In the same vein, on hired labour other means of bargaining should explored. REFERENCES Adegbola, A.A. & Akinbode, I.A. 1986. A review of old and current agricultural development schemes in Nigeria: Lessons for future programme designs. In Agricultural Development in Nigeria. Ife Journal of Agriculture special publication, 8:1–34. Akinkoye, O. 2001. An overview of organization and management of public sector irrigation schemes. Paper presented at the National Workshop on Participatory Irrigation Management organized by National Agricultural Extension and Research Liaison Services (NAERLS), Ahmadu Bellow University, Zaria in collaboration with the Department of Irrigation and Drainage, Federal Ministry of Water Resources, Abuja. 26–30 March 2001:12pp. Alimi, T. (Personal communication). Ayanwale, A.B. 1995. Resource use efficiency in cassava processing in Oyo North Area of Oyo state, Nigeria. Ife Journal of Agriculture 16, 17:123–135. Bos, M.G. 1997. Performance Indicators for irrigation and drainage. Irrigation Drainage Systems 11(2): 119–137. Chukwuma, G.O. 1993. Some Considerations in Developing Irrigation Research Priorities for Nigeira. Proc. National seminar on Irrigation Research Priorities for Nigeria held at the University of Ilorin, Nig. 20–23 April 1993:65–71. Henderson, J.M. & Quandt, R.E. 1980. Microeconomic Theory: A mathematical Approach 3rd ed., McGraw-Hill Kogansha Ltd. Japan, 420pp. Herpar, W.M. 1991. Statistics, London, Pitman Publishing: 501pp. Kolawole, A. 1988. RBRDAs and vulnerability to hunger in Nigeria, the case of the South Chad Irrigation Project. Food Policy 13(4):389–396. Krishna, R. 1963. Farm Supply Responses in India-Pakistan: A case study in the Punjab Region: Economic Journal, Sept, 1963. Kwanashie, M.A., Togun, A., Ajobo, O. & Ingawa, S.B. 2000. Nigeria Water Resources Management Strategies—Economic and Financing. Technical Report, 16pp. Mandal, K.C., Sabur, S.A. & Molla, A.R. 1995. Resource use efficiency of irrigated HYV boro rice cultivation by difference farm size groups and its impact on employment and distribution of income in DTWII project area of Mymensingh Bangladesh J. Agric. Econs, 8(1):71–87. Moore, M.R., Gollehon, N.R. & Carey, M.B. 1994. Multi crop production decisions in western irrigated agriculture: the role of water price, American, Journal of Agricultural Economics, 76(4):859–874. NCAER (National Council of Applied Economics Research) New Delhi 1959. Criteria for fixation of water rents and selection of irrigation projects, London, Asian Publishing House.

Socio-economic performance of Sepeteri irrigation project in Nigeria

299

Nwa, E.U. 1993. Irrigation Research Priorities for Nigeria. Proc. National Seminar held at the University of Ilorin 20–23 April, 1993, Ilorin, Nigeria, Nwa, EU, Pradhan, P. (eds) IIMI, 104pp. Ogunfowora, O., Esang, S.M. & Olayide, E.O. 1979. Resource productivity in traditional agriculture: a case study of four agricultural divisions in Kwara State of Nigeria. Journal of Rural Economics and Development 9(2):119–131. Omezzine, A. & Zaibet, L. 1998. Management of modern irrigation systems in Oman: allocative vs. irrigation efficiency. Agricultural Water management, 37(2):99–107. Omotesho, O.A., Olufe, J. & Oladeji, S.O. 1993. Resource productivity in food crop production in some selected villages of Oyi Local Government Area, Kwara State, Nigeria. Ife Journal of Agriculture 14(15): 90–97. O-ORBRDA 1998. Federal Republic of Nigeria, Ogun-Oshun River Basin and Rural Development Authority 1997 Annual Report, January, 1999. 49pp. Schreiner, D.F., Badger, D.D., Welsh, M.P. & Suprato, A. 1989. Policy Applications in Natural Resource Projects. In Agricultural Policy Analysis Tools for Economic Development (ed) L.Tweeten, London, Westview Press, 279–321.

Theme D: Application of geophysical, GIS, and remote sensing techniques

Mapping vegetation for upscaling transpiration using high-resolution optical satellite and aircraft images in Serowe, Botswana Y.A.Hussin1, D.C.Chavarro1, M.Lubczynski1 & O.Obakeng1,2 1

International Institute for Geoinformation Science and Earth Observation (ITC), Enschede, The Netherlands 2 Geological Survey of Botswana, Lobatse, Botswana Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: High spatial resolution images of multi-spectral digital TETRACAM camera were used to map vegetation for upscaling transpiration from tree-bush-shrub of Serowe, Botswana. The camera was mounted on a small aircraft to collect data in 30, 60 and 100cm spatial resolution. The results were compared with one-meter improved multispectral IKONOS satellite images. The high resolution airborne images show high potential for mapping tree-bushy-shrubby vegetation of the study area for up-scaling transpiration. The spectral characteristics of the high spatial resolution images are similar to IKONOS satellite images, while the spatial characteristics of the high spatial resolution images are much better than the one-meter MS IKONOS satellite images.

1 INTRODUCTION As part of the water cycle in the nature or what is well know as the hydrological cycle, surface water is heated by solar radiation and thus evaporated to the atmosphere. However, water in plant is emitted through leaves by a process called transpiration. Vegetation cover is a major component of the hydrological cycle. It has influence on the hydrology of both ground and surface water and on soils. The illogical use or abuse of natural vegetation can have a major effect and consequently changes the hydrological cycle and produce adverse effects. These effects can be very dramatic in arid and semiarid regions. In these regions the climate is very sever with extremely hot temperatures up to 50°C and high evaporation and low annual rainfall. In Botswana, just like any other semi-arid country, water is scarce and surface water is available only for short periods after the rainy season. The main water supply is groundwater. Vegetation in Botswana, as

Mapping vegetation for upscaling transpiration

303

in any other semiarid ecosystems, is often characterized by tree-bush-shrub-grass savannas and woodlands. The increasing demand for fodder and fuel wood has led to drastic declinations on the vegetation cover. The transpiration from vegetation is contributing to the development of cloud cover and consequently to the precipitation. Thus, the amount of ground water is related to the vegetation cover. Research work in Botswana have proven the assumption that water loss comes from plants relies under the experience of rooting plant systems depth for (68m) Le Maitre et al. 2000. The transpiration measurements where done during dry season were soil evaporation was almost negligible. Accurate ground transpiration estimates become more important being the connecting link between groundwater balance and transpiration model from plant. Therefore, a relationship can be established between the ground water and the biophysical parameters of vegetation cover. Accurate mapping of vegetation cover would lead to assessment of the biophysical parameter of vegetation and consequently to transpiration. The objective of this paper is mapping vegetation cover for up-scaling transpiration using high-resolution optical satellite (IKONOS) and aircraft images (TETRACAM). The study area of this research is located in Serowe, Kalahari, Botswana. 2 VEGETATION MAPPING WITH REMOTE SENSING Forests cover large areas of the global land surface. For many developing countries, it represents an important income source for their economies. Due to over exploitation, forests are currently under constant risk. The protection of forests from disasters (e.g. fire, disease, erosion, deforestation, over grazing) over extensive area is difficult without having any information such as condition, area, species, age classes and volume. With these types of information, it is possible to make a proper management of the forest by identifying and selecting the appropriate area for different management purposes, such as, harvesting, protection, etc. Having all these information collected, there is a need to store the referred information properly, for better and comprehensive use. For this reason, forest maps play an important role in organizing gathered information for further strategies and policies determination in order to make the best use of forest. Different approaches for mapping, like land survey, aerial photographs and satellite imagery can be used depending on the level of detail required and the extension of the area under study. For wide areas, satellite imagery has been shown effective for forest classification and consequently mapping. It is recognized that different satellite or airborne imagery can give different results in terms of information extraction. These different results relate to differences in spatial and spectral resolution. Vegetation mapping involves the evaluation of the existing data and information, collecting field data or ground truth, analysing the data and finally developing the vegetation map and validate it (USGS, 1994). Riquene (2002) have studied the vegetation condition of the current research area using Landsat-TM images and ASTER Optical scanner sensors. The study concluded that ASTER images resulted in better vegetation map than TM because of its higher spatial resolution of the 15 meter than TM of 30 meter. Further more the results showed that ASTER data gave more

Water resources of arid areas

304

reliable vegetation maps than Landsat TM data. Mapanda (2003), following the steps of Riquene research, who used high spatial resolution (4 meters) multi-spectral satellite images of IKONOS in comparison with ASTER images of 15 meter spatial resolution, have concluded that IKONOS images gave much better results in mapping the vegetation cover of Serowe, Botswana. This research is going one step further in using higher spatial resolution of multispectral images. These images were acquired using Tetracam multi-spectral digital camera. The camera was mounted on a small aircraft and has collected images in 30cm, 60cm and 100cm spatial resolution. This paper is presenting the first preliminary results of the use of the Tetracam airborne multi-spectral digital images for mapping tree-bushshrub-grass-savannas and woodlands vegetation cover of part of Serowe, Botswana. These images will hopefully be used for up-scaling transpiration of the vegetation in this area in conjunction with IKONOS MS data. 3 STUDY AREA The study area is located in the Central District, about 275km NE of Gaborone the capital of Botswana. Topography is gentle, which varies from 1060 meters above sea level to approximately 1240. It is characterized to be lower in the east and southeast of the region, and the highest location in the vicinity of the escarpment edge. From these ones the average slope is 5% and it gradually decrease to less than 1% towards the east and southeast. Soils units, which can be found in that region, are related to arenosols, regosols, lixisols, luvisols and vertisols. Arenosols are the most common soil units in the study area. It has low moisture retention capacity than the other soil units. Climate is a semi-arid with a mean annual rainfall of 447mm. Rainfall occurs mainly in the summer fallowed by a dry winter season. Summer season stretches from October to April and the winter begins in May to September (Tyson, 1986) (Obakeng, 2000). Main vegetation type is thought that belong to the Northern Kalahari Tree and Bush Savanna. Trees are mostly of Acacia species, which are characterized by the marked tendency to occur in cluster, and are normally accompanied by a variety of grass species such as Ariatida and Eragrotis. Vegetation communities are determined by location on either sandveld or hardveld areas. Dense vegetation is found within and along river courses. This suggests that the vegetation density is governed by the availability of water, which may be partly controlled by topography and geomorphology (Obakeng, 2000). 4 DATA AND ANALYSIS The airborne multi-spectral data was collected using TETRACAM multi-spectral digital camera, which collects its data in three spectral bands namely red, green and near infrared. The data is collected in a rectangular frame of 1280×1024 pixels. The size of the pixels (e.g. ground resolution) would depend then on the altitude of the aircraft above the ground. The camera would saved the image in DCA format (Digital Camera Format), which is a compressed file that can be un-compressed and transferred to Bitmap format

Mapping vegetation for upscaling transpiration

305

that can be imported to any image processing software. The Airborne data was collected in three different spatial resolutions 30, 60 and 100cm. An area of 10×10km was selected as a study site. It is located in the Hardveld part of the Serowe terrain, on which two multi-spectral IKONOS satellite scenes of November 2001 and February 2002 were collected. These images are collected with 4 spectral bands (blue, green, red, and NIR). The spatial resolution of these data sets is 4 meters. A MS image, so called panchromatic sharpen, was available too. The Pan-Sharpen image is a MS image fused with the Panchromatic image of IKONOS, which has 1 meter spatial resolution. Thus the spatial resolution of the new MS image will be improved to 1 meter. For this study area, two aerial surveys were implemented to collect the multi-spectral digital camera data. The first aerial survey was done in November 2003 and the second one was done in February 2004. These surveys were designed and implemented using Aerial-Photography types of survey. The survey divides the area into flight lines. Within each flight line, images were collected with a front overlap of 20% and a side overlap between flight lines of 20% too. The following data where collected: 1. 30cm spatial resolution: 39 flight lines with a total of 910 images 2. 60cm spatial resolution: 21 flight lines with a total of 333 images 3. One meter spatial resolution: 14 flight lines with a total of 187 images. A qualitative approach was used in the analysis of airborne multi-spectral digital images, which mainly involves visual interpretation, spectral signature measurements, spatial features measurements and comparisons of different spectral and spatial data resolution. The same approach was used with the MS IKONOS satellite data to be compared to the airborne data. 5 RESULTS AND DISCUSSIONS The results presented in this paper are the first preliminary findings of some exploratory analysis of airborne MS digital images. Supervised classification and accuracy assessment was not done because an organized fieldwork to collect ground truth was not done yet in the study area. The authors are planning for one in early May of this year (2004). A spectral signature analysis of the 30cm, 60cm and one-meter spatial resolution images of the MS digital Tetracam airborne camera using the digital interpretation of the false color composite, unsupervised classification, and Normalized Difference Vegetation Index (NDVI) (Figures 1–6) showed that the sample used has 3 different spectral classes which referred to 3 different species of the bushy vegetation in the area. It also showed two high contrasted spectral classes, which refer to a soil and a grass classes the selected sample shown in the mentioned figures. A comparison of the above findings with the same signature analysis to the MS IKONOS image of improved one-meter spatial resolution showed similar results of 3 spectral classes of the bushy vegetation and two other high contrasted classes representing the soil and grass classes.

Water resources of arid areas

306

Figure 1. Color composite image of 30cm spatial resolution of MS airborne MS camera.

Figure 2. NDVI map the 30cm spatial resolution of MS airborne image.

Mapping vegetation for upscaling transpiration

307

Figure 3. Unsupervised classification map of the 30cm spatial resolution of MS airborne image.

Figure 4. Color composite image of 60cm spatial resolution of MS airborne MS camera.

Water resources of arid areas

308

Figure 5. NDVI map the 60cm spatial resolution of MS airborne image.

Figure 6. Unsupervised classification map of the 60cm spatial resolution of MS airborne image.

Mapping vegetation for upscaling transpiration

309

Figure 7. Color composite image of one-meter MS IKONOS satellite image.

Figure 8. NDVI map of one-meter MS IKONOS satellite image.

Water resources of arid areas

310

Figure 9. Unsupervised classification map of one-meter MS IKONOS satellite image. Figures 7–9 show the false color composite, unsupervised classification, and Normalized Difference Vegetation Index (NDVI) of the MS IKONOS satellite image. A general comparison of IKONOS images and the airborne MS image showed that both types of images are close in their spectral characteristics, especially the improved one-meter resolution of IKONOS as compared to the 30 and 60cm resolution images of the airborne data. A spatial analysis of several selected objects (e.g. bushes, trees, soil and grass) on the images showed that when spatial resolution increases the accuracy of the information extracted increase. For example a canopy diameter of a single tree measured on the ground is 5.5 meter. The same tree canopy diameter measured on the one-meter, 60cm and 30cm spatial resolution of the airborne MS images are 7.1, 6.13 and 5.38 meter respectively. This is a clear evidence that the higher the spatial resolution of the image the better the accuracy of the analysis or interpretation output. Moreover, the higher the spatial resolution the higher the amount of details extracted from the images as a results of the interpretation or any image analysis technique (e.g. classification or image transformation such as NDVI). For example, some of the tree species in this study area are likely to grow in clusters structure. These clusters may include 3–5 trees in one cluster. On average the crown diameter of these trees say 5 meters. Consequently a total area of the canopies from such cluster may reach up to 125 square meters. The lower the spatial resolution (e.g. one-meter or lower), the less details that represent such a cluster on the image and vis versa. However, in the case of a small bush or shrub, which, has a canopy of approximately one-meter diameter may not show clearly on the image because of the open crown structure. In such a case the spectral reflectance of the soil will dominate such the

Mapping vegetation for upscaling transpiration

311

reflectance from the canopy of the bush. While when using higher resolution (e.g. 30 and 60cm) the representation or the appearance of a small bush will be possible. Therefore, as the spatial resolution increase the spatial information extracted about an object on the images (e.g. a bush or a tree) will consequently increase. The information will include the surrounding area (e.g. soil or grass) of the tree or the bush targeted. This means that using higher spatial resolution we can define the size and shape of any tree or bush much accurate than using lower resolution. The size of the crown of a bush or tree is effecting the estimation of the transpiration of that bush or tree. Consequently this will effect the process of up-scaling transpiration from the area in general. As far as the spatial resolution is concern, a general comparison of IKONOS images and the airborne MS image showed that the information extracted from the 30cm and 60cm spatial resolution images of the airborne MS images is much better than the onemeter resolution of IKONOS image. 6 CONCLUSIONS The following conclusion remarks can be drawn: – The high spatial resolution 30 and 60cm multi-spectral digital Tetracam images have high potential for mapping tree-bushy-shrubby vegetation of semi-arid area (e.g. Serowe, Botswana) for up-scaling transpiration. – The spectral characteristics of the high spatial resolution images are similar to IKONOS satellite images. – The spatial characteristics of the high spatial resolution images are much better than the one-meter MS IKONOS satellite images.

ACKNOWLEDGMENT This research work was partly supported by the internal research fund of GWFLUX Project at ITC. However, Botswana Geological Survey (BGS) has offered the main financial support of the aerial survey missions, fieldwork logistics and transportation. The authors appreciate and acknowledge the support of Botswana Geological Survey. REFERENCES Le Maitre, D.C., Scott, D.F. & Colvin, C. 2000. Information on interactions between Groundwater and Vegetation relevant to South African Conditions: A review. Groundwater: Past Achievements and Future Challenges, Silili et al. (eds). Balkema, Rotterdam, 959–962. Mapanda, W. 2003. Scaling-up and Mapping Transpiration Using Remote Sensing and GIS: A Tool for Water and Forest Management. Unpublished MSc, ITC—International Institute for Geoinformation Science and Earth Observation, Enschede. Obakeng, O.T. 2000. Groundwater recharge and vulnerability: A case study at the margins of the south-east Central Kalahari Sub-basin, Serowe region, Botswana. Unpublished MSc, ITC— International Institute for Geoinformation Science and Earth Observation, Enschede.

Water resources of arid areas

312

Riquene, A.H. 2002. Vegetation mapping in Arid Zones: A multi-sensor analysis, the relationship between Vegetation Distribution and Environmental Factors: A case study in Serowe, Botswana. Unpublished MSc, ITC—International Institute for Geoinformation Science and Earth Observation, Enschede. Tyson, P.D. 1986. Climatic Change & Variability in Southern Africa. Cape Town, South Africa: Oxford University Press. USGS, 1994, 19 July 2001. Field Methods for Vegetation Mapping. USGS-NPS. Available: http://%20biology.usgs.gov/npsveg/fieldmethods/ [2001, 24 August 2001].

Gravity study on groundwater structure in Central Butana (Sudan) K.M.Kheiralla TU Bergakademis, Freiberg, Germany A.E.Ibrahim El Neelain, University, Sudan Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: A number of isolated outcrops of Pre-Cambrian Basement Complex rocks scatter over the central plains of Sudan. In the Butana region, several hills occur prominent landmarks within the predominantly flat clay plains. The solid geology of the Butana plain is rather concealed under a veneer of variable thickness of superficial clays covering over 70% of its total area. The present study was suggested in an effort to delineated and define the mentioned anomaly in more detail and give a reasonable geological and hydrogeological exploration for its existence. Integrated gravity methods and geological feature were applied to achieve the above aims. A total of 200 gravity points were conducted in the study area and they are compile to 275 gravity points acquired, by Sun Oil Company (approximately 2475Km2 were covered in the study area). Gravity data analysis was performed by “GEOSOFT” packages. The result show that the gravity lows is largely attributed to the occurrence of low-density rocks (granite intruded) into the high-density rocks (green schist) of the Butana region. On the other hand the gravity high zones unambiguously coincide with the areas of known shallow Basement Complex. This gravity low is largely attributed to the occurrence of lowdensity rocks (granitic intrusion) into the high-density rocks (green schist) of the Butana region. Thus for it is not unusual gravity measurements in such areas introduced by granitic into country rock of higher densities, to reflect apparently anomalies of typical sedimentary basins. However such ambiguity can be resolved by computation of the second derivative.

1 INTRODUCTION Geophysical prospecting conducted by Sun Oil Company revealed strong negative gravity anomalies in Butana region. It is suggested that sedimentary basins might have caused these anomalies (Ibrahim, 1993; Ibrahim et al., 1996). In the study area the

Water resources of arid areas

314

anomaly (Around Jebel Mundara) lies partially in the Basement rocks. Therefore this research aims to verify the existence of these anomalies and clearly define their extension. In addition, the study aims at determining the type and the dimensions of the rocks causing these anomalies. The study area lies between the Blue Nile River and River Atbara and occupies the Central Butana area, It is bounded by latitudes 14°60′N and 15°80′N and longitudes 34°00′E and 35°20′E. The study area covers about 10,000km2 (Fig. 1). The study area is generally flat, with a gentle slope to the North. The general altitude of the plain is about 500m above mean sea level (m.s.l). The flat monotony of the plain is occasionally broken by some protruding low to moderately high hills or hill chains, which hardly exceed 200m above the ground surface. The climate of the area is arid to semi arid zone of Sudan, characterized by a short duration of a rainy session in summer (July–September), and along dry season for the rest of the year. The average annual rainfall is about 200mm. Average annual temperature over the Butana area is around 40°C in summer (March–October) and 25°C in winter (November–February).

Figure 1. Location map of Central Butana. Table 1. Geological column for butana region (after Iskander et al., 1993). Formation

Age

Superficial deposits Tertiary volcanic

Quaternary/Recent Tertiary

Gravity study on groundwater structure in Central Butana (Sudan)

EL Burg Basal conglomerates The Basement complex

315

Jurassic to cretaceous Precambrian

2 GEOLOGY OF THE STUDY AREA A number of isolated out crops of Precambrian Basement complex rocks scatter over the central plains of Sudan. In the Butana region, several hills occur and make prominent landmarks within the predominantly flat clay plains. The solid geology of the Butana plain is rather concealed under a veneer of variable thickness of superficial clays covering over 70% of the Butana area. Several metasedimentary sub-parallel belts extend for 10–25km in the form of low to moderately elevated ridges surrounded by Butana clay cover (Iskander et al., 1993). The generalized geological column for Butana region can be summarized as shown in Table 1. The majority of the Basement rocks are concealed under the cover of the Butana clay plain. As mentioned, the structural domain in the area is characterized by northeast trending lineaments (Fig. 2). Exposed structures in the metasediments display complex shearing/faulting and tight folding with dipping axial planes where the axes generally have NE-SW trends and with SW mergence Iskander et al. (1993). 3 GRAVITY METHODS 3.1 Introduction The gravimeter used in this survey is a Lacoste & Romberg gravimeter, model D108, which has calibration constant of 1.0863mGal/div and 200mGal measuring range. A total of 200 gravity readings were measured in the study area following a loop survey, with a spacing of 2km between stations (Fig. 3). An area of approximately 2412km2 has been covered. In addition, about 275 gravity points (approximately 2475km2) acquired, by Sun Oil Company were compiled, in the study.

Water resources of arid areas

316

Figure 2. Geological map of Central Butana area modified after (Ahmed & Ayed, 1996).

Figure 3. Location of measured gravity points.

Gravity study on groundwater structure in Central Butana (Sudan)

317

Elevation of the gravity station was determined by Global Positioning System (Garmin II GPS Model 1999), with ±10m accuracy. The Global Positioning System, (GPS) device was used to determine the altitudes of the gravity stations and also for navigation. Control was provided by the available altitude benchmarks in the area and by elevation contour maps of scale 1:100,000, determined by the Survey Department in Khartoum. In addition the altitudes of the previous data (Sun Oil data), were determined by micro barometric altimeter. In gravity measurements effects are produced by sources not directly related to the geological objectives or interest of the study, therefore certain reductions or corrections are necessary to remove these effects. Such corrections include drift, tidal, latitude and elevation correction. This compensates for the earth tides, generated by the complex gravitational interaction on the Earth by the Sun and Moon. These effects are often calculated from table published by Geophysical Journal worldwide. In this work it was done by “GEOSOFT” computer program. 3.2 Gravity data processing 3.2.1 Production of Bouguer anomaly map Presentation of the corrected gravity values is commonly made in the form of contour maps, particularly where the survey has covered a grid of more-or-less evenly spaced stations. Exceptionally, where well-isolated profiles have been surveyed, to obtain crosssectional information on a structure, then the results may be presented in the form of profiles. A sequence of profiles may be shown in stacked form, in proper relative location, on a plan map. In either case, the horizontal scale of the presentation should be inversely related to the distance between the gravimeter stations. Also, the contour interval (mGals) may be inversely related to the scale of the presentation. Software programs are available, by means of which either contour map or profile presentation may be conveniently and quickly made. In this work, gridding was performed by “RANGRID”, program of the GEOSOFT package (GEOSOFT manual, 1989). “RANGRID” produces a minimum curvature grid from data randomly distributed or along non-parallel traverses. The method utilizes different available interpolation options, (e.g. Akima, Cubic etc) to calculate the value of object function at the grid points (original data) that falls within a circle with a given radius centered at the grid points. RANGRID roughly smoothes gaps in the acquired data. In the present study, although data have been acquired along lines, but their irregular points spacing tend to make the gravity data look randomly scattered thus fore they have been subjected to interpolation or gridding process by “RANGRID” which seemed to be a suitable technique (GEOSOFT manual, 1989). The resolution of the produced Bouguer gravity map depends on the choice of the grid cell size, as demonstrated by comparison between the maps shown in Figures (4, 5, and 6), which have been produced by 0.0025, 0.04, and 0.4 cell sizes respectively. To judge on the optimum cell size that resolves the Bouguer gravity map, variations of GB.A has been plotted against corresponding variation of the cell size. It is clearly that no practical displacement (change) has occurred beyond 0.4 cell size. Thus this cell size

Water resources of arid areas

318

(0.4) is consider as the optimum grid cell size suitable for production of the Bouguer gravity map. Contouring of the observed gravity data was performed by “CONTOUR” program of the GEOSOFT package, whose basic function is to thread contour lines through constant levels, defined in a gridded GEOSOFT data file (GEOSOFT manual, 1989).

Figure 4. Bouguer anomaly map, cell size 0.0025. 3.2.2 Production of the residual map The construction of a residual anomaly map due to local structures is therefore a process by which one removes the regional gravity effects. This task could be performed by numerous methods mentioned in the geophysical literature (Seigel, 1995), however in a broad sense they might be classified into graphical or analytical methods.

Gravity study on groundwater structure in Central Butana (Sudan)

Figure 5. Bouguer anomaly map, cell size 0.04.

319

Water resources of arid areas

320

Figure 6. Bouguer anomaly map, cell size 0.4. The gravity measurements at surface determine the sum of all effect from grass root down to the earth crust. Therefore gravity interpretation frequently begins with some procedure, which separate anomalies of interest from superficial disturbance or from deep regional effects. Various methods may be elaborated to perform the separation of the anomalies in order to emphasize the important and interesting features and to suppress the others. These methods include the following. 3.2.2.1 Graphical methods The regional is far away from measuring points it is represented by a long wavelength anomaly due to deep structure. The regional is sometimes shown as a straight line as a result of smoothing a X-Y plot or contour map anomalies. The regional (long wavelength) anomalies mainly due to the effect of deeper structure (lower crust, mantle and core), while the residual (short wavelength) anomalies represent the shallow structure (near surface and crust structure) and may be upper part of the mantle. The graphical methods involve estimation of the regional field from profile plots or contour maps. The advantage of these methods is that, control could be provided by the available geological information (e.g. Basement depth), obtained from boreholes in the survey area.

Gravity study on groundwater structure in Central Butana (Sudan)

321

3.2.2.2 Filtering techniques One of the most important problems in the interpretation of potential field data is to characterize it into different geological structures. The regional one of these methods is the analytical or the filtering techniques were employed which are characterized by: Various methods may be employed to separate anomalies. These include: regional, residual, derivative, low pass filter and high passfilter. However, the selection of the appropriate method depends on the nature of the Bouguer anomalies and the empirical judgment of the interpreter, which is of course vital. On the other hand several of the analytical methods are commonly used for determining the regional and then the residual fields. Griffin (1949), Agocs (1951), Fajklewics (1959), and Abd el Rahman et al. (1983), used in their respective techniques, linear combinations of the average fields on a number of concentric circles of different radii to represent the residual at the common center. The residual field is given by: R= G−Z=G−(ax+by+c) Where, G is the observed gravity, R is the residual field, Z is the regional field, a, b and c are constants. The condition for the above equation is that ∑R2=min These are called residual, but do not posses any relationship with local anomalous mass (Paul, 1967). The only physical significance of this residual is their proportionality to the second vertical derivative value; hence their zero contours coincide with the zero contours of the second derivatives (Nettleton, 1976). Lately dependable method have been introduced by Paul (1967), for computing the second vertical derivatives, consequently the residual determined by the previous method loose much of their significance. 3.2.2.3 Second vertical derivative The second vertical derivative technique was used as a two dimensional filter for interpretation of potential field data (Dobrin, 1976). The second derivatives that have been applied in the present study are the second vertical derivative of the vertical component of gravity. If we used the symbol “g” to denote gravity and choose axes so that Z is vertical downward, then the second vertical derivative is the quantity d2g/dz2. The importance of the second derivative for potential field interpretation arises from the fact that the double differentiation with respect to depth tends to emphasize the smaller, shallower geological anomalies at the expense of larger, regional features (Elkins, 1951).

Water resources of arid areas

322

Figure 7. Second vertical derivative of gravity anomaly map. On the practical side, the second vertical derivative has its disadvantages as seen on map (Fig. 7), that shows a number of anomalies of no actual existence, but they tend to be an interpretation of contours rather than of observed gravity field. The main objective of applying the derivative in this research study is for the delineation of shallow faults. It is interesting to prove how the regional is completely eliminated by the second or higher vertical derivative. This may be shown as follows:

The condition is that,

∑nCn=0

Hence,

Where, Z0 is the average regional at the canter of a particular grid, the residuals around a circle of radius, “n” on substituting,

is the average of

Gravity study on groundwater structure in Central Butana (Sudan)

323

Therefore, for the second vertical derivative only the equation is given as

This equation clearly shows that, the regional field constitutes nothing to second vertical derivative value, and thus this derivative represents entirely the residual value only. A profile of second vertical derivative of gravity in a direction outward from the center of a negative anomaly usually shows an outer maximum value minimum value ratio r,

and an inner

Bott (1965). The source of the anomaly may be determined by the

In case of a sedimentary basin r>1 while for a granite batholiths r<1. 3.3 Gravity data interpretation There are two basic approaches to gravity interpretation. One is to determine a plausible mass distribution directly from the gravity data (Qualitative interpretation). The other is to assume various models conforming to all known constraints and to match gravity effects predicted for each model with the gravity field that has actually been observed. The model that gives the best fit is then considered to be the most probable (Quantitative interpretation) one even though it cannot provide a definitive sub-surface picture. Running “GRAVRED” of the “GEOSOFT” package, and both graphical and analytical methods, were used for the processing did processing and interpretation of the gravity data. 3.3.1 Qualitative interpretation Interpretation of gravity, especially in qualitative sense is constrained by a number of inherent limitations and fundamental ambiguities (Dobrin, 1976). To reduce (offset) these ambiguities usually other geophysical methods and/or geological and borehole information should be incorporated. Physically, the Bouguer gravity map represents anomalies from the entire vertical and lateral density variation with the earth and may be used to qualitatively deduce geological structure. As shown in the Bouguer gravity map (Fig. 6) a prominent gravity high occurs in the eastern and the western parts of the surveyed area where basement is shallow or crops out. Small variations in gravity values in this region probably reflect density variations within the shallow basement caused by variations in weathering, especially in the NW and SE portion of the area where schist rocks occur. The northern part of Bouguer map shows a rounded-shaped strong anomaly, trending NW and with a minimum value of approximately −60mGal which is referred to as Wad Burwa.

Water resources of arid areas

324

Figure (6) shows that the gravity lows over central part the south of Jebel-Mundara Bouguer anomaly map has an amplitude of about −50mGal, a rounded-shaped strong anomaly and of general contours strike which indicates that the study area is narrowly oriented in a NW-SE structural domain. This anomaly is connected to another gravity low over J.Qeili, which extends NW beyond the border of the study area. The abrupt change in the trend direction of the anomalies from northwest (J.Mundara and Qeili), to northeast (Wad Burwa), may be due to the existence of a structural path that might have facilitated the emplacement of low-density bodies whose effects are expressed by the low gravity anomalies. Figure (8) shows a fault or lithological contact as exhibited by the second vertical derivative profile. 3.3.2 Quantitative interpretation The quantitative interpretation determines the shape of the mass excess or deficiency, which cases the gravity anomaly measured on the earth surface. The interpretation of the (residual) gravity anomaly in relation to the sub-surface causal features can be approached into two ways: (1) Linear inverse problem. (2) Non-linear inverse problem. The linear inverse problem arises when the shape of the body is specified and the problem is then to determine the distribution of density as a function of 2-D or even 3-D form of the anomaly on or above the datum plane. In practice the linear inverse problem receives less attention and at most qualitative gravity interpretation is concerned with non-linear inverse problem. The non-linear method calls for approximation of the geological bodies, which are considered to be the gravity source, by assuming simple geometric model from which the theoretical gravity effect can be compared with the observed (residual) gravity data and the shape of the body can be changed (modified) to minimize the difference between the observed and the computed gravity effects, often by interactive and/or iterative computer inversion methods (Kearey et al., 1988). 3.3.2.1 Density measurements In qualitative interpretation of gravity anomalies, it is necessary to determine the density of the subsurface rocks before one can postulate the shape or structure of the source body. For this reason

Gravity study on groundwater structure in Central Butana (Sudan)

325

Figure 8. Fault or lithological contact along profile A-A′ as exhibits second derivative. Table 2. Rock densities of the basement complex. No. of Density Density Rock sample (gm/cc) (gm/cc) type range mean Locality Source Gneiss 2 Green 40 Schist Granite 5 Granite 7 Syenite 16 Gabbro 4

2.85– 2.85 2.65– 2.98 2.59– 2.68 2.59– 2.65 2.57– 2.68 2.92– 2.95

2.84

Gadaref Author

2.81

J.Qeili

2.64

J.Qeili

2.62

Butana

2.63

J.Qeili

2.93

Ahmed (1968) Ahmed (1968) Author

Ahmed (1968) Es Sada Author

some attention has been drawn to the densities and density contrast between the representative rocks in the study area. In fact the density contrast between the rocks are the primary cause of the measured gravity effect. Densities of main rocks composing Butana (Igneous complex) had been measured by Ahmed (1968) and range from 2.57 to 2.68gm/cc. A value of 2.65gm/cc is considered as

Water resources of arid areas

326

the average for the whole complex. Densities of the country rocks vary from 2.65 to 2.98gm/cc, and a value of 2.81gm/cc is considered to be the average. Generally densities of igneous rocks, increase with decrease of silica content or in other words, it follows the acidity line regardless of the rock being plutonic or volcanic (Table 2). On the other hand densities of the metamorphic rock increase with the degree of metamorphism. 3.3.2.2 Modeling of the observed anomalies Modeling of the anomalies in this study was performed by “Grav2dc” program, written by Cooper (1991). It uses the Talwani et al. (1959) type a logarithm, to calculate the gravitational anomaly over one or more 2 D/2.5 D bodies. The construction of models due to local structures is therefore a process by which one removes the regional gravity effects. It eliminates the regional completely and thus enhances the residual anomaly. Thus this models entirely the residual value only (Figs 8, and 9). 3.3.2.3 Modeling along profile A-A′ The profile passes across the central part of the area, generally trend in a NW direction and extends to 50km. No Basement outcrops along this profile, however it passes through Butana clay

Figure 9. (a, b) Two dimensional model along profile A-A′. cover. The profile displays a rapid decrease in gravity anomaly in NW direction. These suggest emplacement of low-density body (mass deficiency) into the green schist. To

Gravity study on groundwater structure in Central Butana (Sudan)

327

account for this low (−ve) gravity anomaly a model representing a granitic intrusion into the green schist with a density contrast of −0.16gm/cc has been simulated as shown in Figure (9a). The model (Fig. 9a) shows intrusive granitic mass of a thickness of about 2km into the green schist, and bounded by several step faults, especially at SE side. 3.3.2.4 Modeling along profile B-B′ The profile starts at about 5km southeast of Jebel-Mundara. The profile extends for about 60km and generally oriented in a NE-SW direction (Fig. 9b). The surface geology along this profile consists entirely of the granite at Jebel Mundara. Modeling of the observed gravity anomaly has reveled emplacement of a granitic body, which extend to 2.5km depth and bounded by near-surface step faults. The simulated (modeled) granitic body correlates with the shape of the Jebel Mundara low anomalies, which trend in the same direction of the fracture system in the old Metamorphic rocks i.e. NNE-SSW. 4 CONCLUSION The interpreted gravity data in this research were measured to confirm the existence of low gravity anomalies in the area that is referred to as Wad Burwa anomaly (Ibrahim, 1993) or otherwise. Filtering of the gravity data comprises the second vertical derivatives of the gravity anomaly. The techniques have proven to be effective in revealing local features more clearly than their respective potential fields. The vertical derivative enhanced and resolved the regional-residual anomalies more clearly. on the other hand have delineated density boundaries (lithological boundary). Modeling of the anomalous field was performed by gravity inversion program (Cooper, 1991), which simulate two-dimensional geological model of irregular geometry, mostly representing the mode of occurrence of the granitic into the country rocks of the Butana region. The surface outcrop of granite controls the shape of the uppermost part of the model, while the −ve density contrasts between the granitic rock and the other green schist account for the −ve Gravity lows in the study area, which are generally known to be shallow Basement Complex terrain. This gravity low is largely attributed to the occurrence of low-density rocks (granitic intrusion) into the high-density rocks (green schist) of the Butana region. Thus for it is not unusual gravity measurements in such areas introduced by granitic into country rock of higher densities, to reflect apparently anomalies of typical sedimentary basins. However such ambiguity can be resolved by computation of the second derivative (Bott, 1965). By verifiable of their occurrence and contact with host rock, granitic bodies slope outwards, thus display (−ve) second derivative gradient as shown in section 3.7.2.1 of this thesis. On the other hand the gravity high zones un-ambiguity coincide with the areas of known shallow Basement Complex. As has been revealed by the quantitative interpretation, the granitic bodies are bounded by near-surface (shallow) step faults. The appearance of these faults on the ground surface is completely masked by the Butana clay

Water resources of arid areas

328

soil and probably they can be exposed by intercepting deep cut water courses (Wadi or Khor). In such conditions the faults can channel water, from surface runoff, to facilitate occurrence of groundwater storage in the basement (granitic rocks) in the Butana area. Filed work for ground water trotting and checking of interpreted data is crucial to up grade the quality of decision. Concurrently geophysical and geological investigation could be carried out in quest to acquire more precise data in the study area (Wadi and flood delta, etc.). The study area of the Butana region has a lack of water supply for both population and livestock uses, because Basement Complex, which is outcropping, or of shallow depth, dominates it. However, further more detailed gravity work in addition to the seismic survey may lead to reveal saturated depression or fracture zones. The presence of water supply in these zones can be expected. REFERENCES Abd el Rahman, E.M., Yehia, A.Y. & Amin, Y.A. 1983. Methods of determination of the proper regional gravity from Bouguer anomaly profile. E.G S. Proc. of 2nd Ann. M. Agocs, W.B. 1951. Least squares residual anomaly determination. Geophysics, 16:686–696. Ahmed, F. 1968. The geology of the Jebel Qeili, Butana and Jebel Sileitaat-Es-Sufr igneous complex, Nile valley, Central Sudan. Unpublished M.Sc. thesis, Univ. Khartoum. Ahmed, F. & Ayed, M.A. 1996. Applied geophysical and satellite imagery techniques, for ground water studies in Central Butana area; ADS report, 25pp, 10–15. Bannister, A. & Raymond, S. 1989. Surveying Catalog, Singapore, Longman Scientific and Technical. Bott, M.H.P. 1965. A geophysical study of the granite problem. Quart. Journ. Geol Soc. London, 112(445): 45–62. Cooper Ltd. 1991. Program “Grav 2dc”, written by G.R.J. Dep. Geophysics, Witwatersrand, South Africa. Dobrin, M.B. 1976. Introduction to Geophysical Prospecting. Mc Graw-Hill, New York. El kins, T.A. 1951. The second derivative method of gravity interpretation. Geophysics, 16:29–50. Fajklewicz, Z. 1959. The use of Cracovian Computation in estimating regional gravity. Geophysics, 15:&QJ;465–478. GEOSOFT reference manual, 1989. Software for earth sciences. GEOSOFT INC, Toronto, Canada. Griffin, W.R. 1949. Residual gravity in theory and practice. Geophysics, 14:39–56. Ibrahim, A.E. 1993. Interpretation of gravity and magnetic data from the Central Africa rift system, Sudan. Unpublished. Ph.D. Thesis Univ. Leeds, 209pp. Ibrahim, A.E., Ebinger, C.J. & Fairhead, J.D. 1996. Lithospheric extension NW of the Central Africa Shear Zone (CASZ) in Sudan from potential field studies. Tectonophysics, 255:70–97. Iskander, W., Ahmed, A.A., Mokhtar, A. & Fadle, A.S. 1993. Appraisal of mineral and water resources of central Butana, Eastern region-Sudan ADS report 85pp. Kearey, P. & Brooks, M. 1988. An Introduction to Geophysical Exploration. Dep. Geol. Univ. Bristol., 296 pp, ch-6, 138–169. Nettelton, L.L. 1976. Gravity and magnetic in Oil exploration. Mc Graw Hill, New York, 464p, 138–169. Paul, M.K. 1967. A method of computing residual anomalies from Bouguer gravity map by applying relaxation technique. Geophysics, 32:708–719. Seigal, H.O. 1995. High precision gravity guides. Canada, Ontario, L4K 1B5:120pp. Sun Oil Company, 1984. Nile blocks gravity survey. Final report, Unpublished.

Gravity study on groundwater structure in Central Butana (Sudan)

329

Talwani, M.J., Worzel, L. & Landisman, M. 1959. Rapid gravity computations for Two-dimension Bodies with application to the Mendocio submarine fracture zone. J. Geophys. Res., 64:49–59.

Remote sensing and electrical resistivity studies on groundwater structure zones in Central Butana (Sudan) K.M.Kheiralla TU Bergakademie Freiberg, German, Gustav-Zeuner-Str, Freiberg, Germany A.E.Ibrahim El Neelain University, Sudan Water Resources of Arid Areas—Stephenson, Shemang & Chaoka (eds), © 2004 Taylor & Francis Group, London, ISBN 04 1535 913 9

ABSTRACT: Remote sensing techniques combined with the resistivity electrical methods adopted for locating the potential groundwater zones in central Butana. Land sat image is helpful in the location of major deepseated fracture zones. The lineament patterns derived from TM image show strong NE-SW orientations and drainage pattern present in the area. The longer NNE to NE trending feature may be important from a regional hydro-geological point of view, where as the NW trending features are significant in that they intersection the major fault. Groundwater occurrence is mainly due to the secondary porosity, such as weathering, joints, fissures and fracture. Good quantity of groundwater potentials has been identified in the high density of drainage and lineament zone in Butana region, lineaments intersections are important with well yield than are individual. Moderate to good yield of groundwater are tapping from weathered zones and good yield are tapping from fracture zones.

1 INTRODUCTION Groundwater has become an important source of water and has played an importance role in developing industry, agriculture, livestock and domestic purpose. The groundwater condition in a crystalline rock terrain is multivariate because of the heterogeneity of the aquifer, due to the varying composition, compaction, and degree of weathering and density of fracturing. As a result, exploration of groundwater in a crystalline rock terrain has proved to be a complex phenomenon. However, the presence of a vast crystalline rock terrain cannot be neglected as an unfavorable zone. The application of Remote sensing techniques and surface geo-electrical methods is highly helpful for groundwater exploration lies in delineating potential zones of groundwater from a large area. Generally the occurrence of groundwater in crystalline

Remote sensing and electrical resistivity studies on groundwater structure zones

331

rock terrain is associated with the geological structural features like lineaments, fractures/fissure, fault zones. The study area lies between the Blue Nile River and River Atbara, It is bounded by latitudes 14° 30′N and 16°00′N and longitudes 33°30′E and 35°30′E. The study area covers about 20,000km2 (Fig. 1). The study area is generally flat, with a gentle slope to the North. The general altitude of the plain is about 500m above mean sea level (m.s.l). The flat monotony of the plain is occasionally broken by some protruding low to moderately high hills or hill chains, which hardly exceed 200 meters above the ground surface (Kheiralla, 2001). According to the 1993 census in Sudan the total population of the study area is roughly estimated at 30,000 persons, Livestock rising is the major activity of 70% of the Butana population. The ecological conditions as well as the long experience of the inhabitants turn pastoralism as the most worthwhile occupation. The animal population within the area is roughly estimated as some 35,000 heads, composed of about 30,000 sheep and goats, 5,000 cattle and camel (Abd el Ati, 1993). Human and animal populations in the Butana area receive their water

Figure 1. Location map of Central Butana. supply from surface and groundwater. However Groundwater is the only permanent source of water supply in the Butana area. Different than elsewhere in Sudan, the weathered and/or fractured Basement Complex are the main sources of groundwater in

Water resources of arid areas

332

Central Butana region, as they can store and yield reasonable quantities of water from their joints and fractures. The alluvial deposits are none water bearing formation, especially close to the blind deltas of the main Wadis (valley) (Ahmed & Ayed, 1996). However, groundwater occurs in sand alluvial fans are expected to be found under thick cover of the Butana clayey. The surface runoff soon evaporates and/or infiltrates. The crystalline rock terrines, underlying the Butana area is assumed to be groundwater devoid. Under specific geological and hydrogeological condition, the crystalline rock terrines can store and yield appreciable quantities of groundwater variable quality. The weathered and/or fractured crystalline rocks underlying the drainage system from local aquifer zones, in the Butana area the hydro-geological significance of the sandy alluvial deposits is that they act as a membrane through which surface flow can infiltrate to recharge the underlying fractured zones. The study aims to investigate the extent of the influence of the drainage and fractured by the use of Remote sensing and applying geo-electrical methods to delineate the general hydro-geological aspects of the sediments overlying the crystalline rocks for exploration of groundwater. In order to understand the significance of the fracture pattern, geological, hydro-geological, drainage system and lineaments map have been prepared with the help of Land sat TM imagery. An isoresistvity map is prepared by conventional survey of equal apparent resistivity (AB/2=60) and then comparing a lineament map and drainage system map to identify the extent of correlation. 2 GEOLOGICAL SETTING AND HYDROGEOLOGY A number of isolated out crops of Pre-Cambrian Basement complex rocks scatter over the central plains of Sudan (Fig. 2). In the Butana region, several hills occur and make prominent landmarks

Remote sensing and electrical resistivity studies on groundwater structure zones

333

Figure 2. Geological map of Central Butana Area. Table 1. Geological column for Butana region. (After Iskander et al., 1993) Formation

Age

Superficial deposits Tertiary volcanic El Burg Basal conglomerates The Basement complex

Quaternary/Recent Tertiary Jurassic to cretaceous Pre-Cambrian

within the predominantly flat clay plains. The solid geology of the Butana plain is rather concealed under a veneer of variable thickness of superficial clays covering over 70% of the Butana area. Several metasedimentary sub-parallel belts extend for 10–25km in the form of low to moderately elevated ridges surrounded by Butana clay cover (Iskander et al., 1993). The generalized geological column for Butana region can be summarized as shown in Table 1. The majority of the Basement rocks are concealed under the cover of the Butana clay plain. As mentioned, the structural domain in the area is characterized by northeast

Water resources of arid areas

334

trending lineaments. Exposed structures in the metasediments display complex shearing/faulting and tight folding with dipping axial planes where the axes generally have NE-SW trends and with SW mergence. Tectonic activity and the associating metamorphism have resulted in variable folding, faulting and shearing giving rise to complex contact relations between the different basement units and within the metasedimentary assemblages.

Figure 3. Map showing lineaments and rose diagram of Central Butana Area. The structural domain in the area is characterized by northeast trending lineaments (Fig. 3). Exposed structures in the metasediments display complex shearing/faulting and tight folding with dipping axial planes where the axes generally have NE-SW trends and with SW vergence. Iskander et al. (1993) Have interpreted that Riera, Es Subagh and Wad Gidair occupy the limbs of a synform whose vergence is to words SSW with a general N trending axis, while a major anticline occupies the area between Es Subagh in the NW limb and Hosheib-Suruj Jebel El Tawill in the SE limb with a NNE-trending axis and a

Remote sensing and electrical resistivity studies on groundwater structure zones

335

NE vergence. In this search three phases of deformation recognized has reported in the Butana region (Fig. 2) these are: ● F1—Regional foliation, lineation, and shear zones. ● F2—Planar and linear structures, upright folds and faults bricca. ● E-W faulting and fracture cleavage. The tectonic events that terminated by continental collision at the end of the Pan-African developed or reactivated a conjugate set of strike-slip faults and shear zones in both the Nubian and Arabian shield. The fault/shear zones have two main trends NW-SE (Najd trend) and the ENE-WSW (Central African Lineament). In the central Butana regions, major faults and shear

Table 2. Sites proposed for drilling of boreholes. Well Locality S.W.L Expected Aquifer No. (m) yield type (103m3/yr)

Apparent resistivity at depth of 60m (Ωm)

1

Es 45 90 FBC 125 Subagh 2 Es 39 80 FBC 90 Subagh (El Buqaa) 3 Es Sada 25 50 ALL/WBC 60 4 Abu 27 40 WBC 55 Gerad 5 El Fuel 36 70 FBC 115 6 El Bresi 33 65 ALL/WBC 70 7 El Edeid 32 50 FBC 85 El Tawill 8 El Edeid 24 60 FBC 100 El Hamur FBC—Fractured Basement Complex; ALL—Alluvium; WBC—Weathered Basement Complex.

zones display apparently complicated sets of fractures generally follow the regional foliation/ schistosity trends in both the metasediments and the underlying Basement rocks. The structural domain in the north and southeast of Es Subagh area is NNE to NE (Fig. 3). Subordinate N-W and E-W trending faults affect mainly the northwestern part of El Butana (Iskander et al., 1993). The prominent NE trending sets of faults are mainly strike-slip with a dextral sense of movement. Those discontinuities extend up to 70km as attested by the linear-controlled drainage system. Some faults sheared the metasedimentary rock assemblage to significant proportions creating 2–1.5m wide zone

Water resources of arid areas

336

of shearing or brecciation (e.g. Jebel El Rabbda), thrusting the evident in Jebel El Tawill with NE trends and E or SE dip direction. This orientation lineament direction of ophiolite transport. Very similar trends have been suggested for the ophiolite and allochthonous sheets transport in the Engassana Hills (Vail and Duggue, 1986). This deformation created an important fracture, which have provided Es Subagh area in Central Butana with fresh water. To the south, J.El Tawill ultramafic belt has been affected by two boundary thrusts (Fig. 2) 2km apart, resulting into variable degrees of shearing on the eastern and western flanks. Similar sub-basin has been created and provides potable water for the inhabitants of the area. The area has complex hydro-geological conditions owing to complexity in the origin of the rock units encountered. The compact pre-Cambrian suites of rocks are poor aquifers. Groundwater occurs in confined conditions in these rocks due to the development of secondary porosity such as fracturing. Fractured crystalline rocks are less permeable at greater depth because stress variations that cause fractures are larger and, over geological time, occur more frequency near the ground surface. Fractures tend to close at depth because of vertical and lateral stresses imposed by overburden loads and horizontal stresses of tectonic origin. Apparently these basic conditions, which control ground water occurrence in crystalline rocks, apply to a large extent to the Butana area. Groundwater generally occurs in the upper weathered/or fractured zone, which may extend down to 70m depth as indicated by the lithological logs of Es Subagh two boreholes. The sheared rocks, which form the bulk of the aquifer, are composed of acid gneisses, quartzite, marbles and granites. The brecciate marbles in the Butana normally form the good aquifers. The formation easily dissolvable by moving waters, and thus forms wide fissures and cavities, which facilitated groundwater storage and transmission (Iskander et al., 1993). The best aquifer zones comprise the marble cavities along Khor Abu Gimbil, Es Sufeiya, fractured Basement rocks of Adeid El Tawill, Es Subagh, Reira, and El Hagar, and the alluvial deposits along Wadi Abu Grad and Wadi Abu Matariq. Depth to ground water level in the study area varies from 20 to 25m at Reira, between 15 to 50m at Es Subagh, from 20 to 40m below the ground surface at Es Sufeiya, 10 to 20m at El Tawill, and much deeper ground water level (60m below the ground surface) at El Hagar (Table 2). Groundwater in the fracture Basement Complex occurs under free water table conditions and at depths ranging from few meters to over 60 meters below the ground surface. The water table attains its highest level during and shortly after rainy season and drops to its lowest level immediately before the next rainy season. The amplitude of the seasonal or annual fluctuation depends to a great extent on the balance between the recharge and discharge of water to and from the aquifer. Generally productivities of the wells directly correlate the intensity of rainfall in the study area. Noticeably in dry years, the water levels drops to the bottom of the wells or completely dry out. Apparently Groundwater movement in crystalline rocks is non-committal to flow direction, because fracture trends tend to variation that cause movement are largely expected variable to flow direction. Basement outcrops from the main catchments area and ground water movement coincides with the drainage systems controlled structurally. Groundwater moves away from the surface water divide and generally in the eastern and NE direction, ground water moves to joint Atbara River. Eight boreholes have been drilled along watercourses or fractured aquifers at variable depths (Table 2). These include four at Es Subagh, Qeili, Abu Gimbil, Husheib and Umm

Remote sensing and electrical resistivity studies on groundwater structure zones

337

Sarha villages. Only five of the eight boreholes are successful with the total out put hardly exceeding 10m3 per hour, per well. Presently only a few are in use. Two hundreds and fifteen hand-dug wells were excavated into different aquifers (e.g. Fractured, weathered Basement, alluvial or fan deposits) to depths ranging between 20–55m (Kheiralla, 2001). Their static water level varies between 15 and 40m, though wells drilled into the alluvium aquifers are relatively shallower not exceeding about 5m deep. The static water level in these hand-dug wells fluctuates between 3–5m per year, which may be considered as a good sign of recharge. The total out put of the hand-dug wells is estimated to be about 2869*102m3/year. Chemical analysis of samples collected from some hand-dug wells indicates fair to good water quality suitable for human as well as animal consumption. Based on the above estimates, the grand total yield of the existing water sources in the Central Butana is around 33*104m3/year. This amount does not exceed 2/5th(40%) of the actual demand. These indicate an actuate water supply shortage. 3 LINEAMENTS One of the objectives of this work is to delineate the lineaments in more details using the geoelectrical method and Remote sensing technique. Waters (1990) suggests that there are two stages involved in hydro-geological investigat