This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
a (mW m2) NQc
a Ab (mW m3) NAc References
Archean Dharwar (India) Kaapvaal basementd (S. Africa) Zimbabwe (S. Africa) Yilgarn (Australia) Superior (N. America) Slave (N. America) Wyoming (N. America)
36 2.1 44 47 3.5 39 1.5 41 0.9 50 3.5 48.3 5.7
8 81 10 23 70 3 6
1.8 1.34 3.3 0.72 2.3 3.1
Total Archeane
41 0.8
188
Proterozoic Aravalli (India) Namaqua (S. Africa) Gawler (Australia) Sao Francisco craton (Brazil) Braziliane mobile belt (Brazil) Ukrainian Shield Trans-Hudson (N. America) Wopmay (N. America) Grenville (N. America)
68 4.9 61 2.5 94 3 42 5 55 5 36 2.4 42 2.0 90 1.0 41 2.0
7 20 6 3 8 12 49 12 30
48 0.8
675
57 1.5 49 4.4 30 2 58.3 0.5
79 2.6 6 1.3 40 2213
Total Proterozoic
e
Paleozoic Appalachians (N. America) Basement United Kingdon Urals Total Paleozoice
— — — 0.73 1.0 2.1
Roy and Rao (2000) Ballard et al. (1987), Jones (1988) — Jones (1987) 540 Cull (1991), Jaupart and Mareschal (2003) 64 Mareschal et al. (2000) 20 Mareschal et al. (2004) 6 Decker et al. (1980) Nyblade and Pollack (1993)
2.3 3.6 1.5 1.7 0.9 0.73 4.8 0.80
— 0.6 1.2 0.2 0.50 1.0 f
10 90 3 5 7 47 20 17
Roy and Rao (2000) Jones (1987) Cull (1991), Jaupart and Mareschal (2003) Vitorello et al. (1980) Vitorello et al. (1980) Kutas (1984) Rolandone et al. (2002) Lewis et al. (2003) Mareschal et al. (2000) Nyblade and Pollack (1993)
1.9 0.5
50 Jaupart and Mareschal (1999) 6 Lee et al. (1987) Kukkonen et al. (1997) Pollack et al. (1993)
a
Mean one standard error. Standard deviation on the distribution. c Number of sites. d After removing the contribution of the sediments. e Total in the compilation by Nyblade and Pollack (1993) excluding the more recent measurements included here. f Area-weighted average value. b
Table 4 cratons
Regional variations of the heat flux in different
Minimum
Maximum 2
(mW m ) Superior Province Trans-Hudson Orogen Australia Baltic Shield Siberian Shield
22 22 34 15 18
48 50 54 39 46
Minimum and maximum values obtained by averaging over 200 km 200 km windows.
competing mechanisms of crustal extraction from the mantle and crustal recycling. In the Proterozoic provinces, high heat flux and crustal heat production (e.g., Wopmay Orogen, Thompson Belt in the THO, Gawler
Craton in Australia) are always associated with recycled (Archean) crust. By contrast, juvenile Proterozoic crust is characterized by low heat flux (e.g., all the juvenile belts of the THO, the Proterozoic rocks of the Kola peninsula). Within each province, there is regional (on a scale on the order of 400 km) variability which must be considered when calculating geotherms. This is illustrated in Table 4 which shows that the regional variability is high within all provinces. Thus, there is no geotherm characteristic of a single province. To illustrate this point, we have calculated geotherms from selected regions from stable North America. The selection covers the extreme regimes within each age group. The crustal models are described in Table 5 and the variations in thermal conductivity are described in Appendix 4. Geotherms shown in Figure 17 illustrate two points: (1) there is no direct relation between the geotherm and the age as the
236
Heat Flow and Thermal Structure of the Lithosphere
Table 5
Heat flux crustal heat production used for geotherm calculations
Region
Q0
A1
H1
A2
H2
Archean E. Abitibi (1) W. Abitibi (1) Slave (2)
29 45 50
0.4 1.2 1.7
40 20 10
0.4 1.2
20 10
Proterozoic Labrador (3) THO (Flin Flon Belt) Wopmay (4) Grenville (1)
22 40 90 40
0.2 0.3 4.8 0.7
40 8 10 40
1.2 1.0
Paleozoic Appalachians (1)
58
3.1
8
1.1
A3
H3
Moho depth
Tm
0.4
20
40 40 40
325 422 428
12 10
0.25 0.4
20 20
40 40 40 40
287 434 705 440
10
0.4
22
40
426
References: (1) Mareschal et al. (2000); (2) Mareschal et al. (2004); (3) Mareschal et al. (2000); (4) Lewis et al. (2003).
profiles from different age groups overlap; (2) there is also no simple relationship between the surface heat flux and the temperature profiles. For example, there may be no difference in mantle temperature between the Grenville and the Appalachians in spite of the higher surface heat flux in the Appalachians than in the Grenville (58 vs 41 mW m2). The high heat flux in the Appalachians is mostly accounted for by high heat production (3 mW m3) at shallow depth resulting in a differentiation index DI ¼ 2.5. In contrast, the Grenville whose crust is made up of stacked slices from all levels appears to be more homogeneous at crustal scale with DI ¼ 1. Thus, the vertical differentiation of the radioelements must be understood in order to estimate Moho and mantle temperatures. On the scale of the whole North American continent, the average heat production and the crustal differentiation index are positively correlated (Perry et al., 2006). Thus, regions with large heat production (and hence high surface heat flux) are systematically associated with an enriched upper crust. In most cases, this is due to highly radiogenic granites which do not extend very deep, as in the Appalachians province for example. All else being equal, temperatures decrease with increasing differentiation index and increase with increasing heat production. With the correlation between heat production and differentiation index (Figure 14), variations of crustal temperatures are much smaller than for a single universal model for the vertical distribution of radioelements. From the standpoint of large-scale geophysical models, the relations between surface heat flux, Moho heat flux, and Moho temperature are nonlinear and cannot be reduced to simple correlations.
6.05.4.5
Variations of Crustal Thickness
Independent evidence for significant horizontal variations of crustal temperatures is provided by the topography of the Moho discontinuity. The characteristic time of relaxation for topography on an interface between two layers with a density difference is (Chandrasekhar, 1961)
8 eff g
½26
where g is the acceleration of gravity, meff an effective viscosity, and is the wavelength of the interface topography. The value above yields only an order of magnitude because the relaxation time is modulated by a function depending on the geometry and the boundary conditions (Chandrasekhar, 1961). For representative crustal rheologies, temperature differences of 100–200 K that are predicted imply that the effective viscosity varies by up to three orders of magnitude. Thus, relaxation of tectonic deformation proceeds at different rates depending on the crustal heat production. Higher crustal heat production and heat flow during the Archean might thus explain the observation that the Archean Moho is flat, even in regions that have experienced compression. However, a few areas that were deformed during the Proterozoic have preserved thick crustal roots: the Kapuskasing uplift in the Superior Province, the Lynn Lake area in the THO, and the eastern part of the Grenville Front. Thick crust with the same bulk crustal composition than elsewhere would lead to high temperatures at the Moho and in the underlying mantle, which would allow flow in the lower crust and relaxation of the crustal root. The persistence of
Heat Flow and Thermal Structure of the Lithosphere
0
(b)
0
25
25
50
50
75
75
Depth (km)
Depth (km)
(a)
100 125
237
100 125
Slave
Voisey Bay
150
150 West Abitibi
175
Flin Flon (THO) 175
East Abitibi
200
Wopmay
200 0
200 400 600 800 1000 1200 1400 1600 Temperature (°C)
(c)
0
200 400 600 800 1000 1200 1400 1600 Temperature (°C)
0 25 50
Depth (km)
75 100 125 Grenville 150 175
Appalachians
200 0
200 400 600 800 1000 1200 1400 1600 Temperature (°C)
Figure 17 Geotherms for different regions in the Canadian Shield: (a) Eastern Abitibi, western Abitibi, and Slave Province (Archean); (b) Voisey Bay, Flin-Flon Belt, and Wopmay Orogen (Proterozoic), (c) Grenville (Mid-Proterozoic), and Appalachians (Paleozoic). Thick lines are the geotherm calculated for the crustal models in Table 6. Thin dotted lines are for the same models with 2 mW m2 change in Moho heat flow.
crustal roots demonstrates that temperatures have remained low for a very long time (a minimum of 1.8 Gy at both Lynn Lake and Kapuskasing and 1.0 Gy in the Grenville), which can be only
explained by anomalously low crustal heat production. Mareschal et al. (2005) have indeed noted that heat flux was anomalously low in these two areas of the Canadian Shield.
238
Heat Flow and Thermal Structure of the Lithosphere
6.05.4.6
Summary
Continental heat flow is sensitive to the local geology and crustal structure and hence must be used with precaution for studies at the lithospheric scale. In stable continents, high heat flux is always associated with high heat production and an enriched upper crustal layer. Thus, one cannot build geotherms with the same function for the vertical distribution of heat production regardless of the local geological context. Heat flux data alone are not sufficient and must be supplemented by additional information on crustal heat production or mantle heat flux. On a large-scale, three key control variables on lithospheric temperatures are correlated: the average surface heat flux, the average crustal heat production, and the vertical variation of heat production. In contrast, variations in the basal heat flux are small ( 3 mW m2). Steady-state thermal models are only valid if heat flux is less than about 90 mW m2. Higher values imply melting in the crust or weak lithospheric mantle that can deform easily, suggesting that other heat transport mechanisms are effective. In a thick lithosphere, long-term thermal transients are inevitable.
6.05.5 Continental Lithosphere in Transient Thermal Conditions 6.05.5.1
General Features
In tectonically active regions, advection of heat usually dominates over conduction and temperatures are strongly time dependent. Thermal evolution models depend very much on the choice of the boundary conditions at the base of the lithosphere and cannot be assessed against heat flux data for several reasons. One difficulty comes from the variable quality and density of heat flow data in active regions. In the western US, the numerous heat flux data from the Basin and Range and Rio Grande Rift are very noisy because of hydrological perturbations (Lachenbruch and Sass, 1978), a situation reminiscent of young sea floor. Furthermore, the inclusion in the data set of measurements made for geothermal energy exploration has introduced a strong bias towards excessively high values. Far from thermal steady state, one may not use heat flux data to estimate lithospheric temperatures by downward extrapolation of shallow heat flux measurements. In a continent that is being deformed, heat flow and temperatures depend on the competing effects of
crustal thickness changes, which imply changes of crustal heat production, and deformation, which affect the temperature distribution. Thus, erosion or crustal extension initially cause steeper geotherms and enhanced heat flux. After these transient effects decay, the reduced crustal thickness leads to a lower heat flux than initial. Conversely, crustal thickening causes the geothermal gradient and the heat flux to decrease at first and then to increase due to higher crustal heat production. In many cases, heat flux also records shallow processes such as the cooling of recently emplaced plutons. Because crustal composition is often affected by syn or postorogenic magmatism, there is no general rule to predict the final crustal thickness, composition, and heatproduction distribution. Following the cessation of tectonic and magmatic activity, one must distinguish between two types of transients. Crustal temperatures return to equilibrium with local heat sources in less than 100 My. This is followed by a much slower transient associated with re-equilibration of the lithospheric mantle. For thick lithosphere, such transients may last as long as 500 My (Nyblade and Pollack, 1993; Hamdani et al., 1991; Kaminski and Jaupart, 2000) and result in negative or positive heat flow anomalies. Such slow thermal relaxation has two important features. First, it involves deep thermal anomalies whose lateral variations are efficiently smoothed out by heat conduction and which do not lead to spatial variations of surface heat flux over distances <500 km. Second, it is linked to changes of thermal boundary layer thickness which may be detectable by other methods (Jaupart et al., 1998).
6.05.5.2
Compressional Orogens
Unless the crust has anomalous composition, the total radiogenic heat production increases with crustal thickness. Steady-state conditions have not been reached in young orogens where heat flow is also enhanced by erosion. High heat flux values have been measured in Tibet and parts of the Alps (Jaupart et al., 1985). These values imply high temperatures in the shallow crust. Because these variations are of short wavelengths, they have been attributed to the cooling of shallow plutons. One should note that crustal melting and emplacement of granite intrusions in the upper crust modify the vertical distribution of radioelements. Thus, one should not use the same heat-production model
Heat Flow and Thermal Structure of the Lithosphere
before and after orogenesis (Sandiford and McLaren, 2002; Mareschal and Jaupart, 2005). 6.05.5.3
Rifts and Zones of Extension
Crustal extension and lithospheric thinning will instantly result in a steeper temperature gradient and an increase in heat flux. Thermal conduction cannot account for the rapid thinning of the lithosphere and mechanical processes such as delamination or diapiric uprise of the asthenosphere are necessary to account for the rapid development of extension zones (Mareschal, 1983). The thermal effects of extension are transient and after return to equilibrium, the heat flux at the surface of thinned crust will reflect the smaller amount of radioactive elements and hence will be lower than before extension. Further changes of crustal heat production may occur due to the injection of basaltic melts, which are depleted in radioelements with respect to average continental crust. This explains, for example, why heat flux is slightly lower in the 1 Gy Keweenawan rift than in the surrounding Superior Province (Perry et al., 2004). In the Basin and Range Province of the southwestern US (Sass et al., 1994; Morgan, 1983), high heat flux values (110 mW m2) are consistent with an extension rate of 100% (Lachenbruch and Sass, 1978; Lachenbruch et al., 1994). The high temperatures that are implied must cause thermal expansion of the crust and parts of the mantle and hence should lead to an elevated topography. On the other hand, crustal thinning has the opposite effect. The elevation of the Basin and Range Province cannot be accounted for only by the extension. The calculated thermal expansion in the lithospheric mantle is not sufficient to account for the high elevation. According to Lachenbruch et al. (1994), the mantle lithosphere beneath the Basin and Range has been delaminated and not simply stretched. A striking feature of the zones of extension is that the transition between the region of elevated heat flux and the surrounding is as sharp as the sampling allows to determine. This is observed across the boundaries of the Colorado Plateau and the Basin and Range in North America (Bodell and Chapman, 1982), between the East African Rift and the Tanzanian craton (Nyblade, 1997), or between the Baikal Rift and the Siberian craton (Poort and Klerkx, 2004). Where the sampling is sufficient, heat flux exhibits short-wavelength variations. These variations are probably due to shallow
239
magmatic intrusions, a hypothesis well justified by the numerous volcanic edifices that dot such areas; they also may reflect groundwater movement (Poort and Klerkx, 2004). For studies of lithospheric structure, one must separate between a high background heat flux due to extension and local anomalies reflecting shallow magmatic heat input, which requires measurements at close spacings. 6.05.5.4 Thermal Relaxation of Thick Continental Lithosphere Once active deformation and magmatism have ceased, the return to thermal equilibrium takes a long time. The thermal relaxation time depends on the thermal structure at the end of activity, the lateral extent of the perturbed region, and the boundary condition at the base of the lithosphere. 6.05.5.4.1
Sedimentary basins The subsidence of sedimentary basins and passive continental margins provides a good record of the relaxation of thermal perturbations in the lithosphere and is sensitive to lithosphere thickness. Such transients have been recorded in intracratonic basins located away from active plate boundaries and have generated a lot of interest (Haxby et al., 1976; Nunn and Sleep, 1984; Ahern and Mrkvicka, 1984). Subsidence is also affected by tectonic, metamorphic, and eustatic effects. In order to identify these effects, some authors have assumed that the continental lithosphere has a well-defined characteristic cooling time of 60 My (Bond and Kominz, 1991) and that subsidence phases that are significantly longer than this require other causes than thermal effects, such as renewed extension for example. These assumptions are not justified for thick continental lithosphere with long thermal relaxation time. Theoretical subsidence models that have been developed differ by their initial conditions and their basal boundary conditions (McKenzie, 1978; Hamdani et al., 1991, 1994). Although this has not been sufficiently emphasized, the latter is the important factor determining the duration of thermal subsidence. The duration of the subsidence episode varies by a factor of three between various intracratonic basins of North America. For a fixed temperature at the base of the lithosphere, theory would imply that the continental lithosphere thickness is about 115 km and 270 km beneath the Michigan and Williston basins, respectively (Haxby et al., 1976; Ahern and Mrkvicka, 1984). For two
240
Heat Flow and Thermal Structure of the Lithosphere
basins of similar age located on the Precambrian basement of the same continent, such a large difference is surprising. This motivated Hamdani et al. (1994) to investigate the influence of thermal boundary conditions at the base of the lithosphere. They showed that subsidence is slower for a fixed flux than for a fixed temperature and attributed the different subsidence behaviors to different thermal processes at the base of the lithosphere. However, these arguments rely on 1-D thermal models which have recently been questioned (Kaminski and Jaupart, 2000). According to Haxby et al. (1976), for example, the initial perturbation beneath the Michigan basin has a radius of about 120 km, which is less than the thickness of the North American lithosphere. In this case, the assumption of purely vertical heat transfer is not tenable. Accounting for horizontal heat transfer, the solution may be cast in the form of a relationship between the width of the thermal perturbation and lithosphere thickness. No solution can be found for lithosphere thicknesses less than 170 km and the observations are best-fitted for a model with fixed heat flux basal boundary condition (Kaminski and Jaupart, 2000). 6.05.5.4.2 Tectonic and magmatic perturbations
The large relaxation time of thick continental lithosphere might lead one to conclude that all thermal perturbations decay slowly and leave a heat flow anomaly for a long time. In some cases where the thermal perturbation is narrow, a large thickness may in a sense be self-defeating as it enhances lateral heat transfer. Thus, thermal relaxation of some tectonic or magmatic perturbations may in fact be more sensitive to width than to thickness. Gaudemer et al. (1988) and Huerta et al. (1998) have shown that temperatures in orogenic belts depend on belt width and on local values of heat production and thermal conductivity. One consequence is that (P, T, t) metamorphic paths may record belt width as well as other characteristics. Another striking example is provided by flood basalt provinces where large volumes of magma rose through the lithosphere. For a laterally extensive thermal perturbation, one should detect a relict thermal signal for more than 200 My. There is no heat flow anomaly over the Deccan Traps, India, which erupted about 65 My (Roy and Rao, 2000). The same is true over the Parana basin in Brazil which saw the emplacement of large magma volumes 120 My ago (Hurter and Pollack, 1996). It seems that, in both
cases, eruptive fissures are localized in relatively small areas, suggesting that the zone affected by magma ascent may be a few 100 km in width. In this case, thermal perturbations decay rapidly by horizontal heat transport. One consequence is that lithospheric seismic velocity anomalies that are associated with large magmatic events cannot be accounted for by thermal effects and hence reflect compositional variations. In several cases, it seems that the lithosphere has been modified over a large depth interval. For example, a pronounced lowvelocity anomaly of narrow width (120 km) extends through the whole mantle part of the lithosphere beneath the south central Saskatchewan kimberlite field in the THO, Canada (Bank et al., 1998). Similar anomalies have been found beneath the Monteregian-White Mountain-New England hotspot track in northeastern America or beneath the Bushveld intrusion in South Africa (Rondenay et al., 2000; James et al., 2001). 6.05.5.5
Long-Term Transients
The large thickness of continental lithosphere implies very large thermal relaxation times with some interesting consequences. 6.05.5.5.1
Archean conditions The Archean era saw the stabilization of large cratons and the emergence of geological processes that are still active today. In the Archean, crustal metamorphism was biased towards high-temperature–low-pressure conditions in contrast to more recent analogs, indicating that crustal temperatures were higher than today. In apparent contradiction, cratons achieved stability because they had strong lithospheric roots, indicating that temperatures in the lithospheric mantle were not much hotter than today. Proposed mechanisms of formation of lithospheric roots involve either stacking of subducted slabs (Helmstaedt and Schulze, 1989; Abbott, 1991), or melting of mantle over hot spots (Griffin et al., 2003). These mechanisms result in different initial thermal conditions and evolution for the stabilized lithosphere. In the Archean, heat production in the Earth was double the present, which might suggest higher temperatures in the crust and in the mantle as well as higher heat flux at the base of young continental lithosphere. On average, however, the Archean crust of today is associated with less-heat-producing elements than its modern analogs. When corrected
Heat Flow and Thermal Structure of the Lithosphere
for age, the total amount of crustal heat production in Archean times was close to that presently observed in Paleozoic provinces. Save for a few anomalous regions with high radioactivity, crustal heat production in the Archean is thus not sufficient to account for crustal temperatures that are higher than those of modern equivalents. The origin of the high-temperature–low-pressure metamorphic conditions must thus be sought in other mechanisms, perhaps widespread magmatic perturbations. Crustal radioactivity heats the crust in a geologically short time, but a much longer time is required to heat up the lower lithosphere. In Archean times, continental lithosphere was never very old and its thermal structure remained sensitive to initial conditions, that is, conditions which led to the extraction of continental material from the mantle and to the stabilization of thick roots. If the lithospheric mantle is formed by the under-thrusting of subducted slabs beneath the crust, it will initially be colder than in steady state. Mareschal and Jaupart (2005) have estimated the time needed for time-dependent crustal radioactivity to heat up the entire lithosphere. When the half-life of crustal radioactivity is of the same order as the thermal time of the lithosphere, lithospheric temperatures cannot adjust to the timedependent radiogenic heat production. Following isolation of a continental root from the convecting
mantle, the ‘radiogenic’ temperature component at the base of the lithosphere reaches a maximum after 1–2 Gy, depending on lithospheric thickness (Figure 18). The peak temperature is 70% of what one would infer from steady-state models with values of heat production at the time of root stabilization. Thus, temperatures in the crust and deep in the continental root are effectively decoupled for a long time. If the root forms with its initial temperature below steady state, the mantle temperature will always be below steady state for the crustal production. Depending on the mechanism of root formation, the lithospheric mantle could well remain sufficiently cold and strong to preserve Archean features (van der Velden et al., 2005). 6.05.5.5.2 Secular cooling in the lithosphere
In thick continental lithosphere, the timescale for diffusive heat transport is comparable to the halflives of uranium, thorium, and potassium, implying that temperatures are not in equilibrium with the instantaneous rate of radiogenic heat generation. The lithospheric mantle undergoes secular cooling even when thermal conditions at the base of the lithosphere remain steady. The magnitude of transient effects depends on mantle heat production as well
1
λτ = 0.22 λτ = 0.35 λτ = 0.50
0.9 t = ~0.9Gy t = ~1.3 Gy
0.8
241
T/(H h2/2 K)
0.7 0.6 t = ~1.8 Gy 0.5 0.4 0.3 0.2 0.1 0
0
0.2
0.4
0.6
0.8
1
λt Figure 18 Temperature at the base of the lithospheric root after its stabilization beneath the crust. The temperature is scaled to the maximum temperature increase due to crustal heat production at the time of stabilization. is the average decay constant of the radioelements (corresponding to a half-life of about 2.5 Gy). is the thermal relaxation time of the root. The chosen values of correspond to root thickness of 180–250 km.
242
Heat Flow and Thermal Structure of the Lithosphere
as on lithosphere thickness. Even large values of heat production do not introduce large transients in a shallow lithosphere. Conversely, even small values of heat production lead to significant transient effects in a thick lithosphere. In lithosphere that is thicker than 200 km, the geotherm is transient and sensitive to past heat generation. For the same parameters values, and in particular for the same values of present heat production, the deeper part of the temperature profile diverges from a steady-state calculation because of the long time to transport heat to the upper boundary. Depending on the amount of radioelements in the lithospheric mantle, the vertical temperature profile may exhibit significant curvature and may be hotter than a steady-state profile by as much as 150 K (Figure 19). For typical values of heat production in the lithospheric mantle, this secular cooling contributes about 3 mW m2 to the total heat flow. Predicted cooling rates for lithospheric material are in the range of 50–150 K Gy1, close to values reported recently for mantle xenoliths from the Kaapvaal craton, South Africa (Albare`de, 2003; Bedini et al., 2004). One important consequence of such long-term transient behavior stems from the shape of the vertical temperature profile. Applying a steady-state thermal model to xenolith (P, T) data leads to an overestimate of the mantle heat flux and an underestimate of the lithosphere thickness. Temperature (°C) 0
0
200
400
600
800
1000 1200 1400 1600 Ac0 = 08 µW m–3 Am0 = 0.03 µW m–3
Depth (km)
50
Qb = 10 mW m–2
6.05.6 Other Geophysical Constraints on the Thermal Regime of the Continental Lithosphere 6.05.6.1
Constraints from Seismology
The 3-D seismic velocity structure of the upper mantle determined by seismic tomography has shown strong correlation with the geology. In particular, the presence of lithospheric roots beneath cratons is associated with higher seismic velocity and lower temperature than outside. Horizontal differences in seismic velocities can be interpreted in terms of compositional and thermal differences. Within the continents, seismic velocities are higher within cratons than outside. There are also smallerscale differences that cannot be explained in terms of temperature only (Poupinet et al., 2003). Heat flux data and thermodynamic constraints can be used to narrow down the range of mantle temperatures consistent with seismic tomography models. Shapiro and Ritzwoller (2004) inverted surface-wave data to obtain vertical profiles of S-wave velocity through both continents and oceans. For a given compositional model, these data can be converted to temperature. In a given area, the solution domain allows for nonmonotonic variations of temperature with depth, that is, with zones where temperature decreases with depth, which are not physically realistic. Applying the constraint that temperature must always increase with depth leads to a narrower solution domain. The range can be further narrowed down with constraints from heat flux by eliminating solutions outside the range of Moho temperatures allowed by thermal models. One striking result is that this procedure gets rid of the nonphysical solutions with negative vertical temperature gradients (Figure 20).
100
150
Present-day Instantaneous profile
1.5 Gy Steady state
200 Present-day Steady state
250
Figure 19 Transient geotherm with decaying heat sources in the lithospheric mantle. Two steady-state calculations corresponding to the same values of heat production today and to values of heat production at 1.5 Gy. Due to the large relaxation time of thick lithosphere, temperatures are not in equilibrium with radioactive heat sources.
6.05.6.2 Seismicity, Elastic Thickness, and Thermal Regime of the Lithosphere In the oceans, both the effective elastic thickness (Te) and the maximum depth of earthquakes increase with the age of the oceanic lithosphere (Watts, 2001; Seno and Yamanaka, 1996). The dependence of Te on the age of the oceanic lithosphere (at the time of loading) was examined in many studies (Watts, 2001; Lago and Cazenave, 1981; Calmant et al., 1990) that show that, with few exceptions, oceanic Te is given approximately by the depth to the 450 C isotherm of the cooling plate model. The thickness of the
Heat Flow and Thermal Structure of the Lithosphere
(c)
0
100
Depth (km)
Depth (km)
(a)
All models
200
300
Models that satisfy the heat flow constraint 0
100
200
300 4.4
4.6
4.8
5.0
4.4
S-wave velocity (km s–1) (d)
0
100
200
4.6
4.8
5.0
S-wave velocity (km s–1)
Depth (km)
Depth (km)
(b)
243
0
100
200
300
300 0
500 1000 1500 Temperature (°C)
0
500 1000 1500 Temperature (°C)
Figure 20 Top: Vertical profiles of S-wave velocity through the Canadian Shield obtained by diffraction tomography. From Shapiro and Ritzwoller (2004) Bottom: Vertical temperature profiles deduced from the velocity data. The left panel shows the whole solution domain, which includes nonphysical temperature profiles such that temperature decreases with depth at shallow levels. The right panel shows the solutions that are consistent with bounds of the Moho temperature deduced from heat flow studies. All the nonrealistic temperature profiles have been eliminated.
seismogenic layer has been determined for intraplate settings or seaward of deep trenches (Wiens and Stein, 1984; Seno and Yamanaka, 1996). This thickness follows closely the Te estimates suggesting that in the ocean, the brittle-to-ductile transition occurs at <600 C. The effective elastic thickness of the lithosphere is related to the yield strength envelope which is useful to understand how the temperature profile affects the strength of the lithosphere. The depth where the strength begins to decrease corresponds to the transition from brittle to ductile. In the oceans, this depth is strongly controlled by temperature, that is, by the age of the plate (McKenzie et al., 2005). In the continents, the strength profile is complicated by the rheological stratification in the
lithosphere. A relationship between age, thermal regime, and strength of the continental lithosphere was suggested by Karner et al. (1983). This clearly holds for the very young lithosphere and explains differences between the elastic thickness in the Basin and Range and stable North America (Lowry and Smith, 1995). The seismogenic zone is usually shallow (<30 km) beneath the continents suggesting a ductile lithosphere (Maggi et al., 2000). This is inconsistent with the large values of Te (>80 km) observed beneath cratons that require a cold lithosphere. The temperature differences inferred from thermal models are consistent with the very long wavelength variations in elastic thickness. Small-scale variations in Te are much more difficult to account for by the thermal regime.
244
Heat Flow and Thermal Structure of the Lithosphere
6.05.6.3
Depth to the Curie Isotherms
The main sources of magnetic anomalies are present in the crust and not in the mantle. The Curie isotherm for magnetite, 580 C, will normally be located in the upper mantle beneath continents and oceans. However, high surface heat flux and elevated temperatures in the lower crust will cause a shallow Curie isotherm with thinning of the magnetic crust and a local source of magnetic anomaly. Satellite magnetic data are useful to estimate the depth to the Curie isotherm (Hamoudi et al., 1998). The high-quality magnetic data obtained by recent satellite missions have sufficient resolution to be useful for lithospheric studies (Maus et al., 2006). Satellite magnetic data have been used to confirm the elevated lower crustal temperatures beneath the Basin and Range (Mayhew, 1982) or to delineate the edge of the North American craton (Purucker et al., 2002).
6.05.6.4
Thermal Isostasy
In the oceans, long-wavelength bathymetric variations are caused by density variations in the lithosphere. The depth of sea floor below sea level is directly related to the average lithospheric density and temperature. Note that the oceanic geotherm is not in steady state. Crough and Thompson (1977) have applied similar concepts to the continental lithosphere. In the continents, density variations are due more to changing crustal thickness (and composition) than to differences in temperature. Low mantle temperature beneath the cratons should increase the density of the mantle and keep the elevation much lower than the observed mean elevation. This observation led Jordan (1981) to propose that the cratonic mantle is made up of refractory residual mantle with lower density than the off-cratonic mantle. This compositional effect balances the thermal effect to give to the cratons their present elevation. The component of the topography of the continents due to thermal isostasy is usually small, except in regions of extension. In the Basin and Range, where the typical crustal thickness is 30 km, the average elevation of 1750 m requires the uppermantle density to be anomalously low. Differences in temperature can account only for part of the elevation, and low-density magma intrusions are thought to also contribute to the buoyancy of the mantle (Lachenbruch and Morgan, 1990). The high heat flow in the Canadian Cordillera suggests that
thermal isostasy contributes to part of the elevation (Lewis et al., 2003). The buoyancy of the mantle beneath the Colorado Plateau is likely to be in part thermal, although the heat flow is not high in the Plateau, possibly because not enough time has elapsed to allow the effect of higher mantle temperature to be conducted to the surface (Bodell and Chapman, 1982).
6.05.7 Conclusions Ironically, Kelvin’s calculation applies to large parts of the Earth surface and does lead to an accurate prediction of age. The age, however, is not that of the planet but that of its oceanic plates. The failure of this simple model to account for heat flux and bathymetry of old sea floor provides very useful information on the heat transport mechanisms that are active in the mantle and helps addressing the more complex problem of continents. The mechanism which brings heat to the base of the oceanic lithosphere is probably also active beneath the continental lithosphere and explains why continents tend to thermal steady state. Lithosphere thickness is in many ways an illdefined variable. Not only does its value vary from one geophysical method to the next, but from the thermal perspective it may also change depending on the basal boundary condition. Furthermore, it may be different for steady-state and transient thermal models. Such complexities are not merely a problem of vocabulary and reflect important features of heat transport in the upper boundary layer of mantle convection. For lithospheric studies, one should regard heat flow data as affected by large geological noise. The sources of this noise are hydrothermal circulation in the oceans, and crustal radiogenic heat production in the continents. Unfortunately, this noise is locally controlled (i.e., local topography and sediment thickness for oceans, geological evolution and crustal structure for continents). Thus, one may not propose generic lithosphere models valid for continents or oceans of given age without carefully accounting for the local environment. One may not determine lithosphere structure without additional constraints and without a heat transport model. Heat flow data, however, do provide strong constraints on heat transport mechanisms.
Heat Flow and Thermal Structure of the Lithosphere
Appendix 1: Measurement Techniques Heat flux is never directly measured but its vertical component is obtained by the Fourier’s law Q ¼ k
qT qz
½27
Continental or oceanic heat flux measurements thus require the determination of the vertical temperature gradient and thermal conductivity (Beck, 1988; Jessop, 1990). Conventional Land Heat Flow Measurements On land, conventional heat flow measurements are obtained by measuring temperature in drill-holes (usually holes of opportunity, mostly mining exploration). Continuous core samples are routinely kept in mining exploration, and conductivity can be measured on samples from the hole. The divided bar method provides the most robust measurement of the bulk rock conductivity because it involves relatively large samples and is insensitive to small-scale variations in lithology, but it is time consuming and only a limited number of samples can be processed. Continuous measurement of conductivity on the entire core can be made with an optical scanning device (Popov et al., 1999). The heat flow is commonly obtained as the slope of the best-fitting line to the ‘Bullard plot’ of temperature versus thermal resistivity R(z): R ðzÞ ¼
Z 0
z
dz9 kðz9Þ
½28
Alternatively, heat flux can be obtained from Fourier’s law over depth intervals where the conductivity is constant. Both methods give comparable results even when the fit is poor or when heat flow varies between depth intervals. Because the temperature field in the upper 200 m is often perturbed by surface effects, including the effect of recent climate change, reliable heat flow measurements require deep boreholes (at least 300 m). Bottom Hole Temperature (BHT) Data Temperature measurements are also routinely available from oil exploration wells, either as BHT or drillstem tests, with a precision never better than 5–10 K after corrections. In these deep wells, the gradient can thus be estimated with a precision of 10–15%. The other difficulty of these measurements
245
is the lack of core samples for thermal conductivity, which has to be estimated from the lithology or from other physical properties (density, porosity, etc.) that are routinely logged. Although, less precise than conventional methods, these data have provided most of the estimates of heat flux in sedimentary basins and on many continental margins.
Appendix 2: Corrections Heat flux determinations assume that heat is transported vertically in steady state, and thus require no lateral variations in surface boundary conditions or physical properties. Changes in vegetation, the proximity to a lake, topography can distort the temperature field and affect the heat flow estimate (Jeffreys, 1938). Rapid erosion (or sedimentation) also affects the temperature field (Benfield, 1949). These effects are largest near the surface and the error on the heat flow is small in sufficiently deep boreholes. The effect of a lake or change in vegetation cannot be estimated without extensive data coverage in the horizontal direction but topographic effect can be accounted for and corrections can be made. If the erosion rate is known, a correction can also be applied (Carslaw and Jaeger, 1959, p. 388).
Appendix 3: Climatic Effects Temperatures near the Earth surface keep a memory of the past surface boundary conditions. This was understood by Kelvin who tried to use this memory to determine the age of the Earth (Thomson, 1864). For a periodic variations of the surface temperature, the temperature wave is attenuated exponentially as it propagates downward with a skin depth pffiffiffiffiffiffiffiffiffiffiffiffi ¼ T = (Carslaw and Jaeger, 1959, page 66), where T is the period and the thermal diffusivity. The daily and annual temperature cycles are damped over less than 0.5 or 10 m. They do not affect temperature at the depth of land heat flux measurements. Long-term variations in surface temperature could potentially significantly affect the temperature gradient. Birch (1948) had already pointed out that, following the last glacial episode that ended c. 10 000 years BP, surface temperature warming could affect the temperature gradient down to 2000 m and, if not accounted for, lead to underestimating the heat flow. If the time-varying surface
246
Heat Flow and Thermal Structure of the Lithosphere
boundary condition was known, it would be easy to account for it and make a correction. For the part of Canada covered by the Laurentide ice sheet, Jessop (1971) proposed a correction for a detailed climate history of the past 400 000 years with temperature equal to present during the interglacials and to –1 C during the glacial episodes. Because the present temperature of the ground surface in Canada is quite low, the correction is usually small (<10% the heat flux). Measurements in deep boreholes in Canada have indeed shown that heat flow does not increase much at depth and that the effect of last glaciation is small (Sass et al., 1971; Rolandone et al., 2003). These measurements also show that the thermal boundary condition at the base of the glacier might have been quite variable (possibly because it depends on the heat flow). Without detailed information on this past boundary condition, identical corrections have been applied to all the data from Canada. Similar corrections have been applied to the data from Siberia. During the past 200 years, there has been a general warming trend following the ‘Little Ice Ages’ with an acceleration since 1960. This recent warming affects temperature profiles down to 200 m. In regions where heat flux is low and the warming has been particularly strong, the temperature gradients are inverted down to 50–80 m. Borehole temperature profiles have been inverted to determine the surface temperature of the past centuries (Cermak, 1971; Lachenbruch and Marshall, 1986; Lewis, 1992). For measuring heat flux, it is now clear that reliable estimates require relatively deep boreholes (>300 m) to filter out the effect of recent warming and measure a stable temperature gradient over >100 m.
Table 6 Thermal conductivity of some rocks at room temperature Rock type
Mean (W m1 K1)
Min–max (W m1 K1)
Basalt Gabbro Gneiss
2.0 2.2 3.6
1.8–2.5 1.8–2.5 2.0–5.0a
Granite Granulite facies rocks Peridotite
3.2 —
2.8–3.6 3.01–3.48
2.8
2.3–3.4
References Clark (1966) Clark (1966) Clauser and Huenges (1995) Clark (1966) Joeleht and Kukkonen (1998) Clark (1966)
a
Thermal conductivity of gneiss in direction perpendicular to foliation is 0.6 that parallel to foliation.
K ¼ 2:26 –
618:241 255:576 þ K0 – 0:30247 ½29 T T
where K is thermal conductivity (in W m1 K1), T is the absolute temperature, and K0 is the conductivity at the surface (for T ¼ 273 K). Clauser and Huenges (1995) propose similar equations to determine the lattice conductivity. The temperature dependence of conductivity cannot be neglected. When calculated with temperature-dependent conductivity, Moho temperatures are 150 K higher than for constant conductivity. At temperature higher than 1000 K, the radiative component to the thermal conductivity must be included. For mantle rocks, the radiative component can be calculated as (Scha¨rmeli, 1979) Kr ¼ 0:37 10 – 9 T 3
½30
Appendix 4: Physical Properties The calculation of the geotherm requires thermal conductivity to be known. Thermal conductivity depends on composition: it increases with the quartz content (Clauser and Huenges, 1995, and references therein). Table 6 gives values of thermal conductivity for some important rocks and minerals. The lattice conductivity decreases with temperature. Over the range of crustal temperatures, the thermal conductivity can vary by as much as 50%. Durham et al. (1987) have measured the thermal conductivity variations for samples of different crustal rocks and proposed the following law for the thermal conductivity in the crust:
The specific heat of crustal rocks 1000 J kg1 K1. The thermal diffusivity 106 m2 s1, that is, 31.5 m2 y1.
is is
References Abbott D (1991) The case for accretion of the tectosphere by buoyant subduction. Geophysical Research Letters 18: 585–588. Ahern J and Mrkvicka S (1984) A mechanical and thermal model for the evolution of the Williston basin. Tectonics 3: 79–102. Albare`de F (2003) The thermal history of leaky chronometers above their closure temperature. Geophysical Research Letters 30, doi:10.1029/2002GL016484.
Heat Flow and Thermal Structure of the Lithosphere Artemieva IM and Mooney WD (2001) Thermal thickness and evolution of Precambrian lithosphere: A global study. Journal of Geophysical Research 106: 16387–16414. Ashwal LD, Morgan P, Kelley SA, and Percival J (1987) Heat production in an Archean crustal profile and implications for heat flow and mobilization of heat producing elements. Earth and Planetary Science Letters 85: 439–450. Ballard S, Pollack HN, and Skinner NJ (1987) Terrestrial heat flow in Botswana and Namibia. Journal of Geophysical Research 92: 6291–6300. Bank C-G, Bostock MG, Ellis RM, Hajnal Z, and VanDecar JC (1998) Lithospheric mantle structure beneath the TransHudson Orogen and the origin of diamondiferous kimberlites. Journal of Geophysical Research 103: 10103–10114. Bea F and Montero P (1999) Behaviour of accessory phases and redistribution of Zr, REE, Y, th, and U during metamorphism and partial melting of metapelites in the lower crust: An example from the Kinzigite formation of Ivrea–Verbano, northwestern Italy. Geochimica et Cosmochimica Acta 63: 1133–1153. Beck AE (1988) Methods for determining thermal conductivity and thermal diffusivity. In: Haenel R, Rybach L, and Stegena L (eds.) Handbook of Terrestrial Heat Flow Density Determination, pp. 87–124. Dordrecht: Kluwer. Becker K and Davis E (2004) In situ determinations of the permeability of the igneous oceanic crust. In: Davis E and Elderfield H (eds.) Hydrogeology of the Oceanic Lithosphere, pp. 189–224. Cambridge: Cambridge University Press. Bedini R-M, Blichert-Toft J, Boyet M, and Albare`de F (2004) Isotopic constraints on the cooling of the continental lithosphere. Earth and Planetary Science Letters 223: 99–111. Benfield AE (1949) The effect of uplift and denudation on underground temperatures. Journal of Applied Physiology 20: 66–70. Bingen B, Demaiffe D, and Hertogen J (1996) Redistribution of rare earth elements, thorium, and uranium over accessory minerals in the course of amphibolite to granulite facies metamorphism: The role of apatite and monazite in orthogneisses from southwestern Norway. Geochimica et Cosmochimica Acta 60: 1341–1354. Birch F (1948) The effects of Pleistocene climatic variations upon geothermal gradients. American Journal of Science 246: 729–760. Birch F (1965) Speculations on the earth thermal history. Geological Society of America Bulletin 76: 133–154. Birch F, Roy RF, and Decker ER (1968) Heat flow and thermal history in New England and New York. In: An-Zen E (ed.) Studies of Appalachian Geology, pp. 437–451. New York: Wiley-Interscience. Blackwell D and Richards M (2004) Geothermal Map of North America. Tulsa: American Association of Petroleum Geologists. Bodell JM and Chapman DS (1982) Heat flow in the north central Colorado plateau. Journal of Geophysical Research 87: 2869–2884. Bond G and Kominz M (1991) Disentangling middle Paleozoic sea level and tectonic events in cratonic margins and cratonic basins of North America. Journal of Geophysical Research 94: 6619–6639. Bonneville A, Von Herzen RP, and Lucazeau F (1997) Heat flow over Reunion hot spot track: Additional evidence for thermal rejuvenation of oceanic lithosphere. Journal of Geophysical Research 102: 22731–22748. Brady RJ, Ducea MN, Kidder SB, and Saleeby JB (2006) The distribution of radiogenic heat production as a function of depth in the Sierra Nevada batholith, California. Lithos 86: 229–244. Bullard EC (1939) Heat flow in South Africa. Proceeding of the Royal Society of London Series A 173: 474–502.
247
Bullard EC (1954) The flow of heat through the floor of the Atlantic ocean. Proceeding of the Royal Society of London Series A 222: 408–422. Calmant S, Francheteau J, and Cazenave A (1990) Elastic layer thickening with age of the oceanic lithosphere. Geophysical Journal International 100: 59–67. Carlson RL and Johnson HP (1994) On modeling the thermal evolution of the oceanic upper mantle: An assessment of the cooling plate model. Journal of Geophysical Research 99: 3201–3214. Carlson RW, Pearson DG, and James DE (2005) Physical, chemical, and chronological characteristics of continental mantle. Reviews of Geophysics 43: RG1001 (doi:10.1029/ 2004RG000156). Carslaw HS and Jaeger JC (1959) Conduction of Heat in Solids, 2nd edn. Oxford: Clarendon Press. Cermak V (1971) Underground temperature and inferred climatic temperature of the past millenium. Palaeogeography Palaeoclimatology Palaeoecology 98: 167–182. Chandrasekhar S (1961) Hydrodynamic and Hydromagnetic Stability. Oxford: Oxford University Press. Clark SP (1966) Thermal conductivity. In: Clark SP (ed.) Handbook of Physical Constants, pp. 459–482. Boulder: GSA. Clauser C and Huenges E (1995) Thermal conductivity of rocks and minerals. In: Ahrens TJ (ed.) A Handbook of Physical Constants: Rock Physics and Phase Relations, pp. 105–126. Washington: AGU. Courtney R and White R (1986) Anomalous heat flow and geoid across the Cape Verde Rise: Evidence of dynamic support from a thermal plume in the mantle. Geophysics Journal of the Royal Astronomical Society 87: 815–868. Crough ST (1983) Hotspot swells. Annual Review of Earth and Planetary Sciences 11: 165–193. Crough ST and Thompson GA (1977) Thermal model of continental lithosphere. Journal of Geophysical Research 81: 4857–4862. Cull JP (1991) Heat flow and regional geophysics in Australia. In: Cermak V and Rybach L (eds.) Terrestrial Heat Flow and the Lithosphere Structure, pp. 486–500. Berlin: Springer-Verlag. Davaille A and Jaupart C (1994) Onset of thermal convection in fluids with temperature-dependent viscosity: Application to the oceanic mantle. Journal of Geophysical Research 99: 19853–19866. Davies GF (1988) Ocean bathymetry and mantle convection, 1. Large-scale flow and hotspots. Journal of Geophysical Research 93: 10467–10480. Davis EE and Elderfield H (2004) Hydrogeology of the Oceanic Lithosphere. Cambridge: Cambridge University Press. Davis EE and Lister CRB (1974) Fundamentals of ridge crest topography. Earth and Planetary Science Letters 21: 405–413. Davis EE, Chapman DS, Wang K, et al. (1999) Regional heat flow variations across the sedimented Juan de Fuca ridge eastern flank: Constraints on lithospheric cooling and lateral hydrothermal heat transport. Journal of Geophysical Research 104: 17675–17688. Decker ER, Baker KR, Bucher GJ, and Heasler HP (1980) Preliminary heat flow and radioactivity studies in Wyoming. Journal of Geophysical Research 85: 311–321. Doin M-P and Fleitout L (1996) Thermal evolution of the oceanic lithosphere: An alternative view. Earth and Planetary Science Letters 142: 121–136. Doin M-P, Fleitout L, and McKenzie D (1996) Geoid anomalies and the structure of continental and oceanic lithospheres. Journal of Geophysical Research 101: 16119–16136. Dumoulin C, Doin M, and Fleitout L (2001) Numerical simulations of the cooling of an oceanic lithosphere above a
248
Heat Flow and Thermal Structure of the Lithosphere
convective mantle. Physics of the Earth and Planetary Interiors 125: 45–64. Durham WB, Mirkovich VV, and Heard HC (1987) Thermal diffusivity of igneous rocks at elevated pressure and temperature. Journal of Geophysical Research 92: 11615–11634. Fountain DM, Salisbury MH, and Furlong KP (1987) Heat production and thermal conductivity of rocks from the Pikwitonei–Sachigo continental cross-section, central Manitoba: Implications for the thermal structure of Archean crust. Canadian Journal of Earth Sciences 24: 1583–1594. Fourier JBJ (1820) Extrait d’un me´moire sur le refroidissement du globe terrestre. Bulletin des Sciences de la Societe Philomathique de Paris. Gass IG, Thorpe RS, Pollack HN, and Chapman DS (1978) Geological and geophysical parameters for mid-plate volcanism. Philosophical Transactions of the Royal Society of London Series A 301: 581–597. Gaudemer Y, Jaupart C, and Tapponier P (1988) Thermal control on post-orogenic extension in collision belts. Earth and Planetary Science Letters 89: 48–62. Griffin WL, O’Reilly SY, Abe N, et al. (2003) The origin and evolution of Archean lithospheric mantle. Precambrian Research 127: 19–41. Guillou L, Mareschal JC, Jaupart C, Garie´py C, Bienfait G, and Lapointe R (1994) Heat flow and gravity structure of the Abitibi belt, Superior Province, Canada. Earth and Planetary Science Letters 122: 447–460. Gung Y, Panning M, and Romanowiz B (2003) Global anisotropy and the thickness of the continents. Nature 422: 707–710. Hamdani Y, Mareschal J-C, and Arkani-Hamed J (1991) Phase change and thermal subsidence basins. Geophysical Journal International 106: 657–665. Hamdani Y, Mareschal J-C, and Arkani-Hamed J (1994) Phase change and thermal subsidence of the Williston basin. Geophysical Journal International 116: 585–597. Hamoudi M, Cohen Y, and Achache J (1998) Can the thermal thickness of the continental lithosphere be estimated from Magsat data. Tectonophysics 284: 19–29. Harris RN and Chapman DS (2004) Deep-seated oceanic heat flow heat deficits, and hydrothermal circulation. In: Dovis E and Elderfield H (eds.) Hydrogeology of the Oceanic Lithosphere, pp. 311–336. Cambridge: Cambridge University Press. Harris RN, Garven G, Georgen J, McNutt MK, Christiansen L, and Von Herzen RP (2000) Submarine hydrogeology of the Hawaiian archipelagic apron 2. Numerical simulations of coupled heat transport and fluid flow. Journal Geophysical Research 105: 21353–21370. Haxby W, Turcotte D, and Bird J (1976) Thermal and mechanical evolution of the Michigan basin. Tectonophysics 36: 57–75. Haxby WF and Turcotte DL (1978) On isostatic geoid anomalies. Journal of Geophysical Research 83: 5473–5478. Heestand RL and Crough ST (1981) The effect of hot spots on the oceanic age–depth relation. Journal of Geophysical Research 86: 6107–6114. Helmstaedt HH and Schulze DJ (1989) Southern African kimberlites and their mantle sample: Implications for Archean tectonics and lithosphere evolution. In: Ross J (ed.) Geological Society of Australia Special Publication: Kimberlites and Related Rocks: Volume 1. Their Composition, Occurrence, Origin, and Emplacement, vol. 14, pp. 358–368. Sydney: Geological Society of Australia. Hirth G and Kohlstedt DL (1996) Water in the oceanic upper mantle: Implications for rheology, melt extraction and the evolution of the lithosphere. Earth and Planetary Science Letters 144: 93–108.
Hirth G, Evans R, and Chave A (2000) Comparison of continental and oceanic mantle electrical conductivity: Is the Archean lithosphere dry? Geochemistry Geophysics Geosystems 1: 2000GC000,048. Holmes A (1915) Radioactivity and the earth’s thermal history: Part 1. The concentration of radioactive elements in the earth’s crust. Geological Magazine 2: 60–71. Huang J and Zhong S (2005) Sublithospheric small-scale convection and its implications for the residual topography at old ocean basins and the plate model. Journal of Geophysical Research 110: B05404 (doi:10.1029/ 2004JB003153). Huerta AD, Royden LH, and Hodges KV (1998) The thermal structure of collisional orogens as a response to accretion, erosion, and radiogenic heating. Journal of Geophysical Research 103: 15287–15302. Humler E, Langmuir C, and Daux V (1999) Depth versus age: New perspectives from the chemical compositions of ancient crust. Earth and Planetary Science Letters 173: 7–23. Hurter SJ and Pollack HN (1996) Terrestrial heat flow in the Parana´ Basin, southern Brazil. Journal of Geophysical Research 101: 8659–8672. James DE, Fouch MJ, VanDecar JC, and van der Lee S (2001) Tectospheric structure beneath southern Africa. Geophysical Research Letters 28: 2485–2488. Jaupart C (1983a) Horizontal heat transfer due to radioactivity contrasts: Causes and consequences of the linear heat flow– heat production relationship. Geophysical Journal of the Royal Astronomical Society 75: 411–435. Jaupart C (1983b) The effects of alteration and the interpretation of heat flow and radioactivity data – a reply to R. U. M. Rao. Earth and Planetary Science Letters 62: 430–438. Jaupart C and Mareschal JC (1999) The thermal structure and thickness of continental roots. Lithos 48: 93–114. Jaupart C and Mareschal JC (2003) Constraints on crustal heat production from heat flow data. In: Rudnick RL (ed.) Treatise on Geochemistry, The Crust, vol. 3, pp. 65–84. New York: Permagon. Jaupart C, Francheteau J, and Shen X-J (1985) On the thermal structure of the southern Tibetan crust. Geophysics Journal of the Royal Astronomical Society 81: 131–155. Jaupart C, Mareschal JC, Guillou-Frottier L, and Davaille A (1998) Heat flow and thickness of the lithosphere in the Canadian Shield. Journal of Geophysical Research 103: 15269–15286. Jeffreys H (1938) The disturbance of the temperature gradient in the Earth’s crust by inequalities of height. Monthly Notices of the Royal Astronomical Society. Geophysical Supplement 4: 309–312. Jessop AM (1971) The distribution of glacial perturbation of heat flow in Canada. Canadian Journal of Earth Sciences 8: 162–166. Jessop AM (1990) Thermal Geophysics. Amsterdam: Elsevier. Joeleht TH and Kukkonen IT (1998) Thermal properties of granulite facies rocks in the Precambrian basement of Finland and Estonia. Tectonophysics 291: 195–203. Johnson HP and Carlson RL (1992) Variation of sea floor depth with age: A test of models based on drilling results. Geophysical Research Letters 19: 1971–1974. Jones MQW (1987) Heat flow and heat production in the Namaqua mobile belt, South Africa. Journal of Geophysical Research 92: 6273–6289. Jones MQW (1988) Heat flow in the Witwatersrand Basin and environs and its significance for the South African Shield geotherm and lithosphere thickness. Journal of Geophysical Research 93: 3243–3260. Jordan TH (1975) The continental tectosphere. Reviews of Geophysics and Space Physics 13: 1–12.
Heat Flow and Thermal Structure of the Lithosphere Jordan TH (1981) Continents as a chemical boundary layer. Royal Society of London Philosophical Transactions Series A 301: 359–373. Jurine D, Jaupart C, Brandeis G, and Tackley PJ (2005) Penetration of mantle plumes through depleted lithosphere. Journal of Geophysical Research 110: B10104 (doi:10.1029/ 2005JB003751). Kaminski E and Jaupart C (2000) Lithospheric structure beneath the Phanerozoic intracratonic basins of North America. Earth and Planetary Science Letters 178: 139–149. Karner GD, Steckler MS, and Thorne J (1983) Long-term mechanical properties of the continental lithosphere. Nature 304: 250–253. Ketcham RA (1996) Distribution of heat-producing elements in the upper and middle crust of southern and west central Arizona: Evidence from the core complexes. Journal of Geophysical Research 101: 13,611–13,632. Kinzler RJ and Grove TL (1992) Primary magmas of mid-ocean ridge basalts 2. Applications. Journal of Geophysical Research 97: 6907–6926. Klein EM and Langmuir CH (1987) Global correlations of ocean ridge basalt chemistry with axial depth and crustal thickness. Journal of Geophysical Research 92: 8089–8115. Kremenentsky AA, Milanovsky SY, and Ovchinnikov LN (1989) A heat generation model for the continental crust based on deep drilling in the Baltic shield. Tectonophysics 159: 231–246. Kukkonen IT, Golovanova YV, Druzhinin VS, Kosarev AM, and Schapov VA (1997) Low geothermal heat flow of the Urals fold belt: Implication of low heat production, fluid circulation or paleoclimate. Tectonophysics 276: 63–85. Kutas RI (1984) Heat flow, radiogenic heat production, and crustal thickness in southwest USSR. Tectonophysics 103: 167–174. Lachenbruch AH (1970) Crustal temperature and heat production: Implications of the linear heat flow heat production relationship. Journal of Geophysical Research 73: 3292–3300. Lachenbruch AH and Marshall BV (1986) Changing climate: Geothermal evidence from permafrost in Alaska. Science 234: 689–696. Lachenbruch AH and Morgan P (1990) Continental extension, magmatism and elevation; formal relations and rules of thumb. Tectonophysics 174: 39–62. Lachenbruch AH and Sass JH (1978) Models of an extending lithosphere and heat flow in the Basin and Range Province. Memoirs of the Geological Society of America 152: 209–258. Lachenbruch AH, Sass JH, and Morgan P (1994) Thermal regime of the southern Basin and Range Province: 2. Implications of heat flow for regional extension and metamorphic core complexes. Journal of Geophysical Research 99: 22121–22133. Lago B and Cazenave A (1981) State of stress of the oceanic lithosphere in response to loading. Geophysical Journal of the Royal Astronominal Society 64: 785–799. Lago B, Cazenave A, and Marty J-C (1990) Regional variations in subsidence rate of lithospheric plates: Implication for thermal cooling models. Physics of the Earth and Planetary Interiors 61: 253–259. Lee MK, Brown GC, Webb PC, Wheildon J, and Rollin KE (1987) Heat flow, heat production and thermo-tectonic setting in mainland UK. Journal of the Geological Society 144: 35–42. Lewis TJ (1992) (Ed.) Climatic change inferred from underground temperatures. Palaeogeography Palaeoclimatology Palaeoecology 98: 78–282. Lewis TJ, Hyndman RD, and Fluck P (2003) Heat flow, heat generation and crustal temperatures in the northern Canadian Cordillera: Thermal control on tectonics. Journal of Geophysical Research 108: 2316.
249
Li X, Kind R, Yuan X, Wo¨lbern I, and Hanka W (2004) Rejuvenation of the lithosphere by the Hawaiian plume. Nature 427: 827–829. Lister CRB (1977) Estimators for heat flow and deep rock properties based on boundary layer theory. Tectonophysics 41: 157–171. Lister CRB, Sclater JG, Nagihara S, Davis EE, and Villinger H (1990) Heat flow maintained in ocean basins of great age – Investigations in the north-equatorial West Pacific. Geophysical Journal International 102: 603–630. Lowry AR and Smith BB (1995) Strength and rheology of the western US Cordillera. Journal of Geophysical Research 100: 17947–17963. Maggi A, Jackson JA, McKenzie DP, and Preistley K (2000) Earthquake focal depths, effective elastic thickness, and the strength of the continental lithosphere. Geology 28: 495–498. Mareschal JC (1983) Mechanisms of uplift preceding rifting. Tectonophysics 94: 51–66. Mareschal JC and Jaupart C (2004) Variations of surface heat flow and lithospheric thermal structure beneath the North American craton. Earth and Planetary Science Letters 223: 65–77. Mareschal JC and Jaupart C (2005) Archean thermal regime and stabilization of the cratons. In: Benn K, Condie K, and Mareschal JC (eds.) Archean Geodynamic Processes, pp. 61–73. Washington DC: AGU. Mareschal JC, Jaupart C, Cheng LZ, et al. (1999) Heat flow in the Trans-Hudson Orogen of the Canadian Shield: Implications for Proterozoic continental growth. Journal of Geophysical Research 104: 29007–29024. Mareschal JC, Jaupart C, Garie´py C, et al. (2000) Heat flow and deep thermal structure near the southeastern edge of the Canadian Shield. Canadian Journal of Earth Sciences 37: 399–414. Mareschal JC, Nyblade A, Perry HKC, Jaupart C, and Bienfait G (2004) Heat flow and deep lithospheric thermal structure at Lac de Gras, Slave Province, Canada. Geophysical Research Letters 31: L12611 (doi:10.1029/ 2004GL020,133). Mareschal JC, Jaupart C, Rolandone F, et al. (2005) Heat flow, thermal regime, and rheology of the lithosphere in the TransHudson Orogen. Canadian Journal of Earth Sciences 42: 517–532. Marty JC and Cazenave A (1989) Regional variations in subsidence rate of oceanic plates: A global analysis. Earth and Planetary Science Letters 94: 301–315. Maus S, Rother M, Hemant K, et al. (2006) Earth’s lithospheric magnetic field determined to spherical harmonic degree 90 from CHAMP satellite measurements. Geophysical Journal International 164: 319–330. Mayhew MA (1982) Application of satellite magnetic anomaly to Curie isotherm mapping. Journal of Geophysical Research 87: 4846–4854. McDonald GJF (1959) Calculations on the thermal history of the earth. Journal of Geophysical Research 64: 1967–2000. McKenzie D (1967) Some remarks on heat flow and gravity anomalies. Journal of Geophysical Research 72: 6261–6273. McKenzie D (1978) Some remarks on the development of sedimentary basins. Earth and Planetary Science Letters 40: 25–32. McKenzie D and Bickle MJ (1988) The volume and composition of melt generated by extension of the lithosphere. Journal of Petrology 29: 625–679. McKenzie D, Jackson J, and Priestley K (2005) Thermal structure of oceanic and continental lithosphere. Earth and Planetary Science Letters 233: 337–349.
250
Heat Flow and Thermal Structure of the Lithosphere
Michaut C and Jaupart C (2004) Nonequilibrium temperatures and cooling rates in thick continental lithosphere. Geophysical Research Letters 31: L24602 (doi:10.1029/ 2004GL021092). Moore WB, Schubert G, and Tackley PJ (1999) The role of rheology in lithospheric thinning by mantle plumes. Geophysical Research Letters 26: 1073–1076. Morgan P (1983) Constraints on rift thermal processes from heat flow and uplift. Tectonophysics 94: 277–298. Nicolaysen LO, Hart RJ, and Gale NH (1981) The Vredefort radioelement profile extended to supracrustal strata at Carletonville, with implications for continental heat flow. Journal of Geophysical Research 86: 10653–10662. Nielsen SB (1987) Steady state heat flow in a random medium and the linear heat flow–heat production relationship. Geophysical Research Letters 14: 318–322. Nunn JA and Sleep NH (1984) Thermal contraction and flexure of intracratonic basins: A three dimensional study of the Michigan Basin. Geophysical Journal of the Royal Astronominal Society 79: 587–635. Nyblade AA (1997) Heat flow across the East African Plateau. Geophysical Research Letters 24: 2083–2086. Nyblade AA and Pollack HN (1993) A global analysis of heat flow from Precambrian terrains – Implications for the thermal structure of Archean and Proterozoic lithosphere. Journal of Geophysical Research 98: 12207–12218. Oxburgh ER and Parmentier EM (1977) Compositional and density stratification in oceanic lithosphere; causes and consequences. Journal of the Geological Society 133: 343–355. Parker RL and Oldenburg DW (1973) Geophysics–thermal model of ocean ridges. Nature Physical Sciences 242: 137–141. Parsons B and McKenzie D (1978) Mantle convection and thermal structure of plates. Journal of Geophysical Research 83: 4485–4496. Parsons B and Sclater JG (1977) An analysis of the variation of ocean floor bathymetry and heat flow with age. Journal of Geophysical Research 82: 803–827. Perry HKC, Jaupart C, Mareschal JC, Rolandone F, and Bienfait G (2004) Heat flow in the Nipigon arm of the Keweenawan Rift, northwestern Ontario, Canada. Geophysical Research Letters 31: L15607 (doi:10.1029/ 2004GL020,159). Perry HKC, Jaupart C, Mareschal JC, and Bienfait G (2006) Crustal heat production in the Superior Province, Canadian Shield, and in North America inferred from heat flow data. Journal of Geophysical Research 111: B04401 (doi:10.1029/ 2005JB003,893). Pinet C, Jaupart C, Mareschal JC, Gariepy C, Bienfait G, and Lapointe R (1991) Heat flow and structure of the lithosphere in the eastern Canadian shield. Journal of Geophysical Research 96: 19,941–19,963. Pollack HN (1986) Cratonization and thermal evolution of the mantle. Earth and Planetary Science Letters 80: 175–182. Pollack HN and Chapman DS (1977a) On the regional variations of heat flow, geotherms, and lithospheric thickness. Tectonophys 38: 279–296. Pollack HN and Chapman DS (1977b) Mantle heat flow. Earth and Planetary Science Letters 34: 174–184. Pollack HN, Hurter SJ, and Johnston JR (1993) Heat flow from the earth’s interior: Analysis of the global data set. Reviews of Geophysics 31: 267–280. Poort J and Klerkx J (2004) Absence of a regional surface thermal high in the Baikal rift; new insights from detailed contouring of heat flow anomalies. Tectonophysics 383: 217–241. Popov YA, Pribnow DFC, Sass JH, Williams CF, and Burkhardt H (1999) Characterization of rock thermal
conductivity by high resolution optical scanning. Geothermics 28: 253–276. Poupinet G, Arndt N, and Vacher P (2003) Seismic tomography beneath stable tectonic regions and the origin and composition of the continental lithospheric mantle. Earth and Planetary Science Letters 212: 89–101. Purucker M, Langlais B, Olsen N, Hulot G, and Mandea M (2002) The southern edge of cratonic North America: Evidence from new satellite magnetometer observations. Geophysical Research Letters 29(9), doi:10.1029/ 2001GL013645. Ramondec P, Germanovich L, Damm KV, and Lowell R (2006) The first measurements of hydrothermal heat output at 9 50’N, East Pacific Rise. Earth and Planetary Science Letters 245: 487–497. Revelle R and Maxwell AE (1952) Heat flow through the floor of the eastern north Pacific ocean. Nature 170: 199–200. Roberts GO (1979) Fast viscous Be´nard convection. Geophysical and Astrophysical Fluid Dynamics 12: 235–272. Rolandone F, Jaupart C, Mareschal JC, et al. (2002) Surface heat flow, crustal temperatures and mantle heat flow in the Proterozoic Trans-Hudson Orogen, Canadian Shield. Journal of Geophysical Research 107: 2314 (doi:10.1029/ 2001JB000,698). Rolandone F, Mareschal JC, and Jaupart C (2003) Temperatures at the base of the Laurentide ice sheet inferred from heat flow data. Geophysical Research Letters 30(18): 1994 (doi:10.1029/2003GL018046). Rondenay S, Bostock MG, Hearn TM, White DJ, and Ellis RM (2000) Lithospheric assembly and modification of the SE Canadian Shield: Abitibi-Grenville teleseismic experiment. Journal of Geophysical Research 105: 13735–13754. Roy S and Rao RUM (2000) Heat flow in the Indian shield. Journal of Geophysical Research 105: 25587–25604. Rudnick RL and Fountain DM (1995) Nature and composition of the continental crust: A lower crustal perspective. Review of Geophysics 33: 267–309. Rudnick RL and Nyblade AA (1999) The thickness of Archean lithosphere: Constraints from xenolith thermobarometry and surface heat flow. In: Fei Y, Bertka CM, and Mysen BO (eds.) Mantle Petrology; Field Observations and High Pressure Experimentation: A Tribute to Francis R. (Joe) Boyd, pp. 3–11. Houston, TX: Geochemical Society. Russell JK, Dipple GM, and Kopylova MG (2001) Heat production and heat flow in the mantle lithosphere, Slave craton, Canada. Physics of the Earth and Planetary Interiors 123: 27–44. Sandiford M and McLaren S (2002) Tectonic feedback and the ordering of heat producing elements within the continental lithosphere. Earth and Planetary Science Letters 204: 133–150. Sandwell D and Schubert G (1980) Geoid height versus age for symmetric spreading ridges. Journal of Geophysical Research 85: 7235–7241. Sass JH, Lachenbruch AH, and Jessop AM (1971) Uniform heat flow in a deep hole in the Canadian shield and its paleoclimatic implications. Journal of Geophysical Research 76: 8586–8596. Sass JH, Lachenbruch AH, and Galanis SP, Jr. (1994) Thermal regime of the southern Basin and Range: 1. Heat flow data from Arizona and the Mojave desert of California and Nevada. Journal of Geophysical Research 99: 22093–22120. Scha¨rmeli G (1979) Identification of radioactive thermal conductivity in olivine up to 25 kbar and 1500K. In: Timmerhaul KD and Barber MS (eds.) Proceedings of the 6th AIRAPT Conference, pp. 60–74. New York: Plenum. Schutt D and Lesher C (2006) The effects of melt depletion on the density and seismic velocity of garnet and spinel
Heat Flow and Thermal Structure of the Lithosphere lherzolite. Journal of Geophysical Research 111: B05401 (doi:10.1029/2003JB002,950). Sclater JG, Jaupart C, and Galson D (1980) The heat flow through oceanic and continental crust and the heat loss from the earth. Reviews of Geophysics 18: 269–311. Seno T and Yamanaka Y (1996) Double seismic zones, compressional deep trench outer rise events, and superplumes. In: Bebout GE, Scholl DW, Kirby SH, and Platt JP (eds.) Subduction Top to Bottom Monograph 36, pp. 347–355. Washington, DC: AGU. Shapiro NM and Ritzwoller MH (2004) Thermodynamic constraints on seismic inversions. Geophysical Journal International 157: 1175–1188. Silver P (1996) Seismic anisotropy beneath the continents: Probing the depths of geology. Annual Review of Earth and Planetary Sciences 24: 385–432. Solomatov V and Moresi L-N (2000) Scaling of time-dependent stagnant lid convection: Application to small-scale convection on the earth and other terrestrial planets. Journal of Geophysical Research 105: 21795–21818. Stein CA (1995) Heat Flow of the Earth. In: Ahrens TJ (ed.) Global Earth Physics: A Handbook of Physical Constants, pp. 144–158. Stein CA and Stein S (1992) A model for the global variation in oceanic depth and heat flow with lithospheric age. Nature 359: 123–129. Strutt RJ (1906) On the distribution of radium in the Earth’s crust and on internal heat. Proceeding of the Royal Society of London Series A 77: 472–485.
251
Thomson W (1864) On the secular cooling of the Earth. Transaction of the Royal Society Edinburgh 23: 295–311. Turcotte DL and McAdoo DC (1979) Geoid anomalies and the thickness of the lithosphere. Journal of Geophysical Research 84: 2381–2387. Urey HC (1964) A review of atomic abundances in chondrites and the origin of meteorites. Reviews of Geophysics 2: 1–34. van der Velden AJ, Cook F, Drummond BJ, and Goleby BR (2005) Reflections on the Neoarchean: A global prespective. In: Benn K, Mareschal JC, and Condie K (eds.) Archean geodynamics and environments, Geophysical Monograph Series, vol. 164, pp. 255–265. Washington: AGU. Vasseur G and Singh RN (1986) Effects of random horizontal variations in radiogenic heat source distribution on its relationship with heat flow. Journal of Geophysical Research 91: 10397–10404. Vitorello I, Hamza VM, and Pollack HN (1980) Terrestrial heat flow in the Brazilian highlands. Journal of Geophysical Research 85: 3778–3788. Von Herzen RP, Cordery MJ, Detrick RS, and Fang C (1989) Heat flow and the thermal origin of hot spot swells: The Hawaiian Swell revisited. Journal of Geophysical Research 94: 13783–13799. Watts AB (2001) Isostasy and Flexure of the Lithosphere. Cambridge: Cambridge University Press. Wiens DA and Stein S (1984) Intraplate seismicity and stresses in young oceanic lithosphere. Journal of Geophysical Research 89: 11442–11464.
6.06
Lithosphere Stress and Deformation
M. L. Zoback, US Geological Survey, Menlo Park, CA, USA M. Zoback, Stanford University, Stanford, CA, USA Published by Elsevier B.V.
6.06.1 6.06.2 6.06.3 6.06.4 6.06.5 6.06.6 References
Introduction Global Patterns of Tectonic Stress Sources of the Lithospheric Stress Field Absolute Stress Magnitudes and the Critically Stressed Crust Stress Field Constraints on Lithospheric Deformation Concluding Remarks
6.06.1 Introduction The advent of plate tectonics brought a new meaning and understanding to the mechanically defined lithosphere. Lithosphere became synonymous with the Earth’s outer thermal boundary, the mobile ‘plates’ of plate tectonics. Motion of the plates was soon understood to be the result of a balance of forces that both drove and resisted their movement. While convection in the Earth is the ultimate source of the energy to drive plate tectonics, large density inhomogeneities associated with plate subduction and generation of new oceanic lithosphere create the major driving forces. Other potentially important forces include viscous drag at the base of plates (either driving or resisting) as well as frictional resistance along plate boundaries. Lithospheric deformation, the ‘tectonics’ of plate tectonics, is a result of the stress state within the lithosphere. In this chapter we review the evidence for, and the implications of, field and laboratory data on the state of stress in the lithosphere and the forces acting upon and within it. Intensive investigation over the past few decades has revealed that the lithospheric state of stress is remarkably uniform both with depth and over vast regions of plate interiors. The broad uniformity of observed stress orientations and relative magnitudes, even across major bends in old orogenic belts, indicates that the lithospheric stress field is the result of present-day forces and is not due to residual stresses from past tectonic activity. Modeling has shown us that the same forces acting on and within the plates to drive plate motion are the same forces responsible for the state of stress
253 253 257 259 263 269 271
in the lithosphere. Direct measurement of stress magnitudes at depth confirms a simple model in which stress differences in the crust are close to and limited by the frictional strength of the most well-oriented faults. In contrast to the relative uniformity of intraplate stress fields, deformation rates in plate interiors vary by about eight orders of magnitude. We show how knowledge of the stress field, combined with mechanical, thermal, and rheological properties, can be used to constrain both the rate and style of lithospheric deformational processes.
6.06.2 Global Patterns of Tectonic Stress Early attempts to map the state of stress in lithosphere were largely based on earthquake focal mechanisms and very sparse, and usually quite shallow, in situ stress measurements (e.g., Raleigh, 1974; Richardson et al., 1976, 1979; Pavoni, 1961, 1980). Despite a relatively low density of stress data, these early investigations hinted at broad-scale uniformity of stresses. In 1980, Zoback and Zoback introduced an integrated stress mapping strategy utilizing a variety of geologic and geophysical data: earthquake focal plane mechanisms, young geologic data on fault slip and volcanic alignments, in situ stress measurements, and stress-induced wellbore breakouts, and drilling-induced tensile fractures. The initial effort to apply this integrated mapping strategy to the conterminous United States (Zoback and
253
254
Lithosphere Stress and Deformation
Zoback, 1980) revealed two fundamental characteristics of the crustal stress field: 1. To first order, the stress field is uniform with depth throughout the upper brittle crust as indicated by the consistency of stress orientations inferred from the different data types that sampled different volumes of rock at different depths: geologic and in situ stress measurement data sampled the surface or very near surface (less than 1–2 km depth), earthquake focal mechanisms provide coverage for depths between about 5 and 20 km, whereas wellbore breakout and drilling-induced fracture data commonly sample 1–4 km deep and in some cases as deep as 5–6 km, providing a valuable link between the near-surface and the focal mechanism data. 2. Stress orientations are remarkably uniform over broad regions (length scales up to thousands of kilometers), despite large changes in crustal geology, tectonic history, and crustal thickness. A brief description of how the state of stress is inferred from the different types of stress indicators used is given in Appendix 1 : More information on the assumptions, difficulties, and uncertainties of inferring stress orientations from these different data types can be found in Zoback and Zoback (1980, 1991), Zoback et al. (1989). A quality-ranking scheme was developed by Zoback and Zoback (1989, 1991) to assess how reliably an individual determination records the tectonic stress field. The quality-ranking scheme also permits comparison of orientations inferred from very different types of stress indicators that sample different depth intervals. This quality criterion was subsequently utilized in the International Lithosphere Program’s World Stress Map Project, a large collaborative effort of data compilation and analyses by over 40 scientists from 30 different countries (Zoback, 1992). A special issue of the Journal of Geophysical Research (vol. 97, pp. 11703–12014) summarized the overall results of this project as well as presented the individual contributions of many of these investigators in various regions of the world. Today, the World Stress Map (WSM) database includes more than 10 000 entries and is maintained by a Research Group of the Heidelberg Academy of Sciences and Humanities. The global stress orientation data continue to reveal striking regional uniformity (Figure 1(a)). The success of the WSM project has validated this integrated stress mapping strategy using a variety of
data types and demonstrated that with careful attention to data quality, coherent stress patterns over large regions of the Earth can be mapped with reliability and interpreted with respect to large-scale lithospheric processes. Similar to stress orientations, stress regime (relative magnitude of the three principal stresses) also shows marked regional uniformity. Because stress is a tensor quantity, a full description of the state of stress requires information on both magnitudes and orientation. Direct stress measurements at depth as well as the style of faulting at depth revealed by earthquakes indicates that, in general, the state of stress at depth can be described by three principal stresses (S1 > S2 > S3) that lie in approximately horizontal and vertical planes (e.g., Zoback and Zoback, 1989; Zoback, 1992). The magnitude of the vertical principal stress is generally taken as the weight of the overburden. Following Anderson (1951) in his classic paper on faulting, we define stress regime by the relative magnitude of the vertical stress (Sv) to the two maximum and minimum horizontal stresses (respectively, SHmax and Shmin). Three distinct regimes are possible:
•
Normal faulting stress regime When the vertical stress dominates (S1 ¼ Sv), gravity drives normal faulting and creates horizontal extensional deformation: Sv > SHmax > Shmin
½1
•
Reverse faulting stress regime When both horizontal stresses exceed the vertical stress (S3 ¼ Sv) compressional deformation and shortening is accommodated through thrust or reverse faulting: SHmax > Shmin > Sv
½2
•
Strike-slip faulting stress regime An intermediate stress state (S2 ¼ Sv), where the difference between the two horizontal stresses (horizontal shear) dominates deformation, resulting in strike-slip faulting: SHmax Sv Shmin
Relative obtained as well inferred
½3
stress magnitude information can be from direct in situ stress measurements as from the style of tectonic faulting from earthquake focal mechanisms and
Lithosphere Stress and Deformation
255
(a) 180°
210°
240°
270°
300°
330°
0°
30°
60°
90°
120°
150°
180°
60°
60°
30°
30°
0°
0°
–30°
–30°
Method: Focal mechanism Break outs Drill induced frac. Borehole slotter Overcoming Hydro fractures Geol. indicators
–60°
–60°
Regime: NF
SS
TF
U
Quality: A B C (2004) world stress data
180°
210°
240°
270°
300°
330°
0°
30°
60°
90°
120°
150°
180°
(b)
Figure 1 (a) Maximum horizontal stress orientations from the World Stress Map project plotted on a base of average topography. Line lengths of data are proportional to quality. Stress regime are indicated by color: red, normal faulting; green strike-slip faulting; blue, reverse faulting. (b) Generalized stress map mean stress directions based on averages of clusters of data shown in Figure 1(a). A single set of thick inward-pointing arrows indicate SHmax orientations in a reverse faulting stress regime. A single set of outward-pointing arrows indicate Shmin orientations in a normal faulting stress regime. Strike-slip faulting stress regime indicated by thick-inward-pointing arrows (SHmax direction) and thin outward-pointing arrows (Shmin direction). Symbol sizes in all cases are proportional to the number and consistency of data orientations averaged from World Stress Map project. Reproduced from Zoback ML (1992) First- and second-order patterns of strees in the lithosphere: The World Stress Map project. Journal of Geophysical Research 97: 11703–11728, with permission from American Geophysical Union. (The current version of the World Stress Map database can be found at: http://www-wsm.physik.uni-karlsruhe.de/.)
geologic observations of young faulting. While global coverage is quite variable, the relative uniformity of both stress orientation and relative
magnitudes in different parts of the world is striking and permits mapping of regionally coherent tectonic stress regimes.
256
Lithosphere Stress and Deformation
Figure 1(b) presents a generalized version of the Global Stress Map that is quite similar to that presented by Zoback (1992). Tectonic stress regimes are indicated in Figure 1(b) by both color and arrow type. Blue inward-pointing arrows indicate SHmax orientations in areas of compressional (strike-slip and thrust) stress regimes. Red outward-pointing arrows give Shmin orientations (extensions direction) in areas of normal faulting regimes. Regions dominated by strike-slip tectonics are distinguished with green thick inward-pointing, and orthogonal, thin outwardpointing, arrows. Overall, arrow sizes on Figure 1(b) represent a subjective assessment of ‘quality’ related to the degree of uniformity of stress orientation and also to the number and density of data (Zoback, 1992). The data shown in Figures 1(a) and 1(b) reinforce the broad-scale patterns and general conclusions regarding the global database summarized in Zoback et al. (1989):
•
The interior portions of most plates (variously called intraplate and mid-plate regions) are dominated by compression (thrust and strike-slip stress regimes) in which the maximum principal stress is horizontal. Active extensional tectonism (normal faulting stress regimes) in which the maximum principal stress is vertical generally occurs in topographically high areas in both the continents and the oceans. Regional consistency of both stress orientations and relative magnitudes permits the definition of broad-scale regional stress provinces, many of which coincide with physiographic provinces, particularly in tectonically active regions. These provinces may have lateral dimensions on the order of 103–104 km, many times the typical lithosphere thickness of 100–300 km.
• •
Zoback (1992) referred to these broad regions subjected to uniform stress orientation or a uniform pattern of stress orientation (such as the radial pattern of stress orientations in China) as ‘first-order’ stress provinces. For example, the uniform ENE SHmax orientation in mid-plate North America covers nearly the entire continental portion of the plate lying at an average elevation of less than 1000 m (excluding the west coast), and may also extend across much of the western Atlantic basin (Zoback et al., 1986). Thus, here the stress field is uniform over roughly 5000 km in both an E–W direction and an N–S direction. As we will see in Section 6.06.3, most of these first-order stress fields are consistent with the balance of forces acting and within the plates which drive plate motions.
Once the first-order stress patterns are recognized, so called ‘second-order’ patterns or local perturbations can be identified (Zoback, 1992). These second-order stress fields are often associated with specific geologic or tectonic features; for example, lithospheric flexure, lateral strength contrasts, as well as the lateral density contrasts within the lithosphere which give rise to buoyancy forces derived from lateral variations in gravitational potential energy (e.g., Artyushkov, 1974; Fleitout and Froidevaux, 1982, 1983; Sonder, 1990; Fleitout, 1991; Wortel et al., 1991; Coblentz et al., 1994; Jones et al., 1996; Bai et al., 1992; Zoback and Mooney, 1993). These second-order stress patterns typically have wavelengths ranging from 5 to 10 (or more) times the thickness of the brittle upper lithosphere. Stress mapping has shown a number of secondorder processes that can produce rotations of SHmax orientations. Few examples include
•
A 75–85 rotation on the northeastern Canadian continental shelf possibly related to margin-normal extension derived from sediment-loading flexural stresses (Zoback, 1992). An 45 rotation of stress directions along the mid-Norwegian margin, apparently due to lithospheric flexure associated with deglaciation (see Grollimund and Zoback, 2000). A 50–60 rotation within the East African Rift relative to western Africa due to extensional buoyancy forces caused by lithospheric thinning (Zoback, 1992), and an approximately 90 rotation along the northern margin of the Paleozoic Amazonas Rift in central Brazil (Zoback and Richardson, 1996). In this final example, this rotation is hypothesized as being due to deviatoric compression oriented normal to the rift axis resulting from local lithospheric support of a dense mass in the lower crust beneath the rift (‘rift pillow’). An 50 counter-clockwise rotation of SHmax directions in the southern San Joaquin basin of California that is associated with the ‘big bend’ of the San Andreas Fault (Castillo and Zoback, 1995). A clockwise rotation of horizontal principal stress directions (and a progressive decrease in horizontal stress magnitudes) going clockwise around the northern boundary of South America. This large-scale rotation of stress directions appears to be at least in part associated with the gravitational effect of a section of the Caribbean Plate subducted to the south beneath continental S. America (Colmenares and Zoback, 2003).
• •
• •
Lithosphere Stress and Deformation
An even larger-scale example of a ‘second-order’ stress rotation can be found in the stress orientations in much of the westernmost United States which are consistent with the combined effects of right-lateral shear along the San Andreas Plate Boundary and extensional buoyancy forces driven by the topographically high regions of the western United States (e.g., Zoback and Thompson, 1978). Flesch et al. (2000) have shown with lithospheric deformational modeling that this superposition is capable’ of explaining the north–northeast maximum horizontal orientations and relative magnitudes, and can also predict the rates of deformation. This final example demonstrates that much can be learned by using the observed stress data to constrain the sources of stress acting on the lithosphere (as discussed in a later section). It also demonstrates that large-scale lithospheric heterogeneities can be as important as platedriving forces in determining the state of stress within the plates (e.g., Humphreys and Coblentz, 2007).
6.06.3 Sources of the Lithospheric Stress Field The most likely sources of the observed, regionally uniform first-order patterns of stress orientations and relative magnitudes are the large-scale forces acting on and within the plates to drive their motion. Solomon et al. (1975) and subsequent studies by Richardson et al. (1976, 1979) were the first to attempt to predict global intraplate stress orientations and relative magnitudes by modeling plate-driving forces. Following Forsyth and Uyeda (1975) and Chapple and Tullis (1977) the forces they considered included
•
constant value forces acting generally perpendicular to plate boundaries – ridge push (a symmetric force at ridges creating compression far within the plates), slab pull (the balance of the negative buoyancy of the subducting slab and the viscous and frictional resistance to subduction), trench suction (a tractions on plates induced by mantle flow moving toward downwelling at subduction zones), and collisional resistance and tractions on the base of plates – drag forces proportional to plate velocity.
•
For a detailed discussion of each of the above platedriving forces, see Chapter 6.02.
257
These forces were incorporated in thin-shell finite-element models of constant-thickness lithosphere to calculate internal stresses in the plates. Richardson et al. (1979) constrained their results by a relatively sparse sampling of the global stress field based on earthquake focal mechanisms. They concluded that ridge push and net slab pull were of comparable size and that shear tractions on the base of the plates were resistive. Zoback et al. (1989) showed a correlation between SHmax orientation and the azimuth of both absolute and relative plate velocities for several intraplate regions. However, Richardson (1992) demonstrated that the ridge push torque pole is very similar to the absolute velocity pole for most plates; thus, a comparison with absolute velocity trajectories can do little to distinguish between ridge push and basal drag as a source of stress. In fact, comparison between stress directions and local azimuths computed from velocity poles is an overly simplistic approach to evaluating the influence of plate-driving force on the intraplate stress field. At best, these correlations demonstrate the important role of the plate boundary forces and can be used to conclude that the net balance of forces driving the plates also stresses them (Zoback et al., 1989). As the intraplate stress database improved, a number of single-plate finite-element models were produced, following a similar approach to Richardson et al. (1979) and assuming a no-net torque constraint. (Richardson, 1978; Richardson and Cox, 1984; Cloetingh and Wortel, 1985, 1986; Richardson and Reding, 1991; Stefanick and Jurdy, 1992; Meijer and Wortel, 1992; Whittaker et al., 1992; Grunthal and Stromeyer, 1992; Coblentz and Sandiford, 1994; Coblentz and Richardson, 1996; Meijer et al., 1997; Coblentz et al., 1998; Flesch et al., 2000; Govers and Meijer, 2001). In many of these models the ridge push force was often used to calibrate the magnitude of all forces acting on the plates as ridge push is the best quantified of the driving forces since it can be calculated directly from bathymetry and the crust/ lithosphere structure in the oceans. Over time, these models became increasingly more sophisticated and incorporated gravitational potential energy forces (so-called ‘buoyancy forces’) within the plates related to lateral variations in lithospheric density and/or thickness (including realistic representations of ridge push force). Most studies found ridge push to be a significant force and concluded that internal buoyancy body forces due to lateral variations in lithospheric
258
Lithosphere Stress and Deformation
structure and density (lithospheric buoyancy) were also significant. Some studies found significant compressive stress transmitted across convergent margins and transforms, others found drag to be an important balancing force, either resistive or driving, often depending on how the drag force was formulated (e.g., restricted only to continental portions of plates). The influence of slab pull on the intraplate stress field was generally found to be no larger than (and sometimes smaller than) the magnitude of the other forces acting on the plates. While such models are illuminating, they generally suffer from the lack of well-constrained boundary conditions. Hence, the results of single-plate stress modeling tend to be rather nonunique. Lithospheric buoyancy forces arise from lateral variations in both crust and mantle-lid thickness. As pointed out by Fleitout and Froidevaux (1982, 1983) regions of high topography (due either to thick crust or thin mantle lid) have high potential energy, hence have a tendency to ‘spread’ or extend. In contrast, thick, cold mantle lid tends to ‘sink’ into less-dense asthenosphere, generating compression within the upper lithosphere. Several lines of evidence indicate that buoyancy forces due to inhomogeneities in lithospheric density structure can be significant and comparable in magnitude to plate-driving forces. 1. Ridge push (recognized as a key force in determining stress orientation in a number of plates, particularly those with no attached slab) is actually a buoyancy force acting over the entire profile of cooling oceanic lithosphere. 2. The dominance of extensional tectonics in tectonically active areas at high elevations >2 km) (e.g., the western United States, the East African Rift System, and the Baikal Rift) indicates that positive buoyancy forces (positive potential energy) derived from unusually low densities in the upper mantle (and/or thin lithosphere) are probably the primary force in these regions, overcoming general intraplate compression (Zoback and Mooney, 2003). 3. As noted earlier, Flesch et al. (2000) demonstrated that stress orientations in the western US are consistent with a balance of right-lateral shear along the San Andreas Plate Boundary and extensional buoyancy forces generated by the elevated crust and thin lithosphere beneath the western Cordillera. Over the past two decades significant advances in our understanding of the crustal stress field have been
matched by greatly increased knowledge of the three-dimensional (3-D) structure and lateral variations and inhomogeneities in the lithosphere as well as structure and physical properties inferred from tomographic and other seismic studies. This increased knowledge of Earth structure has been incorporated in modeling stress patterns within the plates using basal shear tractions determined from mantle flow models. Bai et al. (1992) were the first to attempt a combined approach that predicted both plate velocities and the stresses within plates using a relatively low-resolution model of mantle density anomalies to induce flow in a Newtonian viscous mantle. Bai et al. also included buoyancy forces within the lithosphere induced by variations of Moho depth. They imposed a no-net torque constraint to compute plate velocities; these velocities were then used as the shear tractions on the base of the plates to determine internal stresses. Their models showed a relatively poor correlation between the predicted stress directions and the observed regional stress field, possibly due to the coarse spacing (15 ) of their grid. However, their models were the first to really investigate the importance of sublithospheric density anomalies on the upper lithosphere stress field. Bird (1998) implemented a global model in which laterally heterogeneous plates of nonlinear rheology were separated by faults with low friction. He found that driving forces that result only from elevation differences between ocean ridges and trenches (balanced by passive basal drag and fault friction) are not compatible with observed plate velocities. His best models to match both stress orientations and plate velocities required forward drag acting on the continents only, with dense descending slabs pulling oceanic plates and stirring the more viscous mantle. Steinberger et al. (2001) also attempted to match plate velocities and stress orientations by calculating a mantle flow field from density structures inferred from seismic tomography. In this model, computed stresses also included the effects of buoyancy forces within the lithosphere due to lateral variations in lithospheric and crustal density and structure. However, their results were somewhat equivocal. While mantle flow models were found to be generally in accord with observed stress directions and plate motion, they also predicted stress directions in the absence of any effects of mantle flow which explained the stress observations nearly as well. In a comprehensive study, Lithgow-Bertelloni and Guyunn (2004) attempted to match the observed intraplate stress patterns with a set of 3-D global finite-element models that included mantle flow,
Lithosphere Stress and Deformation
lithospheric heterogeneity, and topography. They tested two versions of lithospheric heterogeneity – one based directly on seismic and other constraints (Crust 2.0, Laske et al. (2002)) and another assuming a simple model of isostatic compensation. They also implemented mantle tractions computed from two models of mantle density heterogeneity – one based on the history of subduction for the last 180 Myr (which successfully reproduces the present-day geoid and Cenozoic plate velocities) and a second inferred from seismic tomography (Grand et al., 1997). Furthermore, they investigated effects of variable viscosity structure, including the case of a low-viscosity channel between 100 and 200 km depth. The results of this study are also somewhat equivocal. Their mantle traction-only models consistently predict large extensional stresses in the center of the Pacific Plate and over much of the Atlantic – in sharp contrast to the compressional tectonism in these areas inferred from intraplate earthquake focal mechanisms (e.g., Wiens and Stein, 1985). Because their predicted stresses from mantle tractions are a factor two to four times greater than stresses due to lithospheric structure (i.e., buoyancy forces), even when they combine the two sources, the predicted stress field does not match large-scale features of the observed stresses very well, particularly in the oceans. They found the best fits for the observed stress field were predicted from models of lithosphere heterogeneity alone (these models include ridge push forces as it is a buoyancy force resulting from lateral variations of density and structure within the lithosphere). Zoback and Mooney (2003) implemented a simple model tying lithosphere buoyancy to surface elevation to predict stress regime within continental plate interiors. They determined the crustal portion of lithospheric buoyancy (density thickness) using the USGS global database of crustal structure determinations (Mooney et al., 2002; Chapter 1.11). Adopting a mid-ocean ridge as a reference column (following Lachenbruch et al. (1985) and Lachenbruch and Morgan (1990)) they computed elevations due only to the crustal component of buoyancy and found the predicted elevations exceed observed elevations in nearly all cases (97% of the data), consistent with the existence of a cool lithospheric mantle lid denser than the asthenosphere on which it floats. The difference between the observed and predicted crustal elevation is a measure of the decrease in elevation produced by the negative buoyancy of the mantle lid. This negative buoyancy was combined with a simple thermal model for the density of the mantle lid to compute
259
mantle-lid thickness. They then computed gravitational potential energy differences relative to midocean ridges by taking a vertical integral over the computed complete lithosphere density structure. Their results are shown in Figure 2(a) and show broad agreement with observed stress regime data, given in Figure 2(b). They found that thick mantle roots beneath shields lead to strong negative potential energy differences relative to surrounding regions resulting in additional compressive stresses superimposed on the intraplate stresses derived from plate boundary forces – consistent with the dominance of reverse faulting earthquakes in the intraplate shield regions. Areas of high elevation and a thin mantle lid (e.g., western US Basin and Range, East African Rift, and Baikal Rift) are predicted to be in extension, consistent with the observed stress regime in these areas. The simplicity of the observed crustal stress field thus appears most consistent with buoyancy-related forces acting directly on and within the lithosphere – in particular, ridge push and internal lithosphere density heterogeneities. Possible contributions from deep-mantle density heterogeneities, or tractions, from the flow they induce seem to be small or even unresolvable. More detailed stress data coverage and more accurate mantle/lithosphere models will be needed to determine the exact balance of forces, as well as the potential significance of drag forces. The body of work modeling stresses over the past 30 years seems to generally support the broad conclusions of the very first comprehensive modeling attempt by Richardson et al. (1979) – that ridge push and net slab pull (downwelling pull balanced by viscous and frictional resistance to subduction) were of comparable size and that shear tractions on the base of the plates are relatively small. As Richardson and Reding pointed out in 1981 in their modeling of stresses within the North American Plate, the large lateral gradients in stress magnitudes (up to an order of magnitude variation) across large plates, required by models in which drag dominates, are not observed.
6.06.4 Absolute Stress Magnitudes and the Critically Stressed Crust Direct measurements of stress to depths of 8 km have confirmed a simple yet profound model of how stress magnitudes vary with depth within the crust. This simple model is one of frictional faulting equilibrium or ‘critically stressed crust’ in which actual
260
Lithosphere Stress and Deformation
(a)
(b)
Figure 2 (a) Map of computed potential energy differences relative to a reference asthenosphere geoid shown on a base of tectonic provinces after Goodwin (1996). Positive potential energy differences giving rise to deviatoric extensional stresses are shown in red and magenta, weak negative potential energy differences are shown in green, and negative potential energy differences are shown in cyan and blue represent the largest deviatoric compression values implying strong deviatoric compressional stresses. (b) Observed stress regime data from the World Stress Map database. These represent a subset of the entire database and are those data points with stress regime information, primarily earthquake focal mechanisms and geologic stress indicators. Magenta indicates an extensional stress regime, characterized by normal faulting. Green shows data indicating strike-slip faulting. Blue points represent areas of a strongly compressional stress regime, characterized by thrust or reverse faulting. Reproduced from Zoback ML and Mooney WD (2003) Lithospheric buoyancy and continental intraplate stress. International Geology Review 45: 95–118, with permission from American Geophysical Union.
stress differences in the crust are very close to the values required for slip on the most well-oriented, preexisting faults. This model is based on classic Coulomb faulting theory and confirmed by laboratory tests of rock failure. It is also the basis for the linear
portion of lithospheric strength curves first developed by Brace and Kohlstedt (1980) that show maximum stress differences (i.e., strength) as a function of depth; the integral of these stress-difference profiles yields lithospheric strength. As reviewed by Zoback et al.
Lithosphere Stress and Deformation
(2002) and summarized below, the implications of a critically stressed crust are profound. According to Coulomb theory, fault slippage will occur when the shear stress on the fault equals the sum of the inherent fault strength So and the frictional resistance to sliding (the product of , the coefficient of friction on the fault, and n, the stress acting normal to the fault plane): ¼ So þ n
½4
(cf., Jaeger and Cook, 1979). The maximum shear stress is given as (S1 S3)/2 and corresponds to the most well-oriented fault planes for slip. Actual stress magnitudes in the Earth’s crust are modified by internal pore pressure in the rock. The concept of ‘effective stress’ is used to incorporate the influence of pore pressure at depth, a component of effective stress ij is related to the total stress Sij via ½5
where ij is the Kronecker delta and Pp is the pore pressure. By applying the concept of effective stress to Anderson faulting theory, we can predict stress magnitudes at depth for different stress regimes. Twodimensional faulting theory assumes that failure on pre-existing faults is a function only of the difference between the least and greatest principal effective stresses 1 and 3. In this case, the ratio of stress magnitudes of 1 and 3 can be shown to be a constant, related to the frictional coefficient of the most favorably oriented faults:
1 =3 ¼ S1 – Pp = S3 – Pp ¼
2 1=2 2 þ 1 þ ½6
(after Jaeger and Cook, 1971). Since the coefficient of friction is relatively well defined for most rocks and ranges between 0.6 and 1.0 (Byerlee, 1978), eqn [6] indicates that frictional sliding will occur when 1/3 3. Assuming hydrostatic pore pressure (a common assumption for which a complete justification is given later in this section) and that the vertical stress is equal to the weight of the overburden, we can compute absolute stress magnitudes for the three main Andersonian faulting regimes: faulting, extensional regimes: S • normal reverse compressional • S 2.3faulting, S, faulting regimes: (when • strike-slip þS ), S 2.2 S . S
hmin 0.6 Sv,
Hmax
Hmax
regimes:
v
hmin’
Hmax
hmin
This simple model of crustal stress magnitude has been validated by a large number of in situ stress measurements in deep wells and scientific boreholes (see review by Townend and Zoback (2000)). The deepest and most complete set of stress magnitude data (collected to 8 km depth in a scientific borehole in Germany) is shown in Figure 3. The measured stresses indicate a strike-slip stress regime and the stress differences throughout the entire depth range sampled are consistent with frictional faulting equilibrium with a frictional coefficient of 0.7 (dashed line above 10 km) (Zoback et al., 1993; Brudy et al., 1997). Further evidence for such a ‘frictional failure’ stress state is provided by a series of earthquakes that were triggered at 9 km
Sv 1/2
0
0
S1–S3 (MPa) 300 200
100
400 Temp (°C)
Stress orientation
2
50
100 Differential stress magnitudes (0.6 < μ < 0.7)
4
Depth (km)
ij ¼ Sij – ij Pp
261
150
6 200 8 250 N130 °E (±10°)
10
0
10
20
30
Number of induced earthquakes
Brittle –ductile transition?
300
12 350
14
Figure 3 Stress measurements in the German scientific research well, KTB, indicate a strong crust, in a state of failure equilibrium as predicted by Coulomb theory and laboratory-derived coefficients of friction of 0.6–0.7. The arrow at 9.2 km depth indicates where the fluid injection experiment occurred. Reproduced from Zoback MD and Harjes HP (1997) ‘Injection induced earthquakes and crustal stress at 9 km depth at the KTB deep drilling site, Germany.’ Journal of Geophysical Research 102: 18477–18491, with permission from American Geophysical Union.
262
Lithosphere Stress and Deformation
depth in rock surrounding the borehole with extremely low perturbations of the ambient, approximately hydrostatic pore pressure (Zoback and Harjes, 1997). Figure 4 (from Zoback and Townend (2001) and Townend and Zoback (2000)) shows a compilation of stress measurements in relatively deep wells and boreholes from a variety of tectonic provinces around the world. As shown in Figure 4(a), the ratio of the measured maximum and minimum effective stresses (as shown in eqn [6]) corresponds to values predicted from frictional faulting equilibrium with a coefficient of friction ranging between 0.6 and 1.0, the same range as that observed in the laboratory (Byerlee, 1978). Additional data collected at shallower depths in the crust (<3 km) substantiate the observation that the upper crust is critically stressed according to Mohr–Coulomb frictional failure theory (see reviews by McGarr and Gay (1978) and Zoback and Healy (1992)). Figure 4(b) shows the same stress measurement as presented in Figure 3(a) but this time as a function of apparent depth (see Townend and Zoback, 2000). As shown, stress magnitudes increase rapidly with depth in the brittle crust. Two important implications of the data in Figure 4 are worth noting. First, ‘Byerlee’s law’ (i.e.,
(a)
that the coefficient of frictional sliding is in the range 0.6–1.0 in the brittle crust, independent of rock type) was defined on the basis of hundreds of laboratory experiments, yet it appears to correspond to faults in situ equally well. This is a surprising result given the large difference between the size of samples used for friction experiments in the laboratory, the size of real faults in situ, the variability of roughness of the sliding surface, and the idealized conditions under which laboratory experiments are conducted, etc. Second, despite the fact that Earth’s brittle crust contains a spectrum of widely distributed faults, fractures, and planar discontinuities at many different scales and orientations, it appears that stress magnitudes at depth (specifically, the differences in magnitude between the maximum and minimum principal stresses) are limited by the frictional strength of the most well-oriented of these planar discontinuities. Because frictional strength depends on pore pressure, it is important to note that the measured stress data shown in Figure 4 are all associated with essentially hydrostatic pore pressures, suggestive of relatively high permeability throughout the upper crust. By analyzing the results of in situ hydraulic tests conducted at length scales of 10–1000 m and to depths as great as 9 km as well as the migration rates of induced seismicity over distances of up to several
(b)
Figure 4 (a) In situ stress measurements in relatively deep wells in crystalline rock indicate that stress magnitudes seem to be controlled by the frictional strength of faults with coefficients of friction between 0.6 and 1.0. (b) When converted to approximate depth, the deep borehole stress measurements indicate a rapid increase in depth, as seen in lithospheric strength profiles. (a) Reproduced from Zoback MD and Townend J (2001) Implications of hydrostatic pore pressures and high crustal strength for the deformation of intraplate lithosphere. Tectonophysics 336: 19–30, with permission from American Geophysical Union. (b) Reproduced from Townend J and Zoback MD (2000) ‘How faulting keeps the crust strong’ Geology 28(5): 399–402, with permission from American Geophysical Union.
Lithosphere Stress and Deformation
kilometers, Townend and Zoback (2000) inferred upper-crustal permeabilities between 1017 and 1016 m2, three to four orders magnitude higher than that of core samples studied in the laboratory at equivalent pressures. Geothermal and metamorphic data also indicate that the permeability of the upper crust is high (>1018 m2) throughout the brittle regime (Manning and Ingebritsen, 1999). In fact, we believe that the mechanism responsible for creating and maintaining high crustal permeability is fundamentally related to the observation that the crust is in a state of frictional faulting equilibrium. Using data from several scientific boreholes, Barton et al. (1995) demonstrated that optimally oriented planes are hydraulically conductive, whereas nonoptimally oriented planes are nonconductive. This conclusion is supported by data collected subsequently from even deeper boreholes in other tectonic settings (Hickman et al., 1997; Barton et al., 1998) and the 8 km deep borehole in Germany (Ito and Zoback, 2000). Another way of saying this is that the active faults that limit crustal strength are also responsible for maintaining pore pressures at hydrostatic values. These results clearly indicate that critically stressed faults act as fluid conduits and control large-scale permeability (Townend and Zoback, 2000; Zoback and Townend, 2001). Thus, the presence of critically stressed faults in the crust keeps the brittle crust permeable and upper-crustal pore pressures close to hydrostatic values. Three independent lines of evidence support the in situ measurements in indicating a critically stressed crust at frictional faulting equilibrium: widespread occurrence of crustal seismicity • the induced by either reservoir impoundment or fluid
• •
injection (cf. Healy et al., 1968; Raleigh et al., 1972; Pine et al., 1990; and Zoback and Harjes, 1997); earthquakes triggered by small stress changes (0.1–0.3 MPa) associated with other earthquakes (cf. Stein et al., 1992); and relatively small stress drops in crustal earthquakes (1–10 MPa) (Hanks, 1977) typically an order of magnitude smaller than expected stress differences at seismogenic depths (as shown in Figure 4(b)).
All these observations imply that the stress driving and released in earthquake faulting involves relatively small fluctuations around frictional faulting equilibrium values.
263
6.06.5 Stress Field Constraints on Lithospheric Deformation It might seem surprising that the state of stress in the crust is generally in a state of incipient frictional failure, especially in relatively stable intraplate areas. A significant implication of this observation is that observed variations in rates of active lithospheric deformation must be directly linked to lithospheric strength, not to large variations in stress state. One of the strongest parameters controlling integrated lithospheric strength is Moho temperature, (e.g., England and Molnar, 1991). Using surface heat flow as a proxy for Moho temperature, we can observe the strong correlation between lithospheric strength and deformation rate by comparing a map of US heat flow (Blackwell and Steele, 1992; Blackwell et al., 1991) with the US National Seismic Hazard Map (Frankel et al., 2002; (Figure 5). The seismic hazard map is not directly a strain-rate map, it portrays shaking hazard, which is proportional to the rate of seismicity as well as the potential size of likely earthquakes in a region, so it too is a proxy. Nonetheless, it is clear that in general within the regions of warmest crust are those deforming most rapidly. The reason for the correlation between lithospheric strength and rate of deformation can be visualized in terms of a simple conceptual model of deforming lithosphere that is in balance with the forces acting on it. The lithosphere is generally represented as three distinct layers – brittle upper crust, ductile lower crust, and ductile uppermost mantle (Figure 6, from Zoback and Townend (2001) following previous workers). Deformation in the ductile lower crust and upper mantle is governed by power-law creep law (e.g., Brace and Kohlstedt, 1980). Because any applied force to the lithosphere will result in steady-state creep in the lower crust and upper mantle, as long as the ‘threelayer’ lithosphere is coupled, stress will build up in the upper brittle layer due to the creep at deeper levels. Stress in the upper crust builds over time, eventually to the point of brittle failure. The fact that intraplate earthquakes are relatively infrequent results simply from the fact that the rate of ductile strain rate is low in the lower crust and upper mantle (Zoback et al., 2002). The detailed manner in which stress in the lithosphere is related to deformation and deformation rate can be investigated using lithospheric strength
264
Lithosphere Stress and Deformation
Figure 5 Comparison of (a) US heat flow (from Blackwell and Steele, 1992) with (b) the US National Seismic Hazard map (Frankel et al., 2003). Note the strong correlation between regions high crustal temperature and regions of high deformation rate as indicated by high seismic hazard.
Brittle seismogenic zone τ = μ(Sn–Pp) 16 km Ductile lower crust Moho 40 km Ductile upper crust
Plate-driving forces ∼3 × 1012 N m–1
ε = A exp(–Q /RT )ΔSn ductile Figure 6 Schematic illustration of how the forces acting on the lithosphere keep the brittle crust in frictional equilibrium through creep in the lower crust and upper mantle. Reproduced from Zoback MD and Townend J (2001) Implications of hydrostatic pore pressures and high crustal strength for the deformation of intraplate lithosphere. Tectonophysics 336: 19–30, with permission from American Geophysical Union.
envelopes that incorporate brittle strength (based on frictional faulting equilibrium) and power-law creep with appropriate rheologies to represent the ductile
behavior. Depending on thermal regime, there may be zones of brittle deformation in the lower crust and upper mantle as well as the upper crust, if the stress
Lithosphere Stress and Deformation
differences for brittle deformation are lower than those required for ductile deformation. Integrating this differential stress profile over the thickness of the lithosphere gives the cumulative strength (or force/length in these 2-D models) of the lithosphere. Typically, investigators select a strain rate for the lithosphere and compute the total strength (force/ length) required to deform the lithosphere at that rate (e.g., Sibson, 1983; Ranalli and Murphy, 1987; Kohlstedt et al., 1995). Zoback and Townend (2001) proposed an alternate approach to modeling lithospheric deformation by assuming a value for the cumulative force deforming the lithosphere and then calculating the resulting strain rate as a function of temperature and rheology. They applied this force-limited, steady-state deformation model to Temperature (°C) 1000 500
1500
0
60 80
0
Temperature (°C) 500 1000
1500
100
Heat flow = 67 mW m–2 Strain rate = 4 × 10–16 s–1
40 60
100
Differential stress (MPa) 100 0 200 300
0
× 1012
Cumulative strength (N m–1) 0 1 2 3 4
20
40 60 80 Archean cratons
100
Cumulative strength (N m–1) 1 2 3 4
80
20 Depth (km)
Depth (km)
80
60
0
20
60
40
100
0
20
80 Cenozoic and Mesozoic rifts
100
40
0
Depth (km)
40
0
Differential stress (MPa) 100 200 300
20 Depth (km)
Depth (km)
20
0
two end-member intraplate regions that differ markedly in their average surface heat flow, but with identical lithosphere structures – both have a 40 km-thick crust and a 60-km-thick mantle lid (Figure 7). The temperature profiles in Figure 7 were computed assuming simple heat productivity models and downward continuation of surface heat flow and allowing thermal conductivity to be a function of both temperature and depth. A strike-slip/ reverse faulting stress state (i.e., S1 > S2 > S3 Sv) was also assumed, consistent with the general compressional state of stress observed in most mid-plate and intraplate regions (see Zoback (1992)). Pore pressures in the lower crust were assumed to be nearlithostatic, following the arguments presented by Nur and Walder (1990) and consistent with the
Depth (km)
0
0
265
Heat flow = 41 mW m–2 Strain rate = 1 × 10–29 s–1
40 60 80
100
× 1012
Figure 7 A comparison between theoretical temperature, differential stress, and cumulative strength profiles for two representative intraplate regions, an area of moderate heat flow (67 m W m2) and a shield area with very low heat flow (41 mW m2). As detailed in Zoback and Townend (2001) a lithospheric structure composed of a 16-km-thick felsic upper crust (with the rheological properties of dry Adirondack granulite), a 24-km-thick mafic lower crust (dry Pikwitonei granulite), and a 60-km-thick lithospheric mantle (wet Aheim dunite) assumed. Reproduced from Zoback MD, Townend J, and Grollimund B (2002) Steady-state failure equilibrium and deformation of intraplate lithosphere. International Geology Review 44: 383–401, with permission from American Geophysical Union.
266
Lithosphere Stress and Deformation
conclusions of Manning and Ingebritsen (1999) that permeability of the lower crust probably does not exceed 1019 m2 at 30 km depth implying relatively long characteristic diffusion times (>105years) consistent with maintaining near-lithostatic pore pressures. Using the temperature–depth profiles shown in Figure 7, Zoback and Townend calculated differential stresses in the ductile portion of the upper and lower crust and lithospheric mantle and the corresponding ductile strain rate, such that the cumulative area under the stress profile (lithospheric strength) does not exceed the assumed total force/length available to deform the lithosphere, 3 1012 N m1. The rationale for limiting the deforming force/ length to this value is based on the extensive stress modeling summarized in the previous section indicating that the magnitude of the cumulative forces acting on the lithosphere are generally on the order of the ridge push force, which is fairly well constrained between 1–5 1012 N m1 based on oceanic crust and lithosphere density and structure data. The more than 12 orders of magnitude difference in the computed strain rates between the two regions subjected to the same force/length is thus related to their relative strengths. As seen in the upper half of Figure 7, a situation representative of intraplate Cenozoic and Mesozoic rifted crust with moderately high heat flow (67 mW m2) exhibits relatively high strain rates (1016 s1). This is due to the relatively high temperatures in the lower crust and upper mantle, so that relatively little force is required to cause deformation there. Note too that in this case, most of the total tectonic force is carried in the strong brittle crust. On geologic timescales this region would appear to deform as a viscous continuum. In contrast, in a cold shield region (lower half of Figure 7) the lower crust and upper mantle are considerably stronger and the total force available is sufficient to only strain the lithosphere at a rate of 1029 s1, a negligible rate (even over billions of years!) and thus consistent with a rigid plate assumption. Differences in strength of the two regions are largely driven by the >400 K difference in temperature at the Moho (40 km depth) (Figure 7). A test of such a force-limited, steady-state model is that the estimated intraplate lithospheric strain rates not exceed approximately 1017 s1, in order to be consistent with plate tectonic reconstructions assuming ‘rigid’ plates (e.g., Gordon, 1998). Because calculations such as those in Figure 7 involve a large number of parameters (surface heat flow, thermal
conductivity, upper-crustal heat productivity, the frictional coefficient of the crust and the rheological parameters of each layer), Zoback and Townend (2001) treated uncertainties in each of these parameters using a Monte Carlo technique: 1000 estimates of each parameter were drawn at random from normal distributions, and 1000 separate models were constructed. Composite temperature–depth, differential stress–depth, and strength–depth profiles were then constructed by stacking the different models’ results. Figure 8 illustrates the intraplate lithosphere modeling results for surface heat flow of 60 6 mW m2 (mean 10%), representative of stable continental heat flow (Pollack et al., 1993). The uppermost plots (Figure 8(a)–8(c)) display the model results incorporating hydrostatic pore pressures in the upper crust, and the three middle plots (Figure 8(d)–8(f)) display the corresponding results for near-lithostatic pore pressures. Note that the temperature–depth profiles are the same in both cases. At the bottom of Figure 8 is a histogram (Figure 8(g)) illustrating the range of estimated strain rates under each pore pressure condition: the strain rates are distributed log-normally about a geometric mean of approximately 1018 s1 under ‘near-hydrostatic upper-crustal pore-pressure’ conditions, and approximately 1015 s1 for nearlithostatic pore-pressure conditions. This latter value is much too high to be consistent with geologic and geodetic observations, and demonstrates the importance of near-hydrostatic fluid pressures in the upper crust for maintaining the strength of intraplate lithosphere. Although we do not illustrate it here, it is important to note that for very low surface heat flow (<50 5 mW m2) such as is characteristic of Proterozoic and Archean cratonic crust (Pollack et al., 1993), strain rates lower than 1020 s1 are expected under either pore-pressure regime. Thus, at the relatively low strain rates characterizing intraplate regions, there is sufficient plate-driving force available to overcome the integrated strength of the lithosphere, causing ductile deformation and maintaining the ‘strong’ brittle crust in a state of frictional failure equilibrium. One manifestation of high crustal strength is the efficient transmission of tectonic stress over distances of thousands of kilometers in intraplate regions documented by the stress orientation and relative magnitude data. Thus, in essence, upper crust acts as a very efficient stress guide. As an example of how lithospheric strength variations concentrate deformation, consider the partition of deformation along the San Andreas Fault System
Lithosphere Stress and Deformation
Depth (km)
(a)
Temperature (°C)
Depth (km)
Differential stress (MPa) 0
(c)
Cumulative strength (N m–1) 0
20
20
20
40
40
40
60
60
60
80
80
80
100
(d)
(b)
0
0
500
1000
1500
100
200
0
100
400
(e) 0
20
20
20
40
40
40
60
60
60
80
80
80
0
500
1000
1500
100
0
200
400
I = 0.4
0
1
2
3 × 10
0
100
(f)
267
12
0
100
I = 0.9
0
1
2
3 × 1012
(g) 250 ‘Rigid’ plates
Not allowed (VLBl, plate reconstructions)
200
Number
150
Near-Hydrostatic (I = 0.4)
100
Near-lithostatic (I = 0.9)
50
0 –30
–28
–26
–24
–22
–20 –18 –16 Log (strain rate) (s–1)
–14
–12
–10
–8
Figure 8 (a–g) Results of 1000 Monte Carlo strain rate calculations for a strike-slip stress state and surface heat flow of 60 6 mW m2, subject to the constraint that the total strength of the lithosphere is 3 1012 N m1. Reproduced from Zoback MD and Townend J (2001) Implications of hydrostatic pore pressures and high crustal strength for the deformation of intraplate lithosphere. Tectonophysics 336: 19–30, with permission from American Geophysical Union.
in central California. As pointed out by Page et al. (1998), deformation along the San Andreas system is transpressional; in addition to the right-lateral shear accommodating relative motion between the Pacific and North American Plates, appreciable fault-normal
convergence has been occurring since about 3.5 Ma. This convergence has resulted in uplift, folding, and reverse faulting over a broad zone about 100 km wide and corresponding to the Coast Ranges adjacent to the San Andreas.
Lithosphere Stress and Deformation
38
dashed trajectories, interpolated from the results of Flesch et al. (2000). As previously discussed, the Flesch et al. model incorporates the combined effects of buoyancy-related stresses in the western United States (principally due to the thermally uplifted Basin and Range Province) and right-lateral shear in the far field associated with plate interaction. The abrupt change` in the rate of deformation at the eastern boundary between the Coast Ranges and Great Valley coincides with a marked decrease in heat flow, and therefore with lower-crustal and upper-mantle temperatures. Zoback et al. (2002) thus inferred that the rate of deformation is high throughout the Coast Ranges because temperatures in the lower crust and upper mantle are high. In contrast, heat flow in the Great Valley is extremely low (comparable to that of shield areas); hence, the available force is insufficient to cause deformation at appreciable rates. In fact, as revealed by undeformed seismic reflectors corresponding to formations as old as Cretaceous in age, it is remarkable how little deformation has occurred in the Great Valley during the Cenozoic (e.g., Wentworth and Zoback, 1989).
24
°
°
This type of distributed deformation could be categorized as a diffuse plate boundary (e.g., Gordon, 1998; Wessel and Mueller, 2007), but it is interesting to consider more specifically why the transpressional deformation is distributed so broadly, and why there is such an abrupt cessation of this deformation at the eastern boundary between the Coast Ranges and the Great Valley. The sharpness of this transition is particularly distinctive given that the entire region is subject to a relatively uniform compressive stress field acting at a high angle to the San Andreas Fault and subparallel strike-slip faults (Figure 9). This high angle implies that the San Andreas Fault (and perhaps other plate boundaries) have low frictional strength, in marked contrast to the high frictional strength exhibited by intraplate faults (briefly summarized by Zoback (2000)). The stress observations shown in Figure 9 (primarily from wellbore breakout measurements in oil and gas wells and earthquake focal mechanisms not associated with right-lateral slip along transform faults; see Mount and Suppe (1987) and Zoback et al. (1987)) are remarkably consistent with modeled stress directions across the region and shown by
36
268
0°
B′ 23
8°
A′
24
0°
200 km
B 36
°
38
°
A
23 8°
Figure 9 Topographic map of western California in an oblique Mercator projection about the NUVEL 1A North America– Pacific Euler pole (DeMets et al., 1990). In this projection, relative plate motion is parallel to the upper and lower margins of the map. The major right-lateral strike-slip faults comprising the San Andreas Fault System are also shown. The data show the direction of maximum horizontal stress from earthquake focal mechanism inversions (lines with a circle in the middle) or borehole stress measurements (bow-tie symbols). The dashed trajectories are interpolations of stress directions calculated by Flesch et al. (2000) using a model based on lithospheric buoyancy and plate interaction. Reproduced from Townend and Zoback MD (2004), with permission from American Geophysical Union.
Lithosphere Stress and Deformation
6.06.6 Concluding Remarks Several decades of study have produced a surprisingly simple and consistent picture of the lithospheric state of stress:
•
The lithospheric stress field is the result of present-day active tectonics, and not related to residual stresses from past tectonic activity – this is indicated by remarkably uniform stress orientations over broad regions of the lithosphere (scales up to thousands of kilometers) including consistency across major bends in old orogenic belts. The same forces acting on and within the plates to drive plate motion are largely responsible for the stress state within the plates – this is demonstrated by the consistency of broad regional intraplate stress field with stresses predicted by models of these driving forces. Buoyancy forces due to lateral variations in lithospheric structure and density are also a significant contributor to the intraplate stress field. Most regions of intraplate extension are in regions of high topography generated by a thinned mantle lid. Direct stress measurements to depths of 8.1 km in deep wells and scientific research boreholes confirm the fact that stress magnitudes within the brittle crust are controlled by frictional strength. In fact, stress differences within the upper brittle layer indicate a state of incipient frictional failure on well-oriented pre-existing fault planes. Because the brittle crust appears to be in a state of incipient frictional failure, the rate of deformation in a region is determined by the overall strength of the lithosphere, hence young, hot lithosphere (e.g., along active plate boundaries) deforms much more rapidly than colder, stronger lithosphere in mid-plate regions. However, measured stress magnitudes in the two contrasting tectonic regions would be identical. Balancing observed rates of lithospheric deformation with applied forces suggests that most of the stress within the lithosphere is carried in its strong, uppermost brittle crustal layer (from the surface to 15–20 km depth).
•
•
•
•
Appendix 1: Indicators of Contemporary Stress Zoback and Zoback (1980) developed an integrated stress mapping strategy in the lithosphere based on data from a variety of sources: earthquake focal plane mechanisms, young geologic data on fault slip and
269
volcanic alignments, in situ stress measurements, and stress-induced wellbore breakouts, and drillinginduced tensile fractures. Each stress indicator is explained briefly below.
Earthquake Focal Mechanisms While earthquake focal plane mechanisms are the most ubiquitous indicator of stress in the lithosphere, determination of principal stress orientations and relative magnitudes from these mechanisms must be done with appreciable caution. The pattern of seismic radiation from the focus of an earthquake permits construction of earthquake focal mechanisms. Perhaps the most simple and straightforward information about in situ stress that is obtainable from focal mechanisms and in situ stress is that the type of earthquake (i.e., normal, strike-slip, or reverse faulting) defines the relative magnitudes of SHmax, Shmin, and Sv. In addition, the orientation of the fault plane and auxiliary plane (which bound the compressional and extensional quadrants of the focal plane mechanism) define the orientation of the P (compressional), B (intermediate), and T (extensional) axes. These axes are sometimes incorrectly assumed to be the same as the orientation of S1, S2, and S3. For cases in which laboratory-measured coefficients of fault friction of 0.6–1.0 are applicable to the crust, there is usually not a large error if one uses the P, B, and T axes as approximations of average principal stress orientations, especially if the orientation of the fault plane upon which the earthquake occurred is known (Raleigh et al., 1972). However, if friction is negligible on the faults in question, there can be considerable difference between the P, B, and T axes and principal stress directions (McKenzie, 1969). An earthquake focal plane mechanism always has the P and T axes at 45 to the fault plane and the B axes in the plane of the fault. With a frictionless fault the seismic radiation pattern is controlled by the orientation of the fault plane and not the in situ stress field. One result of this is that just knowing the orientation of the P-axis of earthquakes along weak, plate-bounding strike-slip faults (like the San Andreas) does not allow one to define principal stress orientations from the focal plane mechanisms of the strike-slip earthquakes occurring on the fault (Zoback et al., 1987). Principal stress directions can be determined directly from a group of earthquake focal mechanisms (or set of fault striae measurements) through use of inversion techniques that are based on the slip kinematics and the assumption that fault slip will
270
Lithosphere Stress and Deformation
always occur in the direction of maximum resolved shear stress on a fault plane (cf., Angelier, 1990; Gephart and Forsyth, 1984; Michael, 1984). Such inversions yield four parameters, the orientation of the three principal stresses and the relative magnitude of the intermediate principal stress with respect to the maximum and minimum principal stress. The analysis of seismic waves radiating from an earthquake also can be used to estimate the magnitude of stress released in an earthquake (stress drop), although not absolute stress levels (Brune, 1970). In general, stress drops of crustal earthquakes are on the order of 1–10 MPa (Hanks, 1977). Equation (7) can be used to show that such stress drops are only a small fraction of the shear stresses that actually cause fault slip if pore pressures are approximately hydrostatic at depth and Coulomb faulting theory (with laboratoryderived coefficients of friction) is applicable to faults in situ. This is discussed in more detail below.
Geologic Stress Indicators There are two general types of ‘relatively young’ geologic data that can be used for in situ stress determinations: (1) the orientation of igneous dikes or cinder cone alignments, both of which form in a plane normal to the least principal stress (Nakamura, 1977) and (2) fault slip data, particularly the inversion of sets of striae (i.e., slickensides) on faults as described above. Of course, the term ‘relatively young’ is quite subjective but essentially means that the features in question are characteristic of the tectonic processes currently active in the region of question. In most cases, the WSM database utilizes data which are Quaternary in age, but in all areas represent the youngest episode of deformation in an area.
In Situ stress Measurements Numerous techniques have been developed for measuring stress at depth. Amadei and Stephansson (1997) and Engelder (1993) discuss many of these stress measurement methods, most of which are used in mining and civil engineering. Because we are principally interested here in regional tectonic stresses (and their implications) and because a variety of nontectonic processes affect in situ stresses near the earth’s surface (Engelder and Sbar, 1984), we do not utilize near-surface stress measurements in the WSM or regional tectonic stress compilations (these measurements are given the lowest quality in the criteria used by WSM
since they are not believed to be reliably indicative of the regional stress). In general, we believe that only in situ stress measurements made at depths greater than 100 m are indicative of the tectonic stress field at midcrustal depths. This means that techniques utilized in wells and boreholes, which access the crust at appreciable depth, are especially useful for stress measurements. When a well or borehole is drilled, the stresses that were previously supported by the exhumed material are transferred to the region surrounding the hole. The resultant stress concentration is well understood from elastic theory. Because this stress concentration amplifies the stress difference between far-field principal stresses by a factor of 4, there are several other ways in which the stress concentration around boreholes can be exploited to help measure in situ stresses. The hydraulic fracturing technique (e.g., Haimson and Fairhurst, 1970; Zoback and Haimson, 1982) takes advantage of this stress concentration and, under ideal circumstances, enables stress magnitude and orientation measurements to be made to about 3 km depth (Baumga¨rtner et al., 1990). The most common method of determining stress orientation from observations in wells and boreholes are stress-induced wellbore breakouts. Breakouts are related to a natural compressive failure process that occurs when the maximum hoop stress around the hole is large enough to exceed the strength of the rock. This causes the rock around a portion of the wellbore to fail in compression (Bell and Gough, 1979; Zoback et al., 1985). For the simple case of a vertical well drilled when Sv is a principal stress, this leads to the occurrence of stress-induced borehole breakouts that form at the azimuth of the minimum horizontal compressive stress. Breakouts are an important source of crustal stress information because they are ubiquitous in oil and gas wells drilled around the world and because they also permit stress orientations to be determined over a great range of depth in an individual well. Detailed studies have shown that these orientations are quite uniform with depth, and independent of lithology and age (e.g., Castillo and Zoback, 1994). Another form of naturally occurring wellbore failure is drilling-induced tensile fractures. These fractures form in the wall of the borehole at the azimuth of the maximum horizontal compressive stress when the circumferential stress acting around the well locally goes into tension, they are not seen in core from the same depth (Moos and Zoback, 1990; Brudy and Zoback, 1993, 1999; Brudy et al., 1997; Lund and Zoback, 1999; Peska and Zoback, 1995).
Lithosphere Stress and Deformation
References Amadei B and Stephansson O (1997) Rock Stress and Its Measurement. London: Chapman and Hall. Anderson EM (1951) The Dynamics of Faulting and Dyke Formation with Applications to Britain, 206 pp. Edinburgh: Oliver and Boyd. Angelier J (1979) ‘Determination of the mean principal directions of stresses for a given fault population.’ Tectonophysics 56: T17–T26. Angelier J (1984) ‘Tectonic analysis of fault slip data sets.’ Journal of Geophysical Research 89: 5835–5848. Artyushkov EV (1973) Stresses in the lithosphere caused by crustal thickness inhomogeneities. Journal of Geophysical Research 78: 7675–7708. Bai W, Vigny C, Ricard Y, and Froidevaux C (1992) On the origin of deviatoric stresses in the lithosphere. Journal of Geophysical Research 97: 11729–11737. Barton CA, Zoback MD, and Moos D (1995) ‘Fluid flow along potentially active faults in crystalline rock.’ Geology 23: 683–686. Barton CA, Hickman SH, Morin R, Zoback MD, and Benoit D (1998) Reservoir-scale fracture permeability in the Dixie Valley, Nevada, geothermal field: Abstracts Volume, Society of Petroleum Engineers Annual Meeting, Trondheim, Paper Number 47371, p. 315–322. Baumga¨rtner J, Rummel F, Zoback MD (1990) Hydraulic fracturing in situ stress measurements to 3 km depth in the KTB pilot hole VB. A summary of a preliminary data evaluation, in KTB Report 90–6a: 353–400. Bell JS and Gough DI (1979) Northeast–southwest compressive stress in Alberta: Evidence from oil wells. Earth and Planetary Science Letters 45: 475–482. Bird P (1998) Testing hypotheses on plate-driving mechanisms with global lithosphere models including topography, thermal structure and faults. Journal of Geophysical Research 103: 10115–10129. Brace WF and Kohlstedt DL (1980) Limits on lithospheric stress imposed by laboratory experiments. Journal of Geophysical Research 85: 6248–6252. Blackwell DD and Steele JL (1992) Geothermal Map of North America. Geological Society of America DNAG Map No. 006. Blackwell DD, Steele JL, and Carter LS (1991) Heat flow patterns of the North American continent: A discussion of the DNAG geothermal map of North America. In: Slemmons DB, Engdahl ER, and Blackwell DD (eds.) Neotectonics of North America: Geological Society of America DNAG Decade Map, vol. 1, pp. 423–437 (498pp.). Boulder, CO: Geological Society of America. Brudy M and Zoback MD (1993) ‘Compressive and tensile failure of boreholes arbitrarily-inclined to principal stress axes: Application to the KTB boreholes, Germany.’ International Journal Rock Mechanics Mining Sciences 30: 1035–1038. Brudy M and Zoback MD (1999) ‘Drilling-induced tensile wallfractures: Implications for the determination of in-situ stress orientation and magnitude.’ International Journal of Rock Mechanics and Mining Sciences 36: 191–215. Brudy M, Zoback MD, Fuchs K, Rummel F, and Baumga¨rtner J (1997) ‘Estimation of the complete stress tensor to 8 km depth in the KTB scientific drill holes: Implications for crustal strength.’ Journal of Geophysical Research 102: 18453–18475. Brune JN (1970) Tectonic stress and the spectra of seismic shear from earthquakes. Journal of Geophysical Research 75: 4997–5009. Byerlee JD (1978) Friction of rock. Pure and Applied Geophysics 116: 615–626.
271
Castillo D and Zoback MD (1995) ‘Systematic stress variations in the Southern San Joaquin valley and along the White Wolf fault: Implications for the rupture mechanics of the 1952 Ms 7.8 Kern County earthquake and contemporary seismicity.’ Journal of Geophysical Research 100(B4): 6249–6264. Castillo DA and Zoback MD (1994) Systematic variations in stress state in the Southern San Joaquin Valley: Inferences based on well-bore data and contemporary seismicity. American Association Petroleum Geologists Bulletin 78(8): 1257–1275. Chapple WM and Tullis TE (1977) Evaluation of the forces that drive the plates. Journal of Geophysical Research 82: 1967–1984. Cloetingh SAPL and Wortel MJR (1985) Regional stress field of the Indian Plate. Geophysical Research Letters 12: 77–80. Cloetingh SAPL and Wortel MJR (1986) Stress in the IndoAustralian plate. Tectonophys 132: 49–67. Coblentz DD and Richardson RM (1996) Analysis of the South American intraplate stress field. Journal of Geophysical Research 101: 8643–8657. Coblentz DD, Richardson RM, and Sandiford M (1994) On the gravitational potential of the Earth’s lithosphere. Tectonics 13: 929–945. Coblentz DD and Sandiford M (1994) Tectonic stresses in the African plate: Constraints on the ambient lithospheric stress state. Geology 22: 831–834. Coblentz DD, Zhou S, Hillis RR, Richardson RM, and Sandiford M (1998) Topography, boundary forces, and the Indo-Australian intraplate stress field. Journal of Geophysical Research 103: 919–931. Colmenares LB and Zoback MD (2003) stress field and seismotectonics of Northern South America. Geology 31: 721–724. DeMets C, Gordon RG, Argus DF, and Stein S (1990) Current plate motions. Geophysical Journal International 101: 425–478. Engelder T (1993) Stress Regimes in the Lithosphere. Princeton, New Jersey: Princeton University Press. Engelder T and Sbar ML (1984) Near-surface in situ stress: Introduction. Journal of Geophysical Research 89: 9321–9322. England PC and Molnar P (1991) Inferences of deviatoric stress in actively deforming belts from simple physical models. Philosophical Transactions of the Royal Society of London, A 337: 73–81. Fleitout L (1991) The sources of lithospheric tectonic stresses. Philosophical Transactions of the Royal Society of London Series A 337: 73–81. Fleitout L and Froidevaux C (1982) Tectonics and topography for a lithosphere containing density heterogeneities. Tectonics 1: 21–56. Fleitout L and Froidevaux C (1983) Tectonic stresses in the lithosphere. Tectonics 2: 315–324. Flesch LM, Holt WE, Haines AJ, and Shen-Tu B (2000) Dynamics of the Pacific-North American plate boundary in the Western United States. Science 287: 834–836. Frankel AD, Petersen MD, Mueller CS, et al. (2002) Documentation for the 2002 Update of the National Seismic Hazard Maps: USGS Open-File Report 02-420, 33pp. Forsyth DW and Uyeda S (1975) On the relative importance of driving forces of plate motion. Geophysical Journal of the Royal Astronomical Society 43: 163–200. Gephart JW and Forsyth DW (1984) ‘An improved method for determining the regional stress tensor using earthquake focal mechanism data: Application to the San Fernando earthquake sequence.’ Journal of Geophysical Research 89: 9305–9320. Goodwin AM (1996) Principles of Precambrian Geology, 327 pp. London: Academic Press.
272
Lithosphere Stress and Deformation
Gordon RG (1998) The plate tectonic approximation: Plate nonrigidity, diffuse plate boundaries, and global plate reconstructions. Annual Review of Earth and Planetary Sciences 26: 615–642. Govers R and Meijer PT (2001) On the dynamics of the Juan de Fuca plate. Earth and Planetary Science Letters 189: 115–131. Grand S, van der Hilst RD, and Widiyantoro S (1997) Global seismic tomography: A snapshot of convection in the Earth. GSA Today 7: 1–7. Grollimund BR and Zoback MD (2001) ‘Impact of glaciallyinduced stress changes on hydrocarbon exploration offshore Norway.’ American Association of Petroleum Geologists Bulletin 87(3): 493–506. Grunthal G and Stromeyer D (1992) The recent crustal stress field in central Europe – Trajectories and finite-element modeling. Journal of Geophysical Research 97: 11805–11820. Haimson BC and Fairhurst C (1970) In situ stress determination at great depth by means of hydraulic fracturing. In: Sumerton W (ed.) Proceedings of the 11th US Symposium on Rock Mechanics, pp. 559–584. New York: Society of Mining Engineers of AIME. Hanks TC (1977) Earthquake stress drops, ambient tectonic stresses and stresses that drive plate motions. Pure and Applied Geophysics 115: 441–458. Healy JH, Rubey WW, Griggs DT, and Ralieg CB (1968) ‘The Denver earthquakes.’ Science 161: 1301–1310. Hickman SH, Barton CA, Zoback MD, Morin R, Sass J, and Benoit R (1997) ‘In-situ stress and fracture permeability along the Stillwater fault zone, Dixie Valley, Nevada.’ International Journal of Rock Mechanics and Mining Sciences 34: 3–4 (Paper No. 126). Hildenbrand TG (1985) Rift Structure of the Northern Mississippi Embayment from the analysis of gravity and magnetic data. Journal of Geophysical Research 90: 12607–12622. Ito T, Zoback MD, and Peska P (2001) ‘Utilization of mud weights in excess of the least principal stress to stabilize wellbores: Theory and practical examples.’ Society of Petroleum Engineers Drilling and Completions 16: 221–229. James TS and Bent AL (1994) A comparison of Eastern North American seismic strain-rates to glacial rebound strainrates. Geophysical Research Letters 21: 2127–2130. Jones CH, Unruh JR, and Sonder LJ (1996) The role of gravitational potential energy in active deformation in the Southwestern United States. Nature 381: 37–41. Kohlstedt DL, Evans B, and Mackwell SJ (1995) Strength of the lithosphere: Constraints imposed by laboratory experiments. Journal of Geophysical Research 100: 17587–17602. Lachenbruch AH, Sass JH, and Galanis SP, Jr. (1985) Heat flow in southernmost California and the origin of the Salton Trough. Journal of Geophysical Research 90: 6709–6736. Lachenbruch AH and Morgan P (1990) Continental extension, magmatism and elevation; formal relations and rules of thumb. Tectonophysics 174: 39–62. Lithgow-Bertelloni C and Guynn JH (2004) Origin of the lithospheric stress field. Journal of Geophysical Research 109: B01408 (doi: 10.1029/2003JB002467). Lund B and Zoback MD (1999) Orientation and magnitude of in situ stress to 6.5 km depth in the Baltic Shield. International Journal of Rock Mechanics and Mining Sciences 36: 169–190. Manning CE and Ingebritsen SE (1999) Permeability of the continental crust: Implications of geothermal data and metamorphic systems. Reviews of Geophysics 37: 127–150. Meijer PT, Govers R, and Wortel MJR (1997) Forces controlling the present-day state of stress of the Andes. Earth and Planetary Science Letters 148: 157–170. Meijer PT and Wortel MJR (1992) The dynamics of motion of the South American plate. Journal of Geophysical Research 97: 1915–1932.
McGarr A and Gay NC (1978) State of stress in the Earth’s crust. Annual Review of the Earth and Planetary Sciences 6: 558–562. McKenzie DP (1982) The relation between fault plane solutions for earthquakes and the directions of the principal stresses. Bulletin of Seismological Society of America 59: 591–601. Michael AJ (1987) The use of focal mechanisms to determine stress: A control study. Journal of Geophysical Research 92: 357–368. Mooney WD, Prodehl C, and Pavlenkova N (2002) Seismic velocity structure of the continental lithosphere from controlled source data. In: Lee WHK (ed.) International Handbook of Earthquake and Engineering Seismology, vol. 81A, pp. 887–910. San Diego, CA: Academic Press. Moos D and Zoback MD (1990) ‘Utilization of observations of well bore failure to constrain the orientation and magnitude of crustal stresses: Application to Continental Deep Sea Drilling Project and Ocean Drilling Program Boreholes.’ Journal of Geophysical Research 95: 9305–9325. Mount VS and Suppe J (1987) State of stress near the San Andreas Fault: Implications for wrench tectonics. Geology 15: 1143–1146. Mueller B, Zoback ML, Fuchs K, et al. (1992) Regional patterns of tectonic stress in Europe. Journal of Geophysical Research 92: 11783–11803. Nakamura K (1977) Volcanoes as possible indicators of tectonic stress orientation – Principle and proposal. Journal of volcanology and Geothermal Research 2: 1–16. Nur A and Walder J (1990) Time-dependent hydraulics of the Earth’s Crust. The role of fluids in crustal processes. Washington DC: National Research Council 113–127. Page BM, Thompson GA, and Coleman RG (1998) Late Cenozoic tectonics of the Central and Southern Coast Ranges of California. Geological Society of America Bulletin 110: 846–876. Peska P and Zoback MD (1995) ‘Compressive and tensile failure of inclined wellbores and determination of in situ stress and rock strength.’ Journal of Geophysical Research 100(B7): 12791–12811. Pine RJ, Jupe A, and Tunbridge LW (1990) An evaluation of in situ stress measurements affecting different volumes of rock in the Carnmenellis granite. In: Cunha PD (ed.) Scale Effects in Rock Masses, pp. 269–277. Rotterdam: Balkema. Pollack HN, Hurter SJ, and Johnson JR (1993) Heat flow from the Earth’s interior: analysis of the global data set. Reviews of Geophysics 31: 267–280. Pavoni N (1961) Faltung durch Horizontal verschiebung. Ecolgae Geologicae Helvetiae Basel 54: 515–534. Pavoni N (1980) Crustal stress inferred from fault-plane solutions of earthquakes and neotectonic deformation in Switzerland. Rock Mechanics, Supplement 9: 63–68. Raleigh CB (1974) Crustal stress and global tectonics. In: International Society for Rock Mechanics and US National Committee for Rock Mechanics (eds.) Advances in Rock Mechanics: Proceedings of the 3rd Congress, International Society for Rock Mechanics, vol. 1A, pp. 593–597. Washington, DC: National Academy of Sciences. Raleigh CB, Healy JH, and Bredehoeft JD (1972) Faulting and crustal stress at Rangely, Colorado. In: Heard HC (ed.) Geophysical Monograph Series 16: Flow and Fracture of Rocks, pp. 275–284. Washington, DC: AGU. Ranalli G and Murphy DC (1987) Rheological stratification of the lithosphere. Tectonophysics 132: 281–295. Richardson RM (1978) Finite element modeling of stress in the Nazca plate: Driving forces and plate boundary earthquakes. Tectonophys 50: 223–248. Richardson RM (1992) Ridge forces, absolute plate motions, and the intraplate stress field. Journal of Geophysical Research 97: 1739–1748.
Lithosphere Stress and Deformation Richardson RM and Cox BL (1984) Evolution of oceanic lithosphere: A driving force study of the Nazca plate. Journal of Geophysical Research 89: 10043–10052. Richardson RM and Reding LM (1991) North American plate dynamics. Journal of Geophysical Research 96: 12201–12223. Richardson RM, Solomon SC, and Sleep NH (1976) Intraplate stress as an indicator of plate tectonic driving forces. Journal of Geophysical Research 81: 1847–1856. Richardson RM, Solomon SC, and Sleep NH (1979) Tectonic stress in the plates. Reviews of Geophysics 17: 981–1019. Sibson RH (1983) Continental fault structure and the shallow earthquake source. Journal of the Geological Society of London 5: 741–767. Solomon SC, Sleep NH, and Richardson RM (1975) On the forces driving plate tectonics: Inferences from absolute plate velocities and intraplate stress. Geophysical Journal of the Royal Astronomical Society 42: 769–801. Sonder LJ (1990) Effects of density contrasts on the orienation of stresses in the lithosphere: Relation to principal stress directions in the Transverse ranges, California. Tectonics 9: 761–771. Stefanick M and Jurdy DM (1992) Stress observations and driving force models for the South American plate. Journal of Geophysical Research 97(B8): 11905–11913. Stein S, Cloetingh S, Sleep NH, and Wortel R (1989) Passive margin earthquakes, stresses and rheology. In: Gregersen S and Basham P (eds.) Earthquakes at North-Atlantic Passive Margins: Neotectonics and Postglacial Rebound, NATO ASI Series C: pp. 231–259. Dordrecht: Kluwer Academic Publishers. Stein RS, King GC, and Lin J (1992) ‘Change in failure stress on the Southern San Andreas fault system caused by the 1992 magnitude 7.4 Landers earthquake.’ Science 258: 1328–1332. Stein S, Sleep NH, Geller RJ, Wang S-C, and Kroeger GC (1979) Earthquakes along the passive margin of eastern Canada. Geophysical Research Letters 6: 537–540. Steinberger B, Schmeling H, and Marquart G (2001) Large-scale lithospheric stress field and topography induced by global mantle circulation. Earth and Planetary Science Letters 186: 75–91. Townend J (2003) Mechanical Constraints on the Strength of the Lithosphere and plate-bounding Faults. Geophysics. PhD Thesis, 135, Stanford, CA, Stanford University. Townend J and Zoback MD (2000) ‘How faulting keeps the crust strong.’ Geology 28(5): 399–402. Wentworth CM and Zoback MD (1989) The style of late Cenozoic deformation at the eastern front of the California Coast Ranges. Tectonics 8: 237–246. Wessel P and Mueller RD (in press) Treatise on Geophysics, vol.6: Plate tectonics. Wiens DA and Stein S (1985) Implications of oceanic intraplate seismicity for plate stresses, driving forces, and rheology. Tectonophys 116: 143–162. Wortel MJR, Remkes MJN, Govers R, Cloetingh SAPL, and Meijer PT (1991) Dynamics of the lithosphere and the intraplate stress-field. Philosophical Transactions of the Royal Society of London Series A 337: 111–126. Wu P and Johnston P (2000) Can deglaciation trigger earthquakes in North America? Geophysical Research Letters 27: 1323–1326. Zoback ML (1992) First- and second-order patterns of stress in the lithosphere: The World Stress Map project. Journal of Geophysical Research 97: 11703–11728. (The current version of the World Stress Map database can be found at: http://www-wsm.physik.uni-karlsruhe.de/).
273
Zoback MD (2000) Strength of the San Andreas. Nature 405: 31–32. Zoback MD, Apel R, Brudy M, et al. (1993) Upper crustal strength inferred from stress measurements to 6 km depth in the KTB borehole. Nature 365: 633–635. Zoback MD and Harjes HP (1997) ‘Injection induced earthquakes and crustal stress at 9 km depth at the KTB deep drilling site, Germany.’ Journal of Geophysical Research 102: 18477–18491. Zoback MD and Haimson BC (eds.) (1983) Hydraulic Fracturing Stress Measurements, US National Committee for Rock Mechanics, 270. Washington, DC: National Press. Zoback MD and Healy JH (1984) ‘Friction, faulting, and in situ stresses.’ Annales Geophysicae 2: 689–698. Zoback MD and Healy JH (1992) In situ stress measurements to 3.5 km depth in the Cajo’ n Pass scientific-research borehole: Implications for the mechanics of crustal faulting. Journal of Geophysical Research 97: 5039–5057. Zoback ML and Mooney WD (2003) Lithospheric buoyancy and continental intraplate stress. International Geology Review 45: 95–118. Zoback MD, Moos D, Mastin L, and Anderson RN (1985) Wellbore breakouts and in situ stress. Journal of Geophysical Research 90: 5523–5530. Zoback ML, Nishenko SP, Richardson RM, Hasegawa HS, and Zoback MD (1986) Mid-plate stress, deformation, and seismicity. In: Vogt PR and Tucholke BE (eds.) The Geology of North America, vol. M: The Western North Atlantic Region, pp. 297–312. Boulder, CO: Geological Society of America. Zoback ML and Richardson RM (1996) Stress perturbation associated with the Amazonas and other ancient continental rifts. Journal of Geophysical Research 101: 5459–5475. Zoback ML and Thompson GA (1978) Basin and Range rifting in northern Nevada: Clues from a mid-Miocene rift and its subsequent offsets. Geology 6: 111–116. Zoback MD and Townend J (2001) Implications of hydrostatic pore pressures and high crustal strength for the deformation of intraplate lithosphere. Tectonophysics 336: 19–30. Zoback MD, Townend J, and Grollimund B (2002) Steady-state failure equilibrium and deformation of intraplate lithosphere. International Geology Review 44: 383–401. Zoback ML and Zoback MD (1980) State of stress of the conterminous United States. Journal of Geophysical Research 85: 6113–6156. Zoback ML and Zoback MD (1989) Tectonic stress field of the conterminous United States. Geological Society of America Memoir 172: 523–539. Zoback ML, et al. (1989) Global patterns of tectonic stress. Nature 341: 291–298. Zoback MD, Zoback ML, Mount VS, et al. (1987) New evidence on the state of stress of the San Andreas fault system. Science 238: 1105–1111. Zoback MD and Zoback ML (1991) Tectonic stress field of North America and relative plate motions. In: Slemmons DB, et al. (eds.) The Geology of North America, Neotectonics of North America, pp. 339–366. Boulder, CO: Geological Society of America.
Relevant Websites http://mahi.ucsd.edu – UCSD Department of Mathematics. http://www-wsm.physik.uni-karlsruhe.de – World Stress Map Project.
6.07
Magmatism, Magma, and Magma Chambers
B. D. Marsh, Johns Hopkins University, Baltimore, MD, USA ª 2007 Elsevier B.V. All rights reserved.
6.07.1 6.07.2 6.07.2.1 6.07.2.2 6.07.2.3 6.07.3 6.07.3.1 6.07.3.2 6.07.3.2.1 6.07.3.3 6.07.3.4 6.07.4 6.07.4.1 6.07.4.2 6.07.5 6.07.5.1 6.07.6 6.07.6.1 6.07.6.2 6.07.6.3 6.07.6.4 6.07.6.4.1 6.07.6.4.2 6.07.6.4.3 6.07.6.4.4 6.07.6.4.5 6.07.6.5 6.07.6.5.1 6.07.6.6 6.07.7 6.07.7.1 6.07.7.2 6.07.8 6.07.9 6.07.10 6.07.10.1 6.07.10.2 6.07.10.2.1 6.07.10.2.2 6.07.10.3 6.07.10.4 6.07.10.5 6.07.10.6 6.07.10.6.1 6.07.10.6.2 6.07.10.7
Introduction The Nature of Magma Transport Characteristics Phase Equilibria Solidification Fronts Crystals in Magma Solidification Front Crystallization or Phenocryst-Free Magmas Phenocryst-Bearing Magma Kilauea Iki Lava Lake Primitive versus Primary Magmas Historical Note on Solidification Front Fractionation Magma Chambers The Problem: The Diversity of Igneous Rocks George Becker’s Magma Chamber Historical Setting Life Time Lines Initial Conditions of Magmatic Systems Cooling from the Roof Style of Crystal Nucleation and Growth The Critical Connection between Space and Composition The Sequence of Emplacement or Delivery of the Magma Forms of magmatic bodies Internal transport style Eruptive timescales and fluxes Filling times Magmatic deliveries, episodes, periods, and repose times Thermal Ascent Characteristics and The Role of Thermal Convection Superheat Summary of Magmatic Initial Conditions End-Member Magmatic Systems The Sudbury Igneous Complex (SIC) Ferrar Dolerites, Antarctica Lessons Learned from Sudbury and the Ferrar Dolerites Ocean Ridge Magmatism Island Arc Magmatism Introductory Arc Form Spacing of the volcanic centers Arc segmentation Character of the Volcanic Centers Magma Transport Subduction Regime Subducting Plate Internal State Thermal regime Hydrothermal flows The Source of Arc Magma
276 277 278 279 279 281 281 284 285 288 289 289 289 290 292 293 295 296 296 297 300 300 300 301 301 302 302 303 305 306 306 309 313 314 316 316 317 317 318 318 320 320 321 321 322 322
275
276
Magmatism, Magma, and Magma Chambers
6.07.10.7.1 6.07.10.8 6.07.10.9 6.07.10.10 6.07.11 6.07.11.1 6.07.11.2 6.07.11.3 6.07.11.4 6.07.11.5 6.07.12 References
Slab quartz-eclogite Diapirism, Rayleigh–Taylor Instability, and Volcano Spacing Alkali Basalts Posterior to Arcs The Arc Magmatic System Solidification Front Differentiation Processes Introduction Solidification Front Instability Silicic Segregations and Crust Reprocessing in Iceland Sidewall Upflow Fissure Flushing Magmatic Systems
6.07.1 Introduction Magmatism is directly linked to tectonism. Where there is no tectonic activity, as in continental shield regions, there is no magmatism. Tectonic activity is a clear sign of convection in the mantle and often the crust, although the associated length scales may be much different. Relative to conduction of heat, convection moves material at rapid rates. This manifests itself in the inability of rock to cool easily during convection, which, coupled with the condition that much of the lower crust and mantle is already near melting, brings on melting as rock is adiabatically convected through a solidus. Convection through a phase boundary is the major process of producing magma on Earth. This form of melting is progressive, and instabilities associated with the field of affected rock, which depend on the size of the field, the degree of melting, and the nature of the melt itself, simultaneously arise to collect and transport the magma upward to Earth’s surface. Most magmas thus arise through partial melting of source or parent rock. The collected magma may contribute to volcanism or, perhaps more often, become stalled at depth in plutonism. This overall process is called ‘magmatism’. This is the process that has given rise to the diversity of rocks on Earth’s surface and perhaps also to the very structure of Earth itself. Within any process of magmatism there are certain physical and chemical processes that are truly fundamental to shaping the behavior and outcome of the magma and it is these that are considered herein. Volcanoes on a planet reflect the physical processes at depth of magma production, collection, and transport. This is in contrast to magma produced in situ through an externally applied heat source as in partial melting of granitic continental crust by
324 325 326 326 326 326 326 328 328 329 330 331
emplacement of higher-temperature basaltic magma in underplating (Bergantz, 1989) or prolonged heating of wall rock on the conduit of a volcanic system. Meteorite impact is also an effective means of magma production. The Sudbury melt sheet of Ontario, some 35 000 km3 of magma, was produced in 5 min by a 10–12 km bolide 1.85 Ga. Wholesale melting of continental crust produced a magma superheated to 1700 C, which is never found in endogenetic magmatism. Only the prolific volcanism of Io, due to viscous dissipation in tidal pumping by Jupiter, is similarly superheated. The most voluminous and steady magmatisms of Earth, like those of the ocean ridges and Hawaii, erupt magma at or near the liquidus, but never superheated. The low crystallinity of these magmas reflects this high-temperature eruptive state, whereas many island arc magmas, especially the andesitic ones, can be of high crystallinity; the most dangerous ones flirt with the point of critical crystallinity at 55 vol.%. As the phenocryst content approaches maximum packing at critical crystallinity the magma becomes a dilatant solid, expanding upon shear. The volcano becomes, in effect, corked or plugged, and can only erupt explosively. Hightemperature, low-crystallinity basaltic magmas are not generally explosive, but the exceeding mobility of the lavas is dangerous. Cinder cone volcanism, commonly areally sporadic and associated with alkali basalts, is just the reverse. The early phase of volcanism is explosive, almost regardless of crystal content, but the associated lavas are sluggish and immobile. Explosiveness can thus reflect an enhanced volatile content, high crystallinity, or high silica content. The pattern and style of magmatism intimately reflect the nature of the causative process. Globally widespread magmatism, as on Io, reflects planetwide melting as in tidal pumping or a heavy impact flux.
Magmatism, Magma, and Magma Chambers
Linear arrays or strings of magmatism, as along ocean ridges, island arcs, and some ocean islands, reflect melting associated with thermoconvective flows tightly focused within the phase field of the source rock. That is, either the thermal regime or the source material (or both) is spatially focused. Areally widespread, small-volume magmatism as, for example, that of cinder cone fields, reflects a pervasive, marginally focused convective flow, much like a broad low-pressure system in the atmosphere, where melting instabilities are mainly due to local irregularities in the source rock detailed composition. Typical convective flows of this nature are the broad, gently upwelling flows associated with the wedge flows driven by subduction. Widespread, but voluminous magmatisms, as in the silicic volcanism of the western United States over the past 50 My, reflect convective stretching and destabilization of the continental lithosphere. The asthenosphere thermal regime is, in effect, brought to the Moho where vast regions of continental crust of irregular thickness are accessible to melting. Volcanoes themselves, as opposed to impact melt sheets, reflect deep-seated, endogenetic processes, and in a simple way the volcanic edifice is a measure of the hydrostatic (i.e., magmastatic) head of the system. The size of volcanoes also reflects the activity or strength of the system, the mobility of the source relative to the surface plate, the strength (and thus age) of the local lithosphere, and the intensity of the gravitational field. The largest volcanoes in the solar system are in the Tharsis Bulge region of Mars. Olympus Mons is 30 km tall and 850 km in basal diameter. Martian gravity is weak, the lithosphere is very old and strong, and the source has evidently been immobile relative to the surface plate. For a given areal extent, the volcanoes of Venus are the shortest, reflecting the high-temperature, weak nature of the Venuisian lithosphere. This is broadly similar to the volcanoes within the rift zones of Iceland, where isostasy is rapid (1 km My1) due to the thin, hot, and weak crust. The plumbing of magmatic systems begins in the source region and ends at a volcano or pluton. Knowledge of the structure of volcanoes and plutons comes mainly from direct observations of field relations in deeply eroded terrains. The deeper source regions can be sensed, in terms of depth, degree of melting upon extraction, and source material, through geochemistry. Detailed knowledge of the physical state of these systems comes mainly from modeling and rare, well-exposed crustal rocks. Hypocenter distributions, particularly in active,
277
voluminous systems, like Hawaii, are valuable in inferring the deeper geometric form, extent, and sense of connection between eruptions and deep transport. Some styles of harmonic tremor, which sometimes occurs during eruptive stages, seem to indicate the pre-eruptive state of magma held in vertically oriented cylindrical conduits. Ground deformation associated with tumescence prior to eruption consistently indicates through elastic modeling a local, near-surface staging region, which is commonly identified with the concept of a magma chamber. This is reinforced by the very nature and pervasiveness of plutons. Overall, the most commonly held conceptualization of a generic magmatic plumping system has a deep source region linked to a near-surface magma chamber through a poorly discerned transport region. Aside from the gross inferences from seismic studies, the intervening plumbing linking source to near surface, the ascent path, has been difficult to ascertain. Although magma perhaps spends most of its life in this part of the system, it has, per force, been largely ignored physically and chemically due to lack of a clear and detailed conception of this important feature. This basic magmatic architecture, source, ascent path, chamber, and pluton or volcano, is found in different forms in most magmatic systems (see Figure 1). Magma chamber is a particularly important concept. Volcanologists use the mechanics of magma chambers to handle all the major and subtle chemical and textural transitions necessary to link one lava to another. Successive lavas may give a time record of the operation of a magma chamber, although the actual chamber can never be sensed. Plutons are, in a strong sense, actual magma chambers, but the sense of how they operated in real time has been largely lost through protracted crystallization and annealing, and the nature of the connection of plutons to either volcanoes or to the deeper, ascent path of the system is not at all clear.
6.07.2 The Nature of Magma Magma is a viscous fluid consisting most often of polymerized silicate melt, crystals, dissolved gasses, and sometime bubbles and foreign chunks of crystals and rocks. Of all the fluids provided by Nature, magma is perhaps the most intriguing. In traversing the melting range, a span of 200 C, magma viscosity increases by a factor of 1018 (MKS or CGS), which is the largest change of any physical parameter for this temperature change. Crystals spontaneously
278
Magmatism, Magma, and Magma Chambers
Magmatic mush column
Solidification time (with latent heat)
Thickness (m)
10
102
103
1
10 102 103 104 105 Time (years)
Figure 1 Magmatic systems (left) are an integrated collection of sills and connecting conduits linking a source region to the near surface and volcanic centers. The thermal timescales (right) for solidification vary throughout the system due to the local thermal regime and the length scale of the system.
nucleate and grow, modifying both the local buoyancy and melt composition, furnishing swarms of particles of a wide range of sizes that sort by size and density and stick together, welding tightly to form strong, truss-like, networks at many degrees of crystallinity.
6.07.2.1
Transport Characteristics
Heat and mass transport are slow in magma. Thermal (K ) and chemical (D) diffusivities are small (typically 102 and 106 cm2 s1, respectively) and kinematic viscosity (v) is large, making the Prandtl (Pr ¼ v/K), Schmidt (Sc ¼ v/D), and Lewis (Le ¼ K/D) numbers each large as opposed to molten metals with small Pr. That is, vorticity, heat, and mass can each be described by an equation of the form qA q2 A ¼ B 2 qt qXi
½1
where A is vosticity, temperature, or mass concentration, B is kinematic viscosity, thermal diffusivity, or mass diffusivity, and Xi is a spatial dimension. A scaling analysis of this equation relates the characteristic distance () of diffusion associated with each of these processes to the transport property B and time (t), yielding ¼ CðBt Þ1=2
½2
where C is a constant near unity in magnitude. The dimension is a measure of the diffusive layer thickness during the flow of momentum, heat, or mass. And the relative thicknesses of these layers is thus measured by v v ¼ Pr ðPrandtl no:Þ T K
½3
v v ¼ Sc ðSchmidt no:Þ D D
½4
D D ¼ Le ðLewis no:Þ T K
½5
Notice also that Pr ¼ Sc Le. As an example of the value of these relative measures, for magma with a kinematic viscosity of 102 cm 2 s1 moving 1 km along a wall at the rate of 0.1 cm s1, the momentum boundary thickness will be about 400 m. The thermal boundary layer will have a thickness of about 4 m, and the mass boundary layer thickness will be about 4 cm. With Pr being a measure of the relative thicknesses of viscous or momentum to thermal boundary layers, momentum boundary layers are much thicker than thermal boundary layers, making the boundary layer approach for momentum transfer in fluid mechanics generally not a useful approach, unlike in metallurgy where Pr is small. Thermal boundary layer approaches, on the contrary, are highly valuable. Similarly, from the Schmidt and Lewis numbers, chemical boundary
Magmatism, Magma, and Magma Chambers
10
Eruption trajectories
4
Figure 2 The types of boundary layers and their characteristic length scales operating in magmatic systems.
us Liquid ent ic asc
Wet magma
entrop
c
6
Dry magma
Nonis
C(x)
Pressure (kb)
x
Liquid us
8
T(x)
Solidu s
z
Isentropic ascent
Boundary layers Chemical Chemical and thermal
Solidu s
Thermal
279
2
layers due to chemical diffusion are much thinner than both momentum and thermal boundary layers (see Figure 2). Diffusion halos about growing crystals, for example, are exceedingly thin and contamination by xenoliths affect only narrow margins. In many ways, this makes solving magmatic transport problems straightforward. Thin thermal and chemical boundary layers are embedded in wide shear zones where the variation in velocity is broad and gentle. We shall soon see, however, that the unusually strong effects of variable viscosity, as mentioned already, partly mitigate this attractive feature of many magmatic flows. Some of this viscosity effect comes from the increasing silica content of the melt as crystallization proceeds, but by far the most serious effect comes from the increasing concentration of crystals themselves and the interactions among these solids. These characteristics affect all aspects of magma generation, transport, differentiation, and eruption.
6.07.2.2
Phase Equilibria
The melting behavior of a typical basaltic magma, in terms of phase equilibria, as a function of pressure is shown by Figure 3. The high-temperature boundary beyond which (in temperature) no crystals exist is the ‘liquidus’ and the lower boundary below which the magma is completely solid, is the ‘solidus’. In between these boundaries are phase fields delineating the stability of various mineral phases. Two magmas are indicated, one is free of dissolved water (dry magma) and the other (wet magma) contains a few percent dissolved water, where the point of saturation occurs at about 2 kb (0.2 GPa, 200 MPa). The solubility of water (and CO2 and SO2, the other principal volatile
0 900
1000
1100 1200 Temperature (°C)
1300
Figure 3 The general phase equilibria for basaltic magma as a function of pressure for dry and wet magmas. The adiabatic ascent path is indicated for the dry magma and also is the effect of rapid cooling due to superheating and convection. For the wet magma, the near surface liquidus may be below the 1-atm solidus temperature.
species) is directly proportional to the confining pressure, or depth in Earth, and there is also a point (10– 15 kb) at which no more water can be dissolved in the magma. At this point, the magma contains about 25 wt.% water, which is equivalent to the molar proportion of water in seawater (96%). Magma in this condition, should it ever be achieved, is an aqueous solution with a silicate solute. At pressures above the indicated point of saturation at 2 kb, the wet magma is undersaturated with water and the phase diagram is geometrically similar to that of the dry magma, except that it is at a significantly lower temperature. This difference, as will be treated later, has a profound effect on the characteristics of eruption. 6.07.2.3
Solidification Fronts
Between the liquidus and solidus the magma is a physical and chemical mixture of liquid (melt) and crystals (and perhaps bubbles). The properties of this mixture are complex and exceedingly important at every stage of a magma’s life in determining the set of dominant physical and chemical processes involved in shaping the magma’s behavior. The buildup of
280
Magmatism, Magma, and Magma Chambers
crystals with decreasing temperature between the liquidus and solidus for typical basalts is shown by Figure 4. Nucleation and growth of crystals begins at the liquidus, by definition, and with decreasing temperature crystallinity (, or crystal fraction) at first builds up slowly and then more rapidly as crystallinity reaches 50 vol.% after which it again builds increasingly slowly with approach to the solidus. Near the liquidus, where crystallinity is low and crystals are small, the magma is a ‘suspension’. Here the crystals can settle relative to one another without much hindrance from one another, but since they are small, settling is slow. This state holds to a crystallinity of about 25 vol.% where the viscosity of the
Crystallinity (vol.%)
100
0 980
Magma Mush zone
Rigid crust
Suspension zone Capture front
T (°C)
1210
Figure 4 The variation of crystallinity in basaltic magma at any pressure as a function of temperature.
mass (see below) has increased by a factor of 10 over that at the liquidus and the crystals now ‘feel’ the presence of one another as they move. Motion of a single crystal causes a shear flow extending out 10 radii which entrains neighboring crystals. There is also some observational (Marsh, 1998) and experimental (Philpotts and Carroll, 1996) evidence that, especially in plagioclase-bearing basaltic magmas, a ‘chicken-wire’ network of crystals may develop that gives a certain structure and strength to the magma. Beyond the suspension zone where crystals are (ideally) still separated from one another is the ‘mush’ zone, which persists until the crystallinity reaches the state of maximum packing where the crystals are all touching. This occurs when crystallinity reaches about 55 vol.% ( 0.55). At this point, the crystals are tacked together to form a solid framework of some strength (e.g., Marsh, 2002). In essence, the magma is now a rock and can no longer flow as a viscous fluid. The remaining melt, which is interstitial to the crystals, can still move, but as a porous or Darcian flow (e.g., Hersum et al., 2005). Because of this overall strength, which is found in Hawaiian lava lakes to be drillable, this region is called the ‘rigid crust’. These three zones, then, suspension, mush, and rigid crust, make up all magmatic ‘solidification fronts’. A solidification front (see Figure 5) is the
Tsolidus
Rigid crust 55% crystals
Mush zone Capture front 25% crystals
Suspension zone Tliquidus Crystal-free magma
Figure 5 Upper solidification front propagating inward (down). The base or outermost edge of the front is defined by the solidus and the leading, innermost edge is the liquidus. The overall thickness of the front increases with time and the rate of thickening depends on the local thermal regime. The inset depicts the framework-like structure of the crystal at about 30% crystals.
Magmatism, Magma, and Magma Chambers
active zone of crystallization that forms the perimeter of all magmas. Solidification fronts are dynamic in the sense that they continually move in response to the prevailing thermal regime or state of heat transfer from (or sometimes to) the magma. All crystallization, by definition, takes place within solidification fronts. And it is within these fronts that physical and chemical processes take place to modify magma composition and texture and the behavioral character of the magma is determined. Beyond the solidification fronts, in the magma interior, there are no new crystals and hence no processes operate there to separate crystals and melt (more below). As magma approaches the point of maximum crystal packing, where f 0.55, it undergoes a dramatic rheological transition. It becomes a dilatant solid, which means that the mass of crystals and melt upon being sheared expands as neighboring solids try to move outward and around one another to accommodate the shear. When this condition is reached for magma in the throat of a volcano, the volcano becomes plugged and can no longer emit lava. With continued cooling and crystallization in the magma below, gases exsolve that can build pressure to the point of catastrophic failure of the entire edifice. This condition will be discussed again shortly. The dynamics of solidification fronts measure the dynamics, especially the rate of cooling, of every magma from deep chambers to lava flows.
6.07.3 Crystals in Magma All magmas can be separated into two groups according to the sources of the constituent crystals: (1) crystals grown in the present cycle of active solidification fronts, and (2) crystals inherited or entrained from prior crystallization events, including crystals from disaggregated wall rock. It is fundamental to the study of every magmatic rock to recognize the provenance of the constituent crystals. Crystals have historically been separated texturally into phenocrysts and groundmass. Phenocrysts are unusually large or otherwise distinctive crystals and groundmass is commonly tiny crystals interstitial to any phenocrysts. The distinction is most apparent in volcanic rocks, but is also often clear in many plutonic rocks and the textures are called porphyritic. It has long been assumed (e.g., Turner and Verhoogen, 1960; Carmichael et al., 1974) that phenocrysts represent crystallization at depth under slow, protracted cooling, and groundmass crystals the latest rapid
281
phase of cooling associated, in volcanic rocks, with eruptive processes or emplacement processes in plutons. The apparent sharp change in crystal size between phenocrysts and groundmass upon quantitative examination using measured crystal size distributions (CSDs) shows that most crystal population are actually continuous in size from phenocrysts to groundmass. This feature is characteristic of active crystallization where a time span of active nucleation has given rise to large crystals. Yet there are many magmas that have inherited crystals from earlier solidification events and these crystals are valuable indicators of dynamic processes involving crystal entrainment and transport. Recognizing the distinction between these two classes of crystals is of critical importance to understanding magmatic processes. Yet, it has proven elusive in some magmatic bodies to recognize imported crystals, which has led to faulty reasoning in deciphering magmatic history. These two regimes, which for convenience are hence forth referred to as phenocryst free and phenocryst bearing, are next considered in some detail to emphasize this importance, beginning first with crystallization controlled by solidification fronts.
6.07.3.1 Solidification Front Crystallization or Phenocryst-Free Magmas Solidification fronts are packaged between the liquidus isotherm at the leading edge and the solidus isotherm at the trailing edge. Nuclei and superclusters already present begin to grow in earnest at the leading edge of the solidification front and these crystals continue to grow as the front passes through the region of melt. In the strictest sense, the melt is stationary and the solidification front moves through the melt transforming it into a crystalline mass. In the leading part of the solidification front, the suspension zone, crystals can settle easily without hindrance, but individual crystals are small and settle slowly. This settling is well described by Stokes’ law: Vs ¼ Cs
ga2
½6
where CS is a constant (2/9) depending on the shape of the crystal, is density contrast, g is gravity, a is crystal radius, and is viscosity. To escape the solidification front, the crystal must settle faster than the rate of advancement of the solidification front itself. Early in the cooling history, especially for magma emplaced in cool upper crustal
282
Magmatism, Magma, and Magma Chambers
wall rock, solidification fronts move rapidly and it is impossible for newly formed crystals to escape. Only in very slow-moving solidification fronts are crystals sometimes able to escape. And if once a crystal becomes deeply embedded in a solidification front, melt viscosity and hindrance from neighboring crystals increase to the point that escape is not possible. The region where this transition occurs marks the boundary between the suspension and mush zones and is called the ‘capture front’. Crystals outward of the capture front are trapped in the solidification front. As cooling proceeds, the distance between these isotherms defining the solidification front increases with the square root of time, just as in any conductive process. That is, even though solidification fronts involve heat production from latent heat of crystallization, from solutions to Stefan-type problems (e.g., Carslaw and Jaeger, 1959; Mangan and Marsh, 1992; Turcotte and Schubert, 1982) the position S(t) of the front is given exactly by: S ðt Þ ¼ 2b ðKt Þ1=2
½7
where b is a constant and K is thermal diffusivity. The rate of advance is given by the time derivative of this equation. VF ¼
1=2 dS ðt Þ K ¼ b dt t
½8
Equating this with the rate of crystal settling from Stokes’ law gives a relation for minimum size crystal that can escape from the solidification front as a function of time. That is, "
b ðK =t Þ1=2 a ðt Þ ¼ Cs g
" #2=5 CP ðTm – Tw Þ 5H ðÞ
1=2
S ðt Þ ¼ 2bfCF ðKt Þ1=2
S ðt Þ ¼ ðl – fCF Þ2b ðKt Þ1=2
½10
where CP is specific heat, Tm is magma initial temperature, Tw is wall rock initial temperature, and H is latent heat. This measure of crystal size must be compared with the actual size crystal growing in the suspension
½12
The time spent traversing this zone is given by the quotient of [12] and [8], t ðt Þ ¼
S ðt Þ ¼ 2ðl – fCF Þt VF ðt Þ
½13
The time for the solidification front to traverse the suspension zone is a linear function of time and independent of the thermal properties of the system. The size (a(t)) of a typical crystal grown during this time can be found using a general growth law for crystal growth. Although this choice is completely arbitrary, a convenient and realistic formula is the linear relation (e.g., Marsh, 1998; Zieg and Marsh, 2002), aðt Þ ¼ G t ðt Þ ¼ 2G ðl – fCF Þt
½14
Solving this equation simultaneously with that given by eqn [9] gives the time at which the first crystals escape from the solidification front, "
½9
½11
and the difference of these two equations gives an estimate of the thickness of the suspension zone
#1=2
The larger the viscosity the larger the crystal must be to escape, and for density contrast it is just the reverse. The constant b measures the effect of latent heat relative to the enthalpy of the magma, and can be adequately represented from its original transcendental equation as (e.g., Zieg and Marsh, 2002) b ¼
zone as the solidification front thickens. This can be found by estimating how long taken by the capture front isotherm to traverse the suspension zone as a function of time. The position of the capture front isotherm is a large fraction ( fCF 0.95–0.98) of position of the leading edge of the solidification front given by eqn [7],
t ¼
ðb =CS ÞK 1=2 ð2G ðl – fCF ÞÞ2 g
#2=5 ½15
This time can also be inserted in to the growth law of eqn [14] to yield the size of the crystal at the time of escape, which is " aðt Þ ¼
ðb=CS Þð2Gðl – fCF ÞÞ1=2 K 1=2 g
#2=5 ½16
The time of escape and size of crystal escaping are shown by Figure 6. The time of escape is sensitive to the growth rate. For a magma viscosity of 103 P where crystals are growing at 1010 cm s1, escape begins after about 30 years and the crystal size (radius) is about 0.03 cm. As might be expected, the size of the crystal at first escape is relatively insensitive to growth rate. The time of escape is much more sensitive to
Magmatism, Magma, and Magma Chambers
(a)
Time of escape (years)
104
103
ms
te (c
102
–12
0
–1 ) = 1
h ra owt
–11
10
Gr
–10
10
10
1 10
103 104 102 μ, Magma viscosity (CGS)
105
(b)
Crystal size, a (cm)
1
10–1
–8
wth
Gro
0 –1 ) = 1 –9 10 –10 ms c ( 10 rate
10–2
10–3 10
102
103
104
105
μ, Magma viscosity (CGS) Figure 6 (a) The time of escape of a crystal from a solidification front as a function of magma viscosity and crystal growth rate. (b) The size of the first escaping crystal as a function of magma viscosity and crystal growth rate.
283
growth rate, which is understandable. The sooner a crystal can reach the critical size to sink faster than the rate of advancement of the capture front, the sooner it can escape. As magma viscosity increases, the time of first escape and the crystal size at escape each increase. For granitic magmas where the viscosity might be 105–106 P, escaping crystals must be about 1 cm in radius, which takes for the growth rates indicated on the order of 103–104 years. This indicates why differentiation by crystal escape from solidification fronts is unlikely in granitic magmas. Crystal escape from solidification fronts is much more likely in basaltic magmas once the crystals reach about 0.1 mm in size, which can occur after about 30 years when the solidification front has a thickness of about 50 m, depending on the cooling regime. The question then arises what happens to these crystals once they escape the solidification front? Since the magma below the advancing front is the hottest part of the system, crystals entering this region will begin melting and dissolving back into the magma (see Figure 7). This process was examined by Mangan and Marsh (1992), who, through a different approach, found results similar to those above. Fractionation begins after about 25 years when capture front advancement has slowed to 1–2 m yr1. For intrusions thinner than about 30 m, this condition is never achieved and no escape is possible, but for magmatic sheets thicker than about 100 m this process may often occur. They also followed the course of the sinking crystals into the magma interior and estimated the time and distance settled for complete
0
Dimensionless depth, d *
Crystals resorbed before reaching lower crust
0.2
Nuc
Cap
leat
ion
0.4
ture
fron
Crystals reach lower crust
front
Solidification front *
t
*
Variable viscosity
Interior uncrystallized magma 0.6 ion
leat
Nuc
0.8
t fron
ture
Cap
*
t fron Crystals resorbed before reaching lower crust
* Constant viscosity Crystals reach lower crust
1.0 0
0.2
0.4 0.6 Dimensionless time, t *
0.8
1.0
Figure 7 The thickening of upper and lower solidification fronts in a sheet-like body and their relationship to capturing settling crystals.
284
Magmatism, Magma, and Magma Chambers
resorption. Survival through the hotter interior depends also on the proximity of the lower solidification front rising from the floor of the magma. Crystal accumulation on the floor of the magma chamber is not possible until the lower solidification front has advanced to intercept the descending crystals prior to resorption. Interception is unlikely until the magmatic sheet is about 75% solidified, where the depth of interception is about midway through the original magma (Figure 7). At best, a microcumulate layer may develop near the centerline of the body. The net result of crystallization of a magma initially containing only nuclei and no large phenocrysts is a body of rock of essentially uniform composition and with a crystal size that increases inward in accord with the local growth or transit time of the solidification front. Many dolerite (diabase) sills throughout the world ranging in thickness from 1 to 350 m are often of uniform composition with a grain size that increases gradually in response to the slowing of the solidification front as it moves into the body from top and bottom. Actual CSDs can be calculated in detail to characterize this process (see Zieg and Marsh, 2002).
6.07.3.2
differentiate the magma became apparent with the controversy over the origin of the sequence of rock at Shonkin Sag laccolith in north-centeal Montana. It is also from early studies of this body that came the basic concepts on magmatic processes assumed to control magmatic chemical evolution. Although much more on this crucial history will be covered later, it is important here to consider this body to set the stage for understanding this large class of magmas, which often lead to exotically layered bodies. Shonkin Sag is a relatively small circular laccolith, 70 m thick and 3 km in diameter. Yet, in spite of its small size, it is strongly differentiated (Figure 8). The obvious curiosity is that many much thicker and more areally extensive dolerite sills throughout the world, also of basaltic composition, show almost no differentiation. In the initial study of Shonkin Sag, Pirsson (1905) realized that a large mass of crystals had settled and formed a thick pile of cumulates on the floor. Many bodies show similar distributions of a thick pile of cumulate crystals on the floor. And most commonly it has been assumed, like Pirsson, that these crystals grew after emplacement in response to strong cooling from the roof of the sill and simply settled to the floor. When the body was restudied in the 1930s independently by C. S. J. Hurlbut (1939) at Harvard and J. D. Barksdale (1937) at Yale, a controversy sprung up. Hurlbut recognized that the initial magma carried a high population (35 vol.%) of large crystals prior to emplacement. He realized that these crystals did not grow after
Phenocryst-Bearing Magma
That some magmas carry significant loads of large crystals that suddenly settle when the magma comes to rest upon emplacement and strongly chemically
Density (g cm–3) 2.5
2.6
2.7
2.8
2.9
3.0
200
Upper solidification front (plus captured phenos.)
Height above base (ft)
Shonkinite 150
Initial magma stripped of phenos.
Syenite
Melt pillows from cumulates 100
Shonkinite cumulates
Lower solidification front plus phenocrysts (cumulates)
50
0 0
20 40 60 % Mafic minerals
80
Pipes carrying interstitial melt from compacting phenos. to transition zone
Figure 8 The stratigraphic field relations between rock types of Shonkin Sag laccolith.
Magmatism, Magma, and Magma Chambers
emplacement and did not fall to the floor to form the thick pile of cumulates. They fell to the floor because they came in as essentially an instantaneous injection of phenocrysts into the system. This may be the origin of the extremely important idea of the role of magma in carrying crystals. Barksdale, on the other hand, was deeply impressed by the presence of what appeared to be an internal contact where the transition zone rocks (see Figure 8) directly overlay the basal cumulates. This feature in the field does, indeed, look very much like an internal contact that might be formed by the injection of a separate magma. This was the point of contention: Hurlbut said the entire stratigraphic distribution of rocks could be formed in place from an initially crystalladen magma. Barksdale insisted that multiple injections of magma are necessary. These two points of view, in situ differentiation versus multiple injections, are still often points of contention in explaining igneous sequences. And, as will be seen here, there is clearly room and good reason for both views. In short, as the large load of phenocrysts was settling and accumulating in a thick pile on the laccolith floor, interstitial melt was forced upward through a series of pipe-like conduits (Marsh et al., 1991). These pipes carried the melt up, expelling it on top of the compacting cumulate pile. Being too dense to rise higher in the sequence, the expelled essentially crystal-free melt established a series of overlapping pillows of melt forming the transition zone. The expelled melt being in the hottest part of the body, and multiply saturated because of its intimate prior contact with the cumulates crystallized into unusually coarse crystals. The net result is a layer of rock in sharp contrast to the texture of the underlying cumulates of clinopyroxene phenocrysts. For all intents, this looks like an internal contact from a separate injection of magma. And it is, but this is an auto- or internal injection. Barksdale was very impressed by this, and his arguments centered in many facets on this observation. It is important to note that, whereas the basic geologic observation is correct, it is the interpretation of this feature in terms of a magmatic process that is faulty. This is a fairly common pitfall in magmatology, and it appears even more strongly when the meaning of a thick cumulate pile of crystals on the floor of a magmatic body is interpreted (see below). On more general grounds, when magma carrying phenocrysts is emplaced, there is a competition between the rate of crystal settling and the rate of solidification. Just as in the previous section when
285
considering crystal capture and escape by the solidification front, the same process operates here except with crystals distributed vertically throughout the body a large population of crystals escape capture and settle freely into the lower, upcoming solidification front. Consider some end-member situations for magma laden with crystals emplaced as a sheet shown by Figure 9. If the magma could cool instantaneously regardless of its thickness, which physically is impossible, all the crystals would be trapped in place before they could settle any distance at all. The final distribution of crystals would be uniform from top to bottom throughout the sheet. At the other extreme, if the body did not crystallize at all, all the crystals would settle to the floor and form a thick pile of constant abundance like a bed of clastic sedimentary rock. Although very thin (1 m) and very thick (hundreds of meters) sheets of magma may approximate these extremes, actual magmas are somewhere in between these two extremes. Inward advancing solidification fronts from the top and bottom compete with crystal settling. The initial advance of the upper and lower solidification fronts is infinitely rapid, which is what leads to the fine-grained chilled margins around igneous bodies, and any phenocrysts near the margins are immediately captured. But because the rate of solidification front advance decreases inversely with distance into the body (i.e., from eqn [2] an isotherm advance velocity (V ) can be formed that leads to V K/Z, where K is thermal diffusivity and Z is distance into the body), more and more crystals escape capture with depth and all crystals initially in the center of the body settle into the lower rising solidification front. The net result is an S-shaped profile of phenocryst abundance vertically through the body (see Figure 9). This basic form of phenocryst distribution is seen in many magmatic sheets and it is a clear indication of solidification of a phenocryst-charged body of magma. Moreover, there are many variations on this basic theme, and one readily accessible system of this nature that was studied in real time is Kilauea Iki Lava Lake in Hawaii. 6.07.3.2.1
Kilauea Iki Lava Lake Lava lakes form at Hawaii when erupting magma fills naturally occurring pits or sink holes due to collapsing deeper lava tubes. Kilauea Iki Lava Lake formed from the 1959 eruption of a crystal (olivine)-laden basalt, filling a 2-km-wide pit to a depth of about 125 m (e.g., Helz, 1986; Helz et al., 1989). Because the eruptive flux repeatedly waxed and waned during
286
Magmatism, Magma, and Magma Chambers
T Liquidus
Upper contact Upper crust
Depth
Capture front
Active magma Accumulation front Lower crust Lower contact
Depth
0.0
Infinitely fast cooling
0.5
Very slow cooling
Sill cooling
1.0 % phenocrysts
Depth (normalized to total depth)
Tasmanian diabase
% phenocrysts
% phenocrysts
Prehistoric Makaopuhi
Prospect intrusion
Shonkin Sag
0 0.2
Total depth =460m
Total depth =75m
Total depth =120m
Total depth =70m
0.4 0.6 0.8 1 2.8
2.9 3.0 0 Density (g cm–3)
10 20 % olivine
30 0
10 % olivine
20
20
40 60 % mafics
Figure 9 The process of settling of phenocrysts leading to formation of various final distributions of crystals. (a) A schematic depiction of the process. (b) The formation of idealized or end-member distributions. (c) Actual distributions of crystals observed in some intrusions.
filling, the load of large olivine crystals delivered to the lava lake also varied over time. Murata and Richter (1966) found that the magnitude of the eruptive flux at Kilauea is proportional to the size and the abundance of olivine crystals entrained by the erupting magma. In other words, the bigger the eruptive
flux and the stronger the eruption, the more and the larger the olivine carried by the erupting magma. This is like a river in flood stage. The bigger and more dramatic the flood, the bigger the bedload of solids that it can move, and the larger the size of the individual solids that can be moved. The olivines
Magmatism, Magma, and Magma Chambers
entrained are not necessarily phenocrysts but are all sorts of crystals that are found in the plumbing system; some are phenocrysts from earlier magmatic events, others are xenocrysts from peridotitic wall rock, and everything in between. When the crystals in a single sample are examined they go from Fo75 to Fo92, and the latter compositions are clearly from mantle wall rock (e.g., Maaloe et al., 1989, 1992). At Hawaii, these crystals are called ‘tramp’ crystals to reflect their perhaps vagrant and varied origins (e.g., Wright and Fiske, 1971). Tramp crystals are found in systems all over the world at, for example, Jan Mayen (Imsland, 1984), Reunion (Upton and Wadsworth, 1967), and the Ferrar dolerites (see below). Considering that all magmas everywhere must ascend and erupt through crystalline mantle and crust wall rock, which are in effect gravel piles, every magmatic system should contain tramp crystals. The surprising aspect of magma carrying this heterogeneous collection of crystals is that if the lava chemical composition is plotted on a conventional diagram of CaO versus MgO (in wt.%), the variation appears as that due to perfect fractionation of olivine crystals indigenous to the magma (see Figure 10). Wright (1971) has shown that when these relations are considered in more detail for individual eruptions, the variations due to olivine control are even cleaner and tighter. Wright calls these trends olivine control lines. That is, all of the
287
variation in MgO can be effected purely by the addition or loss of olivine tramp crystals. These exotic crystals have a strong effect on magma composition also through diffusional exchange. As soon as a tramp crystal enters the ascending magma, it begins to chemically exchange Fe and Mg with the melt. The melt is much poorer in Mg and richer in Fe than typical tramp crystals. The time for these crystals to come to equilibrium with the melt can be estimated from eqn [2]: t a2/D, where a is crystal radius (1 mm) and D is the chemical diffusivity controlling the exchange (106 cm2 s1). It takes about a month or so to erase the original compositional identity of the tramp crystal and bring it into equilibrium with the melt. The melt has lost Fe to the crystal and gained Mg, which has made the apparent magma composition ‘more primitive’, exactly opposite to a normal (and commonly assumed) liquid line of descent. Conversely, these tramp crystals to have maintained their exotic compositional identity, have been entrained in the melt for a very short time, perhaps only days. Given the eruptive compositional variability of the lavas, especially in terms of the entrained crystal compositions, the magma filling Kilauea Iki Lava Lake varied daily during the eruption. The initial conditions in the lava lake sheet of magma, in terms of the abundance of tramp crystals, is in striking contrast to Shonkin Sag. At Shonkin Sag, the magma ascended from depth and the crystals were
15
Kilauea Iki Lava Lake 12
wt.% CaO
Eruption samples 9 Segregation veins 6 Olivine control lines 3
Solidification front All interstitial glasses (Helz et al.,1989)
0
0
6
12
18
24
30
wt.% MgO Figure 10 Variation of CaO and MgO in samples from Kilauea Iki Lava Lake. Note the olivine control lines, which show the effects of gains and losses in entrained olivine crystals. All compositions at MgO less than about 7 wt.% are from drilling of the upper solidification front. Individual symbols represent whole rock analyses of eruption samples (see, e.g., Helz et al., 1989).
288
Magmatism, Magma, and Magma Chambers
presumably flow sorted and mixed prior to emplacement and the sequence of crystal settling leading to the final S-shaped profile. But at Kilauea Iki the S-shaped profile will also reflect detailed variations due to the eruptive filling process (see Figure 11). Similar small-scale variations in intrusive complexes, as is seen in the Basement Sill of the Ferrar dolerites of the McMurdo Dry Valleys, thus reflect periodic filling and repose times and may be a critical link between the similarity of intrusive and volcanic processes of magma transport. Once the tramp olivine crystals have settled to form a cumulative pile on the floor of the lava lake, they begin to diffusionally exchange with the interstitial melt left in the bed and in a matter of a few months they have lost their original compositional heritage. A later analysis of this series of rocks would be difficult to unravel the true magmatic history due to this late, post-emplacement, re-equilibration. The clearest record of the role of tramp crystals, and their compositional diversity, is best recorded in the rapidly chilled rocks at the upper margin. The sequence of chemical differentiation is controlled by two key processes. First is the loss of the load of entrained tramp crystals once the magma came to rest. The magma in the center of the body after loss of the tramp crystals is a melt containing only tiny crystals that differentiated from 20% MgO to 7 wt.% MgO. It is essential to note that the change in composition took place, in essence, instantaneously as the tramp crystals settled out. This is ‘punctuated crystal fractionation and differentiation’. Second is the fact that after loss of the tramp crystals, the center of the lava lake contains only Kilauea Iki
Total depth = 100 m
0
10
20
30
MgO (wt.%) Figure 11 The S-shaped profile for Kilauea Iki where the detailed variations reflecting variations in eruption flux are clear.
crystal-free melt, and the system now resembles a phenocryst-free system discussed earlier. The system continues to cool and solidify without further differentiation. And if by chance this magma were to be erupted, only the low-viscosity, crystal-free core melt would be eruptible. The ensuing lava would thus show extensive chemical signs of differentiation, but it would mainly be due to the role of the entrained tramp crystals. The dramatic contrast in evolution of magma with and without initial concentrations of large crystals is central to understanding all magmas. Yet the importance of this distinction, and even the recognition of this feature itself, was not appreciated in the formative stages of magmatology. The essential concepts of differentiation and, especially, the evolution of magma in chambers were conceived and put into scientific play, in many ways, on faulty reasoning. The field relations were clearly recognized, but the physical processes giving rise to them were unknown. Through these considerations is where the concept came about of magma chambers, undoubtedly the most universal concept in magmatology. 6.07.3.3
Primitive versus Primary Magmas
When considering a magmatic complex or system, petrologists are literally obsessed with identifying the magmatic composition from which all other related compositions in the suite could have arisen by crystal fractionation processes. Norman Bowen himself, as we will see below, incited this obsession by discovering the crystallization reaction series whereby, starting with basalt, granite could be formed by growing and progressively settling out crystals with silica contents less than that of the adjacent magma. The natural tendency has thus been to automatically take the rock with the highest MgO content and assume it is the primary magma of the series. Magma MgO content has become synonymous with primitiveness. This reflects the direct connection between all basaltic magmas and mantle peridotitic rock. After all, there is no doubt whatsoever that basalts generally come from partial melting of the mantle and every other igneous rock, continents included, comes from basalt. It also reflects the observation that at Hawaii there is a clear succession from MgO-rich picrites right on through to tholeiitic basalts with 5–7 wt.% MgO. It is naturally assumed and inherently adhered to, without even having to defend the proposition, that any basaltic rock with low MgO must be a differentiate. It cannot be a primary composition.
Magmatism, Magma, and Magma Chambers
The general fallacy of this line of reasoning comes from not recognizing the fundamental mechanical role of tramp crystals in deciding magma bulk composition. We shall see repeatedly below that virtually any basalt can be a primary magma regardless of its MgO content. Any such basalt traversing the lithosphere has the opportunity to entrain high concentrations of nonindigenous crystals of olivine or other mafic phases. The new bulk composition certainly appears primitive, but it is not primary, and to assume that it is short-circuits the reasoning process leading to understanding magmatic process. It is much more appropriate to consider magmatic processes as being characterized by a ‘carrier magma’ that is subjected to the mechanical gain and loss of crystals throughout its life. The carrier magma itself can be, within reason, of almost any composition. It certainly need not be inherently primitive, and it may be challenging to recognize the carrier magma, which, depending on the age and size of the system, may be of several different types. This concept is particularly important when considering the role of magma chambers in magmatic processes. 6.07.3.4 Historical Note on Solidification Front Fractionation The possible role of solidification fronts in capturing phenocrysts was first noticed by Pirsson (1905) in studying Shonkin Sag. Two of his students (Osborne and Roberts, 1931) were the first to analytically model the growth of the solidification front and compare this with the rate of settling of crystals due to Stokes’ law. They compared their results to observations at Shonkin Sag, but due to a misnumbered sample the calculated profile did not match that observed. Jaeger and Joplin (1955) made a substantial contribution when they realized that newly grown crystals falling from the advancing solidification front will dissolve as they sink into the hotter interior magma and will not contribute to fractionation. They were taken to task for this point of view by Walker (1956) and Hess (1956) who could not understand Jaeger’s straightforward analysis. Walker and Hess firmly held the assumption that, in essence, ‘‘it is well known by petrologists that crystals grow in the interiors of magmas and commonly settle to fractionate the system.’’ They did not realize that the effect they were thinking of came from systems injected with phenocrysts from the start. J. C. Jaeger was the author with H. S. Carslaw of the fundamental
289
treatise on heat conduction in solids, and his analytical insight into cooling and solidification of magma was clear and considerable. The inability of leading petrologists of the day to recognize the truly original and valuable contributions by Jaeger set back the development of magma physics by several decades. The production of S-shaped profiles was analyzed again by Gray and Crain (1969) and Fuijii (1974).
6.07.4 Magma Chambers Magma chambers are the conceptual cornerstones of every magmatic process. They are high-level staging areas that charge volcanoes for eruption. They are cavernous pools wherein complex patterns and cycles of crystallization take place to transform initially homogeneous melt into spatially differentiated sequences, which may or may not be erupted. They are enormously versatile in size, shape, and function, and can produce layered intrusions, a layered oceanic crust, and homogeneous diabase sills and granitic plutons. The central feature of this dynamic versatility is the pervasive nucleation and growth of crystals throughout the chamber that can be separated, sorted, or re-entrained in a vast array of processes (see Figure 12). This valuable and useful concept of magma chambers has been adhered to for over one hundred years. It is based on an exceedingly fundamental and misleading premise of how silicate melts solidify. Here we examine this concept in detail, to strengthen it fundamentally, and to show that all magmatic systems function by a common, basic, and simple set of processes. 6.07.4.1 The Problem: The Diversity of Igneous Rocks The abundant variety of igneous rocks found on Earth’s surface has long spawned an innate curiosity into the processes responsible for this variety and how these processes might have shaped the very structure of Earth itself. A strong burst of intellectual activity concerning the origin and evolution of magma was brought on by the advent of thin section-making, western exploration in the United States, and by the general intense activity in chemistry and physics taking place at the beginning of the twentieth century. The rapid discovery of new elements, radioactivity, blackbody radiation, and the development of chemical thermodynamics by
290
Magmatism, Magma, and Magma Chambers
Final composition
SiO2
Figure 12 The historically imagined or classical magma chamber showing crystals settling from the center of the chamber leading to a pile of cumulates on the floor and a silica-rich sandwich horizon at the interface of the upper and lower crystallization zones.
J. Willard Gibbs, for example, set the stage for fundamental advances in understanding magma.
6.07.4.2
George Becker’s Magma Chamber
In 1897, George Becker (1847–1919) set forth the basic concept of magma chamber differentiation that is still largely adhered to this day. Becker was mainly a chemist who worked for the US Geological Survey. He was impressed by how a cooled bottle of wine grows ice crystals around the edges and as crystallization proceeds the mother liquor or residual melt toward the center becomes increasingly rich in alcohol, the component with the lowest freezing temperature. If the temperature is progressively lowered, the container could become compositionally zoned to form, in effect, a differentiation sequence, which he said would be a perfect analog of a laccolith. Becker said: ‘‘A bottle of wine or a barrel of cider exposed to low temperature deposits nearly pure ice on the walls, while a stronger liquor may be tapped from the center. If still a lower temperature were applied the central and more fusible portion would also solidify. Such a mass would be, so far as I can see, a very perfect analogue to a laccolith.’’ He had in mind the rock sequence displayed by Shonkin Sag laccolith, which had been discovered in north-central Montana during mapping of the 49th parallel. This style of fractional crystallization was well known from ancient times and was practiced by Native Americans in concentrating maple sap prior to boiling down to syrup. Charles Darwin noticed on his visit to the Galapagos during the Beagle voyage that large olivine crystals had clearly settled to the base of lobes of basalt. Fouque
realized in 1879, in his chemical study of the eruptives of Santorini, that the separation of plagioclase enriched the lava in silica. Crystal fractionation was a common chemistry laboratory technique in the late nineteenth century. Madam Curie used it to isolate radium and Becker, having been trained as a chemist, certainly knew of the effectiveness of this process. Becker’s idea undoubtedly describes the most fundamental concept in magmatic processes. The fact that crystals in multicomponent, multiphase systems always have compositions different than the host liquids allows a systematic evolution in melt composition with crystallization. This feature of solutions is what makes metallurgy, ceramics, and igneous petrology the rich fields that they are. Although this feature of solutions was long known, Becker made the important realization that this process might also gives rise to a ‘spatial’ variation in final rock composition. Spatial variations in rock composition are fundamental to magmatic systems. The downside to Becker’s idea, which has been overlooked in deference to its conceptual usefulness, is that the process really only works with dendritic crystals, which are the rule in metals and aqueous solutions but not silicate magmas. The whole essence of magmatic processes, and rocks themselves in the broadest possible sense, rests on the fact that magmas and rocks consist of very small crystals or clusters of small crystals that grow in essentially parasitic chemical relationships (e.g., Marsh, 1998). This simple subtlety in crystal size and form, known so well by every petrologist, coupled with a spatially varying thermal regime is the basis of all magmatic processes. It simplifies and changes completely Becker’s
Magmatism, Magma, and Magma Chambers
concept of magma chambers, and it exposes a robust continuum linking all magmatic events. Louis Pirsson (1860–1919) picked up on Becker’s idea and significantly extended it to explain the lithologic variations seen at Shonkin Sag laccolith. Pirsson, a petrologist at Yale, R. A. Daly, W. H. Weed, and others visited Shonkin Sag and found it to be a remarkably differentiated, small (70 m) sheetlike body. Although Shonkin Sag will be discussed in more detail later, Pirsson combined Becker’s idea with his own insight on thermal convection in fluids to give a plausible explanation of how Shonkin Sag evolved. His detailed exposition (Pirsson, 1905; pp. 187–188) of this process is at the very root common beliefs of magmatic processes: It seems almost impossible to resist the view that in an enclosed mass of magma sufficiently mobile for local differentiation to take place convection currents due to unequal cooling would occur. On the upper surface and along the outer walls cooling would take place more rapidly; on the floor of the chamber, protected by the heated mass above and with heated rocks below, less rapidly. Thus there would be a tendency along the top and sides for the magma to grow heavier and to descend. Material from the more highly heated central part would tend to rise and replace this, and thus currents would be established in the magma, rising in the center, flowing off to the sides at the top, and descending along the cooler walls. . . . Such currents, once established, would continue as long as sufficient mobility remained in the magma to permit them. At some period crystallization would take place, and this most naturally would begin at the outer walls. It would not begin at the top because the material would arrive there from below at its highest temperature. Moving off toward the sides the material begins to cool and descend and becomes coolest as it nears the floor; here crystallization would commence. The first substance to crystallize is the solvent, which in this case would be the femic minerals, chiefly augite. Part of the material solidified would remain attached to the outer wall and form a gradually increasing crust, and part would be in the form of free crystals swimming in the liquid and carried on in the current. Probably at first, as the liquid moved inward over the floor of the laccolith and became reheated, these crystals would remelt, giving rise to numerous small spots of magma of a different composition, which would slowly diffuse.
291
As time went on, however, there would be a constantly increasing tendency for the crystals to endure; they would be carried greater and greater distances. But as they are solid objects and of greater specific gravity than the liquid, there might be a tendency for the crystals to drag behind and accumulate on the floor of the chamber. Moreover, from the heat set free at the time of their crystallization and from the resulting concentration of the chemically combined water vapor in the magma, the residual liquid would tend to have its mobility kept undiminished, since these would be factors which would tend to counteract the increase in viscosity due to cooling. In this manner it may be possible to understand how there would form a femic marginal crust and a great thickness of the femic material at the bottom of the laccolith. As the cooling went on the edges of the outer crust would rise more and more toward the top, finally spreading over it, and as a result the crust should be thinner on the top than elsewhere, as in the Shonkin Sag laccolith, in which the upper crust of femic rock is still preserved.
Beginning with a crystal-free melt, Pirsson suggested that crystals nucleate and grow inward from the edges of the laccolith, convecting melt moves through the growing marginal band of crystals, chemically exchanging with them freely, and unattached crystals pile up on the floor. He suggested that crystallization will be most intense near the roof where cooling is strongest and that these crystals will be carried downward by convection along the walls and then outward over the floor and deposited as the melt itself keeps circulating and differentiating. The end result is a thick pile of crystals on the floor, a thin rind of crystals welded to the roof, and a layer of highly differentiated rock just below the roof in what later came to be called the sandwich horizon. This basic idea, which was given root by Louis Pirsson to explain Shonkin Sag, has pervaded all of igneous petrology, with only minor modification to this day. This is a powerful, well-reasoned, and clearly stated idea. It is based on seemingly straightforward, but faulty, reasoning, which inherently exists, without any clear enunciation or realization, in the reasoning behind almost every present-day petrologic scenario. The key features of this reasoning are an assumption of the knowledge of the ‘initial conditions’ of the system at the time of emplacement, the ‘spatial style of crystallization’ of magma and the ‘size of crystals’ that can be grown after emplacement. These concepts remain elusive to this day, and
292
Magmatism, Magma, and Magma Chambers
need to be examined in detail. Before dong this, however, in order to understand how this idea flourished, it is essential to appreciate the historical context and activity of the petrologists alive at this time.
6.07.5 Historical Setting The propagation of ideas very often rests on the random happenstance of how people appear and disappear from science. And it is in the mix of expertise lost and gained with the loss and gain of these people that ideas become taken seriously, become modified, and inadvertently become fundamental principles of the science itself. A corollary to this is the remark often attributed to Max Planck that beliefs in science mainly change through the loss by death of the holders of the old, out-of-date ideas. But it also sometimes true that with the loss of these people a vast resource of experience in observation is also lost. And some of these lost observations may be those that present serious obstacles to the acceptance of the newer ideas. This is especially serious in petrology where a great deal of the fundamental observations reside in a life’s experience in studying field relations and gaining impressions of processes that may be impossible to quantify or record in any permanent way. So the science may sometimes suddenly proceed not because of the establishment of a new fundamental principle, but by the inadvertent removal of an obstacle through the loss of understanding of a fundamental observation or of the loss of the observer him/herself. It was this nature of events that largely defined the research field of igneous petrology for all of the twentieth century. And the period of 1905–15, in particular, was perhaps the most critical of all in shaping the concepts of magmatic processes and, especially, magma chambers. Prior to about 1905, arguably the greatest advance in understanding magma was in the development of thin sections and the establishment of the CIPW norm. Thin sections allowed a detailed inspection of the intimate relationships between minerals quenched in the act of crystallizing. Although discovered and developed by the Britons W. Nicol (1768–1851) and H. C. Sorby (1826–1908) (e.g., Dawson, 1992), it was principally the Germans K. H. F. Rosenbusch and F. Zirkel who quickly and systematically developed petrography into a science and trained a cadre of petrographers. The
exceedingly fundamental realization that a rock composition could be expressed in any number of textures and crystal sizes spawned the need for some way to normalize texture, or remove its influence altogether, in order to put all rocks on the same basis for classification. The CIPW norm was the outgrowth of this need, whereby each composition could be represented by its equivalent plutonic texture. This removed the mystery of falsely classifying rocks due to overall crystallinity rather than simply due to chemical composition. The CIPW Norm was a monumental product of the penetrating petrologic insight of G. H. Williams (1856–1894) of Johns Hopkins, W. Cross (1854–1949) of the US Geological Survey, J. P. Iddings (1857–1920) of the University of Chicago, and L. V. Pirsson (1860–1919) of Yale, and the analytical chemistry insight of H. S. Washington (1867–1934) of the Carnegie Institute’s Geophysical Laboratory. Williams died in 1894 before full development and publication of what became the CIPW norm in 1903. These men quantitatively combined rock chemistry with their intuitive understanding of observational phase equilibria to predict with uncanny accuracy the crystallization sequence of virtually every igneous rock. In light of the fact that no systematic experimental phase equilibria were yet known for igneous rocks, this was a monumental and sophisticated scientific achievement. Nowadays, we have the equivalent operation in the form of comprehensive thermodynamic models of crystallization, two of which are, among others, the MELTS program by Ghiorso and associates (Ghiorso et al., 1983) and COMAGMAT by Ariskin and associates (Ariskin, 1999). This is a direct reflection of the rapid development of experimental petrology beginning at about 1910. In 1900, the level of petrologic insight within single individuals from the perspectives of field and thin-section studies was very possibly more sophisticated than it is today. There were, however, three major undeveloped areas that critically impeded the understanding of magmatic processes. These were: (1) mechanics of viscous fluids, (2) experimental phase equilibria, and (3) kinetics of crystal nucleation and growth. First, although George Gabriel Stokes (1819–1903) and Claude-Louis-Marie-Henri Navier (1785–1836) beginning in about 1850 had developed the basic equations of motion of viscous fluids, which set forth a stream of analysis involving both viscous and inviscid fluids, including Stokes’ famous result for the drag on a sphere, it was not until 1919 that J. W. Strutt (1842–1919) (Lord Rayleigh) solved the
Magmatism, Magma, and Magma Chambers
problem of simple thermal convection in a layer heated from below. Second, on phase equilibria, there were many bits and pieces of experimental information scattered about on natural systems starting with James Hall in 1798 through to F. A. Fouque in 1879 and Frederick Guthrie’s (1884) experiments on melting of metal alloys who stated: ‘‘I submit that, according to analogy, we should regard compound rocks and minerals [i.e., multicomponent systems], other than sedimentary rocks, as representing various kinds of eutectic alloys.’’ But there was no general understanding of the structure of phase diagrams for magmas. Many petrologists believed because of the apparent preponderance of granites and basalts that this reflected the presence of pole compositions in the solution chemistry of silicate magmas. That is, in a phase diagram of temperature versus solution silica content, there might exist two eutectic regions, one for basalt and the other for granitic or silicic residual compositions. Cooling and crystallizing magma inevitably ends up at one or the other compositional end point. Earth as a planet is essentially bimodal in granitic continents and basaltic ocean basins, observed R. A. Daly (1871–1957), which reflects this process. And third, there was no general appreciation of how the rate of solidification is related to the size and number of crystals in a rock. The distinction between crystals nucleated and grown in place over phenocrysts carried in with the magma was not appreciated. The role of rapid quenching at the contacts of sills and dikes producing chilled margins, laced with large numbers of small crystals, and slower cooling in the interior producing larger crystals, was clearly long recognized. But there was no realization that inordinately large crystals in small sills like Shonkin Sag could not have grown after emplacement. In fact, there was little understanding of the need to isolate and define the initial conditions of a newly emplaced magma. Although not explicitly discussed until much later, it was apparently assumed that all magmas arrived for emplacement essentially free of crystals. Today, these three areas are each becoming mature fields in and of themselves, but the complex interplay of these areas, considered as one dynamic continuum, still remains as the foremost frontier to understanding magmatic processes.
6.07.5.1
Life Time Lines
The life spans of the principal petrologists of this time are shown by Figure 13. Nearly all of
293
those mentioned so far were born in the middle 1800s and by about 1915 they were either dying or retiring from the field. The new wave of petrologists (e.g., Bowen, Grout, Wager) were emerging from graduate school with a desire to apply new methods, mainly phase equilibria, to solve classical magmatic problems. Although Becker and Pirsson had clearly defined the dynamics of magma chambers, the usefulness of this concept was yet to be measured against the phase equilibria of real silicate systems, which hitherto was completely unknown in any quantitative sense. This entire area was ripe for development in the wake of the seminal experimental work by Arthur L. Day on the plagioclase binary. N. L. Bowen (1887–1955) rapidly became the leader in this area and in 1915 he published one of the most important papers of the century (Bowen, 1915). Armed with a relative wealth of basic information on silicate phase equilibria, diffusion constants, and elementary thermodynamic constants (e.g., heats of fusion), which he had extracted from phase diagrams, Bowen systematically evaluated, one by one, all the various suggested mechanisms thought perhaps to effect the differentiation of magma. His logical, detailed, and quantitative approach of coupling experiment, theory, and field inferences swept away a host of processes as being ineffective. He established fractional crystallization by crystal settling as the primary means of differentiation. That is, prior to knowing silicate phase equilibria, nearly every mechanism known to chemistry had been argued to be important in explaining the diversity of igneous rocks. Diffusional exchange with wall rock, Soret effect, wall rock assimilation, gas fluxing, among many other processes, had been suggested to be important. They were all equal competitors. Bowen evaluated each of these in a fashion that is remarkably similar to the modern method of scaling analysis in applied mathematics. He was forceful and thorough in his conclusions and he left no room for argument. He demolished all competing theories on differentiation. He even attacked the Becker–Pirsson model where the residual convecting melt continually reacts through diffusion with dendritic crystals growing on the walls. Instead, he said that only by crystal growth and settling could significant differentiation proceed. He overlooked the fact that this is exactly what Pirsson himself had said in applying Becker’s wine bottle model to Shonkin Sag.
294
Magmatism, Magma, and Magma Chambers
Year 1800
1820
1840
1860
1880
1900
1920
1940
1960
1980
97 Hutton 17 Werner von Buch
52 von Leonhard 62 Nicol 51 Sorby 26 08 Rosenbusch 36 14 Zirkel 38 12 Becker 47 19 Brogger 51 40 Cross 54 49 Williams 56 94 Iddings 57 20 Pirrson 60 19 Washington 67 35 Barus 56 34 Vogt 58 32 Ossan 59 23 Harker 59 39 Lane 63 48 Fenner 70 49 Daly 71 57 Grout 80 58 Shand 82 57 Knopf 82 66 Bowen 87 55 Niggli 88 53 Alling 88 60 Read 89 70 Waters 05
Figure 13 The names and dates of existence of prominent geologists and petrologists of the late nineteenth and early twentieth centuries.
In this 90-page treatise, Bowen showed (see Figure 14) that it is principally the phase relations of solid solutions and not eutectics that control igneous systems, and that there is therefore no practical end to how far differentiation can proceed as long as crystals can settle or be otherwise physically separated from the melt. There were, however, many petrologists who did not believe that crystals could settle in viscous magma. Bowen not only showed experimentally in a large experimental charge that olivine in picritic magma did clearly settle, but he also deduced the viscosity of melts and used Stokes’ law to calculate settling velocities (Bowen, 1947). He performed novel experiments and produced data crucial to answering the central question posed by the earlier generation. The amount of printed material he turned out was enormous and a great deal of it is still pertinent. The bottom line is that Bowen pushed and promoted his model thoroughly and hard. He said
crystals grow in the middle of a charge or a magma chamber. They settle to the bottom and the residual liquid compositionally evolves from basalt to granite. It is simply an unavoidable fact of silicate phase equilibria. In its simplest dynamic sense, petrologists have stuck with this to the present. No one today would refute that the separation of crystals from melt causes chemical differentiation, but what is still unclear is how this actually takes place in a physical sense. For if all significant nucleation and growth of crystals is in marginal fronts and there are no crystals in the interior, how does crystal–liquid separation take place? To his credit, Bowen did worry about this and he proposed that melt buried in high-crystallinity mushes might be kneaded out by tectonic processes. But later he realized that this would probably be ineffective because the timescale of deformation is generally much longer than that for solidification.
Magmatism, Magma, and Magma Chambers
295
Bowen’s reaction series and solidification front 75 Rhyolite Dacite
K-feldspar and quartz
itic Alb
Andesite
0.6 y
Pla
gio
Visc osit
0.4
60
55
e en rox Py
cla se
-
ite iot db an le ibo ph
Crystallinity, φ
65
Am
0.8
70
Basalt
An
e
ort
hit ic
ivin Ol
0.2
Interstitial melt silica content (wt.%)
1.0 Temperature
0 10
50 103
105
107
109
Melt viscosity (P, CGS) Figure 14 Bowen’s reaction series superimposed on an active basalt solidification front at the roof of a sheet of magma. Variations of composition and viscosity as a result crystallization are also shown.
6.07.6 Initial Conditions of Magmatic Systems All natural processes involve time. And the most critical part of analyzing and understanding magmatic processes involves a clear and concise knowledge of the nature of the system at the outset when it all began. The ensuing process itself, especially including the eventual final outcome, reflects in every detail the initial state of the magma. The final product, more than anything else, simply reflects the initial conditions. The importance of these conditions has been routinely overlooked, and the process conjured up to explain rock sequences, especially complicated ones, are often unrealistic and overly complicated. They are complicated mainly because the assumed initial conditions are so far from the final state of the system and are hence impossibly difficult to sensibly connect in any logical construct. These initial conditions, as mentioned earlier, are: 1. the spatial pattern of cooling and crystallization; 2. the composition and initial state of crystallinity of the magma; 3. the sequence of emplacement of the magma; and 4. the final distribution of crystal sizes throughout the body. These are, in and of themselves, often exceedingly difficult conditions to ascertain without first knowing what evidence in the rocks clearly reflects these
conditions. In early studies, the critical importance of these conditions to the final outcome was unknown and the simplest possible conditions were assumed. In their classic study of the chemical evolution and layering of the Skaergaard Intrusion, for example, Wager and Deer (1939) and Wager and Brown (1968) stated at the outset that crystal-free magma was instantaneously injected, perhaps in a violent fashion. For the initial composition of the magma, they took that composition of the chilled against the wall rock, gabbro. Skaergaard contains about 600 km3 of magma in a body with an aspect ration of about 2:1. This is not a large body of magma – many large dolerite sills are this large. But why does it exhibit such well-formed layering? Is it the form of the container holding the magma? The style of cooling? Or is it the chemical composition of the magma? Given the initial conditions assumed for the formation and condition of the initial magma, a host of processes involving sorting of crystals existing in the initial magma are excluded, as are also processes associated with the emplacement sequence itself. Given the assumptions, Wager and Deer explained Skaergaard, with no (or very few) initial crystals, as crystal growth starting somewhere in the middle of the body in response to strong cooling from the roof. The melt mingles with any crystals that are around the edges and keeps reacting like the bottle of wine so that the middle of the body chemically evolves with time. Cumulous crystals, like snowfalls, build up on the
296
Magmatism, Magma, and Magma Chambers
floor and form sequences of modally layered rocks in response to periodic loading and unloading of the system with crystals due to roofward cooling. The geometric form of the body encourages cascades of crystal-laden currents down the sloping walls. Within the piles of cumulate crystals on the floor, compaction, continued crystal growth, and melt migration physically modify the textures and chemically evolve the rock to fit the final product. Additional liquid may leak out, but eventually we end up with some liquid left at the roof with some rind on the top. This perspective is heavily influenced by Pirsson’s treatise on Shonkin Sag. Wager and Deer got this perspective from Harry Hess who was working on the massive, heavily layered Stillwater Intrusion. (Although Hess’s memoir was published much later in 1960, his incidental reports on his Stillwater work are cited by Wager and Deer.) Faced with similar sequences of modally layered rocks, Hess went back to the early work of Becker, Pirsson, and Grout, and incorporated the overall process spelled out by Pirsson. The important point here is that the Pirsson model is the fundamental idea from which all explanations for the evolution of layered intrusions have emanated. It is thus important to understand this model in more detail in light of discerning the importance of magmatic initial conditions. 6.07.6.1
Cooling from the Roof
When an opening is made in the crust for emplacement of a magmatic body, the upper and lower contacts of country rock initially had exactly the same temperature. The magma is generally much hotter than the country rock and the difference in temperature between magma and country rock is large and virtually identical at upper and lower contacts, regardless of the thickness of the body. The rate of cooling of the magma by conduction through the upper and lower wall rock is also virtually identical. The body thus cools at the same rate from top and bottom. This is shown by considering the central temperature in identical hot sheets emplaced far from, near, and on Earth’s surface. There is no difference, especially in early times, in the rate of cooling for deeply buried and shallow intrusions, and only a sight difference for a body (e.g., lava or lava lake) on the surface. The sheets of magma are so much hotter than their surroundings that it is essentially immaterial what they are surrounded by. This also holds for all the other margins of the body, that is, sidewalls. Other processes of cooling, such as hydrothermal convection, will influence differently
the upper and lower contacts, but mainly only in the later stages of cooling when solidification is no longer important (e.g., Marsh, 1989). The conclusion is, thus, that magmas in general do not cool more strongly from the roof and crystallization is no more intense along the roof than it is elsewhere along the margins of the body. Cooling is characterized by heat flow from all outer margins of the magma at ‘approximately’ similar rates, and this cooling is characterized everywhere by crystallization. For every unit of thermal energy flowing from the magma, a specific mass of crystals is produced in a specific spatial pattern. 6.07.6.2 Style of Crystal Nucleation and Growth In Becker’s original model of magma chamber crystallization, for example, liquid moves freely in and out of the mass of growing crystals, carrying, in effect, nutrients to the crystals. The residual melt becomes increasingly depleted in time in crystalline components, thereby becoming enriched in unwanted or residual components. The center of a crystallizing bottle of wine becomes increasingly rich in alcohol. This is essentially what happens in any system where dendritic crystal growth prevails as in metallic alloys and aqueous solutions. These solutions are also commonly of low viscosity melt and low Prandtl number. Compared to silicates dendritic crystals are unusually large and can span a significant distance within a solidification front. If these dendritic systems were scaled up to the size of typical bodies of magma, dendritic crystals might reach sizes of tens of meters or hundreds of meters long. Large compositional boundary layers develop on the large crystals and are buoyant, highly mobile, and highly effective in carrying ‘used up’ fluid from the crystals. This is very much unlike what we see in silicate magmas. One of the curious things about magma is that the size of crystals in most igneous rocks is on the order of tenths of a millimeter to a centimeter. That reflects the multisaturated, multicomponent nature of silicate magmas where chemical diffusion is slow and the melts are highly viscous (e.g., large Prandtl number). Within thickening solidification fronts, crystals nucleate and grow at the leading edge near the liquidus. Nucleation certainly does not start here, for magmas are essentially dirty systems, laced with nuclei. Only in some very rare systems are magmas ever free of nuclei as an initial condition. This reflects the fact that endogenous magmas are never
Magmatism, Magma, and Magma Chambers
found to be superheated. Nuclei are thus always present and serve as seeds for crystallization regardless of the exact phase involved. Once the cooling front arrives, existing nuclei increase in size outward through the solidification front. The tiny growing crystals are in a parasitic relationship where what chemical components one crystal does not want, another one will nucleate and grow to use. Diffusion is so slow in silicate liquids that if an olivine crystal is growing on the liquidus of tholeiitic basalt, for example, unwanted components like CaO, Al2O3, Na2O, Fe2O3, SiO2, and K2O build up in the surrounding melt. This increases the chemical potentials of clinopyroxene and plagioclase, spawning nucleation and crystal growth. This reflects the common state near multiple saturation in many basaltic magmas, a slight enhancement in concentration of rejected components will saturate the melt in another solid phase. Continual growth of these crystals will enhance nucleation and growth of the other phases and vice versa. This creates local intense clusters of crystals of several phases and, eventually, from these clusters appears single large, well-formed crystals (see Figure 15). The small crystals combine through aggregation and grain boundary migration into large,
297
optically continuous crystals. The process is part of the phenomenon of Ostwald ripening, where touching grains combine in response to lowering the net surface free energy. This is, in effect, a form of hightemperature annealing, and it happens so rapidly that it is difficult to observe except in samples quenched from lava lakes. The long-believed perception of crystals nucleating and growing at widely separate locations and only eventually impinging on one another at the close of solidification is a reflection of an assumption of homogeneous nucleation, which rarely occurs in magma. Heterogeneous nucleation is the rule in magma.
6.07.6.3 The Critical Connection between Space and Composition The variations in melt silica content and viscosity with distance or crystallinity within a typical tholeiitic basalt solidification front are shown by Figure 16. Both silica content and viscosity increase dramatically with the degree of crystallinity or distance outward in the solidification front. It is especially notable that silica enrichment increases gradually for the first 50 vol.% of crystallization, increasing by only about 5 wt.%. It is only outward of this point, at much higher crystallinities, that strong enrichment occurs. But once the level of crystallinity reaches 50 vol.%, the point of critical crystallinity has been reached where the magma is a rigid network of crystals laced with melt. The melt and crystals cannot Solidification front Interstitial melt silica content (wt.%) 50
55
Andesite
Basalt
109
Dacite
Rhyolite
107 105 103
Visco
sity
10
1.0 Figure 15 Thin sections (5 2 cm2) from Makaopuhi Lava Lake (courtesy of Dr. T. L. Wright) at crystallinities of c. 20 and 40 vol.% (upper to lower) showing the tendency for crystallization to occur in dense clusters.
0.8
0.6
0.4
0.2
0
Crystallinity, φ Figure 16 A solidification front on the sidewall of a basaltic magma showing the variations in composition, viscosity, and crystal mass as a function of position.
Melt viscosity (P,CGS)
75 70 65 60
298
Magmatism, Magma, and Magma Chambers
Relationship between phase space and distance space Roof rock Solidus
Silicic segregations Rigid crust 50% solids Mush Capture front Suspension Liquidus
Temperature
be easily separated, and the crystals surely cannot be separated by conventional crystal settling. Andesitic, dacitic, and rhyolitic melts occur at increasingly higher states of crystallinity, where the separation of melt and solids in increasingly difficult. The key to chemical differentiation of the magma is to isolate the enriched melt at a separate body of magma. How does this occur? Out in the center of the body nothing is happening. As seen earlier, crystals will not begin growing there until the inward propagating solidification front arrives. Even small crystals escaping from the suspension zone of the solidification front will have no effect on the overall magma composition. Tiny crystals falling from the suspension zone are resorbed in the inner hotter magma and do nothing to change melt composition. Even a big plume of crystals and slightly cooler melt dropping from the suspension zone, which might survive transit of the hot inner magma, simply enters the suspension zone of the lower solidification front with no net spatial change in the composition of the system. It just goes from one part of the system to another. Low-density melt might escape from the lower solidification front, but the viscous chemical boundary layers on the tiny crystals have little buoyancy and are stabilized by the cluster style of nucleation and growth and the advancement of the solidification front itself. Chemically, nothing is going on out in the center of the body. It is within the solidification fronts that any differentiation will take place. The most critical relationship to appreciate in understanding magmatic processes is the relationship between moving across a conventional phase diagram and moving in space in a magmatic system. Because the temperature field of a magma is closely tied to space, phase diagrams are also tied to space. Moving on a phase diagram, thus, always involves moving in space (i.e., x, y, z) in a magma, and these moves are almost always within solidification fronts because that is where temperature varies the strongest with spatial coordinates. This critical relationship can be readily appreciated by considering a David Walker-style normative ternary phase diagram with the normative components olivine, diopside, and silica at the apices and the systems is saturated everywhere with plagioclase (see Figure 17). These phase relations fit a wide range of basaltic magmas from tholeiites to high-alumina basalts (HABs). There is a single major cotectic boundary separating the two major phase fields of olivine, or alternatively orthopyroxene at higher silica
Crystal-free magma
10 μo
μo (bulk viscosity)
Di
Bulk composition of solidification front
Trapped interstitial melts
+ Plag
CPX
Silicic segregations (continents)
MORB (seafloor)
Ol Ol
1 atm OPX
En
Sil
Figure 17 The fundamental relationship between spatial position (upper) in a solidification front and phase position in the normative basalt ternary system.
contents, and diopside or clinopyroxene. Because of the mapping between temperature and space, any movement on this phase diagram, along the cotectic or anywhere else, involves a move in magma space. Beginning in the middle of the magmatic system, the corresponding point on the phase diagram is the beginning bulk magma composition on the cotectic boundary (see the large block dot). Moving in space from the center outward into the upper solidification front no move occurs on the phase diagram because there are no temperature or composition changes. Once the solidification front is entered there is a choice whether to follow the bulk composition on the phase diagram or only the interstitial melt composition. If the local bulk composition is followed, all points map onto the same initial starting bulk composition (i.e., the same starting black dot). This is because there has been no separation of crystals and melt. If, on the other hand, the local residual melt composition is monitored, then the compositions migrate, with increasing crystallization or depth outward in the solidification front, along the cotectic toward the silica corner of the ternary. The silica corner will, in fact, be closely approached as the very last bit of melt
Magmatism, Magma, and Magma Chambers
magmatic mush column. When the plates move apart, aliquots of essentially crystal-free magma are withdrawn or pulled from the subjacent sill by disrupting the roofward solidification front. No magma comes from within the solidification fronts, so the lavas are generally always just basalts. Once the system is activated in this way, more magma arrives from below, which may contain ‘tramp’ or dynamically entrained crystals from the underlying mush column. Upon entering the axial sill, these crystals will settle from the magma, which is an example of ‘punctuated differentiation’ discussed earlier, and the magma will sit there surrounded by solidification fronts. This process also explains why Hawaiian and similar large-volume basaltic centers in oceanic settings produce such a paucity of silicic differentiates. The summit lavas from Kilauea, for example, are never more fractionated than about 51.5 wt.% SiO2 and 7 wt.% MgO (see Figure 18). If this composition is traced from the chemical variation diagram to a solidification front, it is seen that this composition defines the inner, leading edge of the suspension zone of the solidification front. The more fractionated, more
crystallizes near the solidus at the back end of the advancing solidification front. It is important to realize that these late-stage, highly fractionated melt compositions do not exist separately, in and of themselves, as an actual physical entity that could be erupted as lava or re-emplaced as a pluton. These melts are locked within the crystalline framework of the solidification front and cannot be separated by any simple means; more on these processes later. An understanding of this critical relationship between spatial variations in magma and moves on a phase diagram makes solving, or at least understanding, many fundamental petrologic occurrences straightforward. Magmatism at ocean ridges the world over is predominantly tholeiitic with subtle variations of many kinds. No silicic differentiates are ever seen as lava. The volumetrically overwhelming composition rests on the cotectic and does not move toward the silica corner as might be expected from a viewpoint of classical crystal fractionation theory. No continent-like rocks are produced. As will be seen later, the ocean ridge magmatic system is essentially a sill or series of sills, called the axial ridge magma chamber, perched at the top of a vertically extensive
53
Intrasolidification front liquids (uneruptible)
52 51
Tr
ap
Lavas
50
SiO2
299
Di
pe
dm
elt
(eruptible)
s
49 48 47
MgO (wt.%) 46 25
0
20
15 7
8
9 Temp
erature
10
5 4
5
6
3
n
itio
m
Solid crust ulk
Solidus Solidification front
s po
B
co
elts
dm ppe
EN
SIL
Tra
Liquid MgO
10
Liquidus Magma 20
400
800 Temperature (°C)
1200
Figure 18 The explanation for the absence of lavas erupted from Kilauea summit with silica greater than about 51.5 wt.% is found in tracing the magma composition to the spatial position at the leading edge of the solidification front.
300
Magmatism, Magma, and Magma Chambers
siliceous compositions are, again, locked within the front and are not available by simple means for withdrawal and eruption. It should also be noticed that there are many more primitive, high MgO compositions, which represent, by and large, magmas carrying an abundance of entrained forsteriterich olivine xenocrysts or tramp crystals. Once the magma comes to rest, these crystals fall from the magma as in punctuated differentiation and the usual magma type of 51.5 wt.% SiO2 and 7 wt.% MgO is produced. There are sometimes mildly differentiated lavas erupted from the Kilauea rift system, silicic pods are also found in the crusts of the Hawaiian lava lakes, and on Iceland about 18% of the surface lavas are true rhyolites and process leading to these effects is examined later. We will see bimodal differentiation comes from solidification fronts that tear open and form lenses of silicic liquids or silicic segregations. These even show up in the lava lakes in Hawaii, but the key for recognizing the role of this material as eruptable magma on Earth’s surface is recognizing a means to collecting individual segregations into large, eruptable masses of magma. 6.07.6.4 The Sequence of Emplacement or Delivery of the Magma The fundamental question always lurking about but rarely answered in analyzing exhumed magmatic systems is: How was the system built? What was the overall nature of the event in time span and volumes of individual deliveries of batches of magma that made the system? How did the system operate on a year-to-year basis? These are difficult questions and the postmortem evidence in exhumed magmatic systems is rarely conclusive. Rather, the most direct evidence of the time scales and volumes of individual batches of magma comes from volcanic systems. Volcanic systems, on the other hand, lack all information on the details of the internal arrangement and local physical environments of the system in which the volcanic products formed. It is no wonder, then, that volcanologists and plutonists have widely divergent views on the basic structure and workings of magmatic systems. Yet it is the coupling of this evidence that gives critical insight into the real-time dynamics of the establishment and sustaining of magmatic systems. The quantitative connection between volcanism and plutonism is the product of observed erupted flux and the volume of individual plutonic complexes. This product measures the ‘filling time’ or time to establish a magmatic complex. To estimate
this product, something of the form and volumes of plutonic systems must be appreciated along with the temporal pulse of volcanic systems. 6.07.6.4.1
Forms of magmatic bodies Magmatic systems come in all shapes and sizes, but the most common shape is sheet-like for plutons, sills, and certainly lavas. This reflects the common vertical orientation of the least principal stress in the upper crust. Magma ascends as fluid-filled cracks (dikes) or viscous blobs (diapirs) or as a combination of each. Dike and fissure propagation is common in young or developing systems, where the country rock is cool and brittle, or in systems where the crust is in tension as in continuous rifting as at ocean ridges. Magma ascends due to its low density relative to the wall rock and the influence of gravity. Ascension is a gravitational instability. Ascension is impeded by loss of hydrostatic head (e.g., Ryan, 1994). The leading mass of magma reaches a level of neutral buoyancy, which may be due more to the vertically integrated buoyancy than to merely the local density contrast, and is forced to spread laterally by the continual rise of magma deeper in the rising column, thus forming sheetlike bodies, which in the upper crust are known as sills. Sheet-like forms are fundamental to all magma, regardless of level of emplacement in Earth. Magmatic systems are, to first order, a series of sheets connected by circular conduits and dike-like fissures. Diapirism is more the rule in stationary, welldeveloped magmatic systems, as perhaps beneath older volcanic centers in island arcs where the system has had ample time to develop, and the underlying lithosphere has had ample time to heat up from the long-term passage of many magmatic bodies. But here, too, stalling diapirs mushroom out and go to form sheet-like bodies. 6.07.6.4.2
Internal transport style Once the system is established and has formed a magmatic mush column of some type (Figure 1), the form of magma transport is probably more a combination of many processes driven by density variations reflecting basic differences in chemical composition due to melt composition and the abundance and type of crystals carried by the melt. At every level in a magmatic mush column the process of local transport, although broadly similar overall, may vary over time from diapiric on many length scales to slurry flow as in a pipe to dike flow. The style of the process itself and time series of repetition is certainly chaotic. But since the material is a viscous fluid of a relatively narrow
Magmatism, Magma, and Magma Chambers
range of composition and crystallinity, the outcome, almost regardless of the details of transport, is distinct to the tectonic setting and to a large extent predictable. Because of the strong effect of annealing during solidification, it is almost impossible to gauge the importance of repetitive reinjection during building of a plutonic complex. Internal chilled margins as indicators of the sudden juxtaposition of magma batches of contrasting temperature do not survive solidification, and the strong textural changes that do survive are almost universally interpreted as reflecting internal solidification dynamics. There are few if any true indicators in intrusive complexes of the nature of the time series of deliveries of the batches of magma that collected to make the final magma. As noted above, this is where the fluxes and eruptive timescales are absolutely critical to understanding magmatic systems. 6.07.6.4.3
Eruptive t ime sca le s a nd fl uxe s The observed durations of volcanic eruptions have been summarized by Simkin. By far the most common eruption lasts from about a month to a year, with the overall spectrum of times spanning about 100 years. Measures of eruption flux come from mass balances on both large and small scales. The flux of magma from global ocean ridge systems can be estimated from the product of the average rate of spreading (double sided, 5 cm yr1), the thickness of the oceanic crust (8 km), and the length of the ridge system (40 000 km), which amounts to about 15 km3 yr1. This is clearly a minimum estimate, for the actual amount of magma existing in the system at any time, because a significant amount of magma probably gets trapped in the deeper parts of the system and is recycled back into the mantle source rock. The flux that established the island of Hawaii must have been, on average, about 1 km3 yr1. The age of the island is about 1 My and it has a volume of about 106 km3. The apparent fluxes that established large igneous provinces fall in this range of 1–15 km3 yr1. For examples, consider the Columbia River flood basalts (1 km3 yr1), the Deccan traps (3 km3 yr1), and Ontong Java (15 km3 yr1) (see Coffin and Eldholm, 1992). The volumes of these lava sequences are generally much larger (1–50 106 km3) than any known magmatic intrusions, the largest of which are Bushveld (S. Africa), and Dufek (Antarctica), each at about 500 000 km3. The Sudbury impact melt sheet may originally have had a volume of about 35 000 km3, and Skaergaard is on the order of 650 km3. By far the
301
largest single ‘day’ eruptions are those of silicic pyroclastic ash flows. The Fish Canyon ash flow is about 3000 km3, the Toba Tuff is about 2000 km3, and the two Timber Mountain Tuffs are 1200 and 900 km3 (e.g., Smith and Roobol, 1982). The huge flux of these silicic events undoubtedly reflects the mode of transport of uncapping an overpressured, near-surface system containing large amount so dissolved volatiles. It would be difficult imagining an intrusive even of this magnitude. 6.07.6.4.4
Filling times Given the final volume of a pluton or sill, the filling time is a multiple of the input flux. For a body of the size of Skaergaard (650 km3), for example, the filling time is 45 years if all the magma produced at the ocean ridges (15 km3 yr1) were directed to this location. For a Hawaiian rate (1 km3 yr1), the filling time is longer, 650 years. There is no simple way to decide the actual filling time (U–Th and Po–Pb–Ra isotopic disequilibrium may offer some information) – the only physical constraint comes from rate of solidification. The rate of filling must be significantly greater than the rate of solidification. For these sheet-like systems, it is straightforward to show by scaling the heat equation (see eqn [1]) and including the effect of latent heat (e.g., Jaeger, 1968) that the solidification time (t)is well approximated by the simple formula t ¼ 0:694
ðL=2Þ2 K
½17
where L is the half-thickness of the sheet and K is the thermal diffusivity (e.g., 102 cm2 s1). The halfthickness of a sill or pluton can be estimated by noting that the aspect ratio (n) of sills is 100 or more whereas that of plutons is about 10; there are of course large variances in these values, especially for plutons. Nevertheless, for a given volume (V), of magma the half-thickness of the equivalent rectangular sheet of dimensions n2L n2L 2L is given by L ¼ (V/8n2)1/3. Under this approximation, for a given volume of magma, sills will be thinner than plutons and the solidification time of sills will be significantly smaller than for plutons. The competition in filling time and solidification time for a range of fluxes operating over the characteristic eruptive times found by Simkin is shown by Figure 19. In light of the earlier discussion of the controls of crystallinity on magma fluidity, the calculated time for solidification has been lessened by a factor of 10 to ensure that the body is sufficiently fluid to ensure reinjection without creating an internal chilled margin, or also to
302
Magmatism, Magma, and Magma Chambers
Filling time = Flux × duration
105
Time (yrs)
104
1 –1
3
10
3
m
102 10 1
(k
e s tim ton Plu tion x a flu ific lid ve i o t S up Er
yr
10
)
Observed durations of volcanic eruptions
ls
Sil
Eruption duration
0.1 10
4
Thickness
10
3
105
6
10 plutons, n = 10 n 2L
(cm)
10
4
2L 5
10 sills, n = 100
Figure 19 The time to fill sheet-like plutons and sills of a given aspect ratio (lower axis area) found using observed durations of eruptions and the estimated fluxes from large igneous outpourings. The time of solidification as a function of body size (half-thickness) is also shown as a constraint on the time of filling.
ensure that the sequence of arrivals of magmatic parcels can mix to make a single body. From this constraint, the characteristic thickness of a sill is 100 and 1000 m for a pluton, and the filling times are, respectively, about 10 and 300 years. These are geologically reasonable results, but the actual filling times may be much less. This sequence of events for sills will be revisited below when discussing the Ferrar dolerites. 6.07.6.4.5 Magmatic deliveries, episodes, periods, and repose times
At a more detailed level, it is also clear that the full volume of the body is probably not delivered in a single event, but, in keeping with the typical style of volcanic eruptions, individual batches of magma may arrive serially perhaps every few weeks or months as in a volcanic episode. The accumulated volume from the full series of deliveries amounts to a volcanic episode, which gives rise to the final body of magma. This is an important distinction as a series of episodes may also be strung together to make up a much longer magmatic period giving rise to a sequence of sills or a compound pluton consisting of multiple lobes each associated with one another but having separate cooling times. The episodes themselves reflect volcanic repose times, which for many active volcanoes may occur for tens to hundreds or perhaps thousands of years. Moreover, the longer the series of deliveries, the more chance there is for variations in phenocryst content in the batch of
arriving magma, which, as been discussed for Hawaii, is highly dependent on the magnitude of the eruptive flux. Thus, the larger the final body of magma, the larger is the probable variance in the concentration of large crystals, both phenocryts and tramp crystals, contained in the ensemble of the individual delivered batches of magma. The more variance in the deliveries, the more variance to be expected in the final sequence of the rock within the pluton or sill. It is very likely exactly these variances that give rise to exotic sequences of layering from modal sorting during deposition of crystal-laden deliveries of magma in the building of large igneous bodies. Nevertheless, in this context, what is perhaps also surprising is the abundance in volume flood basalt provinces of extensive lavas containing low concentrations of phenocrysts. 6.07.6.5 Thermal Ascent Characteristics and The Role of Thermal Convection From the earliest models of magma chambers, as in Pirsson’s discussion of Shonkin Sag, convection has always played a role in establishing the final product. The exact nature of this convection, whether it was driven by thermal, compositional, or sedimentation effects, has never been clear. Convection has been a necessary convenience in magmatic histories and it has taken on a multifaceted physical nature, most often without explicitly defining its specific form. The process of ascent, transport, and emplacement of magma is certainly a form of convection or advection. Fluid motions accompanying the deposition of phenocrysts, however complicated this might be due to the dynamics of slurries (e.g., Marsh, 2004), are also certainly a form of convection. But most often, perhaps in response to the concentrated attention on mantle convection in driving plate tectonics, magmatic convection has been assumed to be primarily thermal. The physics of conventional thermal convection in a crystallizing solution is not entirely straightforward as has been summarized by Zieg and Marsh (2005), from which the following summary comes. The importance of thermal buoyancy relative to viscous drag and heat transfer by conduction is a measure of the tendency for thermal convection to occur in a layer of fluid. These effects are collectively measured by the dimensionless Rayleigh number (Ra). The physics of thermal convection in many types of fluids with various forms of heat sources, geometries, and boundary conditions is well known (e.g., Turner, 1973). Relatively little is known, however, about thermal convection in fluids undergoing
Magmatism, Magma, and Magma Chambers
crystallization in marginal solidification fronts, especially in fluids as diverse as silicate magmas (Marsh, 1989). This uncertainty has spurred debate over the occurrence of thermal convection in common magmas (e.g., Marsh, 1991). With no direct experimental observations of magmatic thermal convection, experimental attention has been focused on analog fluids that might be similar to magma. These systems are molten paraffin (Viskanta and Gau, 1982; Brandeis and Marsh, 1989, 1990) and solutions of isopropanol and water (Kerr et al., 1989; Hort et al., 1999). This work shows that in an initially superheated melt convection sets in rapidly and persists only until all the superheat has been dissipated. Thermal convection steadily wanes with approach to the liquidus and ceases at the liquidus. All further cooling is by conduction, which may be the behavior of many crystallizing fluids. The strong contrast with nonsolidifying systems is marked by the kinetics of crystallization (Hort et al., 1999; Marsh, 1996; Hort, 1997). There is indirect evidence that it is also the case in magma in comparing the cooling of Hawaiian lava lakes with lava flows. The growth of the crust of the lava lakes, measured through direct drilling (e.g., Wright and Okamura, 1977), shows the same pattern of growth for lavas as measured by Hon et al. (1994). This is fundamentally significant, as will be seen below, for the magnitude of the Rayleigh number (Ra) depends on the characteristic length scale to the third power, and Ra must be much larger than about 103 for convection to be possible. Since most lavas are quite thin (5 m), Ra is small, certainly subcritical, which suggests the same for the much thicker (100 m) lava lakes. 6.07.6.5.1
Superheat This apparent lack of thermal convection may also be a reflection of the lack of superheat in terrestrial magma, which may erupt at or below the liquidus temperatures. This is because undersaturated magma adiabats are steeper than liquidi, which promotes superheating. The appearance of any superheat leads to an enormous Rayleigh number, which brings on vigorous thermal convection, dissipating the superheat to the wall rock and bringing the temperature strongly back to the liquidus (see Figure 3). With the loss of superheat, thermal convection ceases, and the process repeats itself on a timescale governed by the rate of ascent. The thermal ascent trajectory is an oscillation about the liquidus. For sufficiently rapid ascent rates, magma temperature is, in effect, buffered at the liquidus. When ascent is sufficiently slow, heat loss by conduction
303
sustains magma temperature at subliquidus values, but here the magma flirts with stagnation through critical increases in viscosity due to crystallization (note shading in figure). ‘Dry’ magmas, like at Hawaii, erupt at near-liquidus temperatures. No magma is ever bone-dry, and Hawaiian MgO-rich lavas contain significant amounts of ‘phenocrysts’, which are a combination of so-called ‘tramp’ crystals entrained from earlier crystallization cumulates and even wall rock materials (e.g., Wright and Fiske, 1971; Garcia et al., 1989; Marsh, 2006). The relatively low volatile content makes the ensuing eruption of low explosivity, but the lavas are highly mobile. This is just the reverse for ‘wet’ magmas. Here ‘wet’ means magma containing significant amounts of dissolved volatiles, especially water, which for alkali basalts may mount to 2–3 wt.%. Wet magmas similarly follow the liquidus, but only as long as they remain undersaturated. At the depth of volatile saturation, the liquidus and adiabat cross. At this point, the liquidus, depending on the exact water content, may be at or below the 1 atm solidus. Further adiabatic ascent is, at best, isothermal (allowing for some crystallization; see details below), and the magma undergoes progressive devolatilization, bubble nucleation, bubble coalescence, fragmentation, and extensive ash and tephra formation. The eruption is highly explosive, but the ensuing lavas are highly immobile. By analogy with siliceous systems (e.g., Jaupart, 1998; Dobran, 2001; Melnik et al., 2005), the magma degasses and slowly squeezes out as a viscous sludge. The pervasive lack of superheat in endogenetic magma bodies may thus reflect the efficiency of convection in eliminating superheat as it occurs during ascent (Marsh, 1989). Magmas formed by impacts, however, can become instantly superheated in situ. This results in vigorous magmatic convection that lasts until all the magma has cooled to its liquidus temperature. The initial temperature of the Sudbury impact melt sheet, as will be considered below, was approximately 1700 C, well above the liquidus temperatures. In this context, and also perhaps for the lavas on Io, it is useful to consider the role of thermal convection in superheated magma. The magnitude of the governing Ra at the onset of convection determines the time scales for convection to become fully developed and also for the superheat to be removed from the system. Ra can be estimated directly from the definition Ra ¼
gTL3 v
½18
304
Magmatism, Magma, and Magma Chambers
where g is gravity, T is the initial amount of superheat (500 C for Sudbury), is the coefficient of thermal expansion (5 105 C1), L is the thickness of an individual layer (norite 1 km, granophyre 1.5 km for Sudbury), v is the kinematic viscosity of the melt layers (norite 10 cm2 s1, granophyre 100 cm2 s1), and is the thermal diffusivity (102 cm2 s1). With these values, Ra ¼ 1015 – 1017
½19
For reference, the Ra governing convection in Earth’s mantle is 105, and convection commences in a layer of fluid contained between two solid boundaries heated from below and cooled from above when Ra ¼ 1708. For one solid boundary and one fluid boundary, Ra ¼ 1100, which is appropriate for Sudbury. This implies that convection in the Sudbury melt sheet was initially in the regime of vigorous turbulence (e.g., Khurana, 1988). This means that the thermal regime was chaotic and the layers would have been rapidly mixed and homogenized. It is important to emphasize that, during convection, all the main dynamic features of the system (e.g., convective velocity, rate of heat transfer, etc.) depend intimately on Ra. It is also important to realize that even for a magma with a very small amount of superheat (1 C), Ra could be large if L is sufficiently large. In the above estimate of Ra for Sudbury, reducing T to this level only brings Ra down to about 1012–1013. This suggests that any amount of superheat whatsoever will bring on vigorous convection and rapidly bring the temperature back to the liquidus. Thermal convection cannot be expected as a common process in magma chambers. For completeness, especially in considering ascending magmas, extraterrestrial magmas, and impact melt sheets, it is essential to consider the startup time and rate of cooling in convecting magma. The time (t) for convection in each layer to start and become fully established, for example, is given by (e.g., Jhaveri and Homsy, 1980) L2 B 2=3 t ¼ Ra
½20
where B (350) is a constant and the other symbols are as above. Using Ra as found earlier, thermal convection in the Sudbury melt sheet would have fully developed within about an hour. This reflects the potential strength of convection due to the presence of such a large degree of superheat. The rate of cooling also reflects this extreme vigor of thermal convection.
There are two extremes for the emplacement and cooling environment of magma. The first is magma cooled rapidly at the margins due perhaps to hydrothermal transfer in the roof and footwall, and the second is magma deeply embedded in a purely conductive medium. R apid convec tive coolin g The average temperature (T) in a rapidly convecting layer, strongly cooled from its margins, especially from above, is given by
6. 07.6 .5. 1.(i)
h t i – 1=b T ¼ TL þ To 1 þ
½21
where TL is the liquidus temperature, To is the amount of superheat, t is time, and is a group of parameters that characterizes the overall cooling time. The value of is given by ¼
L2 Ra – b 2Cb 9
½22
The symbols used here are as defined above, except for C and b, which come from the parametric relationship (Nu ¼ CRab) between the rate of heat transfer, as measured by the dimensionless Nusselt number (Nu), and Ra. (Nu measures the magnitude of heat transfer by convection relative to that by conduction.) For high Ra convection, typical values of these constants are b ¼ 1/3 and C ¼ 0.4 (e.g., Turner, 1973; Marsh, 1989). The thermal diffusivity (k9 ¼ Kc/Cp) derived here contains mixed properties; the thermal conductivity (Kc) is of the wall rock and the density () and specific heat (Cp) are of the magma, but a representative value of about 102 cm2 s1 is still appropriate. The drop in temperature, as calculated using eqn [20], is shown in Figure 20. Even in a body as large as the Sudbury impact melt sheet, which may have contained 35 000 km of magma, albeit in a thin sheet (3 km 200 km), the superheat could have dissipated in 10 years. This reflects the great efficiency of thermal convection in removing heat from a layer when there is little resistance to heat flow at the margins. For a body buried deeply in a conductive environment, as in continental crust, cooling is significantly slower. 6.07.6.5.1.(ii) Slower convective cooling in a conductive medium For this extreme, consider
a melt sheet sandwiched between upper and lower rock units that transfer heat only by conduction. To assure that this is the slowest possible rate of convective cooling and to simplify the problem, assume
Magmatism, Magma, and Magma Chambers
305
Temperature (T/Tf)
No crystallization
Superheat Vigorous convection Liquidus No convection Solidus
700° C Vigorous
convection Years
Long Solidification Time 105
Time (kt /L2)
Time
Sudbury impact melt sheet
104 time(yrs)
Rapid loss of superheat
Temperature(T/Tf )
Superheat (no crystals) vs no superheat (crystals)
103 102 10 1 3 10
5 104 10 Thickness (cm) 10L L
Sheets at rifts and ridges
Figure 20 Modes of magmatic cooling for superheated magmas, like impact melt sheets (thermal convection), and for all other magmatic bodies (conduction). The scales are in nondimensional temperature vs nondimensional time, where Tf is the final temperature, t is time, L is body half-thickness, and K is thermal diffusivity.
that the upper and lower layers of country rock are infinitely thick. The temperature of the magma as a function of time is given by (Carslaw and Jaeger, 1959; see also Marsh (1989)) pffiffiffiffiffi T – Tw x ¼ exp hx þ h 2 t erfc pffiffiffiffiffiffiffi þ h t Tm – Tw 4t
½23
where exp and erfc are, respectively, the exponential and complementary error functions, Tm is the initial magma temperature, Tw is the initial wall rock temperature, x is the vertical spatial coordinate, and h ¼ Cp/MCp9, where p is wall rock density, Cp and Cp9 are the specific heats of, respectively, the wall rock and magma, and M is the mass of magma (of density 9) in contact with a unit area of wall rock. For a melt sheet of thickness L cooling from above and below where Cp Cp9, M ¼ 9L/2, and 9, whence h ¼ 2/L, the temperature of the magma over time is found by setting x ¼ 0: pffiffiffiffiffiffi
T – Tw ¼ expð4F Þ erfc 4F Tm – Tw
½24
where F ¼ kt =L2 . Since the magma is always well mixed and of uniform temperature, the contact temperature is also the overall magma temperature. In this example, the magma reaches its liquidus temperature in about 10 000 years; this variation in temperature is shown by Figure 20. From these two bounding examples, it is clear that highly vigorous convection in magma lasts a relatively short period of time, probably between a
few tens of years and a few hundred years. Once convection ceases, the magma continues to cool solely by conduction of heat. Early petrologists thought magmas were commonly intruded in a superheated state, or at least there was confusion on this issue. And, as mentioned already, even though crystal-laden volcanic eruptions are the rule, the meaning of this feature relative to magma chamber evolution was not appreciated. Coupled with the inaccuracy appreciated of assuming the importance of cooling mainly from the roof, these misconstrued initial conditions set the stage for a century of research on magma chamber processes. 6.07.6.6 Summary of Magmatic Initial Conditions Like most Earth processes, including the formation of Earth itself, the initial conditions are well nigh impossible to ascertain. This is especially true for magmatic systems. And because the initial conditions, regarding timing and mode of delivery of volumetric inputs and, especially, the concentration of phenocrysts, are only rarely known, it is critically important not to assume initial conditions that are unrealistic. For in any physical process, identifying the correct or reasonably correct initial conditions is absolutely critical to the outcome. The assumption by early petrologists of large magmatic bodies being ‘instantaneously emplaced free of crystals’ is not only incorrect and untenable, but it is an impossibly
306
Magmatism, Magma, and Magma Chambers
difficult starting point from which to reach the observed end result of an exotically layered body. The most obvious cause of layering in intrusive bodies is the periodic influx of crystal-laden magma (e.g., Marsh, 2006), and not due to some form of periodic overturning related to thermal convection during crystallization. In fact, from all experimental evidence on analog systems and circumstantial field evidence, magmas only thermally convect if they are superheated, which is a state never observed for any endogenetic magma. Convection in response to the density contrasts of periodic injections of deliveries of magma, some containing heavy concentrations of crystals, leading to establishment of a large body, must certainly occur. This convection would clearly be periodic and related to crystal deposition and layer formation. In this fashion, the larger the body, the higher the probability of acquiring textural diversity.
6.07.7 End-Member Magmatic Systems Perhaps the most revealing and insightful means of understanding magmatic processes is to examine unusually complete, well-exposed systems where the initial conditions are especially clear. Here we will examine two such systems that in a strong sense define the end members of magmatic system establishment and evolution. The two examined below are the Sudbury impact melt sheet and the Ferrar dolerite system of the McMurdo Dry Valleys, Antarctica. When deciphering these and any other magmatic system, it is critically important to keep in mind the role of solidification fronts in forming the final product and also the major controls that the initial conditions have on deciding the final outcome. In trying to recognize the initial conditions two features must be sought. One is the concentration of large crystals in the initial magma forming the body, and the other is the number of possible injections that gave rise to the body. 6.07.7.1 (SIC)
The Sudbury Igneous Complex
The Sudbury impact melt sheet is the closest example on Earth of a massive magmatic laboratory bench experiment. This is a body that was emplaced instantaneously free of crystals. These are the exact initial conditions postulated by early petrologists. In this
context alone, it is enormously valuable to examine the end product. The SIC was produced 1.85 Ga in a matter of minutes (5) by impact of 10–12 km bolide (e.g., Grieve et al., 1991). Perhaps originally containing as much as 35 000 km3 of magma 3 km deep and spread across a crater 200 km in diameter, the melt sheet had the aspect ratio of a compact disk. Moreover, the initial temperature, which can be ascertained by several methods, was 1700 C, well beyond the liquidus (1200 C) of the magma (Zieg and Marsh, 2005). The melt was intensely superheated, destroying all existing crystals. For all intents and purposes, this is an infinite sheet of magma. The body today (see Figure 21) has been deformed by continent-to-continent collision during Grenville time, which has folded it from south to north much as one would fold a sandwich. This folding, along with glaciation, has made the stratigraphy particularly accessible as has extensive drilling associated with mining (1.5 billion metric tons) of massive deposits of Ni–Cu sulfide ores along the crater floor. The final section of igneous rock is an upper layer of 2 km of granophyre (70 wt.% SiO2) and a lower 1 km layer of norite (57 wt.% SiO2) separated by a transition zone of quartz gabbro of several hundred meters, which is the densest rock in the entire section (Figure 21). The most surprising aspect of this body is the degree of homogeneity of the norite and granophyre. The same stratigraphy is laterally continuous across the entire body. There is no modal layering anywhere. Trace element and isotopic ratios are much too similar to ascribe to pure chance or a protracted igneous process. For example, isotopic ratios, such as 87Sr/86Sr (Dicken et al., 1999) and La/Sm, Gd/ Yb, and Th/Nb, are nearly identical in the two main units (Lightfoot et al., 1997). Although sill-like in form, which is similar to other large bodies, like Bushveld, Dufek, and Stillwater, it has none of the features of major sills. The average composition is 64% SiO2 but this material is not found as extensive chilled margins at the top and bottom, and there is no systematic chemical progression inward to suggest any form of differentiation by crystal fractionation. These features, among many others (see Zieg and Marsh, 2005), along with unusually uniform textural homogeneity, suggests that, once formed, the body did not evolve at all along the lines long assumed for magmatic bodies. In fact, to a close first approximation, once formed, the main units did not evolve at all. They simply crystallized in place. All of these features are likely natural consequences of the impact process.
Magmatism, Magma, and Magma Chambers
307
81° 00′
81° 30′
Lake Wanapite
SIC
i
Chelmsford Fm Onwatin Fm Black member Green member
Gray member (melt bodies) Basal member
Whitewater Group SIC
Granophyre
SIC 0
5 km
81° 30′
1
0
Sudbury
Canada
Sudbury
Southern Province
USA
0
200 km
46° 30′
46° 30′
N
Onaping Formation
Superior Province
46° 45′
46° 45′
'
Transition zone Norite Sublayer Footwall breccia Shocked and fractured footwall rocks (Sudbury breccia)
81° 00′
Figure 21 Geologic map of the Sudbury Igneous Complex, ON, Canada. At right is the general stratigraphy of the crater. After Zieg MJ and Marsh BD (2005) The Sudbury Igneous Complex: Viscous Emulsion Differentiation of a Superheated Impact Melt Sheet. Bulletin of Geological Society of America 117: 1427–1450.
Impacts of this size generate an outward moving shock wave of 500 GPa (5 Mb), which is equivalent to the pressure at Earth’s center. Temperatures reached about 4000 C at the center of the impact, vaporizing the impactor and adjacent silicate crust. Beyond the vaporization zone an extensive melting front existed and beyond this a fracture front. The production of breccia of all sizes (tens of meters to nanometers), in states of vapor, liquid, and solid, characterizes all aspects of the cratering event. As the impactor penetrated downward, it produced within a few minutes a transient cavity 30 km deep and 90 km in diameter. The target materials from which the breccias were made consisted of highly evolved continental crust. The crust was granitic overall, but in detail it consisted of granitic plutons, gneisses, swarms of diabase dikes and sills, gabbroic plutons, quartzose sediments, and untold other lithologic varieties. These materials form an extensive 3-km-thick sequence of ‘fallback’ breccia (Onaping Formation) that cascaded in from the crater rim and atmosphere as the crater relaxed over a few minutes into the final shallow impact melt sheet (6 200 km). The magmatic part of the sheet is the molten breccia (Zieg and Marsh, 2005). This molten breccia is more properly described as a magmatic viscous emulsion. That is, it consisted of blobs of silicate melt of a wide spectrum of chemical composition and thus density and viscosity. Although certainly not chemically immiscible, small density
and compositional differences gave rise to an exceedingly heterogeneous melt where each blob had the freedom to sink or rise. Large volumes of near-identical composition rapidly coalesced into an extensive ‘continuous phase’ within which were other smaller chemically and physically distinct blobs, which formed a ‘dispersed phase’. Because of the inherent granitic nature of the crust here, the continuous phase was granitic and the dispersed phase basaltic. Once the sheet formed, this superheated magmatic emulsion immediately began to separate and coalesce into an upper granitic layer and a lower basaltic layer. The separation and coalescence process was rapid, taking place within a few years. Although a rapid process, this process also allowed extensive intimate chemical exchange between the dispersed and continuous phases. Melt parcels smaller than a certain critical size (1 cm) rose or fell so slowly that they were chemically resorbed by diffusion into the surrounding melt. Moreover, the vast surface area available to chemical exchange allowed the entire melt sheet to chemically interact by diffusion during coalescence of the emulsion. And once the two layers had separated, each superheated layer went into vigorous thermal convection, which thoroughly homogenized each layer. Prior to this time, the strong upward and downward flow of melt parcels had suppressed thermal convection. This convective phase of cooling lasted at most a few tens of years (Zieg and Marsh,
308
Magmatism, Magma, and Magma Chambers
2005). The net result is two layers of strongly contrasting chemical composition that have intimately and thoroughly exchanged trace chemical components to a high degree and, then, each has been thoroughly homogenized internally by exceedingly vigorous convection. The high temperature and the vigorous convection stripped the entire body of volatiles, making it unusually dry for magmas of these compositions. Once the superheat had been dissipated from each layer through thermal convection, convection ceased and crystallization commenced in the form of inward propagating solidification fronts from the top and bottom. These fronts were unusual in that the contact temperatures at the base and roof were each near the solidus temperature of the respective magmas, granophyre and norite. This pinned the solidus at or near the boundaries while the liquidus, defining the leading edge of the front, freely propagated inward to meet, perhaps coincidentally, at the juncture of the two layers. This allowed for unusually thick solidification fronts to form. This made thick crystal mushes at the top and bottom that acted to retard the descent and rising of solid breccia debris from, respectively, the roof and floor. Rafts of Onaping breccia foundering from the roof fall in the less dense granitic magma and descend until reaching the lower layer of basaltic magma. At the same time, rafts of breccia and individual large blocks from granitic plutons forming the crater floor rose until encountering the upper granitic magma. The net result of the movement of all this SiO2 wt.% 55 65
45 1500
debris is a zone of collection at the interface between the two layers. This debris, some of which further disaggregated and dissolved in surrounding magma, and mafic melt squeezed from slight compaction of the lower solidification combined in the end to make the transition zone. The collection of all this debris formed an unusual environment in terms of chemical compositions and also oxidation state and collection of volatiles. This debris carried hydrous minerals that dehydrated upon heating, further shattering the breccias into smaller fragments and freeing volatiles to collect and form miarolitic cavities. This all produced an oxidized environment of heterogeneous composition of strong density contrasts that may have allowed internal gravity waves to propagate along this interface. The culmination of this collection process occurred as the lower solidification front reached the interface, structurally supporting this unit, which is the densest horizon in the entire body. This strong density gradient and overall structure of the transition zone is remarkably consistent everywhere it has been measured (see Figure 22). The net result of generating a large body of magma instantaneously and free of all crystals is to produce two essentially homogeneous layers with no internal modal sorting or layering. This is a dramatic confirmation of the null hypothesis, which states that given a sheet-like magma free of crystals it will crystallize to homogeneous rock with only slight internal variations (Zieg and Marsh, 2005). Magmas do not become exotically layered and show strong
75
1000
Granophyre
SiO2 Height (m)
500
Granophyre Transition zone
0 –500
Quartz gabbro
–1000
Mafic norite
–1500 2.6
Norite
Norite
2.8 3.0 Density (g cm–3)
3.2
2.6
2.8 Density g (cm–3)
3.0
Figure 22 The variations (left) in density and bulk rock silica content through the Sudbury impact melt sheet, showing the granophyre and norite units, and a detailed view (right) of the density structure in the transition zone where the profiles have been normalized for position to the maximum density at each drill hole location. The remarkable similarity throughout the body suggests a delicate mechanical equilibrium.
Magmatism, Magma, and Magma Chambers
differentiation trends from basaltic to granitic by crystal fractionation within such bodies. Then, how does it happen? It happens by injections of magma laden with large crystals. The extent and degree of layering is a direct indication of the extent of injection of crystal-laden magma. The Sudbury record is also valuable in showing what happens to true granite target rock once it is heated beyond its liquidus and then allowed to crystallize as a dry high silica magma. The surprising result is that the final texture is not a coarse-grained typical granite. It is a fine grained granophyre. This is likely due to the loss of its textural template due to superheating and higher viscosity, and thus slower diffusion rates, under anhydrous conditions. Sudbury is also a prime example of how the continental crust became organized early in Earth history. Ongoing impacts extensively melted the fledgling crust, allowing it to continually reorganize through emulsion sorting and coalescence. As mentioned above, Sudbury is critically important in clearly showing what happens to magma when the initial conditions are explicitly known. In this regard, it is an end-member example. Many thinner diabase sills the world over are examples of this type of magmatic body. 6.07.7.2
Ferrar Dolerites, Antarctica
Another end member is represented by systems that are multiply or serially injected, partly with magma literally choked with large crystals. An excellent example of such a system that is unusually well exposed is the Ferrar dolerites of the McMurdo Dry Valleys, Antarctica. The Ferrar dolerites, like similar systems worldwide, were emplaced with the breakup of Gondwana 180 Ma. Instead of the magma occupying the central rift and being eventually lost with widening of the rift, these magmas were emplaced in the stable west continental shoulder of the rift. The magmas were mainly injected into a thick section of Mesozoic sediments called the Beacon Supergroup, consisting of sandstones and associated sediments including coal deposits. The overall package dips slightly (5 ) to the west and is remarkably unfaulted, intact, and exceedingly well exposed. The McMurdo Dry Valleys are Earth’s most ancient landscapes and have remained since the Early Tertiary essentially unchanged (e.g., Wilch et al., 1993). Deep and long east–west river-cut valleys expose 4 km vertical sections over and area of 10 000 km2 revealing
309
four major mafic sills each 300 m thick. Sill composition is broadly similar to dolerite/diabase sills of this age and type emplaced with the Gondwana breakup worldwide. And with 55 wt.% SiO2, they are similar compositionally to the norites of the lower part of the Sudbury melt sheet, but the overall result could not have been more different. The fundamental reason for this difference is the presence of a rich abundance of large orthopyroxene (Opx) crystals and the piecemeal injection of the system as sills rather than as a single large body (Marsh, 2004). A general geologic map of the Dry Valleys showing the distribution of the dolerite sills is given by Figure 23 (Marsh and Zieg, unpublished). The foursill system is capped by extensive flood basalts called the Kirkpatrick Basalt (e.g., Fleming et al., 1992), which are directly connected in age, composition, and field relations to the underlying sills. The general stratigraphic layout is shown by Figure 24. Besides the unusually clean exposures, the singularly valuable feature of this system is the presence of an extensive distribution of large Opx crystals in the lowermost sill, the Basement Sill, which can be used as tracers to follow sill emplacement and solidification. These Opx crystals in every way play the same role here as do the olivine tramp crystals in Hawaiian lavas and lava lakes. They are crystals from earlier magmatic events deeper in the underlying magmatic mush column, and also possibly from subcontinental mantle wall rock, that have been entrained by the rising magma. Sills and dikes containing similar concentrations of phenocrysts have been long recognized (e.g., Drever and Johnston, 1967; Gibb and Henderson, 1996), and Simkin (1967), building on an idea by Baragar (1960), generalized the concept of flow differentiation to suggest a sequence of ascent and emplacement as sills of crystal-laden magma. This Simkin sequence is also shown in Figure 24. A fundamental property of particles in ascending fluid is the tendency to sort themselves by buoyancy and drag (i.e., size, shape, and density contrast) relative to the walls containing the flow. Heavy particles move away from walls and fall, relatively speaking, in the ascending fluid establishing a central high concentration or tongue of particles. The mechanics of this process is not entirely clear, but is associated with small inertial effects in an otherwise essentially inertia-free flow. Because the crystal tongue falls progressively behind the leading tip of the ascending column of magma, the leading magma and the magma on the margins is the most refined, having been stripped of crystals, and is positioned to cool
310
Magmatism, Magma, and Magma Chambers
M
ille
10 km
D eb en ha m
l
G l
rG
N
ct Vi ia or y lle Va
Bar wick
Vall
ey
Preliminary geologic map Ferrar dolerites L.Vida
ey Vall m a h Bal McKelvey Valley
Labyrinth
Cover
ss Pa ll Bu
OLYMPUS RANGE
L. Vanda
Dry Valleys, Antarctica
Wr
igh
tV
Ice Rock
y alle
Young basalts
Dais Sills Upper Sills
ASGAARD RANGE
Labyrinth Sill Peneplain Sill
Basement Sill
Pea
rse
Country rock Beacon sandstones ‘Granite’ and dikes Val
Friis Hills
ley ey r Vall Taylo
Solitary Rocks Kukri Hills
Taylor Gl Cavendish Rocks
B. D. Marsh and M. J. Zieg
Dept. Earth and Planetary Sciences Johns Hopkins University Baltimore, Maryland 21218 2004
Ferrar Gl ocks dral R Cathe
Figure 23 The distribution of dolerite sills (Ferrar dolerites) throughout the McMurdo Dry Valleys of Antarctica (Zieg and Marsh, unpublished).
and solidify first after emplacement. That is, once ascent stalls and the magma begins to spread laterally, the leading crystal-free magma coats and essentially cauterizes the wall rock as the sill wall
rock is progressively opened. All the crystal-free magma about the sill margins is quickly chilled into fine-grained rock. The tongue of crystals is uniformly deposited along the lower solidification front,
Magmatism, Magma, and Magma Chambers
311
Time and intensity
Injection history Sill filling pulses
Individual sill events Entire Ferrar event
Dry valleys, Antarctica 3–4 km
Kirkpatrick Basalt Beacon sandstone Mt. Fleming sill
Asgard sill Peneplain sill
After Simkin (1967)
Kukri peneplain
Basement sill
Opx Tongue
Irizar granite
150 km
Stagnation
Flow
Figure 24 The stratigraphy of the dolerite sills of the McMurdo Dry Valleys, showing (left) the process of filling a sill by magma laden with entrained crystals, as envisioned by Simkin (1967). The inferred time sequence of filling of individual sills as a series of pulses leading to a prolonged periodic episode of sill formation is depicted at the top.
forming a classic S-shaped profile after solidification. Seminal fluid mechanical experiments realistically depicting this process have been performed by Bhattacharji in 1963. This Simkin sequence of emplacement is found in the Basement Sill, where an extensive Opx tongue is found throughout the Dry Valleys. Although the locus of emplacement of sills is rarely found, the distribution of the Opx tongue gives direct insight into this point. The Opx tongue diminishes in thickness in all directions outward from the general area of Bull Pass, a pass in the Olympus Range forming the north wall of Wright Valley. From this area, a series of petal-like lobes propagated outward from a central funnel-like magmatic feeder zone to establish the Basement Sill. Profiles of MgO concentration through the Basement Sill and the upper sills are shown by Figure 25. The chilled margins and the leading tip of the Basement Sill all have low MgO contents of 7 wt.%, similar to Hawaiian basalts free of large crystals, but the composition at the sill center reaches about 20 wt.% MgO, reflecting the high concentration (40 vol.%) of
entrained Opx crystals. The shape of the MgO concentration profile is parabolic and the sizes and abundance of Opx crystals are also of this form, which is due to the emplacement flow field. Upward in the sequence of sills, the maximum MgO content decreases systematically (Figure 25), with less and less indication of Opx involvement. The compositional progression blends into the composition of the overlying Kirkpartick flood basalts. Coupled with the field relations, the overall system appears to have developed from the top down with the flood basalts first arriving from a plexus of regional dikes. Thickening of the basalts progressively capped the dikes forcing later arriving magma into sills that intruded increasing deeper in concert with the magma density. The Basement Sill intruded last in a slow sluggish fashion due to its high crystal content. This slow ascent and lateral emplacement was aided by heating of the country rock by earlier magma. That the emplacement process was very likely pulsatile is indicated by internal irregularities in the MgO profiles.
312
Magmatism, Magma, and Magma Chambers
1
Mt. Fleming Sill
0.8
Relative height
0.7 0.6
Asgard Sill
0.9
0.5
PPS Pandora's Spire 0.4
Labyrinth
0.3 0.2
West Bull Pass
0.1 0 0.00
2.00
4.00
6.00
8.00
10.00
12.00
14.00
16.00
18.00
20.00
MgO (wt.%) Figure 25 Variations of bulk rock MgO (wt.%) through the McMurdo Dry Valleys sills showing the large concentrations in the Basement Sill, reflecting the high concentration of entrained orthopyroxene, and the monotonically decreasing average composition in each higher in the sequence.
Stopping and restarting the emplacement process, as in volcanic reposes, cause resorting due to the steady advance of the upper and lower solidification fronts. Everywhere within the Opx tongue are clear signs of further detailed crystal sorting between Opx and plagioclase. Plagioclase is the other major mineral phase in the Opx tongue, but it differs greatly in size as it nucleated and grew mainly during ascent and emplacement. It is generally 10 times smaller in grain size than the Opx (2–5 mm). This large contrast in size allows plagioclase, given the chance, to sieve through the Opx and collect in pockets and layers. In most places, this sorting reflects the shearing nature of the magmatic flow and plagioclase forms thin undulating horizontal stringers or schlieren. That is, in shearing of dense particle-laden slurries, the flow dilates as the larger particles attempt to move past one another. The dilation allows the smaller particles to sieve into the shear zones and form deposits or schlieren. There are many processes of this nature that occur in granular materials and all of them lead to sorting or layering of the final assemblage (e.g., Savage and Lun, 1988; Makse et al., 1997). Signs of these processes are everywhere in the Basement Sill
and they culminate in the Dais section in central Wright Valley. Here the Basement Sill literally cascades downward and ponds at the deepest part of the body to form an intricately layered complex called the Dais Intrusion (Marsh, 2004). The layering ranges between Opx-rich zones, called orthopyroxenite or websterite, to almost pure anorthosite layers up to 0.5 m thick and laterally continuous for hundreds of meters. This general style of layering between Opx and plagioclase is common in many large layered intrusions like Bushveld, Stillwater, and Ddufek, but here it has been preserved in an unusually pristine form due to the relatively rapid cooling of a body of this size. Cooling and solidification here took place overall in about 2000 years, but much quicker locally. In the noted large-layered bodies, cooling and solidification, like at Sudbury, took hundreds of thousands of years, during which time the initial textures annealed, destroying detailed indications of how the layers were formed. In most cases, the cause of the layering has been ascribed to internal nucleation and growth of crystals that have been resorted due to magma flow, but not as an obvious result of injection of a crystal-laden slurry.
Magmatism, Magma, and Magma Chambers
Considered overall, this magmatic system is chemically similar to Hawaii. Plots of CaO versus MgO for Hawaii and the Dry Valleys Ferrar system are shown by Figure 26. Each system shows strong MgO control at high concentrations of MgO due to the mechanical addition or loss of olivine (Hawaii) and orthpyroxene (Ferrar). The other major difference in these two systems is in the context of the samples. All of the Hawaiian samples are of erupted lava, and only magmas with less than about 50 vol.% crystals can be erupted. Where exactly they came from in the underlying magmatic column and how they achieved their final chemical state is unknown, although it can be conjectured through geochemical methods. In the Ferrar case, the context of each sample is known exactly. In short, all of the rocks richer than 7 wt.% MgO come from the Opx tongue, with the richest in MgO coming from the central part of the Opx tongue, and all rocks with less than 7 wt.% MgO come from either the chilled margins or the upper sills that are free of entrained Opx. The mechanical and chemical processes by which these compositional states were
16
313
reached can be traced by the spatial context of the rocks themselves within the overall system. In this sense, it is revealing that none of the upper sills, nor the Kirkpatrick Basalt, carry any Opx phenocrysts and there is no other physical indication of the role of Opx in their evolution. It is clearly evident in their chemical compositions, being so similar to the fractionated compositions of the Basement Sill, that these magmas are the product of an underlying magmatic mush column dominated by Opx. An additional fortunate characteristic of the Ferrar system is the occurrence of four separate interconnected massive sills and a contiguous overlying sheet of flood basalts. The overall structure of the system is a stack of sills forming a fir tree-like magmatic system. This has allowed substantial volumes of magma, each perhaps representing a volcanic episode, to be serially emplaced as separate aliquots that cooled relatively rapidly to preserve an accurate magmatic record. If, on the other hand, all of this magma had gone to fill a single reservoir, with the time between each episode being short relative to the full solidification time, a massive layered intrusion would have resulted and all the revealing textures would have been lost to annealing (see Figure 27).
14
6.07.8 Lessons Learned from Sudbury and the Ferrar Dolerites
CaO (wt.%)
12 10 8
6 Ferrar dolerites 4 2 MgO (wt.%)
0
0
5
10
15
20
25
15
CaO (wt.%)
12
9
6
Hawaii lavas
3
MgO (wt.%) 0 0
5
10
15
20
25
Figure 26 A comparison of the field of compositions (CaO versus MgO in wt.%) of the Ferrar dolerites and the lavas from the island of Hawaii.
Magmatic systems are made of individual batches of magma delivered in amounts and over time as expected from volcanic processes. If the accumulated body is made of crystal-free contributions and is roughly sheet like, it will crystallize to an essentially uniform mass of rock, regardless of size. In the strictest sense, there will be many local chemical and textural aberrations due to untold local hydrothermal and solidification front processes, but these are second-order features. Strong textural and chemical variants come from variations in the nature of the injected magma. The bigger the body the more injections necessary to build it, and the more chance there is for strong variations in the crystal content of the incoming magma. Pervasive and exotic layering, clearly, reflects the deposition, sorting, and subsequent chemical annealing, including refinement of the incipient texture (e.g., Boudreau, 1994), of large crystals carried into the body by incoming magma. The old adage that only large bodies cool slowly enough to grow crystals large enough to become layered is untenable. Shonkin Sag at 70 m and the Basement Sill at 330 m negate this
314
Magmatism, Magma, and Magma Chambers
Repeated Injection
Ocean ridge magmatic system
Country rock
Granophyres Unstable SF New injection Earlier roof SF Country rock
Injection history
Time and intensity
Upper solidification front OPX Tongue Earlier OPX tongue with aorthosite Basal solidification front
Layered intrusion for stationary chambers Sill magma
Gabbro
Moho Figure 27 A schematic depiction of the repeated injection of basaltic magmas into a single reservoir where the crystal content can vary strongly from one injection to the next. In the end, this leads to an intricately layered pluton.
premise, whereas the many featureless sills and plutons throughout the world are like this because of the phenocryst-free magma from which they were built. The common occurrences of featureless granitic intrusions achieve this state because of the inefficiency of crystal sorting in highly viscous magma. In essence, the final state of most magmas is not too far different than their state. With this basic understanding of magmatic systems gained from systems well exposed and relatively straightforward to understand, it is useful to extend this insight to other magmatic systems. The two most common and extensive systems of Earth are those represented by Ocean Ridge and Island Arc magmatism. These systems are mainly studied from a purely chemical point of view that, in and of itself, can be misleading as to the physical basis behind the outcome, and the following will emphasize more the physical workings of these systems.
6.07.9 Ocean Ridge Magmatism In concert with the classical concept of magma chambers as giant vats of magma undergoing slow progressive cooling, crystallization, thermal convection,
Mantle Figure 28 The mush column magmatic system at ocean ridges. The system is characterized by a thin sill (axial magma chamber, AMC) capping a vertically extensive system of thin sills and conduits within a massive column of mush. The sheeted dike complex at the top records the history of extraction of magma from the AMC as the plates suddenly pull apart over time periods depicted at the top. Note the small plagiogranite lenses within the upper gabbros formed by solidification front instability.
and crystal deposition into systematic layers reflecting the history of the system, vast reservoirs of magma were expected at the level of the oceanic crust beneath ocean ridges. When these features were sought using seismic methods, they were not found. Instead, thin (50–100 m), wide (2–3 km) ribbons of magma, or sills, were found (Sinton and Detrick, 1992). This spawned a general model of ridge magmatism more consistent with what is found for systems like the Ferrar and even Sudbury, albeit on a smaller scale, and the structure of the oceanic crust is an intimate reflection of the combined process of magmatism in response to plate tectonics. This general system is shown schematically by Figure 28. This depiction stems from the general form of magmatic sill complexes found in the Ferrar system, in old continental crust, in ophiolites (e.g., Nicolas, 1995), and in three-dimensional multichannel seismic studies in the North Atlantic (e.g., Cartwright and Hansen, 2006), and it is also
Magmatism, Magma, and Magma Chambers
consistent with what should be expected on a mechanical basis in this tectonic region. This is a magmatic mush column under steady-state operation in response to mantle upwelling associated with large-scale mantle convection. The mush column stands in partially molten mantle ultramafic rock modally dominated by olivine with subordinate orthopyroxene, clinopyroxene, and plagioclase, with spinel and pyrope-rich garnet at successive higher pressures. At the deeper levels (50–100 km), significant partial melting due to adiabatic decompression (McKenzie, 1984) forms an anastomosing complex of melt veins and channels concentrating upward into a main mush column stalk with periodic sills. The mush column is capped by the axial magma chamber beneath the ridge axis. The system works through hydrostatic head produced in response to, in effect, suction at the top associated with the abrupt splitting and spreading of the lithospheric plates. From an initial state of nearhydrostatic equilibrium, this abrupt motion withdraws magma from the underlying axial magma chamber, some of which erupts as pillow lavas and that in the feeder solidifies as a dike. The loss of hydrostatic equilibrium at the head of the system propagates downward throughout the system, perhaps partly as solitons, drawing magma upward in the mush column and reestablishing stability. Because all parts of the contiguous ridge plates do not move at exactly the same time and to the same degree, the process of melt motion at depth is certain to be complicated. Magma at some depths may at various times be pulled laterally as it ascends, making its overall trajectory significantly nonvertical. The intimate history of this process is recorded in the vast sheeted dike complex, which is a characteristic byproduct of ridge magmatism and is known so well from ophiolites. Because they are random samples of the axial magma chamber, they give an excellent inventory of the general state of this magma, both in terms of composition and crystallinity. These dikes are typically fine grained and of low phenocryst content; the phenocrysts are also mostly plagioclase. This information strongly suggests that the axial sill is a passive body of magma that experiences bursts of withdrawals followed by recharging from below. The basalts erupting at the ocean ridges are tholeiitic basalts, with low K2O (0.2 wt.%), modest TiO2 (1.5 wt.%), and low and unfractionated rare earth elements (REEs), and are commonly called mid-ocean ridge basalts (MORBs). To first order, these basalts are chemically among the most globally
315
uniform of any class of magmas. These are called N-MORBs for normal MORBS. But at a more detailed level of inspection, they show significant variations in many facets, the most common of which is enrichments in K2O and TiO2 and sometimes iron. These are called E-MORBs or enriched MORBs and also sometimes Fe–Ti MORBS. The first-order bulk composition of these basalts reflects the underlying magmatic process of prolonged intimate contact with a vast solid assemblage in the magmatic mush column. In effect, the chemical composition is buffered by the mechanics of the overall process, much as in a water purification system. This uniformity is also a result of the uniformity of mantle composition itself, which has probably not changed drastically over Earth history. That is, at the present rate of ridge magmatism (20 km3 yr1), the mantle will be recycled about once every 50 Gy, which, even allowing for the role of contamination by subduction, suggests the mantle composition has been, to first order, fairly constant. But as a result of initial inhomogeneities, subduction, and other styles of magma generation, there are certainly subtle and significant chemical variations in the mantle source rock. In addition to these variations, there are also regional variations in the thermal regime as a result of the form of mantle convection. These effects cause the depth of the principal melting region to vary in extent and depth, which altogether causes distinct fluctuations in the bulk composition of the magmatic product. Another major factor is the rate of spreading, which may strongly affect strength of the melting anomaly and thus the overall mass and activity of the mush column. It is interesting to consider ridge magmatism from the past point of view of classical magma chambers, where, once emplaced in a large vat, the magma differentiates by crystal growth and settling. This strong uniformity of composition of MORBs is explained by starting with MgO-rich magma produced by partial melting of peridotitic mantle to a degree necessary to produce a large mass of mobile magma. This process would produce a series of magmas ranging from picrites to olivine basalts and tholeiites. Yet picrites (olivine-rich basalt) are not found at ridges, and arguments always persisted that if picrites are parental to MORB then why are they never seen? Why don’t they erupt more often, or at all? This is not an uncommon situation in petrologic studies where critically important parts of the hypothesized dynamic puzzle are not seen, but are postulated to be in a hidden zone, or uneruptible due
316
Magmatism, Magma, and Magma Chambers
to density difficulties, or other factors. On the other hand, to others, the fact that critically important magmatic parts of the system are never seen means simply that they do not exist. In the present understanding of ridge magmatism, the magmas are chemically fractionated by contact with olivine, and other phases, but it comes through intimate, diffusional, contact within the mush column. The lack of presence of an actual powerful magma chamber, as found on chemical grounds, is replaced by a virtual magma chamber, which is an integrated chemical process taking place over the full extent of the mush column.
Where there is no subduction there is no volcanism. There are also some close similarities and these are in the spatial trace of volcanism on the surface; both are long, sharp narrow welts, and the compositions of the most voluminous magmas are to first order similar, tholeiitic basalt and HAB. But unlike at the ridges, where there is no choice in the ultimate source of the magma, in arcs a wedge of asthenosphere sandwiched between the arc lithosphere and the subducting oceanic crust provides almost any number of end members and combinations, with and without continental crust, from which to get magma. As at ridges, the main approach to understanding arc magmatism has been, and is, mainly through geochemistry, but here the focus will be more on the fundamental geophysical controls. Moreover, petrologists have long assumed, as based on early thermal models, that the subducting plate is too cold to melt and that the source of arc magma must be from the mantle above the plate, forming in response to fluxing by volatile rising from the dehydrating plate. A perusal of the principal volcanic arcs of the world (Figure 29) shows a diverse range of geologic
6.07.10 Island Arc Magmatism 6.07.10.1 Introductory Arc magmatism contrasts fundamentally with ocean ridge magmatism. The points of volcanism, the volcanic centers and surrounding crust, can remain fixed for long periods of time (1–10 My, depending on location), and the flux of magma is much lower (102 km3 yr1). Also, there is a universal relationship between the rate of subduction and volcanism.
Subduction zones Plate boundaries Plate motion Island arcs
No
r th
Aleutian arc
Eurasian plate Am
er i
ca
np lat
e
Kamchatka arc Cascades arc
Kurile arc Japan arc
Pacific plate
Izu arc
Lesser Antilles arc
African plate
Mariana arc
Central American arc
Philippine arc
Indonesian arc
Indo-Australian plate
Solomon arc New Habrides arc
Nazca plate Lau basin
South American plate
Tonga arc South American arc Kermadec arc
New Zealand arc
Scotia plate Scotia arc
Antarctic plate Figure 29 Islands arcs of the world.
Magmatism, Magma, and Magma Chambers
environments. There are arcs built on thick continental crust in South America, arcs built on oceanic crust in the Scotia arc southeast off South America, arcs traversing from continental to oceanic crust as in the Aleutian Islands, and arcs on young mixed crust in Kamchatka and Japan. The ages of the arcs also vary a great deal from those like the Aleutians that may be as old as 40 My to those as a young as the Scotia arc at 3 My. This is not to say that the arc has been in continual operation for long periods of time. In fact, from the evidence of ashes deposited on the sea floors around arcs, there is an apparent periodicity of eruptive activity, especially around the Pacific, of about 2 My (e.g., Hein et al., 1978; Scheidegger et al., 1980). Coupled with a record of continuous subduction over much longer periods of time, this may indicate that the arc systems periodically load with magma and then discharge over and over again. The force of the volcanism, even within a single arc, can also vary significantly as can the nature of the dominant lavas. In South America, there are major gaps along the strike of the arc where there is no volcanism. These correlate closely with shallowness in the angle of subduction. If the plate subducts too shallow, essentially adhering to the keel of the continental lithosphere, there is no wedge of asthenosphere and no volcanism. The depth of the top of the subducting plate is most commonly 125 km, but it can be as large as 150–160 km. And the detailed structure of the volcanic front of any arc often intimately reflects the structural morphology of the subducting plate. Arc volcanic fronts are not simple continuous arcs, but are almost always segmented into a collection of individual short segments forming a piecewise continuous arc. And the arc segments are usually bounded by major fracture zones or structural irregularities in the subducting plate.
6.07.10.2.1 centers
Spacing of the volcanic
The basic structure of island arcs in terms of the positions and form of the volcanic centers is exceptionally clear. The arc begins with a single sharp line of volcanoes. This is the volcanic front, as named by A. Sugimura (1968), which is, by far, the principal locus of volcanism. The front fills in to a series of unusually regularly spaced volcanic centers separated by 50–65 km. The relatively young Scotia arc shows this basic developmental form rather clearly (see Figure 30). After existing for about 3 My, secondary volcanic centers may appear behind the front, forming a weak secondary front. These are always smaller and spaced closer to the front than the principal spacing (d ) of centers along the front. There is some sign that this secondary spacing (d9) follows the relation d9 ¼ d cos ( ), where is the angle of dip of the subducting plate (Marsh, 1979a, b). Leskov Is. in the Scotia arc and Amak and Bogoslof islands in the Aleutians are prime examples of these secondary centers. Every arc seems to have them, although
Zavodovski Is.
Leskov Is.
Visokoi Is. 57° Candlemus Is.
Scotia arc Saunders Is. 58° South Sandwich Islands Montagu Is.
6.07.10.2 Arc Form There are two spatial characteristics of island arc volcanism that set it aside from all other forms of volcanism. These are the regular spacing of the arc volcanic centers and the segmentation of the volcanic front. Although when built on continental crust, the form of island arcs are can appear cluttered and unclear, it is always there and clearest in young and tectonically uncomplicated regions.
317
Bristol Is.
59°
Thule Is. Cook Is. Figure 30 Detailed view of the Scotia Arc or South Sandwich Islands. Note the prominent volcanic front, the even spacing of volcanic centers, and the location of the secondary volcanic center, Leskov.
318
Magmatism, Magma, and Magma Chambers
other forms of volcanism especially on continents sometimes complicates identification. On a similar timescale, young centers also appear within the main front at points approximately midway between the older initial and well-developed centers. These centers never reach the strength of the main centers. Bobrof Is. and the doublet Koniuji–Kasatochi islands in the central Aleutians are such late centers.
the correlation of large dots on a map. The alignments are particularly striking and delicate if viewed from the exact summit of an active vent. Stepping away 10 m from a vent within a segment will lose the alignment, which can often be seen for 200 km or more in each direction. Considering that the principal active vent within any center usually migrates slowly (10 km My1) in time away from the original front, that the alignment exists at any time is suggestive of a deep-seated fundamental cause.
6.07.10.2.2
Arc segmentation That the arc volcanic front can be broken up into piecewise continuous segments that may reflect a similar structure in the underlying subducting plate was noticed and developed by Richard Stoiber and Michael Carr (e.g., 1971). Every arc shows this basic form and it is especially clear for the Aleutian arc (see Figure 31). To be clear, these are alignments between only the ‘active vents’ of the front and not simply the volcanic centers, which often contain a series of waning and waxing vents. Although some segments contain only two or three vents, most segments contain many vents, and this is much more than simply
6.07.10.3 Centers
Character of the Volcanic
The general pattern of development of individual volcanic centers often shows a similar progression from initiation to final size and this progression can be arrested at any stage. These stages can be called young, mature, and supercenters. The final form, which is reflected by strength or volume of the center, depends critically on the position of the center within the front and within the arc itself. The
Siberia
Ca na da
Alaska
slo
ai tm Ka
ng ak Pe uli k
Redoubt Iliamna Augustine
n
Ch
Pu
igi
rpl
na
58°58
158°
Douglas
di
Chagulak
Carlisle Cleveland Uliaga
Akutan
Makushin
Okmok
Amlia Fracture Zone
Korovin
Koniuji
Bobrof Adak Fractu re Zone Adagdak Great Sitk in
ic Bdry Tecton
Bdry Am chitka P Tectonic
Figure 31 Segmentation of the Aleutian Island arc volcanic front.
Herbert
174°
Yunaska
Seguam
Kasaatochi
Kanaga Moffett
Tanaga
Gareloi
176°
Amukta
54°
178°
Recheschnoi Bogoslof
ik ge Ma
ak An
iak
ch
ino f m nia Ve
180°
166°
168°
170°
172°
ass
Semisopoc
hnoi
Fr os ty Pa Em vlo mo f S ns ist er
ish
Fourpeak
Vesevidof
l ah td
ture Zo Rat Fra c
60° 60
al
Kiska Segula Little sitk in
178°
ne
Sh
es W
Buldir 52°
e
f vlo
Da
Pa
M 176°
160°
Spurr
154° 156°
162°
an Ak
ut
un Ak
n
cea
ific O
Pac
164°
Kagamil
Fi Pog sh ro er m Is no an i ot sk i Am ak
56°
166°
ak
us
hi
n
go Bo
168°
Ku ka k
f
62 Ber ing Sea
Magmatism, Magma, and Magma Chambers
stronger the rate of subduction, the larger the centers. Centers deeply within a segment are generally larger than those at the ends of segments. Centers begin, as can be seen from the appearance of secondary centers, as essentially large multiple domes; the magma is of low volume, highly crystalline, and sluggish. Others begin with the eruption of a series of highly mobile basaltic lavas interbedded with thin deposits of tephra to form a low-profile shield. Continued growth adds more to the shield and then sharply transitions into a steeper-sided cone of primarily (50–70 vol.%) volcaniclastic debris trussed together with occasional lavas. Low-crystallinity basaltic flows may be thin (1 m, especially near the vent) and of limited extent (hundreds of meters), whereas highcrystallinity andesitic flows may be thick (100 m), massive, and extensive (kilometers). This basic transition from shield to steep cone forms the classic composite cone of island arc, consisting of a stratocone built on the foundation of a shield cone. Many volcanoes go no further in development. Reposes in activity allow erosion to deeply incise the cone; summit craters of tall stratocones commonly become glacial cirques that breach the crater and dig deep U-shaped valleys. In the next phase of activity, the cone reconstructs itself with the lavas and debris filling the topographic voids, making the field relations challenging. This is a mature volcanic center. Further development often results in a series of domes and fumaroles ringing the cone at an elevation of about 50–70 % of the cone relief. This is commonly encouraged by deep erosion that may weaken the cone near the main vent, allowing stagnant, highcrystallinity magma to sluggishly extrude laterally. This signals the solfateric stage of development. These locations also sometimes develop into a series of satellite cones, evicting a series of flows that may mimic development of the central cone. Supercenters develop when activity is particularly strong and the stratocone phase outstrips erosion, forming volcanoes of unusually large relief like Mt. Fuji, Mt. Rainer, and Shishaldin (3000 m) in the Aleutians. Beyond this point, the cone can become structurally unstable due to size and subsurface magma ponding and catastrophic caldera collapse often occurs (e.g., Williams, 1941; Smith and Bailey, 1968). The majority of the cone relief is destroyed through collapse and violent ejection, leaving a lowrelief, distinctive crater as at Crater Lake, Oregon. Further activity builds domes and new cones within the caldera, generally not of a size commensurate to the original cone, but sometimes even larger. This
319
cycle of growth and collapse may repeat itself many times until the locus of volcanism migrates to a new nearby location. The style of eruption for arc volcanoes, unlike for Hawaiian and Icelandic volcanoes where fissure eruptions are common, is most often from the central vent. These central vents are highly concentrated and cylindrical that may fill and empty repeatedly during a eruptive episode. This 500 m wide vent at Korovin in the central Aleutians has been observed empty to a depth of 1.5 km and later brimming with magma and then spilling over into lavas (see Figure 32). Acoustic echoing due to explosions at the base of the vent when evacuated can be heard for hundreds of kilometers. The highly explosive nature of stratocone eruptions may more reflect this concentrated plumbing system coupled with high-crystallinity magmas, making them unusually viscous, than an undue effect of high concentrations of volatiles (as is more often assumed). Large regional dikes are uncommon, although they do sometimes accompany inflation and the precursory events of attending caldera collapse. Small local dikes are common near vents and deeply eroded cones sometimes show dikes radiating from the central vent that have trussed together the volcaniclastics of the stratocone and sometimes served as conduits for flank domes and lavas. Ash flows or ignimbrite deposits, so common to continental volcanism and caldera collapse eruptions, are also common to arc volcanism. But because of their high mobility and erosiveness, evidence of these events are often difficult to find and trace, especially in oceanic environments.
Bobrof Adak Fract ure Zo Adagdak ne Great Sit kin Koniuji
Tanaga Kanaga Moffett
Kasaatochi
Korovin Figure 32 View of the detailed relations between two segments of the Aleutian volcanic front at the island of Adak and the position of the prominent fracture zone in the subducting plate.
320
Magmatism, Magma, and Magma Chambers
6.07.10.4 Magma Transport
6.07.10.5 Subduction Regime Perhaps the most unusual aspect of arc magmatism is the deep presence of the wedge flow of asthenospheric mantle driven by the subducting plate (Figure 33). At shallow depths, the subducting lithosphere is in fault contact with the adjacent arc lithosphere. The depth extent of this contact is commonly 70–100 km, but may be much longer in continental terrains where the lithosphere or tectosphere can be very old and thicker as in some places along the subduction zone of South America. The nature of the arc lithosphere is also not well known in many areas. Below this depth of fault contact, the subducting plate is in contact with the mantle wedge or asthenosphere, a high-viscosity fluid (1017 Pa s). In most regions, this is the so-called seismic ‘low-velocity zone’, reflecting incipient melting at the solidus of the regional mantle country rock. The asthenosphere in this wedge-shaped region adheres to the upper boundary with the arc lithosphere and at the same time is dragged downward by the motion of subduction. The streamlines, depicting the mass transfer, due to this dragging tilt upward in
θ
The strong centralization of arc volcanoes and the lack of dispersion on the surface of vents over periods of millions of years suggest deep magma transfer by diapirism rather than by dike propagation. Deeply eroded terrains also show bulbous plutons and a paucity of dikes. When built on continental crust, expansive welts of plutons form vast linear batholiths, like the Sierra Nevada, formed in the subsurface from foundered magma and from reprocessing of crustal wall rock. The dominant type of magma emitted is also affected by the density of the underlying crust. Arcs on continental terrains are dominantly andesitic (i.e., 60 wt.% SiO2), whereas arcs on oceanic crust are dominated by HAB (50 wt.% SiO2). This sensitivity to density contrast is locally critical in diapiric transport and much less so in dike transport. That is, standing dikes in elastic cracks operate more on an integrated, columnwide, density and the strength of the walls rather than on the local buoyancy at each horizon. The extreme paucity of mantle-like ulramafic xenoliths may also reflect the slow diapiric mode of transfer where the same conduit is used over and over, making the column traveled thermally and chemically insulated from access to primitive lithospheric rock.
Arc plate
Magma ascent path
r te
ng cti
pla
u
bd
Su
er
y elar rayy yal
r
dr udna
n obuo labl
rma m
heer
TTh
Figure 33 (upper) The flow of mantle material within the asthenospheric wedge induced by subduction, with the streams lines shown. (lower) The position and development of the cold thermal boundary layer at the slab–wedge interface; also shown is the strong thinning in the boundary layer near the corner and directly beneath the volcanic front. The inset diagram shows the trajectory of ascent of a diapir of mush from the slab as it traverses the wedge flow and repeatedly enters the same spot within the overlying arc lithosphere.
direct response to the angle of subduction. This flow field is found by solving the biharmonic equation for viscous flow, which comes from the Navier–Stokes equation. A key feature of this flow is that the associated stress field creates a traction or coupling between the two plates that, in effect, welds them together along the lithosphere-to-lithosphere fault contact. This junction is locked in place, and the volcanic front, as we have already seen, reflects this intimate connection. This is an unusual flow in several regards. First, the flow is the same at all distances from the corner. The upper part is essentially a parabolic flow between parallel plates and the lower half is a shear flow. For conservation of mass, all the fluid in the top half must turn and flow through the lower half of the flow field. Along the plane of symmetry fluid moves from one side of the flow to the other. This flow is directed normal to the face of the subducting plate. Although normally at great distance from the plate interface, with approach to the corner this flow becomes arbitrary close to the plate. It is in exactly above this corner region where the volcanic front is found. Second, since the downgoing plate is cold and the flow above it is mostly a shear flow, with streamlines parallel to the plate, a cold thermal boundary
Magmatism, Magma, and Magma Chambers
layer develops near the corner and thickens down the plate. This thickness T can be estimated from eqn [2] as a function of distance (x) from the corner, which is measured by x ¼ t/V, where t is time and V is the subduction velocity. Then, pffiffiffiffih x i1=2 T ¼ C K V
½25
where C is a constant of order 1 and K is thermal diffusivity. The boundary layer thickens downward (see Figure 33) along the plate as x1/2, and it thins or thickens inversely with V1/2. At a common distance from the corner, slow plates have much thicker boundary layers than fast-moving plates. This behavior is very much akin to the growth of the lithosphere itself as it moves away from ocean ridges. In this respect, is also critical to recognize that any diffusive process associated with the plate, including chemical diffusion involving volatiles, will have a similar boundary layer development. The third unusual feature of this environment is the corner region where the motion goes from a fault interface controlled by friction to a fluid– solid interface dominated by a shear flow; friction is also involved but to a lesser extent. In a shear flow like this, heat flow is produced according to q ¼ V , where is the local shear stress ( ¼ (dV/dy), where is viscosity and y is the coordinate normal to the plate). But within the arc plate, with approach to the surface, the effective thermal boundary thickness becomes very large. There are thus two key features in this regime that fundamentally affect the production of magma. One is that the corner flow continually brings into the wedge region hotter mantle material from greater depths, and this material is continually brought in contact with the subducting plate. It continually coats the plate and the cold thermal boundary layer is never allowed to become unusually large, as it would if the corner were stagnant. Second, the thermal boundary layer becomes pinched near the corner as the corner flow turns and becomes a shear flow. This thermal pinch is exactly below the volcanic front, and it, in effect, intimately ties the spatial position of the front to the position of the plate. The thermal regime inside the subducting plate is also central to understanding how magma might be produced. 6.07.10.6 Subducting Plate Internal State Two key processes within the downgoing plate affecting the production of magma are the thermal regime and the transport of hydrothermal fluids. The
321
oceanic crust at the surface of the plate has a strong chemical affinity to arc basalts and is clearly a prime potential location for magma production. But as mentioned already, it is commonly considered too cold to undergo melting. This impression stems from early thermal models (e.g., Toksoz and Bird, 1977) and also from the fact that the slab interface was once on the seafloor, the coldest part of the lithosphere.
6.07.10.6.1
Thermal regime This latter impression can be appreciated by recalling that if two solids at contrasting temperatures T1 and T2 are abruptly brought into contact, the temperature of the interface (Ti) immediately becomes the average of the two initial temperatures, Ti ¼ 0.5(T1 þ T2) (e.g., Turcotte and Schubert, 1982). For seafloor rock at 0 C brought up against mantle wedge at, say, 1300 C, the crust interface would thus achieve a temperature of about 650 C, which is well below the solidus. The problem with this result is that the thermal path that the slab takes between leaving the seafloor and reaching a depth of 125 km must be considered. The section where the two plates are in fault contact is especially critical as the frictional interaction produces heat and there is strong evidence that at shallow depths (0–50 km) the preponderance of earthquakes is along this interface, and at great depths the strong band of earthquakes migrates slightly deeper into the plate to perhaps a depth of 10–20 km below the interface. So, the temperature calculated here is an absolute minimum; frictional heating will preheat the oceanic crust such that when it enters the asthenosphere it will attain a significantly higher temperature. And the inclusion of the flow in the wedge is essential to achieve for any realistic estimate of the prevailing temperature along the slab interface. The full problem is quite involved and of the many studies that of Kincaid and Sacks (1997) is notable for its thoroughness (they also give an extensive review of associated seismic and tectonic studies). They find that near the corner, in the location indicated by the flow in Figure 33, there is positive temperature anomaly of about 80–100 C at the slab–asthenosphere interface; over a thin (maybe only tens of meters), the oceanic crust may be brought to a temperature of about 1350 C. This is enough to initiate melting, and that this spot is exactly below the volcanic front is important. This spot is also tied to the plate-to-plate junction and may well determine the position of the volcanic front.
322
Magmatism, Magma, and Magma Chambers
6.07.10.6.2
Hydrothermal flows Being a solid, the thermal regime internal to the subducting plate is usually considered as solely due to heat conduction and this is certainly true to first order. But there is also hydrothermal circulation in response to dehydration. This reflects extensive hydrothermal circulation starting at the ridge and extensive alteration and exchange between the seawater and the oceanic crust. This exchange takes place over tens of millions of years. The principal effect is to alter the original minerals to hydrothermal equivalents, thus partially hydrating the oceanic crust. For example, olivine is partly altered to serpentine, plagioclase feldspar to forms of sericite, and clinopyroxene to hornblende. Although this alteration can be extensive in local areas, overall these effects probably do not affect more than about 15–20 vol.% of the crust. What is much more pervasive is exchange at the isotopic level involving, for example, oxygen (18O/16O) and strontium (87Sr/86Sr), which can take place at high temperature and leave little to no visual trace. Fresh glassy MORBs erupted at the ridges commonly have, for example, initial 87Sr/86Sr near 0.702 5 and, with ocean water at 0.709 0, after about 30–40 My the oceanic crust, which has migrated to significant distances from the ridge crest, now has an 87Sr/86Sr closer to 0.704 0. These variations in depth are found in ophiolites (e.g., Gregory and Taylor, 1981) where a net increase is found in 87Sr/86Sr, but 18O is increased at the crust top (þ12 per mil) due to low temperature exchange and decreased at the crust base due to high temperature exchange such that the integrated effect is 5.8 mil1, which is the normal value of fresh MORB. Thus there is no net exchange in 18O. As the plate subducts and becomes heated, a new hydrothermal circulation system sets up due to phase transformations associated with dehydration. Because of the large horizontal temperature gradient across the oceanic crust the form of this flow is horizontal and upward, which, in effect, makes it helical. This flow is confined only to parts of the plate where there is sufficient permeability to allow flow. Similar hydrothermal flows have been studied over extensive regions about solidifying plutons and the patterns and basic mechanics of these flows are well known. Permeability is directly tied to the temperature and brittleness of rocks, and from these studies it is abundantly clear that rocks are sealed beyond temperatures of about 700 C. For this reason, this hydrothermal flow is confined wholly within the oceanic crust, which is the coldest, most brittle part of the subducting slab. Moreover, the region has been
(and is being) subjected to extreme fracturing in response to earthquakes; each cubic kilometer experiences about 10 000 events during subduction. This makes a strong permeability guide within the oceanic crust. At the same time, for these same reasons, the overlying asthenosphere in the wedge is tightly sealed; it is hot and weak and has nil permeability. Hydrothermal fluids released from the plate do not enter the overlying mantle wedge – they travel upward and repeatedly cycle across the oceanic crust (see Figure 34). This flow has a critical effect on the chemical nature of the subducting plate. This intense, internally confined flow redistributes the initial variations in 18O and 87Sr/86Sr. Because from the interaction with seawater there has been no net gain in 18O, it is reset to its normal value of 5.8. But because the crust has gained 87Sr from seawater, there is a net increase throughout the crust in 87Sr/86Sr to a value of 0.704 0. The low solubility of isotopes like 143Nd/144Nd in seawater hence does not produce a corresponding effect on the oceanic crust. But perhaps the most remarkable and underappreciated effect of this flow is to produce an unusual skarn as it comes in contact with the slab– mantle interface. The flow cannot penetrate the mantle, but this interface, due to all the ongoing faulting and shear, is certainly a mechanical breccia of both rock types, namely, high-pressure MORB (i.e., quartz eclogite) and garnet peridotite. The hydrothermal flow is in approximate chemical equilibrium with the MORB crust, but far from equilibrium with the peridotite. As soon as it touches this hot heterogeneous rock, it precipitates out phases to bring it closer to equilibrium. Although the nature of this ‘slabskarn’ assemblage is not as yet known with any exactness, it will be rich in trace elements like Ba, Rb, and elements with similar geochemical affinities. Thus the most likely area of magma production has a MORB-like bulk composition with an enrichment of some trace elements. 6.07.10.7
The Source of Arc Magma
Although the more popular source of arc magmas is from melting of peridotitic material within the asthenospheric wedge due to fluxing from volatiles escaping from the subducting plate, this is unlikely for several fundamental reasons. First, because of its high temperature the wedge is sealed to hydrothermal flows. Second, the composition of the predominant basalt found in island arcs has an exceedingly close chemical affinity to subducted
Magmatism, Magma, and Magma Chambers
323
Oxygen
θ sid
ion
sit
up
ld fie e w id o l lf rs ma lowe r the old ro C yd t Ho
87
e
r pe
r
O
Sr 8 / 6 Sr
18
o mp
ab
co
Sl
143
Nd/ 144Nd
H
Figure 34 The flow of helical hydrothermal flow up the coldest, most brittle, and most permeable part of the subducting plate. This flow redistributes the concentrations of certain isotopes and trace elements within the slab. No fluid penetrates the overlying asthenospheric wedge.
MORB and virtually none to the wedge peridotite. Third, the strong focusing of arc volcanic centers at the surface, with very little dispersion, over millions of years, links it to a magmatic source that is virtually static with respect to the plate–plate subduction configuration. This cannot be achieved through magma production within the wedge flow regime. The major elements of magma dominate the mass of the magma, take the most energy to change, and are therefore the most diagnostic deciding source characteristics. The close chemical affinity of melts from the plate to HAB, the arc parental magma, can be seen by comparing HAB melts with melts from both the plate and the wedge. This is shown by Figure 35, where melt composition is shown as a function of degree of melting at 30–40 kb (3–4 GPa). Across each panel are horizontal lines representing HAB composition and intersections or matches with melt from either of the two source rocks is indicated. No matches are found for melt from the peridotitic wedge material, but matches are almost complete for each of the major elements for slab melting. It is particularly notable that slab melts are low in MgO content. This reflects a fundamental chemical property of this source rock in that the MgO content cannot exceed about 7.5 wt.%. This is also a fundamental property of HAB. That is,
arc basalts are characteristically low in MgO. For HAB from mature volcanic centers at SiO2 50 wt.%, MgO content is typically about 4–5 wt.%. Some arc basalts are also found with significantly higher MgO, but these can almost universally be demonstrated to have suffered contamination with tramp olivine debris from lithospheric wall rock. Moreover, these basalts are also almost always from arc supercenters where there has been an unusually large eruptive flux. Large eruptive fluxes, like at Kilauea, lead to strong heating of the lithosphere wall rock, which causes structural weakening and collapse, furnishing peridotitic debris for entrainments by ascending arc HAB. Careful petrographic inspection of basalts with elevated MgO and/ or Ni commonly reveals fragments of unzoned, often slightly strained, highly forsteritic (Fo92) olivine from the lithosphere. Because of the inherent assumption that magmas of high MgO contents are primitive and thus parental to the lower MgO suite, these contaminated basalts are often mistakenly assumed to be fundamental to the arc magmatic system. A second distinctive chemical feature of arc HAB is the elevated Al2O3 content (16–20 wt.%), which is reflected by high modal contents of plagioclase in the plutons and lavas. These high plagioclase contents give arc basalts their characteristic gray appearance.
324
Magmatism, Magma, and Magma Chambers
55 50 SiO2
Qtz. Eclogite
45
Pyrolite SiO2
40
Oxide wt.%
35 MgO
30 High-Al Basalt
25
(No matches)
20
Al2O3
15 CaO 10 5
NaO
0
K2O 0
20
40
FeO MgO
60 Melt %
Al2O3
TiO2
Na2O
TiO2 80
100
FeO
20
40
CaO
60 Melt %
80
100
Figure 35 The compositions (oxides and curves) of melts as a function of degree of melting produced from quartz-eclogite subducted oceanic crust and pyrolite of the asthenospheric wedge. The straight horizontal lines show the composition of typical HAB of island arcs (Aleutians) and the large dots show matching points between the observed HAB and the calculated melts. There is almost perfect agreement with the slab source and no agreement whatsoever with the mantle wedge.
6.07.10.7.1
Slab quartz-eclogite These chemical features also reflect the basic highpressure mineralogy of the subducted oceanic crust. At pressures of 2–3 GPa, the MORB mineral assemblage of plagioclase, olivine, clinopyroxene, and Fe–Ti oxides transforms to quartz, garnet, jadeitic pyroxene, rutile, sanadine, and possibly kyanite, which is a quartz-eclogite. In comparison to the pervasive peridotitic country rock of the mantle at these depths, this quartz-bearing source rock is highly unusual. Melting of this assemblage produces the characteristic arc HAB bulk composition. A notable feature of this assemblage is the high modal content of garnet, which will strongly fractionate the concentration of REEs in the ensuing melt. That this is not observed in arc HAB has long been taken to be evidence that garnet, and hence also the slab itself, is not a reasonable source rock. But this conclusion assumes melt extraction takes place at a sufficiently high pressure to assure that garnet is still a dominant phase, which may be dynamically unlikely. That is, magma ascends because it is buoyant relative to its surroundings. And to become buoyant, or gravitationally unstable, only enough melting to make the mushy assemblage less dense than the surroundings is necessary. Further extensive melting will
take place during ascension. The requisite amount of melting necessary to extract a melt free of solids will be, as in solidification fronts, 50 vol.%. Melt extraction is also possible at lower degrees of melting due to compaction (e.g., Mckenzie, 1984), but in the slab this is probably inoperative because of the highly limited thickness of material involved, which makes compaction weak. Gravitational instability may take place after only about 10–15 vol.% melting, and the mushy assemblage may rise 20–30 km before enough additional melting takes place to enable the solids to repack and free a nearly crystal-free melt. By this point (i.e., 2–2.5 GPa), garnet, although still present, is a relatively minor phase, and the REE pattern fits that observed in the arc HAB. A critically important physical aspect of this process is the unusually narrow liquidus–solidus relation of quartz-eclogite, which may be as small as 50–70 C. This means that once the temperature reaches this point, partial melting, melt production, and instability take place rapidly with relatively little rise in temperature. The process, in effect, has a trigger point that may spontaneously generate magma and diapers once the melting point is reached. This is in striking contrast to peridotitic mantle rock where the melting
Magmatism, Magma, and Magma Chambers
325
Aleutian (Atka) High-Al Basalt Rayleigh–Taylor
Garnet
20
Magmatic instability Ga rn et
Cpx
30
Spacing
2.15
μ1 μ2
a = 0.53 h2
μ1 μ2
d=
2h2
Diapir size 15
Plag + Cpx
1/3
1/4
10 Plag Liquid
Cpx Oliv
ine
5
0
1100
1200
1300
Figure 37 The process of Rayleigh-Taylor gravitational instability when a less dense fluid (black) lies beneath a more dense fluid (clear). Plumes automatically form as the most efficient form of fluid transfer. The equations at right shown the relationship between plume spacing and plume radius and the fluid viscosities and source thickness (h2).
1400
Temperature (°C) Figure 36 Phase diagram for typical Aleutian HAB. Note the rosette phase relations near 20 kb.
range is about 500–600 C (where melting takes place gradually). The corollary of this deep-phase equilibria is the phase equilibria of arc HAB itself, which is shown for an Aleutian basalt by Figure 36 as determined by Baker and Eggler. A striking feature of this phase diagram is the phase rosette near 2 GPa defined by plagioclase, garnet, and clinopyroxene. These are phases also common to quartz-eclogite, which suggests that at this point these two rocks (i.e., HAB and quartz-eclogite) would be in equilibrium. This also suggests, alternatively, that this arc HAB may have been extracted from a mush of quartz-eclogite that rose buoyantly to this level. Similar rosette features were long looked for (to no avail) in MORB phase equilibria to identify the point of extraction from peridotitic mantle, and here in arc HAB this feature seems to have been overlooked.
6.07.10.8 Diapirism, Rayleigh–Taylor Instability, and Volcano Spacing In this process of magmatism triggered by the temperature of the slab surface reaching a critical point, a thin band or ribbon of buoyant mush is produced everywhere beneath the arc on the surface of the subducting plate. A layer of lower-density material
beneath higher-density material initiates a collective gravitational instability that leads to the formation of a family of (ideally) evenly spaced plumes or diapirs of the buoyant material (see Figure 37). The distance between the diapirs is given by d ¼
2h2 1 1=3 2:15 2
½26
where h2 is the buoyant layer thickness and 1 and 2 are, respectively, the viscosity of the overlying mantle and the mush. The radius of the diapir is given by a ¼ 0:53h2
1=4 1 2
½27
If these equations are used in concert along with conservation of mass and the observed spacing of arc volcanic centers, an estimate can be made of the nature of the source region. The source region is thin (50–100 m), highly viscous (1010 Pa s), and the diapirs have a radius of 3 km. Not only do these estimates seem reasonable, but there are dynamic features of this process that fit or correlate well with the general observed style of magmatism. The overall magmatic process is driven by the production of buoyant mush, loading the system to the point of systematic instability, which relatively rapidly unloads the system over and over. This is triggered by simply the temperature at the slab interface reaching above the solidus of the quartz-eclogite. Evenly spaced diapirs of mush rise into the convecting mantle wedge, traverse it, and essentially burn a hole upward through the overlying lithosphere. The hot
326
Magmatism, Magma, and Magma Chambers
diapirs soften the wall rock allowing it to flow around the diapir. Early diapirs undoubtedly stagnate before reaching the surface, but continued use by later diapirs eventually makes a path to the surface. This process chemically and thermally insulates the pathway, allowing later magmas to travel to the surface without suffering undue thermal and chemical contamination. Extraction of standard arc HAB takes place during ascent. Diapirs can traverse the mantle wedge and repeatedly enter exactly the same pathway point at the base of the arc lithosphere because the flow through the wedge, across any vertical section, obeys conservation of mass. Any diapir ascending at constant velocity entering this flow on one margin will exit the flow at exactly the same vertical point on the opposite margin. This will not be so for a magma generated ‘within’ the wedge region. The ribbon of melting on the slab surface will spread down-dip at the rate of subduction. Once the melt ribbon reaches a distance d down-dip, a new instability will commence. For typical subduction rates, this will occur about 3 My after the first instability, which is the time of appearance of the sporadic volcanoes of the secondary arcs, like Leskov in the Scotia arc and Amak in the Aleutians. Because of the angle of dip ( ) of the unstable layer, the spacing of these centers will be d cos( ), which is also observed.
flow induced in the wedge region above the plate. Melting is initiated by this flow, which triggers a collective gravitational instability bringing magma periodically to the surface at prescribed locations over long periods of time. The entire process is buttressed by chemical processes that reflect, but do not control, the physical processes. The detailed chemical processes, at all levels, are best understood through an appreciation of the overall physical dynamics of the full system.
6.07.11 Solidification Front Differentiation Processes 6.07.11.1
Introduction
Scattered behind many island arcs are contemporary alkali basalt volcanoes. These are very much unlike arc magmas in composition and spatial distribution. They are scattered over a vast region, are silica poor, alkali rich, commonly contain ultramafic nodules, and have an isotopic identity closely linking them to peridotitic source rock. Unlike arc HAB, everything about them suggests a mantle, not a slab, identity. These are exactly the characteristic magmas expected to be generated within the flowing mantle wedge. The stream lines of flow in the wedge are inclined upward as the flow moves into the corner region above the subducting plate. Already at the point of melting, this convective rise initiates further melting resulting in sporadic, disorganized volcanism on the surface.
Magmatic systems, regardless of setting, are all broadly similar in the basic nature of the physical and chemical processes that shape the final product. The gain and loss of crystals, both indigenous and exotic or xenocrystic, their type, size, concentration, and length of time in contact with the melt are fundamental to the chemical evolution of the magma. But it is the dynamic role of solidification fronts that sets at any time the prevailing magmatic processes that go to determine the spatial pattern of rocks seen in the field and the temporal eruptive sequence of lavas. Crystal fractionation from the interiors of magma chambers, the classical approach to chemical differentiation, is not a realistic process. We have seen that the fundamental difficulty in achieving high degrees of chemical fractionation is that the silica-rich residual melt resides deep within solidification fronts where it is inaccessible to extraction accumulation, transport, and eruption. Given the observed diversity of igneous rocks on Earth and even at some small volcanoes, there are, clearly, other processes operating within solidification fronts that produce strong, bimodal chemical fractionation effects. The tearing open of solidification fronts under gravitational instability, flushing of residual melt from the front, and the sidewall upflow of residual melt are three key differentiation processes. These are processes that operate on relatively small spatial scales (1–10 m) but lead to planetary-scale differentiation effects.
6.07.10.10
6.07.11.2
6.07.10.9 Alkali Basalts Posterior to Arcs
The Arc Magmatic System
Arc magmatism is a mechanical system controlled by subduction of a MORB-like crust that is intimately coupled, physically and thermally, to the
Solidification Front Instability
A detailed examination of almost any thick basaltic sill reveals in the upper parts the presence of interdigitating lenses of silicic material (see Figure 38).
Magmatism, Magma, and Magma Chambers
327
Solidification front instability
Peneplain sill 300 m
Chill
% SiO2
Roof
40
Silicic segregations
Tsolidus
Silicic melt Diopside
Crust
Mush
200
Suspension
Host dolerites
Solidification front
250
50
60
D e p t Silicic segregation h
Gabbro
Tliquidus
150 Magma
Silicic segregations
Olivine
Silica
Opx
100
Dolerite Dolerite
er
50
m
m
Ha
Segregation
0m
Chill
Figure 38 The process of solidification front instability (SFI) in the Peneplain Sill at Solitary Rocks. The left section shows the location and nature of the silicic segregations along with a normative ternary showing the range of segregation compositions relative to the dolerite sill compositions. The SFI process is shown schematically (upper center) and the nature of an actual segregation is also shown (lower right). Reproduced from Marsh BD (2002) On Bimodal Differentiation by Solidification Front Instability in Basaltic Magmas, I: Basic Mechanics. Geochimica et Cosmochimica Acta v66: 2211–2229, with permission from Elsevier.
The lenses often begin thin and small near the sill roof and increase in thickness and length up to, respectively, 1 m and 20–30 m at a depth of about 25% full sill thickness. These bodies have been called by many names from pegmatitic segregations to silicic segregations to silicic schlieren to granophyres (e.g., Walker, 1956). As these names suggest, silicic segregations are significantly enriched in silica over the bulk composition of the sill, commonly 60–65 wt.% SiO2 versus 55 wt.% SiO2. They are also coarser grained and their physical form often suggests that they fill tears and fractures in the sill. A likely mechanism of formation is by internal failure or tearing of the solidification front (Marsh, 2002). Roof solidification fronts are generally denser than the underlying uncrystallized magma, and as they thicken they will eventually fail under the action of gravity, producing internal gashes that fill with local residual melt. The process is similar to hanging an old rug from the ceiling and fastening
weights along the lower edge. As more and more weight is added, the rug will eventually tear horizontally at the weakest spots. Solidification fronts tear where there is enough strength to fracture, which is within the rigid crust (see Figures 4 and 5). The internal distribution of both weight and strength within the solidification front are determined by the spatial increase in crystallinity with decreasing temperature. To avoid tearing, the local strength of the solidification front must increase proportionally with weight, and strength comes from increasing crystallinity. But crystallinity changes are dictated by phase equilibria and not by strength constraints. The noncovariant variation of strength and crystallinity sets the physical conditions for solidification front instability. With initiation of tearing, local interstitial melt is drawn into the tears. Due to conservation of mass, the infilling melt must come from below the tear, and the texture of the segregation reflects the filling process.
328
Magmatism, Magma, and Magma Chambers
Prior to the tearing, from phase equilibria considerations, the interstitial melt is contained in the solidification front at a crystallinity of about 60–65 vol.%. The melt is multiply saturated with solid phases, and with tearing the local melt can now grow large (1–10 cm) pegmatitic crystals of clinopyroxene, often in symplectic intergrowths with plagioclase. The melt arriving from lower in the solidification front is slightly hotter and as it cools it forms progressively smaller crystals downward in the segregation. The texture of the segregation is, thus, coarse at the top and finer downward, reflecting the filling process. This is just the opposite of expected if the segregations were blobs of residual melt from the lower solidification front that escaped and entered the upper solidification front. Once generated, the silicic differentiates are almost impossible to remove or assimilate back into the basaltic magma. If the whole system became molten they will maintain their physical and chemical identity due to their contrast in viscosity, and being at near-chemical equilibrium they will not be assimilated by diffusion. Due to their density contrast they will, however, given the chance, tend to collect at the top of the system and undergo compaction into large silicic masses. Some large dolerite sills have large bulbous masses of such rock at the upper contacts. The inordinately large volume of these masses relative to the associated sill, the thin chilled margins of dolerite separating the mass from the country rock, and the often sharp contacts between the dolerite and the granophyre all suggest that these bodies have formed from reprocessing of basaltic bodies containing silicic segregations. This may be a process operating on many scales that contributes to building continental crust. It may today be operating at the ocean ridges in the formation of plagiogranite lenses and also in the crust of Iceland to form large bodies of rhyolite. 6.07.11.3 Silicic Segregations and Crust Reprocessing in Iceland Silicic segregations are essentially unavoidable to form within any basaltic solidification front, yet they are spatially distributed and can only be collected if they are freed from the rock through wholesale melting. Large-scale systematic melting as in rifting provides a setting where this process can operate efficiently. The ideal locality for repeated rifting of basaltic terranes is at immobile oceanic volcanic centers where the crust can be
reprocessed over and over. This promotes a strong bimodality of crustal composition. About 18% of the surface rocks of Iceland are highly silicic rhyolites. This is in striking contrast to Hawaii, where there are no rhyolitic rocks at all. Deep drilling in Iceland shows the crust laden with lenses and pockets of silicic segregations and coupled with a local ocean ridge that has periodically relocated into older crust, allowing for widespread remelting and conditions ideal for accumulating large amounts of silicic material. Detailed petrologic studies of the 200 km3 rhyolite mass at Torfajokull Caldera shows that this body was formed by this process (Gunnarsson et al., 1998). Fissures propagating into older crust bring hot basaltic magma into thin warm crust, which promote progressive wholesale melting via melting fronts propagating outward into the country rock. Because of heterogeneous melting the wall rock undergoes large-scale failure, freeing silicic segregations that collect into large masses of rhyolitic magma and sometimes erupt from the same craters as the basalts. This process is depicted by Figure 39. This overall process of collecting silicic segregations is very much akin to the emulsion coalescence that occurred in the Sudbury impact melt sheet. The net result of magmatic processes of this type is a bimodal suite of compositions characterized at each end by a basaltic and a silicic member.
6.07.11.4
Sidewall Upflow
Another means for escape of residual silicic melt from deep within solidification fronts is by upward flow along steep lateral walls. Since all magmas are contained within solidification fronts, some bodies will have tall steep lateral margins where low density residual melt can flow upward under Darcian dynamics and collect at higher levels (see Figure 40). This can only happen if the advance of the solidification front is slow enough to allow upward flow through the porous solidification front. In the upper crust solidification fronts move rapidly in response to high cooling rates, but with increasing depths in the crust cooling becomes increasing slow. This retards the inward advance of the solidus, marking the rear of the solidification front, allowing ample time for upward flow of viscous melt. Unlike the idealized flows that are depicted in Figure 2, these realistic flows occur deep within the solidification front and not along hard walls where residual fluid can essentially stream upward and collect rapidly at
Magmatism, Magma, and Magma Chambers
329
Generation of rhyolitic magma in Iceland through reprocessing of older crust Fissure swarm and explosion craters
Basaltic cone and flows
Caldera and rhyolites
Lavas
Gabbros Basalt
Silicic segregations
Silicic melts Melting front
Melting fronts
Figure 39 Reprocessing of the Icelandic crust due to propagation of basaltic fissures into older crust containing silicic segregations and the formation of large bodies of silicic material from the freeing and collection of these individual silicic segregations.
Sidewall upflow of interstitial melt 60 55 50
Melt SiO2 wt.% Vz = (KD /μ)(ΔPg + P / z)
Return flow
Viscosity
Permeability
Figure 40 The flow up a vertical wall of interstitial residual, silica-rich melt deep with a solidification front. The variations in permeability and melt viscosity are indicated as is the form of Darcy’s equation governing the flow.
critical to note here that both m and KD change strongly with distance outward in the solidification front; increases and KD decreases such that the quantity KD/ decreases strongly with approach to the solidus. This regulates the flow and there is an optimum position within the rear of the solidification front where melt is most efficient at flowing upward. The slower the solidification front moves the more silicic is the optimum melt. In the upper crust, the dominant melt has a composition of about 55 wt.% SiO2, and in the lower crust it approaches 65 wt.%. The buildup of volatiles in the residual melt helps the overall process.
6.07.11.5 the roof of the system. The flow is governed by Darcy’s equation, Vm ¼
KD qP þ g qZ
½28
where Vm is the Darcy velocity, which is a measure of the melt flux per unit area, KD is permeability, is melt viscosity, P is pressure, z is the spatial coordinate, is density contrast, and g is gravity. It is
Fissure Flushing
A broadly similar process to melt flow through the rear of a steep solidification front is the flushing of residual melt from an aging solidification front spanning a fissure or other sheet-like body. The process is depicted by Figure 41. During times of repose in magmatic systems, the driving pressure relaxes, conduit walls press in on the remaining magma, and solidification fronts move in from the lateral walls. The fronts eventually meet and form a bridge of
330
Magmatism, Magma, and Magma Chambers
Interstitial melt flushing by fresh magma
60
50
% Crystals
Melt SiO2 (wt.%)
70 95 85 75 50 5
Figure 41 The process of flushing residual melt from a partially solidified fissure by reactivation of magma flow after a repose period. The curves show the variation in residual melt silica content as a function of position and degree of crystallinity within the fissure.
crystals across the conduit. As solidification proceeds, the interstitial melt becomes progressively enriched in silica as a function of position within the conduit. When the repose period ends and magma flow resumes, magma coursing this conduit will flush the residual melt from the crystalline matrix of solids. This melt will be distinctive in its differentiated chemical composition, which will appear chemically indistinguishable to crystal fractionation, and also in the telltale clots of crystals torn from the solidification front during eviction. It is processes like this that undoubtedly give rise to mildly differentiated magmas at many locations in large, long-lived magmatic systems like Hawaii. Another characteristic of these events is their optimum near-surface location on flanks of the volcano where magma transport in fissures is common. The 1955 Kilauea eruption produced lava with these characteristics (Wright and Fiske, 1971) and the ongoing eruptions
of Puu Oo also show some of these features (e.g., Garcia et al., 1989).
6.07.12 Magmatic Systems Earth’s surface displays a rich diversity of igneous rock from which most other rocks are derived. The majority of igneous rocks are either oceanicfloor basalts or continental granitics. The processes that produce this strong bimodality are physical processes, buttressed by chemical processes, associated with a prevailing tectonic theme. Mantle convection gives rise to seafloor spreading and a distinct style of magma production and evolution in a steady-state standing magmatic mush column capped by a passive thin sill. No continental silicic material is produced. The seafloor is an enormous gabbroic batholith. Silicic noise is produced within the solidification
Magmatism, Magma, and Magma Chambers
fronts of basaltic systems on small local scales. The key to accentuating and enhancing this silicic signal is through systematic reprocessing. Remelting of the oceanic crust in subduction zones, during massive bolide impacts, in areas like Iceland, and in other immobile crustal welts, accumulates this silicic material into viable rock masses. Magmatic processes in and of themselves operate within a vast array of tectonic environments, but the processes involved in the behavior and evolution of magma are finite and understandable within a clear physical and chemical framework. The key to analyzing magmatic systems is to gain an integrated physical perspective of the overall process of magma production, transport, and emplacement or eruption. Magmas chemically evolve through the gain and loss of crystals, which is governed by entrainment of exotic crystals in addition to those nucleated and grown within the magma itself. But it is the understanding of the intimate coupling of spatial physical relations and processes with phase equilibria that furnishes the greatest insight into the true processes that shape the magma and the Earth itself.
Acknowledgments This work on magmatic processes is supported primarily by the National Science Foundation via grant OPP-0440718 to the John Hopkins University (BDM). Unless otherwise noted, all figures are from the book Magma Physics in preparation by the author.
References Ariskin AA (1999) Phase equilibria modeling in igneous petrology: use of COMAGMAT model for simulating fractionation of ferro-basaltic magmas and the genesis of high-alumina basalt. Journal of Volcanology and Geothermal Researach 90: 115–162. Baragar WRA (1960) Petrology of basaltic rocks in part of the Labrador Trough. Bulletin of the Geological Society of America 71: 1589–1644. Barksdale JD (1937) The Shonkin Sag laccolith. American Journal of Science 33: 321–359. Becker GF (1897) Fractional crystallization of rocks. American Journal of Science 4: 257–261. Bergantz GW (1989) Underplating and partial melting: Implications for melt generation and extraction. Science 245: 1093–1095. Bhattacharji S (1967) Mechanics of flow differentiation in ultramafic and mafic sills. Journal of Geology 75: 101–112. Boudreau AE (1994) Mineral segregation during crystal aging in two-crystal, two-component systems. South Africa Journal of Geology 4: 473–485.
331
Bowen NL (1915) The later stages of the evolution of the igneous rocks. Journal of Geology 23: 1–89. Bowen NL (1947) Magmas. Bulletin of the Geological Society of America 58: 263–280. Brandeis G and Jaupart C (1986) On the interaction between convection and crystallization in cooling magma chambers. Earth and Planetary Science Letters 345–361. Brandeis G and Marsh BD (1989) The convective liquidus in a solidifying magma chamber: A fluid dynamic investigation. Nature 339: 613–616. Brandeis G and Marsh BD (1990) Transient magmatic convection prolonged by solidification. Geophysical Research Letters 17(8): 1125–1128. Carmichael ISE, Turner FJ, et al. (1974) Igneous Petrology. New york: McGraw-Hill Book Company. Carr MJ, Stoiber RE, et al. (1973) Discontinuities in the deep seismic zone under the Japanese Arcs. Bulletin of the Geological Society of America 84: 2917–2930. Carslaw HS and Jaeger JC (1959) Conduction of Heat in Solids. Oxford: Clarendon Press. Cartwright J and Hansen DM (2006) Magma transport through the crust via interconnected sill complexes. Geology 34(11): 929–932. Coffin MF and Eldholm O (1992) Volcanism and continental break-up: A global compilation of large igneous provinces. In: Storey BC, Alabaster T, and Pankhurst RJ (eds.) Magmatism and the Causes of Continental Break-up, pp. 17–30. Boulder, CO: Geologial Society of America. Dawson JB (1992) First thin sections of experimentally melted igneous rocks: Sorby’s observations on magma crystallization. Journal of Geology 100: 251–257. Dicken AP, Nguyen T, and Crocket JH (1999) Isotopic evidence for a single impact melting origin of the Sudbury Igneous Complex. In: Dressler BO and Sharpton VL (eds.) Geological Society of America, Special Paper 339: Large Meteorite Impacts and Planetary Evolution II, pp. 361–371. Dobran F (2001) Volcanic Processes, Mechanisms in Material Transport. Kluwer Academic. Drever HL and Johnston R (1967) Picritic minor intrusions. In: Wyllie P (ed.) Ultramafic and Related Rocks, pp. 71–82. New York: John Wiley and Sons, Inc. Fleming TH, Elliott DH, et al. (1992) Chemical and isotopic variations in an iron-rich lava flow from the Kirkpatrick Basalt, north Victoria Land, Antarctica: implications for lowtemperature alteration. Contributions to Mineralogy and Petrology 111: 440–457. Fuijii T (1974) Crystal settling in a sill. Lithos 7: 133–137. Garcia MO, Ho RA, et al. (1989) Petrologic constraints on rift-zone processes Results from episode 1 of the Puu Oo eruption of Kilauea volcano, Hawaii. Bulletin of Volcanology 52: 81–96. Ghiorso MS, Carmichael ISE, et al. (1983) The Gibbs Free Energy of mixing of natural silicate liquids; an expanded regular solution approximation for the calculation of magmatic intensive variables. Contributions to Mineralogy and Petrology 84: 107–145. Gibb FGF and Henderson CMB (1996) The Shiant Isles Main Sill: structure and mineral fractionation trends. Mineralogical Magazine 60: 67–97. Gray NH and Crain IK (1969) Crystal settling in sills: A model for suspension settling. Canadian Journal of Earth Sciences 6: 1211–1216. Gregory RT and Taylor HPJ (1981) An oxygen isotope profile in a section of Cretaceous Oceanic Crust, Samail Ophiolite, Oman: evidence for d18O buffering of the oceans by deep (>5 km) Seawater-hydrothermal circulation at mid-ocean ridges. Journal of Geophysical Research 86(B4): 2737–2755.
332
Magmatism, Magma, and Magma Chambers
Grieve RAF, Stoffler D, et al. (1991) ‘‘The Sudbury Structure: Controversial or Misunderstood?’’ Journal of Geophysical Research 96(E5): 22,753–22,764. Gunnarsson B, Marsh BD, et al. (1998) Generation of Icelandic Rhyolites: Silicic Lavas from the Torfajokull Central Volcano. Journal of Volcanology and Geothermal Research 83: 1–45. Hein JR, Scholl DW, et al. (1978) Episodes of Aleutian ridge explosive volcanism. Science 199: 137–141. Helz RT (1986) Differentiation behavior of Kilauea Iki lava lake, Kilauea volcano, Hawaii: An overview of past and current work. Geochemical Society Special Publication 1: 241–258. Helz RT, Kirschenbaum H, et al. (1989) Diapiric transfer of melt in Kilauea Iki lava lake, Hawaii: a quick, efficient process of igneous differentiation. Geological Society of America Bulletin 101: 578–594. Hersum T, Hilpert M, and Marsh B (2005) Permeability and melt flow in simulated and natural partially molten basaltic magmas. Earth and Planetary Science Letters 237: 798–814. Hess HH (1956) The magnetic properties and differentiation of dolerite sills – discussion of dolerite sills. American Journal of Science 254: 446–451. Ho RA and Garcia MO (1988) Origin of differentiated lavas at Kilauea Volcano, Hawaii: Implications from the 1955 eruption. Bulletin of Volcanology 50: 35–46. Holden GS and Hooper PR (1987) Petrology and chemistry of a Columbia River basalt section, Rocky Canyon, west-central Idaho. Geological Society of America Bulletin 87: 215–225. Hon K, Kauahikaua J, et al. (1994) Emplacement and inflation of pahoehoe sheet flows: Observations and measurements of active lava flows on Kilauea Volcano, Hawaii. Geological Society of America Bulletin 106: 351–370. Hort M (1997) Cooling and crystallization in sheet-like magma bodies revisited. Journal of Volcanology and Geothermal Research 76: 297–317. Hort M, Marsh BD, et al. (1999) Convection and crystallization in a liquid cooled from above: An experimental and theoretical study. Journal of Petrology 40(8): 1271–1300. Hsui AT, Marsh BD, et al. (1983) On melting of the subducted oceanic crust: effects of subduction induced mantle flow. Tectonophysics 99: 207–220. Hurlbut CSJ (1939) Igneous rocks of the Highwood Mountains, Montana Part I The Laccoliths. Bulletin of the Geological Society of America 50: 1043–1112. Imsland P (1984) Petrology, Mineralogy and Evolution of the Jan Mayen Magma System. Reykjavik, Prentsmidjan Oddi. Jaupart C (1998) Gas loss from magmas through conduit walls during eruption. The Physics of Explosive Volcanic Eruptions. Geological Society of London Special Publication 145: 73–90. Jaeger JC (1968) Cooling and solidification of igneous rocks. In: Hess HH and Poldervaart A (eds.) Basalts: The Poldervaart Treatise on Rocks of Basaltic Composition, vol. 2, pp. 503–536. New York: Interscience. Jaeger JC and Joplin GA (1955) Rock magnetism and the differentiation of dolerite sill. Journal of the Geological Society of Australia 2: 1–19. Jaeger JC and Joplin GA (1956) The magnetic properties and differentiation of dolerite sills - discussion. American Journal of Science 254: 443–446. Jhaveri B and Homsy GM (1980) Randomly forced RayleighBenard convection. Journal of Fluid Mechanics 98: 329–348. Kadanoff LP (1991) Complex structures from simple systems. Physics Today 9–10. Kerr RC, Huppert HE, et al. (1989) Disequilibrium and macrosegregation during solidification of a binary melt. Nature 340: 357–367. Khurana A (1988) Rayleigh-Benard experiment probes transition from chaos to turbulence. 17–21.
Kincaid C and Sacks IS (1997) Thermal and dynamical evolution of the upper mantle in subduction zones. Journal of Geophysical Researach 102: 12295–12315. Lightfoot PC, et al. (1997) Geochemical relationships in the Sudbury Igneous Complex: Origin of the Main Mass and Offset Dikes. Economic Geology 92: 2890397. Maaloe S, et al. (1989) Population density and zoning of olivine phenocrysts in tholeiites from Kauai, Hawaii. Contributions to Mineralogy and Petrology. Maaloe S, et al. (1992) The Koloa volanic suite of Kauai, Hawaii. Journal of Petrology 33(4): 761–784. Makse HA, et al. (1997) Spontaneous stratification in granular mixtures. Nature 286: 379–382. Mangan MT and Marsh BD (1992) Solidification front fractionation in phenocryst-free sheet-like magma bodies. Journal of Geology 100: 605–620. Marsh BD (1979a) Island arc development: Some observations, experiments, and speculations. Journal of Geology 87: 687–713. Marsh BD (1979b) Island-Arc Volcanism. American Scientist 67: 161–172. Marsh BD (1982) On the mechanics of igneous diapirism, stoping, and zone melting. Amerian Journal of Science 282: 803–855. Marsh BD (1988) Crystal capture, sorting, and retention in convecting magma. Geological Society of America Bulletin 100: 1720–1737. Marsh BD (1989) On convective style and vigor in sheet-like magma chambers. Journal of Petrology 30: 479–530. Marsh BD (1991) Reply to comments On convective style and vigor in sheet-like magma chambers. Journal of Petrology 32(4): 855–860. Marsh BD (1996) Solidification fronts and magmatic evolution. Mineralogical Magazine 60: 5–40. Marsh BD (1998) On the interpretation of crystal size distributions in magmatic systems. Journal of Petrology 39: 553–599. Marsh BD (2002) On Bimodal Differentiation by Solidification Front Instability in Basaltic Magmas, I: Basic Mechanics. Geochimica et Cosmochimica Acta v66: 2211–2229. Marsh BD (2004) A Magmatic Mush Column Rosetta Stone: The McMurdo Dry Valleys of Antarctica. EOS Transactions American Geophysical Union 85(4723): 497–502. Marsh BD (2006) Dynamics of magmatic systems. Elements 2: 287–292. Marsh BD (2007) Magma Physics. (in preparation). Marsh BD, Gunnarsson B, et al. (1991) Hawaiian basalt and Icelandic rhyolite: Indicators of differentiation and partial melting. Geologische Rundschau 80(2): 481–510. McBirney AR (1993) Igneous Petrology. Boston: Jones and Bartlett Publishers. McBirney AR (1999) Santorini and Its Eruptions, by Ferdinand A. Fouque´ (translated by A. R. McBirney), 495 pp. Baltimore, MD: Johns Hopkins University Press. McKenzie D (1984) The generation and compaction of partially molten rock. Journal of Petrology 25: 713–765. Melnik O, Barmin AA, and Sparks RSJ (2005) Dynamics of magma flow inside volcanic conduits with bubble overpressure buildup and gas loss through permeable magma. Journal of Volcanology and Geothermal Research 143: 53–68. Murata KJ and Richter DH (1966) The settling of olivine in Kilauean magma as shown by lavas of the 1959 eruption. American Journal of Science 264: 194–203. Nicolas A (1995) The Mid-Oceanic Ridges Mountains Below Sea Level. Berlin Heidelberg: Springer-Verlag. Osborne FF and Roberts EJ (1931) Differentiation in the Shonkin Sag laccolith, Montana. American Journal of Science 22: 331–353.
Magmatism, Magma, and Magma Chambers Philpotts AR and Carroll M (1996) Physical properties of partly melted tholeiitic basalt. Geology 24(11): 1029–1032. Pirsson LV (1905) Petrography and Geology of the Igneous Rocks of the Highwood Mountains, Montana. United States Geological Survey Bulletin no. 237. Ryan MP (1994) Neutral-buoyancy controlled magma transport and storage in mid-ocean ridge magma reservoirs and their sheeted-dike complex: A summary of basic relationships. In: Ryan MP (ed.) Magmatic Systems, pp. 97–138. San Diego, California: Academic Press, Inc. Savage SB and Lun CKK (1988) Particle size segregation in inclined chute flow of dry cohesionless granular solids. Journal of Fluid Mechanics 189: 311–335. Scheidegger KF, Corliss JB, et al. (1980) Compositions of deepsea ash layers derived from North Pacific Volcanic arcs: Variations in time and space. Journal of Volcanology and Geothermal Research 7: 107–137. Simkin T (1967) Flow differentiation in the prictic sills of North Skye. In: Wyllie PJ (ed.) Ultramafic and Related Rocks, pp. 64–69. New York: John Wiley and Sons. Sinton JM and Detrick RS (1992) Mid-ocean ridge magma chambers. Journal of Geophysical Research 97(B1): 197–216. Smith AL and Roobol MJ (1982) Andesitic pyroclastic flows. In: Thorpe RS (ed.) Andesites Orogenic Andesites Related Rocks, pp. 415–436. New York: John Wiley and Sons. Smith RL and Bailey RA (1968) Resurgent Cauldrons. In: Coats RR, Hay RL, and Anderson CA (eds.) Studies in Volcanology A Memoir in Honor of Howell Williams pp. 613–662. Geological Society of America. Stoiber RE and Carr MJ (1971) Lithospheric plates, Benioff zones, and volcanoes. Geological Society of America Bulletin 82: 515–522. Sugimura A (1968) Spatial relations of basaltic magmas in island arcs. In: Hess HH and Poldervart A (eds.) Basalts The Poldervaart Treatise on Rocks of Basaltic Composition, vol. 2, pp. 537–572. New York: Interscience Publishers. Toksoz MN and Bird P (1977) Modeling of temperatures in continental convergence zones. Tectonophysics 41: 181–193. Turcotte DL and Schubert G (1982) Geodynamics. New York: John Wiley and Sons.
333
Turner FJ and Verhoogen J (1960) Igneous and Metamorphic Petrology. McGraw-Hill. Turner JS (1973) Buoyancy Effects in Fluids. New York: Cambridge University Press. Upton BGJ and Wadsworth WJ (1967) A complex basaltmugearite sill in Piton des Neiges volcano, Reunion. American Mineralogist 52: 1475–1492. Viskanta R and Gau C (1982) Inward solidification of a superheated liquid in a cooled horizontal tube. WarmeStoffubertrgung 17: 39–46. Wager LR and Brown GM (1968) Layered igneous rocks. San Francisco: W.H. Freeman. Wager LR and Deer WA (1939) Geological investigations in East Greenland, Part III: The petrology of the Skaergaard Intrusion. Kangerdlagssuaq, East Greenland, Medd Eroenl 105(4): 1–346. Walker F (1956) The magnetic properties and differentiation of dolerite sills – Critical discussion. American Journal of Science 254: 433–443. Wilch TI, Lux DR, et al. (1993) Minimal pliocene-pleistocene uplift of the dry valleys sector of the Transantarctic Mountains: A key parameter in ice-sheet reconstructions. Geology 21: 841–844. Williams H (1941) Calderas and their origin. Bulletin of the Department of Geologial Sciences, University of California 25(6): 239–346. Wright TL (1971) Chemistry of Kilauea and Mauna Loa lava in space and time. Geological Survey Professional Paper 735: 40. Wright TL and Fiske RS (1971) Origin of the differentiated and hybrid lavas of Kilauea volcano, Hawaii. Journal of Petrology 12(1): 1–65. Wright TL and Okamura RT (1977) Cooling and crystallization of tholeiitic basalt, 1965 Makaopuhi lava lake, Hawaii. U.S. Geological Survey Professional Paper no. 1004. Zieg MJ and Marsh BD (2002) Crystal Size Distributions and Scaling Laws in the Quantification of Igneous Textures. Journal of Petrology 43(1): 85–101. Zieg MJ and Marsh BD (2005) The Sudbury Igneous Complex: Viscous Emulsion Differentiation of a Superheated Impact Melt Sheet. Bulletin of Geological Society of America 117: 1427–1450.
6.08 Dynamic Processes in Extensional and Compressional Settings: The Dynamics of Continental Breakup and Extension W. R. Buck, Columbia University, Palisades, NY, USA ª 2007 Elsevier B.V. All rights reserved.
6.08.1 6.08.2 6.08.2.1 6.08.2.2 6.08.2.2.1 6.08.2.2.2 6.08.2.2.3 6.08.2.2.4 6.08.2.3 6.08.2.3.1 6.08.2.3.2 6.08.2.3.3 6.08.2.3.4 6.08.2.3.5 6.08.3 6.08.3.1 6.08.3.2 6.08.3.3 6.08.3.4 6.08.3.5 6.08.4 6.08.5 6.08.5.1 6.08.5.2 6.08.5.3 6.08.6 6.08.6.1 6.08.6.2 6.08.6.3 6.08.6.4 6.08.6.5 6.08.7 References
Introduction Processes Affecting the Dynamics of Continental Extension Tectonic Force for Extension Localizing Processes Thermal advection due to stretching Magmatic intrusion Magmatic heat input Cohesion loss Delocalizing Processes Thermal diffusion Viscous flow Local (crustal) isostasy Regional isostasy Additional effects High-Angle versus Low-Angle Normal Faults Rift Shoulder Uplift Low-Angle Fault Development and Stress Rotation Fault Rotation Large Offset of Normal Faults 2-D Models of Fault Formation and Offset Pure versus Simple Shear Rifting Wide versus Narrow Rifts Slow Rifting and Thermal Diffusion Viscous Stresses Local Isostatic Crustal Thinning Dikes versus Stretching to Initiate Rifting Force Available for Driving Rifting Force Needed for Tectonic Rifting Force Needed for Magmatic Rifting The Meaning of Rift Straightness The Distance of Dike/Rift Propagation Conclusions and Future Work
6.08.1 Introduction The earliest ideas about continental drift, which eventually lead to plate tectonics, were based on the observation that the eastern coasts of North and South America matched the shape of the western coasts of Europe and Africa (Wegener, 1929). This implies that the continents somehow break apart. We now understand that plate tectonics involves the creation of new plates at spreading centers and
335 339 339 341 341 341 342 342 343 343 343 344 345 345 345 346 348 349 349 350 351 359 359 360 361 362 365 366 366 370 371 371 372
the return of those plates to the Earth’s interior at subduction zones. No terrestrial spreading centers have been active for more than 200 My and spreading centers are constantly being subducted, as the Chile Rise is now being lost beneath South America. Without the development of new spreading centers, plate tectonics might cease. New spreading centers primarily develop either by the splitting of intact plates or when regions of plate convergence begin to extend. 335
336
The Dynamics of Continental Breakup and Extension
Both plate splitting and extension of convergent regions occur most often in continental, and not oceanic, lithosphere. Rifts are places where the breakup process either did not or has not yet progressed to seafloor spreading. Thus, they give us a snapshot of the early phase of continental extension. Passive margins (sometimes called rifted margins) are places where extended and/or intruded continent grades into the ocean basin formed by postrift spreading, and they give us some constaints on the state of the crust and lithosphere when breakup ended. Over the 30-plus years since the acceptance of plate tectonics, much effort has been made to characterize rifts and rifted margins and understand the processes affecting them. The pattern of faults, magmatic intrusive and extrusive constructs, uplifted rift flanks, and sediment-filled basins are used to reconstruct what happened during the extension of continents. One of the clearest messages from such observations is that continental rifts form with a variety of geometries, faulting patterns, and subsidence histories. For example, some rifts are narrow, like the Red Sea, (e.g., Cochran, 1983a; see Figure 1), and some are wide, like the Basin and Range Province (e.g., Stewart, 1978; see Figure 2). Some areas of apparently narrow rifting, such as metamorphic core
Figure 1 Shaded topographic relief image of the Red Sea viewed from the south with a vertical exaggeration of 20. The green to off-white transition marks the boundary between above and below sealevel. The highlands of the Ethiopian Plateau and Yemen Highlands are in the foreground along with the low-elevation Afar Triangle. The Gulf of Suez is just visible in the distance (made with GeoMapApp).
complexes, do not subside locally (e.g., Coney and Harms, 1984; Davis and Lister, 1988; see Figure 3), while some rifts, like those in East Africa, form deep basins even with modest amounts of extension (e.g., Rosendahl, 1987; Ebinger et al., 1989; see Figure 4). It 44° N
42° N
40° N
38° N
36° N
34° N
32° N
124° W 122° W 120° W 118° W 116° W 114° W 112° W 110° W Figure 2 Shaded topographic relief image of Western North America shown in plain view. The Basin and Range Province is bounded by the Sierra Nevada on the west and the Colorado Plateau on the east (made with GeoMapApp).
The Dynamics of Continental Breakup and Extension
has become accepted that the condition of the lithosphere at the time of rifting, its thermal structure, and crustal thickness, can have a profound effect on the tectonic development of a rift (e.g., Sonder et al., 1987; Braun and Beaumont, 1989; Dunbar and Sawyer, 1989; Buck, 1991; Bassi, 1991). Several decades ago, it was recognized that a few rifts and rifted margins were affected by copius magmatism, such as large parts of the East African Rift or the East Greenland Margin while many others appeared to have been little affected by magmatism, such as the Northern Red Sea or the margin of the US East Coast (Sengor and Burke, 1978). Simple kinematic models of lithospheric ‘stretching’ (e.g., McKenzie, 1978) reproduce the gross pattern of subsidence across rifts and margins. Thus, most models of rifting,
Figure 3 Shaded topographic relief image of the D9 Entrecasteau Islands of the Woodlark Basin. These islands are thought to be the youngest continental metamorphic core complexes on Earth (made with GeoMapApp).
33° E
RUNGWE VOLCANICS
35° E
21
A
30
Kiw
ira
337
A′
R,
Son
gw eR
1000 m 0
1 25
15
D
10° S –75 h hu
A
Malawi Rift morphology
uR
0
50 km
915
152
5
Ru
–75 65
R.
2
uk
R S.
B
–1
uru
B′ 1000 m 0
B 3 B′
20 1525
12° S
4 C
110
C a
200 290 0 38
aR gw an Dw
.
C′ 1000 m 0
5
C′
R.
6
915
474
915
Bu
Figure 4 Topographic relief and profiles across the Malawi Rift of the East African Rift System (from Ebinger et al. (1987)). Note the change in polarity of the half-graben comprising the rift.
338
The Dynamics of Continental Breakup and Extension
central themes of this chapter. These broad questions necessitate that we investigate the possible meaning of smaller-scale structures such as faults and the mechanics of magmatic intrusion, as well as a plethora of observations concerning rift structure and rifted margin history. Given the great variety of rifts and the wealth of data relating to their formation, it is not surprising that controversies have arisen. In this short chapter only a few of these controversies and only a small fraction of the observations relating to them can be discussed. Among the questions that will be touched on are: (1) Do some extensional (i.e., normal) faults slip with low dip angles (i.e., less than 30 )? (2) Are wide rifts formed by slow extension or due to the extension of hot, weak lithosphere? (3) Can nonorogenic areas rift without massive magmatic input? Before discussing these controversies an overview of some of the processes that may affect rifting is given.
including complex numerical ones, treat tectonic stretching and neglect magmatic processes. However, as better data has been collected in rifts (e.g., Maguire et al., 2006) and across rifted margins (e.g., Hinz, 1981; Holbrook et al., 2001) the importance of magmatism has become increasingly clear (Figure 5). A number of margins that were formerly thought to be nonvolcanic, like much of the South American–African Margin, were affected by huge volumes of pre- and synrift volcanism and magmatic intrusion. Seismic observations, and in some cases drilling have been used to identify these volcanic packages since they are buried by thick layers of sediment. Also, many nonvolcanic rifts and margins are along-strike continuations of volcanic rifts. This raises the question of whether some of the nonvolcanic sections were affected by magmatism early in their development. The issues of the initial state of the lithosphere and the influence of magmatism on rift development are
(a) Nonvolcanic rifted margin Georges Bank Basin Drift
0
G–1 G–2
Rift
Depth (km)
Cenozoic
Salt
10
K Ju Tr
Oceanic crust
Post-rift unconformity Continental crust
20 Mantle 30 0
100
200
300 km
(b) Volcanic rifted margin Extensional volcanic passive margin 50–80 km
No crustal extension coeval with plume activity
Eventual external highs Eventual pre-plume sedimentary basin Pre-breakup traps Onshore geology t
SD
Oceanic SDR
Re
xt
?
Thick oceanic crust
ne
Continental crust
zo city
Moho
12–30 km
SDRin
Sills
Post-breakup sediments
o
vel igh-
H
Moho Figure 5 Schematic cross-sections of representative (a) nonvolcanic (from Manspeiserand Cousminier, 1988) and (b) volcanic margins (from Geoffroy, 2005). SDRint and SDRext refer to internal and external seaward dipping reflectors. The high-velocity zone may represent mafic intrusives.
The Dynamics of Continental Breakup and Extension
6.08.2 Processes Affecting the Dynamics of Continental Extension A simple way to study the mechanics of rifting builds on kinematic models. The initial pattern of lithospheric extension is specified and the changes in the force needed to continue extension are estimated. If the force decreases, then extension should stay localized where it initiated. If the force increases, then it should be easier for the locus of extension to migrate laterally, resulting in a wide rift. The advantage of this approach is that many parameter combinations can be tested in a short time. The disadvantage is that underlying assumptions and approximations may not always be valid. Two properties of the lithosphere control how it deforms in response to applied tectonic forces and to the addition of material such as magma. One is the intrinsic strength or yield stress of the lithospheric material. The yield stress is the applied stress required to produce strain at a given rate. The other property is the density distribution in the lithosphere. The density distribution, including topographic relief, controls the gravitational stresses that must be overcome to deform the lithosphere. A number of processes can affect the yield strength and the density distribution of the lithosphere. Models of rifting describe one or more of these processes and how they interact to produce the kinds of rifts that are observed. It is convenient to think of these processes in terms of ones that promote or impede localized deformation of the lithosphere. To do this requires definition of a reference lithospheric state so that the effect of processes can be looked at in light of how much they change the extensional tectonic force needed to continue rifting. The reference lithosphere is taken to be laterally uniform in composition (i.e., crustal thickness) and in thermal structure (so isotherms are horizontal). Since the first infinitesimal increment of deformation does no work against gravity the reference tectonic force for extension involves only the intrinsic strength of the lithosphere.
6.08.2.1
Tectonic Force for Extension
Separation of lithospheric plates requires extensional stresses. At any depth those stresses can cause yielding by fault slip, ductile flow, or magmatic dike intrusion, whichever takes the least stress. Here we assume that the reference state does not involve magmatism and so defer discussion of magma intrusion. The difference
339
between the maximum and minimum principal stresses required to cause rock to deform at a given depth, the yield stress, is taken often to be the lesser of two stresses: the stress to produce brittle failure or the stress to cause ductile flow at a specified strain rate. The extensional state of stress is approximated using the usual assumption that the vertical or z-direction is the largest principal stress and equals the lithostatic stress (Anderson, 1951) given by 1 ðzÞ ¼ g
Z
z
r ðz9Þ dz9
½1
0
where g is the acceleration of gravity and r is the density of rock in the lithosphere. In the crust the assumed density is 2800 kg m 3; and in the mantle, density is 3300 kg m3. At low temperatures and moderate confining pressure, rocks can break on faults. Continued application of stress can cause slip on those faults. If one component of the horizontal stress is the minimum stress (3) then dip-slip faults should form with a normal sense of slip. Following Brace and Kohlstedt (1980), the stress difference (1 3) needed for normal faulting is estimated under the assumption that cohesionless fractures exist in all directions to accommodate fault slip. Then the minimum stress difference for faulting, the yield stress, is f ðzÞ ¼ Bð1 ðzÞ – Pp Þ
½2
where B ¼ 2f/[(1 þ f 2)1/2 þ f ], where f is the coefficient of friction. Pp is the pore pressure. Assuming f ¼ 0.85, which is the average friction coefficient for a wide range of rocks (Byerlee, 1978), makes the constant B ¼ 0.8. If the pore pressure in the rock, Pp, is taken to be hydrostatic then the faulting or brittle yield stress should increase linearly with depth in a constant density lithosphere: f ðzÞ ¼ Cz
½3
For crust with a constant density of 2800 kg m3 the gradient of increase of the faulting yield stress with depth, C, is 14.1 MPa km1. At high temperature, rocks can flow in response to stress differences without forming macroscopic fractures. For such ductile flow the stress difference for _ are found to be ductile flow, d, and strain rate, ", related through a flow law: _ 1=n expðE=nRT Þ d ¼ ð"=AÞ
½4
The Dynamics of Continental Breakup and Extension
0
(a)
–20 Depth (km)
Yield stress
–40 –60 –80
Crustal thickness = 30 km Heat flow = 40 mW m–2
–100 0
200 400 600 800 Stress difference (MPa)
1000
(b) 30
Tectonic force
where T is absolute temperature, R is the universal gas constant, E is activation energy (e.g., Goetze and Evans, 1979), and A is a constant for given material. The ductile yield stress depends on the composition of rock, as well as temperature. Dry anorthite rheology is assumed for the crust and a dry olivine rheology for the mantle. For anorthite, E ¼ 238 kJ mol1, A ¼ 5.6 1023 Pan s1, and n ¼ 3.2; for olivine, E ¼ 500 kJ mol1, A ¼ 1.0 1015 Pan s1, and n ¼ 3 (Kirby and Kronenberg, 1987). To estimate the stress difference for extension (the yield stress) as a function of depth, z, we must specify the temperature profile through the lithosphere. This is done by assuming temperatures are in steady state with a constant heat flow from below and radioactive heat production within the crust. For the illustrative examples shown here the thermal conductivity is set to 2.5 W m1 C1 for the crust and 3.0 W m1 C1 for the mantle. The crustal heat production is set to 3.3 107 W m3, which contributes 10 mW m2 to the surface heat flow for a 30 km thick crust. The mantle heat flow is adjusted to provide a given surface heat flow for a specific crustal thickness. Figure 6(a) shows yield stress profiles for a moderate heat flow temperature profile, assuming a 30 km thick crust. The horizontal force per unit length required to cause tectonic extensional yielding of the entire model lithosphere, FT, is estimated by integrating yield stress over depth (Figure 6(b)). This force depends strongly on the temperature profile and, thus, on the surface heat flow. To extend continental lithosphere with a heat flow of about 40 mW m2, as is seen adjacent to some rifts like the Red Sea (Martinez and Cochran, 1988), may require as much as 30 Tera Nt m1 of tectonic force if no magma were intruded. A number of factors can affect where extensional strain is concentrated in a continental region. Extension may, and probably does, nucleate in a site of a previous weakness. A prime question for us is whether extensional strain remains concentrated in that initial site or migrates to other places. Here we will not consider propagation of rifting orthogonal to the direction of extension, but we will concentrate on simplified versions of two-dimensional (2-D) models concerned with the vertical plane parallel to the direction of extension. We discuss eight processes that may affect concentration or delocalization of strain during continental rifting. Though this may seem like a bewilderingly large number of things that could affect extension, it should become clear that some of these processes can be insignificant. In an effort to show what may control
(TeraNt m–1)
340
Crustal thickness = 30 km
25 20 15
Force to stretch 10 5 0 30
40
50 60 70 80 Heat flow (mW m–2)
90
100
Figure 6 (a) Yield stress profile for diabase rheology crust over olivine rheology mantle for a typical continental thermal profile with an average value of surface heat flow. (b) Tectonic force for extension (the integral of yield stress with depth) as a function of surface heat flow for 30 km thick crust. The horizontal line gives the approximate maximum tectonic force available to drive rifting.
the importance of each process we derive scaling relations that show what parameters should control the effect of the process. The derivations attempt to show the approximate change in horizontal tectonic force needed for continued rifting due to a particular process. Each process is viewed in terms of how it affects the lithospheric strength or the gravitational force to continue extension. For example, a small amount of lithospheric thinning will reduce the strength of the lithosphere in proportion to the initial strength of the lithosphere times the square of the fractional amount of thinning. Local lithospheric thinning will produce gravitational stresses that also make continued extension easier, but the magnitude of the change in gravitational stresses is typically less than a few percent of the weakening effect. Familiarity with
The Dynamics of Continental Breakup and Extension
these scaling relations may make it easier to interpret complex numerical models of continental extension in which many parameters may affect the results.
Localizing processes (a) Necking and thermal advection
εA
Total strain ε
Yield stress Brittle lithosphere
6.08.2.2
Localizing Processes
A
B
Isotherm marking the brittle–viscous transition
The most obvious thing that may keep extension focused in one area is the localized thinning of lithosphere as it extends. Extension implies that material points horizontally offset from each other move apart. Conservation of mass requires vertical motion in response to this lateral extension. This implies some upward movement of the material within the lithosphere and so advection of heat. If this heat advection is significantly faster than thermal diffusion, then isotherms at the base of the lithosphere should move up. The lithosphere is thinned and so it is weakened. This process of ‘necking’ is self-reinforcing. The weaker the thinned or necked area the more concentrated, and perhaps more rapid, the extension. The more concentrated the extensional strain the faster the rift becomes weaker. To quantify the necking-related reduction in brittle lithospheric yield strength, FN, we ignore thermal diffusion and relate lithospheric thinning to extensional strain, ". The base of the brittle lithosphere is the place where the ductile and brittle yield strengths are equal. The ductile stress depends strongly on temperature, so that when heat is advected upward the depth of the base of the lithosphere moves up. The contribution to the yield force (the integral of the yield strength envelope) due to the ductile part of the lithosphere is a small fraction of the total. Thus, in the interest of simplicity all the strength is taken to be in the brittle part of the lithosphere. Assume that the base of the brittle lithosphere, where z ¼ Hb, is marked by an isotherm. The new brittle lithospheric thickness is (1 ")Hb. Since the brittle lithospheric strength is proportional to the square of the brittle layer thickness the change in strength is FN ¼ CHb2 – Cð1 – "Þ2 Hb2 2C"Hb2
Hb
A
6.08.2.2.1 Thermal advection due to stretching
½5
Figure 7(a) illustrates the reduction in strength due to advective lithospheric thinning. In doing this we ignore the effect of strain rate on yield stress, which is treated in the section below on viscous stresses. The key result is that the first-order reduction in strength depends on two times the strain times the initial strength of the lithosphere.
341
B z
ΔFN = 2C Hb εA 2
(b) Magmatic intrusion stress Brittle lithosphere
A Dike
B
Hb
Hc
Decrease in yield strength at constant strain rate Yield stress Brittle yield stress = Cz B
Isotherm marking the brittle–viscous transition
Difference between stress for faulting and diking
z ΔFM = C Hb2– g (ρm – ρf ) (Hb – Hc)2
(c) Magmatic intrusion heating Brittle lithosphere
Yield stress Hb
A
A
B
Isotherm marking the brittle–viscous transition 2 C Hb Δxd(T*m – T(Hb)) T(H b)
(d) Fault weakening – cohesion loss Brittle lithosphere
A
B
Decrease in yield strength at constant strain rate S Yield stress Brittle yield stress = Cz BB AA
Hb
Isotherm marking the brittle–viscous transition ΔFC = Hb S
Brittle yield stress = Cz B
z ΔFI =
Brittle yield stress = Cz
z
Loss of cohesive strength
Figure 7 Schematic illustrations of processes that may lead to localization of strain during continental lithospheric extension. Plots to the right show the approximate distribution of yield stress with depth both at the centre of a rift (a) and at an area unaffected by the rifting (b). The difference in the two curves is marked with vertical hatchers and that area is proportional to the change, here reduction, in the tectonic force needed for continued rifting. The scaling of these force changes is given within ovals. See text for further explanation.
6.08.2.2.2
Magmatic intrusion Dikes are magma intrusions with a thickness much smaller than their width or length. Molten basalt is assumed to be the material filling rift-related dikes since mantle melting can produce basaltic magma and because more felsic dikes might be too high in viscosity to easily propagate. Dikes should form in planes perpendicular to the least principal stress, 3;
342
The Dynamics of Continental Breakup and Extension
for a rift, this should be in vertical planes parallel to the rift (Anderson, 1951). It is assumed that preexisting vertical fractures are prevalent in order to avoid the complications of fracture mechanisms (e.g., Rubin and Pollard, 1987). However, the extra stress needed to break open dikes in unbroken rock should be limited by the rock tensile strength, which would make a small contribution to the tectonic forces estimated here. Neglected also are the viscous stresses associated with the flow of magma in a dike, since the goal is to estimate the minimum stress difference (defined as 1 3, where 1 is the maximum principal stress) required to have magma stop and freeze at a given depth in a dike. Before freezing, magma in a dike can cease moving up or down when the static pressure in the dike equals the horizontal stress at the dike wall (Lister and Kerr, 1991). The vertical pressure variation in a static column of fluid magma is related to its density, f, so for magma emplacement: q3/qz ¼ fg. To specify the level of fluid magma pressure, it is assumed that dikes always cut to the surface, where the pressure is zero. In that case the stress difference required for dike emplacement is M ðzÞ ¼ 1 ðzÞ – gf z
½6
Clearly, with these simplifications, the stress difference for magma to allow extensional separation between blocks of lithosphere depends only on the density difference between the lithosphere and fluid magma. If the crustal rock density and fluid magma density are taken to be equal, the stress difference for crustal diking is zero. In mantle of density m ¼ 3300 kg m3 the stress difference required for intrusion of fliud magma with density f ¼ 2700 kg m3 increases at a rate of 6 MPa km1 of depth into the mantle. If the ductile mantle is too weak to maintain such stresses, then the magma cannot be emplaced at depth and will be extruded. Then the force to emplace magma is just related to the difference in brittle layer thickness and the crustal layer thickness: FM ¼ gðm – f Þ
ðHb – Hc Þ2 2
½7
The decrease in force needed for rifting if there is enough magma to reach the surface is then: FM ¼ FT – FM ¼ ½CHb2 – gðm – f ÞðHb – Hc Þ2 =2 ½8
6.08.2.2.3
Magmatic heat input Basaltic magma is hotter than lithosphere so that dikes intrusion advects heat into the lithosphere and
can weaken it. Magma also releases a considerable quantity of latent heat when it crystallizes. We can define an effective temperature of magma as Tm which is equal to the actual magma temperature plus the latent heat times the specific heat. A reasonable value of the added effective temperature due to latent heat is 300 C so that liquid magma with an actual temperature of 1300 C would have an effective temperature of 1600 C. Since the temperature at the base of the brittle lithosphere is likely to be 600–800 C, the intrusion of magma could heat the rock sufficiently to cause it to flow ductilly. To estimate the reduction in lithospheric thickness on intrusion of a set of dikes of total width xd requires that we specify the width of region of lithosphere that is heated. Taking a conservative estimate that this width of region of intrusion is equal to the lithospheric thickness results in a thinning of the lithosphere of approximately xd(Tm T(Hb))/ T(Hb) given a linear geotherm in the lithosphere. Thus, the reduction in lithospheric strength due to dike intrusions totaling xd in width is FI ¼ 2CHb xd ðT m – T ðHb ÞÞ=T ðHb Þ
½9
The lithospheric strength here is defined as the strength with no further magmatism. Thus, dike intrusions totaling 10 km wide could very significantly reduce the lithospheric thickness and strength and allow rifting to proceed at moderate levels of forces even if no more magma is supplied.
6.08.2.2.4
Cohesion loss Within a rift, extension is not smoothly distributed, at least in the near surface, but is concentrated on normal faults. For faults to accommodate so much strain they must be weaker than the surrounding less-deformed rock. One possibility is that faults have lost some or all of the cohesion of the surrounding rocks. The failure of brittle materials, such as cold rock, has been described by a number of criteria. We assume the Coulomb–Navier criterion which states that failure occurs when the shear stress exceeds cohesion, S, plus a friction coefficient, , multiplied by the normal stress. In deriving the brittle yield stress above we followed Brace and Kohlstedt (1980) and others, who adopt this criterion with S ¼ 0. When the yield criterion is reached the rock may break as cohesion is lost. The optimum orientation of faults is controlled by the friction coefficient. Then for cohesion loss, the yield stress on optimally oriented faults is reduced by an amount equal to
The Dynamics of Continental Breakup and Extension
2S/[(1 þ 2)1/2/2 þ ]. For close to 0.75 the reduction in yield stress is approximately equal to S. Laboratory measured values of rock cohesion, also known as inherent shear strength, range from almost zero for weak sediments to nearly 50 MPa for some igneous rocks (Handin, 1966). The existence of large vertical cliffs requires cohesion in large volumes of rock that are of order 10 MPa. Assuming that strain weakening results in noncohesive faults then they would be weaker by approximately S. Multiplying this cohesion drop by the thickness of the brittle layer gives the amount of reduction in the yield strength of a fault compared to unfaulted rock. Thus, the reduction in force due to cohesion loss is approximately: FC ¼ SHb
½10
For a brittle layer thickness that is over 10 km the reduction in strength due to cohesion loss should be fairly small compared to the strength that remains due to rock friction. It is possible that faults develop a lower frictional coefficient or become weaker due to pore pressure increases, though these are difficult to quantify.
6.08.2.3
Delocalizing Processes
6.08.2.3.1
Thermal diffusion Thermal diffusion can lead to thickening and so strengthening of the brittle lithosphere when the lithosphere is out of thermal equilibrium. Lithosphere may be in equilibrium with basal heat flux, which is presumably delivered by mantle convection. Thermal disequilibrium may be produced by several types of extension-related heat advection, such as lithosphere stretching or dike intrusions. Diffusion tends to return isotherms perturbed by advection to their previous configuration. It is difficult to derive simple analytic expressions for the rate of lithospheric thickening due to thermal diffusion except for the simplest of boundary and initial conditions. The problem of cooling of a halfspace with no heat sources gives a simple result that captures important features of more complex model configurations. For half-space cooling the depth to an isotherm marking the base pffiffiffiffiffi of the lithosphere increases proportional to t where is thermal diffusivity (see Turcotte and Schubert, 2002). Thus, the velocity of lithospheric thickening is proportional to /Hb, where Hb is the thickness of the lithosphere. In a time interval t the lithosphere is thickened by
343
roughly t/Hb. The yield stress at the base of the lithosphere is CHb and the lithospheric strengthening due to thermal diffusion is that stress times the increase in thickness: Fd C "="_
½11
where the time interval t is replaced by the ration of the strain, ", and the strain rate of extension, "._ For a typical rock diffusivity of 106 m2 s1 a million years of cooling would produce an increase in strength of about 5 1011 Nt m1, which is a small fraction of the strength of lithosphere a few tens of kilometers thick. 6.08.2.3.2
Viscous flow By definition the strength of viscous material is proportional to strain rate, or a strain rate raised to a power. This means that when viscous material strains faster its flow stress is larger. This results in stress being delocalized in a viscous layer. A thought experiment can help illustrate this effect. Assume that the strain rate is greater in one region, as shown in Figure 8. Since the viscous flow stresses are greater in the region of high strain rate, the transition between brittle and viscous behavior (sometimes called the brittle–ductile transition) is deeper there. Clearly, where the transition depth is deeper the total yield strength is greater. Thus, the area straining the fastest will be the strongest. The amount of deepening of this transition and the lithospheric stregthening depends on two things: the contrast in strain rate and the length scale for changes in viscosity. At constant stress the vertical distance Ze over which the viscosity changes by a factor e (2.72) is related to the temperature gradient in the region of the transition, dT/dz, and to the temperature there (T0) as Ze ¼ R T02/(E dT/dz). To get an expression for the related force increase we assume constant stress at the transition depth and then estimate the vertical distance change in that depth for a given contrast in strain rate. Also, we assume a linear temperature gradient through the lithosphere so that dT/dz ¼ T0/Hb. Doing this gives a force increase of FV ¼
CHb2 RT02 "_ A ln EðT0 – TS Þ "_ B
½12
where Ts is the surface temperature. "_ B and "_ A are the background and local strain rates, respectively. It is the viscous stress effect that can lead to both folding and boudinage structures in layered rocks
344
The Dynamics of Continental Breakup and Extension
Delocalizing proceses (a) Thermal diffusion lithospheric thickening Yield stress Brittle yield stress = Cz
Brittle lithosphere
Hb
A Isotherm marking the brittle–viscous transition
A
B
B Increase in yield strength
z
ΔFT = C κ·ε 2ε
(b) Viscous stresses
ε·A
Strain · rate ε
ε·B
Yield stress
Brittle yield stress = Cz B
Brittle lithosphere
Hb B
A Brittle–viscous transition depth
A 2
2
· ·)
C HbRT0 ε ln A E(T0 – TS) ε
)
ΔFV =
Increase in yield strength at constant temperature
z
B
(c) Local isostasy
εA
Total ε strain
Pressure Crust
Hc
A
A
B
B
Moho
z
ρ
ΔFL = ρ c(ρm − ρc)gHc εA m 2
(d) Regional isostasy α
Yield stress
w Brittle lithosphere
A
B
Isotherm marking the brittle–viscous transition ΔFR = ρgαw
Hb
B
z
Brittle yield stress = Cz A
Increase in yield strength due to changes in stress field
Figure 8 Same as Figure 7 except that the processes could promote delocalization of strain during extension.
(Ramberg, 1955; Biot, 1961; Smith, 1977). It is this viscous delocalization that has been suggested to contribute to the boudinage-like structure of the wide Basin and Range Province of the western United States (Fletcher and Hallet, 1983), as discussed below. 6.08.2.3.3
Local (crustal) isostasy Local isostasy is an idealized description of how lithosphere floats on underlying fluid asthenosphere
(Watts, 2001). The term ‘local’ implies that the surface elevation at a point depends only on the average density of the column of lithosphere below that point. Essentially, the shear stress on vertical planes is taken to be zero. However, horizontal stresses should be continuous across vertical planes. Vertical stresses at a given depth should depend on the topography and average density of material in a column. Thus, where there is topographic relief in local equilibrium there will be nonzero stress differences (1 3). For example, an elevated area will be in relative tension. Conversely, as low area will be in relative compression as material tries to flow into it. Localized extension results in crustal thinning. Because crust is less dense than underlying mantle, local crustal thinning should produce lowered elevations in the center of a rift (e.g., McKenzie, 1978). This puts the center of the rift into relative compression, and this makes continued extension harder. To estimate the magnitude of this effect we follow previous workers (e.g., Artemjev and Artyushkov, 1971; Fleitout and Froidevaux, 1983) and assume that the wavelength of crustal thickness variations is large compared to crustal thickness. Then, the increase in the tectonic force for extension due to crustal thinning, Fc, equals the integral over depth of the difference in lithostatic pressure, P. This pressure difference equals wcg, where w is the change in surface elevation between the center of the rift and the adjacent, unrifted area. Here 0 is the density of crust and g is the acceleration of gravity. Topographic relief w in local isostatic equilibrium has an amplitude of "aHc(m 0)/m where m is the density of the mantle and "aHc is the amount of crustal thinning at the center of a rift. As long as the strains are not too large then the local crustal buoyancy force, FL approximately equals PH0, can be expressed as FL ¼ c g"a Hc2 ðm – 0 Þ=m
½13
Just as thermal diffusion may act to diminish the effect of thermal advection during lithospheric necking, lower crustal flow can act to even out crustal thickness variations and so reduce the crustal buoyancy effect. The idea that lateral crustal flow may be important in some areas is discussed in Block and Royden (1990), Buck (1991), Bird (1991). Isostatic topography in a rift can be caused by temperature differences. However, the thermal buoyancy force change due to lithospheric stretching is usually much smaller than that due to crustal
The Dynamics of Continental Breakup and Extension
thinning. This is because density changes due to a temperature change of 500 C are 10 times smaller than the density difference between crust and mantle. The thermal buoyancy force change scales with the thermal lithospheric thickness squared. Thus, when the lithosphere is thick, thermal buoyancy effects can be of larger magnitude than the crustal buoyancy force change because the depth extent of temperature anomalies can be much greater than the depth to Moho, as noted by Turcotte and Emerman (1983) and Le Pichon and Alvarez (1984). Thermal and crustal buoyancy are of opposite sign for a rift. 6.08.2.3.4
therefore diminish those delocalizing effects (Burov and Cloetingh, 1997). However, it is difficult to quantify the rate of erosion or to specify either the distribution of sediment or its density. The rest of this chapter is concerned with the several controversies that continue to exercise researchers in the field of continental extension. Each of these controversies can be cast in terms of a pair of endmember concepts and the title of the following sections emphasizes these extreme possibilities.
6.08.3 High-Angle versus Low-Angle Normal Faults
Regional isostasy
Fault offset results in stress changes around and on the fault. Forsyth (1992) argues that those stress changes should inhibit continued offset on the fault. Eventually it could be easier to break a new fault in adjacent lithosphere than for continued slip on an original fault. Figure 8(d) illustrates this effect. As described in a later section the scaling between the force and the vertical deflection is directly proportional to the horizontal wavelength of the response of the lithosphere to vertical loads, represented by the flexural parameter, . It is approximately: FR ¼ gc w
½14
where w is the vertical offset across the fault. 6.08.2.3.5
Additional effects Other factors may have a considerable effect on the geometry of extension, but they are even harder to quantify than the factors discussed here. For example, erosion and sedimentation may reduce both the local isostatic (crustal buoyancy) effect and the regional isostatic effect due to fault offset. This could SW
345
Older Tertiary sedimentary and volcanic rocks
One of the most exciting and contentious areas of work on extensional tectonics in the last 20 years involves the interpretation of subhorizontal normal faults (Figure 9). These ‘low-angle normal faults’ were first recognized in continental metamorphic core complexes (e.g., Coney and Harms, 1984) but now have been seen on the ocean floor near some mid-ocean ridges (e.g., Tucholke et al., 1998, 1997). The excitement centers around whether the faults actively slipped at low dip angles (<30 ) or at greater dips. Under the tenets of classical fault mechanics, normal faults in the brittle upper crust should initiate at dips greater than 45 and should be active at dips of no less than 30 (Anderson, 1951; Byerlee, 1978; Sibson, 1985). Normal faults appear to fall into two distinct populations. High-angle normal faults dip at an angle of >30 and are offset <10 km (Vening Meisnez, 1950; Stein et al., 1988). Low-angle normal faults dip between 30 and subhorizontal, and some appear to have accommodated horizontal throws as much as 50 km or more (e.g., Wernicke, 1992; Davis Younger Tertiary sedimentary and volcanic rocks
NE
Breccia ‘Detachment’
Mylonitic foliation Metamorphic and intrusive rocks 0
km
10
Figure 9 Interpretive cross-section of the Whipple Mountains metamorphic core complex, thought to be a site of large magnitude extension with no local subsidence. The detachment fault appears to have accommodated tens of kilometers extension and is not in a subhorizontal position. From Davis GA (1980) Problems of intraplate extensional tectonics, Western United States. In: Continental Tectonics, pp. 84–95. Washington, DC: National Academy of Science.
346
The Dynamics of Continental Breakup and Extension
and Lister, 1988; Yin and Dunn, 1992; John and Foster, 1993). Before discussing ideas for the origin of low-angle normal faults the question of how topographic relief is produced by offset of high-angle normal faults will be surveyed.
6.08.3.1
Rift Shoulder Uplift
Active rifts are primarily identified by their characteristic topographic relief, with a depression or rift valley surrounded by elevated flanks or shoulders (Figures 1 and 5). Geological and geophysical data indicate that the shoulder of a rift is typically footwall of a major, riftbounding normal fault. Vening-Meisnez (1950) considered how normal fault offset might produce topographic relief. He assumed that the load was related to a fault cutting low-density crust floating on the asthenosphere. In this formulation the force supporting uplift depends on the crustal thickness, hc, and the density contrast between crust and mantle. To quantify the model, imagine that two floating blocks cut by a fault are moved apart and denser fluid upwells to fill the space between them. If the blocks are crust with density c and the fluid mantle has a density m, then the triangular parts of the blocks will be subject to buoyancy forces related to their shape. The upward force on the triangle of the footwall block is given by V0 ¼
ghc2 c ðm – c Þ 2m tan
½15
where g is the acceleration of gravity, hc is the thickness of the crust, and is the dip of the fault cutting the blocks. Applying this load to the end of a thin elastic plate floating on the mantle would produce a vertical deflection of the plate of: wðx Þ ¼
2V0 – x = x cos e m g 2
½16
where is the flexural parameter of the flexed plate (e.g., Turcotte and Schubert, 2002). Neglecting the distance from the point of application of the load and the edge of the region with full crustal thickness the maximum deflection occurs at x ¼ 0 and equals wð0Þ ¼
hc2 c ðm – c Þ 2m tan
the footwall given by eqn [3] is 1.6 km. Even though the load of sediments filling the hanging wall basin would reduce the footwall uplift, the model fit is in reasonable agreement with the observed uplift magnitude for continental rifts. This model does not work for oceanic rifts, since oceanic crust is much thinner and denser than continental crust. However, the topographic relief at midocean ridges is comparable to that seen in continental rifts. Taking hc ¼ 6 km, ¼ 12 km and c ¼ 3000 kg m3 with other values held the same, eqn [17] predicts maximum footwall uplift of 0.16 km. This is far smaller than that observed for oceanic rifts. Vening-Meisnez (1950) was right that the loads produced by normal fault offset should be supported regionally, but he did not consider the load produced by the offset of the Earth’s surface. This load is even more important than the load caused by offset of the Moho, which was treated by Vening-Meisnez (1950). The slip of a normal fault should deflect the surface of the Earth and create topographic relief even if the crustal thickness is zero and the lithosphere has the same density as the asthenosphere. Basic rock mechanics (Anderson, 1951) predicts that shear displacement should be easier than opening displacement at depth in the Earth. The top and bottom sides of a normal fault should remain in contact as the fault shears. The offset of the fault pushes the footwall block up and the hangingwall block down. Flexing in response to the loads should result in a curved fault so it is not immediately clear how to formulate a thin plate flexural approximation to the effect of fault offset. Weissel and Karner (1989) formulated the problem by conceptually turning off gravity when the fault offsets half the Earth’s surface down relative to the other side of the fault. The offsets can be considered loads equal to the vertical offset times gravity times density. When gravity is ‘turned on’ these loads deflect the lithosphere, resulting in a topographic pattern as shown in Figure 10. Using the Weissel and Karner (1989) approach an analytic solution for the deflection of the surface can be derived assuming thin elastic plate flexure theory. For the geometry in which a normal fault drops the hangingwall down (Figure 10) the load distribution is
½17
For a typical (i.e., moderate heat flow) continental region the model appears to work. For such a region it is reasonable to assume hc ¼ 30 km, ¼ 60 km, c ¼ 2900 kg m3, m ¼ 3300 kg m3, so the uplift of
V ðxÞ ¼
g tan x; g tan x0 ;
0 < x < x x > x
½18
where x ¼ 0 is the position of the intersection of the fault with the surface on the footwall (the breakaway) and x ¼ x is the intersection of the fault with
The Dynamics of Continental Breakup and Extension
347
(a) Inital (Andersonian) fault break x=0 x z
Elastic lithosphere
(b) Finite fault offset (no gravity)
Δx
Load, P = ρgΔz
z=0 Δz
(c) Flexural response when gravity is ‘turned on’
New lithosphere freezes on
Base of lithosphere thermally erodes
Figure 10 Schematic of a way to treat the effect of normal fault offset on topographic relief. (a) shows an ideally oriented fault and (b) shows how one side of the fault would move down in the absences of gravity. The magnitude of the load, P, would exist if gravity were ‘turned on’. (c) shows the flexural response to that load.
the surface on the hangingwall. The vertical deflection at position x due to a localized load V at position x9 is V ðx 9Þg – x = ðx 9 – x Þ ðx 9 – x Þ cos e þ sin 4 x9 x
½19
The maximum uplift of the footwall occurs at x ¼ 0 and this can be calculated by integrating eqn [19] from x ¼ 0 to x ¼ 1: tan – x = x x wð0Þ ¼ sin e – cos þ1 ½20 4
Figure 11 plots the value of the breakaway uplift versus horizontal fault offset. The uplift increases with offset up to the point where fault offset x ¼ (/2) then diminishes slightly. The maximum uplift is 0.3 tan . For a fault dip angle of 45 and a flexure parameter ¼ 10 km, the maximum uplift given by eqn [6] is about 3 km. This is on the high-end of the range of uplifts seen for oceanic or continental rifts. Several things can limit the development of topographic uplift across a normal fault, including sedimentary loading of the down-dropped side of a fault to produce the kind of half-graben basin seen for many continental rifts (e.g., Stein et al., 1988).
w (0)/α tan θ
wðx Þ ¼
0.375
0.250
0.125
0 0
1
2
3
4
5
(Δx /α) Figure 11 The results of an analytic calculation (eqn [1]) of the uplift of the footwall breakaway of a normal fault as a function of the horizontal offset, x, of the fault. The offset is normalized by the flexure parameter, , and the vertical uplift is normalized by tan , where is the fault dip angle.
Braun and Beaumont (1989) used numerical models of extension of a 2-D viscous-plastic layer to show that local extensional thinning of the layer (necking) could produce reasonable uplift of rift shoulders. They also interpret their results in terms of a two-stage process in which stretching would produce a topographically low basin without gravity,
348
The Dynamics of Continental Breakup and Extension
100 km
(a) 1 km
Flank uplift Basement (b)
Postrift sediments
Breakup unconformity Synrift and prerift sediments
(c)
Figure 12 Schematic of the mechanism of (a) lithospheric necking and (b) resultant uplift caused by the gravitational response to the lowering of the surface due to plate stretching. (c) illustrates how continued sediment input (combined with thermal subsidence) can lead to subsidence of once-elevated rift flanks. After Braun J and Beaumont C (eds.) (1989) Contrasting styles of lithospheric extension: Implications for differences between basin and range province and rifted continental margins. extensional tectonics and stratigraphy of the north Atlantic margins. American Association of Petrololeum Geologist Memoir 46: 53–79.
but with gravity it uplifts rift shoulders (see Figure 12). Braun and Beaumont (1989) also noted that such uplift might explain the ‘breakup unconformity’ seen at many rifted margins (Figure 13). 6.08.3.2 Low-Angle Fault Development and Stress Rotation On the basis of geological mapping of inactive faults, many authors have contended that slip has occurred along some normal faults with dip angles <30 (Wernicke, 1981; Davis and Lister, 1988; Miller and John, 1988). There is no clear evidence, however, for low-angle normal faults that are active today. Focal mechanism studies indicate the orientation of seismogenic faults. Information such as aftershock locations or the relation between surface fault breaks and hypocenters is needed to determine which nodal plane is the fault plane. Well-constrained fault-plane solutions indicate
Figure 13 Illustration of typically observed rift flank uplift and rift margin syn- and postrift sediments with an intervening ‘breakup unconformity’. From Braun J and Beaumont C (eds.) (1989) Contrasting styles of lithospheric extension: Implications for differences between basin and range province and rifted continental margins. extensional tectonics and stratigraphy of the north Atlantic margins. American Association of Petrololeum Geologist Memoir 46: 53–79.
that most seismogenic normal faults dip at angles >30 ( Jackson, 1987; Thatcher and Hill, 1991), although at least one has been interpreted as a low-angle fault (Abers, 1991). Several authors have suggested that normal faults could initiate with low dips if unusual loads reoriented the tectonic stress field itself (Spencer and Chase, 1989; Yin, 1989; Parsons and Thompson, 1993). Under the right set of regional or local loading conditions, the principal stresses might be rotated to a configuration at least geometrically compatible with low-angle normal faulting under the assumption that faulting occurs at an angle of approximately 30 to the maximum principal stress, a well-established result of Mohr–Coulomb fracture mechanics. These stress rotation models have a strong intuitive appeal. They tie a ubiquitous and puzzling feature of the Basin and Range Province, regional detachment faulting, to conditions known or strongly suspected to have existed there at the onset of extensional deformation: orogenic loading, ductile flow below the brittle layer, and widespread calc-alkaline magmatism. They also demonstrate that unusual boundary conditions can alter stress orientations. The papers advocating initiation of normal faults at low dip angles (Spencer and Chase, 1989; Yin, 1989; Parsons and Thompson, 1993) did not address the question of whether the magnitudes of the reoriented stresses would allow regional low-angle normal faulting under geologically realistic conditions. Wills and Buck (1995) carried out simple analyses designed to test this aspect of several stress-field rotation models. Their
The Dynamics of Continental Breakup and Extension
results show that the areas at which these models predict low-angle normal fault development are the least favorable places for fault slip to occur. They also quantified the magnitude of spatial variations in cohesive strength and pore pressure required to initiate slip on low-angle normal faults, variations that are implausible.
6.08.3.3
Fault Rotation
An alternative to slip on low-angle normal faults is that the upper parts of some actively slipping high-angle normal faults rotate to shallower dips. Spencer (1984) first suggested that the isostatic response to offset of a normal fault would tend to decrease the dip of the fault. However, Spencer (1984) confined his discussion to the rotation of active low-angle faults. Hamilton (1988) and Wernicke and Axen (1988) argued that large rotation of high-angle faults is consistent with the structures seen in two different extensional settings. Buck (1988) also argued that rotation could explain low-angle fault structures, and calculated the flexural response of lithosphere to the loads caused by the offset of a high-angle normal fault. This model produced realistic low-angle fault geometries only when (1) the lithospheric yield strength was finite and (2) when the offset of the model fault was about twice the lithospheric thickness. To get the inactive, up-dip parts of model normal faults to rotate to a low-angle orientation Buck (1988) assumed that large offsets, relative to the lithospheric thickness, could occur on a high-angle normal fault to produce low-angle fault structures. A major question is why such large offsets might occur on some high-angle normal faults and not on others.
6.08.3.4
Large Offset of Normal Faults
The offset of a dip-slip fault produces topography and so changes the stresses around the fault. For a highangle normal fault, the topographic relief should build up quickly as the fault is offset. Vening-Meisnez (1950) recognized this and was among the first to suggest that the stress changes related to normal fault offset could result in new faults being formed. Offset of an elastic layer by slip of one fault produces maximum bending stresses at about a flexural wavelength from the fault. Vening-Meisnez (1950) assumed the next fault would break where the bending stresses at the surface were maximally extensional, and result in a graben. This assumption is reasonable, since the yield stress (the stress needed to break and slip on a fault) is minimum at the surface.
349
A different approach to analyzing the effect of normal fault offset on stresses was suggested by Forsyth (1992). In contrast to Vening-Meisnez (1950) he ignored the direct effect of bending stresses on promoting layer breaking and instead estimated the increase in the average regional tectonic stress in a layer due to the buildup of fault related topography. Forsyth (1992) noted that Anderson’s theory for normal faulting is only valid for infinitesimal fault slip because it only considers the work done overcoming friction on the fault surface. The initial orientation of a fault requiring the least regional stress is the one that dissipates the least friction on the fault per unit of horizontal displacement. When a fault is displaced, work is done in the bending of the lithospheric plate cute by the fault. To estimate this work, Forsyth (1992) approximated the lithosphere as a perfectly elastic layer floating on an inviscid substrate. The deflection of the layer caused by fault offset is estimated by using the thin-plate flexure equation. To do this analytically, the fault is treated as a vertical boundary cutting the lithosphere. The topographic step across the fault is the horizontal fault throw, x, times tan , where is the fault dip. Because of the work done bending the lithosphere, it takes extra horizontal tensional stress to keep the slip occurring on the fault. Forsyth (1992) found that the extra horizontal stress increases linearly with x depends on tan2 . He estimated that after only a few hundred meters of slip on a typical high-angle fault, it is easier to break a new fault rather than to continue slip on the original fault. On a low-angle fault, the extra horizontal stress needed for continued motion builds up more slowly with offset. Forsyth (1992) suggested that an initially low-angle normal fault could build up much more offset than a high-angle fault. Buck (1993) used the approach of relating work building topography to tectonic forces suggested by Forsyth (1992), but he modified the way fault dip was considered (Figure 14). As in the description of topographic relief produced by fault offset, described above, the nonvertical dip of the fault significantly reduces the topography-related work to continue fault displacement. Another important change in formulation was inclusion of the effect of finite yield strength on bending stresses. A Mohr–Coulomb plate will bend more easily than a purely elastic plate, since bending stresses cannot exceed the yield stress. Inclusion of finite yield stress in this model radically lowers the size of the tectonic force increase due to fault related topography (as shown in Figure 15). Using reasonable values for friction and cohesion
350
The Dynamics of Continental Breakup and Extension
τ
σ1
(a)
θ
h
σR
σ3
2kμ
σR σ3 σ σ1 R
Δx
(b)
σ3 = ρgh
τ
σP
σP σR τ
τ0
(c)
2
– σ3
kμ =
ρgh 2kμ
–
σP kμ
σn
2kμ
σR ρgh
σ3 =
σ3 = ρgh –
τ0
σP
σR =
σn
2 τ0 (1 + μ 2)1/2 + μ
σn
(1 + μ 2)1/2 + μ (1 + μ 2)1/2 – μ
Figure 14 Illustration of the effect of plate bending stresses on the stress state required for slip on a normal fault. (a) shows the initial Andersonian stress state for slip on a cohesionless, optimally oriented fault. (b) shows the lowering of the minimum stress, 3, required for slip when the plate bending stress can be considered to reduce the vertical stress by an amount p. (c) shows the regional stress difference needed to break a new fault with a shear strength 0. Here is the friction coefficient. From Buck WR (1993) Effect of lithospheric thickness on the formation of high-and low-angle normal faults. Geology 21: 933–936.
θ = 60°, h = 10 km
6.08.3.5 2-D Models of Fault Formation and Offset
Plate Stress, σp (MPa)
500 Simplified geometry elastic
400
~ tan2 θ
Elastic ~ tan2 θ
300
200
100 Elastic–Plastic ~ tan3/2 θ 0
0
4 8 12 16 Horizontal throw, Δx (km)
20
Figure 15 Results of analytic and thin plate flexure numerical calculation of the change of plate stress with horizontal fault offset for a floating brittle layer 10 km thick and a 60 dipping normal fault. The straight line is from Forsyth (1992) while the elastic curve includes the effect of finite fault dip. The inclusion of finite yield strength of the brittle layer greatly reduces the magnitude of the plate bending stress. From Buck WR (1993) Effect of lithospheric thickness on the formation of high-and low-angle normal faults. Geology 21: 933–936.
for a 10 km thick layer, the Buck (1993) model predicts possibly unlimited fault offset. For a thicker layer the fault offset may be limited to an amount smaller than the layer thickness (Figure 16).
The thin-plate approximation used in the studies described above is clearly not valid for large fault offset. To be confident of internal consistency in normal fault evolution models require treatment of the 2or 3-D stress and strain field. Analog models are useful for simulating the early, small-offset stage of fault development (e.g., Tirel et al., 2006; Withjack and Schlische, 2006; Brun et al., 1994; Corti et al., 2003; Tron and Brun, 1991; Supak et al., 2006), but cannot easily simulate the thermally controlled strength field evolution likely to affect large offset faults. Ideally, models could follow the development of faults in an extending, 3-D, viscous–elastic–plastic layer. Given the numerical cost of such models groups have first developed 2-D models (Figure 17). Early studies numerically simulated normal fault offset in 2-D cross-sections of elastic layers, assuming that the fault offset varies slowly in the third dimension (e.g., Melosh and Williams, 1989; King and Ellis, 1990). These studies assumed a pre-existing weak fault embedded in a purely elastic layer and solved for the topographic relief and stress changes around the offset faults. Given the potential importance of the finite brittle yield strength (here described as plastic deformation) in controlling how lithosphere can bend in response to fault offset, there has been great effort to
The Dynamics of Continental Breakup and Extension
Large-offset, low-angle normal fault Δx Upper-crustal lithosphere
h
Lower-crustal asthenosphere Moho
Small-offset, high-angle normal fault Δx
Lithosphere
h Moho
Asthenosphere Figure 16 Cartoons of possible implications of the calculations of plate bending stresses given in Figure 15. Because plate bending stresses scale with the square of plate thickness while the cohesive strength of the plate scales linearly with thickness a noncohesive fault in a thick layer could accrue very large offset. A noncohesive fault cutting a thick cohesive layer might not achieve large offset before another fault replaced it.
include plastic deformation in numerical models. Braun and Beaumont (1989) and Bassi (1991) treat the lithosphere as a viscoplastic layer but were not concerned with localized fault development. Behn et al. (2002) and Huismans et al. (2005) allowed strain rate or strain-dependent weakening that lead to localized zones of concentrated deformation (model ‘faults’). Poliakov and Buck (1998) adapted a numerical treatment of viscous–elastic–plastic deformation to normal fault development. They showed that a sequence of high-angle faults might form and accommodate extension at a simple model mid-ocean ridge structure. In the Poliakov and Buck (1998) formulation the faults weaken as a function of their offset related strain, up to a maximum amount of weakening. Figure 17 shows results of 2-D numerical experiments of extending an elastic–plastic layer that stays nearly uniform in thickness while a fault forms due to strain-dependent cohesion loss (see Section 6.08.2). The fault offset produces topography very much like that predicted by eqn [19] except that the wavelength of deformation appears to evolve in the early stage of model fault offset. The maximum fault produced
351
topographic relief, shown in Figure 18, also follows the pattern predicted by the simple analytic expression derived above (eqn [20]). Numerical experiments also have addressed the question of fault offset and rotation. Lavier et al. (1999, 2000) used an elastic–plastic formulation and investigated how much fault weakening is needed to get large offset of a normal fault. The base of the extending layer was kept at a constant depth to simulate cooling of asthenosphere pulled up as the footwall of the fault moved out and up. Figure 19 shows the results of one model experiment that produced a large offset fault. The inactive, up-dip part of the footwall side of the fault was rotated into a flat lying position. Lavier et al. (2000) showed large offset faults could only form when the fault weakening was above a minimum level, but that the rate of fault weakening with strain had to be within bounds. If the maximum amount of fault strain weakening is independent of the thickness of the brittle layer being extended then large offset faults will only happen when a layer is thinner than a given thickness. Figure 20 illustrates the force changes due to normal fault offset estimated for thin and thick layers based on the numerical models. Only for a thin layer do we expect to get a large offset fault. Figure 21 (from Lavier and Buck, 2002) shows two model cases of normal fault offset in extending viscous–elastic–plastic layers where the thermal structure is allowed to vary due to advection and diffusion of heat. It shows that a thin layer could develop a large offset fault while multiple, basin-bounding faults develop in a thicker layer. This is consistent with the observation that large offset normal faults are only seen in areas of higher than average heat flow where one expects thinner than average lithosphere.
6.08.4 Pure versus Simple Shear Rifting Rift valleys are partly filled with sediment and rifted margins are characterized by syn- to post-rifting sedimentary units that may be more than 10 km thick. Such thick piles of sediment can only be accommodated if the rifted region subsides by several kilometers. The sediments act as a load that pushes down the surface and amplifies the amount of subsidence that would occur with no sedimentary infill. If a rift basin is narrow compared to the flexural wavelength of lithospheric response, then the load is regionally compensated. As noted in the last
352
The Dynamics of Continental Breakup and Extension
(a) Setup
Yield stress o
x
0
Extension
σy
Compression Co
Brittle layer
Fr
H
z
he
ict
sio
i on
n+
fri
cti
on
z
Inviscid substrate
Depth (km)
2 0 –2 –4 –6 –8 –10
0.0
1.8
3.6
0.0
1.7
3.4
0.0
1.8
3.7
0.0
1.9
3.8
0.0
1.8
3.6
0.0
1.9
3.8
2 0 –2 –4 –6 –8 –10
2 0 –2 –4 –6 –8 –10
2 0 –2 –4 –6 –8 –10
2 0 –2 –4 –6 –8 –10
Increasing extension
Depth (km)
Depth (km)
Depth (km)
Depth (km)
Depth (km)
(b). Results
2 0 –2 –4 –6 –8 –10
Figure 17 (a) Setup for a numerical model of extension of a floating brittle Mohr–Coulomb layer with a single normal fault seeded at the center of the model layer. Grid spacing is 250 m through this 10 km thick layer and a cohesion of 20 MPa is reduced to 2 MPa over a fault offset of 1500 m. (b) Results of numerical calculation showing the development of plastic (brittle strain) and surface topography. Note that the model is similar in setup to the conceptual model pictures in Figure 10.
section, sediment loading can pull down the rift shoulders as it fills in the rift basin. If the rift is wider than the flexural wavelength, then the sediment can be treated as a locally compensated load.
To illustrate local compensation of sediment consider the simple case of an air filled rift basin with an initial depth of di. Imagine that gravity is ‘turned off’ as the basin is filled with sediment. When gravity is
The Dynamics of Continental Breakup and Extension
Positive topography (m)
5000
4000
3000
2000
1000
10 km thin layer 20 km thick layer
0 0
5
10
15
20
Horizontal offset (km) Figure 18 Results of the uplift of the footwall breakaway point as a function of horizontal fault offset. The 10 km layer case is pictured in Figure 17 while the results of a 20 km thick layer with similar setup are also shown. The numerical results fit very well by the analytic estimate shown in Figure 11 for a flexure parameter of about 0.8 times the layer thickness and for a fault dip of 50 .
turned on the sediment load pressure is the fill depth, ds (¼di), times the sediment density, s, times the acceleration of gravity, g. This load will push down the surface and, if we assume that sediment inflow keeps the basin filled to sea level, then the basin will continue to deepen until the sediment load is ‘compensated’. Local compensation implies that the crust and mantle lithosphere floats on fluid mantle asthenosphere. This means that the weight of a column of crust and mantle to an arbitrary ‘depth of compensation’ must be constant. As sediment fills the hole it displaces mantle which is denser than sediment. Equating the column weight before and after basin fill, the ultimate basin fill depth ds is related to the unfilled depth as ds ¼ di m =ðm – s Þ
½21
where m is the density of mantle. Taking s ¼ 2600 kg m3 and m ¼ 3250 kg m 3, a 1 km deep air filled basin would be filled with locally compensated sediment to 5 km. Deep holes might be likely to be filled in by water before being filled with sediment, but this intermediate step does not change the results of this simple analysis. To get thick sedimentary packages requires a depression or hole. Wells drilled into rifted margins showed two interesting things about the thick sedimentary sequences (Figure 22). The sediments were deposited over a long time (tens to hundreds of million years) and the infilling sediments were
353
generally deposited at shallow water depths in a lake or in an ocean basin (Sleep, 1971; Steckler and Watts, 1978). Since compensation should occur on a timescale of thousands of years, these observations imply that the hole the sediments were filling had subsided with time. To understand the mechanism by which rift basins and rifted margins subside we need to understand subsidence of ocean basins. One of the most striking features of the ocean basins is the smooth pattern of depth increase with distance from the mid-ocean ridges. The depth of seafloor increases with the square root of its age (Parsons and Sclater, 1977). This is a clear confirmation of plate tectonics, which holds that the oceanic plates are formed when a hot mantle and magma upwells at a spreading center and then cools as they move away. Since most plates move laterally much faster then they thicken by cooling (except very close to the spreading center) this cooling may be approximated by the cooling of a half-space of hot material. Lithospheric material contracts and so becomes denser as it cools. Thus, as the lithosphere cools and shrinks it will float lower on the asthenosphere. Combining half-space cooling with thermal contraction and local compensation predicts subsidence that is proportional to the square root of lithospheric age (e.g., Langseth and Taylor, 1967; McKenzie, 1967; Parsons and Sclater, 1977). Sleep (1971) noted that the rate of subsidence of rifted margins also is related to the square root of sediment age. This suggested that rift margin subsidence might be controlled by lithosphere cooling down from a hot state at the time of rifting. However, there is still the problem of how to get a initially hot region to cool. Sleep reasoned that if continental lithosphere, with a surface close to sea level, were heated and thinned thermal expansion would float the surface above sea level. The surface would gradually subside back to sea level and no marine sediments could fill in. Something had to happen to lower the surface. The idea favored by Sleep (1971) was erosion of the uplifted surface. If a layer of low-density crust (compared to mantle) were removed then the surface could subside below sea level as cooling proceeded. However, the sediments that would be eroded off the elevated region are not seen. Several authors (Artemjev and Artyushkov, 1971; Salveson, 1978; McKenzie, 1978) suggested that thinning of the lithosphere by pure shear extension created the thermal input needed to explain the later thermal subsidence. Because there is no active
354
The Dynamics of Continental Breakup and Extension ΔF/F(0) = 0.33, Δx f = 1.5 km Topography (m)
1500 1000 500 0 –500 –1000 –1500
VE 4:1
10
30
50 Distance (km)
70
1500 1000 500 0 –500 –1000 –1500
90
Depth (km)
VE 1.5:1 0 2 4 6 8 10
0 2 4 6 8 10
Plastic strain
Topography (m)
0.0
4.5
9.0
1500 1000 500 0 –500 –1000 –1500
1500 1000 500 0 –500 –1000 –1500 10
30
50
70
90
0 2 4 6 8 10
0 2 4 6 8 10
Plastic strain
Depth (km)
Topography (m)
0.0 1500 1000 500 0 –500 –1000 –1500
10
4.5
30
50 Distance (km)
Increasing extension
Depth (km)
Distance (km)
9.0
70
90
0 2 4 6 8 10
1500 1000 500 0 –500 –1000 –1500
0 2 4 6 8 10
Plastic strain 0.0
1.6
3.1
Figure 19 Results of a numerical model of extension of a floating brittle Mohr–Coulomb layer with a single-seeded normal fault. The model is similar to that shown in Figure 17 but the calculated amount of extension is large enough to allow the footwall of the fault to rotate so that abandoned parts of the fault rotate to, and even past, horizontal. The cohesion loss of the fault was a function of strain with a decrease of one-third of the initial layer brittle yield strength occurring linearly with fault offset up to 1.5 km. From Lavier L, et al. (2000) Factors controlling normal fault offset in an ideal brittle layer. Journal of Geophysical Research 105(B10): 23431–23442.
The Dynamics of Continental Breakup and Extension
Thin layer, moderate weakening rate
Force change on fault offset
(a)
~H 2 Perturbation
Bending Fault offset, Δx Total ~ΔCH
Δxc
Weakening
Thick layer, moderate weakening rate
Force change on fault offset
(b)
Bending
~H 2
New fault breaks
Fault offset, Δx Total ~ΔCH Weakening
Δxc Figure 20 Schematic explanation of the reason that thin layers might allow unlimited normal fault offset while a thicker layer would not. The force change due to bending is compared to the assumed (or model input) fault weakening amount. If the summed force change is negative then the fault can continue slip (a) while if the force change becomes positive the fault should be replaced by another fault.
asthenospheric heating of the lithosphere in these models, they are termed ‘passive’. The crustal thinning allows subsidence below sea level, or the original level of the continent, and sediment accumulation. Mckenzie (1978) quantified the effects of uniform pure shear extension or ‘stretching’ of the entire lithosphere. Stretching, should thin both the crust and thermal lithosphere (see Figure 23). Local compensation of crustal thinning gives ‘tectonic subsidence’ on the timescale of the thinning. Lithospheric thinning causes immediate uplift followed by slow ‘thermal subsidence’. In McKenzie’s stretching model the rift is approximated as a rectangular region of extenional pure shear. Pure shear describes homogeneous thinning of an entire block of material by a stretching factor given by the ratio of the initial to final thickness ( > 1 for extensional thinning).
355
If the initial thermal gradient in the lithosphere is given by dT/dz, instantaneous pure shear thinning will result in steepening of the gradient to dT/dz. After rifting, time-dependent conductive cooling of the lithosphere eventually results in re-equilibration of the thermal gradient to its initial value. The stretching model is useful since it gives a 1-D description of the effect of lithospheric extension and so is easy to implement mathematically. McKenzie’s (1978) model neglects lateral heat flow and heat loss during the finite time duration of extension but matches the main characteristics of the subsidence histories of many margins (see Figure 24). The instantaneous uniform pure shear extension model was further generalized by Jarvis and McKenzie (1980) to include the effects of vertical heat loss during a finite duration extension event. They were able to model the heat flow and subsidence through a finite period of lithospheric stretching and thinning and the subsequent evolution after the end of the extension event. In general, the heat flow and subsidence history differed little from the instantaneous extension model, provided the time during which extension occurs is not too long. For a lithosphere thinned to half of its original thickness, for example, they found good agreement with the instantaneous case provided the duration of extension is less than 15 Ma. Though the stretching model gives a simple explanation of the approximate pattern of subsidence seen at most rifts it fails to explain the details of the subsidence. Where constraints on subsidence are good, the ratio of thermal subsidence to tectonic subsidence is greater than predicted by the stretching model (e.g., Royden and Keen, 1980; Davis and Kuszner, 2004; see Figure 25). The problem is even worse if the effects of lateral heat conduction are added to the model (Cochran, 1983b; Alvarez et al., 1984). The syn-rift subsidence is augmented owing to the additional lateral heat loss, and the post-rift subsidence is diminished compared to the 1-D model. It appears that extra lithospheric thinning is needed compared to a standard stretching model. This observation led to the proposal of a two-layer stretching model (Royden and Keen, 1980) in which the subcrustal lithosphere is decoupled from the crust and can be extended by a greater amount than the crust (see Figure 26). Although this model provides a simple way to specify the amount of lithospheric heating independently of the amount of crustal extension, it lacks a physical mechanism for this process.
Hydrothermal cooling H = 10 km, Δxc = 1.5 km, Hk = 6 km, Tk = 400° C
4
Topo (km)
2 0 –1 –3
2 0 –1 –3
0
20
40
60
80
100
120
0 –5 –10 –15 –20 –25 –30 0
140
20
40
Depth (km)
40
60
80
100
120
20
40
60
80
Distance (km)
100
120
Temperature (°C)
0 27.5 26.9 26.2 25.6 25.0 24.4 23.7 23.1 22.5 21.8 21.2 20.6 19.9 19.3 18.7
0 –5 –10 –15 –20 –25 –30 0
100
120
700 650 600 550 500 450 400 350 300 250 200 150 100 50 0
27.5 26.9 26.2 25.6 25.0 24.4 23.7 23.1 22.5 21.8 21.2 20.6 19.9 19.3 18.7
140
0 –5 –10 –15 –20 –25 –30
140
140
Viscosity (log10Pa s)
Depth (km)
700 650 600 550 500 450 400 350 300 250 200 150 100 50 0
0 –5 –10 –15 –20 –25 –30 20
80
Distance(km)
Distance(km)
0
60
3.00 2.79 2.58 2.37 2.17 1.96 1.75 1.55 1.34 1.13 0.92 0.72 0.51 0.30 0.10
Total strain
3.00 2.79 2.58 2.37 2.17 1.96 1.75 1.55 1.34 1.13 0.92 0.72 0.51 0.30 0.10
0 –5 –10 –15 –20 –25 –30
Temperature (°C)
–5
Total strain
Depth (km)
–5
H = 15 km, Δxc = 1.5 km, Hk = 6 km, Tk = 400° C
Viscosity (log10Pa s)
Topo (km)
4
20
40
60
80
100
120
140
0 –5 –10 –15 –20 –25 –30 0
20
40
60
80
100
120
140
Distance (km)
Figure 21 Results of numerical model calculations for extension of thin and thick brittle lithosphere. For both cases, the rheology is viscous–elastic–plastic with the viscous strength depending strongly on temperature. The evolving temperature field is computed under the assumption that hydrothermal circulation cools the shallow part of the lithosphere. Extension of the hotter, thinner lithosphere leads to a single, large-offset fault. With extension of the cooler thicker layer multiple normal faults form. From Lavier L, Buck WR (2002) Half graben versus large-offset low-angle normal fault: The importance of keeping cool during normal faulting. Journal of Geophysical Research 107: 2122.
The Dynamics of Continental Breakup and Extension
Time (My BP)
(a) Local loading 140 0
120
Age (My BP) 100 80 60
40
20
200 0
0
Due to tectonic driving force
150
100
Due to sediments
Subsidence (km)
2000
0
50
Cost B-2 Airy
1
1000
Depth (m)
357
2 3
β=2 4
β=3
5
β=4
3000 6
MOR
7
4000
5000 Figure 22 Subsidence data for the margin of the western Mediterranean showing that subsidence was not driven just by sediment loading. Estimated components of subsidence with time due to the isostatic response to sediment loading compared to the subsidence that would occur without sediment infill, here labeled as being due to an (unspecified) tectonic driving force. From Watts AB and Ryan WBF (1976) Flexure of the lithosphere and continental margin basins. Tectonophysics 36: 25–44.
Figure 24 Comparison of subsidence data for the COST-B2 well on the US East Coast with predictions of the stretching model with different stretching factors,
. From Steckler MS, Berthelot F, Lyberis N, and LePichon X (1988a) Subsidence in the Gulf of Suez: Implications for rifting and plate kinematics. Tectonophysics 153: 249–270.
SNORRI
0
Without Sediment 1 Depth (km)
Initial conditions
T (°C) 0 1000
With Sediment
2
Theoretical
3
Extension Faulting Hinge-zone formation Crustal thinning Lithospheric heating Syn-rift sediments
δ
4 80
Cooling Thermal subsidence Flexure Post-rift sediments Hinge zone Coastal onlap
Figure 23 Schematic illustration of the McKenzie (1978) lithospheric stretching model. From Steckler MS, Berthelot F,Lyberis N, and LePichon X (1988a) Subsidence in the Gulf of Suez: Implications for rifting and plate kinematics. Tectonophysics 153: 249–270.
β
1.6 2.5 1.6 3.0 1.6 5.0 60
40 Age (Ma)
20
0
Figure 25 Subsidence data (dashed and dotted lines) for a well on the Labrador continental margin. Solid lines are predictions of a model with greater stretching of the lithosphere (factor ) compared to stretching of the crust (by a factor ). From Royden L and Keen CE (1980) Rifting process and thermal evolution of the continental margin of eastern Canada determined from subsidence curves. Earth and Planetary Science Letters 51: 343–361.
358
The Dynamics of Continental Breakup and Extension
(b) Uniform extension T (°C) O
(a)
Initial conditions
T1
α/β
(a) Pure shear Brittle–ductile transition
T (°C) T1 O y
a
(c) Nonuniform extension y/δ
T (°C) O T1
(a – y)/β
Crust–mantle boundary
Figure 26 Schematic of the ‘two-layer’ stretching model in which the lithosphere is stretched by a different amount from the crust. From Royden L and Keen CE (1980) Rifting process and thermal evolution of the continental margin of eastern Canada determined from subsidence curves. Earth and Planetary Science Letters 51: 343–361.
Upwelling asthenosphere
Lithosphere–asthenosphere boundary
(b) Simple shear Brittle–ductile transition
Another way to explain the observed increase in thermal relative to initial subsidence is to horizontally offset the locus of crustal versus lithospheric thinning. This is key a feature of the ‘simple shear model’ of lithospheric extension that was advanced by Wernicke (1985). One side of the rift is viewed as the upper plate and the other is the lower plate of gently dipping normal fault or detachment (see Figure 27). Offset of the detachment would lead to very asymmetric syn- and post-rift subsidence. On the upper plate side areas of little crustal thinning occur over places where the lithosphere is greatly thinned. Thus, the upper plate side could be a place of little initial tectonic subsidence and large magnitude thermal subsidence, fitting the general pattern of observed rifted margin subsidence. Several geological and geophysical observations have suggested the importance of simple shear extension in the crust which gives rise to asymmetric structures. Low-angle normal faults exposed at the Earth’s surface have been traced to mid-crustal levels using seismic reflection techniques. Seismic data in the eastern Basin and Range show low-angle normal faults that penetrate the upper and middle crust (Allmendinger et al., 1987). The Bay of Biscay, which has been considered a classic example of a pure shear margin (de Charpel et al., 1978), is interpreted by Le Pichon and Barbier (1987) to show evidence of crustal-scale simple shear along a detachment. Metamorphic core complexes show that rocks can come up from mid-crustal levels apparently in association with extensional shear zones (Davis, 1983). The observation of synthetically dipping normal faults over broad areas (e.g., the southern Basin
Dikes
Half-graben complex
Crust–mantle boundary Detachment fault Upwelling asthenosphere Lithosphere–asthenosphere boundary
Figure 27 Schematic of structures formed by either pure shear or simple shear lithospheric extension. From Lister GS and Davis GA (1989) Models for the formation of metamorphic core complexes and mylonitic detachment terranes. Journal of Structural Geology 11: 65–94.
and Range; e.g., Stewart, 1978) has been taken as evidence for simple shear (Wernicke, 1981). Strong topographic and volcanic asymmetries exist across some rifts and conjugate passive margins, among them the Red Sea Rift (Wernicke, 1985), East African Rift (Bosworth, 1987) and the Southeast Australian–Lord Howe Rise conjugate margin. The direct observations of simple shear extension are confined to the crust. However, the observation of normal faults extending to great depth in the crust combined with observations of topographic asymmetries across rifts have led to the suggestion that normal faults and/or ductile shear zones may extend through the entire lithosphere (Wernicke, 1981, 1985; Lister et al., 1986). Such lithospheric detachment models imply that much of the deformation in
The Dynamics of Continental Breakup and Extension
an extending region occurs as simple shear rather than pure shear. Particularly good data to test the predictions of the simple shear model come from studies of the Northern Red Sea. The asymmetric uplift and volcanism bordering the Red Sea is well known and has been cited as evidence for a through-going lithospheric detachment (Wernicke, 1985). The kinematic history of opening of the northern Red Sea is well constrained by geological and geophysical data. Although rifting in the Red Sea began around the end of the Oligocene–Early Miocene (Cochran, 1983a), evidence from the Gulf of Suez and Gulf of Aqaba (Steckler et al., 1988a) suggests that in the northern Red Sea, most of the extension has occurred within the last 19 My and that it continues to the present. The young age of this rift is particularly important for discriminating between models of its formation because many of the effects are transient and will not be observable tens of millions of years after rifting. Geophysical fieldwork in the northern Red Sea provided excellent heat-flow coverage of this area (Cochran et al., 1986; Martinez and Cochran, 1988). To compare the predictions of the simple shear model to this data Buck et al. (1988) used a numerical technique that solved for 2-D conductive and advective heat transport through time. Simple shear extension of the lithosphere was modeled as occurring along a straight shear zone and the topographic response to simple shear could be described in terms of local isostasy. The long wavelength topographic asymmetry across the Red Sea, which has been cited as evidence for simple shear extension of the lithosphere, was not matched by any of the simple shear model cases. The observed high heat flow anomalies in the Red Sea require a large component of pure shear lithospheric extension centered under the region of maximum crustal extension. In contrast, at the plate separation rate of the northern Red Sea, simple shear extension of the lithosphere along a shallow (<30 ) dip detachment is ineffective in reproducing the observed heat-flow anomalies. Only a narrowing region of pure shear extension can satisfy the width of the rift, and the peak observed heat-flow values of 300 mW m2. Other, older continental margins have been studied in terms of whether various kinematic extensional models can explain their structure and subsidence patterns. However, as conjugate margins have been studied it seems that both rifted margins subside the way this model predicts for the upper plate margin. This has been described by Driscoll and Karner (1998) as the ‘upper plate paradox’.
359
6.08.5 Wide versus Narrow Rifts It has long been recognized that extension of continents can result in either narrow or wide rifts (e.g., England, 1983). For narrow rifts, such as the Rhinegraben (Illies and Greiner, 1978), the Gulf of Suez (Steckler et al., 1988b), the East African Rift System (e.g., Ebinger et al., 1989), the Baikal Rift (e.g., Zorin, 1981), and the Rio Grande Rift (e.g., Morgan et al., 1986), the region of intense normal faulting is on the order of 100 km wide. For wide rifts, the type example of which is the Basin and Range Province of western North America, the region of significant normal faulting is as much as 800 km wide (Hamilton, 1987). The Basin and Range is characterized by small lateral gradients in crustal thickness, while narrow rifts may show large lateral gradients in thickness of crust and in topography. Most rifts are relatively narrow (Figures 1 and 4) and this is generally ascribed to the weakening effect of thermal advection. The lithosphere can be thinned by stretching or heated by injection of magma. Since hot crust and mantle materials flow at lower yield stresses (eqn [4]) a hotter region will be weaker. Extension should stay concentrated in the region of weakest lithosphere. The zone of extensional strain should remain confined to a region that is about as wide as the lithosphere is thick. If a region of focused extension should become effectively stronger than the surrounding region then one might expect the locus of extension to migrate to the weaker regions. If this continued then a region of rifting that is wider than the lithospheric thickness might develop. Three delocalizing processes have been suggested as ways to get wide rifts: (1) thermal diffusion strengthening, (2) viscous strain-rate strengthening, and (3) the change in gravitational stresses due to crustal buoyancy. The scaling relations derived in an earlier section help us to evaluate the potential of each of these mechanisms to produce a wide rift.
6.08.5.1
Slow Rifting and Thermal Diffusion
England (1983) used the thin-layer approximate to investigate whether wide rifts could occur due to thermal diffusion strengthening. He considered the evolution of continental yield strength during extension and noted that, if the mantle is much stronger than the crust, then very slow extension could lead to a strength increase in the extending area. The basic idea is that weak crust is thinned by stretching and
360
The Dynamics of Continental Breakup and Extension
isostatically ‘replaced’ by intrinsically stronger mantle. Diffusive cooling has to be significant to have this compositional strengthening trump the advective weakening effect. To get this effect to produce strengthening during extension England (1983) had to assume that mantle is much stronger than crust and that extension rates were very slow. According to England (1983), as well as other workers (Kusznir and Park, 1987; Sonder and England, 1989), the diffusive strengthening effect explains the occurrence of wide regions of continental extension like the Aegean Sea and the Basin and Range Province. However, observations do not seem to confirm this prediction since some wide rifts did not extend at a slower velocity than did some narrow rifts. Sawyer (1985) questioned the validity of the England (1983) analysis since it did not consider the effect of finite brittle yield strength of the crust and mantle. The yield strength in the England (1983) model only involved the viscous yield strength. Sawyer (1985) has questioned whether the crust– mantle strength contrast is likely to be as large as assumed by England (1983). At a range of temperatures likely to be found at the base of the crust, the ductine yield strength of crust can be thousands of megapascals lower than that of the mantle. Sawyer (1985) noted that the brittle failure stress may be independent of rock type so that the contrast in crust and mantle strength is limited. When even conservative brittle yield strength values are included it greatly lowers the threshold extension rates that would lead to lithospheric strengthening and so widening of a rift. Also, if the crust and mantle are dry, then the viscous yield strength of the crust may not be significantly different from that of the mantle (Kohlstedt et al., 1995; Mackwell et al., 1998). Naturally, the importance of diffusion relative to advection of heat depends strongly on the rate of deformation. Scaling arguments can give a good idea of the strain rates that could produce extensional strengthening. The force for continued extension is reduced due to advective thinning by roughly 2"CHb2. The best case to make the slow-strain-strengthening mechanism work is when the brittle layer thickness Hb is close to the crustal thickness Hc. The force for continued extension is reduced due to advective thinning _ . A ratio of the advective and diffusive by 2C ð"="Þ force changes should be less than 1 for cases where the diffusive effect dominates. This occurs when "_ < /Hc2. For Hc ¼ 40 km and ¼ 10–6 m2 s1 the maximum estimated strain rate is 10–15 s1. This is an extremely low extensional strain rate. To get such a low strain rate
the side of a 40 km wide zone of extension would have to be separating a velocity of less than 1 mm yr1! This scaling shows that the diffusive strengthening effect becomes more important for thinner crust, assuming that the base of the crust is hot enough for the crust there to be ductile and so weaker than the mantle. Still, White (2004) suggests that England’s (1983) mechanism might explain the abandonment of many intracratonic rifts that seem to have extended at very low strain rates. 6.08.5.2
Viscous Stresses
Another possible effect to explain how wide rifts form relates to the stresses needed for yielding by viscous flow. The ductile yield stress in a representative flow law (eqn [4]) depends on strain rate. Thus, any area extending at a high rate will extend at higher stresses. This idea prompted Bassi (1991) to suggest that wide rifts could be a product of viscous stress resisting localization of deformation. It is difficult to consider using simple scaling relations to look at competition between viscous strain-rate strengthing and other effects. The difficulty is that the viscous effect is independent of strain and most weakening (localizing) effects increase with increasing strain. To analyze the effect of viscosity on deformation a linear stability approach is often used. The linearization of the model equations, implicit in these calculations, means that these results are strictly applicable only to the earliest phase of extension when strains are small. These models predict only two forms of extension of an initially uniform lithosphere, both of which could be termed a wide rift (e.g., Fletcher and Hallet, 1983; Zuber and Parmentier, 1986; Martinod and Davy, 1992). The lithosphere extends either in a uniform pure shear mode, or with laterally periodic variation in the rate of extension. This second type of extension, often called ‘lithospheric boudinage’, is suggested to explain the development of a series of basins and ranges in a broad region of continental extension. Linear stability analysis does not predict a narrow rift mode of extension. Since many rifts, like the Rhine graben and the Red Sea are narrow, one wonders what situations lead to their development as compared to wide rifts like the Basin and Range. Zuber and Parmentier (1986) and Bassi (1991) show that a large initial perturbation in lithospheric thickness or strength might cause localization in a single narrow rift, as opposed to a sequence of basins and ranges. The effect of viscous stresses on the distribution of extension in finite amplitude (as opposed to linear
The Dynamics of Continental Breakup and Extension
361
Yield stress (MPa)
(a) Setup 0
Depth (km)
Brittle upper crust Viscous lower crust
0
100 Fric tion
Low
Cohesion plus friction
ate ain r
str
High strain rate
25
20 km 500 0 –500 –1000 –1500 –2000
Depth (km)
Topography (m)
(b) Results
–100
–50
0
50
100
Distance (km) 0 –10 –20 0.0
0.012 5
0.025
Figure 28 (a) Numerical setup and (b) results for a model in which the viscosity of the lithosphere forces a distributed pattern of extension. Color scale indicated brittle plastic strain in a series of local basins that are separated by higher ranges. After Buck WR, et al. (2003) A numerical model of lithospheric extension producing fault bounded basins and ranges. International Geology Review 45(8): 712–723.
stability) numerical model has been treated by a number of workers (e.g., Bassi, 1991; Huismans et al., 2003, 2005; Buck et al., 2003; Nagel and Buck, 2006). Figure 28 shows the distribution of strain in a series of basins and ranges in a model with viscous effects driving delocalization. The viscous effects depend both on the viscosity variation with depth at the base of the lithosphere and on the extent to which viscous material weakens as a functions of strain or strain rate. In no models with large viscous delocalization do we get model large offset faults (e.g., Nagel and Buck, 2006). Thus, the existence of localized large-offset strike slip and normal faults argues against the importance of viscous delocalization. 6.08.5.3
Local Isostatic Crustal Thinning
The gravitational stresses related to local thinning of the crust could lead to wide rifting. To do this, the stretching-related gravitational delocalizing effect has to be larger than the localizing (weakeing) effect of lithospheric advective thinning. The scaling relations derived in Section 6.08.2 give an indication of what controls these two effects. The force for continued extension is reduced due to advective thinning by roughly 2"CHb2. For reasonable friction coefficients
and hydrostatic pore pressures in the crust the brittle yield constant C 14 103 Pa m1. For a brittle layer thickness Hb ¼ 20 km and a strain " ¼ 10%, this scaling estimate is for weakening of 5 1011 Nt m1. As discussed in the ‘local isostasy’ section, the local isostatic gravitational force change for crustal stretching is 2"(c/m) (m c)gHb2. For crust 50 km thick with density 2800 kg m 3 overlying mantle with a density of 3200 kg m 3 a 10% strain would produce a force increase of 1012 Nt m1. Thus, these simple scalings indicate that the wide rift mode should occur when the continental crust is thicker and hotter than normal. Extension of lithosphere with normal crustal thickness and average heat flow gives a narrow rift because the advective weakening dominates. The main question about the crustal buoyancy mechanism for wide rifting concerns the hot lower crust. The mechanism should only produce wide rifting if the lower crust has to be weak, but not too weak. To understand this we need to consider how flow of the lower crust will tend to fill in any area of crustal thinning. Crustal thinning lowers pressures in the crust where the crust is thinned. Surrounding crust will pour into this region if any part of the crust is weak enough to flow. The lower crust should
362
The Dynamics of Continental Breakup and Extension
be the hottest part of the crust and so, if there is little or no compositional stratification of the crust, the lower crust should have the lowest effective viscosity. Hot, low-viscosity lower crust may flow into the region of crustal stretching; thus, reducing the crustal thickness variations needed for buoyancy resistance to continued extension. If the lower crust is extremely hot and weak, then this model predicts a ‘core complex mode’ of extension, in which the locus of upper-crustal extension remains fixed in one place while the lower crust thins over a broad region. To more quantitatively access this idea for wide rifting Buck (1991) combined simple calculations of four of the processes discussed in the earlier section on rift processes: (1) local isostatic crustal buoyancy, (2) lower crustal flow, (3) thermal advection, and (4) thermal diffusion. Extension was assumed to be concentrated in a single region of pure shear stretching and the width of the region equaled the lithospheric thickness. The strain rate of stretching was fixed and changes in the vertical thermal structure of the straining region were computed using a 1-D heat equation. Changes of crustal thickness and buoyancy were calculated using a 1-Dl thin channel equation to describe advective crustal thinning and lower-crustal flow. After a small amount of extension the change in force was computed. If the force decreased the extension was considered to remain in the initial location while if the force increased the extension was taken to migrate. Migration of extension would eventually produce a wide rift. Because of the simplifying approximations of the model, millions of parameter combinations could be tested using modest amounts of computer time. Thus, maps of the wide rifting mode could be plotted in terms of crustal thickness and model heat flow for different crustal rheologies and assumed extension rates (Figure 29). The model predicts that wide rifting could result from crustal buoyancy for reasonable rheologies. The range of conditions for model wide rifts is broadly consistent with observations (Figure 30). The possibility that wide rifting cold be controlled by crustal buoyancy has been suggested by more complete thin-layer calculations (Hopper and Buck, 1996) and by fully 2-D thermo-mechanical calculations (e.g., Christensen, 1992). For all these approaches the wide rift and the ‘core complex’ modes are only predicted for thick, hot crust. As noted in Section 6.08.1 the main place where crust is thickened and heated is in orogens; thus, it is not surprising that most wide regions of rifting like the Basin and Range Province and the Aegean occur in regions where the crust was orogenically
thickened. Indeed, it appears that extension driven by collapse of high mountains can occur even while convergence is still occurring across those mountains (see Dewey, 1988; Willet and Pope, 2004). Wide rifts may just be the result of extension of particularly wide orogenic belts where the crust has become fairly uniformly thick and hot as for the Tibetan Plateau (Royden et al., 1997).
6.08.6 Dikes versus Stretching to Initiate Rifting Areas of initially thin lithosphere should rift at very low levels of tectonic force. Further, some areas of high heat flow and initially thick crust, such as the North American Basin and Range Province, start extending with little or no basaltic volcanism. Models neglecting magmatism do predict the general patterns of observed extensional strain inferred for such areas. It should be noted that these ‘hot’ weak areas are not typical, but may require some kind of preheating. The effect of orogenesis, especially thickening of radiogenic crust, has been implicated as a way of heating regions such as the Basin and Range and the Aegean Sea extensional provinces (e.g., Sonder et al., 1987). However, in areas of low-tonormal heat flow, the earliest phase of rifting is often accompanied by basaltic magmatism. Though it has long been known that many rifts are associated with volcanism (Figure 31), a fundamental change in our understanding of rifted margins has been driven by new observations concerning the volume and flux of magmatism associated with those margins. There has been a gradual realization that most rifted margins are temporally and spatially associated with very large magma input. Seismic data has been one of the key inputs that have lead to a change from emphasis from passive lithospheric stretching to active magma-assisted rifting. It was the seismic imaging of ‘seaward dipping reflectors’ along many margins that led to increased interest in the role of magma during rifting. Many rifts and rifted margins are associated with large igneous provinces, but the igneous provice, where there are thick layers of subareal extruded basalt account for a small fraction of the length of the rift (Figure 32). Seaward-dipping reflectors are seismically bright layers that dip ocean-ward on many rifted margins such as parts of the South Atlantic Margin (Hinz, 1981) and the North Atlantic Greenland Margin (e.g., Holbrook et al., 2001). Drilling into the units has
The Dynamics of Continental Breakup and Extension
363
Qs = 100 mW m–2
Core complex mode
°C 0
MPa 1200 0
0
Lithosphere Asthenosphere
Depth (km)
Temperature
log(Pa s) 600 17
23 Viscosity
Yield strength
20
40
Crust 60 Mantle
Qs = 80 mW m–2 °C
Wide rift mode 0
0
MPa 1200 0
Depth (km)
Temperature
Crust Mantle
log(Pa s) 600 17
Yield strength
23 Viscosity
20
40
Lithosphere 60 Asthenosphere
Qs = 60 mW m–2 °C
Narrow rift mode 0
0
MPa 1200 0
Temperature
log(Pa s) 600 17
Yield strength
23 Viscosity
20
Depth (km)
Crust Mantle
40
60 Lithosphere
80
Asthenosphere
Straining region
40 km
V.E. = 2
Figure 29 Illustration of model results predicting different forms of extension depending on the initial lithospheric thermal structure and crustal thickness. Both the core complex model and the wide rift mode are cases where warm crust is driven by gravity (local isostatic forces) to deform over a region that may be wider than the lithospheric thickness. From Buck WR (1991) Modes of continental lithospheric extension. Journal of Geophysical Research, B, Solid Earth and Planets 96(12): 20161–20178.
confirmed that these are basalt flows (Larsen et al., 1993). These dipping relectors are thought to form the same way tilted basalt flows form on Iceland where flows from a rift center load and depress the surface isostatically. As the rift widens the center of magmatic fissure eruptions moves away from the marging, resulting in the kinds of pattern schematically shown in
Figure 33. The imaging of seaward-dipping reflectors and deeper high-velocity units that may be mafic intrusives (see Figure 5) indicates that many more margins may be considered to be ‘volcanic’ than was previously thought. It is estimated that many millions of cubic kilometers of basalt is intruded and extruded along some rift systems (e.g., Holbrook et al., 2001).
364
The Dynamics of Continental Breakup and Extension
140 50° E
1000 km
Core complex
Wide rift
R
25° N
100 SRG
NB and R
ALT
TIB
R
80
ea
40 20
RGB
Narrow rift NRS
30
Arabia
H
ds
60
Re
Heat flow Qs (mW m–2)
30° E
120
Ad Darb Fault
EAF and BAK
40
50
60
Y
70
A
Crustal thickness (km)
Core complex analogs: ALT Altiplano TIB Tibet
E
Gulf
den
of A
NB and R Northern Basin and Range Narrow rifts: NRS Northern Red Sea SRG Southern Rio Grande RGB Rhinegraben EAF East African Rift BAK Baikal
There are several models for the origin of magma that affects volcanic margins. A key controversy is whether the magmatism caused the rifting or vice versa. The idea that rifting causes volcanism comes from considering the way lithospheric stretching advects hot mantle upward and so can lead to melting of that passively flowing mantle (e.g., White and McKenzie, 1989). However, several lines of evidence support the idea of some kind of active mantle upwelling to trigger rifting. First, volcanism is often spread over regions that do not include apparently significant rifting and therefore stretching (Figures 34 and 35). Dikes appear to radiate from a central location that may be the locus of plume upwelling (Figures 35 and 36). Also, the volcanism seems to predate rifting in many rifts (Figure 37). Rifts and passive margins associated with large igneous provinces tend to be nearly straight. No rift is perfectly straight, but if a great circle can pass through part of every segment of a rift, then that rift could be said to be straight. For many rifts, like the Aden Rift, the small-scale segments of the rift do not parallel the coastline or the border faults of the rift
Figure 31 Map of the distribution of the Ethiopian flood basalts and related dikes along the Red Sea. From Ernst RE and Buchan KL (1997) Giant radiating dyke swarms; their use in identifying pre-Mesozoic large igneous provinces and mantle plumes. In: Mahoney JJ and Coffin MF (eds.) Geophysical Monograph: Large Igneous Provinces; Continental, Oceanic, and Planetary Flood Volcanism, pp. 297–333. Washington, DC: American Geophysical Union.
Landward
2
Seaward
3
4
5 180
200 Distance (km)
Two-way travel time (s)
Figure 30 Predicted modes of extension as a function of crustal thickness and surface heat flow compared to data for regions showing those modes of extension, or in the case of Altiplano and Tibet having the conditions that might lead to core complex formation. From Buck WR (1991) Modes of continental lithospheric extension. Journal of Geophysical Research, B, Solid Earth and Planets 96(12): 20161–20178.
5° N
Africa
220
Figure 32 Example of a multichannel seismic profile with arrow showing the prominent seaward-dipping reflectors along the East Greenland Margin. From Hopper JR, et al. (2003) Structure of the SE Greenland margin from seismic reflection and refraction data; implications for nascent spreading center subsidence and asymmetric crustal accretion during North Atlantic opening. Journal of Geophysical Research, B, Solid Earth and Planets 108: 22.
The Dynamics of Continental Breakup and Extension
365
Spreading center B
WG
NEG BTIP
SEG
Flows Plate
Plate
Lithosphere
Labrador Sea opening
Phase 1
Feeder dikes
(62–58 My)
Figure 33 Illustration of the isostatic model of formation of seaward-dipping reflectors as basalt flows coming out of a rift spreading center. After Palmason G (1980) A continuum model of crustal generation in Iceland: kinematic aspects. Journal of Geophysical Research 47: 7–18.
Mantle plume
Hot mantle plume Mainly continent-based magmatism at widespread localities Picrites and high-Mg basalts common
V G SEG HB Lithosphere
Baffin Is. Phase 2
Gre
(56 My on)
enl
Disko Is.
and
Figure 35 Illustration of the idea that plumes may have fed magma to West and East Greenland during the formation of that large igneous province and those rifted margins. From Saunders AD, et al. (1997) The North Atlantic igneous province. In: Mahoney JJ and Coffin MF (eds.) Large Igneous Provinces; Continental, Oceanic, and Planetary Flood Volcanism, vol. 100, pp. 45–93. Washington, DC: American Geophysical Union.
WG EG
Rockall
1000 km
Mantle plume
B e Europ
Figure 34 Distribution of basalts and dikes for East Greenland (EG), West Greenland (WG), and the British Tertiary Basalt Provinve (B). From Ernst RE and Buchan KL (1997) Giant radiating dyke swarms; their use in identifying pre-Mesozoic large igneous provinces and mantle plumes. In: Mahoney JJ and Coffin MF (eds.) Geophysical Monograph: Large Igneous Provinces; Continental, Oceanic, and Planetary Flood Volcanism, pp. 297–333. Washington, DC: American Geophysical Union.
(e.g., Leroy et al., 2004). This is not surprising since the stress orientation and opening direction can change after the initial phase of opening. The Red Sea is the clearest example of a straight rift with the coastlines showing only slight deviations from a great circle over 2000 km and the axis of spreading and rifting is even straighter than the coast. The South Atlantic Margins of South America and Africa, associated in space and time with the Parana Flood Basalts
(Hinz, 1981), are another very straight margin. The dike intruded and volcanic covered North Atlantic Margins of Greenland (e.g., Holbrook et al., 2001) and the conjugate European margins are also remarkably straight on a scale of thousands of kilometers. A magmatic rift does not have to be volcanic. The Northern Red Sea and the Gulf of Suez lack evidence of massive syn-rift volcanism (forming manykilometer-thick seaward-dipping seismic reflector packages) that have been documented along much of the South Atlantic and Greenland Margins (Hinz, 1981; Mutter and Zehnder, 1988). However, synrift dikes striking parallel to the rift are seen as far north as the Gulf of Suez and valley filling, synrift volcanics are common there as well (Patton et al., 1994). In this section simple estimates of the effect of copious magma on the force to rift will be made and compared to the estimated force available for rifting.
6.08.6.1
Force Available for Driving Rifting
Many authors have discussed ways that plate tectonic and plume-related processes could produce relative tension at a rift (e.g., Forsyth and Uyeda, 1975;
366
The Dynamics of Continental Breakup and Extension
150 km depth gives 5 1012 Nt m1 of rift driving force. Higher elevations and deeper compensation depths are possible, but the force may be spread over a longer rift than the width of the uplifted area. Thus, the average level of rift force is likely to be less than about 5 1012 Nt m1.
1. Dike intrusion and rifting
6.08.6.2 Volcanic rocks
2. Spreading
f el
Passive margin dikes
Sh
Sh el f
Failed arm dikes
Figure 36 Illustration of how giant radiating dike swarms are often seen to relate to the geometry of subsequent rifting. From Ernst RE and Buchan KL (1997) Giant radiating dyke swarms; their use in identifying pre-Mesozoic large igneous provinces and mantle plumes. In: Mahoney JJ and Coffin MF (eds.) Geophysical Monograph: Large Igneous Provinces; Continental, Oceanic, and Planetary Flood Volcanism, pp. 297–333. Washington, DC: American Geophysical Union.
Solomon et al., 1980), and these approaches give similar estimates of rift-driving forces. The simplest of these approaches relates to the uplift that may occur over a region of abnormally hot mantle. The radiation of rift branches away from the uplifted plateau areas of Ethiopia and Yemen are consistent with the extensional driving force being related to the uplift of that region (e.g., Sengor and Burke, 1978; Ernst and Buchan, 1997). For uplift due to a low-density root, the extensional force scales with the magnitude of the uplift, e, and the depth of the low-density compensation layer, d (Spohn and Schubert, 1982; Bott, 1991). The force is roughly emgd, where m is the mantle density and g is the acceleration of gravity. For a root of uniform density, hot mantle between 100 and 200 km, d would be 150 km. Then 1 km of uplift compensated at
Force Needed for Tectonic Rifting
Recall from Section 6.08.2 of this chapter that the tectonic yield strength depends on the thermal structure of the lithosphere and to a lesser extent on the composition of the lithosphere. To compare the force for tectonic extension to that required for magmatic extension it is sufficient to ignore the possibility that crust is weaker in terms of ductile flow than mantle. To the extent that we can ignore the contribution of ductile yield stress to the lithospheric strength and if we take the density of the lithosphere constant with depth we can get a simple estimate of the tectonic force for rifting: FT ¼
CHb2 2
½22
where Hb is the thickness of the brittle lithosphere and C is a constant defined in eqns [2] and [3], and is approximately 1.4 104 Pa m1. Neglecting hydrostatic pore pressures increases the value of C, and so the estimated force, by about 50%, and taking a mantle density for the lithospheric density increases it a further 10%. Assuming the creeping part of the lithosphere contributes to the tectonic force also increases this estimate, but for crust creeping more easily than the mantle the tectonic force is reduced by an amount that depends on the crustal thickness, rheological constants, and thermal state (e.g., Brace and Kohlstedt, 1980; Sawyer, 1985; Kusznir and Park, 1987). For the purpose of making simple comparisons with the force needed to open a magmatic rift in this chapter eqn [4] is sufficient. 6.08.6.3 Rifting
Force Needed for Magmatic
As long as the magma is less dense than the rock, it intrudes then some extensional stress has to be applied to keep the magma from rising to the surface and extruding. Only when the magma is ‘kept down’ by this stress difference will the dike open at depth and so allow plates to move apart (Figure 38). Neglecting any stress needed to open a crack, as
The Dynamics of Continental Breakup and Extension
(a)
367
Siberian Traps
Greenland BTP
Central Atlantic Margin
Deccan Traps NW African Dikes
Emeishan Basalts
Etendeka Traps
Parana Traps
Karoo Traps
Farrar Volcanics
(b)
Trap
Stratoid OMA ?
? Ethiopia 30
40
SDR
20
?
10
2
0
OMA
Greenland 70
60
55 53
?
50
40
30
OMA
Deccan 70
65
60
? Parana
140 138
50
40
OMA 132
127/5
120
110
100
OMA ? Karoo 190
183
180
170
150
160 OMA ?
Central Atl. 210
200
190
180
175?
170
Figure 37 (a) The distribution of large igneous provinces over the last 300 My. (b) shows the temporal relation between the major (trap) phase of vocanism in these provinces compared to the onset of ocean-floor spreading (as seen by oceanic magnetic anomalies, OMA). From Courtillot et al. (1999).
discussed earlier, the minimum stress difference needed to open a magma filled dike is M ðzÞ ¼
Z
Hb
g ðs ðzÞ – f Þ dz
½23
0
Equation [23] is valid only when the fluid magma density is less than or equal to the average
lithospheric density. For example, for constant lithospheric density that is less than the magma density an extensional force would be required to pull dense magma up from the asthenosphere and magma could not rise to the surface and the limits of integration would have to be changed. Such cases are not likely to be important for lithosphere thicker than a
368
The Dynamics of Continental Breakup and Extension
(a)
Minimum stress for opening dike
30
Dike filled with magma of density ρ f Stress σv = g ∫ρs(z) dz σh = Pf
ρs = ρm Mantle lithosphere
σ h = Pf = ρ f gz
20
Tectonic
15 Magmatic Hc = 30 km
10 Estimated rift push
5
Figure 38 The stress distribution for extensional separation of two lithospheric blocks by a vertical magmatic intrusion (a dike). Here the solid lithosphere has a density, s, that is greater than the fluid magma density, f. The yellow region represents crust with a density, c, close to that of basaltic magma in dikes and the green represents mantle lithosphere with a density, m, greater than that of the crust. The horizontal stress, h, equals the pressure in the dike, Pf, while the vertical stress, v, equals the overburden pressure. The area between these two stresses (highlighted in blue) equals the force that must be applied to keep the dike open. From Buck WR (2006) The role of magma in the development of the AfroArabian rift system. In: Yirgu G, Ebinger CJ, and Maguire PKH (eds.) The Afar Volcanic Province within the East African Rift System, vol. 259, pp. 43–54. London: Geological Society.
0 0 (b)
10 20 30 40 50 60 70 80 90 100 Lithospheric thickness, H L(km)
15 Magmatic cases Force (TeraNt m–1)
σh
Depth
σv
Force (TeraNt m–1)
25
ρs = ρc
Crust
10
Hc = 5 km 30 km Estimated rift push
5
55 km
0 0
few kilometers and for basaltic magma densities, so they are not discussed further. The extensional tectonic force to open a magmafilled dike through denser lithosphere with a thickness hl is FM ¼
Z
Hb
M ðzÞ dz
½24
0
Consider the simple case that the densities of the crust and mantle, c and m, respectively, are constant with depth. If the entire crust is brittle down to its base at z ¼ Hc and the thickness of mantle lithosphere is Hm (so the total lithospheric thickness Hb ¼ Hc þ Hm), then Hc2 þ g ð c – f ÞHc 2 Hm þ ðm – f Þ Hm 2
10 20 30 40 50 60 70 80 90 100 Lithospheric thickness, H L(km)
Figure 39 Illustration of estimated force required for lithospheric extension for either tectonic or magmatic rifting as a function of lithospheric thickness. Equation [22] was used to compute the tectonic force and eqn [7] was used to compute the magmatic force. (a) shows the forces as a function of lithospheric thickness assuming a 30 km thick crust. The dashed line is an estimate of the force available to drive lithospheric extension. (b) shows how the crustal thickness affects the estimated magmatic force. It is the mantle lithospheric thickness, which equals the total lithospheric thickness minus the crustal thickness, that controls this force. From Buck WR (2006) The role of magma in the development of the Afro-Arabian rift system. In: Yirgu G, Ebinger CJ, and Maguire PKH (eds.) The Afar Volcanic Province within the East African Rift System, vol. 259, pp. 43–54. London: Geological Society.
FM ¼ g ð c – f Þ
½25
The density of continental crust is not likely to be very different from the density of basaltic magma. If that is true then it takes no force to open a dike in the crust, but it still takes considerable force to open a dike into the mantle lithosphere. For c ¼ f we get
the further simplification that gives eqn [7] (given in an earlier section). For reasonable density values we plot this estimate of the force to dike through the entire lithosphere versus the lithospheric thickness (Figure 39) and we compare this to the quadratic relation between tectonic force and lithospheric thickness for tectonic stretching (eqn [22]). For a density contrast between
The Dynamics of Continental Breakup and Extension
solid mantle and fluid magma of 500 kg m 3 (based on mantle density of 3250 kg m 3 and magma density of 2750 kg m 3) it would take 4 1012 Nt m1 to dike trough a 40 km thick lithospheric mantle layer. To dike through 100 km of mantle lithosphere would take 2.5 1013 Nt m1 and this is considerably more force than is likely to be available to drive rifting. The term magmatic rifting will be used to describe lithospheric extension that is aided by dike intrusion and
369
the force for magmatic rifting is taken to be the force to open dikes through the lithosphere. Dikes may not open through the entire thickness of the lithosphere if insufficient magma is available, and for such cases the force required to rift would be intermediate between the force for tectonic rifting and the force for magmatic rifting. Figure 40 shows how different these forces are likely to be as a function of lithospheric thickness. The tectonic force
Tectonic stretching (a)
°C
Normal lithosphere
0
MPa 1000 0
800
0
Yield stress
Temp 20
Depth (km)
Crust Mantle
40
60
Lithosphere 80
Asthenosphere
(b)
°C
Orogenic lithosphere 0
0
MPa 1000
0
Temp
20
Depth (km)
Crust Mantle
800
Yield stress
40
60
Lithosphere 80
Asthenosphere
(c)
Magmatic extension
°C 0
0
MPa 1000
Temp 20
Mantle
Depth (km)
Crust Dike
0
800
Yield stress
40
60
Lithosphere 80
Asthenosphere Straining region
V. E. = 2. 40 km
Figure 40 Schematic of stresses needed for extension of continental lithosphere. Note the large difference in the yield stress, the stress difference needed to get extensional separation of two lithospheric blocks. Tectonic extension of (a) normal continental lithosphere should require very large yield stresses and correspondingly large tectonic extensional forces. (b) Orogenically lithosphere should have a higher geotherm and much lower lithospheric strength. (c) Magmatic intrusion may allow extension of normal lithoshere at modest tectonic force levels.
370
The Dynamics of Continental Breakup and Extension
depends on the square of the whole thickness of the lithosphere (times 6000 Nt m3), while the magmatic force depends on the square of the mantle lithospheric thickness (times 2500 Nt m3). For reasonable driving force levels, only lithosphere thinner than 30 km thick should rift tectonically (i.e., in the absence of magmatic dike intrusion). For a normal continental crustal thickness of 40 km, the base of the lithosphere could be as deep as 80 km and still allow magmatic rifting at reasonable force levels. Figure 40 shows that lithosphere with thicker continental crust should rift magmatically with much less force than for lithosphere with thin crust. Rifts intruded by large volumes of basaltic magma should subside little during the intrusion phase of extension. If sufficient dike intrusion occurs to heat the lithosphere appreciably then its strength could be lowered to the point where tectonic extentions can proceed at moderate stress levels. Such combined magmatic and tectonic extension could give a rifted margin with less tectonic subsidence and more thermal subsidence than predicted by a McKenzie stretching model. As noted earlier this is the pattern of subsidence most commonly seen at rifted margins. 6.08.6.4
The Meaning of Rift Straightness
The straightness of many rifts may indicate that they were initiated by magmatic dike intrusions. Lithospheric extension can be accomplished by either dike opening or fault slip. Both dikes and faults form in response to stress difference in the lithosphere. Faults form and slip where the shear stress is great enough to overcome the strength of the material and allow brittle deformation. Faults change orientation where the stress orientations change or where there are strength variations. So, rifts where faults accommodate the extension do not have to be straight. They can curve where the stresses change or ‘side-step’ into weaker areas. Dikes are narrow magma filled tension cracks (e.g., Lister and Kerr, 1991) and as such the plane of the dike must be orthogonal to the least principal stress in a strong, brittle layer. When the least principal stress is horizontal the dike is vertical and these are the kinds of dikes discussed here. The key to the straightness of dikes may be that dikes, unlike faults, need a source of magma and a connection to that magma source. The open part of the dike is the conduit connecting the source area to the tip of the dike. If dikes are fed from a distributed magma source
below a brittle layer then the dikes do not have to be straight except on the scale of the layer thickness. One way to produce dikes that are straight over lateral distances longer than the brittle layer thickness is if the magma source is localized. The dikes can only remain connected to the source if they are straight. If the magma does come from a central source then the dike cannot propagate if it loses connection to the magma supply because the connection is the open, unfrozen, straight dike behind the propagating tip. If a dike tip were to step laterally away from the plane of the open dike by more than the width of the dike it would no longer be fed magma. Dikes exposed in ophiolites and bordering rifts are typically about a meter wide (e.g., Varga, 2003), so very small offsets are viable. Straight dikes have been observed propagating from a central magma chamber along a subaerial segment of the Mid-Atlantic Ridge in Iceland. In 1975 an episode of approximately 15 dike intrusion and magma extrusion events began, with the longest dike propagating 70 km from the Krafla central volcano (Tryggvason, 1980). Seismic and geodetic measurements unequivocally show that the dikes are sourced from a central magma chamber that subsided while the dike was propagating (Einarsson, 1991). Dikes on the flanks of active volcanoes, like those propagating down the east rift zone of Kilaeua Volcano in Hawaii, are seen to be straight (e.g., Cervelli et al., 2002). Ancient dikes in the MacKenzie Dike Swarm of Canada are straight and traceable for thousands of kilometers (Fialko and Rubin, 1999). The fact that most mid-ocean ridge segments are straight and nearly orthogonal to the spreading direction may reflect dikes fed from central magma chambers. New data on ridge segments suggests that at the slowest spreading rates there are nonmagmatic segments that are not straight. Dick et al. (2003) note that the slowest spreading centers, such as the Gakkel Ridge in the Arctic and oblique sections of the Southwest Indian Ridge, show an alteration of volcanic segments and nonvolcanic ones. Peridotite samples dredged from the surface of the nonvolcanic segments indicate that the mantle there is stretched with little or no input of magma. When these nonvolcanic segments are along oblique sections of the ridge, the usual pattern of ridge segments orthogonal to transform faults is not seen. Such segments are cut by numerous faults that trend oblique to the spreading direction.
The Dynamics of Continental Breakup and Extension
6.08.6.5 The Distance of Dike/Rift Propagation Dike propagation may be controlled by a dauntingly large number of thermal and mechanical processes. Among them are the pressure and flux of magma coming out of a source region, the viscous resistance to magma flow along the body of the dike and into the tip region, elastic stresses in lithosphere, and the freezing of magma (e.g., Lister and Kerr, 1991; Rubin, 1995). Although the details of dike mechanics are controversial (e.g., Delaney and Pollard, 1982; Fialko and Rubin, 1999; Ida, 1999), there is no doubt that the distance of dike propagation should be related to the supply of magma. The volume of magma intruded into the thick lithosphere along a several-thousand-kilometer-long rift has to be massive. The size of a magma chamber should play a major role in determining how much magma can be supplied to dikes as they propagate. The magma pressure is likely to be reduced as volume is extracted from a magma chamber. Buck et al. (2006) argue that lithosphere-cutting dikes stop when the ‘driving stress’ (defined as the difference between magma pressure and tectonic stress orthogonal to the dike) becomes too small, either because the dike slows and the tip freezes, or because the driving stress is not enough to break open a new section of dike. The larger and shallower the magma chamber the smaller the pressure drop on extraction of a given volume of magma. Large magma chambers should supply large amounts of basalt to a propagating dike and so could be necessary to the production of long dikes. Simple thermal arguments would suggest that the size and depth of a magma chamber should correlate with the flux of magma coming into a region. The greatest known fluxes of magma occur during the geologically short periods when large igneous provinces form. The volumes are up to several million cubic kilometers and the time interval of high rate magma output is 1 My or less (Courtillot and Renne (2003), and references therein). Very large, near-surface magma chambers are likely to form during periods of high melt flux from localized mantle upwellings. Magma chambers the size of large gabbroic layered intrusions found near the centers of some large igneous provinces could feed dikes propagating thousands of kilometers. For example, the Skaergaard layered intrusion of East Greenland is estimated to have a volume of 300 km3 (Nielsen, 2004), sufficient to fill a dike 2000 km long, 50 km high, and 3 m thick.
371
The regions where the mantle lithosphere is very thick may not rift even if copious, high-pressure basaltic magma is present (see Figure 39(b)). Recall that extrusion of magma on the surface limits the maximum magma pressure. Cratonic regions, where the lithosphere may be well over a hundred kilometers thick (Jordan, 1975; Venkataraman et al., 2004), and old oceanic lithosphere, where the mantle lithosphere may be 60 km thick (e.g., Wiens and Stein, 1984) should be too thick to rift. Dikes and associated magmatic rift propagation can stop either because the magma pressure gets too low, or the stress is not sufficient to open a dike through the lithosphere. For the Afro-Arabian Rift System it may be the lack of extensional stress sufficient to open dikes through very thick mantle lithosphere that limits dike propagation. Previous workers have noted that changes in lithospheric strength may have controlled the termination of the Red Sea Branch (e.g., Steckler and ten Brink, 1986) or affected the structure of the East African Branch (e.g. Rosendahl, 1987). Those workers were concerned with the difference in tectonic strength of the lithosphere. It is possible that similar arguments may apply for magma-assisted rifting. It may be the increase in force for magmatic rifting related to mantle lithosphere thickening that limits rift propagation. This is suggested by the observation that the Northern and Southern Branches of the system end close to regions of thick mantle lithosphere. The Northern Red Sea Branch ends close to the Mediterranean Sea where old oceanic lithosphere may be too thick to be cut by dikes.
6.08.7 Conclusions and Future Work The questions discussed here are certainly not settled and new observations and models are clarifying the arguments. Several other major problems of continental extension have not been discussed here. For example, what controls the time interval between major rifting events? What controls the length of rifts? Why are fault patterns at rifts so variable? New observations and improved numerical models should allow us to address these and other rift-related questions. More geological and geophysical data is needed on rift and margin structure and history. The amount and timing of magmatism must be constrained. To understand the fault structures forming margins we need better constraints on actual distribution of faults
372
The Dynamics of Continental Breakup and Extension
and their offset histories. It appears that we are still in the discovery phase of finding out what rifts are like. Continuum numerical models are good for studying some of the problems of large-scale rift structure. New techniques need to be developed to deal with. A great challenge is to combine models of highly localized processes such as fault zone evolution and dike propagation with larger regional scale models of continental extension. On the other end of the spectrum detailed rift models need to be embedded into global-scale numerical models to test ideas about the ways large-scale convective processes may be linked to the formation of continental rifts.
References Abers GA (1991) Possible seismogenic shallow-dipping normal faults in the Woodlark-D’Entrecasteaux extensional province, Papua New Guinea. Geology 19: 1205–1208. Allmendinger RW, Hauge TA, Hanser EC, et al. (1987) Overview of the COCORP 40 N Transect Western United States: The fabric of an orogenic belt. Geological Society of America Bulletin 98: 308–319. Alvarez F, Virieux, and LePichon X (1984) Thermal consequence of lithosphere extension: The initial stretching phase. Geophysical Journal of the Royal Astronomical Society 78: 389–411. Anderson EM (1951) The Dynamics of Faulting and Dyke Formation, with Applications to Britain. London: Oliver and Boyd. Artemjev ME and Artgushkov EV (1971) Structure and isostasy of the Baikal rift and the mechanism of rifting. Journal of Geophysical Research 76: 1197–1211. Audin L, Hebert H, Deplus C, Huchon P, and Khanbari K (2001) Lithospheric structure of a nascent spreading ridge Inferred from gravity data; the Western Gulf of Aden. Journal of Geophysical Research, B, Solid Earth and Planets 106: 26345–26363. Bassi G (1991) Factors controlling the style of continental rifting; insights from numerical modelling. Earth and Planetary Science Letters 105(4): 430–452. Behn MD, Lin J, and Zuber MT (2002) A continuum mechanics model for normal faulting using a strain-rate softening rheology; implications for thermal and rheological controls on continental and oceanic rifting. Earth and Planetary Science Letters 202(3–4): 725–740. Biot MA (1961) Theory of folding of stratified, viscoelastic media and its application in tectonics and orogenesis. Geological Society of America Bulletin 72: 1595–632. Bird P (1991) Lateral extrusion of lower crust from under high topography, in the isostatic limit. Journal of Geophysical Research, B, Solid Earth and Planets 96(6): 10275–10286. Block L and Royden LH (1990) Core complex geometries and regional scale flow in the lower crust. Tectonics 9: 557–567. Bosworth W (1987) Off-axis volcanism in the Gregory rift, east Africa: Implications for models of continental rifting. Geology 15: 397–400. Bott MHP (1991) Ridge push and associated plate interior stress in normal and hot spot regions. Tectonophysics 200: 17–32.
Brace WF and Kohlstedt DL (1980) Limits on lithospheric stress imposed by laboratory experiments. Journal of Geophysical Research 85: 6248–6252. Braun J and Beaumont C (eds.) (1989) Contrasting styles of lithospheric extension: Implications for differences between basin and range province and rifted continental margins. extensional tectonics and stratigraphy of the north Atlantic margins. American Association of Petrololeum Geologist Memoir 46: 53–79. Braun J and Beaumont C (1989) A physical explanation of the relation between flank uplifts and the breakup unconformity at rifted continental margins. Geology 17: 760–764. Brun JP, Sokoutis D, and Van Den Driessche J (1994) Analog modeling of detachment fault systems and core complexes. Geology 22: 319–322. Buck WR (1988) Flexural rotation of normal faults. Tectonics 7: 959–973. Buck WR (1991) Modes of continental lithospheric extension. Journal of Geophysical Research, B, Solid Earth and Planets 96(12): 20161–20178. Buck WR (1993) Effect of lithospheric thickness on the formation of high-and low-angle normal faults. Geology 21: 933–936. Buck WR (2004) Consequences of the asthenospheric variability on continental rifting. In: Karner GD, Taylor B, Driscoll NW, and Kohlstedt DL (eds.) Rheology and Deformation of the Lithosphere at Continental Margins, pp. 1–31. New York: Columbia Univeristy Press. Buck WR (2006) The role of magma in the development of the Afro-Arabian rift system. In: Yirgu G, Ebinger CJ, and Maguire PKH (eds.) The Afar Volcanic Province within the East African Rift System, vol. 259, pp. 43–54. London: Geological Society. Buck WR, Einarsson P, and Brandsdottir B (2006) Tectonic stress and magma chamber size as controls on dike propagation: Constraints from the 1974–1989 Krafla rifting Episode. Journal of Geophysical Research 111: B12404. Buck WR, Lavier LL, and Babeyko A (2003) A numerical model of lithospheric extension producing fault bounded basins and ranges. International Geology Review 45(8): 712–723. Buck WR, Lavier LL, and Poliakov AN (2005) Modes of faulting at mid-ocean ridges. Nature 434: 719–723. Buck WR, Martinez F, Steckler MS, and Cochran JR (1988) Thermal consequences of lithospheric extension: Pure and simple. Tectonics 7: 213–234. Buiter SJH, Babeyko AY, Ellis S, et al. (2006) The numerical sandbox; comparison of model results for a shortening and an extension experiment. In: Buiter SJH and Schreurs (eds.) Geological Society Special Publications: Analogue and Numerical Modelling of Crustal-Scale Processes, vol. 253, pp. 29–64. London: Gological Society. Burov E and Cloetingh S (1997) Erosion and rift dynamics; New thermomechanical aspects of post-rift evolution of extensional basins. Earth and Planetary Science Letters 150(1–2): 7–26. Byerlee JD (1978) Friction of rocks. Pure and Applied Geophysics 116: 615–626. Cervelli P, Segall P, Amelung F, et al. (2002) The 12 september 1999 upper east rift zone dike intrusion at Kilauea Volcano, Hawaii. Journal of Geophysical Research, B, Solid Earth and Planets 107(7): 13. Christensen UR (1992) An Eulerian technique for thermomechanical modeling of lithospheric extension. Journal of Geophysical Research B, Solid Earth and Planets 97(2): 2015–2036. Cochran JR (1983b) Effects of finite rifting times on the development of sedimentary basins. Earth and Planetary Science Letters 66: 289–303.
The Dynamics of Continental Breakup and Extension Cochran JR (1983a) A model for the development of the Red Sea. American Association of Petroleum Geologist Bulletin 67: 41–69. Cochran JR, Martinez F, Steckler MS, and Hobart MA (1986) Conrad deep: A new northern Red Sea deep. Origin and implications for continental rifting. Earth and Planetary Science Letters 78: 18–32. Coney PJ (1980) Cordilleran metamorphic core complexes: An overview. In: Crittenden MD, Jr., Coney PJ, and Davis GH (eds.) Geological Society of America Memoir, Cordilleran Metamorphic Core Complexes, vol. 153, pp. 7–31. Boulder, CO: Geological Society of America. Coney PJ and Harms TA (1984) Cordilleran metamorphic core complexs: Cenozoic extensional relics of Mesozoic compression. Geology 12: 550–554. Corti G, Bonini M, Conticelli S, Innocenti F, Manetti P, and Besse J (2003) Analogue modelling of continental extension: A review focused on the relations between the patterns of deformation and the presence of magma. Earth-Science Reviews 63: 169–247. Courtillot V, Jaupart C, Manighetti I, Tapponnier P, and Sokoutis D (1999) On causal links between flood basalts and continental breakup. Earth and Planetary Science Letters 166: 177–195. Courtillot VE and Renne PR (2003) On the ages of flood basalt events. Comptes Rendus Geoscience 335: 113–140. Davis GA (1980) Problems of intraplate extensional tectonics, Western United States. In: Continental Tectonics, pp. 84–95 Washington, DC: National Academy of Science. Davis GA and Lister GA (1988) Detachment faulting incontinental extension; perspective from the southwestern U.S. Cordillera. In: Clark SP, Burchfield BC, and Suppe J (eds.) Geological Society of America. Special Paper, vol. 218: Processes Incontinental Lithospheric Deformation, pp. 133–159. Boulder, CO: Geological Society of America. Davis GH (1983) Shear zone model for the origin of metamorphic core complexes. Geology 11: 342–347. Davis M and Kusznir NJ (2004) Depth dependent extension at rifted margins. In: Karner GD, Taylor B, and Driscoll NW (eds.) Rheology and Deformation of the Lithosphere at Continental Margins, pp. 92–136. New York: Columbia Univeristy Press. de Charpel O, Guennoc P, Montadert, and Land Roberts DG (1978) Rifting, crustal attenuation and subsidence in the Bay of Biscay. Nature 275: 706–711. Delaney PT and Pollard DD (1982) Solidification of basaltim magma during flow in dike. American Journal of Science 282: 856–885. Dewey JF (1988) Extensional collapse of orogens. Tectonics 7(6): 1123–1139. Dick HJB, Lin J, and Schouten H (2003) An ultraslow-spreading class of ocean ridge. Nature 426(6965): 405–412. Driscoll NW and Karner GD (1998) Lower crustal extension across the Northern Carnarvon basin, Australia: Evidence for an eastward dipping detachment. Journal of Geophysical Research 103(B3): 4975–4991. Dunbar JA and Sawyer DS (1989) How preexisting weaknesses control the style of continental breakup. Journal of Geophysical Research, B, Solid Earth and Planets 94(6): 7278–7292. Ebinger CJ, Deino AL, Drake RE, and Tesha AL (1989) Chronology of volcanism and rift basin propagation: Rungwe volcanic province, East Africa. Journal of Geophysical Research 94: 15785–15803. Ebinger CJ, Rosendahl BR, and Reynolds DJ (1987) Tectonic model of the Malawi rift, Africa. Tectonophysics 141: 215–235.
373
Einarsson P (ed.) (1991) The Krafla rifting episode 1975–1989. In: Gardarsson A and Einarsson A´ (eds.) Na´ttu´ra M’yvatns, (The Nature of lake M’yvatn), pp. 97–139. Reykjavı´c: Icelandic Nature Science Society. England P (1983) Constraints on extension of continental lithosphere. Journal of Geophysical Research 88: 1145–1152. Ernst RE and Buchan KL (1997) Giant radiating dyke swarms; their use in identifying pre-Mesozoic large igneous provinces and mantle plumes. In: Mahoney JJ and Coffin MF (eds.) Geophysical Monograph: Large Igneous Provinces; Continental, Oceanic, and Planetary Flood Volcanism, pp. 297–333. Washington, DC: American Geophysical Union. Fialko YA and Rubin AM (1999) Thermal and mechanical aspects of magma emplacement in giant dike swarms. Journal of Geophysical Research 104(B10): 23033–23049. Fleitout L and Froidevaux C (1983) Tectonic stresses in the lithosphere. Tectonics 2(3): 315–324. Fletcher RC and Hallet B (1983) Unstable extension of the lithosphere: A mechanical model for Basin and Range structure. Journal of Geophysical Research 88: 7457–7466. Forsyth DW (1992) Finite extension and low-angle normal faulting. Geology 20: 27–30. Forsyth DW and Uyeda S (1975) On the relative importance of the driving forces of plate motion. Geophysical Journal of the Royal Astronomical Society 43(1): 163–200. Gans PB (1987) A open-system, two-layer crustal stretching model for the eastern Great Basin. Tectonics 6: 1–12. Geoffroy L (2005) Volcanic passive margins. Geoscience 337: 16. Goetze C and Evans B (1979) Stress and temperature in the bending lithospheree us constrained by experimental rock mechanics. Geophysical Journal of the Royal Astronomical Society 59: 463–478. Hamilton W (1988) Extensional faulting in the death valley region (abstract). Geological Society of of America Abstracts with Programs 20: 165–166. Hamilton WB (ed.) (1987) Crustal extension in the basin and range province, Southwestern United States. In: Coward MP, Dewey JF, and Hancock PL (eds.) Geological Society Speciaa Publication: Continental Extensional Tectonics, pp. 155–176. Oxford: Blackwell Science. Handin J (1966) Strength and ductility. In: Clark SP, Jr. (ed.) Geological Society of America Memoir 97: Handbook of Physical Constraints, pp. 223–290. Boulder, CO: Geological Society of America. Hinz K (1981) A hypothesis on terrestrial catastrophes – Wedges of very thick oceanward dipping layers beneath passive margins. Geologische Jahrbuch Reihe E 22: 3–28. Holbrook WS, Larsen HC, Korenaga J, et al. (2001) Mantle thermal structure and active upwelling during continental breakup in the North Atlantic. Earth and Planetary Science Letters 190: 251–266. Hopper J and Buck WR (1996) Effects of lower crustal flow on continental extension and passive margin formation. Journal of Geophysical Research 101: 20175–20194. Hopper JR and Buck WR (1993) The initiation of rifting at constant tectonic force: The role of diffusion creep. Journal of Geophysical Research 98(16): 16213–16221. Hopper JR, Dahl JT, Holbrook WS, et al. (2003) Structure of the SE Greenland margin from seismic reflection and refraction data; implications for nascent spreading center subsidence and asymmetric crustal accretion during North Atlantic opening. Journal of Geophysical Research, B, Solid Earth and Planets 108: 22. Huismans RS and Beaumont C (2003) Symmetric and asymmetric lithospheric extension; relative effects of frictional-plastic and viscous strain softening. Journal of
374
The Dynamics of Continental Breakup and Extension
Geophysical Research, B, Solid Earth and Planets 108(10): 22. Huismans RS, Buiter SJH, and Beaumont C (2005) Effect of plastic-viscous layering and strain softening on mode selection during clothes extension. Journal of Geophysical Research 110(B2): 17. Ida Y (1999) Effects of the crustal stress on the growth of dikes: Conditions of intrusion and extrusion of magma. Journal of Geophysical Research 104(B8): 17897–17910. Illies JH and Greiner G (1978) Rhinegraben and the Alpine system. Geological Society of America Bulletin 89(5): 770–782. Jackson JA (1987) Active normal faulting and crustal extension. In: Coward MP, Dewey JF, and Hancock PL (eds.) Continental Extensional Tectonics, vol. 28, pp. 3–17. Oxford: Blackwell Scientific. Jackson JA and White NJ (1989) Normal faulting in the upper continental crust: Observations from regions of active extension. Journal of Structural Geology 11: 15–36. Jarvis GT and McKenzie DP (1980) Sedimentary basin formation with finite extension rates. Earth and Planetary Science Letters 48: 42–52. John BJ and Foster DA (1993) Structural and thermal constraints on the initiation angle of detachment faulting in the southern Basin and Range: The Chemehuevi Mountains case study. Geological Society of America Bulletin 105: 1091–1108. Jordan TH (1975) The continental tectosphere. Reviews of Geophysics and Space Physics 13(3): 1–12. King G and Ellis M (1990) The origin of large local uplift in extensional regions. Nature 348: 689–692. Kirby SH and Kronenberg AK (1987) Rheology of the lithosphere: selected topics. Reviews of Geophysics 25: 1219–1244. Kohlstedt DL, Evans B, and Mackwell SJ (1995) Strength of the lithosphere: Constraints imposed by laboratory experiments. Journal of Geophysical Research 100: 17587–17602. Kusznir NJ and Park RG (1987) The extensional strength of the continental lithosphere: Its dependence on geothermal gradient, and crustal composition and thickness. In: Coward MP, Dewey JF, and Hancock PL (eds.) Geological Society of London. Special Publications, vol. 28: Continental Extensional Tectonics, pp. 35–52. London: Geological Society. Langseth MG and Taylor PL (1967) Recent heat flow measurements in the Indian Ocean. Journal of Geophysical Research 85: 3740–3750. Larsen HC, Saunders AD, and Clift PD (1993) Preliminary results from drilling on the SE Greenland margin; ODP Leg 152. AGU 1993 fall meeting, American Geophysical Union. 74: (Supplement): 606. Lavier L and Buck WR (2002) Half graben versus large-offset lowangle normal fault: The importance of keeping cool during normal faulting. Journal of Geophysical Research 107: 2122. Lavier L, Buck WR, and Poliakov ANB (1999) Self-consistent rolling-hinge model for the evolution of large-offset lowangle normal faults. Geology 27(12): 1127–1130. Lavier L, Buck WR, and Poliakov ANB (2000) Factors controlling normal fault offset in an ideal brittle layer. Journal of Geophysical Research 105(B10): 23431–23442. Lavier LL and Manatschal G (2006) A mechanism to thin the continental lithosphere at magma-poor margins. Nature 440(7082): 324–328. LePichon X and Alvarez F (1984) From stretching to subduction in back-arc regions: Dynamic considerations. Tectonophysics 102: 343–357. LePichon X and Barbier F (1987) Passive margin formation by low-angle faulting within the upper crust: the northern Bay of Biscay margin. Tectonics 6: 133–150. Leroy S, Gente P, Fournier M, et al. (2004) From Rifting to Spreading in the Eastern Gulf of Aden; a geophysical survey
of young oceanic basin from margin to margin. Terra Nova 184–192. Lister GS and Baldwin SL (1993) Plutonism and the origin of metamorphic core complexes. Geology 21: 607–610. Lister GS and Davis GA (1989) Models for the formation of metamorphic core complexes and mylonitic detachment terranes. Journal of Structural Geology 11: 65–94. Lister GS, Etheridge MA, and Symonds PA (1986) Detachment faulting and the evolution of passive continetal margins. Geology 14: 246–250. Lister JR and Kerr RC (1991) Fluid-mechanical models of crack propagation and their application to magma transport in dykes. Journal of Geophysical Research 96: 10049–10077. Maguire PHK, Keller GR, and Klemperer SL (2006) Crustal structure of the northern Main Ethiopian Rift from the EAGLE controlled-source survey: A snapshot of incipient lithospheric breakup. In: Yirgu G, Ebinger CJ, and Maguire PKH (eds.) The Afar Volcanic Province within the East African Rift Systems, vol. 259, pp. 269–291. London: Geological Society. Mackwell SJ, Zimmerman ME, and Kohlstedt DL (1998) Hightemperature deformation of dry diabase with application to tectonics on Venus. Journal of Geophysical Research 103: 975–984. Manspeizer W and Cousminer HL (1988) Late Triassic–Early Jurassic synrift basins of the U.S. Atlantic margin. In: Sheridan RE and Grow JA (eds.) The Geology of North America: The Atlantic Continental Margin; U.S., pp. 197–216. Boulder, CO: Geological Society of America. Martinez F and Cochran JR (1988) Structure and tectonics of the northern Red Sea: Catching a continental margin between rifting and drifting. Tectonophys 150: 1–32. Martinod J and Davy P (1992) Periodic instabilities during compression or extension of the lithosphere 1. Deformation modes from an analytical perturbation method. Journal of Geophysical Research 97: 1999–2014. McKenzie DP (1978) Some remarks on the development of sedimentary basins. EPSL 40: 25–32. McKenzie DP (1967) Some remarks on heat flow and gravity anomalies. Journal of Geophysical Research 72(24): 6261–6273. Melosh HJ and Williams CA (1989) Mechanics of graben formation in crustal rocks: A finite element analysis. Journal of Geophysical Research 94(B10): 13961–13973. Miller JMG and John BE (1988) Detachment strata in a tertiary low-angle normal fault terrane, southeastern California: A sedimentary record of unroofing, breaching, and continued slip. Geology 19: 645–648. Morgan P, Seager WR, and Golombeck MP (1986) Cenozoic mechanical and tectonic evolution of the Rio Grande Rift. Journal of Geophysical Research 91: 6263–6276. Mutter JC, Talwani M, and Stoffa PL (1982) Origin of seawarddipping reflectors in oceanic crust off the Norwegian margin by ‘‘subariel seafloor spreading’’. Geology 10: 353–357. Mutter JC and Zehnder CM (1988) Deep crustal structure and magmatic processes: The inception of seafloor spreading in the Norwegian–Greenland sea. In: Morton AC and Parsons B (eds.) Early Tertiary Volcanism and the Opening of the NE Atlantic, pp. 34–38. London: The Geological Society. Nagel TJ and Buck WR (2006) On the mechanics of parallel dipping normal faults. Journal of Geophysical Research 111: B08407. Nagel TJ and Buck WR (2004) Symmetric alternative to asymmetric rifting models. Geology 32: 937–940. Nielsen TFD (2004) The shape and volume of the Skaergaard Intrusion, Greenland; implications for mass balance and bulk composition. Journal of Petrology 45(3): 507–530. Palmason G (1980) A continuum model of crustal generation in Iceland: kinematic aspects. Journal of Geophysical Research 47: 7–18.
The Dynamics of Continental Breakup and Extension Parsons B and Sclater JG (1977) Ocean floor bathymetry and heat flow. Journal of Geophysical Research 82: 803–827. Parsons T and Thompson GA (1993) Does magmatism influence low-angle normal faulting? Geology 21: 247–250. Patton TL, Moustafa AR, Nelson RA, and Abdine AS (1994) Tectonic evolution and structural setting of the Suez Rift. In: Landon SM (ed.) Interior Rift Basins, pp. 9–55. Tulsa, OK: American Association of Petroleum Geologists. Poliakov A and Buck WR (1998) Mechanics of stretching elastic-plastic-viscous layers: Applications to slowspreading mid-ocean ridges. In: Buck WR, Delaney PT, Karson JA, and Lagabrielle Y (eds.) AGU Monograph: Faulting and Magmatism at Mid-Ocean Ridges, vol. 106, pp. 305–324. Washington DC: AGU. Ramberg H (1955) Natural and experimental boudinage and pinch-and-swell structures. Journal of Geology 63: 512–526. Rosendahl BR (1987) Architecture of Continental rifts with special reference to East Africa. Annual Review of Earth and Planetary Sciences 15: 443–503. Royden L and Keen CE (1980) Rifting process and thermal evolution of the continental margin of eastern Canada determined from subsidence curves. Earth and Planetary Science Letters 51: 343–361. Royden LH, Burchfiel BC, and King RW (1997) Surface deformation and lower crustal flow in eastern Tibet. Science 276(5313): 788–790. Rubin AM (1995) Propagation of Magma-filled cracks. Annual Review of Earth and Planetary Sciences 23: 287–336. Rubin AM and Pollard DD (1987) Origins of Blake-Like Dikes in Volcanic Rift Zones. In: Decker RW, Wright TL, and Stauffer PH (eds.) US Geological Survey Professional Paper 1350, pp. 1449–1470. Reston, VA: US Geological Survey. Salveson JO (1978) Variations in the geology of rift basins; a tectonic model. 1978 International Symposium on the Rio Grande Rift; Program and Abstracts. Olsen KH and Chapin CE, Conference Proceedings – Los Alamos Scientific Laboratory. 7487: 82–86. Saunders AD, Fitton JG, Kerr AC, Norry MJ, and Kent RW (1997) The North Atlantic igneous province. In: Mahoney JJ and Coffin MF (eds.) Large Igneous Provinces; Continental, Oceanic, and Planetary Flood Volcanism, vol. 100, pp. 45–93. Washington, DC: American Geophysical Union. Sawyer DS (1985) Brittle failure in the upper mantle during extension of continental lithosphere. Journal of Geophysical Research, B 90(4): 3021–3025. Sengor AMC and Burke K (1978) Relative timing of rifting and volcanism on Earth and its tectonic implications. Geophysical Research Letters 5: 419–421. Sibson RH (1985) A note on fault reactivation. Journal of Structural Geology 7(6): 751–754. Sleep NH (1971) Thermal effects of the formation of Atlantic continental margins by continental breakup. Geophysical Journal of the Royal Astronomical Society 24: 325–350. Smith RB (1977) Formation of folds, boudinage, and mullions in non-Newtonian materials. Geological Society of America Bulletin 88: 312–320. Solomon SC, Richardson RM, and Bergman EA (1980) Tectonic stresses: Models and magnitudes. Journal of Geophysical Research 85: 6086–6092. Sonder LJ and England PC (1989) Effects of a temperaturedependent rheology on large-scale continental extension. Journal of Geophysical Research, B, Solid Earth and Planets 94(6): 7603–7619. Sonder LJ, England PC, Wernicke BP, and Christiansen RL (1987) A physical model for Cenozoic extension of Western North America. In: Coward MP, Dewey JF, and Hancock PL (eds.) Continental Extensional Tectonics,
375
pp. 187–201. London: Geological Society of London (Durham, UK, April 18–20, 1985). Spencer JE (1984) Role of tectonic denudation in warping and uplift of low-angle normal faults. Geology 12(2): 95–98. Spencer JE and Chase CG (1989) Role of crustal flexure in initiation of low-angle normal faults and implications for structural evolution of the Basin and Range province. Journal of Geophysical Research 94: 1765–1775. Spohn T and Schubert G (1982) Convective thinning of the lithosphere; a mechanism for this initiation of continental rifting. Journal of Geophysical Research 87: 4669–4681. Spyropoulos C, Griffith WJ, Scholz CH, and Shaw BE (1999) Experimental evidence for different strain regimes of crack populations in a clay model. Geophysical Research Letters 26(8): 1081–1084. Steckler MS, Berthelot F, Lyberis N, and LePichon X (1988a) Subsidence in the Gulf of Suez: Implications for rifting and plate kinematics. Tectonophysics 153: 249–270. Steckler MS and ten Brink US (1986) Lithospheric strength variations as a control on new plate boundaries: Examples from the northern Red Sea region. Earth and Planetary Science Letters 79: 120–132. Steckler MS and Watts AB (1978) Subsidence of the Atlantictype continental margin off New York. Earth and Planetary Science Letters 41: 1–13. Steckler MS, Watts AB, and Thorne JA (1988b) Subsidence and basin modeling at the U.S. Atlantic passive margin. In: Sheridan RE and Grow Ja (eds.) The Geology of North America, vol. 1–2, The Atlantic Continental Margin: U.S., pp. 399–416. Boulder, CO: Geological Society of America. Stein R, King G, and Rundle J (1988) The Growth of Geological Structures by Repeated Earthquakes 2. Field Exmaples of Continental Dip-Slip Faults. Journal of Geophysical Research 93(B11): 13319–13331. Stewart JH (1978) Basin-range structure in western North America; a review. In: Smith RB and Eaton GP (eds.) Cenozoic Tectonics and Regional Geophysics of the Western Cordillera, pp. 1–31. Boulder, CO: Geological Society of America. Supak SK, Bohnenstiehl DR, and Buck WR (2006) Flexing is not stretching: An analog study of bending induced cracking. Earth and Planetary Science Letters 246: 125–137. Thatcher W and Hill DP (1991) Fault orientations in extension and conjugate strike-slip environments and their implicationsgy. Geology 19: 1116–1120. Tirel C, Brun JP, and Sokoutis D (2006) Extension of thickened and hot lithospheres; inferences from laboratory modeling. Tectonics 25: 1. Tron V and Brun JP (1991) Experimental and numerical modelling of continental deformation. Technophysics 188(1–2): 71–84. Tryggvason E (1980) Subsidence events in the Krafla area, north Iceland 1975–1979. Journal of Geophysics 47: 141–153. Tucholke B, Lin J, and Kleinrock M (1998) Megamullions and mullion structure defining oceanic metamorphic core complexes on the Mid-Atlantic Ridge. Journal of Geophysical Research 103: 9857–9866. Tucholke B, Lin J, Kleinrock M, et al. (1997) Segmentation and crustal structure of the western Mid-Atlantic Ridge flank, 25 degrees 25’-27 degrees 10’N and 0-29 m.y. Journal of Geophysical Research, B, Solid Earth and Planets 102(5): 203–210. Turcotte D and Schubert G (2002) Geodynamics: Application of Continuum Physics to Geological Problems. New York: John Wiley & Sons. Turcotte DL and Emerman SH (1983) Mechanisms of active and passive rifting. Tectonophysics 94: 39–50. Varga RJ (2003) The sheeted dike complex of the Troodos Ophiolite and its role in understanding mid-ocean ridge processes. In: Dilek Y and Newcomb S (eds.) Geological
376
The Dynamics of Continental Breakup and Extension
Society of America, vol. 373, pp. 323–336. Special Paper: Ophiolite Concept and the Evolution of Geological Thought. Vening-Meisnez FA (1950) Les grabens Africains re´sultants de compression ou de tension de la crouˆte terrestre? Memoires, Institut Royal Colonial Belge 21: 539–552. Venkataraman A, Nyblade AA, and Ritsema J (2004) Upper mantle Q and thermal structure beneath Tanzania, East Africa from teleseismic P wave spectra. Geophysical Research Letters 31: 15. Watts AB (2001) Isostasy and Flexure of the Lithosphere. Cambridge, United Kingdom: University of Cambridge. Watts AB and Ryan WBF (1976) Flexure of the lithosphere and continental margin basins. Tectonophysics 36: 25–44. Wegener A (1929) The Origin of Continents and Oceans. London, UK: Methuen. Weissel JK and Karner G (1989) Flexural uplift of rifts flanks due to mechanical unloading of the lithosphere during extension. Journal of Geophysical Research 94: 13919–13950. Wernicke B (1985) Uniform-sense normal simple shear of the continental lithosphere. Canadian Journal of Earth Sciences 22: 108–125. Wernicke B (1981) Low-angle normal faults in the Basin and Range province: Nappe tectonics in an extending crogen. Nature 291: 645–648. Wernicke B and Axen GJ (1988) On the role of isostasy in the evolution of normal fault systems. Geology 16: 848–851. Wernicke BP (1992) Cenozoic extensional tectonics of the U.S. Cordillera. In: Burchfield BC, Lipman PW, and Zoback ML (eds.) The Cordilleran OrogenL Conterminous U.S. pp. 553–583. Boulder, CO: Geological Society of America G-3. White NJ (2004) Using prior subsidence data to infer basin evolution. In: Curtis A and Wood R (eds.) Geological Prior Information; Informing Science and Engineering, vol. 239, pp. 211–224. White RS and McKenzie DP (1989) Magmatism at rift zones: The generation of volcanic continental margins and flood basalts. Journal of Geophysical Research 94: 7685–7729.
White RS and McKenzie DP (1995) Mantle plumes and flood basalts. Journal of Geophysical Research 94: 17543–17585. Wiens DA and Stein S (1984) Intraplate seismicity and stresses in young oceanic lithosphere. Journal of Geophysical Research 89: 11442–11464. Willett SD and Pope DC (2004) Thermo-mechanical models of convergent orogenesis; thermal and rheologic dependence of crustal deformation. In: Karner GD, Taylor B, Driscoll NW, and Kohlstedt DL (eds.) Rheology and Deformation of the Lithosphere at Continental Margins, pp. 179–222. New York, NY: Columbia University Press. Wills S and Buck WR (1997) Stress field rotation and rooted detachment faults: A test of fault initiation models. Journal of Geophysical Research 102(20): 20503–20514. Withjack MO and Schlische RW (2006) Geometric and experimental models of extensional fault-bend folds. In: Buiter SCJ and Schreurs G (eds.) Geological Society Special Publications, Analogue and Numerical Modelling of Crustal-Scale Processes, vol. 253, pp. 285–305. London: The Geological Society. Yin A (1989) Origin of regional, rooted low-angle normal faults: A mechanical model and its tectonic implications. Tectonics 8: 469–482. Yin A and Dunn J (1992) Structural and stratigraphic development of the Whipple–Chemechuevi detachment fault system, southeastern California: Implications for the geometrical evolution of domal and basinal low-angle normal faults. Geological Society of American Bulletin 104(6): 659–674. Zorin YA (1981) The Baikal Rift; an example of the intrusion of asthenospheric material into the lithosphere as the cause of disruption of lithospheric plates. Tectonophysics 73(1–3): 91–104. Zuber MT and Parmentier EM (1986) Lithospheric necking: A dynamic model for rift morphology. Earth and Plantary Science Letters 77: 373–383.
6.09 Dynamic Processes in Extensional and Compressional Settings – Mountain Building: From Earthquakes to Geological Deformation J-P. Avouac, California Institute of Technology, Pasadena, CA, USA ª 2007 Elsevier B.V. All rights reserved.
6.09.1 6.09.2 6.09.2.1 6.09.2.2 6.09.2.3 6.09.2.4 6.09.2.5 6.09.2.6 6.09.2.7 6.09.3 6.09.3.1 6.09.3.1.1 6.09.3.1.2 6.09.3.1.3 6.09.3.1.4 6.09.3.1.5 6.09.3.2 6.09.3.2.1 6.09.3.2.2 6.09.3.2.3 6.09.4 6.09.4.1 6.09.4.2 6.09.4.3 6.09.4.4 6.09.5 6.09.5.1 6.09.5.2 6.09.5.2.1 6.09.5.2.2 6.09.6 6.09.6.1 6.09.6.2 6.09.6.3 6.09.6.4 6.09.6.5 6.09.6.6 6.09.6.7 6.09.7 6.09.7.1 6.09.7.2 6.09.7.3 6.09.7.4
Introduction Geodynamical Setting of the Himalaya The Himalaya as a Result of the India–Asia Collision Variation of Crustal Thickness across the Himalaya Geological Architecture of the Himalayan Range and Southern Tibet Metamorphism Topography across the Himalayan Range Crustal-Scale Structural Models of the Himalaya Geophysical Constraints on the Structure of the Crust Holocene Deformation and Erosion Active Thrusting and Folding in the Sub-Himalaya Structural evolution of the sub-Himalaya River incision across the sub-Himalaya Converting incision rates to uplift rates in the sub-Himalaya Converting uplift rates to horizontal shortening from area balance Converting uplift rates to horizontal shortening from the fault-bend fold model River incision, Erosion, and Uplift across the Range Fluvial incision across the whole range Denudation across the whole range Holocene kinematics of overthrusting along the MFT–MHT Longer-Term Geological Deformation and Exhumation Foreland Deposition: A Record of Underthrusting Structural Evolution of the Thrust Package Exhumation of the Lesser and High Himalaya: A Record of Overthrusting Overthrusting, Underthrusting, and Accretion Kinematic and Mechanical Models of Crustal Deformation Thermokinematic Model of the Evolution of the Range since 15 Ma Modeling Deformation and Surface Processes Based on Continuum Mechanics Model implementation Modeling results Geodetic Deformation and the Seismic Cycle Large Earthquakes in the Himalaya Geodetic Deformation in the Nepal Himalaya Microseismic Activity in the Nepal Himalaya What Controls the Down-Dip End of the Locked Portion of the MHT? A Model of the Seismic Cycle in the Central Nepal Himalaya Geodetic Deformation, Seismic Coupling, and Recurrence of Large Earthquakes in the Himalaya Is Interseismic Strain Stationary? Discussion The Critical Wedge Theory: Does It Apply to the Himalaya? Evidence for Low Friction on the MHT Importance of the Brittle–Ductile Transition Metamorphism during Underthrusting
378 379 379 380 382 386 387 387 387 388 388 389 390 391 392 394 395 395 396 396 396 397 399 400 402 403 403 405 405 408 409 410 411 414 417 417 418 420 421 421 424 425 426 377
378
Mountain Building: From Earthquakes to Geological Deformation
6.09.7.5 6.09.7.6 6.09.7.6.1 6.09.7.6.2 6.09.7.7 6.09.8 References
How Does the Steep Front of the High Himalaya Relate to Tectonics, Erosion, and Climate? The Elevation and Support of Mountain Ranges: Effect of Climate and Lower Crustal Flow Height and width of a critical brittle wedge Effect of ductile deformation in the lower crust The Fate of the Indian Crust and Mantle Lithosphere Conclusions
6.09.1 Introduction Mountain ranges are the most spectacular manifestation of continental dynamics. The fact that some mountain ranges were able to maintain their topography over tens of millions of years, while their erosion was feeding large sedimentary basins, is unambiguous evidence that tectonic forces can cause sustained uplift of subsidence of the continental crust. Geologists noticed quite early on that most mountain ranges are contractional orogens, the result of horizontal contraction of the continental crust, and that they tend to form long belts separating domains with often quite different geological history (e.g., Willis, 1891; Argand, 1924). A rapid tour of active mountain ranges on Earth today shows that contractional mountain ranges can arise in a variety of contexts. Some have formed along converging plate boundaries as the result of collisions which can involve two continents (along the Himalaya, for example, as detailed in this review), a continent and an island arc (in Taiwan) (e.g., Malavieille et al., 2002a), or a continent and an oceanic plateau (in the Southern Alps of New Zealand) (e.g., Walcott, 1998). Contractional mountain ranges can also form along subduction zones without being necessarily collisional features. In the Andes for example, the stresses transmitted across a subduction zone appear to be sufficient to cause trench-perpendicular shortening (e.g., (Lamb, 2006), probably because high heat flow in the back-arc zone weakens the continental lithosphere (Hyndman et al., 2005). Mountain ranges are thus often closely associated with converging plate boundaries. However, active mountain building can also occur far away from plate boundaries, the Tien Shan, in Central Asia, being an outstanding example (e.g., Hendrix et al., 1992 and Avouac, et al., 1993). Mountains are major players in the interactions between solid Earth and its climate. Surface processes
426 428 429 429 430 431 432
participate to the mechanics of mountain building not only because of the isostatic response to erosion (Molnar and England, 1990a), but because they contribute to focalizing crustal deformation leading to a positive feedback between erosion at the surface and shortening of the crust (Avouac and Burov, 1996). Climate may also influence the mechanics of mountain building along subduction zones, through its influence on the amount of sediments delivered to the trench and hence on the mechanical coupling across the plate boundary (Lamb and Davis, 2003). Mountain ranges affect atmospheric circulation and the distribution of precipitation, and consequently drainage patterns (Ruddiman and Kutzbach, 1991; Raymo and Ruddiman, 1992; Ramstein, et al., 1997). Their erosion influences eventually their elevation through isostasy, and influences the chemistry of the atmosphere through a variety of chemical reactions and through burial of organic matter (Kerrick and Caldeira, 1993, 1999; Derry and France-Lanord, 1997; France-Lanord and Derry, 1997). Finally, mountain ranges and their piedmonts are also a primary locus of geohazards, earthquakes, landslides, and floods, in particular. For all these reasons, understanding better orogenic processes is a fundamental issue in geology. The anatomy of mountain ranges, and the tectonic processes at work deep in the crust, might be best studied from the investigation of the exhumed core of ancient orogens. However, a lot of insight on orogenic processes can be gained by ausculting orogens that are actively deforming. The main intent of this chapter is to illustrate that point and show how the study of active processes can shed light on how mountain ranges form and evolve over a geological period of time. This chapter focuses on the Himalaya as it is undoubtedly the the world’s most impressive example of an active collisional orogen. This incomparably long and high mountain arc is the setting of rapid,
Mountain Building: From Earthquakes to Geological Deformation
ongoing crustal shortening and thickening, intense denudation driven by the monsoon climate, and frequent very large earthquakes. The relation of this range to plate tectonics has long been recognized (Dewey and Bird, 1970). As reviewed in this chapter, we now have a reasonably solid understanding of the structure of the range, of its petrometamorphic history, and of the kinematics of its active deformation. The long-term geological history of the range – from several millions to a few tens of millions of years – has been documented by structural, thermobarometric, and thermochronological studies. Morphotectonic investigations have revealed its evolution over the past several thousands or tens of thousands of years. And, finally, geodetic measurements and seismological monitoring have revealed the pattern of strain and stress buildup over several years. This chapter shows that the results of these investigations can be assembled into a simple, coherent picture of the structure and evolution of the range. Some emphasis is put on the key role of surface processes: these processes carved the morphologic features that are used to deduce vertical displacements; they generated the molasse deposits that filled the subsiding foreland basin, providing a stratigraphic record of mountain building; also they influenced the evolution of the range by changing its thermal structure and stress field via redistribution of surface mass. Surface processes
379
therefore contribute to recording the geological history of an orogen, and they also participate in the mountain-building process itself. The Himalaya is consequently one of the best places on Earth where the geological history of a mountain belt can be compared with its current tectonic processes. Although some aspects might be specific to the setting of this mountain range, the tectonic processes at work there are presumably the same as those at work in any other contractional mountain range.
6.09.2 Geodynamical Setting of the Himalaya 6.09.2.1 The Himalaya as a Result of the India–Asia Collision The Himalayan arc and the Tibetan Plateau formed as a result of the collision between India and Asia (e.g., Powell and Conaghan, 1973). The two continents were once separated by the Tethys Sea, which was subducted beneath the southern margin of Asia (Figure 1). The age of the onset of collision is still debated, and is estimated to be beween c. 65 and 45 Ma depending on the chosen approach to that question (e.g., Patriat and Achache, 1984; Jaeger et al., 1989; Searle et al., 1990; Rowley, 1996; DeSigoyer et al., 2000; Ding et al., 2005). In fact, Trench
–100 Ma
Volcanic arc Continental crust India
Tethys
Asia
Upper mantle
Oceanic crust Upper mantle Subduction
–40 Ma
Collision
Suture
Figure 1 Prior to the collision, an ocean (the Tethys Sea) used to separate the northern margin of India and Eurasia. The southern margin of Asia was an active margin with a subduction zone similar, for example, to the Andean subduction zone bordering the western margin of South America. Modified from Malavieille J, Marcoux J, and De Wever P (2002b) L’ocean perdu. In: Museum National D9Histoire Naturelle (France), Avouac J-P, and De Wever P (eds.) Himalaya-Tibet, Le choc des continents, pp. 32–39. Paris: CNRS Editions et Museum national de’Histoire naturelle.
380
Mountain Building: From Earthquakes to Geological Deformation
it seems that the rise of the topography during a continental collision might postdate significantly the onset of the tectonic collision, defined as the time when the continental margin on the subducting plate first encounters the trench. For example, the northern Australia margin has started colliding with the Banda arc in Pliocene resulting in a well-developed fold-and-thrust belt, but no significant mountain range has formed yet (Bowin et al., 1980). This is probably because, in the early stage of a collision, subduction of the cold continental crust can occur. This process, required from the observation of high-pressure metamorphic rocks which were exhumed quite early in the history of the collision (DeSigoyer et al., 2000), has been observed in analog (Chemenda, et al., 2000) and thermomechanical modeling (Toussaint, et al., 2004a). The first stratigraphic evidence of the collisions, namely, the transition from passive margin sedimentation to fore-arc sedimentation, is observed in the Ypresian (50.7 Ma) in the western Himalaya (Beck, et al., 1995) and seems younger than Middle Lutetian (40 Ma) in the central-eastern Himalaya, suggesting that suturing began in the west and migrated east (e.g., Rowley, 1996). Reconstruction of the plate motion of India relative to Asia (Figure 2(a)) shows an abrupt decrease in the convergence rate, probably resulting from the collision, from 15 to 4–5 cm yr–1 at c. 50 Ma (Figure 2(b)) (Molnar and Tapponnier, 1975; Patriat and Achache, 1984). Since that time, India has indented 3000 km into Asia, producing a combination of lateral escape and crustal thickening that has given rise to the highest topographic features on Earth (e.g., Molnar and Tapponnier, 1975; Peltzer and Tapponnier, 1988; Harrison, et al., 1992; Tapponnier et al., 2001) (Figure 3). Although the pattern and kinematics of active faults in Asia and geodetic measurements clearly show a combination of strike-slip faulting and crustal shortening, the respective contribution of these two mechanisms to the overall deformation remains a matter of debate (e.g., Molnar and Tapponnier, 1975; Avouac and Tapponnier, 1993; Larson et al., 1999a; Wang et al., 2001; Zhang et al., 2004) (Figure 4). At present, northern India is moving 35 mm yr1 along an N15 E azimuth (31 mm yr1 along an N10 E azimuth for a point located at the western Himalayan syntaxis and 38 mm yr1 along an N20 E azimuth for a point located at the eastern Himalayan syntaxis) (Bettinelli et al., 2006) (Figure 2(b)). This rate is nearly 25% less than that suggested by the Indian plate motion model derived from the opening of
the Indian Ocean (Figure 2(b)). The unusually large difference between the geodetic measurements and geologic plate model is probably due to the difficulty in taking into account in plate models the zone diffuse intraplate deformation that has developed within the Indian Ocean, probably as a consequence of the intraplate stresses induced by the collision itself (Gordon et al., 1990; Royer and Chang, 1991; Deplus et al., 1998). Crustal shortening across the Himalaya, estimated at 19 2.5 mm yr1 based on geodetic measurements from central and eastern Nepal (Bettinelli et al., 2006) (see Section 6.09.6), absorbs nearly half of the current convergence rate. 6.09.2.2 Variation of Crustal Thickness across the Himalaya The remarkable elevation of the Himalayan and Tibetan Plateau unambiguously results from an extremely thick crust. Various seismic experiments conducted in southern Tibet all suggest a crustal thickness on the order of 80 km (Hirn et al., 1984; Zhao et al., 1993; Mitra et al., 2005). For comparison, the crustal thickness of the Indian Shield is estimated at 40–44 km (Saul et al., 2000; Mitra et al., 2005) (Figure 5). Receiver function studies across the eastern Nepal Himalaya (Schulte-Pelkum, et al., 2005) and northeastern India (Mitra et al., 2005) show the Moho can be traced as a continuous feature across the arc. This observation agrees with the modeling of gravity measurements (Cattin et al., 2001; Hetenyi et al., 2006). The Bouguer anomaly in India and in Tibet primarily indicates local Airy compensation but important deviations from Airy isostasy are observed below the Himalayan range and its foreland (Lyon-Caen and Molnar, 1983, 1985; Jin et al., 1996; Cattin et al., 2001), and are particularly evident in the gravity data across the central Nepal Himalaya (Figure 5). Values more negative than those expected from local isostasy are observed over the Gangetic Plain, indicating some mass deficit there. By contrast, mass excess is indicated below the adjacent Himalayan range. These deviations are the signature of flexural support of the range, meaning that the weight of the Himalaya is supported by the strength of the underthrusting Indian Plate. The steep gravity gradient, on the order of 1.3 mGal km1, beneath the High Himalaya suggests a locally steepening of the Moho’s dip. Flexural modeling of a thin elastic plate overlying an inviscid fluid (Lyon-Caen and Molnar, 1983, 1985; Maggi et al., 2000) successfully reproduces the observed gravity
Mountain Building: From Earthquakes to Geological Deformation
381
(a) 40° N 0 10
30° N
0
20
10
30
20
40
20° N
30 50
10° N
40
60
50
0° 70 60
10° S 80
20° S
70
80
30° S
40° S 50° E
India/Eurasia convergence velocity (mm yr–1)
(b)
60° E
70° E
80° E
90° E
100° E
160 140 120 100 80 Eastern syntaxis
Western syntaxis
60 40
–1
38 mm yr 31 mm yr–1
20 0 80
70
60
50
40
30
20
10
0 Ma
Figure 2 Plate motion of India relative to Eurasia (a) and convergence rates (b) over the last 80 Ma. This kinematics was obtained from the recent synthesis of the magnetic anomalies of the Indian Ocean (Royer and Patriat, 2002) (see Patriat and Segoufin (1988) for a former similar analysis). The India/Africa relative motion derived from the analysis of this data set was referenced to Eurasia through an Africa/North America and North America/Eurasia plate circuit across the Atlantic Sea. Convergence rates were computed at points attached to the Indian Plate located at the current position of the eastern and western Himalaya. The modern velocities computed at this same points from a plate model determined from geodetic measurements (Bettinelli, et al., 2006) are shown for comparison (dashed lines). The abrupt decrease of the convergence rate starting at about 50 Ma is thought to relate to the onset of the India–Asia continental collision (Molnar and Tapponnier, 1975).
382
Mountain Building: From Earthquakes to Geological Deformation
influence of the thermal structure of the range on crustal rheology. The kinematics of underthrusting results in a thermal structure with relatively high temperatures at mid-crustal depths, favoring ductile flow within the crust and implying some decoupling between the upper crust and upper mantle and an abrupt decrease of the apparent flexural rigidity of the lithosphere (Burov and Diament, 1995). If the thermal structure and its influence on crustal rock rheology are taken into account, there is indeed no need for additional forces other than the weight of the topography to result in the downwarping of the Indian lithosphere under the load of the range and sediments in the foreland (Figure 5), as is shown in the thermomechanical model of Cattin et al. (2001).
ITSZ
Figure 3 Sketch showing how indentation of India into Eurasia since the onset of the collision has been absorbed by a combination of crustal thickening and lateral escape.
6.09.2.3 Geological Architecture of the Himalayan Range and Southern Tibet
anomalies. However, this kind of model requires that some forces or momentum be exerted on the flexed plate in addition to the load of the high topography (Lyon-Caen and Molnar, 1983). In order to account for the locally steeper gradient of gravity anomalies (Figure 5), this kind of modeling also requires an abrupt weakening of the Indian Plate beneath the high range. This weakening probably reflects the
Relics of the Tethys Sea can now be traced along the Indus–Tsangpo suture zone (ITSZ) (e.g., Burg, 1983; Searle et al., 1987) well north of the Himalayan summits (Figure 6). To the south, Cambrian to Eocene Tethyan sediments deposited on the northern passive margin of the Indian continent were sutured to the volcanic and plutonic rocks of the once-active margin of Eurasia (Burg et al., 1987; Searle et al., 1987). Now uplifted to elevations of 5000 m, they were
Eurasia
40° N
Tarim Ordos Tibet
20°
N
South China
India
120° E
80° E 20 mm yr–1
100° E
km 0
500
1000
Figure 4 Velocities, relative to stable Eurasia, measured from GPS over a 10-year period. Data from Wang Q, Zhang P-Z, Freymueller JT, et al. (2001) Present-day crustal deformation in China constrained by global positioning system measurements. Science 294: 574–577.
Bouguer anomaly (mGal)
Mountain Building: From Earthquakes to Geological Deformation
0
383
Predicted from local isostasy
–200 Mass deficit
Mass excess Model (Hetenyi et al., 2006)
–400 –600 –500
0
500
1000 NNE
SSW Gangetic plain
Himalaya
Tibet
Depth (km)
0 Crust Moho from HIMNT experiment
–50
Moho from receiver functions
Mantle –100 –500
0 500 Horizontal distance (km)
1000
Figure 5 Bouguer anomalies along an N18 E profile across the Himalaya of central Nepal (modified from (Cattin, et al., 2001) and (Hetenyi, et al., 2006)) (approximately section AA9 in Figure 7). All the data within a 30-km-wide swath were projected onto the section. Several data sets were merged with accuracies ranging from about 0.5 to 7 mGal (Bureau Gravime´trique International database: Van de Meulebrouck, 1983; Abtout, 1987; Sun, 1989). Thick vertical bars show Moho picks on receiver functions beneath Tibet (Kind, 1996) and beneath the Indian Shield (Saul et al., 2000). Also shown is the Moho across eastern Nepal Himalaya determined from the HIMNT experiment (dashed line) (Schulte-Pelkum, et al., 2005) and from the Hi-CLIMB experiment (Hetenyi, et al., 2006). The thick line shows the expected Bouguer anomaly in the case of isostatic compensation of a crust with 2.67 density over an upper mantle with 3.27 density (computed from the topography smoothed with a 50-km-wide gaussian). The thin line is the Moho computed from a mechanical model that accounts for the low-density sediments in the foreland, the rheological layering of the crust, and its dependence on the thermal structure (Cattin et al., 2001).
intensely deformed, probably mostly in the early ‘Alpine’ period of the collision (Burg et al., 1984; Ratschbacher et al., 1994), although there is evidence that thrust faulting in southern Tibet probably persisted until mid-Miocene time (Yin, et al., 1999). A major normal fault separates the Tethyan sedimentary cover from the High Himalayan crystalline units; initially called the North Himalayan Normal Fault (Burg et al., 1984), its later appellation as the the South Tibetan Detachment (STD) is more commonly used in the literature (Burchfiel et al., 1992). Motion on the STD is thought to have resumed sometime between 15 and 20 Ma (Searle et al., 1997). To the south, crustal shortening chiefly resulted from deformation on a limited number of major thrust faults. From north to south, these are the Main Central Thrust (MCT), the Main Boundary Thrust (MBT), and the Main Frontal Thrust (MFT) Faults (Figures 7 and 8) (e.g., Gansser,
1964; Le Fort, 1975a; Nakata, 1989; Yeats et al., 1992; Meigs et al., 1995). These faults were likely activated in a forward propagating sequence. The major Himalayan thrust faults separate domains with contrasting geology: The foreland basin. The Indo-Gangetic Plain is a 200–300-km-wide foreland basin in which a fraction of the material eroded from the neighboring rising topography is trapped. The rest of the material is transported by the Ganges drainage system and ultimately delivered to the Bengal Fan. Several kilometers of Cenozoic molasse deposits have thus accumulated over the Archean to Early Proterozoic metamorphic Indian basement rocks. The foreland fold-and-thrust belt. North of the IndoGangetic Plain, the sub-Himalaya is a zone of thinskinned tectonics bounded to the south by the MFT and to the north by the MBT (e.g., Delcaillau, 1986; Mugnier et al., 1999; Lave´ and Avouac, 2000). Tertiary
384
Mountain Building: From Earthquakes to Geological Deformation
90 E
80 E +
La
+
+
da
+
k
+
h + + +
+ + +
+ + + +
++ +
High Himalaya
Arc Granites Fore-arc sediments
Lesser Himalaya
Ophiolites
Sub-Himalayan Molasse
Tethys sediments
Gangetic plain deposits
+ +
0
+
Ga + ng de
TIBET +
+
Lhasa +
+
Annapurna
Main Central Thrust Delhi
+
A′
+
+
INDIA
+
+ +
+
+
+
+
ne uture Zo
S sangpo Indus–T
Everest
Main boundary Thrust
+
+ + +
H
++
EPT
30 N
500 km
IND
se + +
Katmandu
Main Frontal Thrust
A
Figure 6 Simplified geologic map of Himalayan arc. For clarity High Himalayan nappes over the LH are not represented. Locations of section AA9 and INDEPTH I reflection line (e.g., Hauck et al., 1998) are indicated. Geological map modified from Malavieille J, Marcoux J, and De Wever P (2002b) L’ocean perdu. In: Museum National D9Histoire Naturelle (France), Avouac J-P, and De Wever P (eds.) Himalaya-Tibet, Le choc des continents, pp. 32–39. Paris: CNRS Editions et Museum national de’Histoire naturelle.
siltstones, sandstones, and conglomerates have been scraped off the basement, folded, and faulted at the front of the advancing range, forming a typical foreland fold-and-thrust belt. In central Nepal, the sections along Surai Khola (Corvinus, 1988; Appel and Rossler, 1994),
29° N
TIBET
Tinau Khola (Gautam and Appel, 1994), Arung Khola (Tokuoka et al., 1986), and Bakeya Khola (Harrison et al., 1993) show almost the same sedimentary sequence, 3500–5500 m thick, that paleontological and magnetostratigraphic studies indicate was deposited between 14
A′
29° N Quaternary Siwaliks
Main
28° N
Cen
Lower LH Upper LH
tral T
Granite HHC Crystalline nappe Tothys formations
hrus
Main Bo
t
28° N
undary T
hrust
K B′
Main F rontal T
hrust
B
27° N
27° N
A INDIA 0
100 84° E
200 km 86° E
88° E
Figure 7 Geological map of central and Eastern Nepal with location of section AA9 (Figure 8) and BB9 (Figure 11). K, Katmandu basin. Source data: Department of Mines and Geology, Nepal.
Mountain Building: From Earthquakes to Geological Deformation
385
8000 High Himalaya
Tibet
Elevation (m)
6000 Lesser Himalaya
4000 Sub-Himalaya
2000
Gangetic plain
0
0
100
IT
ST D
CT
ST
50
200
M
Katmandu
M
FT
M
0 Depth (km)
Mahabharat
SZ
Mcho
100
100
0
200
Distance from MFT (km) Upper Siwaliks Middle Siwaliks
Paleozoic metasediments Sub-Himalaya
Indian upper crust
Lower Siwaliks
Lesser Himalaya
Leucogranite Gneisses
Higher Himalaya
Tertiary Jurassic Trias
Tethys formations
Paleozoic
Figure 8 Geological section across central Nepal Himalaya at the longitude of Katmandu. Thick line shows the MHT fault, MHT, which reaches the surface at the front of the Siwalik Hills, where it coincides with the MFT. The MBT fault separates LH metasediments from the molasse deposits of the sub-Himalaya (the Siwalik Hills). The Main Central Thrust (MCT) Fault places the higher-grade metamorphic rocks of the High Himalayan crystalline units over the LH metasediments. Top: Mean, maximum, and minimum elevation within a 50-km-wide swath centered on section AA9 (see Figures 6 and 7 for location of section). Several geographical domains are distinguished from north to south : Tibet, High Himalaya, LH, sub-Himalaya, and the Gangetic Plain.
and 1 Ma. In places, the fold belt involves slices of preTertiary sediments, which are reddish-maroon quartzites and gray shales with some doleritic intrusions (Lave´ and Avouac, 2000), very similar to Vindhyan units drilled at Raxaul (Sastri et al., 1971). Pre-Tertiary units involved in the sub-Himalayan fold belt are also reported 50 km east of the section along the Bagmati, north of the Kamla Khola (Mascle and He´rail, 1982). These observations indicate that the decollement underlying the fold belt must lie on top of the Indian basement, at 5–6 km depth along the section across central Nepal (Figure 8).
The internal zones. North of the MBT, the Lesser Himalayan (LH) units consist of low-grade metasediments: phyllite, quartzite, and limestone of Devonian or older ages (Upreti, 1999) (Figures 7 and 8). In some places, overlying Tertiary units have been preserved, particularly sandstones and siltstones of the Dumri formation (Sakai, 1985). This formation consists of Himalayan foreland sediments that were deposited between c. 16 and 21 Ma, as indicated by dating of detrital muscovite and from preliminary results of a magnetostratigraphic study (DeCelles et al., 1998). The section across the
386
Mountain Building: From Earthquakes to Geological Deformation
6.09.2.4
Himalaya of central Nepal near Katmandu is characterized by a crystalline sheet forming a klippe, usually referred to as the Katmandu Klippe, on top of the LH units (Sto¨cklin et al., 1980). The resistance to erosion of the crystalline units, combined with its uplift along the MBT, is chiefly responsible for the impressive relief of the Mahabharat south of the Katmandu basin (Figure 8). The crystalline units, consisting mainly of schist and gneiss intruded by Late Cambrian to Ordovician granites (Scha¨rer and Allegre, 1983), are overlain by Cambrian to Eocene Tethyan sediments (Sto¨cklin, et al., 1980). Thermobarometric studies indicate that this crystalline sheet was thrust over the Dumri formation, particularly in the Tansen area (Bollinger et al., 2004a). The basal ‘Mahabharat’ thrust can be interpreted as the southern extension of the MCT, although the possibility that it is a distinct thrust fault rooting below the MCT cannot be excluded (e.g., Upreti, 1999). North of the klippe, the foliation in the LH schist decribes a large antiformal structure, called the Pokhara-Gorkha anticlinorium (Peˆcher, 1989). This structure, as shown in Figure 8, has been interpreted as a hinterland dipping duplex structure (Brunel, 1986), and it has been recognized on most sections across the Nepal Himalaya (Schelling and Arita, 1991; Schelling, 1992; DeCelles et al., 2001) and in Kumaon, India (Srivastava and Mitra, 1994). The High Himalayan units consist of amphibolite-grade schist intruded by large leucogranitic plutons. Neodymium isotopic provenance studies suggest that LH metasediments were derived from the Indian craton, while the High Himalayan rocks most probably correspond to an exotic terrane accreted onto India in the Early Paleozoic (Robinson et al., 2001).
Metamorphism
As early as the mid-nineteenth century, field geologists noticed an upward increase of metamorphic grade across the Himalaya (Medlicott, 1864; Mallet, 1874; Oldham, 1883a). A number of field studies have since documented this ‘inverted’ gradient (e.g., Le Fort, 1975b; Peˆcher, 1989). The inverted gradient is most obvious within and close to the MCT zone, especially along the Everest section, which has received particular attention (Lombard, 1953; Brunel and Kienast, 1986; Lombardo and Rolfo, 2000; Catlos et al., 2002; Law et al., 2004). Metamorphic grade increases upward from chlorite to biotite, garnet, kyanite, and sillimanite-grade rocks over a structural distance of a few kilometers (Figure 9). The metamorphic grade of the LH, on the other hand, could not be documented from conventional thermobarometric techniques because of the poor mineralogy of these rocks. This limitation was recently overcome by means of the calibrated RSCM thermometric technique (Beyssac et al., 2004a). This technique uses the degree of graphitization of carbonaceaous matter, measured by Raman microspectroscopy, to determine peak metamorphic temperatures (Beyssac et al., 2002; Bollinger et al., 2004a). These temperatures in the LH typically vary between 550 C near the MCT to 350 C at deeper structural levels (Figure 9). Combined with local structural measurements, these data indicate that the peak metamorphic temperatures increase upward in the section. This apparently inverted gradient, which was first documented near the MCT zone, is observed to actually encompass the topmost 5–7 km of the LH units. The inverted metamorphic gradient has been interpreted as either evidence for a thermal structure with recumbent
Peak metamorphic temperature 700°C 600°C 500°C 400°C
50 km Ky an Ga rn ite et Si llim an ite
C hl
or it
e
Bi
ot it
e
300°C
MCT
Everest + + + + + + +
+
+
+ ++ + + + +
Figure 9 Simplified section of the MCT zone in the Everest area. Metamorphic grade across the MCT zone shows an apparently inverted gradient together, with metamorphic peak temperatures increasing upward (based on Lombard, 1953; Brunel and Kienast, 1986; Lombardo, 2000; Catlos et al., 2002; Law et al., 2004).
Mountain Building: From Earthquakes to Geological Deformation
isotherms (e.g., Le Fort, 1975a), or as the result of postmetamorphic shearing of isograds (e.g., Brunel et al., 1979; Hubbard, 1996), and it has stimulated a variety of thermal modeling studies (Jaupart and Provost, 1985; Molnar and England, 1990b; Royden, 1993; Harrison et al., 1998). 6.09.2.5 Range
Topography across the Himalayan
The easily erodible molasse of the sub-Himalaya forms foothills reaching elevations of a few hundred meters. The very rough relief of these hills attests to rapid erosion and tectonic uplift (Hurtrez et al., 1999). North of the foothills, in the LH domain, the topography of the range reaches elevations up to 2500 m, with the higher topography correlating strongly with remnants of High Himalayan crystalline klippes (Figure 8). In fact, the morphology in the LH is generally ‘inverted’, with synclinal crests and anticlinal valleys (Valdiya, 1964), suggesting moderate tectonic activity. If we focus on the Himalaya of Nepal, we observe that north of a line trending N108 E (roughly coinciding with the trace of the MCT) the topography rises abruptly from elevations of 1000 m to more than 6000 m. The abrupt break in slope that marks the front of the High Himalaya can be traced all along the Nepal Himalaya. A unique feature of the range’s morphology is the position of the front of the high range well to the north of the main bounding thrust faults along the foothills. This feature has inspired a variety of interpretations, which are discussed in Section 6.09.7. 6.09.2.6 Crustal-Scale Structural Models of the Himalaya A variety of crustal-scale sections across the Himalaya have been proposed based on surface geology. One early view was that all the major thrust faults are crustal-scale faults that developed following a forward propagation sequence after the collision along the Indus-Trangpo Suture Zone (ITSZ) (e.g., Le Fort, 1975b; Mattauer, 1975; 1986; Molnar and Lyon-Caen, 1988; Molnar, 1990). According to this view, all faults would be parallel to one another and reach to the Moho. Subsequent studies proposed the alternative that all the major faults root into a common mid-crustal decollement (Brunel, 1983; Schelling and Arita, 1991). This geometry was initially inferred from the LH antiformal structure, and it appears to be consistent with the pattern of exhumation recorded in the
387
metamorphism of the Lesser and Higher Himalayan formations. Indeed, peak metamorphic pressures documented in the Nepal Himalaya never exceed 8 kbar (see Guillot (1999)) for a review), suggesting that rocks were exhumed from depths of 30–35 km at most. This hypothesis forms the basis of the cross-section constructed in Figure 8, which is similar to the ‘balanced’ sections that have been constructed along several transects across the Himalaya of Kumaon (Srivastava and Mitra, 1994), far western Nepal (DeCelles et al., 2001), central Nepal (Robinson et al., 2001), and eastern Nepal (Schelling and Arita, 1991). The main point is that these balanced sections all suggest some duplex structure in the LH, with all the major thrust faults rooting into a mid-crustal decollement beneath the high range. The decollement must have developed after the emplacement of the crystalline thrust sheets. It should be realized that the geometric rigor used to balance these sections may give a misleading impression of accuracy. The following underlying assumptions do not strictly hold in this case: (1) that deformation occurs by bedding slip with constant width and length of the various units maintained, and (2) that the footwall is nondeformable basement. Not only have the LH units developed an intense foliation and experienced a significant amount of pure shear, the basement itself might actually deform as it is underthrust below the range. The details of the balanced sections, though they have the merit of being based on a plausible kinematics, should therefore be regarded with caution.
6.09.2.7 Geophysical Constraints on the Structure of the Crust The structure of the crust across the central Himalaya has been investigated through a variety of techniques, including gravimetry, magnetotellury, and seismology. A major feature revealed by seismic experiments in southern Tibet is a strong mid-crustal reflector at a depth of 35–40 km. The existence of this reflector was first suggested by wide-angle seismic reflection studies (Hirn and Sapin, 1984); later, it was better imaged by the common midpoint (CMP) deep seismic profiles run during the INDEPTH experiment (Zhao et al., 1993; Brown, et al., 1996; Nelson, et al., 1996). As shown in Figure 10, this conspicuous reflector was found to coincide with the mid-crustal decollement inferred from the structural sections across the range, and it was consequently termed the Main Himalayan Thrust fault, or MHT (Brown et al., 1996).
388
Mountain Building: From Earthquakes to Geological Deformation
D ST
CT
DT M
Katmandu
M
FT M
S
INDEPTH
N
20
80
3200
320
32
10
60
1000
40 100
Depth (km)
0
TIB-1
Ωm 0
100 N18° distance from MFT (km)
TIB-3
200
Figure 10 Geophysical constraints on the crustal structure across central Nepal. The conductivity section was obtained from a magnetotelluric experiment carried out along the section AA9 across central Nepal (Lemonnier et al., 1999). Also reported are the INDEPTH seismic sections run about 300 km east of section AA9 (see location in Figure 7) (Zhao et al., 1993; Brown et al., 1996; Nelson et al., 1996). All the thust faults are inferred to root at depth in a subhorizontal ductile shear zone that would correspond to the prominent mid-crustal reflector.
The deep electrical structure of the central Nepal Himalaya was imaged by magnetotelluric sounding (Lemmonnier et al., 1999) (Figure 10). Variations of electrical conductivity in the crust can result from changes in fluid content, pore geometry, or lithology. Conductive zones in the crust are generally thought to reflect well-connected conductive phases, such as brines or melts, or conductive minerals like graphite (Marquis et al., 1995). The section shows a high conductivity in the foreland (30 m) consistent with the geometry of the molassic foreland basin and contrasting with the resistive Indian basement (>300 m) and LH units (>1000 m). A continuous shallow-dipping conductor that can be traced northward coincides relatively well with the position of the decollement beneath the LH inferred from structural studies. It may well reflect some dragging of fluidrich sediments along the thrust fault, a process expected as the rugged topography of the Indian basement underthrusts the LH. Farther north, the magnetotelluric section shows a major conductive zone (30 m) at 15 km depth under the front of the High Himalaya. This zone coincides with the position of the midcrustal ramp beneath the front of the High Himalaya, as well as with a zone of intense microseismic activity (Figure 11) revealed from local seismic monitoring (Pandey et al., 1995, 1999; Cattin and Avouac, 2000). These geophysical data thus indicate that the MHT is a major thrust fault that can be traced nearly
continuously from the MFT, along the foothills, to beneath the high Himalaya and southern Tibet. Contrary to early views, the MCT, MBT, and MFT should not be seen as equivalent major thrust faults cutting through the whole crust, but are more likely splay faults rooting in a single mid-crustal decollement. The geometry of the MHT is characterized by a ramp-and-flat geometry, probably with two major ramps. One ramp is very shallow and corresponds to where the fault emerges with a dip angle of 30 at the surface along the MFT. The position of the other ramp is more conjectural, possibly lying at mid-crustal depths beneath the front of the high range and dipping northward by an estimated 15 .
6.09.3 Holocene Deformation and Erosion 6.09.3.1 Active Thrusting and Folding in the Sub-Himalaya The kinematics of thrusting along the front of the Nepal Himalaya can be interpreted from the study of deformed river terraces. The methodology is reviewed here, followed by a summary of results obtained from the study of terraces along the Bagmati river along section BB9 south of Katmandu basin (see location in Figure 7). For more details and results from other sections, the reader is referred to Lave´ and Avouac (2000).
Mountain Building: From Earthquakes to Geological Deformation
389
260
Elevation (m)
9.1 cal ky BP
T0 210
9.2 cal ky BP 6.1 cal ky BP
T1 2.2 cal ky BP erbed nt riv Prese
T3
160
110 0
–2
2
4
12 10 6 8 Distance from MFT (km)
MFT
14
16
18
20
22
MDT
Figure 11 Structural section with elevation of abandoned terraces along the Bagmati River across the Siwalik Hills south of Katmandu basin, approximately along AA9 section (see location in Figure 7). Also indicated are ages of terrace abandonment from C14 dating after calibration to calendar ages. The age of T3 (in italics) was indirectly inferred from the age of the same terrace level along the nearby Bakeya River. Modified from Lave´ J and Avouac JP (2000) Active folding of fluvial terraces across the Siwaliks Hills, Himalaya of central Nepal. Journal of Geophysical Research 105: 5735–5770.
6.09.3.1.1 Structural evolution of the sub-Himalaya
Several structural cross-sections were constructed along the Bagmati River and nearby rivers (Lave´ and Avouac, 2000). In this area, the Siwalik Hills are divided into twofold belts (Figures 7 and 11). The southern one is the topographic expression of the MFT fault, which reaches the surface, or close to it, along the front of the foothills (Nakata, 1989). The other fold belt, 15 km north of the MFT fold belt, is associated with the Main Dun Thrust (Mugnier et al., 1992) (Figure 11). The MFT fold belt makes a gently inclined monocline with dips varying between 25 and 50 to the NNW (Figure 11). At the front of the fold belt, Lower Siwalik clay- and siltstones are disrupted by meter-scale faults and folds. The section can be balanced easily using the fault-bend fold model (Suppe, 1985). The model assumes that deformation occurs by bed-parallel shear only, so that bed lengths and bed thicknesses are preserved during folding. The observed dip angles then imply a curved
ramp that roots into the decollement on top of the Indian basement (Figures 19 and 10). Restoration of the Bagmati and nearby Bakeya sections suggests a minimum shortening of 11 km due to slip on the MFT, and 12 km of further shortening due to slip on the MDT. The relief of the fold shows that 90% of the material uplifted along the MFT and MDT has been eroded away. So, on average over a geological period of a few million years, denudation must have balanced tectonic uplift. Although this particular section shows the MFT reaching the surface, it should be noted that the MFT is often blind. This is probably because before a fold develops into a fault-bend fold, there is a prior phase of fault-tip folding (Suppe and Medwedeff, 1990; Mitra, 2003). During this phase, the fault remains blind and shortening of the unfaulted sedimentary layer is probably accomplished solely by folding with some component of pure shear likely (as illustrated in Figure 12). As a result, deciphering the kinematic evolution of deformation from structural
390
Mountain Building: From Earthquakes to Geological Deformation
A′
BB′
(a)
t0
B’
B′
AA′ A
(b)
(c)
t1
t2
Figure 12 Diagram showing the formation, abandonment, and warping of fluvial terraces during fold growth. (a) Initially the river has enough stream power to bevel a wide channel at the rate imposed by the growing fold (tectonic uplift minus sedimentation rate at the front of the fold). At time t0, due to some decrease of its stream power (of possibly climate origin), the river starts entrenching into a narrower channel. (b) At time t1, a paired terrace, corresponding to the river channel at time t0 is preserved in the landscape (labeled T0). This terrace overhangs the active riverbed by a quantity that depends on the amount of tectonic uplift induced by the folding between t0 and t1, and on the change of base level due to sedimentation at front of the fold. (c) At time t2, two paired terraces, corresponding to the river channel at time t0 and t1, are preserved. Their geometries reflect cumulative folding since they were abondoned. Modified from Lave´ J and Avouac JP (2000) Active folding of fluvial terraces across the Siwaliks Hills, Himalaya of central Nepal. Journal of Geophysical Research 105: 5735–5770.
and geomorphic studies is challenging, and a more elaborate approach than the one outlined here is needed (Bernard et al., 2007; Daeron et al., 2007; Simoes et al., 2007). 6.09.3.1.2 River incision across the sub-Himalaya
Abandoned fluvial terraces are ubiquitous in the subHimalaya of central Nepal, and have been most extensively surveyed along the Bagmati, Bakeya, Narayani, and Ratu Rivers (Delcaillau, 1986; Lave´ and Avouac, 2000). These terraces are strath terraces, with the bedrock tread (often termed the ‘strath’ (Bull, 1991)) being overlain by a few meters of gravel very similar to the
bedload of the modern rivers. The gravel is generally overlain by overbank sands and silts with, in places, mixed colluvium or fanglomerates fed from adjacent valley slopes. The difference in elevation between the abandoned terrace and the present riverbed provides some direct estimate of the incision since the terrace abandonment. Charcoal samples collected from the various preserved terraces in the study area all yielded Holocene (<10 ka) 14C ages. The chronological data obtained in this way indicate four major episodes of terrace abandonment, dated to 9.2 0.15, 6.1 0.15, 3.7 0.1, and 2.2 0.2 ka (14C ages converted to calendar ages). These dates probably correspond to wet-to-dry
Mountain Building: From Earthquakes to Geological Deformation
climate transitions during the Holocene period, suggesting that these terraces are most probably of climatic origin. (Lave´ and Avouac, 2000). The uppermost level, labeled T0 in Figure 11, is generally the most prominent in the landscape because it corresponds to particularly wide terrace treads. The second most prominent level is T3, dated to 2.2 ka. Along the Bagmati River, the T2 terrace is not well preserved and could not be mapped and measured easily in the field. Figure 11 shows the elevation above the present riverbed of the three other terraces along the Bagmati River. These data provide an estimate of the average incision, i(x, y, t), at any point (x, y) along the river since the abandonment of terrace T at time t of incision: i ðx; y ; t Þ ¼ T ðx; y ; t Þ – r ðx; y Þ
½1
where r(x, y) is the present river profile and T(x, y, t) is the present elevation profile of the abandoned river terrace. 6.09.3.1.3 Converting inc ision rates to upl ift r ates in the s u b-Him ala y a
The geometry of the abandoned terraces in Figure 11 clearly demonstrates that since these terraces were abandoned, they have recorded some amount of folding and faulting associated with thrusting along the MFT. By contrast, the terraces are not affected by any clear deformation across the MDT. They also can be traced farther upstream across the MBT without any evidence of tectonic activity (Lave´ and Avouac, 2000). The determination of tectonic uplift relative to some chosen reference frame, here chosen to be the footwall of the MFT, requires some estimate of the geometry of the riverbed at the time it was deposited (Figure 12). It may simply be assumed that the river has maintained a constant profile during deformation, with river incision counterbalancing tectonic uplift. Accordingly, tectonic uplift since terrace abandonment would then be equal to incision, that is, the difference of elevation between the abandoned terrace and the present river, uðx; y ; t Þ ¼ i ðx; y ; t Þ
½2
This equation assumes that the term that arises from the apparent uplift of the riverbed due to the horizontal component of the displacement field can be neglected. This is a valid assumption because of the generally low stream gradient cutting across the
391
sub-Himalaya. In this case this term is less than 1% of the estimated uplift. Given that on average over many seismic cycles the fold might be assumed to grow more or less steadily, u(x, y, t) may be assumed to be proportional to the time elapsed since the terrace abandonment. It follows that the incision deduced from two terraces abandoned at different times, t and t9, should match: i ðx; y ; t Þ=t ¼ i ðx; y ; t 9Þ=t 9
½3
The incision profiles deduced from T0 and T3 along the Bagmati section are similar, but some mismatch on the northern limb of the fold, due to an incision deficit, was observed (Lave´ and Avouac, 2000). This suggests that the simple assumption of eqn [3] does not hold. Actually, the cumulative uplift since terrace deposition at time t with respect to a point attached to the footwall of the MFT ought to be written as uðx; y; t Þ ¼ i ðx; y; t Þ þ b ðx; y; t Þ
½4
where b(x, y, t) accounts for the change in the geometry of the riverbed with the sign convention illustrated in Figure 13. This term might vary along the section (something ignored in Figure 13) and should be written as b ðx; y; t Þ ¼ pðx; y; t Þ þ q ðx; y; t Þ
½5
where q(x, y, t) is the local base-level change since time t, and p(x, y, t) is the change in elevation at (x, y) due to possible changes of the river gradient and river sinuosity. The term q(x, y, t) is equal to the sedimentation rate (or incision rate) at the front of the MFT (Lave´ and Avouac, 2000). There are no uplifted Holocene terraces south of the MFT, and rivers clearly aggrade where they exit the Siwalik Hills. Based on the magnetostratigraphic sections (Appel et al., 1991; Harrison et al., 1993), the sedimentation rate can be estimated to 0.45 0.5 mm yr1. In the absence of any way to assess possible changes in the stream gradient during the Holocene, this parameter is assumed to be constant and equal to the present 0.27% value. It should be realized that the Holocene trend toward a drier climate in this part of Asia may have forced hydrological modifications, including gradient changes, but these are not easily analyzed. The paleo-channel corresponding to each terrace level can be estimated from the terrace remnants within the steep flanks of the canyon (Lave´ and Avouac, 2000). Study of these channels reveals that
392
Mountain Building: From Earthquakes to Geological Deformation
Elevation above present riverbed
Uplift relative to footwall since terrace abandonment Deposition since terrace abandonment (b > 0)
Incision since terrace abandonment i(x,t )
u(x,t )
0 b(t )
Area:A Folded abandoned fluvial terrace with age t
z2 z
d(z,t ) x z1
x2
x1
Figure 13 Sketch defining the relationship between fluvial incision, i, base-level change, b, and uplift relative to the footwall, u, since the abandonment of fluvial terrace across a growing anticline.
river sinuosity has increased during the Holocene. The river was least sinuous (s ¼ 1.8) at the time of T0 strath beveling. At the time of T1 abandonment, the sinuosity had increased to 2.0. At the time of T3 abandonment, it was 2.1. The present riverbed shows the narrowest and most sinuous channel (s ¼ 2.4). This sinuosity increase must have resulted in a gradual increase of the apparent slope of the river, and hence an incision deficit, on a projection perpendicular to the fold axis. To account for sinuosity changes, p(x, y, t) might be written as pðx; y ; t Þ ¼ s :Lðx; y ; t Þ
½6
where L (x, y, t) is the change in longitudinal distance along the river due to the sinuosity change since t, and s is the channel gradient, which is assumed to be constant along that particular river reach and unchanged with time. If the fold geometry is locally cylindrical and if the x-axis is perpendicular to the azimuth of the axial surface then the uplift rate is only a function of x. It should be emphasized that this section shows that it is possible to estimate tectonic uplift rates from river incision rates, but that these two rates generally differ due to possible base-level and sinuosity changes. In certain cases where the tectonic signal is not as large as in the Himalaya, climate-driven hydrological changes can be the main factor governing river incision (Poisson and Avouac, 2004).
The uplift rate profiles in Figure 14 were computed from eqn [4], taking sinuosity and base-level changes into account. The two curves derived from the two best preserved terraces along that section are identical within the error bars. Together with the uplift rate profiles deduced from other terrace levels along the Bagamati and Bakeya Rivers (Lave´ and Avouac, 2000), these data suggest that the anticline has been growing more or less steadily during the Holocene. The zone of active uplift is 1–14 km wide and the uplift rate reaches a maximum of 11 mm yr1. 6.09.3.1.4 Converting uplift rates to horizontal shortening from area balance
A simple way to deduce horizontal shortening from uplift profiles across the MFT fold is to assume conservation of mass, that is, conservation of area in cross-section if the effect of tectonic transport perpendicular to the section can be neglected (Molnar et al., 1994). Let A be the area between the present profile of an abandoned terrace tread and the profile of the same terrace level at the time, t, it was abandoned (Figure 15). Let h be the thickness of the units detached from the basement. Assuming plane strain and conservation of mass, we derive the mean horizontal displacement, d, necessary to account for the deformation of the terrace: Z d ðt Þ ¼
A ðt Þ ¼ h
x
uðx; t Þdx 0
h
½7
Mountain Building: From Earthquakes to Geological Deformation
12
T3
11 10 Uplift rate (mm yr–1)
393
21 mm yr –1sin θ
9 8
T0
7 6 5 4 3 2 1 0 –2 –1
0
1
2
3
4 5 6 7 8 9 10 11 12 13 14 15 Distance from MFT (km)
Figure 14 Uplift rate along the Bagmati section deduced from terrace T0 and T3 after correction for 0.45 mm yr1 base level change due to sedimentation south of the MFT and sinuosity changes. Dashed line shows the predicted uplift rate pattern assuming fault-bend folding. For a cylindrical fold, it would correspond to U(x, y) ¼ d(t) sin (x, y). It yields a good fit to the observed uplift rates for a horizontal shortening rate across the MFT of 21 mm yr1.
(a)
x
A(t ) = d (t ) × h = ∫o u(x,t ) dx
Mass conservation u
h
d
(b) MFT
d
θ
u
d
Figure 15 (a) Relationship between horizontal shortening and uplift assuming conservation of area. A is the area between the abandoned terrace tread and the profile of the riverbed at the time t it was abandoned. h is the thickness of the units detached from the basement. Conservation of area implies that the cumulative shortening, d, since terrace abandonment is d ¼ A/h. (b) Relationship between horizontal shortening and uplift rates, assuming fault-bend folding. At any point with abscissa x along the section, the fault-bend fold model predicts that local dip angle , and local cumulative uplift, u, should relate to the amount of bed-parallel displacement, d, of the bed out-cropping at that point according to u ¼ d sin . The sketch assumes no bed-parallel shear above the decollement, implying that all beds have experienced the same displacement d. It should be noticed that the relation still holds if d varies with depth.
Mountain Building: From Earthquakes to Geological Deformation
This approach requires a continuous terrace tread across the growing fold. The only such terrace is the T0 level along the Bagmati River. Deformation of that level corresponds to an area of A(T0) ¼ 1.05 0.25 km2. According to the structural description above, the depth of the basement is 5.0 0.3 km beneath the MFT, and it dips northward with a slope of 2.9 (corresponding to a grade of 2.5%). Given that the MFT ramp merges with the decollement 18 km north of its trace at the surface, the depth to the decollement beneath the fold is 5.5 0.3 km. The accumulated shortening since 9.2 0.2 cal ky is then estimated to be d ¼ 192 (þ60/50) m, yielding a shortening rate of 21 (þ7/–6) mm yr1.
6.09.3.1.5 Convert ing up lift rates to hori zont al sh orte ni ng fr om t he f au lt-be nd fold model
Figure 14 suggests that the uplift profiles recorded by the Holocene terraces are closely related to the geological structure of the fold. Indeed, the small variation of the uplift rate within 1 km of the middle of the fold correlates with a variation in bedding dip angles, and thus of the dip of the ramp at depth. The correlation between structural geology and recent uplift is in fact expected for a fault-bend fold. According to this model, the hanging wall accommodates deformation imposed by the fault geometry at depth via bedding plane slip, with no change of the length and width of the beds (Figure 14). It follows that, on the back limb, uplift rate depends on the dip of the fault at depth, which equals the local bedding dip angle. In the case of a fold with cylindrical geometry (with beds striking perpendicular to the section, taken to be parallel to the y axis), assuming that the slip increment since terrace abandonment is small, one may write U ðx; t Þ ¼ dðt Þ sin ðx Þ
½8
where (x) is the dip angle at point (x, y), and d(t) is the displacement measured parallel to the decollement outside the fold zone of the bed outcropping at point (x, y) (again the fold geometry is assumed cylindrical with the axis x being perpendicular to the fold’s strike). The equation holds insofar as the dip angle does not vary significantly over a distance of d, that is, 1=d >>
dðsin Þ dx
The fault-bend model leads to a testable correlation between structural geology and terrace warping. The uplift profiles deduced from terraces T0 and T3 along the Bagmati River are in close agreement with the bedding dip angles, as expected from eqn [9] (Figure 14). This model makes it possible to estimate the cumulative shortening since the abandonment of each terrace by least-squares adjustment of the model to the uplift profiles, with d (t) being the only adjustable variable. Each abandoned terrace yields a point on the plot in Figure 16, which shows d as a function of elapsed time since the abandonment of each terrace. The points are seen to follow a linear trend. A least-squares fit yields 20.4 1 mm yr1, which may be taken to represent the rate of fault slip on the MFT, or alternatively the horizontal shortening rate across the fold over the Holocene period, since the decollement is nearly horizontal (within a few degrees) according the structural model. This rate is much better constrained than the estimate deduced from conservation of mass in the previous paragraph. The 1 mm yr1 uncertainty neglects the uncertainty related to the applicability of the fault-bend fold model, because this source of 250 Holocene shortening across the MFT
Horizontal shortening (m)
394
200
–1
m
yr
T0
m .5 ±1
21
150
T1
100 T2 Bagmati terraces
50 T3
Bakeya terraces
0 0
1
2
3
4
5 6 7 8 Terrace age (ka)
9
10 11 12
Figure 16 Plot of horizontal shortening deduced from the various terrace treads along the Bagmati and Bakeya section as a function of age of terrace abandonment (Lave´ and Avouac, 2000). These data are consistent with a uniform 21 1 mm yr1 shortening rate over the Holocene period. If the fold is assumed to have been growing by incremental deformation during large earthquakes, the mean shortening rate is slightly modified since some amount of elastic straining at the large scale may yet remain to be released, or may not have been released by the time of emplacement of the different terraces. Assuming that these increments can be as large as 5 m of horizontal shortening and that all values between 0 and 5 m are equiprobable, we obtain a long-term averaged shortening rate, offering a best fit to the terrace record, of 21.5 1.5 mm yr1.
Mountain Building: From Earthquakes to Geological Deformation
uncertainty is difficult to incorporate formally. In addition, this estimate ignores the fact that the slip along the MFT is not linear as a function of time, because it has most probably resulted from recurring coseismic slip events. When this stick-slip behavior is taken into account, assuming slip events between 4 and 6 m (with a uniform probability in this range), the long-term slip rate is revised to 21 1.5 mm yr1 (Lave´ and Avouac, 2001). 6.09.3.2 River incision, Erosion, and Uplift across the Range 6.09.3.2.1 Fluvial incision across the whole range
Fluvial incision rate (mm yr–1)
In the previous section we saw that the pattern of river incision could be used to determine tectonic uplift in the sub-Himalaya. The same approach might be applied at the scale of the whole range, although there are several limiting factors. In the LH, strath terraces are sparse. There are, however, prominent fill terraces greater than
395
100 m thick, and probably of Pleistocene age (Iwata, 1976; Fort et al., 1983). Close to the front of the high range, the valleys become steep and narrow and the terraces are no longer preserved. Such a terrace pattern suggests a slow rate of fluvial incision in the LH and an abrupt increase at the front of the high range (Lave´ and Avouac, 2001), as first inferred from the observation that most Himalayan rivers have developed knickpoints where they cross the front of the high range (Seeber and Gornitz, 1983). Due to the difficulty of using the terrace record to determine incision rates, another approach has been adopted. The principle of the approach is that fluvial incision rate can be estimated from the shear stress exerted by the flowing water at the bottom of the channel (Lave´ and Avouac, 2001). The model was calibrated from a few sites in the LH and subHimalaya, where Holocene strath terraces could be dated. This approach has been shown to yield results consistent with the terrace record of river incision at sites where both approaches could be applied. Figure 17 shows the average profiles obtained from river incision
8 6 4 2
KTM11
Elevation (km)
0 6 4 2 0 MFT
MBT
Katmandu
MCT
Depth (km)
0 20
MHT
40 60 0
100 Distance form MFT (km)
200
Figure 17 Rate of river incision across the Himalaya of central Nepal (Lave´ and Avouac, 2001). Yellow shading shows river incision across the Siwalik Hills deduced from three Holocene terraces along the Bagmati River. River incision across the LH (gray shading) was derived from the computation of river incision along the major rivers of central Nepal based on their present hydrological regime. Also shown are erosion rates over the last about 1 Ma derived from fission tracks data from the Annapurna area (Burbank et al., 2003). These data were converted into erosion rates using the thermal model of Henry et al. (1997) which is close to the more recent thermal model KTM11 (Bollinger et al., 2006). Also shown is the surface erosion pattern used in the thermokinematic model KTM11 between 20 Ma and 8 Ma (dashed line) and since 8 Ma (continuous line).
396
Mountain Building: From Earthquakes to Geological Deformation
estimated along six major rivers running across the central Nepal Himalaya. It turns out that river incision in the LH does not exceed a few millimeters per year, and it increases abruptly at the front of the high range to reach values up to 8 mm yr1. The zone of rapid river entrenchment is only 50 km wide, and it coincides with the front of High Himalaya. 6.09.3.2.2 range
Denudation across the whole
In the sub-Himalaya, the rivers cutting across the active folds incise at approximately the rate of tectonic uplift (minus the sedimentation rate, which is only a minor term in this case). Since the present topography accounts for less than 10% of the total volume of rock uplifted by thrusting since fold inception, this implies that denudation must also be in equilibrium with tectonic uplift (Hurtrez et al., 1999). Current denudation rates are estimated for the 11 catchments within the Lesser and Higher central Nepal Himalaya for which measurements of suspended load are available (Lave´ and Avouac, 2001). These rates were found to be smaller by 20% than estimates assuming that denudation equals river incision rates along the main drainages in each watershed. This apparently large misfit is insignificant in view of the large uncertainties in both the measurements of suspended load and estimates of sediment discharge, and may partly be offset by the contribution of the bedload. Mass processes are probably the dominant erosional process in the Himalaya of Nepal, especially in the high range, where the bare, steep hillslopes often show evidence of recent landslides. As has been suggested for the High Himalaya of northern Pakistan, hillslopes in Nepal are probably near their critical slopes for mass movement (Burbank et al., 1996). In this setting, fluvial downcutting is probably the ratelimiting process driving hillslope erosion, and denudation is expected to equal fluvial incision on average. Denudation rates in the Nepal Himalaya can also be compared with the sediment volume delivered to the foreland basins and to the Bengal Fan. The pattern of fluvial incision of Figure 17 predicts a sediment flux of 375 km2 My. This rate may be compared with the (1 0.5) 106 m3 yr1 rate proposed by Metivier et al. (1999) to be eroded from the Himalaya over the last two million years and deposited in the Gangetic Plain and Bengal Fan. Given that the Ganges and Brahmaputra Rivers drain the Himalaya over an 1800-km-long stretch, this value implies 280 110 km2 My1 of denudation on
average across a section of the range. To first order, this suggests that denudation of the whole landscape matches the pace of fluvial downcutting, and that fluxes have not varied dramatically over the Quaternary period. 6.09.3.2.3 Holocene kinematics of overthrusting along the MFT–MHT
The average slip rate on the MFT over the Holocene appears to be very close to the total shortening rate across the range, as measured from GPS. It also compares well with geological estimates of the shortening rate across the LH of Nepal, which range from 16 to 25 mm yr1 (Hauck et al., 1998; DeCelles et al., 2001). This estimate also compares well with the Quaternary shortening rate across the Himalaya deduced from E–W extension in southern Tibet. If this Quaternary extension is assumed to accommodate the lateral variation of the direction of thrusting along the arcuate Himalayan front (Baranowski et al., 1984; Armijo et al., 1986; McCaffrey and Nabelek, 1998), the estimated extension rate of about 10 mm yr1 implies a Late Quaternary shortening rate of 20 10 mm yr1 (Armijo et al., 1986). The observed pattern of erosion is found to closely mimic the uplift predicted by slip along the flat-ramp-flat geometry of the MHT fault. The theoretical pattern of uplift can be simply estimated by assuming that the hanging wall is thrust along the MHT and accommodates the MHT geometry simply by vertical shear (Molnar, 1987). If the topography is approximately steady state, as argued in the previous section, the flexural loading of the Indian Plate does not vary and this model does not require any computation of the flexural response of the lithosphere. This approach predicts an uplift pattern that fits reasonably with the fluvial incision pattern of Figure 17 (Lave´ and Avouac, 2001) as discussed later in this chapter (Figure 32). It thus seems that over the Holocene period, all the shortening across the Himalaya has been absorbed by slip along one major thrust fault.
6.09.4 Longer-Term Geological Deformation and Exhumation We now compare the kinematics of deformation established for the Holocene period with the kinematics of deformation and exhumation over a longer geological period. We review first how these kinematics can be inferred by combining (1) the history of
Mountain Building: From Earthquakes to Geological Deformation
deposition in the foreland, (2) the structural architecture of the Himalayan internal zones, and (3) the cooling histories and metamorphic grades in the Lesser and High Himalaya.
forming a ‘foreland flexural basin’ in which a fraction of the material eroded from the range accumulates with a stratigraphic organization that depends on the kinematics of overthrusting (Lyon-Caen and Molnar, 1985; Lave´, 1997) (Figure 18). Several kilometers of Cenozoic molasse sediments (the Murrees and Siwaliks formations) have thus accumulated on the Precambrian Indian basement. These molasse formations crop out along the foothills, where they were scraped off the basement and folded (Figure 19). They consist of an upward coarsening sequence
6.09.4.1 Foreland Deposition: A Record of Underthrusting Sedimentation in the foreland basin provides indirect information on orogenic growth. As the Himalayan wedge grows, it overthrusts the Indian basement, Age of oldest sediments over basement
397
1 Slope: — Vpr Proximal facies Distal facies Distance from MFT
Vpr
V2 Vpr
x3
x2
α
W
x1 V0
Moho
Figure 18 Sketch showing how sediments prograde over the flexed Indian basement during overthrusting and Himalayan wedge growth. The flexed Indian lithosphere makes a forebulge at about 250–300 km from the Himalayan front where most recent foreland sediments are starting to overlap on the Indian basement (x3). The sediment progradation rate, Vpr, can be estimated by plotting the ages (t1, t2) of the oldest foreland sediments overlying the basement as a function of distance (x1, x2) perpendicular to the range front. Assuming that the foreland’s geometry is entirely constrained by loading due the weight of the Himalayan topography, Vpr depends on the rate of convergence with the front of the high range, VHR, and on the growth of the mountain wedge. W is the width of the orogenic wedge.
Δ x1
Δ x2
h
3
3
h* 2
2
X
1
1
MFT
Figure 19 Sketch showing how dated outcropping sections in the sub-Himalaya can be used to infer horizontal displacement of the foreland due to underthrusting. The present thickness of unit 1, h, represents, after decompaction, the depth of the foreland basin, h, at time of the onset of deposition of unit 2, t12. If the geometry of the basin is assumed to have remained stationary, this section was at that time at a distance (x1þx2) from its present position. We can then estimate by how much the Indian Plate has underthrust the Himalayan topography since time t12.
398
Mountain Building: From Earthquakes to Geological Deformation
magnetostratigraphy (see location in Figure 3). If the geometry of the foreland basin is assumed to be steady state, the stratigraphic thickness, h, of the foreland sediments deposited before t12 can be used to estimate the distance at time t12 of that section from the front of the high range. As shown in Figure 19, this estimate requires some restoration of slip along the MFT. The positions of the transitions from both the Lower to the Middle Siwalik and from the Middle to the Upper Siwalik were used. Altogether, the plot in Figure 20 indicates a rate of progradation, Vpr, of 15 mm yr1 over the last 15 My. Assuming that the geometries of the Himalayan wedge and of the flexed Indian Plate have not changed with time, as initially proposed by LyonCaen and Molnar (1985), Vpr might be taken as equal to the overthrusting rate V0 (the convergence rate between India and southern Tibet) (Figure 20). This reasoning, however, ignores retreat or advance of the mountain front as a result of erosion or of crustal thickening. More recently, Molnar (1987) proposed that internal deformation of the crustal wedge would account for 5 mm yr1 of additional shortening, and revised the thrusting rate to be 15– 20 mm yr–1. A more general formulation should account for internal deformation of the range as well as for
(claystone, siltstone, sandstone, and conglomerates) of upper Miocene to Pleistocene age. As shown in Figure 18, the development of the flexural basin during underthrusting implies some progradation of sedimentary facies and of the contact between the basement and the most distal sediments away from the mountain front. This model implies, as observed, an upward gradation from distal to more proximal facies. If the age of the oldest sediments overlying the preTertiary basement is plotted as a function of the distance perpendicular to the range, the rate of sediment progradation can be deduced (Figure 20). The southward progradation over the last 15 My was estimated with this method to be between 10 and 15 mm yr1 (Lyon-Caen and Molnar, 1985), based on a compilation of all well data in the Gangetic foreland. In Figure 20, only the well data close to central Nepal area were considered. This plot is poorly constrained because the age of the youngest sediments overlying the basement was estimated with very few chronological constraints and with the assumption of constant sedimentation rates (Lyon-Caen and Molnar, 1985). Two additional points were added by Lave´ (1997) based on the stratigraphic sections along the Surai Khola (Appel et al., 1991) and along the Bakeya (Harrison et al., 1993); these were both dated with
Raxaul Mohand
25
20 –1
–1
Age (My)
m
10
15
yr
m
m 5m
yr
1
–1
r
Tihar
my 20 m
10 Kasganj Ujhani 5
0 0 Forebulge
100 200 Horizontal distance (km)
300 High range
Figure 20 Age of oldest Cenozoic foreland deposits overlying the Indian basement as a function of their present distance from the front of the high range (defined from the 3500 m elevation contour line shown in Figure 3). See Figure 2 for location of wells (modified from Lyon-Caen and Molnar, 1985). Two additional points were added to this plot based on the stratigraphic sections along the Surai Khola (Appel et al., 1991) and along the Bakeya (Harrison et al., 1993), as illustrated in Figure 19 (modified from Lave´, 1997).
Mountain Building: From Earthquakes to Geological Deformation
possible erosional retreat of the front of the high range. This is shown in Figure 18 to represent the edge of the overriding load that flexes down the plate ((Avouac, 2003). We then write Vpr ¼ V1 þ
dW dt
½9
where V1 represents the velocity of the front of the high range and W is the width of the accretionary prism (Figure 18). It is assumed that the center of mass of the topographic load flexing the Indian basement depends on the position of the crest of the Himalayan wedge, the width and shape of the orogenic wedge, and the partial support of topography by the buoyant crustal root. For this reason, is a geometric factor between 0 and 1. The rate at which the Indian basement underthrusts the high range is V1. This velocity is hereafter termed the ‘underthrusting rate’. It equals the rate of sediment progradation, Vpr, in the case of an orogenic wedge with constant width. The retreat, if positive, or advance, if negative, of the front of the high range is then V2 ¼ V0 – V1
½10
This velocity represents the rate at which the hanging wall rides over the MHT and is termed here the ‘overthrusting’ rate. This term can be large, and it could have varied during Himalayan orogenesis. In the absence of any crustal thickening or isostatic response, if eroded at a vertical rate of e, the mountain front would apparently retreat by V2 ¼
e tan
½11
The denudation rate in the High Himalaya is estimated to be 4–8 mm yr1 (Lave´ and Avouac, 2001), and the gradient at the front of the high range is 10%. These figures would yield a very rapid retreat of as much as 40–80 mm yr1. In reality, erosion is compensated by some uplift, u, due to isostatic response and crustal thickening, so that the apparent retreat is less and should be written, V2 ¼
e –u tan
½12
Some estimate of V2 is therefore needed, in addition to the description of the internal deformation of the wedge, to fully describe the kinematics of mountain building and its relation to rates of sediment progradation and deposition in the foreland.
399
6.09.4.2 Structural Evolution of the Thrust Package The structural architecture of the Himalaya can be used to infer the kinematics of deformation through the technique of palinpastic restoration (Dahlstrom, 1969). The chronology can be constrained from cross-cutting relationships, and provenance data from the molasse (Siwalik and Dumri formations) and thermochronology. A complete discussion of this issue is beyond the scope of this chapter, but a few key results are reported here. Compressional deformation north of the Himalaya started in the early ‘Alpine’ time of the collision, and there is evidence that crustal thickening in southern Tibet persisted until mid-Miocene time (e.g., (Ratschbacher et al., 1994; Yin et al., 1999)). Farther south, deformation and anatexis in the MCT zone occurred at c. 22 Ma (Copeland et al., 1991; Hodges et al., 1996), an age consistent with a slightly younger cooling in the rocks of the hanging wall attributed to emplacement along the MCT (Copeland et al., 1991; Hodges et al., 1996). Such a timing is also consistent with conventional and isotopic provenance data from the Dumri formation that indicate an early Miocene age for the onset of erosion of the High Himalayan crystalline rocks (Robinson, et al., 2001). Out-of-sequence reactivation of the MCT may have taken place by late Miocene to Early Pliocene time, as indicated from monazite ages of rocks exhumed from mid-crustal depths (Harrison et al., 1997; Catlos et al., 2001). This reactivation has been proposed to be responsible for the inverse thermal gradient and for the steep morphologic front of the High Himalayan range (Harrison et al., 1997, 1998). However, there is no direct, reliable estimate of the timing of thrusting on the MBT. Age estimates and the provenance of Siwalik sandstones in northern India, west of our study area, indicate that exhumation of the LH commenced some 10 My ago (Meigs et al., 1995). This observation was interpreted to indicate motion on the MBT began as early as 10 Ma (Meigs et al., 1995; Huyghe et al., 2001). Alternatively, exhumation of LH rocks could have resulted from the development of the LH duplex, requiring a much younger (possibly 5 Ma) activation of the MBT (DeCelles, et al., 1998; Robinson, et al., 2001). Finally, deformation in the sub-Himalaya due to motion on the MFT and the various thrust faults between the MFT and the MBT has proceeded from mid-Miocene time to the present. The chronological constraints discussed above, when considered with various restored balanced cross-sections across the central Himalaya (Schelling
400
Mountain Building: From Earthquakes to Geological Deformation
(c) STD
MFT DT RT
MBT
MCT
5–0 Ma
(b)
STD RT
DT
MCT
15–6 Ma
(a) MCT
45–16 Ma
Tibetan Himalayan zone
Lesser Himalayan zone
Greater Himalayan zone
Synorogenic foreland-basin deposits
Figure 21 Schematic model of the structural evolution of the Himalayan orogen. This scheme is based mainly on observations along a section across the Himalaya of western Nepal. MCT, Main Central Thrust; DT, Dadeldhura Thrust; RT, Ramgarh Thrust; MBT, Main Boundary Thrust; MFT, Main Frontal Thrust; STD, Southern Tibetan Detachment. 45–16 Ma: Emplacement of the MCT and then of the DT crystalline thrust sheets, coeval with motion on the STD. Deposition of the Dumri and Bhainskati foreland formations. 15–6 Ma: Emplacement of the Lesser Himalayan RT thrust sheet and development of the duplex; deposition of the Lower and Middle Siwaliks. 5–0 Ma: Motion on the MBT and MFT; deposition of the Upper Siwaliks; major phase of exhumation of the LH duplex. Modified from Robinson DM, DeCelles PG, Patchett PJ, and Garzione CN (2001) The kinematic evolution of the Nepalese Himalaya interpreted from Nd isotope. Earth and Planetary Science Letters 192: 507–521.
and Arita, 1991; Srivastava and Mitra, 1994; DeCelles et al., 2001), provide some idea of the kinematics over the last 20 My. Figure 21 presents the kinematic model proposed for far western Nepal (Robinson, et al., 2001). This model also applies, with some minor variations, to central Nepal (Robinson et al., 2003; Bollinger et al., 2004a). Over the last 15 Ma, the Himalayan orogen has developed as a result of frontal accretion at the toe, internal deformation of the wedge, and underplating. Underplating, associated with duplex formation of the LH units at mid-crustal depths, has likely been the dominant process contributing to the volume of the Himalayan wedge since Middle Miocene time.
6.09.4.3 Exhumation of the Lesser and High Himalaya: A Record of Overthrusting Exhumation rates across the range can be used to determine the tectonic evolution of the orogenic wedge. The reader is referred to Bollinger et al. (2004a, 2006) for a more detailed discussion of the approach outlined here. The most complete picture of the chronology of exhumation in the area comes from 39Ar/40Ar ages of muscovite (Figure 22). This technique provides an estimate of the age of a rock sample when it cooled through the muscovite blocking temperature, 350 C. Several authors, in particular Peter Copeland and his colleagues, have analyzed samples
Mountain Building: From Earthquakes to Geological Deformation
401
30
Age (Ma)
25 Slope: 0.23 My km–1
20 15 10 5 0
0
20
40 60 Distance from MBT (km)
80
100
Figure 22 Cooling ages in Central Nepal as a function of distance from MBT along an N18 E section. All samples collected from the LH and the overlying crystalline thrust sheets are reported. The Ar39/Ar40 muscovite ages appear to approximately follow a linear trend with ages increasing southward by about 0.2 My km1. Data from Bollinger et al., (2004a) from Macfarlane et al. (1992); Arita et al. (1997); Arita and Ganzawa (1997); Copeland (1997), and Copeland (pers. comm. (1999) as reported in Bollinger et al. (2004a)). The origin is taken at the MBT. All samples collected from the LH and the overlying crystalline thrust sheets are reported.
Time since cooling below 350°C
collected along the central Nepal section from the high range to the southern edge of the Katmandu Klippe (Figure 22). These data indicate that the trailing edge of the Katmandu thrust sheet cooled below 350 C at c. 20–22 Ma, an age consistent with the onset of deformation and anatexis in the MCT zone. Ages follow a nearly linear trend corresponding to a slope of about 0.23 My km1, gradually decreasing northwards to 5 Ma at the front of the high range (Figure 22). As a first step, we may ignore
accretion and assume steady-state topography and thermal structure defining the reference frame used to describe the kinematics. Given the shallow dip angle of the MHT below the LH and the likelihood of a steep 350 C isotherm, the rocks now outcropping along a section perpendicular to the range must have all crossed the isotherm near the same location (Figure 23). The cooling ages should then follow a linear trend, as observed in Figure 23 (Avouac, 1 Slope: — V2
350°C V1
MHT
V2
Figure 23 Kinematic model for a convergence rate V0, an overthrusting rate V0 VHR, and an underthrusting at rate VHR. The topography and thermal structure are assumed to be in a steady state. This kinematics predicts that cooling ages (for Ar/ Ar dating on muscovites for example) along a section perpendicular to the range should follow approximately a linear trend with an age gradient of 1/V0 VHR. The isotherm corresponding to the Muscovite-blocking temperature for 39Ar/40Ar is estimated at 350 C.
402
Mountain Building: From Earthquakes to Geological Deformation
2003). The slope of the age–distance relationship thus provides an estimate of the rate of overthrusting of the hanging wall, which is the inverse of the age gradient:
depths (Figure 24). The particle path trajectories shown in Figure 25 assume a continuous process of accretion (Avouac, 2003). The top of the underthrusting plate is accreted, while the lower part underthrusts the orogenic wedge. In reality, the process of accretion might be discontinuous. When the overthrusting rate derived from the thermochronological data, 4.3 mm yr1, is added to the rate of underthrusting determined from the sediment progradation rate in the foreland, 15 mm yr1, we obtain an estimate of the shortening rate across the whole range, V0. The shortening rate is thus averaged to be 19.3 mm yr1 over the last 15 Ma. The uncertainty associated with that rate is difficult to estimate because of the number of assumptions made to reach it. It could be as high as 30%. The kinematics may be compared with two endmember cases of a steady-state regime without accretion:
V2 ¼ V0 – V1 4:3 mm yr – 1
6.09.4.4 Overthrusting, Underthrusting, and Accretion In the previous section the development of the foreland basin, the structural architecture of the internal zones, and the geochronological data required that (1) the footwall underthrusts the MHT (Figure 23), (2) part of the footwall be accreted to the hanging wall, and (3) that the wedge itself overthrusts the MHT and is eroded (Figure 24). The kinematics of shortening across a mountain range should thus generally be described in terms of ‘underthrusting’, ‘overthrusting’, and ‘accretion’. The kinematics in Figure 23 might be modified to fit the structural evolution model of Figure 21 by assuming that some accretion occurs by underplating due to the development of a duplex at mid-crustal
1. If the hanging wall is assumed to overthrust an undeformable footwall by 20 mm yr1, there would be no sediment progradation and no
MCT V2
V1
Zone
of ac
MHT
cretio
n
Figure 24 Schematic model of duplexing in the Lesser Himalaya. Thick line shows the geometry of the active thrust fault at a time just before it migrates to a new position farther south (dashed line). This migration induces accretion of material from the footwall to the hanging wall. Thin dashed line show the geometry of other thrust sheets accreted before time t. The position of the mid-crustal ramp, the ‘accretion window’, is taken as a reference to define velocities Vpr and V0 VHR, assuming that it controls the topography of the range and the load flexing the Indian lithosphere.
A B
350°C C D
Figure 25 Particle trajectories expected from the model of duplexation shown in Figure 24, assuming that ramp migration is a continuous process. Dots show approximate positions about every five million years. This model also predicts diachronic exhumation with an age gradient of 1/V0 VHR.
Mountain Building: From Earthquakes to Geological Deformation
subsidence of the foreland, and we should observe a diachronous exhumation corresponding to V2 20 mm yr – 1 Vpr ¼ V1 0 mm yr – 1
2. In the opposite case, if the convergence is assumed to be entirely accommodated by underthrusting of the Indian basement below a rigid hanging wall without any erosion at the surface (a ‘subduction zone’ kinematics), we get V2 0 mm yr – 1 Vpr ¼ V1 20 mm yr – 1
The simple kinematics obtained from the analysis described here lies between these two end members and reconciles the rate of sediment progradation and subsidence of the foreland with the retreat of the mountain front driven by erosion. When comparing this to the Holocene kinematics described in the previous section we conclude that the long-term shortening rate is approximately equal to the Holocene shortening rate. The rate of accretion and the partitioning of Holocene shortening into underthrusting and overthrusting are more difficult to assess. However, the pattern of exhumation provides some keys. These issues will be addressed in the modeling section, in which it is suggested that the partitioning of the convergence for Holocene kinematics is approximately the same as over the long term.
6.09.5 Kinematic and Mechanical Models of Crustal Deformation The data reviewed in the previous sections suggest that the shortening rate across the Himalaya determined from geodetic measurements (19 2.5 mm yr1; Bettinelli et al., 2006) and the averaged rate over the last 15 Ma (20 mm yr1) are remarkably close and equal to the Holocene slip rate on the MHT (21 1.5 mm yr1; Lave´ and Avouac, 2000). These findings suggest that the Himalayan wedge has probably suffered little internal shortening, at least over the Holocene period, and that the shortening rate across the range has not varied much over the last 15 Ma. This section will describe numerical models consistent with these two key hypotheses. To start, an outline of the thermokinematic model of the long-term deformation of Bollinger et al. (2006) is provided. In this approach, the kinematics is prescribed, the thermal
403
evolution is computed, and the model parameters are adjusted by trial-and-error until they fit the observations. This approach provides a way to test more rigorously, and to refine, the kinematic model of Section 6.09.4.4. The next section, describes how the Holocene kinematics can be reproduced from a mechanical model which accounts for both erosion and the effect of temperature on the rheology of the crust (e.g., Cattin and Avouac (2000)). Attempts at reproducing the long-term evolution of a mountain range from thermomechanical modeling are not reviewed here because this work is more relevant to the geological and geophysical data reviewed in the previous section. Readers interested in this kind of modeling – in which the kinematics of deformation is computed from the force balances, assuming some temperature-dependent rheology and that the mechanical and thermal evolutions are coupled – are referred for examples to Beaumont et al. (2004) and Toussaint et al. (2004a, 2004b). 6.09.5.1 Thermokinematic Model of the Evolution of the Range since 15 Ma In this section is briefly described a thermokinematic model that assumes that shortening across the range has been entirely taken up by slip along the MHT, and that the growth of the Himalayan wedge has resulted from underplating only (model KTM11 of Bollinger et al. (2006)). A finite-element model has been used to solve the heat advection–diffusion equation in transient state (Zienkiewicz and Taylor, 1989). The time-varying temperature field is thus computed from the assumed velocity field, boundary conditions, and material properties. The! velocity field satisfies the continuity ! equation ð ? n ¼ 0Þ, which implies continuity of the velocity normal to the fault along the underplating window. Vertical shear along hinge lines is allowed to accommodate variations in accretion rate (Figure 26). Peak T and cooling ages are then computed by particle tracking through the time-dependent temperature field. The thermal structure is computed based on the kinematics sketched in Figure 26, taking into account the effects of erosion, accretion, and shear heating (assuming a low friction on the MHT, as argued below). The topography is assumed to be steady state. For simplicity, the possible flat and ramp geometry is ignored, and the MHT is given a constant dip angle of 10 . A zone of underplating allows transfer of material from the footwall to the hanging wall (Figure 26). The thermal parameters,
404
Mountain Building: From Earthquakes to Geological Deformation
Surface temperature = 0°C
Depth (km)
Equilibrated geotherm
1.0 N W
MHT Zone of
V1
m –3
V2 underpla
ting
2.5 N W m –3
2.5 N W m –3
40
0 km
Moho
0.4 μ W m –3
80 km
120 km 150 km
V1
No conductive heat flow
(x = 300 km)
(x = 50 km) 0
Figure 26 Model geometry, kinematic and thermal parameters used for modeling cooling history corresponding to model KTM11 of Bollinger et al. (2006). Shear stress along the MHT is assumed to be the minimum between the ductile and brittle shear stresses. We assume an effective friction ¼ 0.1, and ductile stresses at the ductile shear zone is taken to be 100 m and obey a Quartz-type flow law (Henry et al., 1997). In the model, shear stresses do not exceed 50 MPa, values consistent with constraints obtained by comparing seismicity and current geodetic strain on maximum deviatoric stresses (Bollinger et al., 2004b). The reader is referred to Bollinger et al. (2006) for more details regarding the modeling.
also indicated in Figure 26, were taken from Henry et al. (1997). The sum V1 þ V2, which represents the convergence rate across the range, V0, was chosen to match the 21 mm yr1 Holocene shortening rate. As inferred from the pattern of 39Ar/40Ar ages, the overthrusting rate was chosen to be V2 ¼ 4 mm yr1, implying an underthrusting rate of V1 ¼ 17 mm yr1. Because the model assumes steady-state topography, the underthrusting rate V1 should be equal to the rate of progradation of sediments in the foreland. The position of the window of underplating was adjusted to fit the position and shape of the accreted LH units and to correspond with peak metamorphic temperatures in the 300–550 C range of observed values in the LH. The position of the window and the rate of underplating determine the gradient of peak metamorphic temperatures and the geometry of the accreted units. Because the topography is assumed to be steady state, they also determine the pattern of erosion at the surface. The metamorphic gradient observed near the front of the high range (Figure 27(a)), where peak metamorphism in the MCT zone was reached over the last 5–10 Ma, is somewhat different from that observed below the klippes farther south (Figure 27(b)), where peak metamorphism was reached c. 15 Ma. The history of underplating was adjusted to fit these different gradients and to meet the structural constraint that the LH duplex developed over the last c. 8 Ma (DeCelles et al., 2001; Robinson et al., 2001) (Figure 21). Accordingly, the model assumes an increase of rate of accretion via underplating at c. 8 Ma.
The model that fits best the data, KTM11, was obtained by trial and error (Bollinger et al., 2006). This model assumes an erosion pattern grossly consistent with that derived from river incision across the Himalaya although rates are somewhat smaller (Figure 17). The model predicts a cooling history consistent with the various thermochrological data, including data from lower-temperature thermochronological techniques than the 39Ar/40Ar data described above (Figure 28). These data are relatively well distributed both in time (spanning the last 20 My) and in space along the section. The predicted peak temperatures also fit the observations reasonably well (Figure 27). It should be noted that the model predicts peak temperatures lower than 350 C in the core of the anticlinorium at structural distances as short as 5 km from the top of the LH; this prediction is consistent with the report that muscovites were not reset for 39Ar/40Ar dating at this distance from the MCT (Wobus et al., 2005; Bollinger et al., 2006). The model thus accounts for the inverted metamorphic gradient and is consistent with the thermochronological data available to date. It also reconciles the kinematics of recent crustal deformation deduced from morphotectonic and geodetic studies with the evolution of the range over the last 20 million years. According to this model the Himalayan range has grown primarily by underplating with little internal shortening of the wedge, through the development of a mid-crustal duplex straddling the brittle–ductile transition on the MHT. This model, however, ignores possible activity along the STD between 16 and 20 Ma. Coeval thrust
Mountain Building: From Earthquakes to Geological Deformation
750
(a)
650 600
RSCM
550
KTM11
500 450 400
300 250
MCT
350
Kippe
Peak temperatures (°C)
700
Lesser Himalaya
200 0
5
Structural distance from the top of the LH (km) (b)
Conventional thermometry
750
RSCM
650
KTM11
600 550
450 400 350 300 250
MCT
500
High Himalaya
Peak temperatures (°C)
700
Lesser Himalaya
200 0
5
Figure 27 Peak temperatures estimated from RSCM thermometry and conventional thermometry at respectively 80 km (a) and 50 km (b) from the front of the range in central western Nepal (data from Bollinger et al. (2004a)) and reference therein). These data show an inverted metamorphic gradient, with the peak metamorphic temperature increasing upward, which encompasses the topmost 7–8 km of Lesser Himalayan units. Red curves show the prediction of model KTM11 (Bollinger, et al., 2006).
motion along the MCT and normal motion along the STD at this time would possibly have allowed rapid exhumation of the HHC as a result of mid-crustal channel flows (e.g., Grujic et al., 1996). This effect is neglected, as most of the data analyzed here pertain to the more recent tectonic history of the range. 6.09.5.2 Modeling Deformation and Surface Processes Based on Continuum Mechanics Recent deformation and the thermometric and thermochronological data from the central Nepal Himalaya all suggest that the shortening across the
405
range has been taken up primarily by thrusting along the MHT fault, with negligible internal shortening of the Himalayan orogenic wedge. Here, this kinematics is shown to be not only plausible from a mechanical point of view, provided that friction along the MHT is low, but also quantitatively consistent with the observed pattern of erosion and the present morphology of range. The modeling should therefore be seen primarily as a test of the internal consistency of the various data sets and as a validation of the hypotheses used to design the models. Furthermore, the modeling provides some insight into the mechanical properties required to fit the various observations, and it sheds light on the coupling between erosion and crustal deformation. 6.09.5.2.1
Model implementation The models discussed here were computed from a two-dimensional (2-D) finite-element code (Hassani et al., 1997) designed to account for the mechanical layering of the crust. The code was modified to incorporate surface processes and the dependency of rheology on local temperature (Cattin and Avouac, 2000). We focus here on the initial model of Cattin and Avouac (2000), referred to here as CA2000. This model has been updated and several types of erosion laws were tested more recently (Godard et al., 2004, 2006). The model CA2000 assumes a prescribed geometry for the MFT, with its characteristic geometry of ramps and flats along which quasi-static frictional sliding is allowed. Outside of the MFT proper, the medium is assumed to deform according to a combination of brittle failure and thermally activated ductile flow. Deformation thus depends on the boundary conditions, assumed rheology, and local temperature. The thermal structure is assumed a priori and was computed from a thermokinematic model with the same parameters as KTM11, but with no underplating taken into account (Cattin and Avouac, 2000). The thermal structure (Figure 30) differs only marginally from that predicted by KTM11 (Figure 29). A 700-km-long section is initially loaded with the present average topography along section AA9 (Figure 30). A fault is introduced with a prescribed geometry representing the flat beneath the LH and the mid-crustal ramp at the front of the High Himalaya (Figure 30). The southern end of the model is fixed. At the northern end, vertical displacements are free and the upper crust is subjected to
406
Mountain Building: From Earthquakes to Geological Deformation
Thermochronological data set
25
15
10
5
0
Apatite Fission track zircon Fission track Mu Ar/Ar Mu Ar/Ar in shear zone
Model KTM11
Ages (My)
20
0
20
40
60
80
100°C 200°C 300°C 400°C 500°C
100
N18° E distance from the MBT (km) Figure 28 Synthesis of thermochronological ages from the Lesser Himalaya and HHC across the Himalaya of central Nepal. Data from Copeland et al. (1991), Macfarlane (1993), Arita et al. (1997), Arita and Ganzawa (1997), Catlos et al. (2001), Wobus et al. (2003); Bollinger et al. (2004a)), and Copeland (pers. comm. (1999) as in Bollinger et al. (2004a)). The data are compared with the predicted cooling ages from model KTM11 for different closure temperatures. See Figure 7 for location of section.
0
100 °C
Depth (km)
200 °C
500 °C
300 °C 400 °C 50
0
100 Distance from the MBT (km)
200
Figure 29 Thermal structure across the central Nepal Himalaya (approximately section AA9 in Figure 7) predicted from model KTM11 (Bollinger et al., 2006). Red thin line show the best-fitting creeping dislocation obtained from the modeling of the GPS data from central Nepal (Bettinelli et al., 2006). The ellipse shows the 1-sigma uncertainty on the position of the updip end of the creeping portion of the MHT (i.e., the down-dip end of the Locked Fault Zone).
20 mm yr1 of horizontal shortening by imposing a southward horizontal velocity of 20 mm yr1 between the surface and 40 km depth. At depths below 40 km, horizontal displacements are null so that the Indian mantle lid does not shorten. It should be noticed that these boundary conditions favor simple shear beneath Tibet. The model is loaded with gravitational body forces (g ¼ 9.81 m s2) and is supported at its base by hydrostatic pressure to allow for isostatic restoring forces. Denudation of the range is modeled using 1-D linear diffusion, which assumes that the southward flux of sediments at the surface is proportional to the local
slope of the topography. Conservation of mass then yields the erosion rate e¼k
q2 h qx 2
½13
where k is the mass diffusivity coefficient, h is the elevation, and x the horizontal distance. The value of k was chosen to ensure an average denudation rate consistent with the 250–675 m2 yr1 flux of sediments eroded from the range (Lave´ and Avouac, 2001) (k ¼ 104 m2 yr1). In the foreland (x 0), the model assumes that sedimentation maintains a flat topography at a constant elevation.
Mountain Building: From Earthquakes to Geological Deformation
407
Depth (km)
CA2000
0
200 400 60
400
–50
600
600
–100 0
200
100
300
400
Free vertical velocity
N18° E distance from MFT (km) Erosion
g
Imposed horizontal velocity
Crust Mantle
Hydrostatic pressure Figure 30 Setting of the numerical model of (Cattin and Avouac, 2000), referred here as CA2000. Bottom shows the boundary conditions. Top shows the assumed steady-state thermal structure which differs only marginally from KTM11 shown in Figure 29 (Bollinger et al. (2006)).
The model assumes a depth-varying viscoelastoplastic rheology. Brittle failure is determined by the Drucker–Prager criterion 1 1 ð1 – 3 Þ ¼ c ðcot Þ þ ð1 þ 3 Þ sin 2 2
½14
where c is the cohesion and is the internal friction angle. Viscous deformation is assumed to obey a powercreep law, consistent with laboratory experiments (Carter and Tsenn, 1987; Kirby and Kronenberg, 1987; Tsenn and Carter, 1987): n
"_ ¼ AP ð1 – 3 Þ expð – EP =RT Þ
½15
where R is the universal gas constant, T is the temperature, EP is the activation energy, and AP and n are empirically determined constants (assumed not to vary with stress and p,T conditions). The ductile flow law thus depends on the rock type and on temperature. The rheology of the mantle is thought to be governed by olivine, and that of the upper crust by quartz (see parameters in Table 1). For the lower crust, the two end-members considered have either a
Table 1 Physical parameters and material properties used in the mechanical models of Cattin and Avouac (2000) Parameter 3
(kg m ) E (GPa) c (MPa) Ap (Pan s1) n Ep (kJ mol1)
Quartz
Diabase
Olivine
2900 20 0.05 10 30 6.03 1024 2.72 134
2900 20 0.25 10 30 6.31 1020 3.05 276
3300 70 0.25 10 30 7 1014 3 510
, density; E, Young’s modulus; , Poisson’s ratio; c, cohesion; , internal friction angle; Ap, power-law strain rate; n, power-law exponent; Ep, power-law activation energy.
soft ‘quartz’ rheology or a stronger ‘diabase’ rheology (Table 1). Fault friction. The MHT is assumed to follow a simple static friction law: jtjj – ? n 0
½16
where t and n are the the shear and normal stress, respectively, on the fault and is the effective friction coefficient.
408
Mountain Building: From Earthquakes to Geological Deformation
6.09.5.2.2
Modeling r esults Figure 31 shows the strain rate field in CA2000, obtained assuming a diabase rheology for the lower crust. The ‘MHT’, along which quasi-static frictional slip occurs, roots into a subhorizontal shear zone that fits the prominent mid-crustal reflector imaged from the INDEPTH profiles. This is a result of the position of the down-dip end of the prescribed fault’s geometry and of the a priori thermal structure which implies a subhorizontal zone locally weakened by higher temperatures (Figure 30). Sensitivity tests have revealed that the effective friction, , along the MHT must be lower than 0.15. If basal friction is too high, the hanging wall shortens as it overrides the MHT. If friction is low enough, the hanging wall can ride over the MHT without much internal deformation. Distributed plastic deformation then only occurs in the vicinity of the kinks along the MHT ramp-and-flat geometry (Figure 31). The horizontal velocities are therefore nearly constant, with the exception of the discontinuity at the point where the fault meets the surface (Figure 32). In CA2000, the 20 mm yr1 of horizontal shortening across the range is partitioned into 15 mm yr1 of underthrusting and 5 mm yr1 of overthrusting (Figure 32). This partitioning is thus quite close to that derived from the long-term kinematic evolution of the range. The model leads to some natural balance between denudation and uplift rate. This balance is seen in Figure 32, in which denudation rate, modeled here from a simple linear diffusion equation, approximately equals the modeled uplift rate and both are close to the pattern of fluvial invision. The model
reproduces reasonably well the zone of localized uplift and erosion observed at the front of the high range. The modeled topography rapidly reaches steady state or close to it. As expressed in eqn [12], uplift and erosion differ at steady state because erosion must compensate for uplift and for the horizontal transport of the hanging wall topography. The difference is small, less than 1 mm yr1, because the overthrusting rate is only 20% of the convergence rate and the slope of the topography is 10–20% at most. In fact, the assumption of a diffusive erosion law leads to a modeled topography in CA2000 that departs somewhat from the observed average topography in the study area (Figure 33). The predicted topography fails in particular in reproducing the break in slope at the front of the High Himalaya. Godard et al. (2004, 2006) were able to solve this problem and predict a more accurate topography by introducing an erosion model (Lave´, 2005) in which river valleys and interfluves are distinguished from each other (Figure 32). Their model, which otherwise is the same as that of Cattin and Avouac (2000), assumes that fluvial incision is proportional to the shear stress in excess of some threshold (Lave´ and Avouac, 2001) and that denudation of hillslope is driven by fluvial incision. An unexpected result of this modeling is that relatively high viscosity in a lower crustal diabase rheology was found to better reproduce the topographic profile across the range than did the softer, quartz-like rheology applied by Cattin and Avouac (2000). This shows that rheology should not be assessed independently of the erosion law.
Depth (km)
CA2000
0 MHT
–50 Moho
–100
0
100 N18° E distance from the MFT (km)
200
0.0 0.1 0.2 0.3 0.4 0.5 Inelastic deformation rate (×10–13 s–1) Figure 31 Shear strain rates over the long-term computed by assuming a diabase rheology for the lower crust and quartz rheology for the upper crust (Cattin and Avouac, 2000). It shows a subhorizontal shear zone that closely follows the midcrustal reflector imaged by the INDEPTH experiment (Zhao et al., 1993).
Mountain Building: From Earthquakes to Geological Deformation
409
(a) 20mm yr–1
Depth (km)
0 20 40 60 80 0
100 N18° E distance from MFT (km)
200
(b) 12 Fluvial incision
(mm an–1)
Denudation
Tectonic uplift
8
Erosion 4
0
(c)
0
100
200
0
100
200
Horizontal velocity (mm yr–1)
30
20
10
0
Figure 32 Velocity, horizontal and vertical displacement rates predicted by the FEM model of Cattin and Avouac (2000). (a) Velocity field relative to India. (b) Predicted uplift and erosion rates are compared to the measured river incision profile of Lave´ and Avouac (2001) and to the erosion rates derived from thermochronological data (Burbank et al., 2003). (c) Horizontal velocity relative to India of sub-Himalaya as derived from the slip rate on the MFT. These data indicate that deformation of the hanging wall of the MHT is negligible. The model (continuous line) fits the pattern provided that friction on the flat portion of the MHT does not exceed 0.15, or eventually 0.3 if a high pore fluid pressure is assumed.
6.09.6 Geodetic Deformation and the Seismic Cycle In the previous sections, crustal deformation was described as a continuous, steady process resulting
from a combination of viscous deformation at depth and quasi-static plastic deformation in the upper crust. This is probably a reasonable approximation for deformation over a geological period of time longer than 10 ka. However, the recurrence
Altitude (km)
7 6 5 4 3 2 1
Altitude (km)
410
7 6 5 4 3 2 1
Mountain Building: From Earthquakes to Geological Deformation
measurements offer a way to assess how this nonstationary deformation proceeds.
Mean topography
6.09.6.1 River elevation
0
200 km
100
Figure 33 Top: Comparison of the average topography across the central Nepal Himalaya with that predicted by model CA2000 and that of Godard et al. (2004). CA2000 assumes a linear diffusion model of erosion and a quartz rheology for the lower crust. Godard’s et al. (2004) model assumes that river erosion drives denudation of hillslopes, and a diabase rheology for the lower crust. Bottom: Comparison of river profiles along major rivers from central Nepal with the theoretical river profile predicted by the modeling of Godard et al. (2004).
of large earthquakes associated with mountain building in the Himalaya (Figure 34) shows that deformation is not a steady process. It must result in part from seismic events or aseismic transients, like afterslip, and from viscous deformation that may not necessarily be stationary. Geodetic
Large Earthquakes in the Himalaya
Four destructive earthquakes have occurred in the Himalayan region since the end of the nineteenth century (Ambraseys and Bilham, 2000; Ambraseys and Douglas, 2004; Hough et al., 2005): the Mw 8.1 1897 Shillong earthquake which did not really occur along the Himalayan arc; the Mw 7.8 Kangra earthquake, which triggered a Mw 7.0 event north of Dehradun; the Mw 8.2 Bihar–Nepal earthquake of 1934; and the Mw 8.5 Assam earthquake of 1950. Considerable uncertainty remains as to the exact location and size of the 1934 Bihar–Nepal earthquake, the most recent large historical event in the study area (e.g., Chen and Molnar (1977); Pandey and Molnar (1988)). Long-period seismic data indicate a released seismic moment of about 4.1 1021 N m, corresponding Mw 8.4 earthquake, could have occurred along a thrust fault with a 5 northward dip (Molnar, 1984). Macroseismic intensities and subsidence of the foreland revealed from leveling data suggest that the earthquake ruptured a 250–300 km along-strike segment of the arc (Bilham et al., 1998) (Figure 34). There is no evidence that the rupture broke the surface (Chander, 1989; Lave´ et al., 2005), and the northward extent of the rupture is not constrained. 90°
80° 20
05
Isl am
ab ad
7.5 < Mw < 8
15 55
8 < Mw
Tibet
19
05 14
30°
13
30° 18
03 150
5
New Delhi
1950
w
1833 19 34
1947 1724
1100
1897
80°
90°
Figure 34 Estimated rupture area of major earthquakes along the Himalaya. Based on Ambraseys (2000), Ambraseys and Bilham (2000), Ambraseys and Jackson (2003), and Bilham (2004). The AD 1100 and AD 1413 events were both documented from paleoseismic studies (Upreti, 2000; Lave´ et al., 2005; Kumar et al., 2006). There is a possibility that the paleoruptures along the front of the northwestern Himalaya, which were only loosely dated to AD 1413, might in fact relate to the 1505 historical earthquake (Kumar, et al., 2006). If so the magnitude of that event might have been close to Mw 9.
Mountain Building: From Earthquakes to Geological Deformation
There are historical indications that Katmandu has been struck repeatedly by severe earthquakes in the past (Rana, 1935; Pant, 2002). Major destruction was reported in 1225, 1344, 1408, 1681, and 1833. The 1833 event, estimated to be Mw 7.7, apparently ruptured the western part of the segment that ruptured again during the 1934 quake (Bilham, 1995), but the extent of damage in Katmandu suggests a slightly smaller or more distant event. The 1255 earthquake was a major disaster that caused the death of King Abhaga Malla and killed about one-third of the population of Katmandu valley (Pant, 2002). Very little is known about the 1344 event, which may have caused the death of King Ari Malla (Pant, 2002). When assessing these historical earthquakes, it should be noted that some palaces and temples in Katmandu valley built after the mid-fifteenth century remained intact until the 1934 event. These constructions had deep foundations (up to 15 m deep) and were reinforced by wooden frames so that they could resist significant shaking. The Temple of Shiva in Bakthapur was one such building, built in CE 1458 but totally destroyed in 1934 (Pandey and Molnar, 1988). Some other temples built less strongly at the end of the fourteenth century or in the early fifteenth century survived until they were destroyed in 1934 suggesting that the 1681 and 1408 events were smaller than the 1934 earthquake. On this basis, the time interval between earthquakes with magnitudes similar to or larger than the 1934 event may vary between 100 and 475 years. Among the large Himalayan earthquakes of the last century, the 1905 Kangra event was probably the most similar to the 1934 event (Figure 34). A recent re-analysis of the seismic waveforms suggests a surface-wave magnitude between 7.5 and 8.2, with the lower estimates being more probable (Ambraseys, 2000). The area between the locations of the 1934 and 1905 events represents a 800-km-long seismic gap. The most recent event that ruptured its eastern portion was probably the 1833 event. A large-magnitude event along the western portion of the Himalaya was reported in 1803 (Oldham, 1883b). The magnitude of this earthquake is estimated to be Mw 7.5 by Ambraseys and Douglas (2003), but might be underestimated, in particular because the historical data are difficult to interpret due to the political situation in the area at that time (Bilham and Wallace, 2005). This earthquake ruptured an arc segment probably between Dehradun and far western Nepal and its moment magnitude may have reached as high as 7.8
411
or 8 (Bollinger, 2002). Further back in time, historical accounts suggest that a ‘giant’ earthquake in 1505, with a magnitude estimated to 8.2 by Ambraseys and Douglas (2003), but possibly greater than 8.5 (Bilham and Wallace, 2005). None of these earthquakes produced known surface ruptures. In fact, the Mw 7.6 earthquake that ruptured the Himalayan front east of the Kashmir basin in 2005 (Avouac et al., 2006) is the only example of a large Himalayan earthquake with surface ruptures. Surface ruptures extend over a distance of about 75 km, and the mean coseismic slip was measured to 4 m. Due to the particular setting of the Hazara syntaxis that is not reviewed here, this earthquake broke along a relatively steep splay fault with a complex geometric relationship to surface geology (Avouac et al., 2006). Paleoseismic investigations suggest that the MFT may in fact only break during even larger earthquakes. According to these studies, the Himalaya may have produced other earthquakes with magnitudes as high as Mw 8.8. Along the Himalayan foothills in Nepal, there is evidence for a 17 (þ5/3) m slip event on the MFT c CE 1100 at locations separated by 240 km along-strike (Nakata et al., 1998; Upreti et al., 2000; Lave´ et al., 2005). Evidence for a similar event with an age loosely constrained to c. CE 1413 was also found in the Kumaon and Garhwal Himalaya (Kumar et al., 2006). In fact, there is a possibility that these paleoruptures might in fact relate to the 1505 historical earthquake (Ambraseys and Jackson, 2003), which would then be inferred to have ruptured the Himalayan front from western Nepal to Garwhal over a distance possibly as large as 800 km. If so the magnitude of that event might have been close to Mw 9. Feldl and Bilham (2006) also argue for the possibility for such extremely large events, which would then need to have a quite long reccurence interval of the order of a thousand years. 6.09.6.2 Geodetic Deformation in the Nepal Himalaya The pattern of active deformation across the Nepal Himalaya has been documented from geodetic measurements (Bilham et al., 1997; Jouanne et al., 1999; Larson et al. 1999b; Chen et al., 2004; Bettinelli et al., 2006). These measurements show that all along the Himalayan front in Nepal, relative displacements between the Gangetic Plain and the LH MFT have been small (Figure 34). Velocities relative to
412
Mountain Building: From Earthquakes to Geological Deformation
northern India rise from 3 mm yr1 around Katmandhu basin to 15 mm yr1 50 km to the north. Data collected in southern Tibet (Bilham et al., 1997; Larson et al., 1999b) and converted to the same reference frame as the data from Nepal (Bettinelli et al., 2006) show that velocities reach to 1517 mm yr1 north of the High Himalaya (Figures 35 and 36). The sparse data available from southern Tibet suggest some additional N–S contraction occurs north of the Himalaya though at a much lower rate than across the range itself (Feldl and Bilham, 2006). At least 70% of the contraction perpendicular to the range is thus confined to a 100-km-wide zone at the front of the High Himalaya, starting 80 km north of the MFT. These data show that although aseismic deformation was going on beneath the High Himalaya and southern Tibet over the last few years, the MFT– MHT has remained locked from where it emerges along the foothills to below the high range. The data can indeed be adjusted from a model in which the MHT is assumed to be fully locked over a width of about 100 km to one in which the MHT moves slightly laterally (Figure 35). Deformation is then computed from creeping dislocations using the formulation of Okada (1992) for point sources in an elastic halfspace (Bollinger et al., 2004b). Comparison with finiteelement modeling has shown that modeling ductile
80°
82°
84°
aseismic shear from this approach is a reasonable first-order approximation (Vergne et al., 2001). The convergence rate in this model varies from 19 2.5 mm yr1 across eastern Nepal (Figure 37) to 14 mm yr1 across western Nepal (Bettinelli et al., 2006). Figure 36 compares the prediction of a simpler 2-D model of interseismic deformation, in which only one rectangular creeping dislocation is considered, with the horizontal velocities measured from Global Positioning System (GPS) and the uplift rates determined from levelling measurements (Jackson and Bilham, 1994b). The position of the best-fitting creeping dislocation coincides remarkably well with the mid-crustal reflector imaged beneath southern Tibet (Figure 38). This observation adds further support to the interpretation of this reflector as the northward continuation of the basal detachment beneath the Himalayan wedge. This model does not predict any significant N–S contraction of southern Tibet due to the assumption of an abrupt transition from the fully locked to the creeping portions of the MHT. If a very gradual transition, several hundreds of kilometers wide, is allowed then the inferred shortening rate gets larger up to 21 mm yr1 or more (Feldl and Bilham, 2006). If such a pattern was stationary over the whole seismic cycle, such a model would then require large earthquake ruptures along the MHT to extend over a several hundred kilometres beneath the Tibetan
86°
88°
30°
30°
28°
28°
km 0 80°
50
100 82°
84°
86°
88°
Figure 35 Horizontal velocities relative to stable India measured from geodetic measurements over the last 15 years (Bettinelli et al., 2006). Black arrows represent velocities determined from campaign measurements. Red arrows show velocities determines from continuous geodetic measurements using GPS and the DORIS system (Tavernier et al., in press). Green arrows show the predicted velocities assuming that the MHT is fully locked over the yellow area (Bettinelli et al., 2006).
Mountain Building: From Earthquakes to Geological Deformation
413
(a) 20
Horizontal velocity (mm yr–1)
2-D model, 16.2 mm yr–1
LHAS
15
10
5
0
–5 0
50
100
150
200
250
300
350
400
450
500
N18° E Distance from SIMR (km) (b) 10
Uplift rate (mm yr–1)
2-D model, 16.2 mm yr–1
5
0
Subsidence Katmandu basin –5 0
50
100
150 km
N18° E Distance from SIMR (km) Figure 36 Comparison of velocities across the Himalaya of central and eastern Nepal with the prediction of a model of interseismic strain computed from a creeping dislocation embedded in an elastic half-space (Bettinelli et al., 2006). All data were projected on section AA9 across central and eastern Nepal. The velocities at the IGS station in Lhassa (LHAS) is shown for reference but was not used to determine the shortening rate across eastern Nepal Himalaya, due to its location well east of the profile. (a) Horizontal velocities relative to stable India. Red dots: cGPS stations and DORIS station at EVEB. Black dots: campaign GPS measurements. (b) Observed (Jackson and Bilham, 1994a) and modelled uplift rates along a levelling profile across Central Nepal running from the Indian border to the Chinese border.
Plateau. This seems improbable to the author who favors the view that the pattern of geodetic contraction across southern Tibet varies during the seismic cycle as a result of viscoelastic relaxation. It would imply that the shortening rate determined from the elastic dislocation modeling described above would be timedependent with an apparent variation reflecting the stress transfer during the interseismic period (Perfettini and Avouac, 2004). Figure 37 shows how the misfit between the observed geodetic deformation across central and eastern Nepal and that modeled from the 3-D model of Figure 35 varies with the convergence rate. The rate, at the 1-sigma confidence level, is
estimated to be 19 2.5 mm yr1. A 2-D model yields a somewhat lower rate estimated at about 16 mm yr1 (Bettinelli et al., 2006). Despite numerous efforts in this field, the geodetic rate of shortening across the range thus remains poorly constrained. This is due in part to the sparsity of geodetic data from southern Tibet and in part to the uncertainty of the plate motion of India. The 2.5 mm yr1 uncertainty stated here takes into account the uncertainties on the original geodetic measurements and also on the plate motion of India. However, as is commonly done in tectonic geodesy, the uncertainty related to the geometry of the elastic dislocation model itself is ignored, as is the uncertainty related to the overall
414
Mountain Building: From Earthquakes to Geological Deformation
3.0 2.9
Holocene rate –1 21 ± 11.5 mm yr
2.8 2.7
Geodetic rate –1 19 ± 2.5 mm yr
χ2
2.6 2.5 2.4 2.3 2.2 2.1 2.0 11
12
13
14 15
16
17
18 19
20
21 22
23 24
25
Shortening rate (mm yr–1) Figure 37 Determination of the best-fitting slip-rate on the MHT determined from the 2-D modeling of the horizontal and vertical geodetic velocities of Figure 36. The 2 represents the misfit between the observed and predicted velocities. The uncertainty on the slip rate was obtained by renormalizing the uncertainties on the data so that the normalized reduced 2 would be unity at minimum (Bettinelli et al., 2006). The geologic slip rate on the MHT and it 1 – uncertainty (yellow shading) is also shown for reference.
INDEPTH
N
T
T
D
ST
Katmandu
T
MB
MC
MF
S
Depth (km)
0 20 40 60 TIB-1
80
0
100 N18° E Distance from MFT (km)
TIB-3
200
Figure 38 Ouline of the structural cross-section across central Nepal and the seismic sections, TIB-1 and TIB-3 of the INDEPTH profile (Zhao et al., 1993; Nelson et al., 1996; Hauck et al., 1998) (see location in Figure 6). Thick continuous line shows the geometry of the creeping dislocation determined from the joined inversion of all geodetic data for central and eastern Nepal (Bettinelli et al., 2006). Hypocenters from well-constrained events recorded during the temporary deployment, in 1995, of three three-component seismic stations in addition to the permanent seismic network (Cattin and Avouac, 2000).
applicability of the model. The uncertainty in the geometry is probably the major worry, and any estimate of geodetic shortening across a mountain range should take into account the highly dependent, unstated uncertainties involved with the geometry. This is not to mention the uncertainties on the elastic properties of the medium itself, which probably vary laterally and with depth in a setting such as the Himalaya. Despite the uncertainties in these estimates, the rates of shortening across the Himalaya determined
from geodesy or from geology over the Holocene and over the longer term appear to be in fairly good agreement (Figure 37). 6.09.6.3 Microseismic Activity in the Nepal Himalaya Local seismic monitoring in the Katmandu area (Pandey et al., 1995, 1999) has revealed intense microseismic activity (Figure 39). Most events cluster in a narrow zone beneath the front of the high
Mountain Building: From Earthquakes to Geological Deformation
415
30°
A'
28°
MFT
3500 m ML magnitude
>5 4-5 2.5-4
km 0
80°
82°
100
84°
A
86°
88°
Figure 39 Seismicity of Nepal recorded by the DMG seismological network between 04/01/1995 and 10/12/1999 (Pandey and Rawat, 1999). The red line follows the 3500 m elevation contour line. Seismic activity north of this zone is mostly related to active normal faulting in southern Tibet. South of this line seismic activity, due to N–S compression, is probably triggered by interseismic straining (Pandey et al., 1995; Pandey and Rawat, 1999).
range at depths between 5 and 30 km. Initial observations suggest that the belt of seismicity ends abruptly to the north as soon as the elevation exceeds 3500 m (Figure 36), suggesting some cutoff effect due to the topography (Avouac, 2003). This zone with high microseismic activity coincides with high uplift rates recorded in leveling measurements ( Jackson and Bilham, 1994b) and lies quite close to the tip of the creeping zone as determined from the modeling of the geodetic deformation (Figure 38). This correlation suggests that the microseismic activity is triggered by stress buildup near the down-dip end of the locked fault zone (Cattin and Avouac, 2000; Bollinger et al., 2004b). The mechanical consistency of the simple model described above was tested using the FEM model CA2000 described in Section 6.09.5. To simulate the effect of the locked fault zone from this model, friction on the MHT was increased slightly (from 0.15 to 0.17) after the regime of quasi-static sliding along the MHT had been reached. At this stage in the model, frictional sliding along the fault was suddenly inhibited while ductile shear extended along the subhorizontal shear zone beneath the high range and southern Tibet. The model successfully predicts horizontal displacement and uplift rates consistent
with geodetic measurements (Figure 40), as well as stress accumulation near the northern edge of the locked fault. It turns out that 83% of the microearthquakes fall within this area of enhanced Coulomb stress (Cattin and Avouac, 2000). The correspondence of seismicity cutoff to a particular elevation (3500 m) can be simply explained by a vertical stress increase inducing a stress tensor change, such that the deviatoric stresses are no longer affected or even decrease during interseismic stress buildup. Figure 41 compares the seismicity relocated from the double-difference technique (courtesy of Sudhir Rajaure) with Coulomb stress change computed from the model of interseismic deformation (Bollinger et al., 2004b). This model accounts for the nonuniform tectonic stress field resulting from the topography. The vertical stress is computed from the topography, and the horizontal principal stress was adjusted to predict a null Coulomb stress change at the location where the seismicity is observed to drop (Figure 41). The deviatoric stresses in the 5–15 km depth range of the seismicity is thus estimated to a maximum of 35 MPa (Bollinger et al., 2004b). Because the seismicity in this region releases an accumulated moment that amounts to, at most, a few
416
Mountain Building: From Earthquakes to Geological Deformation
(a) N
INDEPTH
D
T
Depth (km)
ST
Katmandu
T
MD
MC
T MF
S 0 20
MHT 40 60 80
TIB-1
0
TIB-3
200
100 N18° E distance from MFT (km) –5 –4 –3
–2
–1
0
1
2
3
4
5
Change in Coulomb failure stress (105 Pa)
Vertical velocity (mm yr–1)
(b) 10
Leveling measurements
5
0
–5
Horizontal velocity
(mm yr–1)
(c) 25 Continuous sites 15
Campaign sites
5
–5
Figure 40 (a) Horizontal velocities relative to India along section AA9’ computed from CA2000 (Cattin and Avouac, 2000)(continuous line) compared with velocities estimated from GPS campaigns (Jouanne et al., 1999; Larson et al., 1999) that were available at the time of the publication of CA2000. The red dots shows for comparison the velocities determined from continuous geodetic measurements between 1997 and 2005 (Bettinelli et al., 2006). (b) Measured uplift rates (circles) (Jackson and Bilham, 1994b) and theoretical uplift rates (continuous line) computed from CA2000 (Cattin and Avouac, 2000), assuming a shortening rate of 16 mm yr1. (c) Coulomb stress changes due to 350 years of interseismic strain computed from model CA2000 and observed seismicity projected on section AA9 (white dots) (Cattin and Avouac, 2000).
percentage of the upper crustal strain revealed from geodetic monitoring, it holds that these earthquakes release an amount of strain that is insignificant compared to crustal deformation over the long term. The
measured geodetic deformation therefore must be either permanent and aseismic, or elastic. Significant nonrecoverable deformation of the upper crust in the interseismic period can be excluded, since localized
Mountain Building: From Earthquakes to Geological Deformation
417
30° N –2
C
B
–1
0 bars
1
2
C B C 28° N
A km 0
80° E
50
100
82° E
84° E
86° E
88° E
Figure 41 Coulomb stress variations computed for optimally oriented faults for a spatially varying regional stress field (Bollinger et al., 2004b). The vertical principal stress is assumed to vary in proportion to elevation, for a mean crustal density of 2900 g cm3, and the horizontal stress is assumed uniform. Also reported is the seismicity recorded by the NSC (DMG) that was relocated by the double-difference technique (courtesy of Sudhir Rajaure, DMG/NSC).
slip on the MFT accounts for most, if not all, of the shortening across the range. The deformation measured over the last few years across the Katmandu section thus necessarily reflects elastic strain of the upper crust that will be ultimately released by slip on the MHT. This stress transfer is probably chiefly the result of recurring major earthquakes rupturing the MHT.
6.09.6.4 What Controls the Down-Dip End of the Locked Portion of the MHT? Geodetic and seismic data provide some information on the position of the down-dip end of the locked portion of the MHT. The mechanical model of interseismic strain shown in Figure 38 produces a smooth transition of the fault zone from fully locked to ductilely flowing along its northward continuation, where temperatures exceed 450 C. Close inspection of the velocity field reveals that a portion of the fault near the base of the prescribed fault geometry undergoes stable sliding. However, the exact location of the transition cannot be constrained tightly from the GPS campaign data, and there are a variety of possible sliprate-distribution solutions with a tapering zone of finite width, including the one produced from the mechanical model, that would fit the data equally well. Assuming that the transition is abrupt, it is possible to invert the geodetic data for the location of the up-dip end of the creeping dislocation in the 2-D model of Figures 36 and 38. The location of this point is relatively well constrained and, when
compared to the thermal structure of Figure 12, falls in a temperature range of 250–350 C (though this large range ignores uncertainties in the thermal structure itself). This temperature range suggests that the onset of creeping dislocation could correspond to the transition from slip-weakening friction to aseismic stable sliding. Indeed, according to laboratory experiments and field observations, this thermally activated transition occurs at a temperature of 300–350 C for quartzo-feldspathic rocks (Blanpied et al., 1991; 1995). This is also in keeping with the observation that the down-dip extent of the seismogenic zone generally coincides with the 350 C isotherm, if this temperature is reached above the Moho (Oleskevich et al., 1999). The same rationale proposed for subduction zones (Hyndman et al., 1997), for cases in which the locked fault zone does not extend deeper than the fore-arc Moho, seems to hold for intracontinental megathrust faults as well. Note that the coincidence between the transition to stable sliding and the zone of higher conductivity images from the magnetotelluric sounding experiment described previously also suggest that fluids and possible metamorphic reactions may influence the position of the down-dip end of the locked fault portion.
6.09.6.5 A Model of the Seismic Cycle in the Central Nepal Himalaya While motion along the locked portion of the MHT is probably stick-slip and results from recurring large
418
Mountain Building: From Earthquakes to Geological Deformation
earthquakes, aseismic creep prevails on the deeper part of the interface (Figure 42). The elastic strain accumulated in the upper crust during the interseismic period must be transferred to slip on the MFT at the surface. Some of the largest events along the front of the high range probably do facilitate this strain transfer and activate a portion of the MFT–MHT. For example, analysis of the accelerometric data suggests this may be possible for the 1991 Uttar Kashi events (Cotton et al., 1996). These events did not propagate all the way to the Himalayan piedmont, possibly because the elastic stresses in the surrounding elastic medium were too low to sustain the propagation of the seismic rupture. It might then be conjectured that events of M ¼ 7 contribute to loading of the shallow-dipping decollement below the LH, raising elastic stresses there until one event ruptures all the way to the surface. M 7 events such as the Uttar Kashi might then account for most of the stress transfer to the front of the thrust fault system (Figure 42). If this were the case, the average return period of large earthquakes in the study area could also be estimated from the rate of accumulation of the slip deficit on the locked portion of the MHT. However, postseismic or possibly aseismic creep events may also contribute to some fraction of fault slip. 6.09.6.6 Geodetic Deformation, Seismic Coupling, and Recurrence of Large Earthquakes in the Himalaya We have seen that geodetic measurements and background seismicity can be used jointly to delineate the locked portion of the fault, providing important information on the potential location and size of future earthquakes. The major thrust fault along the Himalaya of Nepal appears to be locked from the Earth’s surface, at the MFT, to beneath the front of the high range, a distance of about 100 km northward (yellow area in Figure 35). We then estimate what proportion of the slip along this fault zone, which remains locked in the interseismic period, is ultimately released by earthquakes. The 3-D model of interseismic strain of Figure 35 allows some estimate of the deficit of slip that is accumulating on the Locked Fault Zone that can then be converted into a rate of accumulation of moment deficit. Assuming a shear modulus of 30 GPa, this estimate is 4.3 1019 Nm yr1 for the entire 900-km-long segment of the Himalayan arc encompassed by the model. This means that, for the
current rate and pattern of interseismic strain, it takes only 7 years on average over the whole area to accumulate an elastic strain equivalent to what was released during the 2005 Mw 7.6 Kashmir earthquake (the moment of this event is estimated by joint inversion of seismic and geodetic data to be 2.8 1020 Nm (Avouac et al., 2006). Similarly, it would take only 100 years on average over the whole area to accumulate an elastic strain equivalent to what was released during the Mw 8.2, 1934 Bihar Nepal, (assuming a moment of 4 1021 Nm). If this earthquake is typical of the largest possible events along this segment of the Himalayan arc and assuming that motion along the LFZ occurs only during large earthquakes, then there should be one such earthquake per century on average, or less-frequent earthquakes with even larger magnitudes. The historical catalog might be used to test whether the known historical earthquakes are consistent with long-term slip on the MHT, assuming it would be only the result of reccurring seismic ruptures. The seismic moment released by historical earthquakes in Nepal need then to be estimated. Let us assume then the 1803 and 1833 earthquakes have released seismic moments of 4 1021 Nm and 2 1021 Nm, respectively, in the upper range of possible values. If this is the upper bound, then we calculate that large earthquakes over the last two centuries have released at most 1022 Nm, or 5 1021 Nm yr1, about the same amount as that required to account for slip on the LFZ resulting only from large seismic events. So, with these assumptions, it is possible that the LFZ might slip only during very large earthquakes. The time period considered might be too small to be representative of the long-term average Himalayan seismicity and it should be realized that widely different conlusions can be reached depending on the time period considered. A similar analysis was carried on by Bilham and Ambraseys (2005) considering the full 2200 km length of the Himalayan arc and the historical seismicity since AD 1500 and assuming full locking of the MHT over a width of 80 km on average. This analysis suggests that historical seismicity would have released only about one-third of the geodetic convergence over the last 500 years. This might be taken to imply that either the largest events are underrepresented in the historical catalog, suggesting that several magnitude-8 earthquakes, or even larger events are overdue in the Himalaya. Alternatively the reasoning is flawed because of a possible overestimation of the average convergence rate, of the
Uplift ratio (mm an–1)
Mountain Building: From Earthquakes to Geological Deformation
10
419
Interseismic period
5 0 0
100
200
Depth (km)
0 Locked zone
20
Frictional stable sliding
40
Ductile
shear
60 80 0
100
200
Uplift (m)
Distance (km) Coseismic Mw < 7–7.5
2 1 0 0
100
200
Depth (km)
0 Ruptured zone
20 40 60 80 0
100
200
Distance (km)
Uplift (m)
3
Coseismic Mw > 8
2 1 0 –1 0
100
200
100 Distance (km)
200
Depth (km)
0 20
Ruptured zone
40 60 80 0
Figure 42 Model of the seismic cycle in the Nepal Himalaya. Slip along the locked portion of the Main Himalayan Thrust (MHT) fault is thought to result from ductile at depth greater than about 20 km, where temperature exceed about 350 C. The shallower portion of the MHT is thought to slip during either blind thrust earthquakes with magnitude typically less than 7.5, or during larger earthquakes breaking the surface along the trace of the MFT. Seismicity of the MHT is thought to play a negligible role in the strain budget.
420
Mountain Building: From Earthquakes to Geological Deformation
degree of interseismic locking and ignorance of possible aseismic transient slip. The uncertainties involved with the various estimated quantities of these reasonings thus provide the possibility that a fraction of the slip on the locked portion of the MHT is in fact aseismic. Monitoring of geodetic deformation in the Himalaya will help address that question in the future, and help constrain better the rate of acumulation of moment deficit, at the scale of the whole Himalayan arc, due to be released by future earthquakes.
6.09.6.7
Is Interseismic Strain Stationary?
Due to stress transfer associated with the seismic cycle, strain rates during the interseismic period should not be stationary. Indeed, there should be some time variation in crustal straining in the period between two large recurring earthquakes due to both frictional afterslip and ductile flow (e.g., Smith and Wyss (1968); Pollitz, et al. (1998); Cohen (1999)). In the case of the Himalaya, a strong asymmetry of the postseismic relaxation should be expected due to the presumably much lower viscosity beneath southern Tibet. In this setting, a viscous strain wave would be expected to propagate north of the range following each large earthquake rupturing at the front of the Himalaya (Bott and Dean, 1973). As a result the apparent shortening rate across the range might appear larger in the early postseismic period than in end of the interseismic period. This mechanism has been analyzed based on a simplified 1-D model of the seismic cycle (Perfettini and Avouac, 2004). In this simple model, the behavior of the thrust fault system is simplified and modeled as a simple springs-and-sliders system loaded by a constant force. Three main domains with different rheologies are considered, with the springs allowing stress transfer between them during the seismic cycle. The upper fault portion is assumed to undergo unstable frictional sliding and is modeled by slider 1 with a state-and-rate friction law (Dieterich, 1979; Rice and Tse, 1986). Its behavior is obtained by solving the following equations: t1 ¼ 1 ð þ a1 log ðV =V Þ þ b1 log ðV =Dc ÞÞ ˜ d=dt ¼ 1 – ðV =Dc Þ
½17 ½18
where t1 is the shear stress; 1 is the effective normal stress set to some arbitrary value of 250 MPa (corresponding to a depth of about 10 km); is the steady-state friction coefficient corresponding to some reference sliding velocity V; a1,b1, and Dc
are frictional parameters that were chosen in order to obtain a coseismic slip of the order of 5 m. The deeper fault portion is assumed to undergo rate-strengthening brittle creep. This process is modeled from a rate-strengthening friction law associated with slider 2, t1 ¼ 2 ð þ a2 log ðV =V ÞÞ
½19
where a2 > 0 to ensure stable sliding. The fault roots at depth in a ductile shear zone with viscosity , arbitrarily assumed to be about w ¼ 5 km thick. This fault portion is modeled from a viscous slider so that t3 ¼ ? V3
½20
where ¼ /w. Motion of the most frontal slider, slider 1, would represent the slip along the seismogenic fault portion. Motion of slider 3 represents the convergence across the fault system. A constant force, F, is applied on slider 3, which was adjusted in order to produce an average slip rate V0 ¼ 20 mm yr1. The behavior of the system is then characterized by 1. the coseismic slip, U; 2. the duration of the interseismic period: t ¼ U/V0 ; 3. the Maxwell relaxation time defined here as: TM ¼ k/ ; 4. the brittle relaxation time defined here as: tR ¼ a2:2/kV0. If the driving force, F, exceeds some critical value, Fc, determined by the frictional coupling between the upper and the lower plate, the system produces recurrent slip events, with a return period T, followed by afterslip (due to the second slider) and viscous relaxation (due to the third slider). If the viscosity is high, that is, if the Maxwell time is of the order of the duration of the interseismic period, slider 3 has a nearly uniform motion. If the Maxwell time is much shorter than the duration of the interseismic period, then some significant variations are observed. More precisely, the model predicts that the shortening rate at the end of the interseismic period, vf, relates to the shortening rate at the begining of the postseismic relaxation, vi, according to vi T t ¼ exp _ vf TM F – Fc
½21
where t is the coseismic stress drop. So if the coseismic stress drop is large compared to the preseismic
Mountain Building: From Earthquakes to Geological Deformation
deviatoric stresses driving viscous flow at depth, variation of strain in the interseismic period might be significant. The Maxwell time is difficult to estimate in the absence of geodetic measurements of postseismic relaxation following Himalayan earthquakes. The modeling of the interseismic and longer-term kinematics of deformation across the range provides some constraints but is not very discriminant (Cattin and Avouac, 2000; Godard et al., 2004). The fact that ‘quartz’-like rheology (Table 1) seems too weak (Godard et al., 2004), places a lower bound on the possible value of the Maxwell time. Given the thermal structure of the range, such a nonlinear rheology leads to a postseismic relaxation similar to the relaxation of a linear viscous body with a Maxwell time of the order of several years. The evidence for time variation of the interseismic strain pattern might be better found by looking at lateral variations of the interseismic strain. This mechanism might indeed explain the apparent lower shortening rate across western Nepal Himalaya, where no large earthquake is known to have occurred for several centuries, than across eastern Nepal Himalaya which ruptured in 1934. Also it might be an explanation for the geodetic rate lower than the the long-term Holocene rate on the MFT. Of course, this discussion which is based on a simple 1-D model is only tentative at this point. In reality, stress transfers occurs in 3-D and coseismic stress drop due to the rupture of a particular arc segment would be compensated by increased stresses on adjacent segment and elastic response of the whole surrounding medium (not only of the thrust sheet as modeled here). Also the northward propagation of the postseismic strain wave is ignored. A proper analysis of this possible mechanism would require better geodetic measurements, especially across southern Tibet, where the gradient of horizontal displacement should vary during the seismic cycle, and 3-D viscoelastic modeling.
6.09.7 Discussion 6.09.7.1 The Critical Wedge Theory: Does It Apply to the Himalaya? The geometry of the Himalayan orogen, together with the early view that the MCT, MBT, and MFT were crustal-scale thrust faults that formed in a forward propagation sequence (Figure 43), has inspired the idea that the Himalaya could be a crustal-scale equivalent of an accretionary prism, an analog to the
Gangetic plain MF
T
MB
Himalaya MC T
T
ST
D
NH
421
Tibet NZ
ITSZ
Moho
Figure 43 Schematic geological section across the Himalaya of central Nepal reflecting early interpretation of the major thrust faults as crustal-scale parallel faults; analogy with the geometry of a sand wedge at front of a bulldozer is shown.
wedge of sand that forms in front of a moving bulldozer (Davis et al., 1983). The formation of this wedge can be easily reproduced in ‘sand-box’ experiments (e.g., Malavieille (1984), Davis et al. (1983)). When sand layers are detached from a rigid basement and pushed by a moving buttress, they form a wedge with a nearly constant slope (Figure 44). This wedge is critical, meaning that any point within the wedge is at the verge of failure. In this experiment, the geometry of the wedge is dictated by the balance between frictional stresses at base of the wedge and stresses induced by the slope of the wedge (Chapple, 1978; Davis et al., 1983; Dahlen and Suppe, 1988, 1990). The self-similar growth of the wedge results from frontal accretion by southward propagation of the deformation front and from thickening and shortening of the wedge (Figure 43). For a homogeneous, noncohesive, brittle wedge obeying the Coulomb–Mohr failure criterion (e.g., Jaeger and Cook, 1979) (which is close to the Drucker–Prager criterion mentioned above), the taper angle, þ , assumed small, is approximately given by (Davis et al., 1983; Dahlen et al., 1984; Barr and Dahlen, 1990) ¼
9b – 29 1 þ 29
½22
where is the slope of the topography and is the slope of the basement. b9 and 9 characterize the basal friction and the compressive strength of the wedge, respectively, and are referred to as the effective basal friction and effective wedge strength (Barr and Dahlen, 1989). These values depend on the
422
Mountain Building: From Earthquakes to Geological Deformation
(a)
1
(b)
2
(c)
3
Figure 44 Example of an analog modeling of the growth of a brittle wedge. A rigid backstop moves to the left over a fixed rigid basement inducing detachment of the sand layers. Note the forward propagation sequence of thrusting. Courtesy of Ste´phane Dominguez, ISTEEM, Montpellier.
ratio l of the pore-fluid pressure to the lithostatic pressure (Hubbert and Rubey, 1959), according to 9b ¼ ð1 – b Þb 9 ¼ ð1 – Þ
sin 1 – sin
½23 ½24
where is the internal friction angle (the coefficient of internal friction is ¼ tan ) and b is the basal coefficient of friction. Equation [16] expresses how the topographic slope of a critical wedge depends on the dip angle of the basement and on the basal friction and strength of the wedge. It shows in particular that the slope of the wedge increases with the basal friction. Without erosion (as in Figure 44), the wedge would grow indefinitely and the force needed to drive deformation would increase as the square of the thickness of the wedge. This is because the force transmitted at the interface between the underthrusting basement and the overthrusting wedge is the integral of the basal shear stress along the length of the basal contact both of which increase linearly with the thickness of the self-similar wedge. If erosion is taken into account, as it should be to reproduce the evolution of a subaerial fold-andthrust belt or of a mountain range as whole, the
wedge reaches some steady-state geometry with particle trajectories like those pictured in Figure 45. It is then possible to derive an analytical expression of the velocity field (Barr and Dahlen, 1989). Horizontal shortening is distributed throughout the prism and the horizontal velocity at the surface can be derived from a simple mass balance (Avouac, 2003), yielding V ðx Þ ¼
V 0 ? h ? ðW – X Þ W ? ðH þ x ? tan Þ
½25
where the origin of the abscissa is taken at the backstop, assumed to be fixed. V0 is the convergence rate, hence the velocity of the leading edge of the wedge. This equation assumes that the erosion rate is uniform. In Figure 45, the predicted horizontal surface velocity was computed for h ¼ 15 km, W ¼ 200 km, tg ( þ ) ¼ 4%, and V0 ¼ 20 mm yr 1. This set of parameters corresponds approximately to the geometry and size of the Himalayan wedge of central Nepal, assuming that about half of the Indian crust is incorporated into the Himalayan wedge. For the wedge to be steady state, a 2-D erosion flux of 300 m2 yr 1 is then required (per unit length of the Himalayan arc). This is in the range of possible values inferred from the sediment budget over the last 2 My at the scale of the whole Himalayan range (Metivier et al., 1999), or from denudation rates
Horizontal velocity (mm yr–1)
Mountain Building: From Earthquakes to Geological Deformation
423
20
10
0
0
50
100 W
150
200 km
K h
Figure 45 Particle trajectories in an steady-state critical accretionary prism subjected to erosion (Barr, et al., 1991). Horizontal surface velocities computed from equation 19 for h ¼ 15 km, W¼200 km, tg(þ) ¼ 4%, and V0 ¼ 20 mm yr1.
derived from river sediment discharge in central Nepal (Lave´ and Avouac, 2001). This value corresponds to an erosion rate of 1.5 mm yr1 on average over the section. Based on the analytical expression of the velocity field, it is possible to compute the thermal structure from a thermokinematic approach and to estimate pressure–temperature–time (‘PTt’) paths (Barr and Dahlen, 1989; Barr et al., 1991). This kinematics predicts that in a steady-state orogenic prism, the peak metamorphic conditions increase from the toe to the rear of the wedge (Figure 45). This model can thus account for the northward increase of metamorphic grade from the Lesser to the High Himalaya across the MCT zone, a metamorphic zonation that is in fact quite common across mountain ranges. The critical wedge model has gained immediate success because it provides a simple and rigorous mechanical framework to analyze the structure and deformation of an orogen and its sensitivity to erosion and to the mechanical properties of the wedge. A number of possible variations around the initial ideas were explored by assuming different possible rheological laws, including the effects of (1) cohesion, (2) the transition to ductile deformation at depth, (3) different kinds of boundary conditions, (4) extending the model into 3-D, and (5) including more realistic erosion laws (e.g., (Dahlen et al., 1984;
Koons, 1989, 1994; Willett et al., 1993; Williams et al., 1994b; Willett, 1999). The data reviewed above show that the kinematics of deformation across the Nepal Himalaya turns out to be quite different from that predicted by the critical taper theory. Crustal shortening across the Nepal Himalaya has resulted in apparently little internal deformation of the orogenic wedge over the Holocene or over the longer geological period modeled with KTM11. In the KTM11 model, erosion maintains a balance between denudation and tectonic uplift, and a steady state can thus be reached through underplating of material. Although the possibility for underplating has been considered in some thermokinematic models (Barr and Dahlen, 1989), there is no theoretical analysis in the literature of why this mode of accretion arises and what controls the respective contribution of underplating and frontal accretion to the total flux of material into a wedge (Figure 46). This issue has been addressed from analog experiments, which have shown that underplating can be reproduced in sand wedge models under some conditions (Gutscher et al., 1998). Recently, Konstantinovskaia and Malavieille (2005) were able to reproduce underplating by the formation of a duplex. The structural evolution in their model shows striking similarities to that of the Himalayan orogen (Figure 47). The key factors in this experiment were (1) the introduction of a weak
424
Mountain Building: From Earthquakes to Geological Deformation
100°C 250°C 400°C
Figure 46 Particle trajectories and thermal structure in a steady-state critical wedge subjected to erosion and underplating. Adapted by Barr TD and Dahlen FA (1989) Brittle frictional mountain building. Part 2: Thermal structure and heat-budget. Journal of Geophysical Research-Solid Earth and Planets 94: 3923–3947.
Zone of maximum exhumation
Deformation stages for model thrust wedge with décollement level (HF6DL).
HF6DL
Synformal thrust stack FT1 FT3
115 cm
FT2
Erosion 6° FT2
68 cm
Proto-wedge
Detachment FT1
21 cm
Décollement level
20 cm
Décollement level
Figure 47 Evolution of an analog sand wedge with a glass bed decollement (Konstantinovskaia and Malavieille, 2005). Note the similarity with the proposed structural evolution of the Himalayan orogen.
horizontal layer of glass microbeads, and (2) erosion at the surface so that the topography was always maintained close to the critical slope (here controlled by the weaker layer). The model leads first to the development of a frontal wedge by frontal accretion, while the rear of the wedge is uplifted due to the growth of a duplex at depth, all accomplished with little shortening (Figure 47). The weak detachment is key to the formation of the duplex since it becomes the roof thrust of the system. There is no shortening of the medium above the roof thrust probably because the critical slope of the topography is controlled by the weak layers and is thus steeper than the slope required for brittle failure of the stronger sand layers above the roof thrust. So, the Himalaya is a case of an orogenic wedge in which the slope of the topography is steeper than, or equal to, the critical slope that would arise if the wedge were homogeneous and everywhere at the verge of brittle failure. Accretion through the development of a duplex at depth, coupled with erosion at the surface, is a viable mechanical explanation as demonstrated from the analog experiments of Konstantinovskaia
and Malavieille (2005). Because the frontal slope results from the erosional retreat of the advancing range front, erosion might play the key role of maintaining a slope steeper than the critical value for brittle deformation of the wedge. Combined with the effect of a weak basal friction, the topography must induce stresses that do not exceed the brittle strength of the medium within the wedge above the MHT. 6.09.7.2 MHT
Evidence for Low Friction on the
The observation that the overhanging wall of the MHT does not shorten during overthrusting allows the possibility of inferring the friction along the MHT. This geometry means that the slope of the topography must be overcritical. If the friction was too high, or the topographic slope too shallow, the wedge would be undercritical and the hanging wall would deform. There would be no slip on the MFT at the surface until the wedge topography reached the critical value. Given the slope of the Himalayan wedge in the LH and the estimated dip angle of the
Mountain Building: From Earthquakes to Geological Deformation
MHT (4 atmost beneath the LH), and assuming an internal friction of 0.85, the effective basal friction on the flat portion of the MHT is inferred from eqns [22] and [23] to be smaller than 0.12 (Cattin and Avouac, 2000). This is in keeping with the numerical model CA2000, which was found to require a basal friction less than 0.13. This value is consistent with the analysis of Davis et al. (1983) who considered the whole Himalayan wedge (which has a steeper slope than the LH portion) and obtained a somewhat larger effective basal friction of 0.25. Another constraint on the friction along the MHT zone comes from the observation that microseismic activity along the front of the high range shows a cutoff at an elevation of 3500 m (Section 6.09.6.3). This threshold is interpreted to correspond to the point where Coulomb stresses no longer increase during the interseismic period of increasing N–S compressive stress. This is achieved if the N–S deviatoric stress is of the order of magnitude of the vertical stress variation due to the increase in elevation across the range. When the elevation rises higher than 3500 m, the principal compressive stress is vertical and decreases during interseismic elastic strain (the effect of the Poisson coefficient is that vertical stress decreases as horizontal compression increases). Isostatic restoring forces and elastic stresses cause difficulties in estimating the effect of topography on the stress field at depth. To first order, the 2500 m elevation difference between the area with intense microseismic activity in the LH and the cutoff elevation must correspond to a zz increase of 70 MPa at most (for a crustal density of 2.6 g cm3). We infer that the shear stresses on the decollement cannot exceed 35 MPa. Given that the decollement lies at a depth of 10 km, this corresponds with a friction of the order of 0.1 (Bollinger et al., 2004b). A classical interpretation of low basal friction below thrust sheets invokes fluid pressurization (Hubbert and Rubey, 1959). Even if a high pore pressure is assumed, up to 0.9 times the lithostatic pressure, the friction on the flat portion of the MFT must be lower than 0.3. This value is still significantly smaller than the 0.6–0.8 value for intact rocks, but is in the range of empirical values inferred for some intracontinental thrust faults or along subduction zones (Davis, et al., 1983; Lamb, 2006). The low friction on the MHT could well be the result of high pore fluid pressure within the shear zone. This seems a reasonable scenario if some foreland sediments are dragged along the interface, as the relatively high conductivity along the MHT might
425
suggest (Figure 10). Fluid pressurization could then result from sediment compaction during shear (Sleep and Blanpied, 1992; Segall and Rice, 1995). Given that the MHT probably slips mainly during seismic slip events, the low apparent friction could well be due to dynamic effects such as frictional melting, hydrodynamic lubrication, or, in the presence of fluids, thermal fluid pressurization (Kanamori and Brodsky, 2004). Once these dynamic weakening mechanisms trigger an earthquake at some point on the MHT, the rupture could propagate along the MHT even if the preseismic stress is far from the critical stress needed for static slip.
6.09.7.3 Importance of the Brittle–Ductile Transition The variation with depth of rheology is key to understanding a number of aspects of the Himalayan orogen, including its seismicity, its pattern of geodetic deformation, and even its topography. Regarding this last point, it is noteworthy that the transition from the Himalayan range to the flat Tibet Plateau takes place close to where the mechanical coupling between the wedge and the underthrusting basement transitions from brittle to ductile behavior. The transition probably leads to a reduction of the basal shear stress and thus to a shallower critical topographic slope (Williams et al., 1994a). This feature arises naturally in thermomechanical models of orogens ((Toussaint et al., 2004a), as it does in the FEM models discussed in Section 6.09.5 (Figure 33). Temperature is probably the major controlling factor on this transition in behavior, as demonstrated from the discussion of geodetic deformation and its relation to the thermal structure. In this regard, observational data from the Himalaya are remarkably consistent with the wealth of laboratory data on the rheology of crustal rocks (e.g., Kohlstedt et al., 1995; Marone, 1998). As we have seen, the development of a mid-crustal duplex is a key feature in the structural evolution of the Himalaya. Although this feature can be reproduced from sand-box experiments (Konstantinovskaia and Malavieille, 2005), it is probable that, in nature, development of a duplex relates to the brittle–ductile transition, as suggested from the analysis of ancient exhumed analog systems (Dunlap et al., 1997). It is possible that rheological layering due to the thermal structure of an active orogen might be a primary factor allowing this kinematics, as suggested from the
426
Mountain Building: From Earthquakes to Geological Deformation
observation of a similar kinematics in thermomechanical modeling (Toussaint et al., 2004a). 6.09.7.4 Metamorphism during Underthrusting The rocks cropping out in the High Himalaya and in the LH have without question experienced metamorphic reactions during the development of the orogen. We have seen that model KTM11 predicts a PTt path in reasonably good agreement with the estimated peak temperatures and cooling histories along the section. The portion of the crust that underthrusts the high range and southern Tibet should also experience petrological transformations as the rocks are transported to greater depth and higher temperatures. Returning now to the interpretation of the electrical conductivity of Figure 10 (Lemmonier et al., 1999), for the comparison with the thermal structure (Figure 29) is instructive. Because the mid-crustal zone of high conductivity does not follow any particular isotherm, in particular since it does not extend to the north, a thermometamorphic control model can be excluded. This conductive body, which becomes prominent where the underthrusting footwall reach temperatures 350 C, may instead result from the presence of fluids released by metamorphic reactions, especially those involving chlorite breakdown. The isotopic composition of hot springs located at the front of the high range, on top of the conductivity anomaly, do suggest some contamination by metamorphic fluids (Kotarba, 1986; Evans et al., 2001). The fluids released by metamorphic reactions would percolate upward through the zone where intense microseismic activity is triggered by interseismic strain, allowing for some connectivity of the fluid phase that is required for it to be detectable by magnetotelluric sounding. The thermal structure might also be used to estimate how the crustal density should vary given the predicted PT conditions along the section (Henry et al., 1997; Cattin et al., 2001) (Figure 48). A prediction of the model is that (provided that the kinetics of metamorphic reaction can be neglected) the underthrusted crust should enter the granulite and then the eclogite facies. However, the resulting density distribution would be inconsistent with the observed gravity. The misfit between the predicted and observed Bouguer anomaly is as high as 150 mGal in southern Tibet (Model 1 in Figure 48), suggesting that eclogitization does not occur as extensively as
this model would imply. A much better fit of the gravity measurements is obtained if the lower crust is assumed to be metastable (Model 2 in Figure 48). Recent seismological investigations across the eastern Nepal Himalaya did reveal seismic velocities in the lower crust suggesting partial eclogitization may be as high as 30% (Schulte-Pelkum et al., 2005). As argued by Cattin et al. (2001) and Jackson et al. (2004), one possible explanation of this is that eclogitization is limited by the availability of water, which is known to be the critical factor limiting metamorphic processes in the lower crust (Austrheim, 1990). Eclogitization of the lower crust beneath the High Himalaya and southern Tibet might proceed slowly, controlled by the availability of fluid and its circulation. It is possible that some fraction of the underthrusting Indian crust becomes eclogitized and subducted into the mantle, as shown in the cartoon of Figure 49. 6.09.7.5 How Does the Steep Front of the High Himalaya Relate to Tectonics, Erosion, and Climate? The sudden 2–3 km increase in the topography at the front of the high range, almost 100 km north of the trace of the Main Frontal Fault (Figure 8), has been interpreted in various ways. Because this area coincides with a zone of localized ongoing uplift, as revealed from leveling data, the topographic step has been attributed to active thrusting at the front of the high range (Bilham et al., 1997). There is indeed evidence for some neotectonic activity at the front of the high range (Hodges et al., 2004). It has been proposed that the lower crust of Tibet would be extruded from below the high plateau toward the front of the high range due to coupling between intense localized erosion and uplift (Beaumont et al., 2001). The steep front of the High Himalaya has also been interpreted as the preserved morphologic signature of a Late Miocene reactivation of the MCT zone (Harrison, et al., 1997). Others have proposed that headward regressive erosion along the rivers cutting the edge of the Tibetan Plateau would have induced uplift of the Himalayan peaks through isostatic rebound, enhancing orographic precipitation and hence denudation (Molnar, 1990; Burbank, 1992; Montgomery, 1994; Masek, 1994b) (Figure 50). The morphology of the front of the High Himalaya of Nepal undoubtedly reflects some dynamic equilibrium between tectonic deformation and surface processes. Indeed, in the absence of
Mountain Building: From Earthquakes to Geological Deformation
Model 1
10 Depth (km)
427
2670 2750
50
2850 2950
90 3050
Model 2
Depth (km)
10
3150 3250
50
3350
density (kg m–3) 90
–100
(mgal)
–200 –300 Model 1 –400
Mo
de
–500 0
100
200
l2 300
500
400
Distance from the MFT (km)
Figure 48 Density model deduced from a steady-state thermopetrological model showing eclogitization of the lower crust beneath the Himalayan range (Cattin et al., 2001) computed based on the petrogenetic grid of Bousquet et al. (1997). White lines show constraints on Moho depths derived from INDEPTH (dashed line from seismic profile and vertical bars from receiver functions). The Bouguer anomaly (dashed line) associated with the density of Model 1, is found to be inconsistent with the data (gray shading), by as much as 150 mGal under Tibet. It therefore suggests that eclogitization does not occur as south as shown in Model 1. The dashed curve exhibits a warped anomaly (100–250 km) with wavelength and amplitude strikingly comparable with those observed in the data north of the high range (200–350 km). Model 2 (black line) was modified model assuming that eclogitization is delayed by 6.5–8.5 My, possibly due to the kinetics of petrological changes.
Himalayan wedge Suture
Tibet
Lithosphere
Pure shear deformation
Indian crust Mantle
Asian crust
Moho
Partial eclogitization? Eclogitized crust?
Figure 49 Cartoon of the collision zone illustrating the possibility for partial eclogitization of the underthrusted Indian crust and continental subduction. A fraction of the Indian crust is incorporated into the Himalayan wedge and is ultimately eroded away. The remaining fraction is either injected beneath southern Tibet, as suggested, for example, by Zhao and Morgan (1985) or eclogitized and subducted, as suggested, for example, by Le Pichon et al. (1992).
428
Mountain Building: From Earthquakes to Geological Deformation
Enhanced erosion >> uplift
Convergence >> uplift
Uplift >> enhanced precipitations Enhanced precipitations >> enhanced erosion
Figure 50 Coupling between mountain uplift, climate, and erosion. Modified from Masek JG, Isacks BL, and Fielding E (1994) Rift flank uplift in Tibet: Evidence for a viscous lower crust. Tectonics 13: 659–667.
tectonic uplift, the various rivers cutting across the Himalaya would be retreating at very different rates and the edge of the Tibetan Plateau, instead of having the straight and sharp edge that is observed today, would rapidly evolve a sinuous geometry (Lave´ and Avouac, 2001). The steep relief and the high uplift rates in the High Himalaya must therefore correlate with a zone of more localized uplift. This reasoning implies that the present morphology of the front of the high range more probably reflects present tectonics and erosion than it does a Late Miocene reactivation of the MCT (Harrison et al., 1997; Catlos et al., 2001). It is improbable that a significant contribution to uplift takes place by active thrusting at the front of the high range, as suggested by Bilham et al. (1997); Hodges et al. (2004); Wobus et al. (2005)). This scenario would imply some amount of shortening rate to be added to the 21 mm yr1 absorbed at the MFT, and the total shortening rate across the range would subsequently exceed the geodetic rate significantly. The idea that the Himalaya would currently be uplifted by extrusion of the lower crust from beneath the high plateau (Beaumont et al., 2001) can therefore also be excluded. However, there is geological evidence that some kind of channel-flow tectonics may have happened in the past (Grujic, et al., 1996), at the time of coeval motion on the MCT and STD. As we have seen, the present topography of the range can be reproduced reasonably well if localized uplift is assumed to result simply from thrusting over
a mid-crustal ramp (with a dip angle on the order of only 15 ). This is demonstrated from the success of the FEM models in reproducing the observed topography as well as the observed pattern of erosion across the range (Figures 32 and 33) (Godard et al., 2006). The available data on active tectonics in the Himalaya and on fluvial incision are thus consistent with the simple view that active tectonics in the Himalaya are primarily controlled by localized slip along the MHT, with topography being close to steady state, as denudation driven by fluvial downcutting balances tectonic uplift. 6.09.7.6 The Elevation and Support of Mountain Ranges: Effect of Climate and Lower Crustal Flow It has long been recognized that isostasy is a major factor allowing for the support of high topography (Airy, 1855; Pratt, 1855). In this view, some deficit of mass at depth compensates for the excess of mass due to the topography (by virtue of Archimedes’ Principle). In the case of the Himalaya and southern Tibet, the deficit of mass is clearly due primarily to the exceptional Moho depth, which permits an almost local isostatic balance (Figure 5). At first glance, this means the elevation of the range is directly related to the thickness of the crust. But this is only an approximation because the presence of elastic bending stresses means that balance is only achieved at a regional scale. The mass deficit below
Mountain Building: From Earthquakes to Geological Deformation
the foreland and the mass excess below the high topography is a direct manifestation of this effect (e.g., Lyon-Caen and Molnar, 1985). 6.09.7.6.1 Height and width of a critical brittle wedge
Let us consider first that the critical brittle wedge theory applies at the scale of the whole orogen. Let us then imagine a brittle wedge over a basement at isostatic balance (the inclusion or exclusion of flexural stresses in the basement makes no difference in this reasoning), with a rigid backstop. Examples of such numerical experiments can be found in Hilley and Strecker (2004) and Whipple and Meade (2004). The width and height of such an orogen depends on these two factors in a rather simple way, which is amenable to analytical formulations. If the parameters characterizing erosion rates are kept constant (‘climate’ is assumed constant), the wedge will tend toward a steady-state geometry in which the eroded flux balances the accreted flux. Accordingly, the volume and maximum height of the orogenic wedge depends on the rate of accretion of material. If ‘climate’ changes the internal deformation of the wedge, the wedge adjusts its width to that the critical slope is still maintained and the eroded and accretion flux balance each other. In this case the climate change would have no impact on the erosion flux. Similarly, if the accreted flux is varied the wedge will also adjust its width. This means that, except for transients, a change in erosion flux can only relate to a change in tectonic forcing. It should be noted that the tectonic forcing in this reasoning is described in terms of an accretion flux that depends on the convergence rate, which is assumed constant. The assumption of a constant convergence rate implies that any change in the wedge geometry has to be associated with a change in the driving force. Namely, if climate change is assumed to lead to more rapid erosion, the wedge will shrink so that the force transmitted by friction along the basal decollement will decrease (as the square of the width or height of the orogen). The assumption of a constant convergence rate might be appropriate in the case of a small orogen, for example Taiwan, which would not affect the force balance at the scale of the tectonic plates. In the case of a mountain range the size of the Himalaya, the assumption of a constant driving force can be argued as more appropriate. Let us now characterize tectonic forcing as the force applied to the system. In this case, a constant
429
tectonic forcing implies constant wedge geometry. If climate change leads to increased erosion, the convergence rate would need to increase so that the accretion flux would still compensate erosion. According to this reasoning, the height and width of a critical brittle wedge is ultimately determined by the force applied to the system, that is, the horizontal forces associated with plate tectonics (which presumably vary less rapidly than climate). In the case of the Himalaya, it would be primarily the force arising from the excess buoyancy of the Indian Ocean ridge system that is transmitted across the Indian continent (excluding the eventual effect of drag at the base of the Indian lithosphere) (e.g., Sandiford et al., 1995). Climate then determines both the erosion flux (with some complex transfer function that depends on the erosion law) and the convergence rate. 6.09.7.6.2 Effect of ductile deformation in the lower crust
In the real world, an orogenic wedge cannot be considered a brittle wedge at the scale of the crust. One implication of the dominantly ductile rheology is that an isostatically compensated mountain range is not in hydrostatic balance. The result is that the ductile lower crust will tend to be expelled from below the high topography (Gratton, 1989; Bird, 1991; Avouac and Burov, 1996; McKenzie and Jackson, 2002) (Figure 51). This process is advocated to explain the eastward growth of the Tibetan Plateau, for example (e.g., Clark and Royden, 2000; Clark et al., 2004). In the absence of erosion, a mountain range collapses with a decay time that depends only on the rheology of the lower crust. Let us consider an isolated mountain range, one the size of the Tien Shan, for example, that is 300 km wide and not associated with a plateau like the Himalaya. It would take between 3 and 10 Ma for an initial average relief of 3000 m to be reduced to half as a result of this process alone (Avouac and Burov, 1996). The decay time increases with the wavelength of the topography (Bird, 1991). If we now consider the same isolated mountain range submitted only to erosion, the relief would also decay with a similar or somewhat shorter decay time, assuming erosional parameters consistent with the curent (<1 mm yr1) erosion rate in this particular mountain range (Avouac and Burov, 1996; Me´tivier and Gaudemer, 1997). Now, when crustal shortening and erosion are considered together, a positive feedback arises (Avouac and Burov, 1996) (Figure 51). Rather than resulting in a
430
Mountain Building: From Earthquakes to Geological Deformation
No erosion
Erosion Sedimentaion
Sedimentaion
Moho
Figure 51 Coupling between mountain building and erosion (Avouac and Burov, 1996). In the absence of erosion any mountain range is expected to collapse as a result of ductile flow in the lower crust (top). In the presence of erosion and sedimentation, a positive feedback can arise favoring focused shortening below the mountain range, where vertical advection of heat results in thermal weakening of the crust (bottom). In that case, a mountain range can grow until some steady-state geometry is reached.
topography decaying even more rapidly, this feedback allows an initial topography to grow until it reaches some steady state. The dynamic equilibrium between denudation and tectonic uplift then arises naturally from the coupling between surface processes and ductile flow at depth, and also contributes to localizing the deformation (Avouac and Burov, 1996). It should be noted that the thermal evolution of the range is a key factor in allowing for this feedback to happen. The hotter and weaker crust below the range deforms more easily, favoring localized deformation there, and enhanced erosion leads to upward advection of heat and hence further thermal weakening of the crust below the range (Koons, 1987). In this case, the relationship between the geometry of the orogen and climate forcing is more complex than it is in the case of a critical brittle wedge, because the geometry of the range is dependent on strain rates (due to the power-law rheology ascribed to the ductile crust) and the relative contribution of brittle and ductile stresses to the force balance varies as the geometry of the wedge and its
thermal structure change. Qualitatively, the model yields that if the system is submitted to constant driving forces, like enhanced erosion due to climate, the shortening rate across the range should increase just as in the reasoning based on the critical brittle wedge model. The elevation and width of the orogen should then be reduced to compensate for the strainrate hardening effect, which is presumably small. The coupling described here is more complex than the isostatic adjustment of an eroding range in local isostasy described by some authors (e.g., Molnar and England, 1990a), but it does include such a component. The coupling leads to the additional fundamental implication that a change in climate might also induce a change in convergence rate. To a first order, the topography of a range should directly gauge the driving tectonic forces and depend only marginally on climate. Again, if the system is to be driven by constant tectonic forces, climate would affect primarily the erosion flux. If the system is assumed to be driven by a constant convergence rate, then the erosion flux would depend only weakly on climate. 6.09.7.7 The Fate of the Indian Crust and Mantle Lithosphere It is clear that, following a continental collision such as the India–Asia collision, most of the crust is detached from the mantle lithosphere. A fraction, estimated at 1/3, of the Indian crust is then incorporated into the Himalayan wedge and is ultimately eroded away when the orogenic wedge volume reaches steady state. The remaining fraction is either injected beneath southern Tibet, as suggested by Zhao and Morgan (1985), or eclogitized and subducted (Figure 49). These are only crude estimates, and what happens to the lithospheric mantle remains poorly known. It is probable that at the onset of the collision subduction continues for a while until the oceanic slab breaks away and sinks into the mantle under its own weight. Slab breakoff seems necessary because the subducting continental mantle lithosphere is already quite hot and weak and will therefore tend to resist subduction, due to its buoyancy, and instead neck easily (Davies and Vonblanckenburg, 1995). There is indication that slab break did indeed occur in the early stages of the India–Asia collision (Kohn and Parkinson, 2002). It is also possible that the mantle lithosphere may have continued to subduct (Mattauer, 1986), as shown in the cartoon of Figure 49, and as argued
Mountain Building: From Earthquakes to Geological Deformation
from the interpretation of tomographic results (Replumaz et al., 2004). It should be noted that while continental subduction might be taking place beneath southern Tibet, it is not associated with the kind of seismicity that characterizes oceanic subduction zones (e.g., the ‘Benioff zone’). This is not surprising, because the subducting mantle lithosphere is probably too hot to be seismogenic. Also, the presumably small density contrast between the subcontinental lithosphere and the upper mantle makes it difficult to detect from tomographic studies. Another possibility is that the mantle lithosphere thickened homogeneously (England and Mckenzie 1983; Houseman and England 1986) and ultimately became unstable and was removed by convection (Houseman et al., 1981). It has been proposed that convective removal of the mantle lithosphere beneath Tibet would have occurred at c. 8 Ma, leading to a sudden increase of the elevation of the plateau, which would in turn have strengthened the Monsoon (Molnar et al., 1993). This scenario seems at odds with recent advances in paleo-altymetry, which suggests that the Tibetan Plateau had reached close to its present elevation probably before c. 15 Ma (Garzione et al., 2000; Spicer et al., 2003; Currie et al., 2005), or even before 35 Ma (Rowley and Currie, 2006).
6.09.8 Conclusions The Himalaya is a unique setting in which to address a variety of geological problems associated with mountain-building processes. Some progress over the last decades in understanding these processes results from various multidisciplinary efforts that have provided important information on the subsurface structure of the range and on the kinematics of deformation over different timescales. At this point, the available data can be assembled in a relatively consistent way in light of our current understanding of the mechanics of the lithosphere and of the seismic cycle. The kinematics observed in the Nepal Himalaya, described in some detail in this chapter, might be specific to some aspect of the Himalayan setting. Other mountain ranges located in a different climatic and tectonic setting might evolve quite differently, and analyzing these differences would probably help in the understanding of some aspect of the mechanics governing collisional mountain building. One example of a persisting puzzle is that shortening across the
431
Himalaya of central Nepal has been absorbed by localized thrusting on only one single major fault along its piedmont, the MFT, but it is clear that shortening across the Tien Shan is absorbed by thrusting on a number of faults distributed within the range and along its northern and southern piedmont (Thompson et al., 2002). The example of Taiwan is closer to that of an intermediate Himalaya, with shortening being absorbed on several thrust faults within the western foothills (Simoes et al., in press). The mechanical reasons for these differences are unclear and need to be clarified. In addition, the focus has been put here on active deformation and on geological deformation over the last c. 15 Ma. This time covers only half the life of the Himalayan orogen. A number of fundamental geodynamic processes occurred earlier and have also imprinted the Himalayan orogen, in particular through their influence on magmatism. The Tethys oceanic slab probably detached early on in subduction (Kohn and Parkinson, 2002), and likely induced a significant change in the balance of forces across the collision zone. Some crustal material underthrusted to depths >100 km were exhumed quite early in the history of the collision (Guillot et al., 1997; DeSigoyer et al., 2000); this scenario requires a tectonic regime fundamentally different from the one that has prevailed over the last 15 Ma. Finally, slip on a major normal fault, the STD, coeval with the onset of motion on the MCT at c. 20 Ma, led to a rapid episode of unroofing of the High Himalaya just before the current tectonic regime was established. The observation that nearly half the convergence between stable Eurasia and India is currently being absorbed by thrusting along the Himalayan arc, is a clear indication that deformation of the brittle upper crust is highly localized. This finding does not conflict with the observation that the zone with finite strain across the Himalaya and southern Tibet is more than 100 km wide. Indeed the locus of active localized deformation has moved with time as material was accreted onto the orogen. In the author’s view, highly localized active deformation most probably also reflects localized strain at the scale of the lithosphere, possibly related to continental subduction. Also, it has been argued based on thermomechanical modeling that surface erosion not only contributes to localize deformation within the crust (Avouac and Burov, 1996) but also favors continental subduction and inhibits the possibility for thickening of the mantle and its subsequent delamination (Pysklywec, 2006).
432
Mountain Building: From Earthquakes to Geological Deformation
References Abtout A (1987) La de´termination des anomalies des champs de potentiel (magne´tisme, gravime´trie) et la structure de la lithosphe`re continentale en Asie. Airy GB (1855) On the computations of the effect of the attraction of the mountain masses as disturbing the apparent astronomical latitude of stations in geodetic surveys. Philosiophical Transactions of the Royal Society of London 145: 101–104. Ambraseys N (2000) Reappraisal of north-Indian earthquakes at the turn of the 20th century. Current Science 79: 1237–1250. Ambraseys N and Bilham R (2000) A note on the kangra M-s ¼ 7.8 earthquake of 4 April 1905. Current Science 79: 45–50. Ambraseys NN and Douglas J (2004) Magnitude calibration of north Indian earthquakes. Geophysical Journal International 159: 165–206. Ambraseys N and Jackson D (2003) A note on early earthquakes in northern India and southern Tibet. Current Science 84: 570–582. Appel E and Rossler W (1994) Magnetic polarity stratigraphy of the Neogene Surai Khola section (Siwaliks, SW Nepal). Himalayan Geology 15: 63–68. Argand E (1924) La tectonique de l’Asie. 13th International Geological Congress: Brussels 170–372. Appel E, Rosler W, and Corvinus G (1991) Magnetostratigraphy of the Miocene Pleistocene Surai-Khola Siwaliks in west Nepal. Geophysical Journal International 105: 191–198. Arita K, Dallmeyer RD, and Takasu A (1997) Tectonothermal evolution of the Lesser Himalaya, Nepal: Constraints from Ar-40/Ar-39 ages from the Kathmandu Nappe. Island Arc 6: 372–385. Arita K and Ganzawa Y (1997) Thrust tectonics and uplift proccess of the Nepal Himalaya revealed from fission-track ages (in Japanese with English abstract). Journal of Geography (Tokyo Geographical Society) 106: 156–167. Armijo R, Tapponnier P, Mercier JL, and Han TL (1986) Quaternary extension in southern Tibet – Field observations and tectonic implications. Journal of Geophysical ResearchSolid Earth and Planets 91: 13803–13872. Austrheim H (1990) The granulite eclogite facies transition – A comparison of experimental work and a natural occurrence in the bergen arcs, western Norway. Lithos 25: 163–169. Avouac JP (2003) Mountain building, erosion and the seismic cycle in the Nepal Himalaya. In: Dmowska R (ed.) Advances in Geophysics, pp. 1–79. Amsterdam: Elsevier. Avouac JP, Ayoub F, Leprince S, Konca O, and Helmberger DV (2006) The 2005, M-w 7.6 Kashmir earthquake: Sub-pixel correlation of ASTER images and seismic waveforms analysis. Earth and Planetary Science Letters 249: 514–528. Avouac JP and Burov EB (1996) Erosion as a driving mechanism of intracontinental mountain growth. Journal of Geophysical Research-Solid Earth 101: 17747–17769. Avouac J and Tapponnier P (1993) Kinematic model of active deformation in central Asia. Geophysical Research Letters 20: 895–898. Avouac JP, Tapponnier P, Bai M, You H, and Wang G (1993) Active thrusting and folding along the northern Tien-Shan and Late Cenozoic rotation of the Tarim relative to Dzungaria and Kazakhstan. Journal of Geophysical Research-Solid Earth 98: 6755–6804. Baranowski J, Armbruster J, Seeber L, and Molnar P (1984) Focal depths and fault plane solutions of earthquakes and active tectonics of the Himalaya. Journal of Geophysical Research 89: 6918–6928. Barr TD and Dahlen FA (1989) Brittle frictional mountain building. Part 2: Thermal structure and heat-budget. Journal
of Geophysical Research-Solid Earth and Planets 94: 3923–3947. Barr TD and Dahlen FA (1990) Constraints on friction and stress in the Taiwan fold-and-thrust belt from heat flow and geochronology. Geology 18: 111–115. Barr TD, Dahlen FA, and McPhail DC (1991) Brittle frictional mountain building. Part 3: Low-grade metamorphism. Journal of Geophysical Research-Solid Earth and Planets 96: 10319–310338. Beaumont C, Jamieson RA, Nguyen MH, and Lee B (2001) Himalayan tectonics explained by extrusion of a lowviscosity crustal channel coupled to focused surface denudation. Nature 414: 738–742. Beaumont C, Jamieson RA, Nguyen MH, and Medvedev S (2004) Crustal channel flows. Part 1: Numerical models with applications to the tectonics of the Himalayan-Tibetan orogen. Journal of Geophysical Research-Solid Earth 109: 2004. Beck RA, Burbank DW, Sercombe WJ, et al. (1995) Stratigraphic evidence for an early collision between northwest India and Asia. Nature 373: 55–58. Bernard S, Avouac JP, Dominguez S, and Simoes M (2007) Kinematics of fault-related folding derived from a sanbox experiment. Journal of Geophysical Research 112: B03S12. Bettinelli P, Avouac JP, Flouzat M, et al. (2006) Plate motion of India and interseismic strain in the Nepal Himalaya from GPS and DORIS measurements. Journal of Geodesy 80: 567–589. Beyssac O, Bollinger L, Avouac JP, and Goffe B (2004) Thermal metamorphism in the lesser Himalaya of Nepal determined from Raman spectroscopy of carbonaceous material. Earth and Planetary Science Letters 225: 233–241. Beyssac O, Goffe B, Chopin C, and Rouzaud J-N (2002) Raman spectra of carbonaceous material in metasediments: A new geothermometer. Journal of Metamorphic Geology 20: 859–871. Bilham R (1995) Location and magnitude of the 1833 Nepal earthquake and its relation to the rupture zones of contiguous Great Himalayan earthquakes. Current Science 69: 101–128. Bilham R (1997) The elusive height of Mount Everest: Everest. National Geographic Society 26–27. Bilham R (2004) Earthquakes in India and the Himalaya: Tectonics, geodesy and history. Annals of Geophysics 47: 839–858. Bilham R and Ambraseys N (2005) Apparent Himalayan slip deficit from the summation of seismic moments for Himalayan earthquakes, 1500–2000. Current Science 88: 1658–1663. Bilham R, Blume F, Bendick R, and Gaur VK (1998) Geodetic constraints on the translation and deformation of India: Implications for future great Himalayan earthquakes. Current Science 74: 213–229. Bilham R, Larson K, and Freymueller J (1997) GPS measurements of present-day convergence across the Nepal Himalaya. Nature 386: 61–64. Bilham R and Wallace K (2005) Future Mw > 8 earthquakes in the Himalaya: Implications from the 26 Dec 2004 Mw ¼ 9.0 earthquake on India’s eastern plate margin. Geological Survey of India Special Publication 85: 1–14. Bird P (1991) Lateral extrusion of lower crust from under high topography, in the isostatic limit. Journal of Geophysical Research-Solid Earth and Planets 96: 10275–10286. Blanpied ML, Lockner DA, and Byerlee JD (1991) Fault stability inferred from granite sliding experiments at hydrothermal conditions. Geophysical Research Letters 18: 609–612. Blanpied ML, Lockner DA, and Byerlee JD (1995) Frictional slip of granite at hydrothermal conditions. Journal of Geophysical Research-Solid Earth 100: 13045–13064.
Mountain Building: From Earthquakes to Geological Deformation Bollinger L (2002) De´formation de l’Himalaya du Ne´pal, PhD Thesis, 400 pp, University of Paris-Sud XI. Bollinger L, Avouac JP, Beyssac O, et al. (2004a) Thermal structure and exhumation history of the lesser Himalaya in central Nepal. Tectonics 23: (doi:10.1029/2003TC001564). Bollinger L, Avouac JP, Cattin R, and Pandey MR (2004b) Stress buildup in the Himalaya. Journal of Geophysical Research 109: (doi:10.129/2003JB002911). Bollinger L, Henry P, and Avouac JP (2006) Mountain building in the Nepal Himalaya: Thermal and kinematic model. Earth and Planetary Science Letters 244: 58–71. Bott MHP and Dean DS (1973) Stress diffusion from plate boundaries. Nature 243: 339–341. Bousquet R, Goffe B, Henry P, LePichon X, and Chopin C (1997) Kinematic, thermal and petrological model of the Central Alps: Lepontine metamorphism in the upper crust and eclogitisation of the lower crust. Tectonophysics 273: 105–127. Bowin C, Purdy GM, Johnston C, et al. (1980) Arc-continent collision in Banda Sea region. Aapg Bulletin-American Association of Petroleum Geologists 64: 868–915. Brown LD, Zhao W, Nelson KD, et al. (1996) Bright spots, structure, and magmatism in southern Tibet from INDEPTH seismic reflection profiling. Science 274: 1688–1690. Brunel M (1983) Etude pe´tro-structurale des chevauchements ductiles en Himalaya (Ne´pal oriental et Himalaya du NordOuest), The`se doctorat d’E´tat Thesis, Paris VII, Paris. Brunel M (1986) Ductile thrusting in the Himalaya – Shear sense criteria and stretching lineations. Tectonics 5: 247–265. Brunel M, Colchen M, Le Fort P, Mascle G, and Peˆcher A (1979) Structural analysis and tectonic evolution of the central Himalaya of Nepal. In: Saklani PS (ed.) Structural Geology of the Himalaya, pp. 247–264. New-Delhi: Today and Tomorrow’s Printers and Publishers. Brunel M and Kienast JR (1986) Etude pe´tro-structurale des chevauchements ductiles himalayens sur la transversale de l’Everest-Makalu (Ne´pal oriental). Canadian Journal of Earth Sciences 23: 1117–1137. Bull W (1991) Geomorphic Response to Climatic Change, 326 pp. New York: Oxford University Press. Burbank DW (1992) Causes of recent Himalayan uplift deduced from deposited patterns in the Ganges Basin. Nature 357: 680–683. Burbank DW, Blythe AE, Putkonen J, et al. (2003) Decoupling of erosion and precipitation in the Himalaya. Letters to Nature 426: 652–655. Burbank DW, Leland J, Fielding E, et al. (1996) Bedrock incision, rock uplift and threshold hillslopes in the northwestern Himalaya. Nature 379: 505–510. Burchfiel BC, Zhiliang C, Hodges KV, et al. (1992) The southern Tibetan detachment system, Himalayan Orogen: Extension contemporaneous with and parallel to shortening in a collisional mountain belt. Geological Society of America Special Paper 269: 1–41. Burg JP (1983) Himalayan orogen and global tectonics seen from the Tsangpo suture zone of Tibet (China). In: Sinha AK (ed.) Himalayan Orogen and Global Tectonics, pp. 35–44. New-Delhi: Oxford & IBH Publishing. Burg JP, Brunel M, Gapais D, Chen GM, and Liu GH (1984) Deformation of leucogranites of the crystalline Main Central Sheet in southern Tibet (China). Journal of Structural Geology 6: 532–542. Burg JP, Leyreloup A, Girardeau J, and Chen GM (1987) Structure and metamorphism of a tectonically thickened continental-crust – The Yalu Tsangpo suture zone (Tibet). Philosophical Transactions of the Royal Society of London Series A-Mathematical Physical and Engineering Sciences 321: 67–86.
433
Burov EB and Diament M (1995) The effective elastic thickness (T-E) of continental lithosphere – What does it really mean? Journal of Geophysical Research-Solid Earth 100: 3905–3927. Byerlee J (1987) Friction of rocks. Pure and Applied Geophysics 116: 615–626. Carter NL and Tsenn MC (1987) Flow properties of continental lithosphere. Tectonophysics 136: 27–63. Catlos EJ, Harrison TM, Kohn MJ, et al. (2001) Geochronologic and thermobarometric constraints on the evolution of the Main Central Thrust, central Nepal Himalaya. Journal of Geophysical Research-Solid Earth 106: 16177–16204. Catlos EJ, Harrison TM, Manning CE, et al. (2002) Records of the evolution of the Himalayan orogen from in situ Th–Pb ion microprobe dating of monazite: Eastern Nepal and western Garhwal. Journal of Asian Earth Sciences 20: 459–479. Cattin R and Avouac JP (2000) Modeling mountain building and the seismic cycle in the Himalaya of Nepal. Journal of Geophysical Research-Solid Earth 105: 13389–13407. Cattin R, Martelet G, Henry P, Avouac JP, Diament M, and Shakya TR (2001) Gravity anomalies, crustal structure and thermo–mechanical support of the Himalaya of central Nepal. Geophysical Journal International 147: 381–392. Chander R (1989) Southern limits of major earthquake ruptures along the Himalaya between longitudes 75-degrees and 90degrees-E. Tectonophysics 170: 115–123. Chapple WM (1978) Mechanics of thin-skinned fold-and-thrust belts. Geological Society of America Bulletin 89: 1189–1198. Chemenda AI, Burg JP, and Mattauer M (2000) Evolutionary model of the Himalaya-Tibet system: Geopoem based on new modelling, geological and geophysical data. Earth and Planetary Science Letters 174: 397–409. Chen QZ, Freymueller JT, Wang Q, Yang ZQ, Xu CJ, and Liu JN (2004) A deforming block model for the present-day tectonics of Tibet. Journal of Geophysical Research-Solid Earth 109: B01403. Chen WP and Molnar P (1977) Seismic moments of major earthquakes and average rate of slip in central Asia. Journal of Geophysical Research 82: 2945–2969. Clark MK and Royden LH (2000) Topographic ooze: Building the eastern margin of Tibet by lower crustal flow. Geology 28: 703–706. Clark MK, Schoenbohm LM, Royden LH, et al. (2004) Surface uplift, tectonics, and erosion of eastern Tibet from largescale drainage patterns. Tectonics 23: TC1006. Coblentz DD, Zhou SH, Hillis RR, Richardson RM, and Sandiford M (1999) Topography, boundary forces, and the Indo-Australian intraplate stress field. Journal of Geophysical Research-Solid Earth 103: 919–931. Cohen SC (1999) Numerical models of crustal deformation in seismic zones. Advances in Geophysics 41: 134–231. Copeland P (1997) The when and where of the growth of the Himalaya and the Tibetan plateau. In: Ruddiman WF (ed.) Tectonic Uplift and Climatic Change, pp. 19–40. New York: Plenum Press. Copeland P, Harrison TM, Hodges KV, Maruejol P, Lefort P, and Pecher A (1991) An early pliocene thermal disturbance of the main central thrust, central Nepal – Implications for Himalayan tectonics. Journal of Geophysical Research-Solid Earth and Planets 96: 8475–8500. Corvinus G (1988) The Mio-Plio-pleistocene litho- and biostratigraphy of the Surai khola Siwaliks in west Nepal: First results. Comptes Rendus des Se´ances de l’Academie des Sciences Paris, Se´rie D 306: 1471–1477. Cotton F, Campillo M, Deschamps A, and Rastogi BK (1996) Rupture history and seismotectonics of the 1991 Uttarkashi, Himalaya earthquake. Tectonophysics 258: 35–51. Currie BS, Rowley DB, and N Tabor J (2005) Middle Miocene paleoaltimetry of southern Tibet: Implications for the role of
434
Mountain Building: From Earthquakes to Geological Deformation
mantle thickening and delamination in the Himalayan orogen. Geology 33: 181–184. Daeron M, Avouac JP, and Charreau J (2007) Modeling the shortening history of a fault-tip fold using structural and geomorphic records of deformation. Journal of Geophysical Research 112: B03S13. Dahlen FA (1990) Critical taper model of fold-and-thrust belts and accretionary wedges. Annual Review of Earth and Planetary Sciences 18: 55–99. Dahlen FA and Suppe J (1988) Mechanics, growth and erosion of moutain belts. Special Paper Geological Society of America 218: 161–178. Dahlen FA, Suppe J, and Davis D (1984) Mechanics of fold-andthrust belts and accretionary wedges: Cohesive Coulomb theory. Journal of Geophysical Research 89: 10087–10101. Dahlstrom CDA (1969) Balanced cross sections. Canadian Journal of Earth Science 6: 743–757. Davies JH and Vonblanckenburg F (1995) Slab breakoff – A model of lithosphere detachment and its test in the magmatism and deformation of collisional orogens. Earth and Planetary Science Letters 129: 85–102. Davis D, Suppe J, and Dahlen FA (1983) Mechanics of fold-andthrust belts and accretionary wedges. Journal of Geophysical Research 88: 1153–1172. DeCelles PG, Gehrels GE, Quade J, Ojha TP, Kapp PA, and Upreti BN (1998) Neogene foreland basin deposits, erosional unroofing, and the kinematic history of the Himalayan foldthrust belt, western Nepal. Geological Society of America Bulletin 110: 2–21. DeCelles PG, Robinson DM, Quade J, et al. (2001) Stratigraphy, structure, and tectonic evolution of the Himalayan fold-thrust belt in western Nepal. Tectonics 20: 487–509. Delcaillau B (1986) Dynamique et e´volution structurale du pie´mont frontal de l’Himalaya: Les Siwaliks du Ne´pal oriental. Revue de Ge´ologie Dynamique et de Ge´ographie Physique 27: 319–337. Deplus C, Diament M, Hebert H, et al. (1998) Direct evidence of active deformation in the eastern Indian Ocean plate. Geology 26: 131–134. Derry LA and France-Lanord C (1997) Himalayan weathering and erosion fluxes: Climate and tectonic controls. In: Ruddiman WF (ed.) Tectonic Uplift and Climatic Change, pp. 289–312. New York: Plenum. DeSigoyer J, Chavagnac V, Blichert-Toft J, et al. (2000) Dating the Indian continental subduction and collisional thickening in the northwest Himalaya: Multichronology of the Tso Morari eclogites. Geology 28: 487–490. Dewey JF and Bird JM (1970) Bird, mountain belts and new global tectonics. Journal of Geophysical Research 75: 2625–2647. Dieterich JH (1979) Modeling of rock friction. Part 1: Experimental results and constitutive equations. Journal of Geophysical Research 84: 2161–2168. Ding L, Kapp P, and Wan XQ (2005) Paleocene-Eocene record of ophiolite obduction and initial India-Asia collision, south central Tibet. Tectonics 24: TC3001. Dunlap WJ, Hirth G, and Teyssier C (1997) Thermomechanical evolution of a ductile duplex. Tectonics 16: 983–1000. England PC and McKenzie DP (1983) A thin viscous sheet model for continental deformation. Geophysical Journal of the Royal Astronomical Society 73: 523–532. Evans MJ, Derry LA, Anderson SP, and France-Lanord C (2001) Hydrothermal source of radiogenic Sr to Himalayan rivers. Geology 29: 803–806. Feldl N and Bilham R (2006) Great Himalayan earthquakes and the Tibetan plateau. Nature 444: 165–170. Fort M, Freytet P, and Colchen M (1983) The quaternary sedimentary evolution of the intra-montane basin of Pokhara in relation to the Himalaya Midlands and their hinterland (west central Nepal). In: Sinha AK (ed.)
Contemporary Geoscientific Researches in Himalaya, pp. 91–96. Dehra-Dun, India: Wadia Institute of Himalayan Geology. France-Lanord C and Derry LA (1997) Organic carbon burial forcing of the carbon cycle from Himalayan erosion. Nature 390: 65–67. Gansser A (1964) Geology of the Himalaya, 289 pp. London: Interscience Publishers. Garzione CN, Dettman DL, Quade J, DeCelles PG, and Butler RF (2000) High times on the Tibetan plateau: Paleoelevation of the Thakkhola graben, Nepal. Geology 28: 339–342. Gautam P and Appel E (1994) Magnetic polarity stratigraphy of Siwalik Group sediments of Tinau khola section in west central Nepal, revisited. Geophysical Journal International 17: 223–234. Godard V, Cattin R, and Lave J (2004) Numerical modeling of mountain building: Interplay between erosion law and crustal rheology. Geophysical Research Letters 31: (doi:10.1029/ 2004GL021006). Godard V, Lave J, and Cattin R (2006) Numerical modelling of erosion processes in the Himalaya of Nepal: Effects of spatial variations of rock strength and precipitation. In: Buiter SJH and Schreurs G (eds.) Geological Society, Special Publications : Analogue and Numerical Modelling of Crustal-Scale Processes, pp. 341–358. London: Geological Society of London. Gordon RG, Demets C, and Argus DF (1990) Kinematic constraints on distributed lithospheric deformation in the equatorial Indian-Ocean from present motion between the Australian and Indian Plates. Tectonics 9: 409–422. Gratton J (1989) Crustal shortening, root spreading, isostasy, and the growth of orogenic belts: A dimensional analysis. Journal of Geophysical Research 94: 15627–15634. Grujic D, Casey M, Davidson C, et al. (1996) Ductile extrusion of the higher Himalayan crystalline in Bhutan: Evidence from quartz microfabrics. Tectonophysics 260: 21–43. Guillot S (1999) An overview of the metamorphic evolution in central Nepal. Journal of Asian Earth Sciences 17: 713–725. Guillot S, De Sigoyer J, Lardeaux JM, and Mascle G (1997) Eclogitic metasediments from the Tso Morari area (Ladakh, Himalaya): Evidence for continental subduction during the India-Asia convergence. Contributions to Mineralogy and Petrology 128: 197–212. Gutscher MA, Kukowski N, Malavieille J, and Lallemand S (1998) Material transfer in accretionary wedges from analysis of a systematic series of analog experiments. Journal of Structural Geology 20: 407–416. Harrison TM, Copeland P, Hall SA, et al. (1993) Isotopic preservation of himalayan/tibetan uplift, denudation, and climatic histories of two molasse deposits. The Journal of Geology 101: 157–175. Harrison TM, Copeland P, Kidd WSF, and Yin A (1992) Raising Tibet. Science 255: 1663–1670. Harrison TM, Grove M, Lovera OM, and Catlos EJ (1998) A model for the origin of Himalayan anatexis and inverted metamorphism. Journal of Geophysical Research 103: 27017–27032. Harrison TM, Ryerson FJ, Le Fort P, Yin A, Lovera OM, and Catlos EJ (1997) A late Miocene–Pliocene origin for the central himalayan inverted metamorphism. Earth and Planetary Sience Letters 146: E1–E8. Harrison TM, Yin A, and Ryerson FJ (1998) Orographic evolution of the Himalaya and Tibet. In: Crowley TJ and Burke K (eds.) Tectonic Boundary Conditions for Climate Reconstructions. New York: Oxford University Press. Hassani R, Jongmans D, and Chery J (1997) Study of plate deformation and stress in subduction processes using
Mountain Building: From Earthquakes to Geological Deformation two-dimensional numerical models. Journal of Geophysical Research 102: 17951–17965. Hauck ML, Nelson KD, Brown LD, Zhao W, and Ross AR (1998) Crustal structure of the himalayan orogen at 90deg east longitude from project INDEPTH deep reflection profiles. Tectonics 17: 481–500. Hendrix MS, Graham SA, Caroll AR, et al. (1992) Sedimentary record and climatic implications of recurrent deformation in the Tien Shan: Evidence from Mesozoic strata of the north Tarim, south Junggar, and Turpan basins, northwest China. Geological Society of America Bulletin 104: 53–79. Henry P, Le Pichon X, and Goffe B (1997) Kinematic, thermal and petrological model of the Himalaya; constraints related to metamorphism within the underthrust Indian crust and topographic elevation. Tectonophysics 273: 31–56. Hetenyi G, Cattin R, Vergne J, and Nabelek JL (2006) The effective elastic thickness of the India Plate from receiver function imaging, gravity anomalies and thermomechanical modelling. Geophysical Journal International 167: 1106–1118. Hilley GE and Strecker MR (2004) Steady state erosion of critical Coulomb wedges with applications to Taiwan and the Himalaya. Journal of Geophysical Research 109: B011411. Hirn A, Jobert G, Xu ZX, and Gao EY (1984) Main features of the upper lithosphere in the unit between the High Himalaya and the Yarlung Zangpo Jiang suture. Annales Geophysicae 2: 1984. Hirn A and Sapin M (1984) The Himalayan zone of crustal interaction. Annales Geophysicae 39: 205–249. Hodges KV, Parrish R, and Searle MP (1996) Tectonic evolution of the central Annapurna range, Nepalese Himalaya. Tectonics 15: 1264–1291. Hodges KV, Wobus C, Ruhl K, Schildgen T, and Whipple K (2004) Quaternary deformation, river steepening, and heavy precipitation at the front of the Higher Himalayan ranges. Earth and Planetary Science Letters 220: 379–389. Hough S, Bilham R, Ambraseys N, and Feldl N (2005) Revisiting the 1897 Shillong and 1905 Kangra earthquakes in northern India: Site response, Moho reflections and a Triggered earthquake. Current Science 88(10): 1632–1638. Houseman GA and England PE (1986) Finite strain calculations of continental deformation. Part 1: Method and general results for convergent zones. Journal of Geophysical Research 91: 3651–3663. Houseman GA, McKenzie DP, and Molnar P (1981) Convective instability of a thickened boundary-layer and its relevance for the thermal evolution of continental convergent belts. Journal of Geophysical Research 86: 6115–6132. Hubbard MS (1986) Thermobarometric constraints on the thermal history of the Main Central Thrust zone and Tibetan slab, eastern Nepal. Journal of Metamorphic Geology 7: 19–30. Hubbard MS (1996) Ductile shear as a cause of inverted metamorphism: Example from the Nepal Himalaya. The Journal of Geology 104: 493–499. Hubbard MS, Royden L, and Hodges K (1991) Constraints on unroofing rates in the High Himalaya, eastern Nepal. Tectonics 10: 287–298. Hubbert MK and Rubey WW (1959) Role of fluid pressure in mechanics of overthrust faulting. Part I: Mechanics of fluid-filled porous solids and its application to overthrust faulting. Geological Society of America Bulletin 79: 115–166. Hurtrez J-E, Lucazeau F, Lave’J, and Avouac J-P (1999) Investigation of the relationships between basin morphology, tectonic uplift, and denudation from the study of an active fold belt in the Siwalik Hills, central Nepal. Journal of Geophysical Research 104: 12779–712796.
435
Hyndman R, Currie CA, and Mazzotti SP (2005) Subduction zone backarcs, mobile belts, and orogenic heat. GSA Today 15: 4–10. Hyndman RD, Yamano M, and Oleskevich DA (1997) The seismogenic zone of subduction thrust faults. The Island Arc 6: 244–260. Huyghe P, Galy A, Mugnier J-L, and France-Lanord C (2001) Propagation of the thrust system and erosion in the Lesser Himalaya: Geochemical and sedimentological evidence. Geology 29: 1007–1010. Iwata S (1976) Late Pleistocene and Holocene moraines in the Sagarmatha region, Khumbu Himal. Journal of the Japanese Society of Snow and Ice 38: 115–119. Jackson JA, Austrheim H, McKenzie D, and Priestley K (2004) Metastability, mechanical strength, and the support of mountain belts. Geology 32: 625–628. Jackson M and Bilham R (1994a) 1991–1992 GPS measurements across the Nepal Himalaya. Geophysical Research Letters 21: 1169–1172. Jackson M and Bilham R (1994b) Constraints on Himalayan deformation inferred from vertical velocity fields in Nepal and Tibet. Journal of Geophysical Research 99: 13897–13919. Jaeger JC and Cook NGW (1979) Fundamentals of Rocks Mechanics, 3rd edn. London: Chapman and Hall. Jaeger JJ, Courtillot V, and Tapponnier P (1989) Paleontological view of the ages of the deccan traps, the cretaceous tertiary boundary, and the India-Asia collision. Geology 17: 316–319. Jaupart C and Provost A (1985) Heat focussing, granite genesis and inverted metamorphic gradients in continental collision zones. Earth and Planetary Science Letters 73: 385–397. Jin Y, McNutt MK, and Zhu Y (1996) Mapping the descent of Indian and Eurasian plates beneath the Tibetan Plateau from gravity anomalies. Journal of Geophysical Research 101: 11275–11290. Jouanne F, Mugnier JL, Pandey M, et al. (1999) Oblique convergence in Himalaya of western Nepal deduced from preliminary results of GPS measurements. Geophysical Research Letters 26: 1933–1936. Kanamori H and Brodsky EE (2004) The physics of earthquakes. Reports on Progress in Physics 67: 1429–1496. Kerrick DM and Caldeira K (1993) Paleoatmospheric consequences of Co2 released during early Cenozoic regional metamorphism in the Tethyan Orogen. Chemical Geology 108: 201–230. Kerrick DM and Caldeira K (1999) Was the Himalayan orogen a climatically significant coupled source and sink for amostpheric CO2 during the Cenozoic? Earth and Planetary Science Letters 173: 195–203. Kind R, Ni J, Zhao W, et al. (1987) Mid-crustal Low-velocity zone beneath the southern Lhasa block: Results from the INDEPTH-II earthquake recording program. Science 274: 1692–1694. Kirby SH and Kronenberg AK (1987) Rheology of the lithosphere: Selected topics. Reviews of Geophysics 25: 1219–1244. Kohlstedt DL, Evans B, and Mackwell SJ (1995) Strength of the lithosphere: Constraints imposed by laboratory experiments. Journal of Geophysical Research 100: 17587–17602. Kohn MJ and Parkinson CD (2002) Petrologic case for Eocene slab breakoff during the Indo-Asian collision. Geology 30: 591–594. Konstantinovskaia E and Malavieille J (2005) Erosion and exhumation in accretionary orogens: Experimental and geological approaches. Geochemistry Geophysics Geosystems 6: Q02006. Koons P (1994) Three-dimensional critical wedges: Tectonics and topography in oblique collisional orogens. Journal of Geophysical Research 99: 12301–12315. Koons PO (1987) Some thermal and mechanical consequences of rapid uplift – An example from the southern Alps, New-Zealand. Earth and Planetary Science Letters 86: 307–319.
436
Mountain Building: From Earthquakes to Geological Deformation
Koons PO (1989) The topographic evolution of collisional mountain belts – A numerical look at the southern alps, New-Zealand. American Journal of Science 289: 1041–1069. Kotarba M (1986) Origin and possibilities of utilization of thermal waters in the central Nepal Himalaya. In: Le Fort P, Colchen M, and Montenat C (eds.) E´volution des Domaines Oroge´niques d’Asie Me´ridionale (de la Turquie a` l’Indone´sie), pp. 183–189. Nancy: Sciences de la Terre. Kumar S, Wesnousky SG, Rockwell TK, Briggs R, Thakur VC, and Jayangondaperumal R (2006) Paleoseismic evidence of great surface-rupture earthquakes along the Indian Himalaya. Journal of Geophysical Research 111 (doi: 10.1029/2004JB003309). Lamb S (2006) Shear stresses on megathrusts: Implications for mountain building behind subduction zones. Journal of Geophysical Research 111: B07401 (doi: 07410.01029/ 02005JB0003916). Lamb S and Davis P (2003) Cenozoic climate change as a possible cause for the rise of the Andes. Nature 425: 792–797. Larson K, Bu¨rgmann R, Bilham R, and Freymueller JT (1999a) Kinematics of the India-Eurasia collision zone from GPS measurements. Journal of Geophysical Research 104: 1077–1093. Larson KM, Bu¨rgmann R, Bilham R, and Freymueller JT (1999b) Kinematics of the India-Eurasia collision zone from the GPS measurements. Journal of Geophysical Research 104: 1077–1093. Lave´ J (2005) Analytic solution of the mean elevation of a watershed dominated by fluvial incision and hillslope landslides. Geophysical Research Letters 32: L11403. Lave´ J (1997) Tectonique et Erosion: L’apport de la Dynamique Fluviale a` l’e´tude Sismotectonique de l’Himalaya du Ne´pal Central, The`se de doctorat thesis, Paris 7. Lave´ J and Avouac JP (2000) Active folding of fluvial terraces across the Siwaliks Hills, Himalaya of central Nepal. Journal of Geophysical Research 105: 5735–5770. Lave´ J and Avouac JP (2001) Fluvial incision and tectonic uplift across the Himalaya of central Nepal. Journal of Geophysical Research 106: 26561–26591. Lave´ J, Yule D, Sapkota S, et al. (2005) Evidence for a great medieval earthquake (approximate to 1100 AD) in the central Himalaya, Nepal. Science 307: 1302–1305. Law RD, Searle MP, and Simpson RL (2004) Strain, deformation temperatures and vorticity of flow at the top of the Greater Himalayan slab, Everest Massif, Tibet. Journal of the Geological Society 161: 305–320. Le Fort P (1975a) Himalaya: The collided range: Present knowledge of the continental arc. American Journal of Science 275A: 1–44. Le Fort P (1975b) Himalaya: The collided range: Present knowledge of the continental arc. American Journal of Science 275A: 1–44. Le Pichon X, Fournier M, and Jolivet L (1992) Kinematics, topography, shortening, and extrusion in the India-Eurasia collision. Tectonics 11: 1085–1098. Lemmonier C, Marquis G, Perrier F, et al. (1999) Electrical structure of the Himlaya of central Nepal: High conductivity around the mid-crustal ramp along the MHT. Geophysical Research Letters 26: 3261–3264. Lombard A (1953) Pre´sentation d’un profil ge´ologique du Mt. Everest a la plaine du Gange (Ne´pal oriental). Bulletin of the Society of Belge de Ge´ologie Paleontologie De Hydrogeologie, Brussels 42: 123–128. Lombardo B and Rolfo F (2000) Two contrasting eclogite types in the Himalaya: Implications for the Himalayan orogeny. Journal of Geodynamics 30: 37–60. Lyon-Caen H and Molnar P (1983) Constraints on the structure of the Himalaya from an analysis of gravity anomalies and a flexural model of the lithosphere. Journal of Geophysical Research 88: 8171–8191.
Lyon-Caen H and Molnar P (1985) Gravity anomalies, flexure of the Indian plate and the structure, support and evolution of the Himalaya and Ganga Basin. Tectonics 4: 513–538. Macfarlane AM (1993) Chronology of tectonic events in the crystalline core of the Himalaya, Langtang national park, central Nepal. Tectonics 12: 1004–1025. Macfarlane A, Hodges KV, and Lux D (1992) A structural analysis of the main central thrust zone, Langtang National Park, Central Nepal Himalaya. Geological Society of America Bulletin 104: 1389–1402. Maggi A, Jackson JA, McKenzie D, and Priestley K (2000) Earthquake focal depths, effective elastic thickness, and the strength of the continental lithosphere. Geology 28: 495–498. Malavieille J (1984) Modelisation experimentale des chevauchements imbriques: Application aux chaines de montagnes. Bulletin de la Societe geologique de France XXVI(1): 129–138. Malavieille J, Lallemand SE, Dominguez S, et al. (2002a) Arccontinent collision in Taiwan: New marine observations and tectonic evolution. Geological Society of America Special Paper 358: 189–213. Malavieille J, Marcoux J, and De Wever P (2002b) L’ocean perdu. In: Museum National D9Histoire Naturelle (France), Avouac J-P and De Wever P (eds.) Himalaya-Tibet, Le choc des continents, pp. 32–39. Paris: CNRS Editions et Museum national de’Histoire naturelle. Mallet FR (1874) On the geology, etc., of the Darjiling district and the Westen Duars. Memoirs Geolological Survey India 11. Marone C (1998) Laboratory-derived friction laws and their application to seismic faulting. Annual Review of Earth and Planetary Sciences 26: 643–696. Marquis G, Jones AG, and Hyndman RD (1995) Coincident conductive and reflective middle and lower crust in southern British Columbia. Geophysical Journal International 120: 111–131. Mascle GH and He´rail G (1982) Les Siwaliks: Le prisme d’accre´tion tectonique associe´ a` la subduction intracontinentale himalayenne. Ge´ologie Alpine, Grenoble 58: 95–103. Masek JG, Isacks BL, and Fielding E (1994a) Rift flank uplift in Tibet: Evidence for a viscous lower crust. Tectonics 13: 659–667. Masek JG, Isacks BL, Gubbels TL, and Fielding EJ (1994b) Erosion and tectonics at the margins of continental plateaus. Journal of Geophysical Research 99: 13941–13956. Mattauer M (1975) Sur le me´canisme de formation de la schistosite´ dans l’Himalaya. Earth and Planetary Science Letters 28: 144–154. Mattauer M (1986) Intracontinental subduction, crust–mantle de´collement and crust-stacking wedge in the Himalaya and other continental belts. In: Coward MP and Ries AC (eds.) Geological Society of London, Special Publications: Collision Tectonics, pp. 37–50. London: Geological Scociety of London. McCaffrey R and Nabelek J (1998) Role of oblique convergence in the active deformation of the Himalaya and southern Tibet plateau. Geology 26: 691–694. McKenzie D and Jackson J (2002) Conditions for flow in the continental crust. Tectonics 21: 1055. Medlicott HB (1864) On the geological structure and relations of the southern portion of the Himalayan range between river Ganges and Ravee. Memoir Geological Society of India 3: 1–206. Meigs AJ, Burbank DW, and Beck RA (1995) Middle-late Miocene (>10 Ma) formation of the main boundary thrust in the western Himalaya. Geology 23: 423–426. Me´tivier F and Gaudemer Y (1997) Mass transfer between eastern Tien Shan and adjacent basins (central Asia):
Mountain Building: From Earthquakes to Geological Deformation Constraints on regional tectonics and topography. Geophysical Journal International 128: 1–17. Metivier F, Gaudemer Y, Tapponnier P, and Klein M (1999) Mass accumulation rates in Asia during the Cenozoic. Geophysical Journal International 137: 280–318. Mitra S (2003) A unified kinematic model for the evolution of detachment folds. Journal of Structural Geology 25: 1659–1673. Mitra S, Priestley K, Bhattacharyya AK, and Gaur VK (2005) Crustal structure and earthquake focal depths beneath northeastern India and southern Tibet. Geophysical Journal International 160: 227–248. Molnar P (1984) Structure and tectonics of the Himalaya: Constraints and implications of geophysical data. Annual Review of Earth and Planetary Science 12: 489–518. Molnar P (1987) Inversion of profiles of uplift rates for the geometry of dip-slip faults at depth, with examples from the Alps and the Himilaya. Annales Geophysicae Series BTerrestrial and Planetary Physics 5: 663–670. Molnar P (1990) A review of the seismicity and the rates of active underthrusting and deformation at the Himalaya. Journal of Himalayan Geology 1: 131–154. Molnar P, Brown ET, Burchfiel BC, et al. (1994) Quaternary climate-change and the formation of river terraces across growing anticlines on the north flank of the Tien-Shan, China. Journal of Geology 102: 583–602. Molnar P and England P (1990a) Late Cenozoic uplift of mountain ranges and global climate change: Chicken or egg?. Nature 346: 29–34. Molnar P and England P (1990b) Temperatures, heat flux, and frictional stress near major thrust fault. Journal of Geophysical Research 95: 4833–4856. Molnar P, England P, and Martinod J (1993) Mantle dynamics, uplift of the Tibetan Plateau, and the Indian Monsoon. Reviews of Geophysics 31: 357–396. Molnar P and Lyon-Caen H (1988) Some simple physical aspects of the support, structure, and evolution of mountain belts. In: Processes in continental lithospheric deformation. Geological Society of America Special, Paper 218: 179–207. Molnar P and Tapponnier P (1975) Cenozoic tectonics of Asia: Effects of a continental collision. Science 189: 419–426. Montgomery DR (1994) Valley incision and uplift of mountain peaks. Journal of Geophysical Research 99: 13913–13921. Mugnier J-L, Leturmy P, Mascle G, et al. (1999) The Siwaliks of western Nepal. Part 1: Geometry and kinematics. Journal of Asian Earth Sciences 17: 629–642. Mugnier JL, Mascle G, and Faucher T (1992) La structure des Siwaliks de l’Ouest Ne´pal: Un prisme d’accre´tion intracontinental. Bulletin de la Socie´te´ Ge´ologique de France 163: 585–595. Nakata T (1989) Active faults of the Himalaya of India and Nepal. In: Malinconico LL, Jr. and Lillie R (eds.) Tectonics of the Western Himalaya, pp. 243–264. Boulder, CO: Geological Society of America. Nakata T, Yagi H, Okumura K, Upreti BN, et al. (1998) First successful paleoseismic trench study on active faults in the Himalaya. EOS Transactions of American Geophysical Union 79: F615. Nelson KD, Zhao W, Brown LD, et al. (1996) Partially molten middle crust beneath southern Tibet: Synthesis of project INDEPTH results. Science 274: 1684–1688. Okada Y (1992) Internal deformation due to shear and tensile faults in a half-space. Bulletin of the Seismological Society of America 82: 1018–1040. Oldham RD (1883a) The geology of Jaunsar and the lower Himalaya. Records of the Geological Survey of India 16: 193–198.
437
Oldham T (1883b) Catalog of Indian Earthquakes. Memoir of the Geological Survey of India 19: 163–215. Oleskevich DA, Hyndman RD, and Wang K (1999) The updip and downdip limit to great subduction earthquakes: Thermal and structural models of Cascadia, South Alaska, SW Japan and Chile. Journal of Geophysical Research 104: 14965–14991. Pandey MR and Molnar P (1988) The distribution of intensity of the Bihar-Nepal earthquake of 15 january 1934 and bounds on the extent of the rupture zone. Journal of Nepal Geological Society 5: 23–45. Pandey MR, Tandukar RP, Avouac J-P, Lave´ J, and Massot JP (1995) Interseismic strain accumulation on the Himalaya crustal ramp (Nepal). Geophysical Research Letters 22: 741–754. Pandey MR, Tandukar RP, Avouac J-P, Vergne J, and He´ritier T (1999) Seismotectonics of Nepal himalaya from a local seismic network. Journal of Asian Earth Sciences 17: 703–712. Pandey P and Rawat RS (1999) Some new observations on the Amritpur granite series, Kumaun lesser Himalaya, India. Current Sciences 72: 296–299. Pant MR (2002) A step toward a historical seismicity of Nepal. Adarsa 2: 29–60. Pathier E, Fielding EJ, Wright TJ, Walker R, and Parsons BE (2006) Displacement field and slip distribution of the 2005 Kashmir earthquake from SAR imagery. Geophysical Research Letters 33: L20310. Patriat P and Achache J (1984) India-Eurasia collision chronology has implications for crustal shortening and driving mechanism of plates. Nature 311: 615–621. Patriat P and Segoufin J (1988) Reconstruction of the central Indian-Ocean. Tectonophysics 155: 211–234. Peˆcher A (1989) The metamorphism in central Himalaya. Journal of Metamorphic Geology 7: 31–41. Peltzer G and Saucier F (1996) Present-day kinematics of Asia derived from geologic fault rates. Journal of Geophysical Research 101: 27943–27956. Peltzer G and Tapponnier P (1988) Formation and evolution of strike-slip faults rifts, and basins during the India–Asia collision. An experimental approach. Journal of Geophysical Research 93: 15085–15117. Perfettini H and Avouac JP (2004) Stress transfer and strain rate variations during the seismic cycle. Journal of Geophysical Research-Solid Earth 109: B06402. Poisson B and Avouac JP (2004) Holocene hydrological changes inferred from alluvial stream entrenchment in North Tian Shan (Northwestern China). Journal of Geology 112: 231–249. Pollitz FF, Burgmann R, and Segall P (1998) Joint estimation of afterslip rate and postseismic relaxation following the 1989 Loma Prieta earthquake. Journal of Geophysical ResearchSolid Earth 103: 26975–26992. Powell CMA and Conaghan PJ (1973) Plate tectonics and the Himalaya. Earth and Planetary Science Letters 20: 1–12. Pratt JH (1855) On the attraction of the Himalaya Mountains, and of the elevated regions beyond them, upon the plumb line in India. Philosophical Transactions of the Royal Society of London 145: 53–100. Pysklywec RN (2006) Surface erosion control on the evolution of the deep lithosphere. Geology 34: 225–228 (doi: 210.1130/ G21963.21961). Ramstein G, Fluteau F, Besse J, and Joussaume S (1997) Effect of orogeny, plate motion and land sea distribution on Eurasian climate change over the past 30 million years. Nature 386: 788–795. Rana BSJB (1935) Nepalako mahabhukampa (1990 sala) Nepal’s great earthquake. Ratschbacher L, Frisch W, Liu G, and Chen C (1994) Distributed deformation in southern and western Tibet during and after
438
Mountain Building: From Earthquakes to Geological Deformation
the India-Asia collision. Journal of Geophysical Research 99: 19917–19945. Raymo ME and Ruddiman WF (1992) Tectonic forcing of Late Cenozoic climate. Nature 359: 117–122. Replumaz A, Karason H, van der Hilst RD, Besse J, and Tapponnier P (2004) 4-D evolution of SE Asia’s mantle from geological reconstructions and seismic tomography. Earth and Planetary Science Letters 221: 103–115. Rice JR and Tse ST (1986) Dynamic motion of a single degree of freedom system following a rate and state dependent friction law. Journal of Geophysical Research 91: 521–530. Robinson DM, DeCelles PG, Garzione CN, Pearson ON, Harrison TM, and Catlos EJ (2003) Kinematic model for the Main Central thrust in Nepal. Geology 31: 359–362. Robinson DM, DeCelles PG, Patchett PJ, and Garzione CN (2001) The kinematic evolution of the Nepalese Himalaya interpreted from Nd isotope. Earth and Planetary Science Letters 192: 507–521. Rowley DB (1996) Age of initiation of collision between India and Asia: A review of stratigraphic data. Earth and Planetary Science Letters 145: 1–13. Rowley DB and Currie BS (2006) Palaeo-altimetry of the late Eocene to Miocene Lunpola basin, central Tibet. Nature 439: 677–681. Royden LH (1993) The steady state thermal structure of eroding orogenic belts and accretionary prisms. Journal of Geophysical Research 98: 4487–4507. Royer JY and Chang T (1991) Evidence for relative motions between the Indian and Australian plates during the last 20 My from plate tectonic reconstructions – implications for the deformation of the Indo-Australian plate. Journal of Geophysical Research – Solid Earth and Planets 96: 11779–11802. Royer JY and Patriat P (2002) L’Inde part a la deriv". In: Museum National D9Histoire Naturelle (France), Allegre CJ, Avouac J-P, and De Wever P (eds.) Himalaya-Tibet, Le choc des continents, pp. 25–31. Paris: CNRS Editions et Museum national de’Histoire naturelle. Ruddiman WF and Kutzbach JE (1991) Plateau uplift and climatic-change. Scientific American 264: 66–75. Sakai H (1985) Geology of the Kali Gandaki supergroup of the lesser himalaya in Nepal. Memoirs of the Faculty of Science, Kyushu University, Series D, Geology 25: 337–397. Sandiford M, Coblentz DD, and Richardson RM (1995) Ridge torques and continental collision in the Indian–Australian plate. Geology 23: 653–656. Sastri VV, Bhandari LL, Raju ATR, and Datta AK (1971) Tectonic framework and subsurface stratigraphy of the Ganga basin. Journal of Geological Society of India 12: 222–233. Saul J, Kumar MR, and Sarkar D (2000) Lithospheric and upper mantle structure of the indian shield, from teleseismic receiver functions. Geophysical Research Letters 27: 2357–2360. Scha¨rer U and Allegre CJ (1983) The Palung granite (Himalaya); high resolution U–Pb systematics in zircon and monazite. Earth and Planetary Science Letters 63: 423–432. Schelling D (1992) The tectonostratigraphy and structure of the eastern Nepal Himalaya. Tectonics 11: 925–943. Schelling D and Arita K (1991) Thrust tectonics, crustal shortening, and the structure of the far-eastern Nepal Himalaya. Tectonics 10: 851–862. Schulte-Pelkum V, Monsalve G, Sheehan A, et al. (2005) Imaging the Indian subcontinent beneath the Himalaya. Nature 435: 1222–1225. Searle MP, Parrish RR, Hodges KV, Hurford MW, Ayres MW, and Whitehouse MJ (1997) Shisha Pangma leucogranite, South tibetan Himalaya: Field relations, geochemistry, age, origin, and emplacement. The Journal of Geology 105: 295–317.
Searle MP, Pickering KT, and Cooper DJW (1990) Restoration and evolution of the intermontane Indus Molasse Basin, Ladakh Himalaya, India. Tectonophysics 174: 301–314. Searle MP, Windley BF, Coward MP, et al. (1987) The closing of Tethys and the tectonics of the Himalaya. Geological Society of America Bulletin 98: 678–701. Seeber L and Gornitz V (1983) River profiles along the Himalayan arc as indicators of active tectonics. Tectonophysics 92: 335–367. Segall P and Rice JR (1995) Dilatancy, compaction, and slip instability of a fluid-infiltrated fault. Journal of Geophysical Research-Solid Earth 100: 22155–22171. Simoes M, Avouac JP, Chen YG, et al. (2007) Kinematic anaylsis of the Pakuashan fault-tip fold, West Central Taiwan: Shortening rates and age of folding inception. Journal of Geophysical Research 112: B03S14. Sleep NH and Blanpied ML (1992) Creep, compaction and the weak rheology of major faults. Nature 359: 687–692. Smith SW and Wyss M (1968) Displacement on the San Andreas fault subsequent to the 1966 Parkfield earthquake. Bulletin of the Seismological Society of America 58: 1966–1973. Spicer RA, Harris NBW, Widdowson M, et al. (2003) Constant elevation of southern Tibet over the past 15 million years. Nature 421: 622–624. Srivastava P and Mitra G (1994) Thrust geometries and deep structure of the outer and lesser Himalaya, Kumaon and Garhwal (India): Implications for evolution of the Himalayan fold-and-thrust belt. Tectonics 13: 89–109. Sto¨cklin J, Bhattarai KD, Chettri VS, and Bhandari AN (1980) Geologic map of Kathmandu area and Central Mahabharat Range. Department of Mines and Geology, Kathmandu, Nepal. Sun W(1989) Bouguer gravity anomaly map of the People’s Republic of China. Suppe J (1985) Principles of Structural Geology, 537 pp. Englewood cliffs, NJ: Prentice-Hall. Suppe J and Medwedeff DA (1990) Geometry and kinematics of fault-propagation folding. Eclogae Geologicae Helvetiae 83: 409–454. Tapponnier P, Xu ZQ, Roger F, et al. (2001) Oblique stepwise rise and growth of the Tibet plateau. Science 294: 1671–1677. Tavernier G, Fagard H, Feissel-Vernier M, et al. (2005) The International DORIS Service IDS. Advances in Space Research 36(3): 333–341 (doi: 10.1016/j.asr.2005.1003.1102). Thompson SC, Weldon RJ, Rubin CM, Abdrakhmatov K, Molnar P, and Berger GW (2002) Late quaternary slip rates across the central Tien Shan, Kyrgyzstan, central Asia. Journal of Geophysical Research-Solid Earth 107: 2203. Tokuoka T, Takayasu K, Yoshida M, and Hisatomi K (1986) The Churia (Siwalik) group of the Arung khola area, west central Nepal. Memoirs of the Faculty of Science, Shimane University 20: 135–210. Toussaint G, Burov E, and Avouac JP (2004a) Tectonic evolution of a continental collision zone: A thermomechanical numerical model. Tectonics 23. Toussaint G, Burov EB, and Jolivet L (2004b) Continental plate collision: Unstable vs. stable slab dynamics. Geology 32: 33–36. Tsenn MC and Carter NL (1987) Upper limits of power law creep of rocks. Tectonophysics 136: 1–26. Upreti BN (1999) An overview of the stratigraphy and tectonics of the Nepal Himalaya. Journal of Asian Earth Sciences 17: 577–606. Upreti BN, Nakata T, Kumahara Y, et al. (2000) The latest active faulting in southeast Nepal: In Active Fault Research for the New Millenium. In: Okumura KTK and Goto H (eds.) Proceedings of the Hokudan International Symposium and School on Active Faulting, pp. 533–536. Hiroshima , Japan: Letter Press Ltd.
Mountain Building: From Earthquakes to Geological Deformation Valdiya KS (1964) The unfossiliferous formations of the Lesser Himalaya and their correlation, paper presented at 22nd International Geological Congress, Delhi. Van de Meulebrouck J (1983) Reconnaissance geophysique des structures crustales de deux segments de chaiˆne de collision: Le Haut Allier (Massif Central franc¸ais) et le Sud Tibet (Himalaya). Vergne J, Cattin R, and Avouac J-P (2001) On the use of dislocations to model interseismic strain and stress build-up at intracontinental thrust faults. Geophysical Journal International 147: 155–162. Walcott RI (1998) Modes of oblique compression: Late Cenozoic tectonics of the South Island of New Zealand. Reviews of Geophysics 36: 1–26. Wang Q, Zhang P-Z, Freymueller JT, et al. (2001) Present-day crustal deformation in China constrained by global positioning system measurements. Science 294: 574–577. Whipple KX and Meade BJ (2004) Controls on the strength of coupling among climate, erosion, and deformation in twosided, frictional orogenic wedges at steady state. Journal of Geophysical Research-Earth Surface 109: F01011. Willett SD (1999) Orogeny and topography: The effects of erosion on the structure of mountain belts. Journal of Geophysical Research 104: 28957–28981. Willett S, Beaumont C, and Fullsack P (1993) Mechanical model for the tectonics of doubly vergent compressional orogens. Geology 21: 371–374. Williams CA, Conners C, Dahlen FA, Price EJ, and Suppe J (1994a) Effect of the brittle–ductile transition on the topography of compressive mountain belts on Earth and Venus. Journal of Geophysical Research-Solid Earth 99: 19947–19974. Williams CA, Connors C, Dalhen FA, Price EJ, and Suppe J (1994b) The effect of brittle–ductile transition on the
439
topography of compressive mountain belts on Earth and Venus. Journal of Geophysical Research 99: 19947–919974. Willis B (1891) The Mechanics of Appalachian Structure, 281 pp. US Geological Survey 13th Annual Report 1891-1892, part 2. Wobus C, Heimsath A, Whipple K, and Hodges K (2005) Active out-of-sequence thrust faulting in the central Nepalese Himalaya. Nature 434: 1008–1011. Wobus CW, Hodges KV, and Whipple KX (2003) Has focused denudation sustained active thrusting at the Himalayan topographic front? Geology 31: 861–864. Yeats RS, Nakata T, Farah A, Mizra MA, Pandey MR, and Stein RS (1992) The Himalayan frontal fault system. In: DeJong KA and Farah A (eds.) Special Issue: Geodynamics of Pakistan, Geological Survey of Pakistan: Seismicity of the Hazara Arc in Northern Pakistan: Decollement vs. Basement Faulting. Annales Tectonicae pp. 85–98. Supplement to Volume VI. Yin A, Harrison TM, Murphy MA, et al. (1999) Tertiary deformation history of southeastern and southwestern Tibet during the Indo-Asian collision. GSA Bulletin 111: 1644–1664. Zhang PZ, Shen Z, Wang M, Gan WJ, Burgmann R, and Molnar P (2004) Continuous deformation of the Tibetan Plateau from global positioning system data. Geology 32: 809–812. Zhao WL and Morgan J (1985) Uplift of Tibetan plateau. Tectonics 4: 359–369. Zhao W, Nelson KD, and Project INDEPTH Team (1993) Deep seismic-reflection evidence continental underthrusting beneath southern Tibet. Nature 366: 557–559. Zienkiewicz OC and Taylor RL (1989) The Finite Element Method, 4th edn. New York: McGraw-Hill.
6.10
Fault Mechanics
C. H. Scholz, Columbia University, Palisades, NY, USA ª 2007 Elsevier B.V. All rights reserved.
6.10.1 6.10.2 6.10.2.1 6.10.2.2 6.10.3 6.10.3.1 6.10.3.2 6.10.3.3 6.10.3.4 6.10.3.5 6.10.3.6 6.10.3.7 6.10.3.8 6.10.3.9 6.10.3.10 6.10.4 6.10.4.1 6.10.4.1.1 6.10.4.1.2 6.10.4.1.3 6.10.4.2 6.10.4.2.1 6.10.4.2.2 6.10.5 6.10.5.1 6.10.5.1.1 6.10.5.1.2 6.10.5.1.3 6.10.5.2 6.10.6 6.10.6.1 6.10.6.2 6.10.7 6.10.7.1 6.10.7.2 6.10.7.3 6.10.7.4 6.10.7.4.1 6.10.7.4.2 6.10.7.5 6.10.8 6.10.8.1 6.10.8.2 References
Introduction Elementary Fault Theory Anderson’s Theory of Faulting Overthrust Faults and the Hubert–Rubey Theory Fracture Mechanics of Faults Linear Elastic Fracture Mechanics The Dugdale–Barenblatt Model Critical Fault Tip Taper (CFTT) Model Experimental Studies of Fault Propagation Fault Scaling Laws and the Mechanics of Fault Growth Displacement–Length Scaling Fault Tip Taper Scaling Process Zones and Their Scaling Cataclasite Zone Scaling Interpretation of the Scaling Laws in Terms of Crack Models Fault Interactions Mode III Interactions: Normal Faults Pinning Coalescence Nucleation inhibition Mode II Interactions Interactions at strike-slip jogs Strike-slip splay faults Fault Populations Observations of Fault Populations Power law distributions Exponential distributions Periodic distributions The Formation of Fault Populations Strain and Faulting Calculation of Brittle Strain from Fault Data Fault Rotation and Lockup Fault Rocks and Structures Shallow Schizosphere Deep Schizosphere Brittle-Plastic Transition Region Plastosphere Shear Zones Mylonite fabric elements Shear localization and strain softening in mylonite zones Synoptic Model for Faults and Shear Zones The Strength of Faults Direct Evidence for Fault Strength The Weak San Andreas Fault Fallacy
442 442 442 443 445 445 446 446 447 447 447 450 450 453 455 456 456 456 458 459 460 460 463 463 465 465 467 467 467 469 469 471 472 472 473 474 475 475 475 476 477 477 478 479
441
442
Fault Mechanics
6.10.1 Introduction
but at the short (earthquake) timescale, they interact. Thus for large faults, which include such shear zones, both the brittle/frictional and ductile/plastic regions must be considered as interacting parts of the same fault system, and both should be discussed in any treatment of faults. Much of the material in this chapter is expanded and updated from Scholz (2002).
Continental crust may be considered to be composed of two mechanical units: an upper region, the schizoshere, which responds to large deformations in a brittle manner, and below that, the plastosphere, which deforms by plastic flow. There is a gradual transition between the two. The schizosphere constitutes the upper 10 to several tens of kilometers of the crust, depending on the heat flow and hence age of the crust. Below 1–2 km, stresses are compressive and macroscopic deformation is by shear fracture. Such shear fractures, which have friction between their walls, are called faults. The oceanic lithosphere may be similarly divided. Brittle deformation of the schizosphere takes place on two timescales. The long timescale corresponds to the formation and propagation of faults and the short timescale to the nucleation and propagation of earthquakes. Earthquakes are dynamically propagating shear cracks that propagate on faults due to a stick-slip frictional instability. The shear displacement on faults is usually the result of the cumulative slip of many earthquakes. Because of these different timescales, the two phenomena, faulting and earthquakes, are usually studied separately and by different disciplines. This chapter is concerned primarily with faults, which are treated as quasistatic features. When a fault reaches the dimensions of the schizosphere, it will generate, by virtue of the stress concentration at its base, a ductile shear zone within the plastosphere, which will further develop and deepen as the fault continues to grow. This ductile shear zone thus becomes an integral part of the structure. For compatibility of their boundary conditions, the two must act as a unit at long timescales,
6.10.2 Elementary Fault Theory The first recognition of faults as tectonic agents was made by Nicolaus Steno in his Prodromus of 1669, in his discussion of the origin of mountains (Adams, 1938). Faults were known by early miners, in fact, fault is an English mining term (Lyell, 1832). Lyell was the first to associate earthquakes with faults, in a study of the earthquakes of 1783 in Calabria. However, the modern theory of faulting did not arise until the early twentieth century. 6.10.2.1
Anderson’s Theory of Faulting
Early in his career, E.M. Anderson proposed that faults are brittle fractures that occur according to the Coulomb criterion (Anderson, 1905). The Coulomb criterion relates the shear stress to normal stress n on a plane within the material ¼ 0 þ n
where the parameters 0 and are called the cohesion and coefficient of internal friction, respectively. The angle ¼ tan1 is called the angle of internal friction. Figure 1 shows this criterion with a Mohr circle representation. From the Mohr circle it can be seen that failure will occur on two conjugate planes σ1
τ
σ3
τo φ
σ3
2θ
σ1 + σ3 2
½1
σ1
θθ
σ3
σn σ1
Figure 1 Mohr diagram illustrating the Coulomb fracture criterion and the orientation of shear fractures.
Fault Mechanics (a) σNS
443
(b) 30° 30°
30
II 30 II
I VI
I
60
60
N (σv) III
60
III
60 IV
V
VI
V IV
σEW (σv)
Figure 2 (a) Slice through the Coulomb fracture envelop at a fixed level of vertical stress v. (b) Orientation of faults for the various sides of the fracture envelop. From Scholz CH (2002) The Mechanics of Earthquakes and Faulting, 2nd edn. Cambridge: Cambridge University Press.
oriented at acute angles on either side of the maximum principal compressive stress 1, ¼ =4 – =2
½2
with opposite senses of shear. The intersection of the two planes is in the direction of the intermediate principal stress, 2. Recognizing the condition that near the free surface one of the principal stresses must be vertical, Anderson showed that the three classes of faults, normal, thrust, and strike-slip result from the three types of inequality that arise between the three principal stresses. Figure 2(a) shows a section through a Coulomb fracture envelope at a fixed value of the vertical principal stress v. The two horizontal principal stresses are assumed to be north–south and east– west for convenience, denoted NS and EW. Each side of this six-sided envelope defines a different regime of faulting, as shown in Figure 2(b). Side 1, for example, is the failure locus for NS > V > EW, hence for strike-slip faults striking at an acute angle from N, as shown in Figure 2(b). In the figure it is assumed that ¼ 0.6, hence ¼ 30 . At the end of his career, Anderson (1951) produced a monograph in which, to support his hypothesis, he described many examples from Britain of conjugate fault systems in which the angular relationships conform to the expectations of the Coulomb criterion. The angular relations of conjugate faults to the stress field apply only to the stress state at the time of the formation of the faults. In many cases of the old faults
discussed by Anderson, it is not possible to prove the simultaneity of their origin. Furthermore, as will be discussed later, faults rotate away from the 1 direction as they accrue slip, so that their included angle at a later time may be larger than given in eqn [2]. Figure 3 shows an example of currently active faults in Japan. The figure depicts the Izu Peninsula, which lies on the Philippine Sea Plate and is currently colliding with the plate (unknown) on which NE Japan sits (Nakamura et al., 1984; Somerville, 1978). Many sets of conjugate strike-slip faults occur on Izu, of which the ones that have recently ruptured in large earthquakes with known slip directions are shown. The alignments of parasitic cone emanating from the Hakone and Oshima volcanoes bisect the angle between the conjugate faults. These cones overlie dikes which are produced by hydraulic fracturing and hence lie parallel to 1 (Anderson, 1936; Nakamura, 1969), thus providing independent evidence for the stress direction.
6.10.2.2 Overthrust Faults and the Hubert–Rubey Theory An important category of faults that does not conform with Andersonian mechanics are overthrusts. These are nearly horizontal thrust faults, common in foreland fold and thrust belts, that are observed to have transported their upper plates large distances over their lower plates. These faults do not conform in their orientation with Anderson’s theory, and they also
444
Fault Mechanics
Fuji
hu
ns
Ho
Hakone Sagami Bay I930 M = 7.0 ro iT m ga Sa
IZU gh Su
r ug aT rou
h ug
I980 M = 6.7
35° N
Oshima
I978 M = 7.0 I974 M = 6.9
0
20 km
139° E Figure 3 Map of the Izu Peninsula, Japan, showing the faults ruptured in recent strike-slip earthquakes and the orientation of parasitic volcanoes (open circles) associated with major volcanoes (filled triangles). From Scholz CH (2002) The Mechanics of Earthquakes and Faulting, 2nd edn. Cambridge: Cambridge University Press.
l
σh
Z
Figure 4 Diagram illustrating the Hubert–Rubey theory of overthrust faulting. From Scholz CH (2002) The Mechanics of Earthquakes and Faulting, 2nd edn. Cambridge: Cambridge University Press.
present a fundamental mechanical difficulty. Figure 4 shows a schematic diagram of a thrust sheet on height z and length l is pushed along a friction surface by a horizontal stress h. The force balance for this is F ¼ h ¼ ðz – pÞl
½3
where is friction, p is pore pressure, and a simple effective stress law is assumed. Let v ¼ gz, the lithostatic overburden, and let pore pressure p ¼ gz, where is a constant. Then the maximum length of a block that can be pushed without h exceeding the compressive strength of the rock C is l ¼ C =ð1 – Þg
½4
If we assume that p is hydrostatic, then ¼ 0.4. Assuming reasonable values of C ¼ 200 MPa, ¼ 0.85 and ¼ 2.5, we obtain from eqn [4] that l ¼ 16 km. Many overthrust sheets, however, are observed to be much longer than that, in the range of 50–100 km. In some cases overthrusts, called decollements, slide on a layer of a weak ductile material such as salt or gypsum. The Jura overthrust in Switzerland, for example, moved on a gypsum layer, which was plastically deformed to form a mylonite. In such cases, the right-hand side of [3] is replaced by maxl, where max is the yield strength of the weak material. Then [4] becomes l ¼ Cz=max
½5
and because C >> max, l >> z. Hubbert and Rubey (1959) suggested that under certain circumstance p could exceed hydrostatic, and, in the case of overthrusts, could reach the lithostatic pressure without hydrofracturing intervening. Thus as in eqn [4] approaches 1, the constraint on the length of the overthrust sheet vanishes. The mechanism by which pore pressure could be so elevated was controversial at the time. Voight
Fault Mechanics
(1976) contains a collection of papers debating this issue. It is now well known, from oilfield data, that such overpressures are found in sedimentary basins that contain abundant shale that can seal in the pressures produced by compaction of the sediments. Typical geological terranes in which this mechanism, and overthrusts, occur are foreland fold and thrust belts and the accretionary prisms of subduction zones.
(a)
445
Δu
x c/2 (b)
Δu
6.10.3 Fracture Mechanics of Faults Anderson’s theory is concerned only with the orientation of faults with respect to the principal stress directions. He did not consider how faults grow. Faults occur in lengths ranging from a few centimeters to hundreds of kilometers, so they must grow in length as they accrue slip. To understand this behavior, we need to treat faults as shear cracks, which is treated here in a simplified form.
x c/2-s (c)
c/2
Δu
FTT x c/2
6.10.3.1
Linear Elastic Fracture Mechanics
This form of fracture mechanics assumes the fault to be a crack in a linear elastic medium. Because the stress and displacement distributions around such a crack have been solved analytically, this theory is the most attractive mathematically and is the one most often used. For a more complete treatment see, for example, Lawn and Wilshaw (1975), and, for geological applications, Pollard and Segall (1987). In this model the crack tip is assumed to be a mathematically sharp slit. The displacement D between the crack walls is given by D ¼
2ð1 – vÞ 2 ðL – x 2 Þ1=2
½6
where L is the length of the crack, x is the position along its length, v and are elastic constants, and the stress-drop ¼ (y f), the yield stress of the intact rock less the residual friction stress. The displacement distribution is thus elliptical, as shown in Figure 5(a). The maximum displacement is thus Dmax
2ð1 – vÞ ¼ L
½7
The stress at the tip of the crack is ¼ K ð2r Þ – 1=2 f ðÞ
½8
Figure 5 (a–c) Three types of crack models discussed in the text: (a) elastic model; (b) Dugdale model; (c) small-scale yielding model. From Scholz CH (2002) The Mechanics of Earthquakes and Faulting, 2nd edn. Cambridge: Cambridge University Press.
where r is the distance from the crack tip, K is called the stress intensity factor, and f () is a function that describes the dependence with the angle from the plane of the crack. The functions f () depend on crack mode. They can be found tabulated in standard sources such as Lawn and Wilshaw (1975). The stress intensity factor is given by K ¼ ðLÞ1=2
½9
The Griffith fracture criterion is included by assuming that the crack propagates when K reaches a critical value Kc defined by the point at which the critical energy release rate Gc equals the surface energy Gc ¼ Kc2 =2ð1 þ uÞ ¼ 2
½10
where is the specific surface energy of the material. This is the Griffith equilibrium equation. Notice that [9] indicates that K increases with L. Because the right hand side of [10] is constant, equilibrium requires that decrease as the crack grows: that
446
Fault Mechanics
is, the crack weakens as it lengthens. This feature results in the Griffith instability: if the driving stress is maintained the crack will run without limit after reaching its critical length. Combining [7] with [9], we obtain Dmax ¼
2ð1 – uÞ 1=2 L Kc
breakdown zone (y f) remains constant during fault growth. Thus, unlike the elastic crack model, the fault remains of constant strength during growth. The displacement at the inner end of the breakdown zone, D0, scales with L. The breakdown energy is G ¼ ðy – f ÞD0
½11
Because Kc is constant at equilibrium, this implies that Dmax scales as L1/2.
Using the expression for D0 from Cowie and Scholz (1992a), [13] becomes G ¼
6.10.3.2
The Dugdale–Barenblatt Model
The elastic crack model, while convenient for calculating stress and displacement fields around faults, has several problems when it comes to applying it to crack propagation. The first is the square root singularity in the crack tip stress [8]. No real materials, of course, can support such a stress singularity. Materials such as steel and rock are likely to inelastically yield in some volume around the crack tip by plastic flow or microcracking and thus relax the singularity. Dugdale (1960) and Barenblatt (1962) were the first to address this problem. They modified the elastic crack model by assuming a breakdown, or yielding region s in the vicinity of the tip. They did this by imposing a resistive cohesion stress up to the yield stress y in the breakdown region. The resulting displacement distribution is shown in Figure 5(b). Cowie and Scholz (1992a) applied this model to faulting. There are two ways of scaling s. If one assumes that s is constant, then the breakdown regions become comparatively small as the fault grows and the behavior approaches that of the elastic crack model. This is not realistic, however, because as the L increases the size of the stress concentration at the tip increases, so we should expect that s will increase likewise. For an elastic, purely plastic material, for example, the radius of the zone of yielding at the crack tip is (Atkinson, 1987), ry ¼ ð1=2Þð=y Þ2 ¼
L a 2 8 y
½12
where a is the applied stress and y the yield stress. Because the ratio of these two is scale independent, we see that ry scales with L. Therefore, we should scale s with L in the Dugdale–Barenblatt model. For this case, Cowie and Scholz (1992a) showed that the stress in the
½13
2ð1 – vÞLðy – f Þ2 cos 2 lnðsec 2 Þ
½14
where cos 2 ¼ ðhıiL – sÞ=L. Thus G scales with L, rather than being constant, as in the elastic crack model. In this case, the right-hand side of the equilibrium equation scales with L, so that there is no Griffith instability: the fault grows quasistatically. Kc then scales with L1/2 (see eqn [10]) and so the equivalent of [11] will show that D scales linearly with L. Specifically, ð1 – vÞðy – f Þ Dmax fnctð2 Þ ¼ 2 L
½15
we see that the faults are self-similar and that the Dmax/L ratio is a strain that corresponds to the stress drop (y f).
6.10.3.3 Model
Critical Fault Tip Taper (CFTT)
The Dugdale–Barrenblatt model is unrealistic in that, for mathematical facility, it restricts the inelastic deformation to the plane of the crack. A more realistic elastic-plastic model that allows inelastic deformation within a volume surrounding the crack tip, is called the critical tip opening angle (CTOA) model (Kanninen and Popelar, 1985). Because we are applying this to faults rather than opening mode cracks, we call it the CFTT model. This is a numerical model so that we can discuss only its results. The displacement profile for the CFTT model is shown in Figure 5(c). The displacements near the crack center are similar to the elastic crack, but near the tips they taper linearly, as the name suggests. This CFTT remains constant as the fault propagates in a material of uniform y and the value of the taper increases with y. Like the Dugdale–Barenblatt model, both G and Dmax scale linearly with L. These predictions agree very well with experimental results for cracking in ductile materials like steel and
Fault Mechanics
447
plateau following (f) in Figure 6 reflects the residual friction as the fully formed fault slides. Subsequent microscopic study of one of their experiments by Moore and Lockner (1995) showed that a brittle process zone about 40 mm wide had formed around the fault (Figure 7). The crack density increases exponentially as the fault is approached (Figure 7(a)). The cracks are primarily intergranular tensile cracks. Their orientation peaks at about 30 from the fault on its compressional side (Figure 7(b)). These results show that inelastic deformation occurs near the tip of the fault and the fault evidently forms by the breakdown of this process zone when a critical crack density occurs within it. In the next section we will examine field data to see if this process occurs for natural faults.
aluminum. Under the conditions of fault propagation, rock does not undergo crystalline plasticity like those materials. Rock is rather a granular aggregate of several anisotropic crystalline phases which results in a highly heterogeneous stress field at the grain scale. At sufficiently high stresses, this will result in dilatant microcracking at the grain scale. This should occur near the tips of faults, resulting in a high concentration of inelastic deformation there, forming what is known as a brittle process zone.
6.10.3.4 Experimental Studies of Fault Propagation It is difficult to study fault propagation in the laboratory because in compression tests an instability usually occurs just at the beginning of the falling part of the stress–strain curve, so that data cannot be taken beyond that point. Lockner et al. (1991) were able to avoid this by using a dynamic feedback system to stiffen the system and surpress the instability. Figure 6 shows the complete stress-strain curve for one of their experiments and the locations of acoustic emissions during each stage. In stage (a) on the ascending part of the loading curve, acoustic emissions produced by dilatant microcracking (Scholz, 1968b), are uniformly distributed in the central region of the test specimen. The linear concentration of acoustic emissions in (b) indicates the nucleation of a fault, which subsequently propagates through the sample amid a cloud of acoustic emissions. The stress
6.10.3.5 Fault Scaling Laws and the Mechanics of Fault Growth Because faults are shear cracks with friction, as displacement D accrues on the fault, it must also grow in length L. In order to identify the proper crack model to use for faults, the scaling laws between D and L and other fault parameters must be determined. In this section, the observations will be presented and then discussed in the light of crack models. This will lead to an understanding of how faults propagate. 6.10.3.6
Displacement–Length Scaling
Early attempts to find a relationship between D and L consisted of collecting data on D and L from the
600
b c
a Westerly PCONF = 50 MPa
d e
400
σΔ(MPa)
f
200
0
0
1 2 Shortening (MM)
3
(a)
(b)
(c)
(d)
(e)
(f)
Figure 6 Left, complete stress–strain curve for failure of granite. Right, orthogonal views of the locations of acoustic emission for each stage of the stress–strain curve. From Lockner DA, Byerlee JD, Kuksenko V, Ponomarev A, and Sidorin A (1991) Quasi-static fault growth and shear fracture energy in granite. Nature 350: 39–42.
448
Fault Mechanics
Crack density across fracture
(a)
Total crack density (mm/mm2)
50 40
Compressional side
Dilational side
30 20 10
Uniform density at peak stress Initial crack
0 –50
density
0 Distance from fault (mm)
(b)
50
A-2–A-3 5
Fault plane
Cylinder axis
Crack length (mm) per square mm granite
4
3
2
1
0 90°
60°
30°
0° –30° Orientation (°)
–60°
–90°
Figure 7 (a) Crack density as a function of distance from the fault formed in a experiment like that shown in Figure 6. (b) Orientation of cracks as a function of angle from the rock cylinder and fault. From Moore DE and Lockner DA (1995) The role of microcracking in shear-fracture propagation in granite. Journal of Structural Geology 17: 95–114.
literature and trying to fit it to a relationship of the form D ¼ Ln
½16
Walsh and Watterson (1988) and Marrett and Allmendinger (1991) concluded that n is 2.0 and 1.5, respectively. However, for theoretical reasons, it is not possible for n to exceed 1, regardless of the crack model considered. If n > 1, eqns [7] and [15] show that stress drop will have to increase with length. The strength of the fault would thus increase with length. If this were the case faults would not grow: the rock would crush at the grain scale. Cowie and Scholz (1992a, 1992b) studied this scaling relation from the viewpoint of the
Dugdale–Barenblatt elastic-plastic crack model. They concluded that for this model, n ¼ 1 and that should increase linearly with rock strength. This latter conclusion indicates that the attempt to fit a universal law of the type given by eqn [16] is incorrect: rather, because faults occur in different rock types and depths that there should be a family of curves with different values of that should describe the scaling. They were able to show that the data of Marrett and Almendinger were consistent with n ¼ 1: the greater slope found by Marrett and Almendinger was because the longer faults sampled deeper into the crust and hence stronger rocks and because their data included plate boundaries, which have no tips and hence need not obey the scaling law.
Fault Mechanics
1000
0.02
Dmax
100
Dave D (m)
Normalized displacement
449
0.01
10
1
Slopes = 1
0
1 0.00 0.0
0.5 Normalized distance along fault
1.0
Figure 8 Displacement profiles normalized to fault length for normal faults in the Volcanic Tablelands, eastern California. The faults range in length from 690 to 2200 m. Modified from Dawers NH, Anders MH, and Scholz CH (1993) Growth of normal faults – Displacement-length scaling. Geology 21: 1107–1110.
In order to separate the variables in eqn [16], Dawers et al. (1993) studied a population of isolated (noninteracting) normal faults in a 200 m thick welded tuff in the Volcanic Tablelands of eastern California. In that locality there was a wide range of lengths of faults in a single rock type and which propagated at the surface, so the strength variable should be constant. They were also careful to select faults that were not interacting with other faults. As will be discussed in Section 6.10.5, the interaction of faults through their stress fields can severely distort their displacement profiles and hence their value of . The along strike displacement profiles for a number of their faults, ranging in length from 690 to 2200 m, normalized to fault length, is shown in Figure 8. A very good data collapse is achieved, indicating that these faults are self-similar. A plot of D versus L is shown in Figure 9. These data show that n ¼ 1 with a cross-over phenomenon occurring where L > 200 m, the layer thickness. When fault length is less that the layer thickness, the displacement profiles are peaked, whereas when L > 200 m, they are more flat-topped, as in Figure 8. This difference results in a slight change in at the cross-over. Thus, when the faults are 3-D cracks, propagating in two directions, they are not selfsimilar with the 2-D cracks that have breached the
10
100 1000 Length (m)
10 000 100 000
Figure 9 Displacement length scaling for the Volcanic Tableland faults. Modified from Dawers NH, Anders MH, and Scholz CH (1993) Growth of normal faults – Displacement-length scaling. Geology 21: 1107–1110.
brittle layer thickness and propagate only in the horizontal direction. A summary diagram showing D–L scaling using the global data set is shown in Figure 10. Here, the data has been edited much more carefully than in the earlier compilations of Walsh and Watterson and Marrett and Almendinger. For example, faults that are plate boundaries have been removed. Plate boundary faults do not have defined lengths and hence do not need to obey this scaling law. A ridge–ridge transform fault, for example, is stress-relieved by the spreading ridges at both ends and hence can accrue slip with no need to grow longer because it does not have tips where displacement must go to zero, producing there a large stress concentration. The linear scaling of D with L is still apparent in Figure 10, though with more scatter than in Figure 9. Much of this has to do with strength variations between the different faults, because the scaling parameter is a stress drop between the yield strength of the rock and the residual friction. For example, the longer faults (> 10 km) penetrate the crust and therefore rupture crystalline rock at high pressure and therefore have higher values. The faults with < 102 are in soft sedimentary rocks like sandstones and shales. Another source of scatter is that there has been no attempt to separate interacting from noninteracting faults (Schlische et al., 1996), which also influences the value. The effects of fault interactions will be discussed in Section 6.10.5.
450
Fault Mechanics
107 106 105 104
102
=
10
–1
L
101 D
100
10
–3
L
Displacement (m)
103
Elliott (1976)–T–29 Villemin et al. (1995)–N–26 Opheim and Gudmusson (1989)–N–7 Krantz (1988)–N–16 Walsh and Watterson (1987)–N–34 Dawers et al. (1993)–N–15 Peacock (1991)–SS–20 Peacock (1991)–N–20 Muroaka and Kamata (1983)–N–15 McGrath (1992)–N–39 Schlishe et al. (1996)–N–201
D
=
10–1 10–2 10–3 10–4 10–3
10–2
10–1
100
101
102
103
104
105
106
107
Length (m) Figure 10 Summary diagram displacement-length scaling for faults. Modified from Schlische RW, Young SS, Ackermann RV, and Gupta A (1996) Geometry and scaling relations of a population of very small rift-related normal faults. Geology 24: 683–686.
6.10.3.7
Fault Tip Taper Scaling
The data in Figure 8 indicate that for noninteracting faults in a uniform rock type the displacement profiles are symmetric with linear, scale-invarient fault tip tapers (FTTs). The linear displacement tapers of faults has also been noticed elsewhere (Burgmann et al., 1994; Cartwright and Mansfield, 1998; Cowie and Shipton, 1998) FTT is plotted versus length for a variety of faults and earthquake ruptures in Figure 11 (Scholz and Lawler, 2004). Figure 11(a) shows the tapers from the Volcanic Tablelands data illustrated in Figures 8 and 9. These data show that FTT is scale independent over the two orders of magnitude range of fault length. Figure 11(b) shows data from other sites, separated into isolated and interacting faults and earthquake ruptures. The interacting faults are overlapping subparallel normal faults that will be discussed in the next section. The tapers for interacting rupture tips are always greater, by up to a factor of 10, than for the isolated ruptures. The scatter of FTT for the isolated faults is much greater than for the data from the Volcanic
Tablelands because they are in a variety of rock types. The data are broken out by rock type in Figure 12. There we see that, among the isolated faults, FTT is greater for faults in granite than for faults in sedimentary rock, averaging 8 102 and 1 102, respectively. The tapers in the welded tuff (Figure 11(a)) are intermediate, averaging 5 102. Thus, like the D/L ratio, FTT increases with rock strength. Both parameters for earthquake ruptures are about two orders of magnitude less than for faults. The D/L ratio for earthquake ruptures is in the range 105104 (Scholz, 2002, p. 207). This is because the for earthquakes is the difference between static and dynamic friction and hence is much smaller than stress drip for faulting.
6.10.3.8
Process Zones and Their Scaling
Just as in the case of the laboratory experiments shown in Figure 7, faults are known to have brittle process zones, or regions of intense microcracking
Fault Mechanics
451
(a)
Taper
100
10–1
10–2 1 10
103
102
104
Fault length (m)
(b) 100
10–1
10–2 Taper
1
10–3
10–5
10–3
Isolated exterior earthquake tip Isolated interior earthquake tip Isolated fault tip
2 3
4
Interacting exterior earthquake tip Interacting interior earthquake tip Interacting fault tip
10–1
101 Length (m)
103
105
Figure 11 (a) Fault tip tapers for isolated faults from the Volcanic Tablelands vs. length. (b) The same for other faults and earthquakes, sorted according to interating or not. Modified from Scholz CH and Lawler TM (2004) Slip tapers at the tips of faults and earthquake ruptures. Geophysical Research Letters 31: L21609.
surrounding them (Anders and Wiltschko, 1994; Chernyshev and Dearman, 1991). Scholz et al. (1993) proposed that the process zone is dominated by dilatant grain scale microcracks formed by the fault tip stresses. Such cracks are opening mode (tensile) cracks that form parallel to the local 1 direction. An empirical dilatancy–stress relation (from Scholz, 1968a) was convolved over the fault tip stresses as the fault propagated past the point in question. This predicts an exponential increase of crack density as the fault is approached. For a mode II (strike-slip) shear crack the crack tip stresses are asymmetric across the fault. As shown in Figure 13, this predicts
different orientation of microcracks on either side: with a maximum peaked at about 30 from the fault on the compressional side, and 70 on the extensional side. These model results agree quite well with the experimental results of Moore and Lockner (1995) shown in Figure 7. To test this model for natural faults, Vermilye and Scholz (1998) measured microcrack densities in traverses across several strike-slip faults in quartzite in the Shawangunk Mountains of New York. This rock type was chosen because quartz has no cleavage that would influence microcrack orientation and because microcracks are well preserved in it. Microcrack
452
Fault Mechanics
1.000 00
0.100 00
Log (taper)
0.010 00
Crustal scale
0.001 00
Igneous plutons Sandstone-shale/ layered sedimentary Siltstone/layered sedimentary
0.000 10
Limestone/layered sedimentary Coal/sedimentary Volcanic tuff
0.000 01 0.001
0.01
0.1
1
10
1000 100 Log (length) (m)
10 000
100 000
1 000 000
Figure 12 The same as Figure 11(b) excepted sorted according to rock type. Modified from Scholz CH and Lawler TM (2004) Slip tapers at the tips of faults and earthquake ruptures. Geophysical Research Letters 31: L21609.
density is plotted versus log distance from the fault in Figure 14 for two faults, one 40 m long (circles) and the other 2 m long (squares). Microcrack density increases exponentially as the faults are approached, in agreement with the theory and experimental results. A nearby exposure of a dexteral strike-slip fault exposed on a cliff face is shown in Figure 15. Lower hemisphere contoured microcrack pole plots are shown for several locations, projected onto the horizontal surface. Background microcracks measured several meters from the fault (MC16) show a girdle pattern. This is interpreted as dilatant microcracks parallel to the regional stress, inclined about 25 to the fault. The near-fault microcracks (MC1 and MC2) have different orientations on either side of the fault, those in MC1 striking at low angles to the fault and those in MC2 at high angles. This is consistent with a dextral fault propagating toward the viewer, with MC1 being on the compressional side and MC2 on the extensional side. At the tip of the fault (MC5-W and MC5-E) the microcracks are fault parallel. This is interpreted as due to tensile mode release of fault dilation during unroofing from the 5–7 km depth at which this fault was formed.
The data in Figure 14 indicate that the maximum microcrack density, which occurs where the rock breaks down to cataclasite, does not depend on fault length, but that the width of the process zone P, as measured by the distance from the fault at which microcrack density reaches the background level, does depend on L. Data collected from various sources (Figure 16) indicates that this scaling is linear. In porous rocks such as the Berea and Navaho sandstones, in which the porosity is nearly 20%, shearing induces porosity reduction rather than dilatancy (Lockner et al., 1992). This results in the formation of deformation bands, local regions of reduced porosity (Aydin and Johnson, 1978; Mair et al., 2000). The damage zones of faults in such rocks consist of numerous deformation bands, with little increase of microcrack density and the damage zone width increases with throw on an individual fault (Shipton and Cowie, 2001). The deformation thus appears to consist mainly of grain sliding and rotation, resulting in compaction, rather than internal microcracking in grains, and the damage zone development does not appear primarily as a crack tip phenomenon but increases with strain.
Fault Mechanics
453
0.06
Y distance
0.03
0.00
–0.03
–0.06 –0.06
–0.03
0.00
0.03
0.06
X distance 12.0 11.0 10.0 9.0 8.0
Density
7.0 6.0 5.0 4.0 3.0 2.0 1.0 0.0 0.0050 0.0075 0.0100 0.0125 0.0500 0.0175 0.0200 0.0225 0.0250 0.0275
Distance Figure 13 Predictions of the process zone model of Scholz et al. (1993). Top, stresses around a strike-slip fault tip: dashed, compression, solid, tension. Lower left, orientation of expected microcracks on each side of the fault. Lower right, expected crack density as a function of distance from the fault. Modified from Scholz CH, Dawers NH, Yu JZ, and Anders MH (1993) Fault growth and fault scaling laws – Preliminary-results. Journal of Geophysical Research, Solid Earth 98: 21951–21961.
6.10.3.9
Cataclasite Zone Scaling
The cataclasite zone is the band of ground rock fragments found in the core of the fault. If the grain size is clay sized it is often called fault gouge, which may or may not contain clay minerals. Some material of this type is initially formed in the innermost part of the process zone where the crack density reaches a critical value to fragment the rock and form the fault.
As displacement occurs on the fault, additional cataclasite is formed by frictional wear. A simple model for steady-state wear predicts (Scholz, 2002, p. 78) that the thickness of wear material T scales as T ¼
D 3h
½17
where is the normal stress, D fault displacement, is a dimensionless parameter called the wear
454
Fault Mechanics
W
E
70
70
60
60
50
50
40
40
30
30
20
20
10
10
0 10 000
1000
0 10 1 1 10 Distance from fault (mm)
100
100
1000
10’000
Figure 14 Microcrack density as a function of distance from two strike-slip faults in the Shawangunk quartzite, New York, one 40 m long (circles) the other 2 m long (squares). Dashed line is the background density. Modified from Vermilye JM and Scholz CH (1998) The process zone: A microstructural view of fault growth. Journal of Geophysical Research, Solid Earth 103: 12223–12237.
13
SW
up
NE
11
su
es
Pr
•
re
4
n
tio
lu
so
Micro veins
3
MC16
14 2 15 1
MC2
σ1
MC1
16
Plane of stereographic projections
8
a
MC5-W
5b
0
mm 100
MC5-E
Figure 15 Contoured pole plots of microcrack orientions around a strike-slip fault in the Shawangunks. From Vermilye JM and Scholz CH (1998) The process zone: A microstructural view of fault growth. Journal of Geophysical Research, Solid Earth 103: 12223–12237.
Fault Mechanics
10 000 Process zone width (m)
1000 100 10 1 0.1 0.01 0.1
1
10
100 10 000 Fault length (m)
1 000 000
Shawangunk faults, this study Anders and Wiltschko, 1994 Chernyshev and Dearman, 1991 Brock and Engelder, 1977 Little, 1995 P = .016 * L R ∧ = 0.91
Figure 16 Scaling of process zone width with fault length. From Vermilye JM and Scholz CH (1998) The process zone: A microstructural view of fault growth. Journal of Geophysical Research, Solid Earth 103: 12223–12237.
SAF
10
–1
10 0
10
–2
log D (slip, m)
–3
4
–4 –4
0 log T (thickness, m)
4
Figure 17 Scaling of cataclasite zone thickness with fault displacement. From Scholz CH (1987) Wear and gouge formation in brittle faulting. Geology 15: 493–495.
coefficient, and h is a measure of the rock strength or friability. This wear law is confirmed by experiment (Wang and Scholz, 1994; Yoshioka, 1986). Observational data for faults, shown in Figure 17, indicate linear scaling between T and D (Evans, 1990; Hull, 1988; Scholz, 1987). There is a large scatter in the data, which may reflect vatiations in the other parameters in eqn [17]. Even for a single fault there is a large variation in T along strike: for example, along
455
the San Andreas Fault (SAF in Figure 17) the cataclasite thickness varies by more than an order of magnitude. This may reflect variations of rock type and fault geometry and structure. Wilson et al. (2005) have pointed out a zone some 70–100 m wide adjacent to the San Andreas Fault consisting of rock that has been pulverized in situ. This is well outside the cataclasite zone in this region. They argue that this is due to dynamic pulverization produced by the passage of earthquakes, either by the rupture tip stress field or fault normal loading and unloading. They report that the surface energy in this process is an appreciable fraction of the work of faulting, but this has been contradicted by similar work on the nearby Punchbowl Fault (Chester et al., 2005).
6.10.3.10 Interpretation of the Scaling Laws in Terms of Crack Models The scaling laws for faults are summarized in Table 1. In all cases the scaling laws are linear. Typical values are given by order of magnitude. In all cases, the actual value of the scaling parameters depends upon rock strength. The D/L ratio increases with strength, and the T/L and P/L ratios should decrease with strength. The scale-independent FTT increases with strength. We can now compare these observations with the crack models discussed earlier. The elastic crack model can immediately be eliminated: it does not predict linear D–L scaling, a linear fault tip displacement taper, or a process zone. Both the Dugdale– Barenblatt and CFTT models predict linear D–L scaling but the presence of a linear FTT and a volumetrically extended process zone are consistent only with the CFTT model. This model predicts also that G scales linearly with L, which is independently confirmed by the finding that P scales linearly with L. This model also predicts a linear FTT and that FTT should increase with rock strength. A consistent picture emerges regarding the mechanism of fault propagation. The crack tip stress field produces pervasive cracking in the adjacent rock. This is primarily in the form of dilatant microcracks at the grain scale: opening mode cracks parallel to the local 1 direction. Most cracking is at the grain scale because it is at that scale that stress heterogeneity, caused by elastic mismatches between grains, is highest, and hence where the highest stress concentrations are found (Scholz, 1968a). Macroscopic deformation features, such as secondary faults and pressure solution
456
Fault Mechanics
Table 1
L D P T
Typical values of fault scaling parameters Length (L)
Displacement (D)
Process zone width (P)
Cataclasite zone width (T)
1 102 102 104
102 1 1 102
102 1 1 102
104 102 102 1
bands, are also found within the process zone, and exhibit the same falloff with distance from the fault as the microcracks, but they are not nearly as ubiquitous (Vermilye and Scholz, 1998). When the microcrack density reaches a critical level, the rock fragments into a cataclasite on which the fault may then slide frictionally, and the process continues. This is the breakdown fault-forming process as originally visualized by Cowie and Scholz (1992a) and observed in the laboratory experiments of Lockner et al. (1991) and Moore and Lockner (1995). It results in a wide zone of distributed fractures containing a narrow zone of comminuted rock within which the main shearing occurs (Chester et al., 1993; Chester and Logan, 1986). According to the scaling parameters of Table 1, the cataclasite zone width is typically about 102 of the process zone width, although, as mentioned above, the former may vary considerably along strike. That however is a 2-D view, whereas the actual scaling is 3-D. Imagine a fault of finite length that initiated at its center and propagated bilaterally in both directions. Because the width of the process zone scales with the length of the fault at the time the fault passed the particular point in question, it is narrowest in the center of the fault and widens to maxima at both tips. On the other hand, the width of the cataclasite zone scales with fault slip, so it will be widest at the fault center and narrowest at the tips. However no one has yet made enough fault traverses to verify these conclusions.
6.10.4 Fault Interactions Slip on a fault produces stress changes in the surrounding regions. One example is shown in Figure 18 (Das and Scholz, 1981), which shows the change of fault parallel shear stress adjacent to a vertical strike-slip fault. Because faults usually grow in populations or systems rather than as isolated features, they commonly grow within the stress change region of other faults. This will affect the behavior of the fault. If it propagates into a region of stress enhancement, it
will be induced to propagate more readily. If it propagates into a stress-drop region, its propagation will be impeded and it may become pinned, unable to propagate farther. The displacement distribution on the fault, as mentioned earlier, will become distorted from what would be expected for a crack in a uniform stress field. Crack tip stress for mode II cracks, such as the strike-slip fault shown in Figure 18, are different from that of mode III cracks, such as normal faults, so the stress interactions will differ between them and we treat them here separately. Fault forming in a given tectonic environment, and hence stress field, are usually subparallel, so the most common interactions, and those the most studied, are between subparallel faults. With one exception, we will be limited to these types of interactions.
6.10.4.1 Faults
Mode III Interactions: Normal
The most extensive studies of fault interaction have been done for subparallel normal faults. This is because for normal faults, if they are not highly eroded, one can easily measure the slip distribution from scarp height. To do this for strike-slip faults one needs offset markers, which are seldom frequent enough. 6.10.4.1.1
Pinning Peacock (1991) and Peacock and Sanderson (1994) pointed out that the displacement distribution of overlapping normal faults is, unlike the examples shown in Figure 8, asymmetric. The peak in displacement is shifted toward the overlapping end and the FTT at that end is steeper than at the distal end. In Figure 19 is shown the displacement profiles for three different times for the main bounding fault system of the East African rift on the west side of Lake Malawi. These profiles were obtained by seismically backstripping seismic profiles in the lake at three time markers several million years apart (Contreras et al., 2000). For the two way travel time (TWTT), one second is approximately 1 km. This fault system is comprised of three overlapping subparallel normal faults, the South
Fault Mechanics
457
–19 16 25 21
30
11
–69 6
–
–19
–14 6
0.0
–9
–
–4
0.0
Figure 18 Change in fault parallel shear stress due to slip on a strike-slip fault. Stress drop 100 MPa, contours in MPa. From Das S and Scholz CH (1981) Off-fault aftershock clusters caused by shear stress increase? Bulletin of the Seismological Society of America 71: 1669–1675.
3.0
N
S
T W T T (s)
North fault
Central fault
South fault
2.0
1.0
0.0 25
50
75
100 125 150 Along-strike distance (km)
175
200
225
Figure 19 Displacement profiles for the three main bounding faults on the west side of the Malawi rift at three times separated by several million years. From Contreras J, Anders MH, and Scholz CH (2000) Growth of a normal fault system: Observations from the Lake Malawi basin of the east African rift. Journal of Structural Geology 22: 159–168.
fault and the North fault offset 5 and 7 km to the east of the Central fault, respectively. The temporal information allows us to observe the progression of interaction between the indivual faults as slip accrues
on them. As time progresses, the overlap of the South and Central faults increases and as it does so, the FTT at the north end of the South fault steepens, whereas the FTT on its noninteracting southern tip remains
458
Fault Mechanics
much less steep. Its displacement profile becomes more and more asymmetric and most of the propagation is to the south, the rate of propagation of the overlapping tip becoming progressively slower as the overlap increases, until finally the fault becomes completely pinned. Much the same thing happens with the North fault, whereas the Central fault becomes pinned at both ends, accumulating displacement without becoming longer. The effect of the pinning is that the D/L ratio of this fault grows with time. This increased D/L ratio is not because the fault is becoming stronger, but because its tips are prevented from growing because they are in the stress-drop shadow of the overlapping faults. This is one of the effects that introduces scatter in D/L plots like Figure 10. The high FTTs at interacting tips shown in Figure 12(b) are also the result of this pinning mechanism. Gupta and Scholz (2000a) studied the systematics of this interaction from a population of centimeter to decimeter scale normal faults exposed on the bedding planes of siltstones in the Solite Quarry in North Carolina. They measured the degree of disturbance of the displacement profiles for interacting tips compared to a noninteracting standard. They found this to be a function of offset and separation in such a way that the more highly interacting tips were deeper in the stress-drop region of the adjacent fault. This is shown in Figure 20, where 25 2 segments 3 segments 4 + segments N and C Malawi C and S Malawi
Standard deviations
20 15 10 5 0 300
250
200 150 100 Stress drop (MPa)
50
0
Figure 20 Distortion of the fault tip of interacting normal faults, in standard deviations from noninteracting faults vs. the magnitude of their stress shadows calculated from an elastic model. The lines and symbols show the temporal change in the same measurements from the Malawi fault segments shown in Figure 19. From Gupta A and Scholz CH (2000a) A model of normal fault interaction based on observations and theory. Journal of Structural Geology 22: 865–879.
the displacement anomaly is measured in standard deviations from the standard and the stress drop is calculated from an elastic crack model using the offset and separation. A simple diagram illustrating the pinning mechanism is shown in Figure 21. The stress field for the driving shear stress of a mode III crack is symmetric, with a stress-drop region (solid contours) on either side of the crack and stress increase regions beyond the tips. Imagine fault F9 to be stationary and being approached by a propagating fault F. Fault F will first be attracted to fault F9 by the enhanced fault tip stress (negative stress drop) of the latter. Once fault F passes the tip of F9, it will enter its stress-drop region and be repulsed, that is, its further growth will be inhibited. In order to continue to propagate, the stress concentration at its tip must increase, to overcome both the rock strength and the stress drop from fault F9. This stress concentration is determined by the FTT, so this must become steeper at the interacting end, and steepen further as tip grows deeper into the stress shadow. As a result, the displacement profile becomes asymmetric. A similar result has been obtained by modeling using the elastic crack model (Willemse, 1997; Willemse et al., 1996). As faults more likely interact than not, asymmetric profiles are more common than symmetric ones (Manighetti et al., 2004). 6.10.4.1.2
Coalescence Figure 22 shows the summed displacement profiles for the three interacting Malawi faults shown in Figure 19. The dashed profile is the expected profile for a single fault with a net length equal to the three fault segments. Notice that as the faults grow, and their interactions increase, their summed profiles become closer to that of a single fault. This indicates that the interacting faults increasingly act as a single unit. This coalescence is a gradual process, it can occur by the faults ‘soft-linking’ through their stress fields, or through connecting by means of secondary faults, that is, hard linking. Figures 23 and 24 show three faults in the Volcanic Tablelands that are in the process of coalescing (Dawers and Anders, 1995). In the vicinity of the relay ramps between the overlapping tips of the major segments many smaller faults have formed, some of which cross the relay ramps obliquely. Their combined effect (Figure 24) is to eliminate the displacement minima that would be expected in the relay ramp regions, thus smoothing out the overall displacement profile so that it appears more like that on a single long fault, just as in the case shown in Figure 22. One of course sees only a portion of the linking phenomena in
Fault Mechanics
459
Region of repulsion
Δσ
F –Δσ
F′
Region of attraction
FTT′
FTT
σy
Isolated fault
σt
σf σC = σy – σt where σC α F T T Interacting fault
σC′ = σy – (σt – Δσ) F T T′ > F T T
Figure 21 A diagram illustrating the fault pinning interaction.
N
S
2.5 3.0
TWT T (s)
2.5 2 1.5 1 0.5 0 0
25
50
75
100
125
150
175
200
225
250
Figure 22 Summed displacement profiles for the three Malawi faults of Figure 19 at different time intervals. The dashed curve is the expected profile for a single fault with length equal to the summed lengths. Modified from Contreras J, Anders MH, and Scholz CH (2000) Growth of a normal fault system: Observations from the Lake Malawi basin of the east African rift. Journal of Structural Geology 22: 159–168.
a section such as shown in this example, one needs information in three dimensions to observe the total picture (Crider and Pollard, 1998). 6.10.4.1.3
Nucleation inhibition If a fault lies entirely within the stress-drop region of another, larger, fault, it will become inactive. Figure 25
shows several examples for normal faults in the Solite Quarry (Ackermann and Schlische, 1997). In this case the larger faults predated the smaller ones, as there is an absence of the smaller ones in their stress-drop regions (stress shadows). Otherwise there would be inactivated small faults in their stress shadows and one would not be able to observe this phenomenon.
460
Fault Mechanics
These three interaction mechanisms are illustrated in Figure 26. As we shall see in Section 6.10.6, these three mechanisms, working together in a system of growing faults, result in the development of populations of faults with distinct statistical properties.
Northern segment
Dmax = 93 m
6.10.4.2
Mode II Interactions
Unlike mode III cracks, mode II cracks, such as strikeslip faults in horizontal section, have asymmetric stress fields on either side of their tips. One side is extensional, the other compressional. There are therefore different interactions for overlapping jogs depending on their sign, compressional or extensional.
Middle segment
6.10.4.2.1 Southern segment
Study area
0
1 km
Figure 23 Map view of interacting normal faults in the Volcanic Tablelands, California. Modified from Dawers NH and Anders MH (1995) Displacement-length scaling and fault linkage. Journal of Structural Geology 17: 607–614.
Interactions at strike-slip jogs Jogs are offsets of strike-slip faults and are the primary reason for the segmentation of such faults. If the sign of the stepover between the fault segments (right or left) is the same as the sign of the fault (right lateral or left lateral) the jog is extensional; if opposite, it is compressional. Both kinds of jogs are observed. Figure 27 shows a number of small scale jogs. Figure 27(a) is a schematic diagram of an extension jog in the Sierra Nevada granite (Martel and Pollard, 1989). In that locality the strike-slip faults are reactivated joints, so they have very small D/L ratios as compared to features that originated as strike-slip faults. The jog forms at what they call splay faults, wing cracks that originate from the tips of the fault
100 90
Normalized single-segment profiles
Total throw
80 70
Mapped scarps
Throw (m)
60 50 40 30 20 10 0 0
1000
2000 3000 4000 Distance (m) from S fault-tip
5000
6000
7000
Figure 24 Displacement profiles for the faults of Figure 23. The summed throw and the expected displacement profile for a single noninteracting fault is also shown. From Dawers NH and Anders MH (1995) Displacement-length scaling and fault linkage. Journal of Structural Geology 17: 607–614.
Fault Mechanics
461
Crack interactions
Forbidden zone
1. Nucleation inhibition
2. Pinning
Figure 25 Photos of nucleation inhibition in small normal faults from the Solite Quarry, North Carolina. Top: A normal fault footwall (FW) with a dearth of smaller faults in its stress shadow region, outlined in chalk. Bottom: View of three larger faults with no small faults in their stress shadow regions – faults are highlighted with chalk for visibility. From Ackermann RV and Schlische RW (1997) Anticlustering of small normal faults around larger faults. Geology 25: 1127–1130.
on the extensional side, and propagate across the gap between the faults. When this occurs a rhomboidal cavity is formed (Figure 27(b)). Such jogs have been modeled numerically by Martel et al. (1988), Burgmann and Pollard (1994), and Pachell and Evans (2002) and with an analog model by Dooley and McClay (1997). They occur on a wide range of scales, forming pullapart basins at the larger scale. One of the larger of these is the basin containing the Dead Sea on the Dead Sea transform (Al-zoubi and Brink, 2002). At these crustal scale cases, inward facing normal faults lie in the place of the wing cracks shown in Figure 27(a). Aydin and Nur (1982) claim that pull-apart basins are self-similar structures, always being about twice as long as wide. However, they may have various temporal histories, which can produce some variability in this (Wakabayashi et al., 2004). The Hanmer Basin on the Hope Fault in New Zealand, for example, has a particularly complex history (Wood et al., 1994).
3. Coalescence (hard linking) Figure 26 Illustration of the modes of interactions of overlapping normal faults.
Figure 27(c) shows a rhombic cavity cross-cut by a ‘shunt’ fault. The 1992 Landers earthquake ruptured across such a shunt fault, the Kickapoo Fault, that crosses the extensional jog between the Johnson Valley and Homestead Valley Faults in California (Figure 28) (Sowers et al., 1994). Slip on the right-lateral Johnson Valley and Homestead Valley Faults both tapered to zero in their overlap regions (Figure 28, lower panel) but slip on the Kickapoo Fault was just enough to keep the overall summed slip constant at about 3 m (Figure 28, middle panel). Thus, the action of the shunt fault precluded the development of much extension at this jog during the earthquake. Notice the very high tip tapers of the Kickapoo Fault as it approaches (but does not touch) the other faults. This FTT is about a factor to 10 greater than that for noninteracting earthquake ruptures (Figure 11(b)). A small compressive jog in quartzite is shown in Figure 28(d). Within the rhombic deformation zone, fine pressure solution lamellae occur in orientations highlighted by the white lines. These are perpendicular to the maximum compression direction in the
462
Fault Mechanics
(b)
(a)
10 cm
Splay crack
Slip patch Boundary fault ahead of slip patch
(c)
(d) 10 cm
10 cm
Figure 27 Illustrations of jogs in strike-slip faults. (a) Schematic diagram of jog formation between joints reactivated in shear (from Martel and Pollard, 1989). (b) An extensional jog in a left lateral fault. (c) An extensional jog containing a shunt fault. (d) A compressional jog. White lines highlight the orientation of pressure solution selvages. Jogs (b), (c), and (d) are from the Shawangunk quartzite.
Homestead Valley Fault 1 km N
Kickapoo Fault
Johnson Valley Fault
3 2 1m
Figure 28 The Kickapoo Fault is a shunt fault between the right lateral Homestead Valley and Johnson Valley Faults activated in the 1992 Landers earthquake. Top, map view; bottom, slip on each fault during the earthquake; middle, summed slip on the three faults. Modified from Sowers JM, Unruh JR, Lettis WR, Rubin TD (1994) Relationship of the Kickapoo Fault to the Johnson Valley and Homestead Valley Faults, San-Bernardino County, California. Bulletin of the Seismological Society of America 84: 528.
Fault Mechanics
jog. At the crustal scale, thrusts or anticlines (Campagna and Aydin, 1991) occur with these orientations. They have been studied with analog models by McClay and Bonora (2001).
6.10.4.2.2
Strike-slip splay faults Splay faults are common in strike-slip systems. Prominent examples from the Alpine and San Andreas systems in New Zealand and California are compared in Figure 29, where the figures have been rotated such that their relative plate motion direction are nearly parallel. Both the Alpine Fault and the Big Bend region of the San Andreas Fault are oblique in a transpressive sense from the plate motion direction. The Marlborough faults of New Zealand, such as the Hope, Awatere, and Wairau and the splay faults of southern California, such as the San Jacinto, Elsinor, and San Clemente, are, in contrast, pure strike-slip faults oriented parallel to the plate motion direction. Similarly, in northern California, the Calaveras and Hayward faults splay from the San Andreas where it too is transpressive. In Figure 30 Global Positioning System (GPS) velocities and the geologic slip rates for the various faults in the San Andreas system in southern California are shown. In this region only 22 of the entire 50 mm yr1 plate motion is taken up in the San Andreas Fault proper, the rest occurring on the splay faults (Bourne et al., 1998b). A very similar slip distribution occurs in the New Zealand case (Bourne et al., 1998a). A detail of the junction between the San Jacinto and San Andreas Faults is shown in Figure 31. The GPS data indicate that north of this juncture the San Andreas moves at a rate of 34 mm yr1 and 12 mm yr1 of that is transferred to the San Jacinto. It is common to think of such transfers as kinematic, that is, the slip divides between two contacting faults, a part continuing on the master fault, the other part on the splay. This would require that the splay offset the master fault, which Figure 31 shows not to be the case. This cannot be because the San Jacinto is too young to have produced an offset, because its net slip is about 20 km, measured some 100 km south of the junction (Kendrick et al., 2002). Furthermore, close examination shows that the San Jacinto Fault does not touch the San Andreas. If it is considered to be continuous with the Glen Helen Fault, it curves to near parallelism with the San Andreas before disappearing. At its tip, of course, the slip must be zero. This noncontacting and nonoffsetting nature of splay junctions is a feature of all the splays of the San Andreas and Alpine Fault systems.
463
No one has attempted to model the splay interaction, but some conjectures can be made. If the slip transfer is not kinematic, it must be dynamic, that is, facilitated by the interacting stress fields of the two faults. The splay fault presumably forms later than the master fault and is prevented from touching it by the stress shadow of the master fault. Its tip is therefore pinned. The slip rate on the San Jacinto Fault must be zero at its tip and increase with distance to the south, up to its full 12 mm yr1. Similarly, the slip rate on the San Andreas Fault must not suddenly decrease, but gradually do so as the slip rate on the San Jacinto Fault increases. As net slip increases on the San Jacinto Fault, the slip gradient at its tip increases, producing an ever increasing stress concentration in the vicinity of its tip. This will cause its tip to propagate, but at an ever decreasing velocity as it goes deeper into the stress shadow of the San Andreas Fault. Some effects of this resulting large stress concentration can be seen in Figure 31. Just NE of the tip, where the stresses should be extensional, is the San Bernadino Basin. This 1.7 km deep basin has basement vertically offset 1.2 km, down to the east, on the San Jacinto Fault (Stephenson et al., 2002). On the other, compressive side, are the uplifted San Gabriel Mountains and the Cucamunga thrust fault. Is the Cucamunga thrust just another transverse range bounding fault, or does it owe its existence to the San Jacinto Fault tip stress field? In its corresponding position at the nearly identical splay junction of the Hope and Alpine Faults, the 1996 M 6.7 Arthur’s Pass earthquake occurred. It had a thrust mechanism, dipping in to the Alpine Fault (Abercrombie et al., 2000).
6.10.5 Fault Populations It has long been known that earthquakes obey a well defined size distribution, the Gutenberg–Richter relation. When defined in terms of seismic moment it is a power law. This kind of distribution is typical of fractal objects, and it can be argued from a geometrical point of view that this arises because they are self-similar (Turcotte, 1992). Earthquakes are selfsimilar: like faults the slip in an earthquake scales linearly with its length (Scholz, 1982). We might then expect faults to have populations with similar power law statistical properties. It turns out, though, that faults can have more than one kind of population. In this section we describe these populations and the mechanisms by which they are generated.
464
Fault Mechanics
Figure 29 The Alpine Fault system (top) compared with the San Andreas system (bottom). The top map has been rotated for comparison. From Yeats R and Berryman K (1987) South Island, New Zealand and transverse Ranges, California, a seismotectonic comparison. Tectonics 6: 363–376.
Fault Mechanics
(a)
465
244°
240°
40° Utah
Nevada
°
38°
C
0 24
r ifo al
36°
a
ni
Arizona
° 33
34° 32°
PV Ba ja ca lifo
SC
°
30°
ia rn
1 24
28°
°
32
47 mm a–1
SA
EL
SJ
2°
(b)
24
50
San Clemente Palos Verdes
Elsinore
San Andreas San Jacinto
Pacific Plate
Strike parallel velocity/mm a–1
3 mm a–1 6 mm a–1
40
6 mm a–1
30 12 mm a–1
20
22 mm a–1
10
0
North American Plate
–200
–150
–50
–100
0
50
100
Figure 30 (a) Map of the San Andreas Fault system in southern California, with location of GPS sites as dots. (b) Strike parallel velocities from GPS data. Staircase is geological slip rates of the faults. Curve is velocities expected from a model. From Bourne SJ, England PC, and Parsons B (1998b) The motion of crustal blocks driven by flow of the lower lithosphere and implications for slip rates of continental strike-slip faults. Nature 391: 655–659.
6.10.5.1
Observations of Fault Populations
6.10.5.1.1
Power law distributions Power law distribution for faults have the form N ðLÞ ¼ aL – c
½18
where N(L) is the number of faults greater than length L. This is a fractal distribution and has been described
many times for faults (for a collection of papers on this topic, see Cowie et al. (1996)). However, the data are seldom convincing that this is the correct distribution or what is the correct value of the exponent (Cladouhos and Marrett, 1996). A major problem is that the data are usually obtained from mapping, which has very little dynamic range. An example is shown in Figure 32. There, the size distribution of
466
Fault Mechanics
faults in the Volcanic Tablelands was measured at three scales. For each scale range there is a limited straight section below which the data roll-off because of a lack of resolution of the smallest faults. At the top end there is another roll-off, for which there are two
Sa
nA
nd
re
as
Fa u
lt
10 km
G
le
n
H
el
en
San Gabriel Mnts.
Fa u
lt
Cucamonga Fault Ly t
le
C
re
ek
Sa nB ern ard ino
Fa u
lt
lt
au oF
int
ac nJ
Sa
Ba sin
Figure 31 Map of the confluence of the San Jacinto Fault with the San Andreas Fault.
causes. The first is data censoring: the faults become larger than the map size and hence their lengths are underestimated. The second problem is that because the number of faults measured was limited, a cumulative distribution was plotted. Burroughs and Tebbens (2001) showed that a cumulative plot of a power law distribution with an upper cutoff produces a roll-off at the upper end and an overestimate of the exponent of the straight segment. As a result of these various problems, plots such as Figure 32 are not very convincing of a power law and overestimate the exponent. These problems are overcome in Figure 33. The data are from normal faults in the low plains of Venus, obtained by the synthetic aperture radar of the Magellan spacecraft (see Scholz, (1997) for an illustration of the image used). There is no dynamic range problem because the image was 200 200 km and the pixel size was 70 m. The image analysis software neglected faults that touched the edge of the image, which removed the censoring problem, and there were enough faults in the population, 1750, that a frequency plot could be used, avoiding the Burroughs and Tebbens problem. One can see that it is clearly a power law distribution with an exponent of 2. The equivalent exponent for a cumulative distribution would then be 1, a little smaller than the exponent
Faults per unit area ≥ L (no sq km–1)
100.000
10.000
1.000
1 0.100 2 3 0.010 4
0.001 1000
10 000
100 000 Length (m)
1 000 000
10 000 000
Figure 32 The length distribution of faults for the Volcanic Tablelands mapped at three scales. From Scholz CH, Dawers NH, Yu JZ, and Anders MH (1993) Fault growth and fault scaling laws – Preliminary-results. Journal of Geophysical Research, Solid Earth 98: 21951–21961.
Fault Mechanics
104
467
where NT is the total number of faults and is the reciprocal of the mean fault length.
Normal faults, Venus 31N332 y = 1.8209e+09*x ^ (–2.0222) R = 0.98493
log n(L)
1000
6.10.5.1.3
100 10 1 100
104
1000
105
log L(m) Figure 33 Frequency plot of the length distribution of faults of the low plains of Venus. Taken from a synthetic aperture radar image from the Magellan mission. From Scholz CH (1997) Earthquake and fault populations and the calculation of brittle strain. Geowissenshaften (15): 124–130.
10 000
N ≥ L0
1000 H : λ = 0.13
100 10 1 0
D : λ = 0.2
10
20
30 40 L0 (km)
50
60
70
Figure 34 Length distribution of faults on the flanks of the East Pacific Rise, from sonar swath map data. From Cowie PA, Scholz CH, Edwards M, and Malinverno A (1993) Fault strain and seismic coupling on midocean ridges. Journal of Geophysical Research, Solid Earth 98: 17911–17920.
of 1.2 obtained in Figure 32, illustrating the overestimation problem with cumulative distributions.
6.10.5.1.2
Exponential distributions Cowie et al. (1993) were the first to discover that exponential fault distributions exist. Their data came from sonar swath maps of the normal faults on the flanks of the East Pacific Rise and are shown in Figure 34 as semi-log plots. The data clearly do not fit a power law, instead they are fit by N ðLÞ ¼ NT expð– LÞ
½19
Periodic distributions It has long been known that joints in layered sedimentary rock are often evenly spaced with a spacing proportional to the bed thickness (Price, 1966, p. 144). Periodic distributions like this are seldom seen for faults. One example, shown in Figure 30, are the subparallel strike-slip faults of the San Andreas system in southern California. There the faults have a fairly even spacing of 40–50 km. Another example are the 100 km long normal faults that form the evenly spaced basins and ranges in Nevada. Hu and Evans (1989) experimentally studied the cracking of adhesive brittle films on ductile substrates subject to tension. They found that the cracks reached a saturated population of evenly spaced system sized cracks. The spacing of the cracks is proportional to the thickness of the brittle layer. They showed that the spacing is such that the cracks occur just outside on one another’s stress shadow. Their results are consistent with those for the faults in Figure 30, which are spaced about three times the brittle thickness of 15 km. Soliva et al. (2006), in a study of a population of normal faults restricted to a single sedimentary layer, found a periodic distribution with closer spacing than that expected for a layer bounded at one edge by the free surface. As we shall see below, these different populations are transition regimes that occur as brittle strain increases. Brittle strain is the strain due to faulting and is determined by summing the moments of the faults, as will be discussed in Section 6.10.7. Thus the power law distributions shown in Figures 32 and 33, were for regions in which the brittle strain was 1–2%. The data for the East Pacific Rise had strains of 5% for the data at 3 S (lower curve) and 10% at 13–15 N (upper curve).
6.10.5.2 The Formation of Fault Populations Cowie et al. (1995) studied a model simulating a brittle elastic plate undergoing stretching. For large strains this model converges on a single system size crack, but at small strains power law (fractal) size distributions were observed forming in geometrically fractal populations.
Fault Mechanics
Spyropoulos et al. (2002), continuing this approach, studied a numerical model of stretching a brittle layer supported by a ductile substrate. It was a discrete spring-block model, illustrated in Figure 35. The blocks were connected by leaf springs to a lower substrate that acted as a ductile layer which was stretched. Each block could rupture, following a slip weakening fracture criterion. The strength of slip weakening was the only tunable parameter in the model, and over a broad range it produced very similar results. Figure 36 shows the fraction of active sites versus strain normalized to disorder. The fraction of active sites is the sum of crack lengths that are opening over a time window. This parameter first increases rapidly with strain as many new cracks are nucleated. The rate of increase in active sites gradually decreases until it reaches a peak, then the number of active sites gradually declines to an asymptotic level at high strain. As new cracks nucleate and grow, the number of sites in stress shadows increases, so that the probability of nucleation of new cracks decreases and probability of the death of old cracks and of the coalescence of cracks both increase,
leading to the peak and eventual decrease in active sites. Thus the rising part of the curve is dominated by nucleation and growth of cracks, and the falling part is dominated by coalescence and the inactivity of small cracks caught in the stress shadows of longer cracks. Figure 37 shows the evolution of the frequency size distribution of cracks with strain in the rising part of the curve in Figure 36. It begins as exponential, mimicking the noise introduced initially in the 0.7 0.6 Fraction of active sites
468
Coalescence Nucleation
0.5 0.4 Saturation
0.3 0.2 0.1
y x
0
2
4 6 Strain/disorder
8
10
Figure 36 Fraction of active sites vs. strain for the model shown in Figure 35. From Spyropoulos C, Scholz CH, and Shaw BE (2002) Transition regimes for growing crack populations. Physical Review E 65: 056105 (doi: 10.1103/ PhysRevE.65.056105).
ky kx
h
0
103 102
x Block motion
101
Top view
100 10–1
kx
h
x
R
z
10–2 10–3
kz
10–4
x Lower layer motion Side view Figure 35 Schematic diagram of a model of fracture of a brittle layer over a ductile substrate. From Spyropoulos C, Scholz CH, and Shaw BE (2002) Transition regimes for growing crack populations. Physical Review E 65: 056105 (doi: 10.1103/PhysRevE.65.056105).
10–5 10–6 10–2
10–1
100
101
102
103
L
Figure 37 Length distribution of faults from the model at low strains. From Spyropoulos C, Scholz CH, and Shaw BE (2002) Transition regimes for growing crack populations. Physical Review E 65: 056105 (doi: 10.1103/PhysRevE.65.056105).
Fault Mechanics
model, and evolves asymptotically to a power law shown as the heavy line. The dashed line has slope – 2 for comparison. The distribution of sizes for the large strain falling part of the curve is shown in Figure 38. Here, it becomes an exponential distribution for all but the smallest cracks. At saturation, the crack distribution was dominated by evenly spaced system-size cracks. This numerical model was validated by physical models in the form of a clay cake stretched on a rubber substrate by Spyropoulos et al. (1999). Both the peaked form of the crack density versus brittle strain curve and the evolution from power law to exponential size distribution were observed. Ackermann et al. (2001) reproduced these results in similar experiments. Gupta and Scholz (2000b) observed this transition for the Asal rift in Djibouti. Using data from a digital elevation model, they found that the brittle strain increased to a maximum of 12% in the center of the rift. As shown in Figure 39(a), the strain was accommodated at first by increasing fault density, but after 6% strain, the fault density no longer increased. After that point further strain was accommodated by an increasing D/L ratio (Figure 39(b)) indicating that all faults had become pinned at 6% strain. The size distribution was power law for low strains and exponential for high strains (Figure 40). Because the faults in the Asal Rift are highly interacting, their displacement profiles are often highly contorted (Manighetti et al., 2001).
6.10.6 Strain and Faulting 6.10.6.1 Calculation of Brittle Strain from Fault Data The brittle strain due to faulting may be determined by summing the geometric moments of the faults in a population, where geometric moment is defined as M ¼ uA
Mij ¼ M ðui nj þ uj ni Þ
½21
where ui is the unit vector in the direction of displacement and ni is the unit normal to fault plane. The formula (Kostrov, 1974) for calculating the net strains from a volume V of rock that have been active over a geologic period is N 1 X ½Mij k 2V k¼1
"ij ¼
½22
To illustrate this compact tensor equation, consider the simple case of a plate of brittle thickness W and length and width l1 and l2 being extended in the x1 direction by a population of parallel normal faults of dip j (Figure 41). The mean displacement uˆ1 of the right-hand face is uˆ1 ¼
N X uk cos jL2 sin j k
W l2
½23
for small faults that do not rupture the entire thickness and are assumed to be equant of side L. We then have
102 101
"11 ¼
0
10
N uˆ1 cos j sin j X ¼ uk L2k l1 V k¼1
½24
a simple form of eqn [22]. If the faults become large such that they rupture the entire thickness of the brittle layer, they will have a width W ¼ W/sin j and the equivalent expression becomes
10–1 R
½20
where u is the mean fault displacement and A is fault area. The geometric moment is the magnitude of the tensor moment,
k¼1
10–2 10–3 10–4
"11 ¼
10–5 10–6
469
0
20
40
60
80
100
120
140
L Figure 38 Length distribution of faults from the model at high strains. From Spyropoulos C, Scholz CH, and Shaw BE (2002) Transition regimes for growing crack populations. Physical Review E 65: 056105 (doi: 10.1103/ PhysRevE.65.056105).
N cos j X uk Lk A k¼1
½25
where V and A are the volume and area of the deformed region, respectively. Cowie et al. (1993) used this method, with exactly this geometry, to calculate the fraction of seafloor spreading that results from normal faulting. The strain shown in Figure 39 was also calculated in this way. In the former study the data source was from sonar swath
470
Fault Mechanics
(b) Transect AA′ Transect BB′
35 × 10–3
0.8
Displacement/ length
km fault length/sq. km
(a)
0.6 0.4 0.2 0.0
30 25 20 15 10 5
0
2
4 6 8 Percent extension
10
0
12
2
4 6 8 10 Percent extension
12
Figure 39 Fault data from the Asal Rift, Djibouti. (a) Fault density vs. brittle strain. (b) Displacement/length ratio vs. strain. From Gupta A and Scholz CH (2000b) Brittle strain regime transition in the Afar depression: Implications for fault growth and seafloor spreading. Geology 28: 1087–1090.
(a)
(b) Extension < 8% Extension > 8%
100
100
8 6
Cumulative number
Cumulative number
8 6 4 2
10
8 6 4
4 2
10
8 6 4 2
2
1
1 4
5 6 78
2
10 length (km)
3
4 5 6 7
10 20 30 40 50 60 70 length (km)
Figure 40 The length distribution of faults for the Asal Rift, at low and high strains. (a) In log–log coordinates. (b) in semi-log coordinates. From Gupta A and Scholz CH (2000b) Brittle strain regime transition in the Afar depression: Implications for fault growth and seafloor spreading. Geology 28: 1087–1090.
mapping and in the latter case from a digital elevation model. They therefore had quite good information on both u and L. For more incomplete data sets, for example, from geological mapping, several scaling results can be included to make the procedure much less demanding (Scholz and Cowie, 1990). It is usually easier to determine fault length than displacement. Since we have already established the linear scaling u ¼ L, it is not necessary to measure u and L independently for every fault. One needs only to determine the scaling parameter for a few faults in the region and then to apply it throughout. A further simplification results from a property of fault populations. For most continental settings, the strains are low and the fault populations are power law. Integrating
over eqn [18] and using the scaling law we find that for large faults (L W ) N X
Mk ¼ aW C
Z
k¼1
dN ðLÞ 2 aW C 2 – C L dL ¼ L dL 2–C
½26
and for small faults (L W ) N X k¼1
Mk ¼ aW C
Z
dN ðLÞ aW C 3 – C ½27 L3 dL ¼ L dL 3–C
We have previously found (Figures 32 and 33) that the value of C is approximately (or perhaps exactly) 1. Inserting this value into the proper integrals above, we see that they are convergent. Almost all the moment is carried by the longest faults, so we need only evaluate
Fault Mechanics
the integrals at their upper bounds. The remaining moment in the smaller faults can be calculated (Scholz and Cowie, 1990) so we need not be concerned with the lack of resolution at the smaller scales. 6.10.6.2
Fault Rotation and Lockup
As faults accrue displacement, they rotate such that the angle between the fault plane and 1 increases. The most familiar cases are where systems of subparallel faults occur, so-called bookshelf faulting (Savage et al., 2004; Sigmundsson et al., 1995; Tapponnier et al., 1990), but any individual fault will also rotate in this manner. Martel (1999) has x2 I2
w* û1 x1 I1
ϕ
n n
w*
I1 Figure 41 Schematic diagram for calculating fault strain. From Scholz CH (1997) Earthquake and fault populations and the calculation of brittle strain. Geowissenshaften (15): 124–130.
Thrust earthquake dips
(a) 10
Optimal angle μ = 0.6
ð1 – pÞ ð1 þ cot R Þ ¼ ð3 – pÞ ð1 – tan R Þ
Normal fault earthquake dips 6
Lockup angle
μ = 0.6
Lockup dip
μ = 0.6
5
4
½28
where p is pore pressure and is the friction coefficient. This stress ratio has a minimum at the optimum angle for fault reactivation R and goes to infinity at the lockup angle 2R . This is the maximum angle of the fault to 1. This angle may be reached as p approaches 3 but may not be exceeded, because p cannot exceed 3 without hydrofracturing occurring and the consequent draining of the pore fluid. We can evaluate this for dip-slip faults because in that case we can assume one of the principal stresses is vertical and the others horizontal. Figure 42 shows the dip angles of large reverse (a) and normal (b) fault earthquakes in which the fault plane is known unambiguously from aftershock, geodetic, or surface faulting evidence ( Jackson, 1987; Sibson and Xie, 1998). Also shown are the optimal and lockup angles for a typical laboratory friction coefficient of 0.6. The lockup angle predicted by this friction value agrees well with the data. This is strong evidence that this laboratory value is appropriate for faults. The most common angles are somewhat farther from the fault than the optimum angle, as expected from finite rotation. The fact that some faults are active at or near the lockup angle indicates that p does sometimes approach 3. Direct evidence for this comes from the (b)
6
Optimal dip μ = 0.6
4 3 2
2 0
analyzed fault rotation at an infinitesimal scale with an elastic crack model: the basic physics behind it is correct for the finite scale. This rotation means that the fault becomes progressively less favorable oriented for slip. Eventually it will become frictionally locked. The stress condition for reactivating a fault inclined at an angle R to 1 is (Sibson, 1985)
Number
Number
8
471
1 5
15
25
35
45 55 Dip,°
65 75
85
95
0
5
15 25 35 45 55 65 75 85 95 Dip,°
Figure 42 Histogram dip angles of active (a) thrust and (b) normal faulting earthquakes. From Scholz CH (2000) Evidence for a strong San Andreas fault. Geology 28: 163–166.
Fault Mechanics
relationship of veins to high angle reverse faults near the lockup angle (Sibson et al., 1988). Because faults are embedded in an elastic medium, when they rotate we might expect there to be an elastic restoring force that resists further rotation and hence causes the fault to lockup before reaching the lockup angle as calculated above. This problem has been evaluated in the case of crustal scale normal faults (Buck, 1988; Cowie et al., 2005; Scholz and Contreras, 1998). Displacement on such a normal fault results in flexure of the crust, which is responsible for the footwall uplift and hanging wall basins that arise. For further displacement, the shear stress on the fault must exceed both friction and the elastic flexural restoring stress. Thus the lockup in this case will occur before the one solely due to friction given by eqn [28]. These analyses indicate that such faults have a maximum displacement and hence length, which are functions of the seismogenic thickness and the effective elastic thickness of the crust. There of course may be other reasons why faults stop propagating; stress shadowing, as mentioned earlier, may cause this, as well as changes in thermal structure or tectonic loading (Bull et al., 2006; Cowie et al., 2005).
6.10.7 Fault Rocks and Structures
Fault zone structure
(1)
(2)
(3) (4) (3)
(2)
(1)
1) Undeformed host rock Fault zone
2) Damage zone 3) Foliated cataclasite 4) Ultracataclasite layer
Fault-core
Figure 43 Schematic diagram of the structure of the San Gabriel Fault, California at a depth of 3–5 km. From Chester FM, Chester JS, Kirschner DL, Schulz SE, and Evans JP (2004) Structure of large-displacement strike-slip fault zones in the brittle continental crust. In: Karner G, Taylor B, Driscoll N, and Kohlstedt D (eds.) Rheology and Deformation of the Lithosphere at Continental Margins, pp. 223–260. New York: Columbia University Press.
100
Relative intensity
472
40 60 80 20 Microfractures
Rocks found in fault zones are characteristic of them, indicating modes of deformation unique to faults. Snoke et al. (1998) have produced a photographic atlas of such rocks. Fault related rocks are, of course, dependent upon their protolith and they vary with depth. Here we attempt to restrict the discussion to faults in crystalline protoliths. The discussion is divided into sections of increasing depth: shallow schizosphere, deep schizosphere, brittle-plastic transition region, and plastosphere. 6.10.7.1
Shallow Schizosphere
In the upper half of the seismogenic layer, faults are completely brittle structures. A schematic diagram of the components of such a fault is shown in Figure 43. A broad damage zone surrounds a fault core consisting of cataclasites, often foliated, containing narrower bands of ultracataclasites within which may be found discrete sliding surfaces (Chester et al., 1993). The San Gabriel Fault, an ancestral strand of the San Andreas Fault in California, has been studied by Chester et al. (1993) and Chester et al. (2004). This fault has a strike-slip offset of 20 km and is exposed at a depth of 2–5 km. Its damage zone is several hundred meters
0 –2
–1
0 1 log distance (m)
2
3
Figure 44 Microfracture density as a function of distance from the San Gabriel Fault. From Chester FM, Chester JS, Kirschner DL, Schulz SE, and Evans JP (2004) Structure of large-displacement strike-slip fault zones in the brittle continental crust. In: Karner G, Taylor B, Driscoll N, and Kohlstedt D (eds.) Rheology and Deformation of the Lithosphere at Continental Margins, pp. 223–260. New York: Columbia University Press.
wide, the cataclasite zone several meters wide and the ultracataclasite zone width is in the centimeter to decimeter range. Both the cataclasites and ultracataclasites are often foliated, with the foliation near parallel to the fault. Almost all shearing has occurred in the fault core, with it being concentrated in the ultracataclasite layer and along slip surfaces. This narrowness of the zone of principal slip has been emphasized by Sibson (2003), who provided many examples. The density of microcracking and mesoscopic features decreases exponentially with distance from the ultracataclastic layer (Figure 44). This is similar to that observed for process zones (Figure 14), so it is
Fault Mechanics
473
(a) 10 000 Host rock Protocataclasite and cataclasite Ultracataclasite
1000 100 10
(b) 1
10 000
D=2
1000 100 0.1
10
(c)
1
10
10 000 Particle density (N(D)/A)
D=2 1000 100 0.1
10
1
10
1 D=2 0.1 0.01 0.001
1 0.01 0.1 Particle diameter (mm)
10
Figure 45 Particle size distribution for three different distances from the San Gabriel Fault. From Chester FM, Chester JS, Kirschner DL, Schulz SE, and Evans JP (2004) Structure of large-displacement strike-slip fault zones in the brittle continental crust. In: Karner G, Taylor B, Driscoll N, and Kohlstedt D (eds.) Rheology and Deformation of the Lithosphere at Continental Margins, pp. 223–260. New York: Columbia University Press.
likely that the damage zone is the wake of a process zone produced by the fault tip, perhaps augmented by further damage from earthquakes, as suggested by Wilson et al. (2005). There is also an increase in mineralization and volatile content as the ultracataclasite layer is approached (Chester et al., 2004). The grain size distribution of the cataclastic rocks is power law, or fractal, with the fractal dimension increasing with strain (Figure 45, see also Di Toro and Pennacchioni (2005)). This occurs due to the production of fine grains at the expense of coarse grains during comminution, as expected from theory (Sammis et al., 1986) and experiment (Marone and Scholz, 1989). The Punchbowl Fault, another large displacement (40 km) strike-slip forerunner of the San Andreas Fault exposed to a depth of 2–4 km, has been described by Chester and Chester (1998) and Chester et al. (2004). It has two fault cores, about 125 m apart. Deformation increases as each core is
approached, in a similar way as the San Gabriel Fault, but remains high in the sliver between the two cores. The northern of these two cores occurs at the contact between crystalline basement and sandstone. Ultracataclasite derived from each protolith are juxtaposed in the core but not mixed, suggesting that slip occurs on a discrete principal fault surface. There is no evidence, in the form of pseudotachylytes, clastic dikes, or explosion breccia, of seismic rupture for either of these faults. Indeed, the delicacy of the structures suggests that slip, at the depths exposed, may have been aseismic. 6.10.7.2
Deep Schizosphere
The Gole Larghe Fault zone of northern Italy, is exposed at a depth of 9–11 km and an estimated ambient temperature of 250–300 C (Di Toro and Pennacchioni, 2005). This fault, with a net
474
Fault Mechanics
Gole Larghe Fault zone Lobbia Glacier 2800
55
Altitude (m)
0m
2400
2000 0
450
900 Distance (m)
1350
1800
Figure 46 Schematic diagram of the Gole Larghe Fault, Italy at a depth of 9–11 km. Fault parallel lines indicated pseudotachylyte veins. From Di Toro G and Pennacchioni G (2005) Fault plane processes and mesoscopic structure of a strong-type seismogenic fault in tonalites (Adamello batholith, Southern Alps). Tectonophysics 402: 55–80.
displacement of 1100 m, is extremely well preserved on glacially polished surfaces. It does not show the extreme localization of slip associated with the San Gabriel and Punchbowl Faults described above. It appears as a 550 m wide zone in which hundreds of pseudotachylyte veins and cataclasite layers are distributed evenly throughout the host tonalite (Figure 46). Pseudotachylytes are the products of friction melting and hence are evidence of ancient seismic faulting (Sibson, 1975). The pseudotachylytes in the Gole Larghe Fault often reactivated cataclasite layers which had been indurated prior to reactivation. The cataclasites in turn reactivated joints. The pseudotachylyte veins are not reactivated, however, suggesting that each represents an individual earthquake. The veins have thicknesses in the millimeter-centimeter range and are associated with displacements in the decimeter to meter range. If the entire work of faulting, d goes into melting, then their thickness/displacement (t/d) should be proportional to (Di Toro and Pennacchioni, 2005). Their thickness increases with displacement, but like Sibson’s (Sibson, 1975) data, the scaling appears closer to d t2 than linear, implying that decreases with d. Di Toro et al. (2005a, 2005b) attempted to determine this and other earthquake source properties from them, but with limited success because the displacement associated with the cataclasite phase of faulting could not be separated from that in the purely pseudotachylyte phase. Di Toro et al. (2005a, 2005b) were able to infer, however, the direction and velocity of propagation from the orientation of pseudotachylyte injection veins.
Although pseudotachylytes sometimes occur at shallower depth (Sibson, 2003), they are primarily found in the deeper, higher stress part of the schizoshere. Like the case of the Gohle Larghe Fault, deep, pseudotachylyte bearing faults are not localized but are often hundreds of meters to kilometers wide with distributed deformation (Allen, 2005; Grocott, 1981; Swanson, 1988). Di Tora and Pennacchioni (2005) suggest that the delocalization mechanism of such faults is the strong welding by pseudotachylytes that prevents their reactivation. Pachell and Evans (2002) studied a smaller displacement (100 m) fault in the Sierra Nevada of California that was active at the same depth range as the Gole Larghe Fault but with distinctively different properties. It was a brittle feature consisting of fractures spanning a width of meters to tens of meters, with abundant jogs and with little developed cataclasite. They did not report any pseudotachylyte veins. In contrast to the cases discussed above, many crustal scale large displacement strike-slip faults exposed in the middle crust are several kilometers wide, with no single concentrated core in which the bulk of the slip occurs. Examples are the Carboneras Fault in southeastern Spain (Faulkner et al., 2003) and the Damxung-Jiali shear zone in Tibet (Edwards and Ratschbacher, 2005). Both of these shear zones have multiple anastamosing zones of concentrated shear extending over widths of 3–5 km. The Tibetan case is clearly seismogenic: multiple pseudotachylyte veins, 1–10 mm thick, are scattered throughout the depth range 5–15 km. Both these faults cut sedimentary and metasedimentary rocks, often having phyllitic textures, which may have a bearing on their different development. 6.10.7.3
Brittle-Plastic Transition Region
Within the brittle-plastic transition regime, pseudotachylytes can be observed to penetrate into mylonites (Camacho et al., 1995; Lin et al., 2005; Passchier, 1984). Lin et al. (2005) describe pseudotachylytes in the Woodroffe thrust, a 1.5 km thick mylonitized shear zone separating granulite facies from amphibolite facies gneisses. The shear zone, exposed at a depth of 25–30 km, contains large volumes of millimeters to centimeter scale pseudotachylyte veins. They are of two types, cataclasite related, similar to those described above, and mylonite related. The mylonite-related veins penetrate into mylonites and ultramylonites and are themselves
Fault Mechanics
overprinted by subsequent mylonitization, with foliation parallel to that of the mylonites. The earthquakes that produce these pseudotachylyte veins nucleate in the deep schizophere but are able to propagate to some distance into the plastically deforming mylonite zone by virtue of the extremely high strain rates at the tip of the propagating rupture, which temporarily embrittles the rock. Plastic deformation during the subsequent slow shearing of the interseismic period resulted in the mylonitization of the pseudotachylyte veins (e.g., Tse and Rice, 1986). The cataclastic-related veins overprint the myloniterelated ones, and were produced subsequent to the unroofing of the fault through the brittle-plastic transition.
6.10.7.4
Plastosphere Shear Zones
Major faults have been seismically imaged to extend well into the plastosphere as ductile shear zones. Examples include the San Andreas Fault, which has been imaged to the Moho by Henstock et al. (1997) and Parsons (1998) and the Wind River thrust, which has been imaged continuously to 24 km by Smithson et al. (1978). In deeply exposed terrain, these have been found to be mylonite belts and can be found to be traced all the way from greenschist grade all the way to granulite grade metamorphism (Bak et al., 1975a; Bak et al., 1975b; Hanmer et al., 1995). 6.10.7.4.1
Mylonite fabric elements Mylonites are products of plastic deformation and have distinctive fabrics. Figure 47 shows a schematic diagram of a shear zone and its associated fabric elements. If the material within the shear zone deforms cataclastically or plastically, then, regardless
σ
Rock
τ Gouge/mylonite C R1
T R2 S
σ1 σ3
Figure 47 Schematic diagram of fabric elements in a cataclasite or mylonite shear zone. From Scholz CH (2002) The Mechanics of Earthquakes and Faulting, 2nd edn. Cambridge: Cambridge University Press.
475
of the magnitude of the external stresses, the internal stresses will always be pure shear, as shown by Byerlee and Savage (1992). If the material is a cataclasite or mylonite, it will develop a foliation on the plane of maximum shortening (S) which, with finite strain, will rotate to within parallelism with the shear zone. In the cataclastic regime, Riedel shears develop in the R1 and R2 direction. In both regimes, zones of concentrated shear may develop in the C orientation (Chester et al., 1993; Lister and Snoke, 1984; Lister and Williams, 1979; Simpson and Schmid, 1983) and in the plastic regime, a foliation or cleavage, called C9, may also develop in the R1 direction. Mylonites have a single foliation with a lineation on the foliation plane. They are the product of shear under semi-brittle to plastic condition. They may form from any protolith, but here we will restrict the discussion to those formed from grantite rock. The foliation is initially on the plane of maximum shortening but with finite strain rotates toward the plane of shear, of the shear zone itself (Figure 48). The liniation is in the direction of shear. S–C mylonites have concentrated shear on C surfaces, which are hence shear discontinuities (Figure 48(b)). 6.10.7.4.2 Shear localization and strain softening in mylonite zones
The localization of shear in mylonite zones indicates that the mylonites must be undergoing some kind of strain softening. Although there has been some experimental work on the formation of mylonites in analog materials (Kawamoto and Shimamoto, 1998), there has been little discussion of the mechanism for this softening in the literature. We therefore take this opportunity to speculate on it, in the particular case of granitoid mylonite zones. Consider the case of a shear zone in the greenschist metamorphic facies, that is, in the temperature range of approximately 300–450 C. Under those conditions, quartz can flow plastically and feldspar cannot. Thus the material we are trying to shear is composed of mainly (say 70–80%) of feldspar, which behaves as a hard strong brittle material, with the rest consisting of quartz and some other ductile minerals like mica, which behave as ductile and relatively soft components. The question is, what controls the strength, the strong component or the weak component? This depends on the rock fabric: the way the different components are arranged (Jordan, 1988). The starting granite has an igneous fabric, the quartz grains are arranged tightly and randomly in the interstices
476
Fault Mechanics
s
c
Distance across zone
(a)
(b)
γ
(c)
γ
γ
Figure 48 Rotation of fabric elements and shear strain in three types of shear zones. (a) mylonite zone, (b) S-C mylonite, (c) cataclasite zone. From Scholz CH (2002) The Mechanics of Earthquakes and Faulting, 2nd edn. Cambridge: Cambridge University Press.
between the more frequent feldspar grains. In this state, the feldspar determines the strength of the rock, just as the sand and gravel determines the strength of concrete, rather than the soft mortar, which only serves to hold the aggregate together. All the forces in the material are being transmitted between the contacts of the hard grains. Once shearing begins, the mylonitic fabric begins to be developed, which changes the ordering of the mineral phases. The ductile minerals, quartz and mica, form undulous sheets between layers of feldsar (such sheets being called ‘quartz ribbons’ because of their appearence in thin section). Almost all subsequent deformation of the rock takes place on these quartz sheets: after an intitial period of cataclastis, the feldspars, often forming augen, act as rigid particles. Thus the strength of the rock will decrease progressively as the mylonitic fabric develops and rotates toward the plane of shear, on which it will have the greatest resolved shear stress. So, although this softening occurs with increase in shear strain, it should be more properly called fabric softening rather than strain softening. It has also been observed, however, that mylonite zones, like brittle fault zones, widen with net shear across them (Hull, 1988). This is a form of strain delocalization, which implies some sort of hardening mechanism. The shear zone is widening by adding new mylonite from its boundaries as net shear increases. This presumable means the high strain mylonites in the center of the shear zone have hardened and can no longer provide sufficient strain (Means, 1984). This may be strain hardening of the usual type, produced by high dislocation density.
In the few cases where one can trace a ductile shear zone with depth, it is usually found that it broadens with depth (Bak et al., 1975a, 1975b). The Nagssugtoqidian shear zone in Greenland, for example, reported to have a strike-slip shear offset in the 100 km range, is almost 25 km wide in the granulite facies depth range (Bak et al., 1975a, 1975b). This broadening may result from the effect of increasing temperature in weakening and making the mylonites less fabric softening.
6.10.7.5 Synoptic Model for Faults and Shear Zones A model that illustrates many properties of fault zones is shown in Figure 49 (Scholz, 1988). The temperature-depth scale is based on the geotherm measured for the San Andreas Fault, and hence may differ in other regions. A granitic protolith is assumed. The left panel shows the expected fault rocks, ranging from cataclasites in the upper, seismically active region, to mylonites below the onset of quartz plasticity at 11 km. Pseudotachylytes occur in the lower part of the former region and the upper part of the latter. Although mylonites are formed within the Greenschist regime, the rock there flows in a semibrittle manner: full plasticity does not occur until feldspar begins flowing in the upper part of amphibolite metamorphism (Simpson, 1985). Thus the strength profile on the right panel does not show the typical parrot beak shape obtained by extrapolating laboratory plasticity data from high temperature (e.g., Brace and Kohlstedt, 1980). The inner panel indicates the
Fault Mechanics
Geologic features
Friction rate behavior (A–B)
Fault rock mechanisms
(–)
450°
22 kmT2
(+)
Unstable zone
Abrasive wear Alternat ing zone
Long term
Nucleation zone (seismogenic layer) Dynamic
Typical hypocenter of large earthquake
Lower stability transition Maximum rupture depth (large earthquake)
Plastic flow
T3 – Greenschist
Onset of quartz plasticity
Strength
Upper stability transition
Adhesive wear
300° C 11 kmT1
Cataclastites
Temperature Depth (S. A. geotherm)
Pseudotachylyte
Clay gouge
Mylonites
T4
Seismic behavior
477
? ?
Feldspar plasticity
?
Amphibolite
Figure 49 Synoptic fault model. See text for discussion. From Scholz CH (1988) The brittle-plastic transition and the depth of seismic faulting. Geologische Rundschau 77: 319–328.
0
0
50
(σ1 – σ3) (MPa) 150 100
200
250
Hydraulic fracturing
2 Depth (km)
seismic behavior interpreted from rate/state variable friction laws. Earthquakes can only nucleate within the unstable region where the friction parameter A-B is negative. They can, however, dynamically rupture shallower or deeper than that. Thus there is a region of alternating plasticity and seismic slip just below the nucleation zone. For more discussion see Scholz (2002).
4 6
Combined analysis
Drilling induced fractures
8
μ = 0.65
6.10.8 The Strength of Faults 10
Faults are, of course, planes of weakness in the crust. The strength of a fault, f ¼ (n p) may be calculated as a function of depth provided that we know the correct value of friction coefficient and pore pressure p. In this section we will examine evidence pertaining to these parameters. 6.10.8.1
Figure 50 Stress as a function of depth in the KTB borehole in Germany. Line is the predicted stress for a friction of ¼ 0.65. Modified from Brudy M, Zoback MD, Fuchs K, Rummel F, and Baumgartner J (1997) Estimation of the complete stress tensor to 8 km depth in the KTB scientific drill holes: Implications for crustal strength. Journal of Geophysical Research, Solid Earth 102: 18453–18475.
Direct Evidence for Fault Strength
The only direct evidence for the absolute value of fault strength comes from stress measurements made in deep (>1 km) boreholes in the crystalline basement. Results from the deepest of these, the KTB borehole in Germany (Brudy et al., 1997), are shown in Figure 50. They indicate that shear stress increases with depth consistent with a friction coefficient of ¼ 0.65, and that pore pressure is hydrostatic at all depths. This same result has universally been found in all deep borehole measurements, which have been made in a variety of tectonic settings: extensional,
compressional, transcurrent, plate boundary, and intraplate (Lund and Zoback, 1999; McGarr and Gay, 1978; McGarr et al., 1982; Shamir and Zoback, 1992; Townend and Zoback, 2000; Townend and Zoback, 2001; Zoback and Harjes, 1997; Zoback and Healy, 1984; Zoback and Healy, 1992; Zoback et al., 1980). The straightforward interpretation of these results is that stresses in the continental crust are limited by the ubiquitous presence of favorably oriented faults with friction 0.6 0.7. This hypothesis was verified by a fluid injection test near the bottom of the
478
Fault Mechanics
ψ (degrees)
KTB hole which produced microearthquake activity with induced pore pressures of only 1% of the ambient stresses (Zoback and Harjes, 1997). This value of agrees with laboratory measurements (Byerlee, 1978). It is quite expected that the friction coefficient should be the same on the field and laboratory scale. Friction is scale independent (Amonton’s 1st law) and it is independent of lithology (Byerlee, 1978) and temperature (Blanpied et al., 1995), up to the onset of plasticity. That pore pressure is hydrostatic in the crystalline rock of the crust is a consequence of the high permeability of the crust resulting from the presence of faults (Townend and Zoback, 2000). This simple result, that the strength of the crust is determined by Byerlee’s law with hydrostatic pore pressures, thus comes as no surprise (except to seismologists, who from the time of Tsuboi (1933) had believed that the strength of faults was the stress drop of earthquakes, 10 MPa). This result is consistent with the orientation of active faults with respect to 1, as shown in Figure 42. Information regarding the strength of faults relative to the surrounding crust can be obtained from stress orientation data. If a fault is very weak with respect to the surrounding crust, then it will act as a principal plane and then 1 must rotate, within several fault depths, to near perpendicularity with it. On the other hand, if the fault strength is comparable with the surrounding crust, then 1 must rotate in the opposite manner and approach the fault at an angle less than the lockup angle, which for ¼ 0.6 is 60 . Hardebeck and Hauksson (1999) did this exercise for all of southern California. They found that the strong fault sense of stress rotation prevailed at all crossing of the faults of the San Andreas system. An example is shown in Figure 51. Unfortunately they misinterpreted their data. The correct interpretation was given by Scholz (2000). Stress measurements from paleopiezometers such as dislocation density indicate that shear stresses of SW Fort Tejon profile 120 100 80 60 40 20 –120 –100 –80 –60 –40 –20 0 20 40 Distance from San Andreas Fault (km)
NE
60
80
Figure 51 Angle between 1 and the San Andreas Fault in a profile across the fault in the vicinity of Fort Tejon. From Scholz CH (2000) Evidence for a strong San Andreas fault. Geology 28: 163–166.
order 100 MPa also are typical of ductile shear zones beneath the seismogenic zone (Kohlstedt and Weathers, 1980). At the Stack of Glencoul on the Moine thrust, the differential stress remains constant with distance for 100 m from the fault even though strain increases greatly as the center of the shear zone is approached (Weathers et al., 1979). Unfortunately, for the past 35 years, a pall of controversy has been cast over this topic, which has been more misleading than illuminating. 6.10.8.2 Fallacy
The Weak San Andreas Fault
Brune et al. (1969) measured heat flow in a profile across the San Andreas Fault in the Mojave sector and no sign of a heat flow anomaly centered on the fault of the type expected if heat transport was by conduction. They concluded that the shear stress on the fault could not be greater than that necessary to account for the stress drop in earthquakes, 20 MPa, in accord with the traditional seismological belief. Explaining this so-called San Andreas heat flow paradox became a major research program in the US It has been, however, a singularly unsuccessful one: no explanation for a weak San Andreas Fault has been proposed that is not either internally contradictory or inconsistent with observations. Proponents of the weak San Andreas Fault hypothesis brush aside the borehole stress measurements discussed above with the argument that it is only well developed faults like the San Andreas Fault that are weak, without defining the term ‘well developed’ in a manner that can be tested. The stress orientation data mentioned above (Scholz, 2000) is dismissed with the argument that the lack of a heat flow anomaly is definitive (Zoback, 2000). To be sure, if there is friction, there must be heat generated. But to say that there is no heat flow anomaly associated with the San Andreas Fault is incorrect. There is in fact a pronounced heat flow anomaly associated with the San Andreas Fault, shown in Figures 52 and 53 (Lachenbruch and Sass, 1980). Figure 52 shows the data projected onto a profile across the fault. A conductive anomaly is drawn for comparison: this certainly does not match the data, which shows a much broader anomaly. Figure 53 shows the data projected along strike. Its most striking feature is its smooth decrease to zero as the Mendocino triple junction is approached. The San Andreas Fault has been propagating to the north for the past 30 my (Atwater, 1970), and its
Fault Mechanics
Pacific Plate
North American Plate
125
2.0
84
1.0
42
0
100 80 60 40 20
0
Heat flow (mWm–2)
Heat flow (HFU)
3.0
Mojave region
125
2.0
84
1.0
42
–2 Heat flow (mW m )
Heat flow (HFU)
California Coast Ranges
seconds every 150 years or so, a thermal flux 109 times higher. It is hard to imagine that this heat would not be advected by transport of water away form the fault at depth, where the stresses are highest. The reader should be aware that the arguments expressed in this section represent the author’s opinions and they are distinctively in the minority at present. For a more detailed discussion of this entire issue, see Scholz and Hanks (2004) and Scholz (2006).
20 40 60 80 100
Figure 52 Heat flow data projected onto a profile perpendicular from the San Andreas Fault. They are compared with the prediction of models of conductive heating from fault friction. Modified from Lachenbruch A and Sass J (1980) Heat flow and energetics of the San Andreas fault zone. Journal of Geophysical Research 85: 6185–6222.
3.0
479
500 0 1000 Distance from the Mendocino Triple Junction (km)
Figure 53 The heat flow data projected along strike of the San Andreas Fault. Modified from Lachenbruch A and Sass J (1980) Heat flow and energetics of the San Andreas fault zone. Journal of Geophysical Research 85: 6185–6222.
northern tip has just reached Cape Mendocino, where it has zero age and displacement. If frictional work is what generates the heat in the anomaly in Figure 53, then its decrease as Cape Mendocino is approached is just what is expected. The other mechanisms proposed to explained this (Lachenbruch and Sass, 1980) have failed the test of time (Scholz and Hanks, 2004). The fallacy in the weak San Andreas Fault hypothesis is the assumption that conduction is the mechanism of heat transport. There is no known crustal scale conductive heat flow anomaly. The crust is simply too permeable to support one. The thermal generating power of the fault is v, the shear stress times the sliding velocity. The conduction models such as Brune et al. (1969) are steady-state models that assume that v is 34 mm yr1, the geologic slip rate of the San Andreas Fault. However, the San Andreas Fault does not slip this way but at v 1m s1 for several
References Abercrombie RE, Webb TH, Robinson R, McGinty PJ, Mori JJ, and Beavan RJ (2000) The enigma of the Arthur’s Pass, New Zealand, earthquake 1. Reconciling a variety of data for an unusual earthquake sequence. Journal of Geophysical Research, Solid Earth 105: 16119–16137. Ackermann RV and Schlische RW (1997) Anticlustering of small normal faults around larger faults. Geology 25: 1127–1130. Ackermann RV, Schlische RW, and Withjack MO (2001) The geometric and statistical evolution of normal fault systems: an experimental study of the effects of mechanical layer thickness on scaling laws. Journal of Structural Geology 23: 1803–1819. Adams FD (1938) The Growth and Development of the Geological Sciences. Baltimore, MD: Williams & Wilkins. Al-zoubi A and ten Brink U (2002) Lower crustal flow and the role of shear in basin subsidence: An example from the Dead Sea basin. Earth and Planetary Science Letters 199: 67–79. Allen JL (2005) A multi-kilometer pseudotachylyte system as an exhumed record of earthquake rupture geometry at hypocentral depths (Colorado, USA). Tectonophysics 402: 37–54. Anders MH and Wiltschko DV (1994) Microfracturing, paleostress and the growth of faults. Journal of Structural Geology 16: 795–815. Anderson EM (1905) The dynamics of faulting. Transactions of the Edinburgh Geological Society 8: 340–387. Anderson EM (1936) The dynamics of the formation of conesheets, ring-dykes, and cauldron-subsidences. Proceedings of the Royal Society of Edinburgh 56: 125–128. Anderson EM (1951) The Dynamics of Faulting, 2nd edn. Edinburgh, UK: Oliver and Boyd. Atkinson BK (ed.) (1987) Introduction to fracture mechanics and its geophysical applications. In: Fracture Mechanics of Rock, pp 1–26. London: Academic Press. Atwater T (1970) Implications of plate tectonics for the Cenozoic tectonic evolution of western North America. Geological Society of America, Bulletin 81: 3513–3536. Aydin A and Johnson AM (1978) Development of faults as zones of deformation bands and as slip surfaces in sandstone. Advances in Applied Mechanics 116: 922–929. Aydin A and Nur A (1982) Evolution of pull-apart basins and their scale independence. Tectonics 1: 91–105. Bak J, Korstgard J, and Sorensen K (1975a) Major shear zone within Nagssugtoqidian of West Greenland. Tectonophysics 27: 191–209. Bak J, Sorensen K, Grocott J, Korstgard JA, Nash D, and Watterson J (1975b) Tectonic implications of precambrian shear belts in Western Greenland. Nature 254: 566–569. Barenblatt GI (1962) The mathematical theory of equilibrium cracks in brittle fracture. Advances in Applied Mechanics 7: 55–80.
480
Fault Mechanics
Blanpied ML, Lockner DA, and Byerlee JD (1995) Frictional slip of granite at hydrothermal conditions. Journal of Geophysical Research,Solid Earth 100: 13045–13064. Bourne SJ, Arnadottir T, Beavan J, et al. (1998a) Crustal deformation of the Marlborough fault zone in the South Island of New Zealand: Geodetic constraints over the interval 1982–1994. Journal of Geophysical Research,Solid Earth 103: 30147–30165. Bourne SJ, England PC, and Parsons B (1998b) The motion of crustal blocks driven by flow of the lower lithosphere and implications for slip rates of continental strike-slip faults. Nature 391: 655–659. Brace WF and Kohlstedt D (1980) Limits on lithospheric stress imposed by laboratory experiments. Journal of Geophysical Research 85: 6248–6252. Brudy M, Zoback MD, Fuchs K, Rummel F, and Baumgartner J (1997) Estimation of the complete stress tensor to 8 km depth in the KTB scientific drill holes: Implications for crustal strength. Journal of Geophysical Research,Solid Earth 102: 18453–18475. Brune J, Henyey T, and Roy R (1969) Heat flow, stress, and rate of slip along the San Andreas fault, California. Journal of Geophysical Research 74: 3821–3827. Buck WR (1988) Flexural rotation of normal faults. Tectonics 7: 959–973. Bull JM, Barnes PM, Lamarche G, et al. (2006) High-resolution record of displacement accumulation on an active normal fault: Implications for models of slip accumulation during repeated earthquakes. Journal of Structural Geology 28: 1146–1166. Burgmann R and Pollard DD (1994) Strain accommodation about strike-slip-fault discontinuities in granitic rock under brittle-to-ductile conditions. Journal of Structural Geology 16: 1655–1674. Burgmann R, Pollard DD, and Martel SJ (1994) Slip distributions on faults – Effects of stress gradients, inelastic deformation, heterogeneous host-rock stiffness, and fault interaction. Journal of Structural Geology 16: 1675–1690. Burroughs SM and Tebbens SF (2001) Upper-truncated power laws in natural systems. Pure and Applied Geophysics 158: 741–757. Byerlee JD (1978) Friction of rocks. Pure and Applied Geophysics 116: 615–626. Byerlee JD and Savage JC (1992) Coulomb plasticity within the fault zone. Geophysical Research Letters 19: 2341–2344. Camacho A, Vernon RH, and Fitz Gerald JD (1995) Large volumes of anhydrous pseudotachylyte in the Woodroffe Thrust, Eastern Musgrave Ranges, Australia. Journal of Structural Geology 17: 371–383. Campagna DJ and Aydin A (1991) Tertiary uplift and shortening in the basin and range – The Echo Hills, Southeastern Nevada. Geology 19: 485–488. Cartwright JA and Mansfield CS (1998) Lateral displacement variation and lateral tip geometry of normal faults in the Canyonlands National Park, Utah. Journal of Structural Geology 20: 3–19. Chernyshev SN and Dearman WR (1991) Rock Fractures. London: Butterworth-Heinemann. Chester FM and Chester JS (1998) Ultracataclasite structure and friction processes of the Punchbowl fault, San Andreas system, California. Tectonophysics 295: 199–221. Chester FM, Chester JS, Kirschner DL, Schulz SE, and Evans JP (2004) Structure of large-displacement strikeslip fault zones in the brittle continental crust. In: Karner G, Taylor B, Driscoll N, and Kohlstedt D (eds.) Rheology and Deformation of the Lithosphere at Continental Margins, pp. 223–260. New York: Columbia University Press.
Chester FM, Evans JP, and Biegel RL (1993) Internal structure and weakening mechanisms of the San Andreas fault. Journal of Geophysical Research,Solid Earth 98: 771–786. Chester FM and Logan JM (1986) Implications for mechanical properties of brittle faults from observations of the Punchbowl fault zone, California. Pure and Applied Geophysics 124: 79–106. Chester JS, Chester FM, and Kronenberg AK (2005) Fracture surface energy of the Punchbowl fault, San Andreas system. Nature 437: 133–136. Cladouhos TT and Marrett R (1996) Are fault growth and linkage models consistent with power-law distributions of fault lengths?. Journal of Structural Geology 18: 281–293. Contreras J, Anders MH, and Scholz CH (2000) Growth of a normal fault system: Observations from the Lake Malawi basin of the east African rift. Journal of Structural Geology 22: 159–168. Cowie PA and Scholz CH (1992a) Physical explanation for the displacement-length relationship of faults using a post-yeild fracture mechanics model. Journal of Structural Geology 14: 1133–1148. Cowie PA and Scholz CH (1992b) Displacement-length scaling relationship for faults: Data synthesis and discussion. Journal of Structural Geology 14: 1149–1156. Cowie PA, Knipe RJ, and Main IG (1996) Special issue: Scaling laws for fault and fracture populations – Analyses and applications – Introduction. Journal of Structural Geology 18: R5–R11. Cowie PA, Scholz CH, Edwards M, and Malinverno A (1993) Fault strain and seismic coupling on midocean ridges. Journal of Geophysical Research,Solid Earth 98: 17911–17920. Cowie PA and Shipton ZK (1998) Fault tip displacement gradients and process zone dimensions. Journal of Structural Geology 20: 983–997. Cowie PA, Sornette D, and Vanneste C (1995) Multifractal scaling properties of a growing fault population. Geophysical Journal International 122: 457–469. Cowie PA, Underhill JR, Behn MD, Lin J, and Gill CE (2005) Spatiotemporal evolution of strain accumulation derived from multiscale observations of Late Jurassic rifting in the northern North Sea: A critical test of models for lithospheric extension. Earth and Planetary Science Letters 234: 401–419. Crider JG and Pollard DD (1998) Fault linkage: Threedimensional mechanical interaction between echelon normal faults. Journal of Geophysical Research,Solid Earth 103: 24373–24391. Das S and Scholz CH (1981) Off-fault aftershock clusters caused by shear stress increase? Bulletin of the Seismological Society of America 71: 1669–1675. Dawers NH and Anders MH (1995) Displacement-length scaling and fault linkage. Journal of Structural Geology 17: 607–614. Dawers NH, Anders MH, and Scholz CH (1993) Growth of normal faults – Displacement-length scaling. Geology 21: 1107–1110. Di Toro G, Nielsen S, and Pennacchioni G (2005a) Earthquake rupture dynamics frozen in exhumed ancient faults. Nature 436: 1009–1012. Di Toro G and Pennacchioni G (2005) Fault plane processes and mesoscopic structure of a strong-type seismogenic fault in tonalites (Adamello batholith, Southern Alps). Tectonophysics 402: 55–80. Di Toro G, Pennacchioni G, and Teza G (2005b) Can pseudotachylytes be used to infer earthquake source parameters? An example of limitations in the study of exhumed faults. Tectonophysics 402: 3–20. Dooley T and Mcclay K (1997) Analog modeling of pull-apart basins. American Association of Petroleum Geologists. Bulletin 81: 1804–1826. Dugdale DSJ (1960) Yielding of steel sheets containing slits. Journal of the Mechanics and Physics of Solids 8: 100–115.
Fault Mechanics Edwards MA and Ratschbacher L (2005) Seismic and aseismic weakening effects in transtension: Field and microstructural observations on the mechanics and architecture of a large fault zone in SE Tibet. In: Bruhn D and Burlini L (eds.) HighStrain Zones: Structure and Physical Properties, pp. 109–141. London: The Geological Society of London. Evans JP (1990) Thickness displacement relationships for fault zones. Journal of Structural Geology 12: 1061–1065. Faulkner DR, Lewis AC, and Rutter EH (2003) On the internal structure and mechanics of large strike-slip fault zones: Field observations of the Carboneras Fault in southeastern Spain. Tectonophysics 367: 235–251. Grocott J (1981) Fracture geometry of pseudotachylyte generation zones: A study of shear fractures formed during seismic events. Journal of Structural Geology 3: 169–178. Gupta A and Scholz CH (2000a) A model of normal fault interaction based on observations and theory. Journal of Structural Geology 22: 865–879. Gupta A and Scholz CH (2000b) Brittle strain regime transition in the Afar depression: Implications for fault growth and seafloor spreading. Geology 28: 1087–1090. Hanmer S, Williams M, and Kopf C (1995) Modest movements, spectacular fabrics in an intracontinental deep-crustal strikeslip-fault – Striding-Athabasca Mylonite Zone, NW Canadian Shield. Journal of Structural Geology 17: 493–507. Hardebeck JL and Hauksson E (1999) Role of fluids in faulting inferred from stress field signatures. Science 285: 236–239. Henstock TJ, Levander A, and Hole JA (1997) Deformation in the lower crust of the San Andreas fault system in northern California. Science 278: 650–653. Hu MS and Evans AG (1989) The Cracking and decohesion of thin-films on ductile substrates. Acta Metallurgica 37: 917–925. Hubbert MK and Rubey WW (1959) Role of fluid pressure in the mechanics of overthrust faulting. Bulletin of the Geological Society of America 70: 115–166. Hull J (1988) Thickness-displacement relationships for deformation zones. Journal of Structural Geology 10: 431–435. Jackson JA (1987) Active normal faulting and crustal extension. In: Coward M, Dewey J, and Hancock P (eds.) Continental Extensional Tectonics, pp. 3–18. London: Blackwell. Jordan P (1988) The rheology of polymineralic rocks – An approach. Geologische Rundschau 77: 285–294. Kanninen MF and Popelar CH (1985) Advanced Fracture Mechanics. Oxford: Oxford University Press. Kawamoto E and Shimamoto T (1998) The strength profile for bimineralic shear zones: An insight from high-temperature shearing experiments on calcite–halite mixtures. Tectonophysics 295: 1–14. Kendrick KJ, Morton DM, Wells SG, and Simpson RW (2002) Spatial and temporal deformation along the northern San Jacinto fault, southern California: Implications for slip rates. Bulletin of the Seismological Society of America 92: 2782–2802. Kohlstedt DL and Weathers MS (1980) Deformation-induced microstructures, paleopiezometers, and differential stresses in deeply eroded fault zones. Journal of Geophysical Research 85: 6269–6285. Kostrov B (1974) Seismic moment and energy of earthquakes, and seismic flow of rock. Izvestiya, Academy of Sciences, USSR Physics, Solid Earth 1: 23–40. Lachenbruch A and Sass J (1980) Heat flow and energetics of the San Andreas fault zone. Journal of Geophysical Research 85: 6185–6222. Lawn BR and Wilshaw TR (1975) Fracture of Brittle Solids. Cambridge: Cambridge University Press. Lin AM, Maruyama T, Aaron S, Michibayashi K, Camacho A, and Kano KI (2005) Propagation of seismic slip from brittle to
481
ductile crust: Evidence from pseudotachylyte of the Woodroffe thrust, central Australia. Tectonophysics 402: 21–35. Lister G and Snoke A (1984) S-C mylonites. Journal of Structural Geology 6: 617–638. Lister GS and Williams PF (1979) Fabric development in shear zones: Theoretical controls and observed phenomena. Journal of Structural Geology 1: 283–297. Lockner DA, Byerlee JD, Kuksenko V, Ponomarev A, and Sidorin A (1991) Quasi-static fault growth and shear fracture energy in granite. Nature 350: 39–42. Lockner DA, Byerlee JD, Kuksenko V, Ponomarev A, and Sidorin A (1992) Observations of quasistatic fault growth from acoustic emissions. In: Evans B and Wong T-F (eds.) Fault Mechanics and Transport Properties of Rock, pp. 3–32. San Diego, CA: Academic Press. Lund B and Zoback MD (1999) Orientation and magnitude of in situ stress to 6.5 km depth in the Baltic Shield. International Journal of Rock Mechanics and Mining Science 36: 169–190. Lyell C (1832) Principles of Geology. London: J. Murray. Mair K, Main I, and Elphick S (2000) Sequential growth of deformation bands in the laboratory. Journal of Structural Geology 22: 25–42. Manighetti I, King G, and Sammis CG (2004) The role of off-fault damage in the evolution of normal faults. Earth and Planetary Science Letters 217: 399–408. Manighetti I, King GCP, Gaudemer Y, Scholz CH, and Doubre C (2001) Slip accumulation and lateral propagation of active normal faults in Afar. Journal of Geophysical Research, Solid Earth 106: 13667–13696. Marone C and Scholz CH (1989) Particle-size distribution and microstructures within simulated fault gouge. Journal of Structural Geology 11: 799–814. Marrett R and Allmendinger RW (1991) Estimates of strain due to brittle faulting – Sampling of fault populations. Journal of Structural Geology 13: 735–738. Martel SJ (1999) Mechanical controls on fault geometry. Journal of Structural Geology 21: 585–596. Martel SJ and Pollard DD (1989) Mechanics of slip and fracture along small faults and simple strike-slip-fault zones in granitic rock. Journal of Geophysical Research 94: 9417–9428. Martel SJ, Pollard DD, and Segall P (1988) Development of simple strike-slip-fault zones, Mount Abbot quadrangle, Sierra-Nevada, California. Geological Society of America Bulletin 100: 1451–1465. Mcclay K and Bonora M (2001) Analog models of restraining stepovers in strike-slip fault systems. American Association of Petroleum Geologists, Bulletin 85: 233–260. McGarr A and Gay NC (1978) State of stress in the earth’s crust. Annual Review of Earth and Planetary Science 6: 405–436. McGarr A, Zoback MD, and Hanks TC (1982) Implications of an elastic analysis of in situ stress measurements near the San Andreas fault. Journal of Geophysical Research 87: 7797–7806. Means WD (1984) Shear zones of type I and II and their significance for reconstructing rock history. Geological Society of America Abstract with Programs 16: 50. Moore DE and Lockner DA (1995) The role of microcracking in shear-fracture propagation in granite. Journal of Structural Geology 17: 95–114. Nakamura K (1969) Arrangement of parasitic cones as a possible key to a regional stress field. Bulletin of the Volcanology Society of Japan 14: 8–20. Nakamura K, Shimazaki K, and Yonekura N (1984) Subduction, bending, and eduction. present and quatemary tectonics of the northem border of the Philippine Sea plate. Bulletin de la Societe Geologique de France 26: 221–243.
482
Fault Mechanics
Pachell MA and Evans JP (2002) Growth, linkage, and termination processes of a 10-km-long strike-slip fault in jointed granite: The Gemini fault zone, Sierra Nevada, California. Journal of Structural Geology 24: 1903–1924. Parsons T (1998) Seismic-reflection evidence that the Hayward fault extends into the lower crust of the San Francisco Bay Area, California. Bulletin of the Seismological Society of America 88: 1212–1223. Passchier C (1984) The generation of ductile and brittle defommation bands in a low-angle mylonite zone. Journal of Structural Geology 6: 273–281. Peacock DCP (1991) Displacements and segment linkage in strike-slip-fault zones. Journal of Structural Geology 13: 1025–1035. Peacock DCP and Sanderson DJ (1994) Geometry and development of relay ramps in normal-fault systems. American Association of Petroleum Geologists Bulletin 78: 147–165. Pollard DD and Segall P (1987) Theoretical displacements and stresses near fractures in rock: With applications to faults, joints, dikes, and solution surfaces. In: Atkinson BK (ed.) Fracture Mechanics of Rock, pp. 277–348. London: Academic Press. Price NJ (1966) Fault and Joint Development in Brittle and SemiBrittle Rock. Oxford: Pergamon. Sammis CG, Osborne R, Anderson J, Banerdt M, and White P (1986) Self-similar cataclasis in the formation of fault gouge. Pure and Applied Geophysics 124: 53–78. Savage JC, Svarc JL, and Prescott WH (2004) Interseismic strain and rotation rates in the northeast Mojave domain, eastern California. Journal of Geophysical Research, Solid Earth 109, doi: 10.1029/2003JB002705. Schlische RW, Young SS, Ackermann RV, and Gupta A (1996) Geometry and scaling relations of a population of very small rift-related normal faults. Geology 24: 683–686. Scholz CH (1968a) Microfracturing and the inelastic deformation of rock in compression. Journal of Geophysical Research 73: 1417–1432. Scholz CH (1968b) Experimental study of the fracturing process in brittle rock. Journal of Geophysical Research 73: 1447–1454. Scholz CH (1982) Scaling laws for large earthquakes: Consequences for physical models. Bulletin of the Seismological Society of America 72: 1–14. Scholz CH (1987) Wear and gouge formation in brittle faulting. Geology 15: 493–495. Scholz CH (1988) The brittle-plastic transition and the depth of seismic faulting. Geologische Rundschau 77: 319–328. Scholz CH (1997) Earthquake and fault populations and the calculation of brittle strain. Geowissenshaften 15: 124–130. Scholz CH (2000) Evidence for a strong San Andreas fault. Geology 28: 163–166. Scholz CH (2002) The Mechanics of Earthquakes and Faulting, 2nd edn. Cambridge: Cambridge University Press. Scholz CH (2006) The strength of the San Andreas Fault: A critical analysis. In: Abercrombie R, McGarr A, Kanamori H, and Toro GD (eds.) Radiated Energy and the Physics of Earthquake Faulting. Wahington, DC: American Geophysical Union. Scholz CH and Contreras JC (1998) Mechanics of continental rift architecture. Geology 26: 967–970. Scholz CH and Cowie PA (1990) Determination of total strain from faulting using slip measurements. Nature 346: 837–839. Scholz CH, Dawers NH, Yu JZ, and Anders MH (1993) Fault growth and fault scaling laws – Preliminary-results. Journal of Geophysical Research, Solid Earth 98: 21951–21961. Scholz CH and Hanks TC (2004) The strength of the San Andreas fault: A discussion. In: Karner GD, Taylor B, Driscoll NW, and Kohlstedt DL (eds.) Rheology and
Deformation of the Lithosphere at Continental Margins, pp. 261–283. New York: Columbia University Press. Scholz CH and Lawler TM (2004) Slip tapers at the tips of faults and earthquake ruptures. Geophysical Research Letters 31: L21609. Shamir G and Zoback MD (1992) Stress orientation profile to 3.5 km depth near the San Andreas fault at Cajon Pass, California. Journal of Geophysical Research 97: 5059–5080. Shipton ZK and Cowie PA (2001) Damage zone and slip-surface evolution over mu m to km scales in high-porosity Navajo sandstone, Utah. Journal of Structural Geology 23: 1825–1844. Sibson RH (1975) Generation of pseudotachylyte by ancient seismic faulting. Geophysical Journal of the Royal Astronomical Society 43: 775–794. Sibson RH (1985) A Note on fault reactivation. Journal of Structural Geology 7: 751–754. Sibson RH (2003) Thickness of the seismic slip zone. Bulletin of the Seismological Society of America 93: 1169–1178. Sibson RH, Roberts F, and Poulsen KH (1988) High-angle reverse faults, fluid pressure cycling, and mesothermal goldquartz deposits. Geology 16: 551–555. Sibson RH and Xie GY (1998) Dip range for intracontinental reverse fault ruptures: Truth not stranger than friction? Bulletin of the Seismological Society of America 88: 1014–1022. Sigmundsson F, Einarsson P, Bilham R, and Sturkell E (1995) Rift-transform kinematics in south Iceland – Deformation from global positioning system measurements, 1986 to 1992. Journal of Geophysical Research, Solid Earth 100: 6235–6248. Simpson C (1985) Deformation of granitic rocks across the brittle–ductile transition. Journal of Structural Geology 5: 503–512. Simpson C and Schmid SM (1983) An evaluation of criteria to deduce the sense of movement in sheared rock. Geological Society of America, Bulletin 94: 1281–1288. Smithson SB, Brewer J, Kaufman S, Oliver J, and Hurich C (1978) Nature of wind river thrust, Wyoming, from cocorp deep-reflection data and from gravity data. Geology 6: 648–652. Snoke AW, Tullis J, and Todd VR (1998) Fault-Related Rocks. Princeton, NJ: Princeton University Press. Soliva R, Benedicto A, and Maerten L (2006) Spacing and linkage of confined normal faults: Importance of mechanical thickness. Journal of Geophysical Research, Solid Earth 111: B01402 (doi:10.1029/2004JB003507). Somerville P (1978) The accommodation of plate collision by deformation in the Izu block, Japan. Bulletin of the Earthquake Research Institute, University of Tokyo 53: 629–648. Sowers JM, Unruh JR, Lettis WR, and Rubin TD (1994) Relationship of the Kickapoo Fault to the Johnson Valley and Homestead Valley Faults, San-Bernardino County, California. Bulletin of the Seismological Society of America 84: 528. Spyropoulos C, Griffith WJ, Scholz CH, and Shaw BE (1999) Experimental evidence for different strain regimes of crack populations in a clay model. Geophysical Research Letters 26: 1081–1084. Spyropoulos C, Scholz CH, and Shaw BE (2002) Transition regimes for growing crack populations. Physical Review E 65: 056105 (doi: 10.1103/PhysRevE.65.056105). Stephenson WJ, Odum JK, Williams RA, and Anderson ML (2002) Delineation of faulting and basin geometry along a seismic reflection transect in urbanized San Bernardino Valley, California. Bulletin of the Seismological Society of America 92: 2504–2520. Swanson MT (1988) Pseudotachylyte-bearing strike-slip duplex structures in the Fort Foster Brittle Zone, S Maine. Journal of Structural Geology 10: 813–828.
Fault Mechanics Tapponnier P, Armijo R, Manighetti I, and Courtillot V (1990) Bookshelf faulting and horizontal block rotations between overlapping rifts in Southern Afar. Geophysical Research Letters 17: 1–4. Townend J and Zoback MD (2000) How faulting keeps the crust strong. Geology 28: 399–402. Townend J and Zoback MD (2001) Implications of earthquake focal mechanisms for the frictional strength of the San Andreas fault system. In: Holsworth RE (ed.) Geological Society of London Spec. Publication 186: The Nature and Tectonic Significance of Fault Zone Weakening, pp. 13–21. London: Blackwell. Tse S and Rice J (1986) Crustal earthquake instability in relation to the depth variation of frictional slip properties. Journal of Geophysical Research 91: 9452–9472. Tsuboi C (1933) Investigation of deformation of the crust found by precise geodetic means. Japanese Journal of Astronomy and Geophysics 10: 93–248. Turcotte DL (1992) Fractals and Chaos in Geology and Geophysics. Cambridge: Cambridge University Press. Vermilye JM and Scholz CH (1998) The process zone: A microstructural view of fault growth. Journal of Geophysical Research, Solid Earth 103: 12223–12237. Voight B (1976) Mechanics of Thrust Faults and Decollements. Stroudsburg, PN: Dowden, Huchinson, and Ross. Wakabayashi J, Hengesh JV, and Sawyer TL (2004) Fourdimensional transform fault processes: Progressive evolution of step-overs and bends. Tectonophysics 392: 279–301. Walsh JJ and Watterson J (1988) Analysis of the relationship between displacements and dimensions of faults. Journal of Structural Geology 10: 238–347. Wang WB and Scholz CH (1994) Wear processes during frictional sliding of rock – A theoretical and experimental-study. Journal of Geophysical Research, Solid Earth 99: 6789–6799. Weathers MS, Bird JM, Cooper RF, and Kohlstedt DL (1979) Differential stress determined from deformation-induced microstructures of the Moine Thrust Zone. Journal of Geophysical Research 84: 7495–7509.
483
Willemse EJM (1997) Segmented normal faults: Correspondence between three dimensional mechanical models and field data. Journal of Geophysical Reseach, Solid Earth 102: 675–692. Willemse EJM, Pollard DD, and Aydin A (1996) Threedimensional analyses of slip distributions on normal fault arrays with consequences for fault scaling. Journal of Structural Geology 18: 295–309. Wilson B, Dewers T, Reches Z, and Brune J (2005) Particle size and energetics of gouge from earthquake rupture zones. Nature 434: 749–752. Wood R, Pettinga J, Bannister S, LaMarche G, and McMorran T (1994) Structure of the hanmer strike-slip basin, Hope fault, New Zealand. Geological Society of America, Bulletin 106: 1459–1473. Yeats R and Berryman K (1987) South Island, New Zealand and transverse Ranges, California, a seismotectonic comparison. Tectonics 6: 363–376. Yoshioka N (1986) Fracture energy and the variation of gouge and surface roughness during frictional sliding of rocks. Journal of Physics of the Earth 34: 335–355. Zoback MD (2000) Strength of the San Andreas. Nature 405: 31–32. Zoback MD and Harjes H-P (1997) Injection induced earthquakes and crustal stress at 9 km depth at the KTB deep drilling site, Germany. Journal of Geophysical Research 102: 18477–18491. Zoback MD and Healy JH (1984) Friction, faulting, and in situ stress. Annales Geophysicae 2: 689–698. Zoback MD and Healy JH (1992) In situ stress measurements to 3.5 km depth in the Cajon Pass scientific research borehole: Implications for the mechanics of crustal faulting. Journal of Geophysical Research 97: 5039–5057. Zoback MD, Tsukahara H, and Hickman S (1980) Stress measurements at depth in the vicinity of the San Andreas fault: Implications for the magnitude of shear stress at depth. Journal of Geophysical Research 85: 6157–6173.
6.11 Tectonic Models for the Evolution of Sedimentary Basins S. Cloetingh, Vrije Universiteit, Amsterdam, The Netherlands P. A. Ziegler, University of Basel, Basel, Switzerland ª 2007 Elsevier B.V. All rights reserved.
6.11.1 6.11.2 6.11.2.1 6.11.2.1.1 6.11.2.1.2 6.11.2.1.3 6.11.2.1.4 6.11.2.1.5 6.11.2.1.6 6.11.2.1.7 6.11.2.2 6.11.2.2.1 6.11.2.2.2 6.11.2.2.3 6.11.3 6.11.3.1 6.11.3.2 6.11.4 6.11.4.1 6.11.4.2 6.11.4.3 6.11.4.4 6.11.4.5 6.11.5 6.11.5.1 6.11.5.2 6.11.5.3 6.11.6 6.11.6.1 6.11.6.2 6.11.6.2.1 6.11.6.2.2 6.11.6.3 6.11.6.3.1 6.11.6.4
Introduction Tectonics of Extensional and Compressional Basins: Concepts and Global-Scale Observations Extensional Basin Systems Modes of rifting and extension Thermal thinning and stretching of the lithosphere: concepts and models Syn-rift subsidence and duration of rifting stage Postrift subsidence Finite strength of the lithosphere in extensional basin formation Rift-shoulder development and architecture of basin fill Transformation of an orogen into a cratonic platform: the area of the European Cenozoic Rift System Compressional Basins Systems Development of foreland basins Compressional basins: lateral variations in flexural behaviour and implications for palaeotopography Lithospheric folding: an important mode of intraplate basin formation Rheological Stratification of the Lithosphere and Basin Evolution Lithosphere Strength and Deformation Mode Mechanical Controls on Basin Evolution: Europe’s Continental Lithosphere Northwestern European Margin: Natural Laboratory for Continental Breakup and Rift Basins Extensional Basin Migration: Observations and Thermomechanical Models Fast Rifting and Continental Breakup Thermomechanical Evolution and Tectonic Subsidence During Slow Extension Breakup Processes: Timing, Mantle Plumes, and the Role of Melts Postrift Inversion, Borderland Uplift, and Denudation Black Sea Basin: Compressional Reactivation of an Extensional Basin Rheology and Sedimentary Basin Formation Role of Intraplate Stresses Strength Evolution and Neotectonic Reactivation at the Basin Margins during the Postrift Phase Modes of Basin (De)formation, Lithospheric Strength, and Vertical Motions in the Pannonian–Carpathian Basin System Lithospheric Strength in the Pannonian–Carpathian System Neogene Development and Evolution of the Pannonian Basin Dynamic models of basin formation Stretching models and subsidence analysis Neogene Evolution of the Carpathians System Role of the 3-D distributions of load and lithospheric strength in the Carpathian foredeep Deformation of the Pannonian–Carpathian System
486 490 490 491 493 496 498 504 505 508 519 519 520 523 525 525 529 535 535 540 542 544 544 546 548 550 552 555 559 561 561 563 566 567 575
485
486
Tectonic Models for the Evolution of Sedimentary Basins
6.11.7 6.11.7.1 6.11.7.2 6.11.7.3 6.11.7.4 6.11.7.4.1 6.11.8 References
The Iberia Microcontinent: Compressional Basins within the Africa–Europe Collision Zone Constraints on Vertical Motions Present-Day Stress Regime and Topography Lithospheric Folding and Drainage Pattern Interplay between Tectonics, Climate, and Fluvial Transport during the Cenozoic Evolution of the Ebro Basin (NE Iberia) Ebro Basin evolution: a modeling approach Conclusions and Future Perspectives
6.11.1 Introduction In this chapter we review the formation and evolution of sedimentary basins in their lithospheric context. To this purpose, we follow a natural laboratory approach, selecting some well-documented basins of Europe. We begin with a brief outline of the evolution of tectonic modeling of sedimentary basin systems since its inception in the late 1970s. We subsequently review key features of the tectonics of rifted and compressional basins in Section 6.11.2. These include the classification of extensional basins into Atlantic type, back-arc, syn- and postorogenic rifts. This is followed by a discussion of thermal thinning of the lithosphere, doming and flood basalts, aspects of particular importance to volcanic rifted margins. We discuss the record of vertical motions during and after rifting in the context of stretching models developed to quantify rifted basin formation. As discussed in Section 6.11.2.1.5., the finite strength of the lithosphere has an important effect on the formation of extensional basins. This applies both to the geometry of the basin shape as well as to the record of vertical motions during and after rifting. We also address the tectonic control on postrift evolution of extensional basins. The concept of strength of the lithosphere has also important consequences for compressional basins. The latter include foreland basins as well as basins formed by lithospheric folding. In Section 6.11.3, we focus on thermomechanical aspects of sedimentary basin formation in the context of large-scale models for the underlying lithosphere. We highlight the connection between the bulk rheological properties of Europe’s lithosphere and the evolution of some of Europe’s main sedimentary basins. In Section 6.11.4, we investigate thermomechanical controls on continental breakup and associated
579 581 585 586 588 590 593 596
basin migration processes using the NW European margin as a natural laboratory. We specifically address relationships between rift duration and extension velocities, thermal evolution, and the role of mantle plumes and melts. This is followed by a brief discussion of compressional reactivation and its consequences for postrift inversion, borderland uplift, and denudation. In Section 6.11.5, we further develop the treatment of polyphase deformation of extensional basins taking the Black Sea Basin as a natural laboratory. We concentrate on rheological controls on basin formation affecting the large-scale basin stratigraphy and rift shoulder dynamics. We also discuss the role of intraplate stresses and lithospheric strength evolution during the postrift phase and consequences for neotectonic reactivation of the Black Sea basin system. In Section 6.11.6, we give an overview on the interplay of extension and compression in the Pannonian–Carpathian basin system of Central Eastern Europe. We begin with a review of temporal and lateral variations in lithospheric strength in the region and its effects on late-stage basin deformation. This is followed by summary of models proposed for the development of the Pannonian–Carpathian system. We also present results of three-dimensional (3-D) modeling approaches investigating the role of 3-D distributions of load and lithospheric strength in orogenic arcs. In doing so, we focus on implications of these models for a better understanding of polyphase subsidence in the Carpathian foredeep. For our discussion of lithospheric folding as a mode of basins formation, and for the interplay between lithosphere and surface processes in a compressional setting, we have selected the Iberian microcontinent, located within the Africa–Europe collision zone. In the first part of Section 6.11.7, we review constraints on vertical motions, present-day stress regime and interaction between surface
Tectonic Models for the Evolution of Sedimentary Basins
transport and vertical motions for Iberia at large. This is followed by a more detailed treatment of tectonic controls on drainage systems using the Ebro basin system of NE Iberia as a natural laboratory. The closing section, Section 6.11.8, draws general conclusions and addresses future perspectives. The origin of sedimentary basins is a key element in the geological evolution of the continental lithosphere. During the last decades, substantial progress was made in the understanding of thermomechanical processes controlling the evolution of sedimentary basins and the isostatic response of the lithosphere to surface loads such as sedimentary basins. Much of this progress stems from improved insights into the mechanical properties of the lithosphere, from the development of new modeling techniques, and from the evaluation of new, high-quality datasets from previously inaccessible areas of the globe. The focus of this chapter is on tectonic models processes controlling the evolution of sedimentary basins. After the realization that subsidence patterns of Atlantic-type margins, corrected for effects of sediment loading and palaeo-bathymetry, displayed the typical time-dependent decay characteristic of ocean-floor cooling (Sleep, 1971), a large number of studies were undertaken aimed at restoring the quantitative subsidence history of basins on the basis of well data and outcropping sedimentary sections. With the introduction of backstripping analysis algorithms (Steckler and Watts, 1982; Bond and Kominz, 1984), the late 1970s and early 1980s marked a phase during which basin analysis essentially stood for backward modeling, namely reconstructing the tectonic subsidence from sedimentary sequences. These quantitative subsidence histories provided constraints for the development of conceptually driven forward basin models. For extensional basins this commenced in the late 1970s, with the appreciation of the importance of the lithospheric thinning and stretching concepts in basin subsidence (Salveson, 1976). After initiation of mathematical formulations of stretching concepts in forward extensional basin modeling (McKenzie, 1978), a large number of basin fill simulations focused on the interplay between thermal subsidence, sediment loading, and eustatic sealevel changes. To arrive at commonly observed more episodic and irregular subsidence curves the smooth postrift subsidence behaviour was modulated by changes in sediment supply and eustatic sea-level fluctuations. Another approach frequently followed was to input a subsidence curve, thus rendering the basin modeling package essentially a tool to fill in an adopted
487
accommodation space (Burton et al., 1987; Lawrence et al., 1990). For the evolution of extensional basins this approach made a clear distinction between their syn-rift and postrift stage, relating exponentially decreasing postrift tectonic subsidence rates to a combination of thermal equilibration of the lithosphere– asthenosphere system and lithospheric flexure (Watts et al., 1982). A similar set of assumptions were made to describe the syn-rift phase. In the simplest version of the stretching model (McKenzie, 1978), lithospheric thinning was described as resulting from more or less instantaneous extension. In these models a component of lithosphere mechanics was obviously lacking. On a smaller scale, tilted fault block models were introduced for modeling of the basin fill at the scale of half-graben models. Such models essentially decouple the response of the brittle upper crust from deeper lithospheric levels during rifting phases (see, e.g., Kusznir et al., 1991). A noteworthy feature of most modeling approaches was their emphasis on the basin subsidence record and their very limited capability to handle differential subsidence and uplift patterns in a process-oriented, internally consistent manner (see, e.g., Kusznir and Ziegler, 1992; Larsen et al., 1992; Dore´ et al., 1993). To a large extent the same was true for most of compressional basin modeling. The importance of the lithospheric flexure concept, relating topographic loading of the crust by an overriding mountain chain to the development of accommodation space, was recognized as early as 1973 by Price in his paper on the foreland of the Canadian Rocky Mountains thrust belt (Price, 1973). Also, here it took several years before quantitative approaches started to develop, investigating the effects of lithospheric flexure on foreland basin stratigraphy (Beaumont, 1981). The success of flexural basin stratigraphy modeling, capable of incorporating subsurface loads related to plate tectonic forces operating on the lithosphere (e.g., Van der Beek and Cloetingh, 1992; Peper et al., 1994) led to the need to incorporate more structural complexity in these models, also in view of implications for the simulation of thermal maturation and fluid migration (e.g., Parnell, 1994). The necessary understanding of lithospheric mechanics and basin deformation was developed after a bridge was established between researchers studying deeper lithospheric processes and those who analyzed the record of vertical motions, sedimentation, and erosion in basins. This permitted the development of
488
Tectonic Models for the Evolution of Sedimentary Basins
basin analysis models that integrate structural geology and lithosphere tectonics. The focus of modeling activities in 1990s was on the quantification of mechanical coupling of lithosphere processes to the near-surface expression of tectonic controls on basin fill (Cloetingh et al., 1995a, 1995b, 1996, 1997). This invoked a process-oriented approach, linking different spatial and temporal scales in the basin record. Crucial in this was the testing and validation of modeling predictions in natural laboratories for which high-quality databases were available at deeper crustal levels (deep reflection- and refraction-seismic) and the basin fill (reflection-seismic, wells, outcrops), demanding a close cooperation between academic and industrial research groups (see Watts et al., 1993; Sassi et al., 1993; Cloetingh et al., 1994; Roure et al., 1996b). Figure 1 displays the global distribution of sedimentary basins (Laske and Masters, 1997; see also: Bally and Snelson, 1980, Schlumberger, 1991). Figure 2 gives the location of extensional and compressional basins in Europe that were selected as examples to discuss advances in data-interactive quantitative basin modeling. Applying a consistent modeling approach to different basins provides an opportunity to compare in this chapter important key parameters of basin evolution (Table 1), shedding light on the tectonic controls underlying observed similarities and differences in basin histories. In the following, we review recent advances in modeling the initiation and evolution of sedimentary basins in their lithospheric context on a global and more regional scale. To the latter purpose we follow a natural laboratory approach selecting some welldocumented basins of Europe. Below we first start with a summary of key features of the tectonics of
Figure 1 Sedimentary basins of the World. Onshore basins are shown in green, offshore basins are lavender and shown for a maximum water depth of 1000 m. Modified from Schlumberger (1991). World oil reserves – charting the future. Middle East Well Evaluation Reivew 10: 7–15.
rifted basins on a global scale. In doing so, we introduce in Section 6.11.2 basic concepts for the tectonics of basin formation. We begin with a review of the dynamics of extensional basin systems. First we discuss the genetic types of extensional basins followed by an overview of characteristics of thermal thinning of the lithosphere, doming and flood basalt extrusion, aspects of particular importance to volcanic rifted margins. We examine the record of vertical motions during and after rifting in the context of stretching models developed to quantify rifted basin development. In this section we proceed with thermomechanical aspects of extensional sedimentary basin development in the context of large-scale models for the underlying lithosphere. We also address the tectonic control on postrift evolution of extensional basins. In Section 6.11.2.2 we review the development and evolution of compressional basins in a lithospheric context. We focus on foreland basins and basins that evolved as a result of large-scale compressional folding of intraplate continental lithosphere. Specifically, we address the role of flexure in compressional basin evolution and the interplay between lithosphere dynamics and topography in compressional basin systems. In Section 6.11.3, we highlight the connection between the bulk rheological properties of the lithosphere and the evolution of some of Europe’s main sedimentary basins. These include some of the bestdocumented sedimentary basin systems of the world. As discussed in Section 6.11.2, the finite strength of the lithosphere plays an important role in the development of extensional and compressional basins. This applies both to the geometry of the basin shape as well as to the record of vertical motions during and after rifting. As pointed out above, the concept of strength of the lithosphere also has important consequences for compressional basins. The latter include foreland basins as well as basins formed in response to lithospheric folding. Section 6.11.3 sets the stage for Sections 6.11.4– 6.11.7 in which basin modeling studies carried out in a number of selected natural laboratories are described. In each of these sections a particular aspect of tectonic processes controlling the evolution of sedimentary basins is examined. In this way, a review of recent advances and data interactive modeling, integrating process modeling with geological and geophysical data is given, covering key aspects of passive-margin evolution, back-arc basins, foreland basins and basins formed by lithospheric folding. As will become clear from these sections, polyphase evolution of basin systems is the rule rather than an exception.
Tectonic Models for the Evolution of Sedimentary Basins
–
–
–
489
+
° 60
–
+
+
2
+
–
+ –
50°
– + 1
+
– 4
+
3
–
+
–
40°
– 5
350°
+
0°
10°
20°
30°
Figure 2 Topographic map of Europe, showing intraplate areas of Late Neogene uplift (circles with plus symbols) and subsidence (circles with minus symbols). Boxes show sedimentary basin systems discussed in this chapter. (1) European Cenozoic rift system and adjacent areas (Section 6.11.2); (2) Northwest European continental margin (Section 6.11.4); (3) Black Sea Basin (Section 6.11.5); (4) Carpathian–Pannonian Basin System (Section 6.11.6); (5) Iberian microcontinent (Section 6.11.7).
Table 1
Key parameters of basin evolution
Symbol
Name
Value
a c Ta
Initial lithosphere thickness Initial crustal thickness Asthenospheric temperature Thermal diffusivity Surface crustal density Surface mantle density Water density Sediment densities
100 km 30 km 1333 C
k c m w s
Thermal expansion factor Crustal stretching factor Subcrustal stretching factor
106 m2 s1 2800 kg m3 3300 kg m3 1030 kg m3 2100– 2650 kg m3 3.2 105 K1
In Section 6.11.4, we investigate thermomechanical controls on continental breakup and associated basin migration processes using the NW European margin as a natural laboratory. The volcanic rifted margins of the northern Atlantic are probably the best documented in the world as a result of a major concentrated effort by academia and industry, the latter in the context of hydrocarbon exploration and production.
A particularly striking feature is the very long duration of the rifting phases prior to continental breakup. This margin system also allows the quantification of controls on rifted margin topography and compressional reactivation during the post breakup phase. In doing so, we specifically address relationships between rift duration and extension velocities, thermal evolution and the role of mantle plumes and melts. This is followed by a brief discussion on compressional reactivation and its consequences for postrift inversion, borderland uplift and denudation. In Section 6.11.5, we further develop the treatment of polyphase deformation of extensional basins taking the Black Sea Basin as a natural laboratory. An intriguing feature of this basin system is the concentration of postrift deformation at its margins, without affecting the basin center. We concentrate on rheological controls on basin formation and its consequences for large-scale basin stratigraphy and rift shoulder dynamics. We also discuss the role of intraplate stresses and lithospheric strength evolution during the postrift phase and implications for neotectonic reactivation of the Black Sea Basin system. In Section 6.11.6, we give an overview on the interplay between extension and compression in the
490
Tectonic Models for the Evolution of Sedimentary Basins
Pannonian–Carpathian basin system of Central Eastern Europe. This area is the site of pronounced contrasts of lithospheric strength between the Pannonian area, which is probably underlain by the hottest and weakest lithosphere of Europe, and the particularly strong East-European Platform lithosphere bounding the Carpathian arc to the East. Noteworthy features of the Pannonian system are the short duration of rifting phases in a back-arc setting affected by extensional collapse and subsequent compressional reactivation. The Carpathian arc is associated with probably one of the deepest foredeeps of the world, the Focs¸ani Basin that developed in front of the bend zone of the Carpathians in an area that is strongly affected by neotectonics and seismicity. This basin contains more than 9 km of Neogene sediments. This system offers also a unique opportunity to study basin evolution in the aftermath of continental collision. We begin with a review of temporal and lateral variations in lithospheric strength in this region and its effects on late-stage basin deformation. This is followed by a summary of models proposed for the development of the Pannonian–Carpathian system. In the second part of Section 6.11.6, we present the results of 3-D modeling approaches investigating the role of 3-D distribution of loads and lithospheric strength in orogenic arcs. In doing so, we focus on implications of these models for a better understanding of polyphase subsidence in the Carpathian foredeep. For our discussion on lithospheric folding as a mode of basins development, and for the interplay between lithosphere and surface processes in a compressional setting, we have selected the Iberian microcontinent, located within the Africa–Europe collision zone. In the first part of Section 6.11.7, we review constraints on vertical motions, present-day stress regime and interaction between surface transport and vertical motions for Iberia at large. This is followed by a more detailed treatment of tectonic controls on drainage systems using the Ebro Basin system of NE Iberia as a natural laboratory.
6.11.2 Tectonics of Extensional and Compressional Basins: Concepts and Global-Scale Observations 6.11.2.1
Extensional Basin Systems
Tectonically active rifts, palaeo-rifts and passive margins form a group of genetically related extensional basins that play an important role in the
spectrum of sedimentary basin types (Bally and Snelson, 1980; Ziegler and Cloetingh, 2004; see also Buck, this volume (Chapter 6.08)). Extensional basins cover large areas of the globe and contain important mineral deposits and energy resources. A large number of major hydrocarbon provinces are associated with rifts (e.g., North Sea, Sirt and West Siberian basins, Dniepr–Donets and Gulf of Suez grabens) and passive margins (e.g., Campos Basin, Gabon, Angola, Mid-Norway and NW Australian shelves, Niger and Mississippi deltas; Ziegler, 1996a,b). On these basins, the petroleum industry has acquired large databases that document their structural styles and allow detailed reconstruction of their evolution. Academic geophysical research programs have provided information on the crustal and lithospheric configuration of tectonically active rifts, palaeo-rifts, and passive margins. Petrologic and geochemical studies have advanced the understanding of riftrelated magmatic processes. Numerical models, based on geophysical and geological data, have contributed at lithospheric and crustal scales toward the understanding of dynamic processes that govern the evolution of rifted basins. Below we summarize basic concepts on dynamic processes that control the evolution of extensional basins. A natural distinction can be made between tectonically active and inactive rifts, and rifts that evolved in continental and oceanic lithosphere. Tectonically active intracontinental (intraplate) rifts, such as the Rhine Graben, the East African Rift, the Baikal Rift and the Shanxi Rift of China, correspond to important earthquake and volcanic hazard zones. The globe-encircling mid-ocean ridge system forms an immense intraoceanic active rift system that encroaches onto continents in the Red Sea and the Gulf of California. Rifts that are tectonically no longer active are referred to as palaeo-rifts, aulacogens, inactive or aborted rifts and failed arms, in the sense that they did not progress to crustal separation. Conversely, the evolution of successful rifts culminated in the breakup of continents, the opening of new oceanic basins and the development of conjugate pairs of passive margins. In the past, a genetic distinction was made between ‘active’ and ‘passive’ rifting (Sengo¨r and Burke, 1978; Olsen and Morgan, 1995). ‘Active’ rifts are thought to evolve in response to thermal upwelling of the asthenosphere (Dewey and Burke, 1975; Bott and Kusznir, 1979), whereas ‘passive’ rifts develop in response to lithospheric extension driven by far-field stresses (McKenzie, 1978). It is, however,
Tectonic Models for the Evolution of Sedimentary Basins
questionable whether such a distinction is justified as the study of Phanerozoic rifts revealed that riftrelated volcanic activity and doming of rift zones is basically a consequence of lithospheric extension and is not the main driving force of rifting. The fact that rifts can become tectonically inactive at all stages of their evolution, even if they have progressed to the Red Sea stage of limited sea-floor spreading (e.g., Bay of Biscay-Pyrenean rift), supports this concept. However, as extrusion of large volumes of rift-related subalkaline tholeiites must be related to a thermal anomaly within the upper mantle, a distinction between ‘active’ and ‘passive’ rifting is to a certain degree still valid, though not as ‘black and white’ as originally envisaged. Rifting activity, preceding the breakup of continents is probably governed by forces controlling the movement and interaction of lithospheric plates. These forces include plate boundary stresses, such as slab pull, slab roll-back, ridge push and collisional resistance, and frictional forces exerted by the convecting mantle on the base of the lithosphere (Forsyth and Uyeda, 1975; Bott, 1993; Ziegler, 1993; see also Wessel and Mu¨ller, this volume (Chapter 6.02)). On the other hand, deviatoric tensional stresses, inherent to the thickened lithosphere of young orogenic belts, as well as those developing in the lithosphere above upwelling mantle convection cells and mantle plumes (Bott, 1993) do not appear to cause, on their own, the breakup of continents. However, if such stresses interfere constructively with plate boundary and/or mantle drag stresses, the yield strength of the lithosphere may be exceeded, thus inducing rifting (Figure 3). It must be understood that mantle drag forces are exerted on the base of a lithospheric plate if its velocity and direction of movement differs from the velocity and direction of the mantle flow. Mantle drag can constructively or destructively interfere with plate boundary forces, and thus can either contribute towards plate motion or resist it. Correspondingly, mantle drag can give rise to the buildup of extensional as well as compressional intraplate stresses (Forsyth and Uyeda, 1975; Bott, 1993; Artemieva and Mooney, 2002). Although the present lithospheric stress field can be readily explained in terms of plate boundary forces (Cloetingh and Wortel, 1986; Richardson, 1992; Zoback, 1992), mantle drag probably contributed significantly to the Triassic–Early Cretaceous breakup of Pangea, during which Africa remained nearly stationary and straddled an evolving upwelling and radial outflowing mantle
491
convection cell (Pavoni, 1993; Ziegler, 1993; Ziegler et al., 2001). Mechanical stretching of the lithosphere and thermal attenuation of the lithospheric mantle are associated with the development of local deviatoric tensional stresses, which play an increasingly important role during advanced rifting stages (Bott, 1992; Ziegler, 1993). This has led to the development of the concept that many rifts go through an evolutionary cycle starting with an initial ‘passive’ phase that is followed by a more ‘active’ stage during which magmatic processes play an increasingly important role (Wilson, 1993a; Burov and Cloetingh, 1997; Huismans et al., 2001a). However, nonvolcanic rifts must be considered as purely ‘passive’ rifts. 6.11.2.1.1
Modes of rifting and extension
Atlantic-type rifts Atlantic-type rift systems evolve during the breakup of major continental masses, presumably in conjunction with a reorganization of the mantle convection system (Ziegler, 1993). During early phases of rifting, large areas around future zones of crustal separation can be affected by tensional stresses, giving rise to the development of complex graben systems. In time, rifting activity concentrates on the zone of future crustal separation, with tectonic activity decreasing and ultimately ceasing in lateral graben systems. In time and as a consequence of progressive lithospheric attenuation and ensuing crustal doming, local deviatoric tensional stresses play an important secondary role in the evolution of such rift systems. Upon crustal separation, the diverging continental margins (pericontinental rifts) and the ‘unsuccessful’ intracontinental branches of the respective rift system become tectonically inactive. However, during subsequent tectonic cycles, such aborted rifts can be tensionally as well as compressionally reactivated (Ziegler et al., 1995, 1998, 2001, 2002). Development of Atlantic-type rifts is subject to great variations mainly in terms of duration of their rifting stage and the level of volcanic activity (Ziegler, 1988, 1990b, 1996b). 6.11.2.1.1.(i)
6.11.2.1.1.(ii) Back-arc rifts Back-arc rifts are thought to evolve in response to a decrease in convergence rates and/or even a temporary divergence of colliding plates, ensuing steepening of the subduction slab and development of a secondary upwelling system in the upper plate mantle wedge above the subducted lower plate lithospheric slab (Uyeda and McCabe, 1983). Changes in convergence rates between colliding plates are probably an expression of changes in plate interaction. Back-arc rifting can
492
Tectonic Models for the Evolution of Sedimentary Basins
Preexisting lithospheric discontinuities control location of rift Down-welling mantle
Lateral mantle outflow
Deviatoric tension Diverging mantle flow
Deviatoric tension
Lithosphere
Mantle drag
Ridge push
Mantle drag plus ridge push
Displacement vector Mantle flow direction Stress vector Figure 3 Diagram illustrating the interaction of shear-traction exerted on the base of the lithosphere by asthenospheric flow, deviatoric tension above upwelling mantle convection cells and ridge push forces. Modified from Ziegler PA, Cloetingh S, Guiraud R, and Stampfli GM (2001). Peri-Tethyan platforms: constraints on dynamics of rifting and basin inversion. In: Ziegler PA, Cavazza W, Robertson AHF, and Crasquin-Soleau S (eds.) Me´moires du Museum National d’Histoire Naturelle 186: Peri-Tethys Memoir 6: Peri-Tethyan Rift/Wrench Basins and Passive Margins, pp. 9–49. Paris: Commission for the Geological Map of the World.
progress to crustal separation and the opening of limited oceanic basins (e.g., Sea of Japan, South China Sea, Black Sea). However, as convergence rates of colliding plates are variable in time, back-arc extensional basins are generally short-lived. Upon a renewed increase in convergence rates, back-arc
extensional systems are prone to destruction by back-arc compressional stresses (e.g., Variscan Rheno-Hercynian Basin, Sunda Arc and East China rift systems, Black Sea domain; Uyeda and McCabe, 1983; Cloetingh et al., 1989; Jolivet et al., 1989; Ziegler, 1990a; Letouzey et al., 1991; Nikishin et al., 2001).
Tectonic Models for the Evolution of Sedimentary Basins 6.11.2.1.1.(iii) Syn-orogenic rifting and wrenching Syn-orogenic rift/wrench deformations
can be related to indenter effects and ensuing escape tectonics, often involving rotation of intramontane stable blocks (e.g., Pannonian Basin: Royden and Horva´th, 1988; Late Carboniferous Variscan fold belt: Ziegler, 1990b), as well as to lithospheric overthickening in orogenic belts, resulting in uplift and extension of their axial parts (Peruvian and Bolivian Altiplano: Dalmayrac and Molnar, 1981, Mercier et al., 1992). Furthermore, collisional stresses exerted on a craton may cause far-field tensional or transtensional reactivation of preexisting fracture systems and thus the development of rifts and pull-apart basins. This model may apply to the Late Carboniferous development of the Norwegian–Greenland Sea rift in the foreland of the Variscan Orogen, the PermoCarboniferous Karoo rifts in the hinterland of the Gondwanan Orogen, the Cenozoic European rift system in the Alpine foreland and the Neogene Baikal rift in the hinterland of the Himalayas (Ziegler et al., 2001; De`zes et al., 2004). Under special conditions, extensional structures can also develop in forearc basins (e.g., Talara Basin, Peru; see Ziegler and Cloetingh, 2004). Postorogenic extension Extensional disruption of young orogenic belts, involving the development of grabens and pull-apart structures, can be related to their postorogenic uplift and the development of deviatoric tensional stresses inherent to orogenically overthickened crust (Stockmal et al., 1986; Dewey, 1988; Sanders et al., 1999). The following mechanisms contribute to postorogenic uplift: (1) locking of the subduction zone due to decay of the regional compressional stress field (Whittaker et al., 1992); (2) roll-back and ultimately detachment of the subducted slab from the lithosphere (Fleitout and Froidevaux, 1982; Bott, 1993; Andeweg and Cloetingh, 1998); and (3) retrograde metamorphism of the crustal roots, involving, in the presence of fluids, the transformation of eclogite to less dense granulite (Le Pichon et al., 1997; Straume and Austrheim, 1999). Although tensional collapse of an orogen can commence shortly after its consolidation (e.g. the Variscan Orogen: Ziegler, 1990b; Ziegler et al., 2004), it may be delayed by as much as 30 My, as in the case of the Appalachian-Mauretanides (Ziegler, 1990b). The Permo-Triassic West Siberian Basin that is superimposed on the juncture of the Uralides and Altaides is an example of a postorogenic tensional basin that began to subside shortly after the consolidation of these orogens (Nikishin et al., 2002; 6.11.2.1.1.(iv)
493
Vyssotski et al., 2006). In addition, modifications in the convergence direction of colliding continents and the underlying stress reorientation can give rise to the development of wrench fault systems and related pull-apart basins, controlling early collapse of an orogen, such as the Devonian collapse of the Arctic– North Atlantic Caledonides (Ziegler, 1989b; Ziegler and De`zes, 2006) or the Stephanian–Early Permian disruption of the Variscan fold belt (Ziegler et al., 2004). The Basin and Range Province of North America is a special type of postorogenic rifting. Oligocene and younger collapse of the US Cordillera is thought to be an effect of the North American craton having overridden at about 28 Ma the East Pacific Rise in conjunction with rapid opening of the Atlantic Ocean (Verall, 1989). In the area of the southwestern US Cordillera, regional compression waned during the late Eocene and the orogen began to collapse during late Oligocene with main extension occurring during the Miocene and Pliocene (Parsons, 1995). By contrast, the Canadian Cordillera remained intact. During the collapse of the US Cordillera, the heavily intruded, at middle and lower levels ductile crust of the Basin and Range Province was subjected to major extension at high strain rates, resulting in uplift of ductilely deformed core complexes by 10–20 km. The area affected by extension, crustal thinning, volcanism, and uplift measures 1500 1500 km (Wernicke, 1990). The Eo-Oligocene magmatism of the Basin and Range Province bears a subductionrelated signature, suggestive of an initial phase of back-arc extension, whereas lithospheric mantleand asthenosphere-derived magmas play an increasingly important role from Miocene times onward, presumably due to the opening of asthenospheric windows in the Farralon slab during its detachment from in the lithosphere and gradual sinking into the mantle (Parsons, 1995). 6.11.2.1.2 Thermal thinning and stretching of the lithosphere: concepts and models
Mechanical stretching of the lithosphere, triggering partial melting of its basal parts and the upper asthenosphere, is followed by segregation of melts and their diapiric rise into the lithosphere, an increase in conductive and advective heat flux and consequently an upward displacement of the thermal asthenosphere– lithosphere boundary. Small-scale convection in the evolving asthenospheric diapir may contribute to mechanical thinning of the lithosphere by facilitating
494
Tectonic Models for the Evolution of Sedimentary Basins
lateral ductile mass transfer (Figure 4) (Richter and McKenzie, 1978; Mareschal and Gliko, 1991). Progressive thermal and mechanical thinning of the higher-density lithospheric mantle and its replacement by lower-density asthenosphere induces progressive doming of rift zones. At the same time, deviatoric tensional stresses developing in the lithosphere contribute to its further extension (Bott, 1992). Although major hot-spot activity is thought to be related to a thermal perturbation within the asthenosphere caused by a deep mantle plume, smaller-scale ‘plume’ activity may also be the consequence of lithospheric stretching triggering by adiabatic decompression partial melting in areas characterized by an anomalously volatile-rich asthenosphere/ lithosphere (Wilson, 1993a; White and McKenzie, 1989). In this context, it is noteworthy that extension-induced development of a partially molten asthenospheric diapir that gradually rises into the lithosphere causes by itself further decompression of the underlying asthenosphere and consequently more extensive partial melting and melt segregation at progressively deeper levels. Thus, the evolving diapir may not only grow upward but also downward. Similarly, acceleration of plate divergence and the ensuing increase in sea-floor spreading rates probably causes at spreading axes partial melting and melt segregation at progressively deeper asthenospheric
levels reaching down to 80–100 km, as imaged seismic tomography (Anderson et al., 1992). Melts, which intrude the lithosphere and pond at the crust–mantle boundary, provide a further mechanism for thermal doming of rift zones (Figure 4). Emplacement of such asthenoliths, consisting of a mixture of indigenous subcrustal mantle material and melts extracted from deeper lithospheric and upper asthenospheric levels, may cause temporary doming of a rift zone and a reversal in its subsidence pattern, such as for instance the MidJurassic Central North Sea arch (Ziegler, 1990b) or the Neogene Baikal arch (Suvorov et al., 2002). Depending on the applicability of the ‘pure-shear’ (McKenzie, 1978) or the ‘simple-shear’ model (Wernicke, 1985), or a combination thereof (Kusznir et al., 1991), the zone of upper crustal extension, corresponding to the subsiding rift, may coincide with the zone of lithospheric mantle attenuation (pure- and combined-shear) or may be laterally offset from it (simple-shear, see Figure 5). Under conditions of Pure shear
McKenzie (1978)
Simple shear
Wernicke (1981)
Mantle plume model 0
0
50
50
100
100
Tensional failure model
Combined shear
0
0
50
50
100
100
Syn-rift sediments Prerift sediments
Mantle lithosphere
Crust
Flow pattern
Figure 4 Rift models.
Barbier et al. (1986)
Crust
Convecting asthenosphere
Mantle lithosphere
Zone of ductile deformation
Asthenosphere
Figure 5 Lithospheric shear models.
Tectonic Models for the Evolution of Sedimentary Basins
pure-shear lithospheric extension, magmatic activity should be centered on the rift axis where in time mid-oceanic ridge basalt (MORB) type magmas can be extruded after a high degree of extension has been achieved. By contrast, under simple-shear conditions, magmatic activity is asymmetrically distributed with respect to the rift axis and MORB-type extrusives may occur on one of the rift flanks (e.g., the Basin and Range Province (Jones et al., 1992), the Red Sea (Favre and Stampfli, 1992), and the Ethiopian Rift (Kazmin, 1991). A modification to the pure-shear model is the ‘continuous depth-dependent’ stretching model which assumes that stretching of the lithospheric mantle affects a broader area than the zone of crustal extension (Figure 6) (Rowley and Sahagian, 1986). In both models it is assumed that the asthenosphere wells up passively into the space created by mechanical attenuation of the lithospheric mantle. In depthdependent stretching models this commonly gives rise to flexural uplift of the rift shoulders, and in the flexural cantilever model, which assumes ductile deformation of the lower crust, this produces footwall uplift of the rift flanks and intrabasin fault blocks (Kusznir and Ziegler, 1992). By the same mechanism, the simple-shear model predicts asymmetrical doming of a rift zone or even flexural uplift of an arch located
to one side of the zone of upper crustal extension (Wernicke, 1985). A modification to the simple-shear model envisages that massive upper crustal extensional unloading of the lithosphere causes its isostatic uplift and passive inflow of the asthenosphere (Wernicke, 1990). The structural style of rifts, as defined at upper crustal and syn-rift sedimentary levels, is influenced by the thickness and thermal state of the crust and lithospheric mantle at the onset of rifting, by the amount of crustal extension and the width over which it is distributed, the mode of crustal extension (orthogonal or oblique, simple- or pure-shear) and the lithological composition of the pre- and syn-rift sediments (Cloetingh et al., 1995b; Ziegler, 1996b). A major factor controlling the structural style of a rift zone is the magnitude of the crustal extensional strain that was achieved across it and the distance over which it is distributed ( factor). Although quantification of the extensional strain and of the stretching factor is of basic importance for the understanding of rifting processes, there is often a considerable discrepancy between estimates derived from upper crustal extension by faulting, the crustal configuration and quantitative subsidence analyses (Ziegler and Cloetingh, 2004).
Uniform stretching model
δ crust = β mantle lithosphere
McKenzie (1978)
Discontinuous, depth-dependent stretching model
δ crust < β mantle lithosphere Royden and Keen (1980) Beaumont et al. (1982)
Continuous, depth-dependent stretching model
δ crust = β mantle lithosphere Rowley and Sahagian (1986)
Crust Figure 6 Lithospheric stretching models.
495
Mantle lithosphere
496
Tectonic Models for the Evolution of Sedimentary Basins
6.11.2.1.3 Syn-rift subsidence and duration of rifting stage
The balance of two mechanisms controls the syn-rift subsidence of a sedimentary basin. First, elastic/isostatic adjustment of the crust to stretching of the lithosphere and its adjustment to sediment loading causes subsidence of the mechanically thinned crust (Figures 5 and 6) (McKenzie, 1978; Keen and Boutilier, 1990). Depending on the depth of the lithospheric necking level, this is accompanied by either flexural uplift or down warping of the rift zone (Figure 7) (Braun and Beaumont, 1989; Kooi, 1991; Kooi et al., 1992). Second, uplift of a rift zone is caused by upwelling of the asthenosphere into the space created by mechanical stretching of the lithosphere, thermal upward displacement of the asthenosphere– lithosphere boundary, thermal expansion of the lithosphere and intrusion of melts at the base of the crust (Figure 4) (Turcotte and Emermann, 1983). Thus, the geometry of a rifted basin is a function of the elastic/isostatic response of the lithosphere to its mechanical stretching and related thermal perturbation (Van der Beek et al., 1994). The duration of the rifting stage of intracontinental rifts (aborted) and passive margins (successful rifts) is highly variable (Figures 8 and 9) (Ziegler, 1990b, Ziegler et al., 2001, Ziegler and Cloetingh, 2004). Overall, it is observed that, in time, rifting activity concentrates on the zone of future crustal separation with lateral rift systems becoming inactive. However, as not all rift systems progress to crustal separation, the duration of their rifting stage is obviously a function of the persistence of the controlling stress field. On the other hand, the time required to achieve crustal separation is a function
Strength envelope
Kinematic modeling
Crust Mantle Depth of necking
of the strength (bulk rheology) of the lithosphere, the buildup rate, magnitude and persistence of the extensional stress field, constraints on lateral movements of the diverging blocks (on-trend coherence, counteracting far-field compressional stresses), and apparently not so much of the availability of preexisting crustal discontinuities that can be tensionally reactivated. Crustal separation was achieved in the Liguro-Provenc¸al Basin after 9 My of crustal extension and in the Gulf of California after about 14 My of rifting, whereas opening of the Norwegian– Greenland Sea was preceded by an intermittent rifting history spanning some 280 My (Ziegler, 1988; Ziegler and Cloetingh, 2004). There appears to be no obvious correlation between the duration of the rifting stage (R) of successful rifts (Figure 9), which are superimposed on orogenic belts (LiguroProvenc¸al Basin, Pyrenees R ¼ 9 My; Gulf of California, Cordillera R ¼ 14 My; Canada Basin, Inuitian fold belt R ¼ 35 My; Central Atlantic, Appalachians R ¼ 42 My; Norwegian–Greenland Sea, Caledonides R ¼280 My) and those which developed within stabilized cratonic lithosphere (southern South Atlantic R ¼ 13 My; northern South Atlantic R ¼ 29 My; Red Sea R ¼ 29 My; Baffin Bay R ¼ 70 My; Labrador Sea R ¼ 80 My). This suggests that the availability of crustal discontinuities, which regardless of their age (young orogenic belts, old Precambrian shields) can be tensionally reactivated, does not play a major role in the time required to achieve crustal separation. However, by weakening the crust, such discontinuities play a role in the localization and distribution of crustal strain. Moreover, by weakening the lithosphere, they Regional isostatic response
Upward flexure
Strength envelope Crust Depth of necking Mantle
Downward flexure
Figure 7 Concept of lithospheric necking. The level of necking is defined as the level of no vertical motions in the absence of isostatic forces. Left panel: kinematically induced configuration after rifting for different necking depths. Right panel: subsequent flexural isostatic rebound.
Western approaches, UK
497
R 210 60
R 110163 85
D
R 133
137
West shetland trough
Sirt, Libya
Midland valley, UK
Oslo graben, Norway
Tucano graben, NE Brazil
Euphrates graben, Syria
Muglad, Soudan
Abortive rifts
Dniepr-Donets, Ukraine
Cent. North Sea
Tectonic Models for the Evolution of Sedimentary Basins
248
270
115
R 110 30
R 95 45
D D
R 65 305
R 54 256
R 28 120
R 20 65
R 7.5
D
362.5 370
85
148
310
370
140
248
140
Figure 8 Duration of rifting stage of ‘abortive’ rifts (palaeo-rifts, failed arms). Vertical columns in My; numbers on side of vertical columns indicate onset and termination of rifting stage in My; numbers under R on top of each column give the duration of rifting stage in My; stars indicate periods of main volcanic activity; D indicates periods of doming. Modified from Ziegler PA and Cloetingh S (2004). Dynamic processes controlling evolution of rifted basins. Earth-Science Reviews 64: 1–50.
R 280 S 55 55
21.5 30.5
D 6 20
165 118
R 80 S 52
R 70 S 27
Norwegian-Greenland Sea
S. Rockall trough
Mozambic-Somali basin
North Atlantic
Labrador Sea
Baffin Bay
95
R 130 S 16
R 122 S 118
D
R 155 S 15
80 55
R 42 S 182
R 29 S5
R 14 S6
R9 S5
Central Atlantic
Red Sea
Gulf of California
Provencal-Ligurian basin
Successful rifts
182
D 5
34
224
125
160
240
295
260
330
Figure 9 Duration of rifting stage of ‘successful’ rifts. Legend same as Figure 8. Numbers beside letter ‘S’ indicate duration of seafloor spreading stage in My. Modified from Ziegler PA and Cloetingh S (2004). Dynamic processes controlling evolution of rifted basins. Earth-Science Reviews 64: 1–50.
498
Tectonic Models for the Evolution of Sedimentary Basins
contribute to the preferential tensional reactivation of young as well as old orogenic belts (Janssen et al., 1995; Ziegler et al., 2001). 6.11.2.1.4
Postrift subsidence Similar to the subsidence of oceanic lithosphere, the postrift subsidence of extensional basins is mainly governed by thermal relaxation and contraction of the lithosphere, resulting in a gradual increase of its flexural strength, and by its isostatic response to sedimentary loading. Theoretical considerations indicate that subsidence of postrift basins follows an asymptotic curve, reflecting the progressive decay of the rift-induced thermal anomaly, the magnitude of which is thought to be directly related to the lithospheric stretching value (Sleep, 1973; McKenzie, 1978; Royden et al., 1980; Steckler and Watts, 1982; Beaumont et al., 1982; Watts et al., 1982). During the postrift evolution of a basin, the thermally destabilized continental lithosphere reequilibrates with the asthenosphere (McKenzie, 1978; Steckler and Watts, 1982; Wilson, 1993b). In this process, in which the temperature regime of the asthenosphere plays an important role (ambient, below, or above ambient), new lithospheric mantle, consisting of solidified asthenospheric material, is accreted to the attenuated old continental lithospheric mantle (Ziegler et al., 1998). In addition, densification of the continental lithosphere involves crystallization of melts that accumulated at its base or were injected into it, subsequent thermal contraction of the solidified rocks and, under certain conditions, their phase transformation to eclogite facies. The resulting negative buoyancy effect is the primary cause of postrift subsidence. However, in a number of basins, significant departures from the theoretical thermal subsidence curve are observed. These can be explained as effects of compressional intraplate stresses and related phase transformations (Cloetingh and Kooi, 1992; Van Wees and Cloetingh, 1996; Lobkovsky et al., 1996). 6.11.2.1.4.(i) Shape and magnitude of rift-induced thermal anomalies The shape and
dimension of rift-induced asthenosphere–lithosphere boundary anomalies essentially controls the geometry of the evolving postrift thermal-sag basin (Figure 5). Thermal sag basins associated with aborted rifts are broadly saucer-shaped and generally overstep the rift zone, with their axes coinciding with the zone of maximum lithospheric attenuation. Pure-shear-dominated rifting gives rise to the classical ‘steer’s head’ configuration of the syn- and
postrift basins (White and McKenzie, 1989) in which both basin axes roughly coincide, with the postrift basin broadly overstepping the rift flanks. This geometric relationship between syn- and postrift basins is frequently observed (e.g., North Sea Rift: Ziegler, 1990b; West Siberian Basin: Artyushkov and Baer, 1990; Dniepr-Donets Graben: Kusznir et al., 1996, Stephenson et al., 2001; Gulf of Thailand: Hellinger and Sclater, 1983; Watcharanantakul and Morley, 2000; Sudan rifts: McHargue et al., 1992). Such a geometric relationship is compatible with discontinuous, depth-dependent stretching models that assume that the zone of crustal extension is narrower than the zone of lithospheric mantle attenuation (Figure 6). The degree to which a postrift basin oversteps the margins of the syn-rift basin is a function of the width difference between the zone of crustal extension and the zone of lithospheric mantle attenuation (White and McKenzie, 1988) and the effective elastic thickness (EET) of the lithosphere (Watts et al., 1982). Conversely, simple-shear dominated rifting gives rise to a lateral offset between the syn- and postrift basin axes. An example is the Tucano Graben of northeastern Brazil, which ceased to subside at the end of the rifting stage, whereas the coastal Jacuipe–Sergipe– Alagoas Basin, to which the former was structurally linked, was the site of crustal and mantle-lithospheric thinning culminating in crustal separation and subsequent major postrift subsidence (Karner et al., 1992; Chang et al., 1992). Moreover, the simple-shear model can explain the frequently observed asymmetry of conjugate passive margins and differences in their postrift subsidence pattern (e.g., Central Atlantic and Red Sea: Favre and Stampfli, 1992). Discrepancies in the postrift subsidence of conjugate margins are attributed to differences in their lithospheric configuration at the crustal separation stage. At the end of the rifting stage, a relatively thick lithospheric mantle supports lower plate margins, whereas upper plate margins are underlain by a strongly attenuated lithospheric mantle and partly directly by the asthenosphere. Correspondingly, lower plate margins are associated at the crustal separation stage with smaller thermal anomalies than upper plate margins (Ziegler et al., 1998; Stampfli et al., 2001). These differences in lithospheric configuration of conjugate simple-shear margins have repercussions on their rheological structure, even after full thermal relaxation of the lithosphere, and their compressional reactivation potential (see further ahead; Ziegler et al., 1998) (Table 2).
Tectonic Models for the Evolution of Sedimentary Basins Table 2 Rheological and thermal parameters of crust and lithospheric materials, adopted for the rheological models shown in Figure 11 Heat production (106 W m3)
Layer
Rheology
Conductivity (W m1 K1)
Sediments (models A,B,C) Sediments (model D) Upper crust Lower crust Lithospheric mantle
Quartzite
2
0.2
Quartzite
1.4
1.8
Quartzite Diorite Olivine
2.6 2.6 3.1
1.88 0.5 0
The magnitude of postrift tectonic subsidence of aborted rifts and passive margins is a function of the thermal anomaly that was introduced during their rifting stage and the degree to which the lithospheric mantle was thinned. Most intense anomalies develop during crustal separation, particularly when plume assisted and asthenospheric melts well up close to the surface. The magnitude of thermal anomalies induced by rifting that did not progress to crustal separation depends on the magnitude of crustal stretching (-factor) and lithospheric mantle attenuation (-factor), the thermal regime of the asthenosphere, the volume of melts generated and whether these intruded the lithosphere and destabilized the Moho (Figure 4). After 60 My, about 65%, and after 180 My about 95% of a deep-seated thermal anomaly associated with a major pullup of the asthenosphere–lithosphere boundary have decayed (mantle-plume model (Figure 4)). Thermal anomalies related to intralithospheric intrusions (tensional-failure model (Figure 4)) have apparently a faster decay rate. For instance, the Mid-Jurassic North Sea rift dome had subsided below the sea level, 20–30 My after its maximum uplift, that is, well before crustal extension had terminated (Ziegler, 1990b; Underhill and Partington, 1993). 6.11.2.1.4.(ii) Stretching factors derived from quantitative subsidence analyses The thick-
ness of the postrift sedimentary column that can accumulate in passive margin basins and in thermalsag basins above aborted rifts is not only a function of the magnitude of the lithospheric mantle and crustal attenuation factors and , but also of the crustal
499
density and the water depth at the end of their rifting stage, as well as of the density of the infilling postrift sediments (carbonates, evaporites, clastics). Moreover, it must be kept in mind that during the postrift cooling process the flexural rigidity of the lithosphere increases gradually, resulting in the distribution of the sedimentary load over progressively wider areas. Quantitative postrift subsidence analyses of extensional basins, neglecting intraplate stresses, are thought to give a measure of the thermal contraction of the lithosphere and, conversely, of lithospheric stretching factors, as suggested by the McKenzie (1978) model and its early users (e.g., Sclater and Christie, 1980; Barton and Wood, 1984). Such analyses, assuming Airy isostasy, yield often considerably larger -factors than indicated by the populations of upper crustal faults (Ziegler, 1983, 1990b; Watcharanantakul and Morley, 2000). This discrepancy is somewhat reduced when flexural isostasy is assumed (Roberts et al., 1993). However, as, during rifting, attenuation of the lithosphere is not only achieved by its mechanical stretching but also by convective and thermal upward displacement of the asthenosphere–lithosphere boundary, the magnitude of a thermal anomaly derived from the postrift subsidence of a basin cannot be directly related to a mechanical stretching factor. Moreover, intraplate compressional stresses and phase transformations in the lower crust and lithospheric mantle have an overprinting effect on postrift subsidence and can cause significant departures from a purely thermal subsidence curve (Cloetingh and Kooi, 1992; Van Wees and Cloetingh, 1996; Lobkovsky et al., 1996). Furthermore, during rifting the ascent of mantlederived melts to the base of the crust can cause destabilization of the Moho, magmatic inflation of the crust and its metasomatic reactivation and secondary differentiation (Morgan and Ramberg, 1987; Mohr, 1992; Stel et al., 1993; Watts and Fairhead, 1997). For instance, lower crustal velocities of 6.30–7.2 km s1 and densities of 3.02 103 kg m3 characterize the highly attenuated crust of the Devonian Dniepr-Donets rift almost up to the base of its syn-rift sediments, probably owing to its syn-rift permeation by mantle-derived melts (Yegorova et al., 1999; Stephenson et al., 2001). Moreover, accumulation of very thick postrift sedimentary sequences can cause phase transformation of crustal rocks to granulite facies, with the resulting densification of the crust accounting for accelerated basin subsidence. This process can be amplified in the
500
Tectonic Models for the Evolution of Sedimentary Basins
cases of eclogite transformation of basaltic melts that were injected into the lithospheric mantle or that had accumulated at its base (Lobkovsky et al., 1996). However, it is uncertain whether large-scale eclogite formation can indeed occur at crustal thicknesses of 30–40 km (Carswell, 1990; Griffin et al., 1990). Although phase transformations, entailing an upward displacement of the geophysical Moho, are thought to occur under certain conditions at the base of very thick Proterozoic cratons (increased confining pressure due to horizontal intraplate stresses and/or ice load, possible cooling of asthenosphere, Cloetingh and Kooi, 1992), it is uncertain whether physical conditions conducive to such transformations can develop in response to postrift sedimentary loading of palaeorifts and passive margins as, for example, suspected for East Newfoundland Basin (Cloetingh and Kooi, 1992) and the West Siberian Basin (Lobkovsky et al., 1996). As rifting processes can take place intermittently over very long periods of time (e.g., Norwegian– Greenland Sea rift: Ziegler, 1988), thermal anomalies introduced during early stretching phases start to decay during subsequent periods of decreased extension rates. Thus, the thermal anomaly associated with a rift may not be at its maximum when crustal stretching terminates and the rift becomes inactive. Similarly, late rifting pulses and/or regional magmatic events may interrupt and even reverse lithospheric cooling processes (e.g., Palaeocene thermal uplift of northern parts of Viking Graben due to Iceland plume impingement: Ziegler, 1990a, 1990b; Nadin and Kusznir, 1995). Therefore, analyses of the postrift subsidence of rifts that have evolved in response to multiple rifting phases spread over a long period need to take their entire rifting history into account. Intraplate compressional stress, causing deflection of the lithosphere in response to lithospheric folding, can seriously overprint the thermal subsidence of postrift basins. The example of the Plio-Pleistocene evolution of the North Sea shows that the buildup of regional compressional stresses can cause a sharp acceleration of postrift subsidence (Figure 10) (Cloetingh, 1988; Cloetingh et al., 1990; Kooi and Cloetingh, 1989; Kooi et al., 1991; Van Wees and Cloetingh, 1996). Similar contemporaneous effects are recognized in North Atlantic passive-margin basins (Cloetingh et al., 1987, 1990) as well as in the Pannonian Basin (Horva´th and Cloetingh, 1996). Also the late Eocene accelerated subsidence of the Black Sea Basin can be attributed to the buildup of a regional compressional stress field (Robinson et al., 1995; Cloetingh et al., 2003).
E17-1
A12-1 G17-1
F18-1
F14-1 P15-1
Q8-2
0m
500
P
Tr
Jurassic
Cretaceous
Tertiary
Figure 10 Tectonic subsidence curves of southern North Sea, showing accelerated subsidence during PlioPleistocene. Postrift stage starts during the Cretaceous. Modified from Kooi H, Hettema M, and Cloetingh S (1991). Lithospheric dynamics and the rapid Pliocene-Quaternary subsidence in the North Sea basin. Tectonophysics 192: 245–259.
At present, many extensional basins are in a state of horizontal compression as documented by stress indicators summarized in the World Stress Map (Zoback, 1992) and the European Stress Map (e.g., Go¨lke et al., 1996). The magnitude of stress-induced vertical motions of the lithosphere during the postrift phase, causing accelerated basin subsidence and tilting of basin margins, depends on the ratio of the stress level and the strength of the lithosphere inherited from the syn-rift phase. Moreover, horizontal stresses in the lithosphere strongly affect the development of salt diapirism, accounting for local subsidence anomalies (Cloetingh and Kooi, 1992) and have a strong impact on the hydrodynamic regime of rifted basins
Tectonic Models for the Evolution of Sedimentary Basins
(Van Balen and Cloetingh, 1993, 1994) by contributing to the development of overpressure, as seen in parts of the Pannonian Basin (Van Balen et al., 1999). Glacial loading and unloading can further complicate postrift lithospheric motions, a better insight into the nature of which is required (Solheim et al., 1996). These considerations indicate that stretching factors derived from the subsidence of postrift basins must be treated with reservations. Nevertheless, quantitative subsidence analyses, combined with other data, are essential for the understanding of postrift subsidence processes by giving a measure of the lithospheric anomaly that was introduced during the rifting stage of a basin and by identifying deviations from purely thermal cooling trends. 6.11.2.1.4.(iii) Postrift compressional reactivation potential Palaeo-stress analyses give evidence for
changes in the magnitude and orientation of intraplate stress fields on time scales of a few million years (Letouzey, 1986; Philip, 1987; Bergerat, 1987; De`zes et al., 2004). Thus, in an attempt to understand the evolution of a postrift basin, the effects of tectonic stresses on subsidence must be separated from those related to thermal relaxation of the lithosphere (Cloetingh and Kooi, 1992). In response to the buildup of far-field compressional stresses, rifted basins, characterized by a strongly faulted and thus permanently weakened crust, are prone to reactivation at all stages of their postrift evolution, resulting in their inversion (Ziegler, 1990a; Ziegler et al., 1995, 1998, 2001). Only under special condition can gravitational forces associated with topography around a basin cause its inversion (Bada et al., 2001). Rheological considerations indicate that the lithosphere of thermally stabilized rifts, lacking a thick postrift sedimentary prism, is considerably stronger than the lithosphere of adjacent unstretched areas (Ziegler and Cloetingh, 2004). This contradicts the observation that rift zones and passive margins are preferentially deformed during periods of intraplate compression (Ziegler et al., 2001). However, burial of rifted basins under a thick postrift sequence contributes by thermal blanketing to weakening of their lithosphere (Stephenson, 1989; Van Wees, 1994), thus rendering it prone to tectonic reactivation. In order to quantify this effect and to assess the reactivation potential of conjugate simple-shear margins during subduction initiation, their strength evolution was modeled and compared to that of oceanic crust (Ziegler et al., 1998). By applying a 1-D two-
501
layered lithospheric stretching model, incorporating the effects of heat production by the crust and its sedimentary thermal blanketing (Table 2), the thermomechanical evolution of the lithosphere was analyzed in an effort to predict its palaeo-rheology (Van Wees et al., 1996; Bertotti et al., 1997). For modeling purposes, a time frame of 100 My was chosen. Of this, the first 10 My (between 100 and 90 My in Figure 11) correspond to the rifting stage, culminating in separation of the conjugate upper and lower plate margins, and the following 90 My to the seafloor spreading stage during which oceanic lithosphere is accreted to the diverging plates. For modeling purposes it was assumed that the prerift crustal and mantle-lithosphere thickness are 30 and 70 km, respectively, and that at the end of the rifting stage the upper plate margin has a crustal thickness of 15 km ( ¼ 2) and a remaining lithospheric mantle thickness of 7 km ( ¼ 10) ( and are, respectively, the crustal and subcrustal stretching factor (Royden and Keen, 1980)), whilst the lower plate margin has a crustal thickness of 10 km ( ¼ 3) and a lithospheric mantle thickness of 63.6 km ( ¼ 1.1) (Figures 11(a) and 11(b)). Results show that through time the evolution of strength envelopes for lower and upper plate passive margins differs strongly. In principle, during rifting, increased heating of the lithosphere causes its weakening; this effect is most pronounced at the moment of crustal separation. However, upper and lower plate margins show a very different evolution, both during the rifting and postrift stage. At the moment of crustal separation, upper plate margins are very weak due to strong attenuation of the mantle-lithosphere and the ascent of the asthenospheric material close to the base of the crust. During the postrift evolution of such a margin, having a crustal thickness of 15 km, the strength of the lithosphere increases gradually as new mantle is accreted to its base and cools during the reequilibration of the lithosphere with the asthenosphere (Figures 11(a) and 12(a)). In contrast, the evolution of a sediment starved lower plate margin with a crustal thickness of 10 km is characterized by a syn-rift strength increase due to extensional unroofing of the little attenuated mantle-lithosphere; the strength of such a margin increases dramatically during the postrift stage due to its progressive cooling (Figures 11(b) and 12(b)). At the timeframe of 0 My, a sediment starved lower plate margin is considerably stronger than the conjugate upper plate margin for which a sedimentary cover of about 7 km was assumed. The strength evolution of an upper plate margin is initially controlled
502
Tectonic Models for the Evolution of Sedimentary Basins
(a)
Strength (MPa)/temperature 1500
1000
500
0
–500
–1000
Ext
Sediments
Comp
Crust
20
40
60
100 Ma
90 Ma
60 Ma
30 Ma
0 Ma
(b) 1500
1000
500
0
–500
–1000
Ext
Crust
Comp
20
40
60
100 Ma
60 Ma
30 Ma
0 Ma
1500
Ext
1000
500
0
–500
–1000
(c)
90 Ma
Comp 20
40
60
90 Ma
30 Ma
0 Ma
1500
1000
Ext
500
0
–500
–1000
(d)
60 Ma
Sediments
Comp 20
40
60
90 Ma
60 Ma
30 Ma
0 Ma
Figure 11 Depth-dependent rheological models for the evolution of lower plate and upper plate passive margins and oceanic lithosphere. For modeling parameters see Tables 2 and 3. (a) Upper plate passive margin ( ¼ 2, ¼ 10), characterized at end of postrift phase by complete sediment fill of accommodation space (s ¼ 2100 kg m3, 7.084 km sediments). (b) Lower plate passive margin ( ¼ 3, ¼ 1.1), marked by sediment starvation at end of postrift phase (1 km sediments). (c) Oceanic lithosphere with a thin sedimentary cover of 1 km. (d) Oceanic lithosphere with a gradually increasing sedimentary cover reaching a maximum of 15 km. Modified from Ziegler PA, Van Wees J-D, and Cloetingh S (1998). Mechanical controls on collision-related compressional intraplate deformation. Tectonophysics 300: 103–129.
Tectonic Models for the Evolution of Sedimentary Basins
503
that it approaches the strength of a sediment filled lower plate margin (Figure 12). To test the effects of sediment infill and thermal blanketing on the strength evolution of upper and lower plate passive margins, a wide range of models were run assuming that sediments completely fill the tectonically created accommodation space (sediment overfilled (Figure 12)), adopting different sediment densities and corresponding sediment thickness variations (Table 3). Results show that a thick syn- and postrift sedimentary prism markedly reduces the integrated strength of a margin. However, despite
by the youthfulness of its lithospheric mantle and its thicker crust, and later by the thermal blanketing effect of sediments infilling the available accommodation space. On the other hand, the strength of oceanic lithosphere, that is covered by thin sediments only, increases dramatically during its 90 My evolution and ultimately exceeds the strength of both margins, even if these are sediment starved (Figures 11(c) and 12). However, the strength of 90 My old oceanic lithosphere that has been progressively covered by very thick sediments is significantly reduced (Figure 11(d)) to the point
(a) e
pher
lithos
40
30
)
re (delta
osphe eanic lith
Oc
20
Sed
100
90
80
70
ed tarv
nt s
ime
60
Sediment overfilled
50
40
30
20
10
10
0
Integrated strength (TN m–1)
nic Ocea
0
(b) re
phe
40
Oce
Sed
ime
a)
re (delt
osphe
ic lith Ocean
ed tarv
nt s
30
Sediment overfilled
20
10
Integrated strength (TN m–1)
hos
lit anic
0 100
90
80
70
60
50 Age (Ma)
40
30
20
10
0
Figure 12 Integrated compressional strength evolution of sediment-starved and sediment-filled (a) upper plate, and (b) lower plate passive margins, compared to the integrated strength evolution of oceanic lithosphere with thin and thick sediment cover as in Figure 11. Shaded areas demonstrate strong sensitivity of integrated strength to sediment infilling, ranging from sediment starvation (highest strength values) to complete fill of accommodation space (dark shading, lowest strength values). Curves in dark shaded area correspond to different sediment densities and related range in sediment thickness (see Table 3). Modified from Ziegler PA Van, Wees J-D, and Cloetingh S (1998). Mechanical controls on collisionrelated compressional intraplate deformation. Tectonophysics 300: 103–129.
504
Tectonic Models for the Evolution of Sedimentary Basins
Table 3
Sediment-overfilled scenario (up to water surface), adopted in Figure 12 Accumulated sediment fill (m) Tectonic air-loaded subsidence (m)
s ¼ 2100 kg m3
s ¼ 2500 kg m3
s ¼ 2650 kg m3
Upper plate
Syn-rift Postrift
1474 2375
4396 7084
7065 11384
9148 14739
Lower plate
Syn-rift Postrift
2599 3332
7752 9939
7752 15972
16130 20679
the strong sediment fill effect on the integrated strength, earlier identified first-order differences between upper and lower plate margins remain. Compared to oceanic lithosphere, both with a 1 km sedimentary cover (Figures 11(c) and 12) and a 15 km thick cover (Figure 11(d)), a sediment overfilled upper plate margin (applying S ¼ 2100 kg m3) is characterized by lower integrated strength values throughout its evolution. However, for a lower plate margin conditions are dramatically different. Up to 20–70 My after rifting, the lower plate integrated strength values are significantly higher than those for oceanic lithosphere, both with a thin and a thick sedimentary cover. From these strength calculations it is evident that at any stage the upper plate margin is weaker than oceanic lithosphere and the conjugate lower plate margin. This suggests that the upper plate margin is the most likely candidate for compressional reactivation and the initiation of a subduction zone. For realistic sediment density and infill values (S 2500 kg m3), also the upper plate margin tends to grow stronger in time, indicating that after a prolonged postrift stage (>70 My) localization of deformation and subduction along such a margin, instead of on the adjacent continent, should be facilitated by weakening mechanisms that are not incorporated in our standard rheological assumptions for the lithosphere, such as preexisting crustal and mantle discontinuities and the boundaries between old and newly accreted lithospheric mantle. The modeling shows that the compressional yield strength of passive margins can vary considerably, depending on their lithospheric configuration, sedimentary cover and thermal age. Lower plate margins, at which much of the old continental mantle-lithosphere is preserved, are considerably stronger than upper plate margins at which asthenospheric material has been accreted to the strongly attenuated old mantle-lithosphere. Although mature oceanic lithosphere is characterized by a high compressional yield
strength, it can be significantly weakened in areas where thick passive-margin sedimentary prisms or deep-sea fans prograde onto it (e.g., Gulf of Mexico: Worrall and Snelson, 1989; Niger Delta: Doust and Omatsola, 1989; Bengal fan: Curray and Moore, 1971). 6.11.2.1.5 Finite strength of the lithosphere in extensional basin formation
In recent approaches to extensional basin modeling the implementation of finite lithospheric strengths is an important step forward (see Section 6.11.3 for a review). Advances in the understanding of lithospheric mechanics (e.g., Ranalli, 1995; Burov and Diament, 1995) demonstrate that early stretching models which assumed zero lithospheric strength during rifting are not valid. Moreover, the notion of a possible decoupling zone between the strong upper crust and the strong lithospheric upper mantle (see Section 6.11.3.1) is also important in the context of extensional basin formation. During the past decades the relative importance of pure shear (McKenzie, 1978) and simple shear (Wernicke, 1985) mode of extension has been a matter of debate. Figure 13 suggests that in the presence of a weak lower crustal layer decoupling of the mechanically strong upper Upper crust MSC
Lower crust Mantle lithosphere
MSL Asthenosphere Strong
Weak
Figure 13 Kinematic model for extension of rheologically stratified lithosphere. See strength profile on left side of diagram. MSC and MSL indicate the base of the mechanically strong crust and mechanically strong lithosphere, respectively. Reston TJ (1990) The lowest crust and the extension of the continental lithosphere; kinematic analysis of BIRPS deep seismic data. Tectonics 9: 1235–1248.
Tectonic Models for the Evolution of Sedimentary Basins
crust from the even stronger upper lithospheric mantle, the zone and symmetry of upper crustal extension, does not necessarily have to coincide with the zone and symmetry of lithospheric mantle extension. This is particularly the case if the upper crust is weakened by preexisting discontinuities favouring its simple shear extensional deformation (Ziegler, 1996b). A better appreciation of the role played by rheology during basin formation and the advent of corresponding modeling capabilities during the past few years has increasingly shifted the focus of attention away from these end members of lithospheric extension. Finite element models have explored the large-scale implications of a finite lithospheric strength and particularly its sensitivity to the presence of fluids in the crust and lithospheric mantle (Braun and Beaumont, 1989; Dunbar and Sawyer, 1989; Govers and Wortel, 1993; Bassi, 1995). These dynamic models, which require intensive computing, are expensive to run, and thus are not suitable for an industry user-oriented environment. However, they have provided the background for a more userfriendly class of kinematic models targeted at modeling rift-shoulder uplift and basin fill (Cloetingh et al., 1995c). These kinematic models invoke the concept of lithospheric necking around one of its strong layers during extensional basin formation (see Braun and Beaumont, 1989; Kooi et al., 1992; Spadini et al., 1995b). Figure 7 illustrates the basic features of these models and their relation to the strength distribution within the lithosphere. In the presence of a strong layer in the subcrustal mantle, the level of lithospheric necking is deep, inducing pronounced rift-shoulder topography. This type of response is to be expected if extension affects cold and correspondingly strong intracontinental lithosphere; it is commonly observed in intracratonic rifts and rifted margin such as, for example, the Red Sea and the Trans-Antarctic Mountains (Cloetingh et al., 1995c). For Alpine/Mediterranean basins, which developed on a weak lithosphere with a thickened crust, the necking level is generally located at shallower depths (Figure 14). An example of such a situation is found in the Pannonian Basin (Van Balen et al., 1995, 1999; Horva´th and Cloetingh, 1996) where the necking level is located at depths between 5 and 10 km within the upper crust. In this case the strength of the lithospheric upper mantle has decreased to almost zero (see also Figure 7). Important exceptions to this general pattern do occur, however, such as in the Southern Tyrrhenian Sea, for which our modeling indicates a deep necking level to fit observational
505
data (Spadini et al., 1995a, 1995b, Spadini and Podladchikov, 1996). This is primarily attributed to the fact that the Southern Tyrrhenian Basin developed essentially on Hercynian lithosphere with significant bulk strength of its mantle component. In other areas, this concept has led to a better understanding of lateral variations in basin structure and sedimentary fill within one and the same basin. This is exemplified by the Black Sea (see Section 6.11.5), where modeling deciphered important variations in necking level and thermal conditions between its eastern and western sub-basins that can be attributed to differences in the timing and mode of their development (Spadini et al., 1996, 1997, Robinson et al., 1995, Cloetingh et al., 2003). The interdependence of the necking depth and such parameters as prerift crustal and lithospheric thicknesses, the EET of the lithosphere and strain rates was investigated in a comparative study on several basins, using the same modeling technology (Cloetingh et al., 1995b, 1995c). Results of this study are summarized in Figure 14, illustrating that for Alpine/Mediterranean basins the position of the necking level depends primarily on the prerift crustal thickness and strain rate, whereas the key controlling factors in intracratonic rifts appear to be the prerift lithospheric thickness and strain rate. This figure also illustrates where other basin formation processes have played a role. For instance, the Saudi Arabia–Red Sea margin (point 7), is characterized by the presence of plume-related activity in the upper mantle that explains the systematic misfit of this case. The importance of the prerift lithosphere rheology for the subsequent basin geometry and the patterns of vertical motions are evident from these models. Moreover, they demonstrate that the better we are able to constrain the prerift evolution of an area, the greater the chance we have to define the parameter range that has to be to adopted for large-scale syn-rift mechanics. 6.11.2.1.6 Rift-shoulder development and architecture of basin fill
As discussed above, the finite strength of the lithosphere has important implications for the crustal structure of extensional basins and the development of accommodation space in them. The development of significant rift-shoulder topography in response to lithospheric extension has drawn attention to the need to constrain the coupled vertical motion of the shoulders and the subsiding basin (Kusznir and Ziegler, 1992). Whereas the standard approach in basin analysis focused until recently primarily on
Tectonic Models for the Evolution of Sedimentary Basins
(b) 45
40
40
35
35
30
30
Z-neck (km)
(a) 45
Z-neck (km)
506
25 20 15
25 20 15
10
10
5
5
0
0 25
30 35 40 45 Prerift crustal thickness (km)
50
40
40
35
35
30
30
Z-neck (km)
(d) 45
Z-neck (km)
(c) 45
25 20 15
100 125 150 175 200 Prerift lithosphere thickness (km)
225
75
100 125 150 175 200 Prerift lithosphere thickness (km)
225
25 20 15
10
10
5
5
0
75
0 5
15
25
35 45 EET (km)
55
65
Figure 14 Correlation diagrams for the relationship between: (a) necking depth and prerift crustal thickness; (b) necking depth and prerift lithosphere thickness; (c) EET and necking depth; and (d) necking depth and strain rate. Squares and diamonds indicate data from Alpine/Mediterranean basins and intracratonic rifts, respectively. Numbers refer to the following basins: 1 Gulf of Lion; 2 Valencia trough; 3 southern Tyrrhenian Sea; 4 Pannonian basin; 5 North Sea; 6 Baikal rift; 7 Saudi Arabia Red Sea; 8 Trans-Antarctic Mountains; 9 Barents Sea; 10 East African Rift; 11 western Black Sea; 12 eastern Black Sea. Modified from Cloetingh S, Durand B, and Puigdefabregas C (eds.) (1995d) Special Issue on Integrated Basin Studies (IBS) – A European Commission (DGXII) project. Marine and Petroleum Geology 12(8): 787–963.
the subsiding basin, treating sediment supply as an independent parameter, necking models highlight the need for linking sediment supply to the riftflank uplift and erosion history. To this end, a twofold approach was followed. The first research line aimed at constraining the predicted uplift histories by geothermochronology. Modeling of the distribution of fission-track length permits to backstack the eroded sediments from their present position in the basin to their source on the rift shoulder in an effort to obtain a better reconstruction of the rift-shoulder geometries (e.g., Van der Beek et al., 1995; Rohrman et al., 1995). This has led to a better understanding of the timing and magnitude of rift-shoulder uplift, for example, on the Norwegian margin, shedding light on the observed relationship between on-shore uplift and the presence of thick Late Cenozoic sedimentary
wedges in the adjacent offshore basin (see also Section 6.11.4). A second research line focused on the development of a model for basin fill simulation, integrating the effects of rift-shoulder erosion through hill-slope transport and river incision with sediment deposition in the basin. As illustrated in Figure 15 (see Van Balen et al., 1995), these models predict the progradation of sedimentary wedges into extensional basins and the development of hinterland basins having a distinctly different stratigraphic signature than predicted by standard models, which invoke stretching and postrift flexure, and which are commonly applied in the existing packages. Testing of this new model against a number of rifted margins around the world demonstrates that erosion of the rift-shoulder topography, created during extension of a lithosphere
Tectonic Models for the Evolution of Sedimentary Basins
507
Rift shoulder (drainage divide)
(a) Model boundary
Shore-line
Hinterland basin
Prerift elevation Basement
Sedimentary basin
Advection/diffusion Diffusion (b)
+ Erosion of topsets
Eroding and retreating rift shoulder
–
w
Flexuralebound = Relative sealevel change
Basinward migrating shore-line Prograding sedimentary wedge
Figure 15 (a) Cartoon of the rift shoulder erosion model. The isostatic response to necking of the lithosphere during extension causes a flexurally supported rift shoulder. In a landward direction the rift shoulder is flanked by a flexural down warp, the hinterland basin. The erosion products of the rift shoulder are transported to the offshore rifted basin and the hinterland basin. When the hinterland basin is completely filled, the sediments are taken out at the location indicated by ‘model boundary’. These sediments are reintroduced into the model at the shoreline position in the offshore basin, establishing conservation of mass in the model. (b) Cartoon illustrating the mechanism of initial postrift coastal offlap at passive margins. Flexural rebound in response to erosional unloading at the rift shoulder causes uplift extending far into the offshore basin. The uplift causes erosional truncation of the topsets of the sedimentary wedge. Coastal onlap will occur when the rate of erosion-induced uplift is less than the subsidence caused by sedimentary loading, thermal contraction and the flexural response to the increase in rigidity. w shows the instantaneous flexural uplift (þ) and subsidence () pattern caused by rift shoulder erosion. Modified from Van Balen RT, Van der Beek PA, and Cloetingh S (1995) The effect of rift shoulder erosion on stratal patterns at passive margins: implications for sequence stratigraphy. Earth and Planetary Science Letters 134: 527–544.
with a finite strength, eliminates to a large extent the need to invoke eustatic sea-level changes to explain the most commonly encountered large-scale stratigraphic features of rifted basins and associated hinterland basins. On the scale of individual half-grabens, faulting exerts the main control on the basin architecture. Modeling studies have focused on the coupling of fault block tilting and the flexural behaviour of the extending lithosphere, as illustrated in Figure 16. In a first step, modeling technology was developed to quantify the thermal effects of faulting on an extending lithosphere (Ter Voorde and Bertotti, 1994). Subsequently, models were developed that link the evolution of individual half-grabens to the extensional response of the deeper lithospheric (Ter Voorde and Cloetingh, 1996). The second step aimed at validating and testing these models against closely constrained natural examples, such as the well-exposed Mesozoic Southern Alpine rifted margin, where extensive radiometric dating and the
application of Ar/Ar laser-probing techniques permitted to constrain with high accuracy the thermal evolution of the extending lithosphere at sub-basin and basin-wide scales (Bertotti and Ter Voorde, 1994). Testing and validation of the stratigraphic modeling component was carried out by a case-history study of the Oseberg field in the Norwegian part of the Viking graben (Ter Voorde et al., 1997). This study demonstrated the need to establish a regional framework, linking the mechanics of the Oseberg block to the crustal evolution of adjacent areas, as a prerequisite for more detailed reservoir modeling. Considering the notion that in syn-rift basins different spatial and temporal scales are by their very nature linked, ignorance of constraints and structural information on surrounding areas will severely limit the quality of reservoir modeling. Modeling provides a quantitative tool to assess tectonic controls on synrift depositional sequences at a sub-basin scale (see also Nottvedt et al., 1995). The amount of footwall uplift of an individual fault block appears to depend
508
Tectonic Models for the Evolution of Sedimentary Basins
Basinwide scale
Large-scale faulting Regional flexural deformation
Moho uplift due to distributed thinning
Fault-block scale
Local scale Footwall uplift
Unconformity Onlap
Figure 16 Finite difference model developed for modeling of extensional tilted fault-blocks and large-scale deformation of the lithosphere. Modified from Ter Voorde and Cloetingh S (1996) Numerical modelling of extension in faulted crust: effects of localized and regional deformation on basin stratigraphy. Geological Society, London, Special Publications 99: 283–296.
directly on the lithospheric necking level, controlling the large-scale response of the lithosphere to extension, and, therefore, should not only be attributed to factors restricted to the sub-basin scale. 6.11.2.1.7 Transformation of an orogen into a cratonic platform: the area of the European Cenozoic Rift System
The European Cenozoic Rift System (ECRIS) extends from the shores of the North Sea to the Mediterranean and transects the French and German parts of the Late Palaeozoic Variscan Orogen (Figure 17) (Ziegler, 1990b, 1994; De`zes et al., 2004). In the ECRIS area, the deeply eroded Variscan Orogen was, and partly is still covered by extensive Late Permian and Mesozoic platform sediments. These were disrupted during the evolution of ECRIS owing to rift-related uplift of the Rhenish and Bohemian Massifs, the Vosges-Black Forest Arch and the Massif Central, thus exposing parts of the Variscan Orogen and providing insight into its architecture. This offers a unique opportunity to assess processes controlling the transformation of the Variscan Orogen into a cratonic platform (see Ziegler et al., 2004 for details and comprehensive references). Main elements of ECRIS are the Lower Rhine (Roer Valley), Hessian, Upper Rhine, Limagne, Bresse, and Eger grabens. The Lower Rhine and Hessian grabens transect the external parts of the Varsican Orogen, the Rheno-Hercynian thrust belt. The Upper Rhine, Bresse, and Limagne grabens cross-cut the internal
parts of the Variscan Orogen (Figure 18), consisting of the Mid-German Crystalline Rise and the SaxoThuringian, Bohemian-Armorican and MoldanubianArverno-Vosgian zones, that are characterized by basement-involving nappes and a widespread syn- and postorogenic magmatism. The Eger graben is superimposed on the eastern parts of the Saxo-Thuringian zone. In the ECRIS area, the depth to Moho varies between 24 and 30 km and increases away from it to 34–36 km and more (Figure 19) (Ziegler and De`zes, 2006). The thickness of the lithosphere decreases from about 100–120 km in the Bohemian Massif and along the southern end of the Upper Rhine Graben to 60–70 km beneath the Rhenish Massif and Massif Central, and appears to increase to some 120 km or more in the Western Netherlands and beneath the Paris Basin (Babuska and Plomerova, 1992, 1993; Sobolev et al., 1997; Goes et al., 2000a, 2000b). Mantle tomography images beneath the ECRIS area a system of upper asthenospheric low velocity anomalies, interpreted as plume heads that have spread out above the 410 km discontinuity (Spakman and Wortel, 2004; Sibuet et al., 2004), and from which secondary, relatively weak plumes rise up under the Rhenish Massif (Ritter et al., 2001) and the Massif Central (Granet et al., 1995). The present crustal and lithospheric configuration of Western and Central Europe bears no relationship to the major structural units of the Variscan Orogen, but shows close affinities to ECRIS (Ziegler and
Tectonic Models for the Evolution of Sedimentary Basins
4°
0° 52°
8°
A
16°
12° HG
LRG
f
TF
ssi
Ma
EG
h
nis
e Rh
509
Bohemian Massif
OW FP
Paris Paris basin
URG
48° Armorican
BF
VG
Massif
s e Ba
48°
S
JU
LG
in
A'
R
AA
BG
ass Mol
P Massif central
L A
44°
EN
AP
44°
S NE NI
PY
RE
NE
0°
ES
4°
8°
12°
Figure 17 Location map of ECRIS in the Alpine foreland, showing Cenozoic fault systems (black lines), rift-related sedimentary basins (light gray), Variscan massifs (cross pattern) and Cenozoic volcanic fields (black). Interrupted barbed line: Alpine deformation front. BF Black Forest, BG Bresse Graben, EG Eger (Ore) Graben, FP Franconian Platform; HG Hessian grabens, LG Limagne Graben, LRG Lower Rhine (Roer Valley) Graben, URG Upper Rhine Graben, OW Odenwald; VG Vosges. Modified from De`zes et al., (2004).
De`zes, 2006). However, development of the Variscan Orogen involved major crustal shortening and subduction of substantial amounts of supra-crustal rocks, continental and oceanic crust and lithospheric mantle (Ziegler et al., 1995, 2004). By analogy with modern examples, such as the Alps (Schmid et al., 1996, 2004; Stampfli et al., 1998), the Variscan Orogen must have been characterized at the time of its end-Westphalian consolidation (305 Ma) by a significantly thickened crust and lithosphere. Subsequently, its orogenically destabilized lithosphere reequilibrated with the asthenosphere so that by Late Mesozoic time cratonic crustal and lithospheric thicknesses of about 28–35 km and 100–120 km, respectively, were established. Processes controlling postorogenic modification of the Variscan lithosphere have been variably attributed to Permo-Carboniferous slab detachment, delamination of the lithospheric mantle, crustal extension, and plume activity and subsequent thermal relaxation of the lithosphere controlling the subsidence of an intracratonic basin system (Lorenz and Nicholls, 1984; Ziegler, 1990b; Henk, 1999; Henk et al., 2000; Van
Wees et al., 2000; Prijac et al., 2000). In order to assess the relative importance of processes contributing toward the postorogenic transformation of the Variscan lithosphere to its craton-like configuration, the evolution of the different Varsican units transected by ECRIS was reviewed, crustal reflection-seismic profiles inspected and quantitative subsidence curves developed for selected wells penetrating the sedimentary cover of the Variscan basement and these compared to a theoretical thermal decay curve. During the last 310 My, the tectonic setting of the Variscan domain changed repeatedly. Following the late Westphalian (305 Ma) consolidation of the Variscan Orogen its Stephanian–Early Permian collapse (305–280 Ma) was controlled by wrench faulting and associated magmatic activity. During Late Permian to Cretaceous times, large parts of the Variscan domain were incorporated into an intracratonic basin system. This was affected during the latest Cretaceous and Palaeocene by an important pulse of intraplate compression that relates to early phases of the Alpine orogeny. At the same time, an array of
510
Tectonic Models for the Evolution of Sedimentary Basins
0°
4°
8°
12°
14°
Rheno-Hercynian Zone Rheno-Hercynian zone
52°
52°
ryst. rise an c erm G iann zzoonnee inggia id riln M hu T o-
Sa x
A
Bohemian Bohemian zone Zone
Moldanubian Moldanubian Zone zone
48°
48° ArmoricanZone zone Armorican
A'
ArvernoArvernoVosgian Vosgian Zone zone 44°
Ligerian Ligerian zone Zone
0°
4°
8°
12°
Figure 18 Variscan tectonic framework with superimposed ECRIS fault pattern. Modified from Ziegler PA, Schumacher ME, De´zes P, van Wees J-D, and Cloetingh S (2004) Post-Variscan evolution of the lithosphere in the Rhine Graben area: constraints from subsidence modelling. In: Wilson M, Neumann E-R, Davies GR, Timmerman MJ, Heeremans M, and Larsen BT (eds.) Geological Society, London, Special Publications, 223: Permo-Carboniferous Magmatism and Rifting in Europe, pp. 289–317. London: Geological Society, London.
mantle plumes impinged on the lithosphere of Western and Central Europe, such as the NE Atlantic and Iceland plumes and precursors of the Massif Central and Rhenish plumes. The resulting increase in the potential temperature of the asthenosphere caused a renewed destabilization of the lithosphere, as evidenced by Palaeocene injection of mafic dykes in the Massif Central, Vosges-Black Forest and Bohemian Massif, reflecting low-degree partial melting of the lithospheric thermal boundary layer at depths of 60–100 km. Starting in mid-Eocene times, ECRIS developed in the foreland of the evolving Alpine and Pyrenean orogens, with crustal extension and continued plume activity causing further destabilization of its lithosphere–asthenosphere system (Ziegler, 1990b; De`zes et al., 2004, 2005; Ziegler and De`zes, 2006, 2007). In terms of defining boundary conditions for modeling the postorogenic evolution of the Variscan
lithosphere, information on its Late Carboniferous (305 Ma), late Early Permian (280 Ma), endCretaceous (65 Ma) and present configuration is required. Whilst present crustal thicknesses are closely constrained, control on lithospheric thickness is less reliable. The Late Carboniferous, late Early Permian and end-Cretaceous–Palaeocene lithospheric configurations can, however, only be inferred from circumstantial evidence. In this respect, vertical movements of the crust, derived from its sedimentary cover, provide constraints for assessing the postorogenic evolution of the Variscan lithosphere. Below the Late Palaeozoic and Mesozoic evolution of the ECRIS area is summarized, results of quantitative and forward modeled subsidence analyses on selected wells presented, and a model proposed for the postorogenic evolution of the lithosphere (see Ziegler et al., 2004). The time scales of Menning et al., (2000) and Menning (1995) were
Tectonic Models for the Evolution of Sedimentary Basins
0° 52°
4°
12°
8° 30
34
16° 32
30
36
32
511
30 32
38
36 36
34 30 32
30
38 40
26
48°
34
26
48°
36
28 40 32 30
30
30
44°
28 30
34 28 30
28 26 26
32 50
0°
14
4°
44°
30
48 40
20
30 14
22 20
30
28 22
26
12
8°
12°
8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 Kilometres
Projection: Lambert Azimuthal Equal Area; Centre:04i.00″/48i:00″; Region:W/E/N/S = 350i/28i/62i/34i; Ellipsoide wgs-84
Figure 19 Depth map of Moho discontinuity (2 km contour interval) in the Alpine foreland, constructed by integration of published regional maps. Red lines: superimposed Cenozoic fault systems. Interrupted barbed line: Alpine deformation front. Modified from De`zes P and Ziegler PA (2004) Moho depth map of western and central Europe. EUCOR-URGENT homepage: http://www.unibas.ch/eucor-urgent (acessed Jul 2007).
adopted for the Carboniferous and Permo-Triassic, respectively, and for later times the scale of Gradstein and Ogg (1996). 6.11.2.1.7.(i) Variscan Orogen Evolution of the Variscan Orogen, which forms part of the Hercynian mega-suture along which Laurussia and Gondwana were welded together, involved the stepwise accretion of a number of Gondwana-derived terranes (e.g., Saxo-Thuringian, Armorican-Bohemian, and Moldanubian terranes) to the southern margin of Laurussia and ultimately the collision of Africa with Europe (Ziegler, 1989a, 1989b, 1990b; Tait et al., 1997). ECRIS transects the suture between the RhenoHercynian foreland and the Saxo-Thuringian terrane, and the more internal sutures between the SaxoThuringian and Bohemian, and the Bohemian and
Moldanubian terranes (Figure 18). The triple junction of the Upper Rhine, Roer, and Hessian grabens is superimposed on the south-dipping Rheno-Hercynian/SaxoThuringian suture. The Upper Rhine Graben transects the south-dipping Saxo-Thuringian/Bohemian suture in the northern part of the Vosges and Black Forest (Lalaye-Lubin-Baden-Baden zone), and the north-dipping Bohemian/ Moldanubnian suture in the southern part of the Black Forest (Badenweiler-Lenzkirch zone) (Eisbacher et al., 1989; Franke, 2000; Hegner et al., 2001). To the southwest, the latter links up with the Mt. du Lyonnais suture that is cross-cut by the Bresse and Limagne Grabens (Lardeaux et al., 2001). Total Carboniferous lithospheric shortening in the area transected by ECRIS presumably exceeded 600 km. At the end-Westphalian termination of the Variscan orogeny, the crustal and lithospheric configuration of
512
Tectonic Models for the Evolution of Sedimentary Basins
the future ECRIS area was heterogeneous and presumably marked by a considerable topographic relief. Whereas the Rheno-Hercynian zone was underlain by continental foreland lithosphere, the SaxoThuringian and Moldanubian zones were characterized by an orogenically thickened lithosphere that was thermally destabilized by widespread granitic magmatism. In the internal zones of the Variscan Orogen, crustal thicknesses probably ranged between 45 and 60 km with crustal roots marking the RhenoHercynian/Saxo-Thuringian, Saxo-Thuringian/Bohemian, and Bohemian/Moldanubnian sutures. By end-Westphalian times, a major south-dipping continental lithospheric slab extended from the Variscan foreland beneath the Rheno-Hercynian/SaxoThuringian suture in the area of the Mid-German Crystalline Rise. Similarly, a north-dipping, partly oceanic subduction slab was probably still associated with the Bohemian/Moldanubnian suture, whereas the south-dipping Saxo-Thuringian/Bohemian slab had already been detached during mid-Visean times (Figure 20(a)).
Central, Montagne Noire: Vanderhaeghe and Teyssier, 2001) and subsidence of often narrow, fault-bounded basins, implying high crustal stretching factors, large intervening areas were not significantly extended (Van Wees et al., 2000; Ziegler et al., 2004). Whilst exhumation of the Variscan Internides had commenced already during the main phases of the Variscan orogeny, regional uplift of the entire orogen and its foreland commenced only after crustal shortening had ceased. Stephanian–Early Permian erosional and tectonic exhumation of the Variscan Orogen, in many areas to formerly mid-crustal levels (Burg et al., 1994; Vigneresse, 1999), can be attributed to a combination of such processes as wrench deformation, heating of crustal roots and related eclogite to granulite transformation (Bousquet et al., 1997; Le Pichon et al., 1997), detachment of subducted slabs, upwelling of the asthenosphere, thermal attenuation of the lithospheric mantle, and magmatic and thermal inflation of the remnant lithosphere (Figure 20(b)).
6.11.2.1.7.(ii) Stephanian–Early Permian disruption of the Variscan Orogen End-Westphalian consoli-
6.11.2.1.7.(ii).(a) Permo-Carboniferous magmatism and lithospheric destabilization The widespread
dation of the Variscan Orogen was followed by its Stephanian–Early Permian wrench-induced collapse (305–280 Ma), reflecting a change in the GondwanaLaurussia convergence from oblique collision to a dextral translation (Ziegler, 1989a, 1989b, 1990b). Continental-scale dextral shears, such as the Tornquist-Teisseyre and the Bay of Biscay fractures zones, were linked by secondary sinistral and dextral shear systems. These overprinted and partly disrupted the Variscan Orogen and its northern foreland. Wrench tectonics and associated magmatic activity abated in the Variscan domain and its foreland at the transition to the Late Permian (Ziegler, 1990b; Marx et al., 1995; Ziegler and Stampfli, 2001; Ziegler and De`zes, 2006). Stephanian–Early Permian wrench-induced disruption of the rheologically weak Variscan Orogen was accompanied by regional uplift, widespread extrusive and intrusive magmatism peaking during the Early Permian, and the development of a multidirectional array of transtensional trap-door and pull-apart basins containing continental clastics (Figure 21). Basins developing during this time span underwent a complex, polyphase evolution, including late-stage transpressional deformation controlling their partial inversion. Although Stephanian–Early Permian wrench faulting locally caused uplift of core complexes (e.g., Massif
Stephanian–Early Permian (305–285 Ma) alkaline intrusive and extrusive magmatism of the Variscan domain and its foreland is mantle-derived and locally shows evidence of strong crustal contamination (Bonin, 1990; Bonin et al., 1993; Marx et al., 1995; Benek et al., 1996; Breitkreuz and Kennedy, 1999; Neumann et al., 1995, 2004). Melt generation by partial melting of the uppermost asthenosphere and lithospheric thermal boundary layer was probably triggered by a rise in the potential temperature of the asthenosphere and by its localized transtensional decompression. Wrench-induced detachment of subducted lithospheric slabs presumably caused a reorganization of the mantle convection system and upwelling of the asthenosphere. Mantle-derived mafic melts, which had ascended to the base of the crust, underplated it, inducing crustal anatexis, and the intrusion of fractionally crystallized granitic to granodioritic–tonalitic melts into the crust. Combined with erosional and locally tectonic unroofing of the Variscan crust, interaction of mantle-derived melts with the felsic lower crust contributed to a reequilibration of the Moho at depths of 28–35 km and locally less. By MidPermian times (280 Ma), some 25 My after consolidation of the Variscan Orogen, its crustal roots had disappeared.
Tectonic Models for the Evolution of Sedimentary Basins
RhenoHercynian
(a)
Mid-german cristalline High
SaxoThuringian
Bohemian
513
Moldanubian
N
S
Westphalian
Kraichgau trough
Saar-Nahe trough
(b) N
Schramberg Burgundy trough trough
Kraichgau trough
(c) N
S
Burgundy trough
S
100–120 km
Mesozoic sediments
Permo-Carboniferous troughs
Granitoids
Mantle-derived melts
Continental crust
Oceanic crust
Old mantle-lithosphere
New mantle-lithosphere
Asthenosphere Figure 20 Conceptual model for Late Palaeozoic and Mesozoic evolution of the lithosphere in the ECRIS area along transect A-A’(not to scale). For location of transect see Figures 17 and 18. Modified from Ziegler PA, Schumacher ME, De´zes P, van Wees J-D, and Cloetingh S (2004) Post-Variscan evolution of the lithosphere in the Rhine Graben area: constraints from subsidence modelling. In: Wilson M, Neumann E-R, Davies GR, Timmerman MJ, Heeremans M, and Larsen BT (eds.) Geological Society, London, Special Publications, 223: Permo-Carboniferous Magmatism and Rifting in Europe, pp. 289–317. London: Geological Society, London.
6.11.2.1.7.(ii).(b) Permo-Carboniferous evolution of the ECRIS Zone During the Stephanian and Early
Permian, a system of essentially ENE–WSW trending transtensional intramontane basins developed in the ECRIS area (Figure 21). Subsidence of these
basins, which contain thick continental clastics and volcanics, involved reactivation of the Variscan structural grain, predominantly by dextral shear. The Saar-Nahe Trough is superimposed on the Rheno-Hercynian/Saxo-Thuringian and partly on
514
Tectonic Models for the Evolution of Sedimentary Basins
0°
4°
8°
12°
16°
52° 52°
KT SB
SN
48°
48° BU
44° 0°
4°
8°
12°
Figure 21 Stephanian–Early Permian tectonic framework of ECRIS area, showing sedimentary basins (horizontally hatched), major volcanic fields (cross hatched), major sills (black) and fault systems with superimposed Variscan tectonic units (solid lines) and Alpine deformation front (interrupted barbed line). Abbreviations: BU Burgundy Trough, KT Kraichgau Trough, SB Schramberg Trough, SN Saar-Nahe Trough. Black dots show location of analysed wells. Modified from Ziegler PA, Schumacher ME, De´zes P, van Wees J-D, and Cloetingh S (2004). Post-Variscan evolution of the lithosphere in the Rhine Graben area: constraints from subsidence modelling. In: Wilson M, Neumann E-R, Davies GR, Timmerman MJ, Heeremans M, and Larsen BT (eds.) Geological Society, London, Special Publications, 223: Permo-Carboniferous Magmatism and Rifting in Europe, pp. 289–317. London: Geological Society, London.
the Saxo-Thuringian/Bohemian sutures. The Kraichgau Trough broadly reflects reactivation of the Saxo-Thuringian/Bohemian Lalaye–Lubin– Baden–Baden suture, whereas the Schramberg and Burgundy troughs are associated with the Bohemian/ Moldanubnian Lenzkirch–Badenweiler–Mt. du Lyonnais suture. Subsidence of these basins was coupled with uplift and erosion of intervening highs, amounting, for instance, at the margin of the Saar-Nahe Trough to as much as 10 km. At the same time, NNE–SSW trending Variscan shear zones were sinistrally reactivated, partly outlining the Cenozoic Upper Rhine and Hessian grabens. In the area of the Variscan Internides, the widespread occurrence of a reflection-seismically laminated 10–15 km thick lower crust is mainly attributed to Permo-Carboniferous injection of mantle-derived basic sills. Moreover, truncation of the crustal orogenic fabric by the Moho (Meissner and Bortfeld, 1990) speaks for contemporaneous magmatic destabilization of the crust–mantle boundary. In the internal Variscan zones, no mantle reflectors related to subducted crustal
material (Ziegler et al., 1998) could be detected despite dedicated surveys (Meissner and Rabbel, 1999). This is thought to reflect delamination and/or strong thermal thinning of the lithospheric mantle (Figure 20(b)). Combined with widespread magmatic activity, these phenomena testify to a major thermal surge that can be related to the detachment of the subducted south-dipping continental Rheno-Hercynian lithospheric slab beneath the Mid-German Crystalline Rise and the north-dipping Moldanubian slab in the area of the Lenzkirch–Badenweiler–Mt. du Lyonnais suture, causing upwelling of the asthenosphere into the space formerly occupied by these slabs (Figure 20(b)). This triggered partial melting of the asthenosphere and remnant lithospheric mantle, ascent of melts to the base of the crust and anatexis of lower crustal rocks (model of Davies and Von Blanckenburg, 1996). In conjunction with the ensuing reorganization of mantle flow patterns, a not-very-active mantle plume apparently welled up to the base of the lithosphere in the eastern parts of the future Southern Permian Basin, causing strong thermal attenuation of the lithospheric
Tectonic Models for the Evolution of Sedimentary Basins
mantle and Moho destabilization (Bayer et al., 1999; Van Wees et al., 2000; Ziegler et al., 2004).
Europe, comprising the North Sea rift, the Danish– Polish Trough and the graben systems of the Atlantic shelves. Stress fields controlling their evolution changed repeatedly during the Jurassic and Early Cretaceous (Ziegler, 1988, 1990b; Ziegler et al., 2001; Ziegler and De`zes, 2006). Although the ECRIS area was only marginally affected by Mesozoic rifting, minor diffuse crustal stretching probably contributed towards the subsidence of the Kraichgau, Nancy-Pirmasens, Burgundy, and Trier troughs (Figure 22). Triassic and Jurassic reactivation of Permo-Carboniferous faults, controlling subtle lateral facies and thickness changes, is also evident in the Paris Basin and in the area of the Burgundy Trough. On the other hand, Mesozoic crustal extension played a more important role in the subsidence of the West Netherlands Basin and in the lower Rhoˆne Valley. In an effort to quantify in the wider ECRIS area Late Permian and Mesozoic vertical crustal movements, tectonic subsidence curves were constructed for selected wells of the Paris Basin, Upper Rhine Graben, and
6.11.2.1.7.(iii) Late Permian and Mesozoic thermal subsidence and rifting By late Early
Permian times (280 Ma), magmatic activity abated and thermal anomalies introduced during the PermoCarboniferous began to decay. Combined with progressive degradation of the remnant topography and cyclically rising sea levels, this accounted for the subsidence of increasingly larger areas below the erosional base level and the development of a new intracratonic basin system. However, in large parts of Western and Central Europe thermal reequilibration of the lithosphere–asthenosphere system was overprinted and partly interrupted by the Triassic onset of a new rifting cycle that preceded and accompanied the step-wise breakup of Pangea. Major elements of this breakup system are the southward propagating Arctic–North Atlantic and the westward propagating Neotethys rift systems. Simultaneously, a multidirectional rift system developed in Western and Central 0°
4°
515
8°
12°
16°
2000
1000 1500
SP 52°
1000 500
52°
1500
WN
1000 500
HD TB Wiesloch-Neibheim Lyon-la-Forêt Bourneville
PB
48°
NP
KT
Trois-Fontaines Champotran
48°
Freiburg
FP
Otterbach ?
Aalen
Trochtelfingen
Benken
Sennely
BU
44° 0°
4°
8°
12°
Figure 22 Isopach map of restored Triassic series, contour interval 500 m, showing location of analyzed wells and Variscan (solid barbed line) and Alpine (interrupted barbed line) deformation fronts. White: areas of nondeposition; horizontally hatched: not mapped area. Abbreviations: BU Burgundy Trough, FP Franconian Platform, GG Glu¨ckstadt Graben, HD Hessian Depression, KT Kraichgau Trough, NP Nancy-Pirmasens Trough, PB Paris Basin, PT Polish Trough, SP Southern Permian Basin, TB Trier Basin, WN West Netherlands Basin. Modified from Ziegler PA, Schumacher ME, De´zes P, van Wees J-D, and Cloetingh S (2004) PostVariscan evolution of the lithosphere in the Rhine Graben area: constraints from subsidence modelling. In: Wilson M, Neumann E-R, Davies GR, Timmerman MJ, Heeremans M, and Larsen BT (eds.) Geological Society, London, Special Publications, 223: Permo-Carboniferous Magmatism and Rifting in Europe, pp. 289–317. London: Geological Society, London.
516
Tectonic Models for the Evolution of Sedimentary Basins
Franconian Platform (Ziegler et al., 2004), applying the back stripping method of Sclater and Christie (1980). These curves, similar to those by Loup and Wildi (1994), Prijac et al., (2000) and Van Wees et al., (2000), show that reequilibration of the lithosphere with the asthenosphere commenced during the late Early Permian (280 Ma) and continued throughout Mesozoic times. Moreover, they show superimposed on the long-term thermal subsidence trends intermittent and generally local Mesozoic subsidence accelerations (Figure 23). These are interpreted as reflecting either tensional reactivation of PermoCarboniferous fault systems or compressional deflection of the lithosphere (Cloetingh, 1988) under stress fields related to far-field rifting and wrench activity.
Depth (m)
(a)
0
300
250
200
150
100
50
0
Age (Ma)
250 500 750 1000
Air-loaded tectonic subsidence curve
(b)
Uplift (m)
1000 750 500 250
Subsidence (m)
0
300
250
200
150
100
50
0
Age (Ma) 250 500 750 1000
Modeled subsidence curve
Figure 23 (a) Air-loaded tectonic subsidence curve and (b) modeled subsidence curve for well Bourneville, Paris Basin. For locations see Figure 22. Black squares: control points derived from penetrated sedimentary sequence. The positive part of the modeled subsidence curve reflects uplift of the crust in response to thermal thinning and/or delamination of the mantle-lithosphere; its negative part reflects thermal subsidence of the crust during re-equilibration of the lithosphere/asthenosphere system. Modified from Ziegler PA, Schumacher ME, De´zes P, van Wees J-D, and Cloetingh S (2004) Post-Variscan evolution of the lithosphere in the Rhine Graben area: constraints from subsidence modelling. In: Wilson M, Neumann E-R, Davies GR, Timmerman MJ, Heeremans M, and Larsen BT (eds.) Geological Society, London, Special Publications, 223: Permo-Carboniferous Magmatism and Rifting in Europe, pp. 289–317. London: Geological Society, London.
Temporal and spatial variations in these subsidence accelerations relate to differences in the orientation of preexisting crustal discontinuities and changes in the prevailing stress field. Nevertheless, overall subsidence patterns reflect long-term reequilibration of the lithosphere–asthenosphere system. 6.11.2.1.7.(iv) Tectonic subsidence modeling To
define the end-Early Permian configuration of the lithosphere, the tectonic subsidence curves were compared to a theoretical thermal decay curve, applying a numerical forward/backward modeling technique which automatically finds the best-fit stretching parameters for the observed subsidence data (Van Wees et al., 1996, 2000). Forward/backward modeling of tectonic subsidence is based on lithospheric stretching assumptions ( ¼ crustal stretching factor, ¼ lithospheric mantle stretching factor) under which the lithosphere is represented by a plate with constant temperature boundary conditions, adopting a fixed basal temperature (McKenzie, 1978; Jarvis and McKenzie, 1980; Royden and Keen, 1980). For thermal calculations, a 1-D numerical finite difference model was used, adopting parameters as given by Van Wees et al., (2000), that allows for incorporation of finite and multiple stretching phases, as well as for crustal heat production effects and conductivity variations (Van Wees et al., 1992, 1996, 2000; Van Wees and Stephenson, 1995). Differential stretching of the crust and lithospheric mantle can be applied to simulate thermal attenuation of the latter. Input parameters for forward/backward modeling of the observed subsidence curves include the prerift crustal and present lithospheric thickness, and for each stretching phase its timing, duration and mode of lithospheric extension (uniform ¼ (McKenzie, 1978); two-layered < (Royden and Keen, 1980)). In iterative steps modeling parameters are changed until a good fit is obtained between the observed and modeled subsidence curves. Best-fit stretching parameters thus determined give a measure of the thermal perturbation of the lithosphere during the Permo-Carboniferous and subsequent tensional events that interfered with the reequilibration of the asthenosphere–lithosphere system. Modeling of the lithosphere evolution in the ECRIS area is based on the concept that after the Permo-Carboniferous thermal surge (300–280 Ma) the temperature of the asthenosphere returned rapidly to ambient levels (1300 C), at which it remained until the end-Cretaceous renewed flareup
Tectonic Models for the Evolution of Sedimentary Basins
Subsidence (m)
Uplift (m)
(a)
Subsidence (m)
Uplift (m)
(b)
Uplift (m)
(c)
Subsidence (m)
of plume activity. In the forward/backward model lithospheric thicknesses of 100–120 km were adopted that, according to Babuska and Plomerova (1992, 1993), are representative for areas not affected by Cenozoic rifting. Furthermore, as most of the analyzed wells are located outside Permo-Carboniferous troughs, initial crustal thicknesses were assumed to be close to the present values. The subsidence curves were modeled with PermoCarboniferous differential crustal and lithospheric mantle extension (attenuation), allowing factors to attain significantly greater values than factors. The high factors represent the effects of delamination and thermal thinning of the lithospheric mantle. On the other hand, the temporary Mesozoic subsidence accelerations were successfully modeled with uniform lithospheric extension ( ¼ ). The modeled subsidence curves (Figure 24) demonstrate that, after an initial uplift phase between 300 and 280 Ma, which gives a measure of Stephanian–Early Permian lithospheric thinning, the subsequent evolution of the lithosphere was governed by the long-term decay of thermal anomalies introduced during the Permo-Carboniferous tectonomagmatic cycle. This is in accordance with the assumption that the temperature of the asthenosphere had returned to ambient levels around 280 Ma. Good fits between observed and modeled tectonic subsidence curves were obtained, assuming initial crustal thicknesses of 30–35 km, final lithospheric thicknesses of 100–120 km, and a Permo-Carboniferous ‘stretching’ phase that spanned 300–280 Ma and involved decoupled crustal extension and attenuation of the lithospheric mantle. This assumption is compatible with the concept that during the PermoCarboniferous reequilibration of the crust–mantle boundary crustal extension played only locally a significant role. In this respect it is noteworthy that the important Permo-Carboniferous troughs, which occur on the Massif Central and the Bohemian Massif and beneath the Franconian Platform, do not coincide with major Late Permian and Mesozoic depocenters, whereas no major Permo-Carboniferous basins are located under the Southern Permian Basin and Paris Basin depocenters (Ziegler, 1990b). Indeed, no obvious relationship is evident between the distribution and orientation of Permo-Carboniferous troughs and the geometry of the superimposed Late PermianMesozoic thermal-sag basins (cf. Figures 21 and 22). This suggests that during the Permo-Carboniferous tectonomagmatic cycle, uniform and/or depth-dependent lithospheric extension was, on a regional scale, only a contributing but not the dominant mechanism of
517
1000 Trois-Fontaines Wiesloch-Neib. Freiburg i. Br. Otterbach
750 500 250 0
300
250
200
150
100
50
0
250 500 750 1000 1000
Lyon-la-Fort Bourneville Sennely Champotran
750 500 250 0
300
250
200
150
100
50
0
250 500 750 1000 1000 Aalen Benken Trochtelfingen
750 500 250 0
300
250
200
150
100
50
0
250 500 750 1000
Figure 24 Modeled subsidence curves for (a) Upper Rhine Graben and Lorraine area, (b) Paris Basin, and (c) Franconian platform. For location of wells see Figure 22. Modified from Ziegler PA, Schumacher ME, De´zes P, van Wees J-D, and Cloetingh S (2004) Post-Variscan evolution of the lithosphere in the Rhine Graben area: constraints from subsidence modelling. In: Wilson M, Neumann E-R, Davies GR, Timmerman MJ, Heeremans M, and Larsen BT (eds.) Geological Society, London, Special Publications, 223: Permo-Carboniferous Magmatism and Rifting in Europe, pp. 289–317. London: Geological Society, London.
crustal and lithospheric mantle thinning, as advocated for the Paris Basin by Prijac et al., (2000). By contrast, lithospheric stretching may have played a somewhat more important role in the evolution of the Hessian Depression, Nancy-Pirmasens and Burgundy system of Late Permian and Mesozoic basins that is superimposed on a Basin-and-Range type array of PermoCarboniferous troughs (Figures 21 and 22).
518
Tectonic Models for the Evolution of Sedimentary Basins
Modeled subsidence curves indicate that during the Permo-Carboniferous tectonomagmatic cycle the lithospheric mantle was significantly attenuated and that factors attained values in the range of 1.8–10. As these values are subject to large lateral variations, they reflect that thinning of the lithospheric mantle was heterogeneous and generally more intense in areas that evolved into Mesozoic depocenters than in areas marginal to them. Similarly, in areas that remained positive features throughout much of Mesozoic times, such as the Bohemian and Armorican Massifs, the lithospheric mantle was apparently not significantly thinned during the Permo-Carboniferous and retained a thickness of 70–100 km, as well as an orogen (subduction)-related anisotropy (Babuska and Plomerova, 2001; Judenherc et al., 2002). Sensitivity studies indicate that best fits between observed and modeled subsidence curves are obtained when the thickness of the thermal lithosphere at its end-Mesozoic equilibration with the asthenosphere is set at 100 or 120 km. Yet, even at these values, Permo-Carboniferous factors and the end-Early Permian thickness of the remnant lithospheric mantle (RLM) vary significantly (e.g., Trochtelfingen: 100 km lithosphere: ¼ 3.02, RLM 23.2 km; 120 km lithosphere: ¼ 1.96, RLM 43.3 km). For the Paris Basin, the best fit between observed and modeled subsidence curves was achieved with a lithosphere thickness of 120 km, whereas for the Upper Rhine Graben and the Franconian Platform best fits were obtained with a lithosphere thickness of 100 km. Whereas a 100 km lithosphere thickness is compatible with the Palaeocene plume-related segregation depth of olivine-melilitic partial melts in the Vosges, Black Forest, and Bohemian Massif (Wilson et al., 1995), the apparently greater lithosphere thickness beneath the Paris Basin remains enigmatic. In view of the above, values given in Table 1 for the end-Early Permian thickness of the RLM ought to be regarded as rough approximations. Nevertheless, it is evident that substantial PermoCarboniferous thinning of the lithospheric mantle provided the principal driving mechanism for the Late Permian and Mesozoic subsidence of thermalsag basins that developed in the ECRIS area. On a regional scale, modeled Permo-Carboniferous crustal extension was relatively low. Under the assumption of initial crustal thicknesses of 30–35 km, automated modeling yielded factors of 1.04–1.13 and crustal thicknesses close to present values. The minor, intra-Mesozoic subsidence accelerations, which overprint the long-term thermal
subsidence curves, were successfully modeled by uniform lithospheric extension with cumulative ¼ values in the range of 1.01–1.07. As corresponding extensional faulting is generally poorly documented, stress-induced deflections of the lithosphere (Cloetingh, 1988) may have contributed to some of these subsidence anomalies. Summarizing, in the wider ECRIS area the following sequence of dynamic processes controlled the transformation of the overthickened crust and lithosphere of the Variscan Orogen to present-day crustal and lithospheric thicknesses: (1) Stephanian–Early Permian wrench faulting caused disruption of the Variscan Orogen, detachment of subducted lithospheric slabs and upwelling of the asthenosphere, giving rise to widespread mantlederived magmatic activity. During this thermal surge, partial delamination and thermal thinning of the lithospheric mantle, thermal inflation of the remnant lithosphere and interaction of mantle-derived partial melts with the lower crust accounted for the destruction of the 45–60 km deep crustal roots of the Variscan Orogen and its regional uplift. By end-Early Permian times the crust was thinned down on a regional scale to 27–35 km, mainly by magmatic processes and erosional unroofing and only locally by mechanical stretching, whilst the thickness of the mantlelithosphere was reduced to 9–40 km in areas that evolved into Late Permian to Mesozoic depocentres, whereas it retained a thickness of 40–90 km beneath slowly subsiding areas and persisting highs. There is no relationship between the degree of lithospheric thinning and the different Variscan tectonic units. (2) During the late Early Permian the temperature of the asthenosphere returned rapidly to ambient levels. With this, reequilibration of the lithosphere with the asthenosphere commenced and persisted during the Mesozoic, controlling the subsidence of a system of intracratonic thermal-sag basins. As there is no clear relationship between the distribution of Permo-Carboniferous troughs and the geometry of the superimposed intracratonic Late Permian to Mesozoic thermal-sag basins, PermoCarboniferous thermal thinning and delamination of the lithospheric mantle provided the principal driving mechanism for their subsidence. Minor intra-Mesozoic tensional events hardly disturbed the asthenosphere–lithosphere system of the ECRIS area where by end-Cretaceous times the lithosphere had reequilibrated with the asthenosphere at depths of 100–120 km.
Tectonic Models for the Evolution of Sedimentary Basins
(3) The lithosphere–asthenosphere system of the ECRIS area became destabilized again at the transition from the Cretaceous to the Palaeocene in conjunction with a phase of major intraplate compression that was accompanied by the impingement of mantle plumes. With the late Eocene activation of ECRIS, crustal extension, and particularly Neogene increased plume activity, caused further destabilization of its lithosphere–asthenosphere system (De`zes et al., 2004). As a result, the present thickness of the lithosphere decreases from 100–120 km in areas flanking ECRIS to 60–70 km beneath parts of the Rhenish Massif and the Massif Central (Babuska and Plomerova, 1993). 6.11.2.2
Compressional Basins Systems
Below we discuss basic concepts and observations for large-scale compressional basin systems. In doing so, we focus on foreland basins and basins initiated by lithospheric folding. 6.11.2.2.1
Development of foreland basins Foreland basins owe their existence to the capacity of the lithosphere to support loads, such as the topography of orogenic wedges or subducted lithospheric slabs. The lithosphere deforms by flexurally bending downwards over areas often exceeding the spatial scale of these loads. The width of the resulting depression, the foreland basin, provides information on the mechanical strength of the underlying lithosphere. Under an imposed load, a zero-strength lithosphere would simply sink vertically into the mantle (Airy isostasy), thus not accounting for the development of a foreland basin. By contrast, a mechanically strong lithosphere, characteristic of cratonic forelands, allows
for the subsidence of wide and relatively shallow foreland basins (flexural isostasy). Such wide foreland basins are associated with the Canadian Rocky Mountains of Alberta and British Columbia and the Appalachian fold-and-thrust belts that are superimposed on the North American Proterozoic cratonic crust (e.g., Beaumont, 1981; Quinlan and Beaumont, 1984). By contrast, the much narrower Alpine foreland basins of Europe (Mascle et al., 1998) developed on considerably younger lithosphere that equilibrated with the asthenosphere after the Variscan orogeny and was modified by Mesozoic rifting activity. The time elapsed between rifting and the syn-orogenic flexural deformation of this foreland lithosphere was very variable (Pyrenees: ca. 30 Ma; Apennines: ca. 130 Ma). The relatively modest widths of the foreland basins of the Carpathians (Zoetemeijer et al., 1999; Mat¸enco et al., 1997b), Pyrenees (Millan et al., 1995), the Betic Cordillera (Van der Beek and Cloetingh, 1992), the Apennines (Zoetemeijer et al., 1993), and the eastern Alps (Andeweg and Cloetingh, 1998) reflect a relatively weak lithosphere. In addition to topographic loading by orogenic wedges and loading by the sedimentary fill of the foreland basins, additional forces operate on the lithosphere, such as slab-pull, slab-detachment and slab roll-back, and play an important role in shaping the geometry of prowedge and retrowedge foreland basins (Figure 25) (e.g., Millan et al., 1995; Ziegler et al., 2002). Moreover, depending on mechanical coupling/decoupling of the orogenic wedge and the foreland lithosphere, horizontal compressional stresses can be exerted onto the latter, influencing the geometry of an evolving flexural foreland basin (Ziegler et al., 2002). In a first generation of flexural foreland basin models, the different loading components and forces were
Pre-existing Orogen structures Retro-wedge Pro-wedge Foreland foredeep foredeep Foreland
Oceanic crust
Continental crust
Oceanic crust
519
Orogenic load
Horizontal stresses
Mantle lithosphere
Zone of mechanical interactions
Hidden load (slab pull)
Figure 25 Conceptual model illustrating forces controlling the subsidence of retro- and pro-wedge foreland basins and showing zone of potential mechanical coupling between the upper and lower plate. Modified from Ziegler PA, Bertotti G, and Cloetingh S (2002). Dynamic processes controlling foreland development – the role of mechanical (de)coupling of orogenic wedges and forelands. In: Bertotti G, Schulmann K, and Cloetingh SAPL (eds.) EGU St. Mueller Special Publication Series, 1: Continental Collision and the Tectono-Sedimentary Evolution of Forelands, pp. 17–56. EGU.
520
Tectonic Models for the Evolution of Sedimentary Basins
represented by a system of vertical loads, shear forces, and bending moments (Royden, 1988, 1993). This approach is illustrated in Figure 26. A characteristic feature of these models is that horizontal stresses were neglected. Nevertheless, these have been shown to play a potentially important role in the geometry of foreland basins and the architecture of their sedimentary fill (Peper et al., 1994). These models were utilized to extract information on the mechanical properties of the lithosphere that was treated as an elastic plate rather than as a depth-dependent rheologically stratified beam. Consequently, the models yielded rough estimates for the effective elastic thickness of the lithosphere and a first approximation for the bulk rheology of the continents. This approach permitted to examine the effect of the time elapsed since the last thermal perturbation of the lithosphere (generally a rifting phase) on its deflection during the subsequent foreland basin development phase. In this respect, Desegaulx et al., (1991) explained the rapid Cenozoic subsidence of the Aquitaine foreland basin (SW France) in terms of preorogenic syn-rift extensional thinning and associated thermal perturbation of the Effective plate end
ρ
win(x)
w
Sea level
ρ
c
ρ
Moho
m
ρ
c
Topographic load Flexure amplified by infilling material
w (x) Flexural bulge Sea level
(ρc– ρw) Moho
crustal root
Subsurface M load P
(ρ – ρ )
(ρc– ρm)
m
c
Effective plate end
Figure 26 Concept of lithospheric flexure during foreland basin evolution. Deflection of the subducting lower plate lithosphere is the combined result of topographic loading by the mountain chain of the upper plate, the weight of the sediments in the foreland basin, and the weight of the subducted slab. The shape of the flexural foreland depression is controlled by the interplay between the lithospheric strength of the lower plate, the magnitude of the differential loads imposed on it, and the level of collisionrelated compressional stresses transmitted into it. Modified from Royden (1993); see also Ziegler et al., (2002).
lithosphere and its postrift thermal subsidence and sedimentary loading. The thermomechanical age concept provides the framework for effective elastic thickness (EET) estimates of the lithosphere. Figure 27 was constructed on the basis of a large data set for Eurasian foreland basins (Cloetingh and Burov, 1996) and illustrates the general trend of increasing elastic plate thickness with increasing thermomechanical age. Also plotted in Figure 27 are predictions for the bulk rheology of the lithosphere based on extrapolations from rock mechanical data, constrained by crustal geophysical data and thermal models. A characteristic feature of these models is the incorporation of a quartz-dominated upper crustal rheology and an olivinecontrolled mantle rheology (see also Figure 11). These models are cast in terms of the depth to the base of the mechanically strong upper part of the crust (MSC) and the mechanically strong part of the upper mantle lithosphere (MSL). Analysis of Figure 27 demonstrates that the mechanical properties of the crust are little affected by its age-dependent cooling, whereas the thickness and strength of the lithospheric mantle is very strongly age dependent. Depth- and temperature-dependent rheological models for the lithosphere show that the mechanically weak lower crust separates the mechanically strong upper crust and upper lithospheric mantle (e.g., Watts and Burov, 2003). The EET bands for the mechanically strong upper crust and upper lithospheric mantle describe the integrated EET of the lithosphere that has a bearing on its response to loads imposed on it. The degree of coupling and/or decoupling between these two mechanically strong layers and the scatter of data points reflects, to a large extent, the importance of mechanical weakening of the lower crust by tectonic stresses (Cloetingh and Burov, 1996). At the same time it was realized that rheological decoupling of the upper crust and lithospheric mantle plays an important role in the structural style of intraplate compressional deformations (e.g., Rocky Mountains and Bohemian Massif; Ziegler et al., 1995, 2002). 6.11.2.2.2 Compressional basins: lateral variations in flexural behaviour and implications for palaeotopography
Modeling of compressional basins followed essentially the same philosophy as the modeling of extensional basins. Initial lithosphere-scale models focused on the role of flexural behaviour of the lithosphere during foreland basin development (e.g., Zoetemeijer et al.,
Tectonic Models for the Evolution of Sedimentary Basins
0
200
0
400
600
800
1000
1200
1600
1800
2000
2200
2400
AAR S.A M
Tc
100°
200° 300 °
JURA
–50
200° Base of the mechanical upper crust
E.A
50 0° 60 0°
VE
NB
TA-DZ
NHD
S.B.S
400° UR 500°
70 0°
–100
Tc
hcl
300°
EIFEL URA
C.B.S
0°
0° 90
700° INDIA
° 00 10
–150
N.B.S FE
600°
80
Depth, z (km)
1400
521
800°
Te
Deepest seismicity
°
00
11
900°
Base of the mechanical mantle h m
1000° 12 00
–200
°
1100°
MSC MSL
1200°
0
200
400
600
800
1000 1200 1400 Time/age, t (Ma)
1600
1800
2000
2200
2400
Figure 27 Compilation of observed and predicted values of effective elastic thickness (EET), depth to bottom of mechanically strong crust (MSC), and depth to bottom of mechanically strong lithospheric mantle (MSL) plotted against the age of the continental lithosphere at the time of loading and comparison with predictions from thermal models of the lithosphere. Labeled contours are isotherms. Isotherms marked by ‘solid lines’ are for models that account for additional radiogenic heat production in the upper crust. ‘Dashed lines’ correspond to pure cooling models for continental lithosphere. The equilibrium thermal thickness of the continental lithosphere is 250 km. ‘Shaded bands’ correspond to depth intervals marking the base of the mechanical crust (MSC) and the mantle portion of the lithosphere (MSL). ‘Squares’ correspond to EET estimates, ‘circles’ indicate MSL estimates, and ‘diamonds’ correspond to estimates of MSC. ‘Bold letters’ correspond to directly estimated EET values derived from flexural studies on, for example, foreland basins, ‘Thinner letters’ indicate indirect rheological estimates derived from extrapolation of rock-mechanics studies. The data set includes (I): Old thermo-mechanical ages (1000–2500 Ma): northernmost (N.B.S.), central (C.B.S.), and southernmost Baltic Shield (S.B.S.); Fennoscandia (FE); Verkhoyansk plate (VE); Urals (UR); Carpathians; Caucasus, (II): Intermediate thermo-mechanical ages (500–1000 Ma): North Baikal (NB); Tarim and Dzungaria (TA-DZ); Variscan of Europe: URA, NHD, EIFEL; and (III): Younger thermo-mechanical ages (0–500 Ma): Alpine belt: JURA, MOLL (Molasse), AAR; southern Alps (SA) and eastern Alps (EA); Ebro Basin; Betic rifted margin; Betic Cordilleras. Modified from Cloetingh S and Burov E (1996) Thermomechanical structure of European continental lithosphere: constraints from rheological profiles and EET estimates. Geophysical Journal International 124: 695–723.
1990; Van der Beek and Cloetingh, 1992). These studies drew on data sets consisting of wells, gravity data, and deep seismic profiling, such as the ECORS profiles through the Pyrenees, completed in the 1980s. Roure et al., (1994, 1996a) and Ziegler and Roure (1996) give a detailed discussion on constraints provided by deep seismic data on the bulk geometry of Alpine belts. Flexural modeling was backed up by large-scale studies on the rheological evolution of continental lithosphere (Cloetingh and Burov, 1996) that demonstrated in compressional settings a direct link between the mechanical properties of the lithosphere, its thermal structure, and the level of regional intraplate stresses. The inferences drawn from large-scale flexural modeling provided feedback for subsequent analysis
on sub-basin scales. For example, modeling predictions for the presence of weak lithosphere in the Alpine belt invoke steep downward deflection of the lithosphere, favouring the development of upper crustal flexure-induced synthetic and antithetic tensional faults (Ziegler et al., 2002). Such fault systems are observed on reflection-seismic profiles in the Alpine Molasse Basin of Germany and Austria (Ziegler, 1990b) and in the Carpathian foreland basin of Poland (Oszczypko, 2006), the Ukraine (Izotova and Popadyuk, 1996), and Romania (Mat¸enco et al., 1997a). This flexure-induced upper crustal normal faulting has caused weakening of the lithosphere. Integrated flexural analysis of a set of profiles across the Ukranian Carpathians and their
522
Tectonic Models for the Evolution of Sedimentary Basins
foreland has demonstrated an extreme deflection of the lithosphere, leading almost to its failure, and very pronounced offsets on upper crustal normal faults (see Figure 28) (Zoetemeijer et al., 1999). Following studies on the palaeorheology of the lithosphere, constrained by high-quality thermochronology in the Central Alps (Okaya et al., 1996) and Eastern Alps (Genser et al., 1996), the importance of large lateral variations in the mechanical strength of mountain belts became evident. This pertains particularly to a pronounced strength reduction from the external part of an orogen towards its internal parts. As a result, foreland basins will develop in the external zone with its stronger lithosphere, whereas in the internal zones of the orogens pull-apart basins will develop
Gravity anomaly (mgal)
(a)
on low-strength lithosphere (Nemes et al., 1997, Cloetingh et al., 1992). The consequences of lateral flexural strength variations of the lithosphere were explored by a modeling study that was carried out along a transsect through the NE Pyrenees that is well constrained by crustal-scale seismic control and an extensive field-derived database (Verge´s et al., 1995). Apart from investigating the present configuration of foreland basin and quantifying the present-day mechanical structure of the lithosphere underlying the southern Pyrenees fold-and-thrust belt, the relationship between palaeotopography and flexural evolution of the orogen was analyzed (Millan et al., 1995). This novel approach led to a set of testable predictions on palaeotopography (Figure 29) and sediment supply to the foreland basin.
Bouguer gravity 100
0
–100 80
(b)
100
120 Distance (km)
140
160
Ukrainian outer carpathians SW
SKIBA UNIT
CARPATHIAN FOREDEEP INNER
0 Stryjska-1
Height (km)
180
OUTER
Dashava-100 Bolochiv-3 Dashava-107
Zhuravno-9
Misun-1
–5
Shevchenkovo-1
Vyshkiv-1
Wells Structural interpretation Model prediction
–10
60
NE PLATFORM
80
100
120 Distance (km)
140
160
180
10°E 15°E 20°E 25°E 30°E 50°N
50°N
45°N
45°N
100°E
15°E
20°E
25°E
30°E
Figure 28 Cross section illustrating (a) Bouguer gravity anomaly and (b) steep down bending (b) of the East-European platform lithosphere beneath the Ukranian segment of the Carpathian fold-and-thrust belt. The extreme curvature of the downbent lithosphere is associated with major upper crustal normal faulting. Narrow foreland basins are characteristic for foreland flexures involving weak lithosphere. Modified from Zoetemeijer et al., (1999).
Tectonic Models for the Evolution of Sedimentary Basins
Erosion/sedimentation
Eastern Pyrenees Partially restored cross section (Middle Eocene) Flexure (different paleo-elevations: 1000 m, 2000 m)
Deflection (km)
5
S
523
Upper crust
Strong
Lower crust
Strong Weak
N
V
0 –5
Upper mantle
–10
Weak
o Moh
V
Strong
Weak
–15 200
160
120 80 Distance (km)
40
0
40
0
EET (km)
Effective elastic thickness 40 20 0 200
160
120 80 Distance (km)
Figure 29 Modeled flexure for a partly restored section (Middle Eocene) in the eastern Pyrenees. Top: Lithospheric flexure computed for low (1000 m) and moderate palaeotopographic elevations (2000 m). Dots represent observed deflection of the lithosphere. Dashed line: Fit obtained for 1000 m elevation, bending moment M ¼ 7 1016 N and shear force V ¼ 1.6 1011 N m1. Solid line: Fit obtained for 2000 m elevation, bending moment M ¼ 5 1016 N and shear force V ¼ 1 1011 N m1. Increasing elevations are accompanied by decreasing subduction forces. Bottom: EET values for the model with the best flexural fit. An overall thicker EET is required for moderate to low estimates of palaeo-elevation. Modified from Millan H, Den Bezemer T, Verge´s J, et al., (1995) Paleo-elevation and EET evolution at mountain ranges: inferences from flexural modelling in the eastern Pyrenees and Ebro basin. Marine and Petroleum Geology 12: 917–928.
6.11.2.2.3 Lithospheric folding: an important mode of intraplate basin formation
Folding of the lithosphere, involving its positive as well as negative deflection (see Figure 30), appears to play a more important role in the large-scale neotectonic deformation of Europe’s intraplate domain than hitherto realized (Cloetingh et al., 1999). The large wavelength of vertical motions associated with lithospheric folding necessitates integration of available data from relatively large areas (Elfrink, 2001), often going beyond the scope of regional structural and geophysical studies that target specific structural provinces. Recent studies on the North German Basin have revealed the importance of its neotectonic structural reactivation by lithospheric folding (Marotta et al., 2000). Similarly, the Plio-Pleistocene subsidence acceleration of the North
Figure 30 Schematic diagram illustrating decoupled lithospheric mantle and crustal folding, and consequences of vertical motions and sedimentation at the Earth’s surface. V is horizontal shortening velocity; upper crust, lower crust, and mantle layers are defined by corresponding rheologies and physical properties. A typical brittle-ductile strength profile (in black) for decoupled crust and upper mantle– lithosphere, adopting a quartz–diorite–olivine rheology, is shown for reference.
Sea Basin (Figure 10) is attributed to stress-induced buckling of its lithosphere (Van Wees and Cloetingh, 1996). Moreover, folding of the Variscan lithosphere has been documented for Brittany (Bonnet et al., 2000), the adjacent Paris Basin (Lefort and Agarwal, 1996), and the Vosges–Black Forest arch (De`zes et al., 2004; Ziegler and De`zes, in press). Lithospheric folding is a very effective mechanism for the propagation of tectonic deformation from active plate boundaries far into intraplate domains (e.g., Stephenson and Cloetingh, 1991; Burov et al., 1993; Ziegler et al., 1995, 1998, 2002). At the scale of a microcontinent that was affected by a succession of collisional events, Iberia provides a well-documented natural laboratory for lithospheric folding and the quantification of the interplay between neotectonics and surface processes (Cloetingh et al., 2002). An important factor in favour of a lithosphere-folding scenario for Iberia is the compatibility of the thermotectonic age of its lithosphere and the wavelength of observed deformations. Well-documented examples of continental lithospheric folding also come from other cratonic areas. A prominent example of lithospheric folding occurs in the Western Goby area of Central Asia, involving a lithosphere with a thermotectonic age of 400 Ma. In this area, mantle and crustal wavelengths are 360 km and 50 km, respectively, with a shortening rate of 10 mm yr1 and a total amount of shortening of, 200–250 km during 10–15 My (Burov et al., 1993; Burov and Molnar, 1998). Quaternary folding of the Variscan lithosphere in the area of the Armorican Massif (Bonnet et al., 2000) resulted in the development of folds with a
524
Tectonic Models for the Evolution of Sedimentary Basins
0
11
22
33
Method: Food mechanism Breakouts Drill. induced frac. Borehole slotter Overcoring Hydro. fractures Geol. indicators Reglme: NF
SS TF Quality:
U
A B C All depths (2003) World stress map
64
56
48
40
World Stress Map Rel. 2003 Heidelberg Academy of Sciences and Humanities Geophysical Institute, University of Karlsruhe
Projection: Mercator
Figure 31 Present-day stress map of Europe showing orientation of maximum horizontal stress axes (SHmax). Different symbols stand for different stress indicators; their length reflects the data quality, ‘A’ being highest. Background shading indicates topographic elevation (brown high, green low). This map was derived from the World Stress Map database (http:// www.world-stress-map.org/).
wavelength of 250 km, pointing to a mantle-lithospheric control on deformation. As the timing and spatial pattern of uplift inferred from river incision studies in Brittany is incompatible with a glacioeustatic origin, Bonnet et al., (2000) relate the observed vertical motions to deflection of the lithosphere under the present-day NW–SE directed compressional intraplate stress field of NW Europe (Figure 31). Stress-induced uplift of the area appears to control fluvial incision rates and the position of the main drainage divides. The area located at the
western margin of the Paris Basin and along the rifted Atlantic margin of France has been subject to thermal rejuvenation during Mesozoic extension related to North Atlantic rifting (Ziegler and De`zes, 2006; Robin et al., 2003) and subsequent compressional intraplate deformation (Ziegler et al., 1995), also affecting the Paris Basin (Lefort and Agarwal, 1996). Leveling studies in this area (Lenoˆtre et al., 1999) also point towards its ongoing deformation. The inferred wavelength of these neotectonic lithosphere folds is consistent with the general relationship
Tectonic Models for the Evolution of Sedimentary Basins
0
Iberia
Central Asia Upper crustal folding
100
Arctic Canada
200 Wavelength (km)
525
Brittany
300
Iberia Iberia (south) Central Asia
400
Mantle folding
Iberia (north)
500
Whole lithosphere folding
600 Central Australia
700
Trans Continental Arch of North America
800 0
200
400
600 800 Time/age (ma)
1000
1200
1400
Figure 32 Comparison of observed (solid squares) and modeled (open circles) wavelengths of crustal, lithospheric mantle and whole lithospheric folding in Iberia (Cloetingh et al., 2002) with theoretical predictions (Cloetingh et al., 1999) and other estimates (open squares) for wavelengths documented from geological and geophysical studies (Stephenson and Cloetingh, 1991; Nikishin et al., 1993; Ziegler et al., 1995; Bonnet et al., 2000). Wavelength is given as a function of the thermo-tectonic age at the time of folding. Thermo-tectonic age corresponds to the time elapsed since the last major perturbation of the lithosphere prior to folding. Note that neotectonic folding of Variscan lithosphere has recently also been documented for Brittany (Bonnet et al., 2000). Both Iberia and Central Asia are characterized by separate dominant wavelengths for crust and mantle folds, reflecting decoupled modes of lithosphere folding (Cloetingh et al., 2005). Modified from Cloetingh S, Burov E, Beekman F, Andeweg B, Andriessen PAM, Garcia-Castellanos D, de Vicente G, and Vegas R (2002) Lithospheric folding in Iberia. Tectonics 21(5): 1041 (doi:10.1029/2001TC901031).
that was established between the wavelength of lithospheric folds and the thermotectonic age of the lithosphere on the base of a global inventory of lithospheric folds (Figure 32) (see also Cloetingh and Burov, 1996; Cloetingh et al., 2005). In a number of other areas of continental lithosphere folding, smaller wavelength crustal folds have also been detected, for example, in Central Asia (Burov et al., 1993; Nikishin et al., 1993). Thermal thinning of the mantle-lithosphere, often associated with volcanism and doming, enhances lithospheric folding and appears to control the wavelength of folds. Substantial thermal weakening of the lithospheric mantle is consistent with higher folding rates in the European foreland as compared to folding in Central Asia (Nikishin et al., 1993), which is marked by pronounced mantle strength (Cloetingh et al., 1999).
6.11.3 Rheological Stratification of the Lithosphere and Basin Evolution 6.11.3.1 Lithosphere Strength and Deformation Mode The strength of continental lithosphere is controlled by its depth-dependent rheological structure in which
the thickness and composition of the crust, the thickness of the lithospheric mantle, the potential temperature of the asthenosphere, and the presence or absence of fluids, as well as strain rates play a dominant role. By contrast, the strength of oceanic lithosphere depends on its thermal regime, which controls its essentially age-dependent thickness (Kusznir and Park, 1987; Cloetingh and Burov, 1996; Watts, 2001; see also Watts, this volume (Chapter 6.01) and Burov, this volume (Chapter 6.03)). Figure 33 gives synthetic strength envelopes for three different types of continental lithosphere and for oceanic lithosphere at a range of geothermal gradients (Ziegler and Cloetingh, 2004). These theoretical rheological models indicate that thermally stabilized continental lithosphere consists of the mechanically strong upper crust, which is separated by a weak lower crustal layer from the strong upper part of the mantle–lithosphere that in turn overlies the weak lower mantle–lithosphere. By contrast, oceanic lithosphere has a more homogeneous composition and is characterized by a much simpler rheological structure. Rheologically, thermally stabilized oceanic lithosphere is considerably stronger than all types of continental lithosphere. However, the strength of oceanic lithosphere can be
526
Tectonic Models for the Evolution of Sedimentary Basins
(a)
(b)
Strength (MPa) −1000 0 10
−500
0
500
1000
Strength (MPa)
1500
−1000 0
Unextended thick shield-type lithosphere
UC: quartz (dry)
20
40 50
Moho UM: olivine (dry)
60
100
Tension
LC: diorite (wet) Moho
60
100
−500
0
500
1000
10
Depth (km)
30
30 LC: diorite (wet)
60 70 80 90 100
Moho UM: olivine (dry) 0 Ma 25 Ma 50 Ma 75 Ma 100 Ma 200 Ma
50 60
100
0 Ma 25 Ma 50 Ma 75 Ma 100 Ma 200 Ma
Tension
Compression
0
250 500 750 1000 1250 1500 Temperature (oC)
Strength (MPa) 1500 −1000 −500 0
0
500
1000 1500 2000 2500 Oceanic lithosphere
10 20
Depth (km)
1500 Extended cratonic lithosphere
Moho
90
250 500 750 1000 1250 1500 Temperature (oC)
1000
40
Compression
(e)
500
UM: olivine (dry)
80
0
0
LC: diorite (wet)
70
Tension
−500
UC: quartz (dry)
20
50
250 500 750 1000 1250 1500 Temperature (oC)
Strength (MPa)
Unextended young orogenic lithosphere
UC: quartz (dry)
Compression
0
−1000 0
1500
20
40
Tension
(d)
Strength (MPa) −1000 0
0 Ma 25 Ma 50 Ma 75 Ma 100 Ma 200 Ma
80
250 500 750 1000 1250 1500 Temperature (oC)
1500 Unextended normal cratonic lithosphere
UM: olivine (dry)
90 0
1000
50
Compression
(c)
Depth (km)
20
70 0 Ma 25 Ma 50 Ma 75 Ma 100 Ma 200 Ma
90
10
UC: quartz (dry)
500
40
70 80
0
10
30
LC: diorite (wet)
Depth (km)
Depth (km)
30
−500
Olivine (dry)
30 40 50 60 70 80 90 100
100 Ma 200 Ma 300 Ma 400 Ma 500 Ma
Tension
Compression
0 250 500 7501000 1250 1500
Temperature (oC)
Figure 33 Depth-dependent rheological models for various lithosphere types and a range of geothermal gradients, assuming a dry quartz/diorite/olivine mineralogy for continental lithosphere (Ziegler, 1996a; Ziegler et al., 2001). (a) Unextended, thick-shield-type lithosphere with a crustal thickness of 45 km and a lithospheric mantle thickness of 155 km. (b) Unextended, ‘normal’ cratonic lithosphere with a crustal thickness of 30 km and a lithospheric mantle thickness of 70 km. (c) Unextended, young orogenic lithosphere with a crustal thickness of 60 km and a lithospheric mantle thickness of 140 km. (d) Extended, cratonic lithosphere with a crustal thickness of, 20 km and a lithospheric mantle thickness of 50 km. (e) Oceanic lithosphere. Modified from Ziegler PA, Cloetingh S, Guiraud R, and Stampfli GM (2001) Peri-Tethyan platforms: constraints on dynamics of rifting and basin inversion. In: Ziegler PA, Cavazza W, Robertson AHF, and Crasquin-Soleau S (eds.) Me´moires du Museum National d’Histoire Naturelle 186: Peri-Tethys Memoir 6: Peri-Tethyan Rift/Wrench Basins and Passive Margins, pp. 9–49. Paris: Commission for the Geological Map of the World.
Tectonic Models for the Evolution of Sedimentary Basins
seriously weakened by transform faults and by the thermal blanketing effect of thick sedimentary prisms prograding onto it (e.g., Gulf of Mexico, Niger Delta, Bengal Fan; Ziegler et al., 1998; see also Figure 13). The strength of continental crust depends largely on its composition, thermal regime and the presence of fluids, and also on the availability of preexisting crustal discontinuities (see also Burov, this volume (Chapter 6.03)). Deep-reaching crustal discontinuities, such as thrust- and wrench-faults, cause significant weakening of the otherwise mechanically strong upper parts of the crust. As such discontinuities are apparently characterized by a reduced frictional angle, particularly in the presence of fluids (Van Wees, 1994), they are prone to reactivation at stress levels that are well below those required for the development of new faults. Deep reflection-seismic profiles show that the crust of Late Proterozoic and Palaeozoic orogenic belts is generally characterized by a monoclinal fabric that extends from upper crustal levels down to the Moho at which it either soles out or by which it is truncated (Figure 34) (see Ziegler and Cloetingh, 2004). This fabric reflects the presence of deep-reaching lithological inhomogeneities and shear zones. The strength of the continental upper lithospheric mantle depends to a large extent on the thickness of the crust but also on its age and thermal regime (see Jaupart and Mareschal, this volume (Chapter 6.05)). Thermally stabilized stretched continental lithosphere with a 20 km thick crust and a lithospheric mantle thickness of 50 km is mechanically stronger than unstretched lithosphere with a 30 km thick crust and a 70 km thick lithospheric mantle (compare
Figures 33(b) and 33(d)). Extension of stabilized continental crustal segments precludes ductile flow of the lower crust and faults will be steep to listric and propagate towards the hanging wall, that is, towards the basin center (Bertotti et al., 2000). Under these conditions, the lower crust will deform by distributing ductile shear in the brittle–ductile transition domain. This is compatible with the occurrence of earthquakes within the lower crust and even close to the Moho (e.g., southern Rhine Graben: Bonjer, 1997; East African rifts: Shudofsky et al., 1987). On the other hand, in young orogenic belts, which are characterized by crustal thicknesses of up to 60 km and an elevated heat flow, the mechanically strong part of the crust is thin and the lithospheric mantle is also weak (Figure 33(c)). Extension of this type of lithosphere, involving ductile flow of the lower and middle crust along pressure gradients away from areas lacking upper crustal extension into zones of major upper crustal extensional unroofing, can cause crustal thinning and thickening, respectively. This deformation mode gives rise to the development of core complexes with faults propagating toward the hanging wall (e.g., Basin and Range Province: Wernicke, 1990; Buck, 1991; Bertotti et al., 2000). However, crustal flow will cease after major crustal thinning has been achieved, mainly due to extensional decompression of the lower crust (Bertotti et al., 2000). Generally, the upper mantle of thermally stabilized, old cratonic lithosphere is considerably stronger than the strong part of its upper crust (Figure 33(a)) (Moisio et al., 2000). However, the
NW
SE
RHENOHERC Devonian
SAXOTHURINGIAN Tertiary
TAUNUS WETTERAU Friedberg Hanau 0
Cryst - Bsmt
Bunter
SPESSART Laufach
MOLDANUBIAN Muschelkalk
Keuper
FRANCONIAN PLATFORM Main Wurzburg Tauber Rothenburg
Impact breccia Malm SWAB-ALP RIES Dinkelsbuhl Nordlingen Donauw rth 0
5
5
10 TWT
10 TWT 0
527
250 Km
Figure 34 Crustal fabric of the Variscan Orogen as imaged by the DEKORP 2-S reflection–seismic line, South Germany. Modified from Behr HJ and Heinrichs T (1987) Geological interpretation of DEKORP 2 –S: A deep seismic reflection profile across the Saxothuringian and possible implications for late Variscan structural evolution of Central Europe. Tectonophysics 142: 173–202.
528
Tectonic Models for the Evolution of Sedimentary Basins
occurrence of upper mantle reflectors, which generally dip in the same direction as the crustal fabric and probably are related to subducted oceanic and/or continental crustal material, suggests that the continental lithospheric mantle is not necessarily homogenous but can contain lithological discontinuities that enhance its mechanical anisotropy (Vauchez et al., 1998; Ziegler et al., 1998). Such discontinuities, consisting of eclogitized crustal material, can potentially weaken the strong upper part of the lithospheric mantle. Moreover, even in the face of similar crustal thicknesses, the heat flow of deeply degraded Late Proterozoic and Phanerozoic orogenic belts is still elevated as compared to adjacent old cratons (e.g., Pan African belts of Africa and Arabia; Janssen, 1996). This is probably due to the younger age of their lithospheric mantle and possibly also to a higher radiogenic heat generation potential (a) −1000 0 10
−500
Strength (MPa) 0 500 1000
of their crust. These factors contribute to weakening of former mobile zones to the end that they present rheologically weak zones within a craton, as evidenced by their preferential reactivation during the breakup of Pangea (Ziegler, 1989a, 1989b; Janssen et al., 1995; Ziegler et al., 2001). From a rheological point of view, the thermally destabilized lithosphere of tectonically active rifts, as well as of rifts and passive margins that have undergone only a relatively short postrift evolution (e.g., 25 Ma), is considerably weaker than that of thermally stabilized rifts and of unstretched lithosphere (Figures 33 and 35, Ziegler et al., 1998). In this respect, it must be realized that, during rifting, progressive mechanical and thermal thinning of the lithospheric mantle and its substitution by the upwelling asthenosphere is accompanied by a rise in geotherms causing progressive weakening of the (b)
1500
−1000 0
Normal continental lithosphere
UC: granite
−500
Strength (MPa) 0 500 1000
1500
Extended, thermally rejuvenated lithosphere
UC: granite
10 LC: o-pyroxene
20
LC: o-pyroxene
Moho
20
UM: dunite
30
Moho
30 40
Depth (km)
Depth (km)
40 UM: dunite
50
50
60
60
70
70
80
80 Tension
90
Compression
Tension
Wet
90
Dry
Dry
100
Compression
Wet
0
250 500 750 1000 1250 1500 o
Temperature ( C)
100
0
250 500 750 1000 1250 1500
Temperature (oC)
Figure 35 Depth-dependent rheological models for dry and wet, unextended ‘normal’ cratonic lithosphere and stretched, thermally attenuated lithosphere, assuming a quartz/diorite/olivine mineralogy. (a) Unextended, cratonic lithosphere with a crustal thickness of 30 km and a lithospheric mantle thickness of 70 km. (b) Extended, thermally destabilized cratonic lithosphere with a crustal thickness of, 20 km and a lithospheric mantle thickness of 38 km. Modified from Ziegler PA, Cloetingh S, Guiraud R, and Stampfli GM (2001) Peri-Tethyan platforms: constraints on dynamics of rifting and basin inversion. In: Ziegler PA, Cavazza W, Robertson AHF, and Crasquin-Soleau S (eds.) Me´moires du Museum National d’Histoire Naturelle 186: Peri-Tethys Memoir 6: Peri/Tethyan Rift/Wrench Basins and Passive Margins, pp. 9–49. Paris: Commission for the Geological Map of the World.
Tectonic Models for the Evolution of Sedimentary Basins
extended lithosphere. In addition, its permeation by fluids causes its further weakening (Figure 35). Upon decay of the rift-induced thermal anomaly, rift zones are, rheologically, considerably stronger than unstretched lithosphere (Figure 33). However, accumulation of thick syn- and postrift sedimentary sequences can cause by thermal blanketing a weakening of the strong parts of the upper crust and lithospheric mantle of rifted basins (Stephenson, 1989). Moreover, as faults permanently weaken the crust of rifted basins, they are prone to tensional as well as compressional reactivation (Ziegler et al., 1995, 1998, 2001, 2002). In view of its rheological structure, the continental lithosphere can be regarded under certain conditions as a two-layered viscoelastic beam (Figure 13) (Reston, 1990, ter Voorde et al., 1998). The response of such a system to the buildup of extensional and compressional stresses depends on the thickness, strength and spacing of the two competent layers, on stress magnitudes and strain rates and the thermal regime (Zeyen et al., 1997; Watts and Burov, 2003). As the structure of continental lithosphere is also areally heterogeneous, its weakest parts start to yield first, once tensional as well as compressional intraplate stress levels equate their strength. 6.11.3.2 Mechanical Controls on Basin Evolution: Europe’s Continental Lithosphere Studies on the mechanical properties of the European lithosphere revealed a direct link between its thermotectonic age and bulk strength (Cloetingh et al., 2005, Cloetingh and Burov, 1996; Pe´rezGussinye´ and Watts, 2005). On the other hand, inferences from P- and S-wave tomography (Goes et al., 2000a, 2000b; Ritter et al., 2000, 2001) and thermomechanical modeling (Garcia-Castellanos et al., 2000) point to a pronounced weakening of the lithosphere in the Lower Rhine area owing to high upper mantle temperatures. However, the Late Neogene and Quaternary tectonics of the Ardennes–Lower Rhine area appear to form part of a much wider neotectonic deformation system that overprints the Late Palaeozoic and Mesozoic basins of NW Europe. This is supported by geomorphologic evidence and the results of seismicity studies in Brittany (Bonnet et al., 1998, 2000) and Normandy (Lagarde et al., 2000; Van Vliet-Lanoe¨ et al., 2000), by data from the Ardennes–Eifel region (Meyer and Stets, 1998; Van Balen et al., 2000), the southern parts of the Upper Rhine Graben (Nivie`re and Winter, 2000), the
529
Bohemian Massif (Ziegler and De`zes, 2005, in press), and the North German Basin (Bayer et al., 1999). Lithosphere-scale folding and buckling, in response to the buildup of compressional intraplate stresses, can cause uplift or subsidence of relatively large areas at timescales of a few million years and thus can be an important driving mechanism of neotectonic processes. For instance, the Plio-Pleistocene accelerated subsidence of the North Sea Basin is attributed to down-buckling of the lithosphere in response to the buildup of the present-day stress field (Van Wees and Cloetingh, 1996). Similarly, uplift of the Vosges–Black Forest arch, which at the level of the crust–mantle boundary extends from the Massif Central into the Bohemian Massif (Figure 17), commenced during the Burdigalian (18 Ma) and persisted until at least early Pliocene times. Uplift of this arch is attributed to lithospheric folding controlled by compressional stresses originating at the Alpine collision zone (Ziegler et al., 2002; De`zes et al., 2004; Ziegler and De`zes, 2005, in press). An understanding of the temporal and spatial strength distribution in the NW European lithosphere may offer quantitative insights into the patterns of its intraplate deformation (basin inversion, up-thrusting of basement blocks), and particularly into the pattern of lithosphere-scale folding and buckling. Owing to the large amount of high-quality geophysical data acquired during the last 20 years in Europe, its lithospheric configuration is rather well known, though significant uncertainties remain in many areas about the seismic and thermal thickness of the lithosphere (Babuska and Plomerova, 1992; Artemieva and Mooney, 2001; Artemieva, 2006). Nevertheless, available data help to constrain the rheology of the European lithosphere, thus enhancing our understanding of its strength. So far, strength envelopes and the effective elastic thickness of the lithosphere have been calculated for a number of locations in Europe (e.g., Cloetingh and Burov, 1996). However, as such calculations were made for scattered points only, or along transects, they provide limited information on lateral strength variations of the lithosphere. Although lithospheric thickness and strength maps have already been constructed for the Pannonian Basin (Lankreijer et al., 1999) and the Baltic Shield (Kaikkonen et al., 2000), such maps were until recently not yet available for all of Europe. As evaluation and modeling of the response of the lithosphere to vertical and horizontal loads requires
530
Tectonic Models for the Evolution of Sedimentary Basins
an understanding of its strength distribution, dedicated efforts were made to map the strength of the European foreland lithosphere by implementing 3-D strength calculations (Cloetingh et al., 2005). Strength calculations of the lithosphere depend primarily on its thermal and compositional structure and are particularly sensitive to thermal uncertainties (Ranalli and Murphy, 1987; Ranalli, 1995; Burov and Diament, 1995). For this reason, the workflow aimed at the development of a 3-D strength model for Europe was twofold: (1) construction of a 3-D compositional model and (2) calculating a 3-D thermal cube. The final 3-D strength cube was obtained by calculating 1-D strength envelopes for each lattice point (x, y) of a regularized raster covering NW Europe (Figure 36). For each lattice point, the appropriate input values were obtained from a 3-D compositional and thermal cube. A geological and geophysical geographic database was used as reference for the construction of the input models.
For continental realms, a 3-D multilayer compositional model was constructed, consisting of one mantle-lithosphere layer, 2–3 crustal layers and an overlying sedimentary cover layer, whereas for oceanic areas a one-layer model was adopted. For the depth to the different interfaces several regional or European-scale compilations were available that are based on deep seismic reflection and refraction or surface wave dispersion studies (e.g., Panza, 1983; Calcagnile and Panza, 1987; Suhadolc and Panza, 1989; Blundell et al., 1992; Du et al., 1998; Artemieva et al., 2006). For the base of the lower crust, we strongly relied on the European Moho map of De`zes and Ziegler (2004) (Figure 37). Regional compilation maps of the seismogenic lithosphere thickness were used as reference to the base of the thermal lithosphere in subsequent thermal modeling (Babuska and Plomerova, 1993, 2001; Plomerova et al., 2002). Figure 38(a) shows the integrated strength under compression of the entire lithosphere of Western and Thermal models
Compositional models
Cold
Sediments Upper Lower
crust
o oh M
crust
Mantle lithosphere
Hot
Local geotherm
Asthenosphere
Strength models
Local geotherm
Base thermal lithosphere 1300°C isotherm
Mechanically strong crust (MSC)
Base crust (Moho) Mechanically strong mantle lithosphere Figure 36 From crustal thickness (top left) and thermal structure (top right) to lithospheric strength (bottom): conceptual configuration of the thermal structure and composition of the lithosphere, adopted for the calculation of 3-D strength models. Modified from Cloetingh S, Ziegler PA, Beekman F, Andriessen PAM, Mat¸enco L, Bada G, Garcia-Castellanos D, Hardebol N, De´zes P, and Sokoutis D (2005) Lithospheric memory, state of stress and rheology: Neotectonic controls on Europe’s intraplate continental topography. Quaternary Science Reviews 24: 241–304.
0
356
352
348
531
24
20
16
12
8
4
344
Tectonic Models for the Evolution of Sedimentary Basins
60
60
56
56
52
52
48
48
44
44
36
16
12
8
4
0
36
356
40
352
40
10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62 Kilometres
Projection: lambert azimuthal equal area; center: 04°.00"/48°.00"; Region: W/E/N/S = 350°/28°/62°/34°; Ellipsoide wgs-84
Figure 37 Depth map of Moho discontinuity (2 km contour interval) for Western Europe, constructed by integration of published regional maps. For data sources, see http://comp1.geol.unibas.ch/. Red lines (solid and stippled) show offsets of the Moho discontinuities. Modified from De`zes P and Ziegler PA (2004) Moho depth map of western and central Europe. EUCOR-URGENT homepage: http://www.unibas.ch/eucor-urgent (acessed Jul 2007).
532
Tectonic Models for the Evolution of Sedimentary Basins
10°0'0"W
0°0'0"E
Moreover, a broad zone of weak lithosphere characterizes the Massif Central and surrounding areas. The presence of thickened crust in the area of the Teisseyre–Tornquist suture zone (Figure 37) gives rise to a pronounced mechanical weakening of the lithosphere, particularly of its mantle part. Whereas the lithosphere of Fennoscandia is characterized by a relatively high strength, the North Sea rift system corresponds to a zone of weakened lithosphere. Other areas of high lithospheric strength are the Bohemian Massif and the London–Brabant Massif both of which exhibit low seismicity (Figure 39). A pronounced contrast in strength can also be noticed between the strong Adriatic indenter and the weak Pannonian Basin area (see also Figure 38). Comparing Figure 38(a) with Figure 38(b) reveals that the lateral strength variations of Europe’s
10°0'0"E
20°0'0"E
60°0'0"N 30°0'0"E
50°0'0"N
(a)
40°0'0"N
50°0'0"N
60°0'0"N
Central Europe, whereas Figure 38(b) displays the integrated strength of the crustal part of the lithosphere. As is evident from Figure 38, Europe’s lithosphere is characterized by major spatial mechanical strength variations, with a pronounced contrast between the strong Proterozoic lithosphere of the East-European Platform to the east of the Teisseyre–Tornquist line and the relatively weak Phanerozoic lithosphere of Western Europe. A similar strength contrast occurs at the transition from strong Atlantic oceanic lithosphere to the relatively weak continental lithosphere of Western Europe. Within the Alpine foreland, pronounced NE–SE trending weak zones are recognized that coincide with such major geologic structures as the Rhine Graben System and the North Danish–Polish Trough, which are separated by the high-strength North German Basin and the Bohemian Massif.
20°0'0"E Fault Normal fault 10°0'0"E
Thrust fault
10°0'0"W
Integrated strength (1012 N m–1)
Figure 38 (Continued )
0°0'0"E
~0
10
20
30
40 50
85
10°0'0"W
0°0'0"E
10°0'0"E
20°0'0"E
533
60°0'0"N 30°0'0"E
50°0'0"N
(b)
40°0'0"N
50°0'0"N
60°0'0"N
Tectonic Models for the Evolution of Sedimentary Basins
Fault
20°0'0"E
Normal fault
10°0'0"E
Thrust fault
10°0'0"W
Integrated strength (1010 N m–1)
0°0'0"E
1
10
20 30 40 60 100 200
Figure 38 Integrated strength maps for intraplate Europe. Contours represent integrated strength in compression for (a) total lithosphere and (b) crust. Adopted composition for upper crust, lower crust, and mantle is based on a wet quartzite, diorite, and dry olivine composition, respectively. Rheological rock parameters are based on Carter and Tsenn (1987) The adopted bulk strain-rate is 1016 s1, compatible with constraints from GPS measurements (see text). The main structural features of Europe are superimposed on the strength maps. Modified from Cloetingh S, Ziegler PA, Beekman F, Andriessen PAM, Mat¸enco L, Bada G, Garcia-Castellanos D, Hardebol N, De´zes P, and Sokoutis D (2005) Lithospheric memory, state of stress and rheology: Neotectonic controls on Europe’s intraplate continental topography. Quaternary Science Reviews 24: 241–304.
intraplate lithosphere are primarily caused by variations in the mechanical strength of the lithospheric mantle, whereas variations in crustal strength appear to be more modest. The variations in lithospheric mantle strength are primarily related to variations in the thermal structure of the lithosphere that can be related to thermal perturbations of the sublithospheric upper mantle imaged by seismic tomography (Goes et al., 2000a), with lateral variations in crustal thickness playing a secondary role, apart from Alpine domains which are characterized by deep crustal roots. High strength in the EastEuropean Platform, the Bohemian Massif, the
London–Brabant Massif and the Fennoscandian Shield reflects the presence of old, cold, and thick lithosphere, whereas the European Cenozoic Rift System coincides with a major axis of thermally weakened lithosphere within the Northwest European Platform. Similarly, weakening of the lithosphere of southern France can be attributed to the presence of tomographically imaged plumes rising up under the Massif Central (Granet et al., 1995; Wilson and Patterson, 2001). The major lateral strength variations that characterize the lithosphere of extra-Alpine Phanerozoic Europe are largely related to its Late Cenozoic
Tectonic Models for the Evolution of Sedimentary Basins
0°0'0"E
10°0'0"E
20°0'0"E
60°0'0"N 30°0'0"E
50°0'0"N
10°0'0"W
40°0'0"N
50°0'0"N
60°0'0"N
534
20°0'0"E Fault Normal fault 10°0'0"E
Thrust fault
10°0'0"W
Earthquake epicentre Integrated strength (1010 N m–1) 0°0'0"E 1
10
20
40 60 100 200
Figure 39 Distribution of crustal seismicity superimposed on map of integrated strength for the European crust (see Figure 38b). Earthquake epicenters from the NEIC data center (NEIC, 2004), queried for magnitude >2 and focal depths <35 km.
thermal perturbation as well as to Mesozoic and Cenozoic rift systems and intervening stable blocks, and not so much to the Caledonian and Variscan orogens and their accreted terranes (De`zes et al., 2004, Ziegler and De`zes, 2006). These lithospheric strength variations (Figure 38(a)) are primarily related to variations in the thermal structure of the lithosphere and, therefore, are compatible with inferred variations in the EET of the lithosphere (see Cloetingh and Burov, 1996; Pe´rez-Gussinye´ and Watts, 2005). The most important strong inliers in the lithosphere of the Alpine foreland lithosphere correspond to the Early Palaeozoic London–Brabant Massif and the Variscan Armorican, Bohemian and West-Iberian Massifs. The strong Proterozoic Fennoscandian–
East-European Craton flanks the weak Phanerozoic European lithosphere to the northeast, whereas the strong Adriatic indenter contrasts with the weak lithosphere of the Mediterranean collision zone. Figure 39 displays on the background of the crustal strength map the distribution of seismic activity, derived from the NEIC global earthquake catalogue (USGS). As is obvious from this figure, crustal seismicity is largely concentrated on the presently still active Alpine plate boundaries, and particularly on the margins of the Adriatic indenter. In the Alpine foreland, seismicity is largely concentrated on zones of low lithospheric strength, such as the European Cenozoic rift system, and areas where preexisting crustal discontinuities are reactivated under the presently prevailing NW-directed stress field, such as
Tectonic Models for the Evolution of Sedimentary Basins
the South Armorican shear zone (De`zes et al., 2004; Ziegler and De`zes, in press) and the rifted margin of Norway (Mosar, 2003). It should be noted that the strength maps presented in Figure 38 do not incorporate the effects of spatial variations in the composition of crustal and mantle layers. Future work will address the effects of such second-order strength perturbations, adopting constraints on the composition of several crustal and mantle layers provided by seismic velocities (Guggisberg et al., 1991; Aichroth et al., 1992) and crustal and upper-mantle xenolith studies (Mengel et al., 1991; Wittenberg et al., 2000).
6.11.4 Northwestern European Margin: Natural Laboratory for Continental Breakup and Rift Basins The Atlantic margin of Mid-Norway (Figure 40) is one of the best documented continental margins of the world owing to the availability of a wealth of highquality industry data and intense research collaboration between industrial, governmental, and academic institutions (Torne´ et al., 2003; Mosar, 2003). For this reason, the Mid-Norway margin serves as a natural laboratory for studying process controlling the development of a passive margin, the crustal separation phase of which was accompanied by extensive magmatic activity. Moreover, it permits to analyze, for passive margins, enigmatic features such as its postbreakup partial inversion and erosional truncation in near-shore areas in conjunction with uplift of the Norwegian Caledonides, the controlling mechanisms of which have since long been debated. Studies on the North Atlantic margins have led to the development of a new generation of models for controls on continental breakup and the subsequent evolution of ocean–continent boundary zones, quantification of melting processes, lateral migration of rifting activity and associated vertical motions (Figure 41) (see also Cloetingh et al., 2005). It appears that breakup processes have set the stage for subsequent tectonic reactivation of the margin under a compressional stress regime and overprinting upper-mantle thermal anomalies (Ziegler and Cloetingh, 2004). Systematic quantification of uplift and erosion of the marginal Caledonian highlands by thermogeochronology and other techniques (Figure 42) has revealed pronounced along-strike variations in the magnitude of their late-stage uplift and related geomorphologic development.
535
6.11.4.1 Extensional Basin Migration: Observations and Thermomechanical Models In the literature, several examples of extensional basins have been described in which the locus of extension shifted in time toward the zone of future crustal separation (e.g., Bukovics and Ziegler, 1985; Ziegler, 1988, 1996b; Lundin and Dore´, 1997). These basins typically consist of several laterally adjacent fault-controlled sub-basins the main sedimentary fill of which often differs by several tens of millions of years. As such, they reflect lateral changes in their main subsidence phases that can be attributed to a temporal lateral shift of the centers of lithosphere extension. A well-documented example of this phenomenon is the Mid-Norway passive continental margin. Prior to the final Late Cretaceous–Early Tertiary rifting event that culminated in continental breakup, the Mid-Norway Vøring margin was affected by several rifting events (Ziegler, 1988; Bukovics and Ziegler, 1985; Skogseid et al., 1992). These resulted in the subsidence of a sequence depocenters located between the Norwegian coast and the continent– ocean boundary (Figure 43). On a basin-wide scale, this passive margin consists of the Late Palaeozoic– Triassic Trøndelag Platform, located in the east, the Jurassic–Early Cretaceous Vøring Basin in the central part of the shelf, and adjacent to the continent– ocean boundary the marginal extended zone that was active during the final stretching event preceding continental breakup at the Paleocene–Eocene transition (Lundin and Dore´, 1997; Reemst and Cloetingh, 2000; Skogseid et al., 2000; Osmundsen et al., 2002). The Nordland Ridge is a classical footwall uplift that is associated with the border fault of the Vøring Basin, whilst the Vøring Marginal High is associated with the continent–ocean boundary. Although there is still considerable uncertainty about the age of the oldest syn-rift sedimentary sequences involved in the different deep-seated fault blocks of the Mid-Norway margin and the width of the area that was affected by the pre-Jurassic extensional phases (Gabrielsen et al., 1999; Mosar et al, 2002), it is generally agreed that in time extension shifted westward, away from the Trøndelag Platform towards the Vøring Basin and ultimately centered on the continental breakup axis (e.g., Bukovics and Ziegler, 1985; Lundin and Dore´, 1997; Reemst and Cloetingh, 2000). Other areas in which extensional strain concentrated in time toward the rift axis or the future
Tectonic Models for the Evolution of Sedimentary Basins
20°
18°
6°
4°
70°
FC
. re Z actu
70°
TF
0
0°
72°
TKF
50
2°
8°
30°
Barents sea
ja Fr Se n
72°
24°
22°
28°
0
10°
16°
14°
12°
26°
–5
536
0
0
0 0
N
B
0
–50
CO
0
0 –5
0
yen ma
50
Jan
–5 0
n tra rm
ba
00
for
Vø
B
64°
Trø
nd
CO
0
ela gp
66°
–1
m
rin
–50
sfo
g
66°
ULO
n si
lat
68°
68°
0
–5 0
Oceanic lithosphere
–100
64°
F
0
B
0
0 –5
62°
–5 0
SB
0
Møre basin
0
Gra
–5
60°
00
50
60
–50
–50
OGR
UWG
50 40 30 20 10 0 –10 –20 –30 –40 –50 –60 –70 –80 –90 –100 –110 –120 –130 –140
0
ing
–1
Vik
60°
ES
0
ben
øFC
62°
58° 58°
0
SK
2° 0
0
0
56° 56° 0
6°
8°
10°
12°
14°
16°
18°
mGal
Tectonic Models for the Evolution of Sedimentary Basins
(a) Trondelag platform
537
(b) Donna terrace
0
2 km 6507/ 7-1 6507/3-1 6609/5-1 6507/2-1
6609/11-1 6508/5-1 6610/ 7-1 6610/ 7-2 (c) Nordland ridge
(d) Ras and traena basin
0
P Tr
6607/12-1 6607/5-1
6609/ 7-1 6507/6-1
2 km Jur 200
Cret 150 100 Age (Ma)
Tert 50
P Tr
Jur 200
Cret 150 100 Age (Ma)
Tert 50
Figure 41 Tectonic subsidence curves for four sub-basins and platforms of the Mid-Norwegian margin. Note the simultaneously occurring subsidence accelerations (colored bars). The Mesozoic subsidence accelerations reflect pronounced rift-related tectonic phases, whereas the Late Neogene subsidence acceleration (light yellow bar) coincides with a major reorganization of the Northern Atlantic stress field. Modified from Reemst, P (1995) Tectonic modeling of rifted continental margins; Basin evolution and tectono-magmatic development of the Norwegian and NW Australian margin. PhD thesis, Vrije Universiteit, Amsterdam, 163p.
breakup axis are, for instance, the Viking Graben of the North Sea (Ziegler, 1990b), the nonvolcanic Galicia margin and the passive rifted ancient South Alpine margin (e.g., Manatschal and Bernoulli, 1999; Bertotti et al., 1997). Several hypotheses have been proposed to explain this rift migration by invoking the principle that rifting occurs where the lithosphere is weakest (Steckler and Ten Brink, 1986). A mechanism for limiting extension at a given location was studied, for example, by England (1983), Houseman and England (1986), and Sonder and England (1989), who found that cooling of the continental lithosphere
during stretching may increase its strength, so that deformation shifts to a previously low strain region (Sonder and England, 1989). With this mechanism, other effects such as changes in plate boundary forces are not required to explain basin migration. Migration of the extension locus has also been attributed to temporally spaced multiple stretching phases, separated by periods during which the lithosphere is not subjected to extension. Under such a scenario, the lithosphere that was weakened during an earlier stretching phase needs sufficient time to cool and regain its strength and to become indeed stronger than the adjacent unextended area before
Figure 40 Bouguer gravity anomaly map of the onshore parts of the Scandinavian North Atlantic passive margin, showing superimposed rift-related normal fault systems. The width of the passive margin between the continent-ocean boundary (COB), marked by the dashed fat black line, and the innermost boundary fault system (IBF), marked by the dashed fat red line, ranges from 550 km in the south to over 700 km in Mid-Norway to 165 km north of Lofoten. Note the onshore extension of the rift system. The wiggly black barbed line marks the position of the Caledonian Thrust Front. Modified from Mosar J (2003). Scandinavia’s North Atlantic passive margin. Journal of Geophysical Research 108: 2630.
538
Tectonic Models for the Evolution of Sedimentary Basins
256 Ma BH-F18
269 Ma BH-F34 320 Ma BH-F35
205 Ma BH-F16
306 Ma BH-F37
214 Ma BH-F22 145 Ma BH-RAM2
164 Ma BH-SE2
166 Ma BH-RAM1 128 Ma BH-N33 142 Ma BH-N34
125 Ma BH-Værøy010
M a
124 Ma BH-Røst2
14 2
171 Ma BH-TJ11
174 Ma TR-RT3 191 Ma TR-RT4 315 Ma JPS-5
300 Ma TR-S11
14 2M a
M a
11 0
137 Ma MR-93-5
119 Ma MR-93-3 106 Ma MR-JOT17
122 Ma MR-93-1 113 Ma MR-EID17
142 Ma MR-T17
224 Ma MR-M5
184 Ma MR-92-15
178 Ma MR-M6
181 Ma MR-92-1
217 Ma MR-OR11 183 Ma 164 Ma MR-O32 MR-T14
171 Ma MR-92-3
185 Ma MA-OR52 186 Ma MR-OR54 183 Ma 209 Ma MR-OR22 MR-K1
a 5M 20
239 Ma MR-T3B
194 Ma MR-OR14 217 Ma MR-B9
M a 509 Ma YH-BY17 376 Ma CC-C9506
170 Ma MR-O28 199 Ma MR-O33
258 Ma CC-9502 284 Ma PA-M2 178 Ma MR-OR20 176 Ma MR- Ø6
235 Ma 190 Ma CC-N9611 228 Ma MR-M7 210 Ma PA-OR15 196 Ma CC-9501 MR- Ør12 238 Ma 217 Ma CC-N9610 MR- Ør10 179 Ma 188 Ma CC-S9623 161 Ma MR-OR46 MR-Ør2 216 Ma CC-T245 208 Ma CC-S9819 313 Ma CC-B14
292 Ma CC-9505
246 Ma CC-S9625
502 Ma
282 Ma GM-98043
365 Ma GM-98038 393 Ma GM-98036
a 5M 20
100
200
Andriessen - VU
300 Ma GM-99002 352 Ma GM-98046 783 Ma GM-98047 372 Ma GM-98048 493 Ma GM-98049
Hendriks - VU Huigen - VU
371 Ma GM-98050
Murrell - VU
Rohrman - VU
Paleogene Cretaceous Jurassic
110 Ma
Grønlie - NGU/Statoil
142 Ma
160 Ma CC-9902
Triassic
205 Ma
300 400 km
Stiberg - Sintef
248 Ma
Permian 290 Ma
Carboniferous
Cederbom - Göteborg
354 Ma
Devonian Silurian
417 Ma
Lorencak - Melbourne
495 Ma
Lehtovaara - Turku
Ordovician 0
1027 Ma ML-17-73
Redfield - NGU/ VU Neogene
a M
164 Ma CC-9903 172 Ma CC-SÅ9629 189 Ma CC-9805
449 Ma ML-A220
AFT age contours 8 24
189 Ma CC-9905 231 Ma CC-9802 240 Ma CC-9807
249 Ma GM-98039
29 0M a
282 Ma CC-S9626 196 Ma CC-P9904
646 Ma ML-A223
510 Ma ML-19-73
331 Ma GM-98029 320 Ma GM-98028 756 Ma ML-9-73 400 Ma GM-99001
17 CC-C9509 Ma 354 Ma
257 Ma CC-P9908
325 Ma CC-A9632 281 Ma 242 Ma CC-9801 CC-A9635
20 5M a
381 Ma CC-C9508 4
426 Ma ML-A596
390 Ma 665 Ma GM-99003 GM-98033
442 Ma GM-98031
213 Ma CC-P9906
161 Ma CC-P9910
231 Ma CC-9808
375 Ma GM-98025
345 Ma GM-99004
480 Ma GM-98050
260 Ma CC-9504 261 Ma CC-S9624
219 Ma CC-S9621 313 Ma CC-S9620
174 Ma CC-9901
524 Ma ML-A121
338 Ma CC-C9507 443 Ma CC-N9617
263 Ma CC-N9614
211 Ma CC-S9622 149 Ma CC-P9909
a 5M 20
388 Ma CC-N9616
253 Ma CC-9503 197 Ma CC-N9612 241 Ma CC-N9613
783 Ma GM-99013
454 Ma GM-99005
Ma 417
195 Ma MR-OR53 163 Ma MR-OR9
168 Ma PA-OR39
480 Ma GM-98024
464 Ma GM-Lavia-A
273 Ma YH-MA1 221 Ma MR-Ø3
511 Ma GM-98021 463 Ma GM-99008
a M
205 Ma
98 Ma MR-EID16
361 Ma GM-99010
368 Ma GM-99009
5 49
202 Ma MR-92-14
402 Ma CC-SÅ9628
337 Ma GM-99011
459 Ma 527 Ma GM-98022 ML-A82
428 Ma GM-98023
347 Ma YH-BY16
607 Ma ML-A339
584 Ma GM-99012
348 Ma GM-98018 403 Ma GM-98019
354 Ma YH-BY15
8 24
143 Ma TR-RT15
354 Ma
Ma 545
112 Ma TR-S2 93 Ma TR-S1 75 Ma JPS-15
110 Ma MR-93-6
317 Ma CC-SÅ9627
617 Ma GM-98011RC
635 Ma GM-98009
301 Ma GM-98017
336 Ma GM-98016
M a
648 Ma GM-98010
658 Ma ML-35-73
305 Ma YH-BY12 378 Ma YH-BY13
812 Ma ML-30-73
495 Ma
153 Ma MR-93-4
224 Ma AG-FT2 214 Ma 201 Ma AG-AG5 AG-AG3 229 Ma AG-FT5 140 Ma TR-RT8 215 Ma 142 Ma YH-MA9 TR-RT10 132 Ma TR-RT11 178 Ma TR-RT13 244 Ma 202 Ma YH-MA6 TR-RT14 250 Ma JPS-8 Ma
a 0M 29
41 7
a M
205 Ma
217 Ma YH-V7
167 Ma AG-AG4 175 Ma TR-RT5 144 Ma AG-AG1
197 Ma JPS-10 147 Ma TR-S6 165 Ma TR-S5 159 Ma TR-S4
550 Ma GM-98012
5 49
255 Ma JPS-12 236 Ma JPS-13
282 Ma YH-BY9
315 Ma JPS-3
164 Ma TR-S9
471 Ma GM-99023
527 Ma GM-99033
178 Ma TR-RT1
547 Ma GM-99024 311 Ma GM-SI202
435 Ma GM-99032
324 Ma JPS-2
267 Ma TR-S10
248 Ma JPS-11 226 Ma JPS-14
a 5M 49
Ma 290
Ma 417
280 Ma TR-S16
280 Ma GM-99030 268 Ma BH-GRM3
54 5M a
307 Ma JPS-1
163 Ma TR-S8 282 Ma JPS-6 144 Ma TR-RT9 302 Ma JPS-7 142 Ma TR-RT7 160 Ma TR-RT6
264 Ma GM-KOV596
243 Ma BH-P2 114 Ma BH-N57 144 Ma BH-N60
35 4
230 Ma TR-S15
217 Ma TR-S12
490 Ma GM-99027
290 M a
256 Ma GM-MT7
278 Ma GM-KOV32 287 Ma BH-F46
191 Ma TR-RT2
205 Ma
206 Ma TR-S13
318 Ma BH-MT4 304 Ma BH-MT3
365 Ma BH-F48
225 Ma BH-L1
20 5M a
208 Ma TR-S14
215 Ma GM-MT6
215 Ma GM-MT5
124 Ma BH-N5 124 Ma BH-N41
111 Ma BH-N46 89 Ma BH-N48
142 Ma
129 Ma BH-N15
120 Ma BH-F53
a M
180 Ma BH-N39
296 Ma BH-F28
170 Ma BH-F1
134 Ma BH-N2 126 Ma BH-N1
147 Ma BH-H7
337 Ma BH-MT1
289 Ma BH-F42
8 24
111 Ma BH-KF7 129 Ma BH-KF14
247 Ma BH-LAN3 196 Ma BH-OP7
320 Ma BH-MT2
237 Ma BH-F40
214 Ma BH-F6
203 Ma BH-T4 173 Ma BH-SE4 143 Ma BH-N11
127 Ma BH-AND3
149 Ma BH-N23 180 Ma BH-KF2
216 Ma BH-T9
a 0M 29
238 Ma BH-RVN9 177 Ma BH-MOO17
239 Ma BH-F13 288 Ma 268 Ma BH-F25 BH-F11
168 Ma BH-N13
Cambrian
545 Ma
Figure 42 Digital elevation model of Scandinavia. Contour lines show vertical motions in My derived from AFT analyses.
the onset of the next stretching event. Obviously, this concept requires a long period of tectonic quiescence between successive rifting events. Bertotti et al., (1997) showed in a model for the thermomechanical evolution of the South Alpine rifted margin that its strongly thinned parts indeed could have been stronger than the remainder of the margin. This is compatible with rheological considerations which suggest that stretched and thermally stabilized lithosphere, characterized by a thinned crust and a
considerably stronger lithospheric mantle, is considerably stronger than unextended lithosphere (Figure 33) (Bertotti et al., 1997). This hypothesis requires sudden time-dependent changes in the magnitude and possibly the orientation of intraplate stress fields, controlling the different stretching and nonstretching phases. Nevertheless, evidence for tensional reactivation of rifts which were abandoned millions of years ago suggests that crustal-scale faults permanently weaken their lithosphere to the degree
Tectonic Models for the Evolution of Sedimentary Basins
CO Vø
rin
B
g
h
ig al h
margin
rw
ay
Trøndelag Platform
No
No
rd
lan
dr
idg
e
Vøring basin
0
100 km
Permo-Triassic rift zone Jurassic-Cretaceous rift zone Breakup related rift zone
Figure 43 Sketch map of rift zones on the Mid-Norwegian margin showing timing of main rift activity. Modified from Skogseid J and Eldholm O (1995) Rifted continental margin off mid-Norway. In: Banda E, Talwani E, and Torne´ M (eds.) Rifted Ocean-Continent Boundaries, pp. 147–153. Dordrecht: Kluwer Academic.
that under given conditions they are prone to tensional and compressional reactivation (Ziegler and Cloetingh, 2004). Sawyer and Harris (1991) and Favre and Stampfli (1992) have described for the evolution of the Central Atlantic rift a gradual concentration of extensional strain towards the future zone of crustal separation. This phenomenon is attributed to the syn-rift gradual rise of the lithospheric isotherms and a commensurate upward shift of the lithospheric necking level and the intracrustal brittle/
0 0
1/2 Vext
Depth (km)
20
σ (MPa) xx
200
400
539
ductile deformation boundary. As a result of this, extensional strain concentrates in time on the thermally more intensely weakened part of the lithosphere, generally corresponding to the rift axis, thus causing narrowing of the rift and abandonment of its lateral graben systems (Ziegler and Cloetingh, 2004). In the case of the Central Atlantic, this process was enhanced by the impingement of a short-lived, though major, mantle plume at the Triassic–Jurassic transition (Wilson, 1997; Nikishin et al., 2002). In the following, we discuss the implications of a viscoelastic plastic finite element model for extension of a lithosphere with an initially symmetric upper-mantle weakness (for details see Van Wijk and Cloetingh, 2002). When large extension rates are applied, focusing of deformation takes place, causing lithospheric necking and eventually continental breakup. Hereafter, this will be referred to as ‘standard’ rifting. A different evolution of deformation localization takes place when the lithosphere is extended at low strain rates. In this case, the necking area may start to migrate laterally, and hence prevent continental breakup. Modeling results will be compared to observations on basin migration at the Mid-Norway, Galicia and ancient South Alpine margins. The modeled domain consists of an upper and lower crust and a lithospheric mantle (Figure 44) to which the rheological parameters of granite, diabase and olivine, respectively, have been assigned (see Table 4) (Carter and Tsenn, 1987). In order to facilitate deformation localization, the crust was thickened by 2 km in the center of the domain, forming a linear feature parallel to the future rift axis. This
T = 0°C Upper crust (granite) Lower crust (diabase) 1/2 Vext
40
Mantle lithosphere (olivine) 60 80 100 120
T = 1333°C
Figure 44 Pre-extensional configuration of the lithospheric model and horizontal deviatoric stress field. vext is extension velocity. Note thicker crust in the central part of the model domain. Upon volume preserving stretching of the lithosphere, the width of the model domain increases while its thickness decreases. Modified from Van Wijk JW and Cloetingh S (2002) Basin migration caused by slow lithospheric extension. Earth and Planetary Science Letters 198: 275–288.
540
Tectonic Models for the Evolution of Sedimentary Basins
Table 4
Material parameter values for the numerical modeling results presented in Figures 45–51
Parameter
Layer
Value
Density
Upper crust Lower crust Mantle lithosphere
2700 kg m3 2800 kg m3 3300 kg m3 105 K1 106 W m3 1050 J kg1 K1 2.6 W m1 K1 3.1 W m1 K1 3.3 1010 Pa 12.5 1010 Pa 2 1010 Pa 6.5 1010 Pa 3.3 3.05 3.0 186.7 kJ mol1 276 kJ mol1 510 kJ mol1 3.16 1026 Pa n s1 3.2 1020 Pa n s1 7.0 1014 Pa n s1 30 0 20 106 Pa 125 km
Thermal expansion Crustal heat production Specific heat Conductivity Bulk modulus Shear modulus Power law exponent ‘n’
Activation energy ‘Q’
Material constant ‘A’
Internal friction angle Internal dilatation angle Cohesion (Initial) thickness of model domain H
causes localized weakening of the upper mantle and a corresponding strength reduction of the lithosphere. A lithosphere with this configuration is thought to be typical for orogenic belts after postorogenic thermal equilibration of their thickened crust (Henk, 1999). The Mid-Norway margin is superimposed on the Caledonian orogen, the crust of which was presumably still thickened prior to the Late Carboniferous onset of rifting (Skogseid et al., 1992). Appendix 1 discusses fundamental aspects of the adopted dynamical model for slow extension. The model is not prestressed and is subjected to horizontal extension applied at its right and left edges (Figure 44). The tested constant extension rates range between 3 and 30 mm yr1 and are compatible with present-day plate velocities derived from the Global Positioning System (Argus and Heflin, 1995). The surface of the model is unconstrained whilst for its base a vertical velocity component is prescribed calculated from the principle of volume conservation. Temperatures are calculated using the heat conduction equation. The initial geotherm is in steady state. The temperature at the surface is 0 C, and at the base of the model (125 km depth) it is 1333 C. The heat
Crust Mantle lithosphere Crust Mantle lithosphere Crust Mantle lithosphere Upper crust Lower crust Mantle lithosphere Upper crust Lower crust Mantle lithosphere Upper crust Lower crust Mantle lithosphere
flow through the right and left sides of the domain is zero. The crustal heat production is constant (Table 4). 6.11.4.2 Fast Rifting and Continental Breakup As rift structures resulting from various high extension rates (‘standard necking cases’) do not differ significantly, one representative test in which the lithosphere was stretched with a total velocity of about 16 mm yr1 is discussed here. At the onset of lithospheric extension deformation localizes in the center of the domain where the initial mantle weakness was introduced. Thinning of the crust and mantle lithosphere concentrates here, and mantle material starts to well up (Figure 45). Thinning of the crust and mantle lithosphere continues, eventually resulting in continental breakup after 27 My of stretching. Continental breakup is here defined as occurring when the crust is thinned by a factor 20, though another factor or another definition for continental breakup could have been chosen. This definition corresponds to about 40–50% of extension of the model domain at continental breakup. Thinning factors for the crust () and
Tectonic Models for the Evolution of Sedimentary Basins
t = 2 My
T (°C)
0 Depth (km)
20
1200 1000 800 600 400 200
40 60 80 100 120
t = 10 My
t = 20 My
0
0
20
1200 20 1000 800 40 600 400 60 200
40 60 80
500
1000 (km) 0
1200 1000 800 600 400 200
80
100 0
541
500
1000 (km)
0
500
1000 (km)
Figure 45 Thermal evolution of the lithosphere 2, 10, and 20 My after the onset of stretching at a rate of vext 16 mm yr1. Note the changing horizontal and vertical scales in the panels indicating the changing sizes of the model domain upon stretching. Modified from Van Wijk JW and Cloetingh S (2002) Basin migration caused by slow lithospheric extension. Earth and Planetary Science Letters 198: 275–288.
Crustal thinning 25 9 8 7 6 5 4 3 2 1.5 1.2
Time (My)
20
15
10
5
0
0
500
1000 (km)
Mantle thinning 25 2 1.9 1.8 1.7 1.6 1.5 1.4 1.3 1.2 1.1 1
20
Time (My)
lithospheric mantle () are shown in Figure 46. The base of the lithospheric mantle is the 1300 C isotherm, and and are defined as the ratio between the initial and the present thickness of the crust or lithospheric mantle, respectively. The total (integrated) strength of the lithosphere during rifting is shown in Figure 47. The center of the model domain, on which the initial mantle weakness was imposed, is the weakest part; this continues to be so until continental breakup. Integrated strength values of the continental lithosphere vary between 1012 and 1013 N m1 with the higher values characterizing Precambrian shields (Ranalli, 1995; England, 1986). In our model the strength of the lithosphere falls within this range. During rifting, the strength of the lithosphere decreases with time, owing to its thinning and heating as a consequence of its stretching and nonlinear rheology. In other cases with higher constant extension rates the localization of deformation is comparable to that discussed above. In all cases, deformation concentrates on one zone in which thinning continues until continental breakup is achieved. The duration of the rifting stage preceding continental breakup depends on the extension velocity. When the lithosphere is stretched at grater velocities, it takes less time to reach continental breakup (Figure 48). When extension velocities are less than 8 mm yr1, stretching of the lithosphere does not lead to continental breakup. The dependence of rift duration on potential mantle temperature is discussed in Van Wijk et al., (2001). The configuration of the rift shows no clear dependence on the tested stretching velocities (see also Bassi, 1995). The tendency for the lithosphere to neck (or to focus strain) is weaker with decreasing extension velocities. When stresses exerted on the lithosphere
15
10
5
0
0
500
1000 (km)
Figure 46 Evolution of thinning factors of crust () and mantle lithosphere () for vext 16 mm yr1. Breakup after 27 My. Width of the model domain (horizontal axis) vs. time (vertical axis). The width of the model domain increases as the lithosphere is extended. Modified from Van Wijk JW and Cloetingh S (2002) Basin migration caused by slow lithospheric extension. Earth and Planetary Science Letters 198: 275–288.
542
Tectonic Models for the Evolution of Sedimentary Basins
25
1012 N m–1 7.2 6.8 6.4 6.0 5.6 5.2 4.8 4.4
Time (My)
20
15
10
5 0
500
1000 (km) RH Figure 47 Evolution of lithosphere strength ( 0 (z) dz), in N m1, for vext 16 mm yr1. Modified from Van Wijk JW and Cloetingh S (2002) Basin migration caused by slow lithospheric extension. Earth and Planetary Science Letters 198: 275–288.
40
20
igratio
n
60
*
Rift m
Breakup time (My)
80
* * 5
20 15 25 10 Extension velocity (mm yr–1)
*
* 30
Figure 48 Rift duration until continental breakup as a function of total extension velocity (stars). Rifting at larger extension rates eventually results in breakup, while at lower rates syn-tectonic cooling prevails causing rift migration before breakup is achieved. Modified from Van Wijk JW and Cloetingh S (2002) Basin migration caused by slow lithospheric extension. Earth and Planetary Science Letters 198: 275–288.
are smaller, the rate of strain localization also slows down, mantle upwelling is slower and syn-rift lateral conductive cooling plays a more important role. When the lithosphere is stretched at high rates, upwelling of mantle material is fast (almost adiabatic) with little or no horizontal heat conduction. The fast rise of hot asthenospheric material further reduces the strength of the lithosphere in the central region, with the consequence that lithospheric deformation and thinning accelerates even further. The result is a short rifting stage preceding continental breakup.
6.11.4.3 Thermomechanical Evolution and Tectonic Subsidence During Slow Extension The lithosphere reacts differently to low extension rates (Van Wijk and Cloetingh, 2002). Results of a representative test with an extension rate of 6 mm yr1 over a period of 100 My have been selected to illustrate the thermal evolution of the lithosphere (Figure 49). During the first 30 My after the onset of stretching, deformation localizes in the center of the model domain where the lithosphere was preweakened (Figure 44). Mantle material wells up and a sedimentary basin developed. Subsequently, as lithospheric stretching proceeds, temperatures begin to decrease (see panels 45 My and 50 My, Figure 49), in contrast to what happened in the standard necking case shown in Figure 45. Development of the upwelling zone ceases and the lithosphere cools in the center of the domain. Cooling of the central zone continues while after 70 My increasing temperatures are evident on both sides of the previously extended central zone. As these new upwelling zones develop further (Figure 49, 110 My panel) two new basins develop adjacent to the initial, first-stage basin. Temperatures in the lithosphere are now lower below the initial basin as compared to the surrounding lithosphere; a ‘cold spot’ is present in an area that underwent initial extension (see also Figure 50). Surface heat flow values reflect this thermal structure; the surface heat flow values are lower in the first-stage basin (66 mW m2) than in the surrounding areas (75 mW m2) at 110 My. Thinning factors for the crust and mantle lithosphere are shown in Figure 50. Thinning of the crust starts, as expected, in the central weakness zone of the domain where a single basin forms, with a maximum thinning factor of 1.85 for the crust. Crustal thinning in the central basin continues until about 65 My after the onset of stretching, at which time the locus of thinning shifts towards both sides of the initial basin. The zones of second-stage maximum thinning are located at a distance of about 500 km from the center of the initial rift. Although stretching of the lithosphere continues after 65 My, extensional strain is no longer localized in the initial rift basin but is centred on the two flanking new rifted basins. The thinning factor of the mantle lithosphere reflects this behaviour. During the first 45 My of stretching upwelling mantle material causes to be larger in the central zone of the domain than in its surroundings. Thereafter, however, temperatures decrease rapidly in the central zone (see also Figure 49) as
Tectonic Models for the Evolution of Sedimentary Basins
t = 30 My 1300
20 Depth (km)
t = 45 My T (°C) 0
0
1100
20
40
700
60
500
80
100
300
t = 50 My T (°C) 0
T (°C)
1300
1300
1100
20
1100
900
900
40
700
900
40
700
500
60
543
300
500
60
300
100
80
100
80
100 500
0
0
1000 (km)
t = 60 My
500
Depth (km)
1100 900 700
40
500 300 100
60
500
T (°C) 1300
1300 1100
20
900
1100
20
900
700
40
500
700 500
40
300
300
60
1000 (km)
t = 110 My T (°C) 0
1300
20
0
t = 70 My T (°C) 0
0
1000 (km)
100
100
60
80
80 0
500
1000 (km)
0
500
1000 (km)
0
500
1000 (km)
Figure 49 Thermal evolution of the lithosphere for a migrating rift at an extension rate of vext 6 mm yr1, times in My after the onset of stretching. Modified from Van Wijk and Cloetingh (2002).
new upwelling zones develop on its flanks with decreasing in the area of the first basin and increasing beneath the new basins. Figure 51 gives, for the modeling domain, the evolution of its surface topography as well as synthetic tectonic subsidence curves for its three rifted basins. After 35 My of stretching, the central basin reaches a maximum depth of about 760 m that would be amplified by sediment loading. Subsequently, the two lateral basins begin to subside whilst the central basin is uplifted. Ziegler (1987) defined basin inversion as the reversal of the subsidence patterns of a sedimentary basin in response to the buildup of compressional or transpressional stresses. According to this definition, no basin inversion takes place in our model as the lithosphere remains in a tensional setting. The synthetic tectonic subsidence curve for the central basin (right panel in Figure 51(b)) shows a clear reversal of its subsidence pattern under a stretching regime. The relative uplift is considerable that, depending on the location within this basin, amounts to approximately 800–1700 m. The left panel in Figure 51(b) shows the subsidence curve for one of the second-stage rifts that started to subside after about 50 My. The central panel shows the subsidence curve for the ‘transition zone’ between the
initial basin and the second-stage basin that formed after rift migration. This area displays continuous uplift. The total strength of the lithosphere, obtained by integrating the stress field over the thickness of the lithosphere (Ranalli, 1995), is shown in Figure 52. After the onset of stretching the central part of the model domain progressively weakens, reaching its minimum strength by 30 My. Thereafter its strength increases, but remains lower than the rest of the domain. From 55 My onward, the smallest values of lithospheric strength occur on both sides of the central basin beneath the second-stage rifts. By comparing the strength of the lithosphere with its thermal structure (Figure 49), the strong dependence of strength on the temperature is evident. From this modeling study it is concluded that lateral rift migration may occur under conditions of long-term low-strain-rate lithospheric extension. Whether such a model can be applied to the MidNorway margin needs to be further analyzed. This requires a step-wise palinspastic restoration of a set of structurally and stratigraphically closely controlled cross sections through the Mid-Norway margin and its conjugate East Greenland margin, permitting quantification of extensional strain rates through time.
544
Tectonic Models for the Evolution of Sedimentary Basins
Crustal thinning 100 1.8
Time (My)
80
1.7 1.6
60
1.5 1.4
40
1.3 1.2
20
1.1
0 0
500
1000 (km)
Mantle thinning 100 1.7
Time (My)
80
1.6 1.5
60
1.4 1.3
40
1.2 1.1
20 0 0
500
1000 (km)
Figure 50 Evolution of thinning factors for the crust () and lithospheric mantle () at an extension rate of vext 6 mm yr1. No continental breakup. The asymmetry (upper panel) is caused by the asymmetry of the finite element grid that was used. We used triangular-shaped elements that resulted in a not perfectly symmetric mesh and initial Moho topography. The wiggles in the lines (lower panel, upper right) are due to interpolation inexactnesses while calculating lithospheric mantle thinning. Modified from Van Wijk JW and Cloetingh S (2002) Basin migration caused by slow lithospheric extension. Earth and Planetary Science Letters 198: 275–288.
6.11.4.4 Breakup Processes: Timing, Mantle Plumes, and the Role of Melts The Mid-Norway margin is a representative part of the Norwegian–Greenland Sea rift along which crustal separation between NW Europe and Greenland was achieved in earliest Eocene times (Mosar et al., 2002; Torsvik et al., 2001). The Norwegian–Greenland Sea rift, which had remained intermittently active for some 280 Ma from the Late Carboniferous until the end of the Paleocene, is superimposed on the Arctic–North
Atlantic Caledonides (Ziegler, 1988; Ziegler and Cloetingh, 2004). During the Devonian and Early Carboniferous, orogen-parallel extension controlled the collapse of the Caledonian Orogen, the subsidence of pull-apart basins and uplift of core complexes (Braathen et al., 2002; Eide et al., 2002). During the subsequent rifting stage, tensional reactivation of Caledonian and Devonian–Early Carboniferous crustal discontinuities played an important role in the structuring of the Mid-Norway margin (Mosar, 2003). During the evolution of the Norwegian–Greenland Sea rift, initially a broad zone was affected by crustal extension. In time, rifting activity concentrated progressively on the zone of future crustal separation with lateral elements becoming step-wise abandoned (Ziegler, 1988, 1990b; Mosar et al., 2002; Ziegler and Cloetingh, 2004). On the Mid-Norway margin, syn-rift sediments attain thicknesses of up to 10 km with postrift series reaching thicknesses of 2–3 km (Osmundsen et al., 2002). Whereas Late Carboniferous to early Late Cretaceous crustal extension was not accompanied by volcanic activity, volcanism flared up during the Campanian– Maastrichtian, peaked during the Paleocene and terminated on the Mid-Norway margin upon earliest Eocene crustal separation when it centred on the newly developing system of seafloor spreading axes. The Paleocene development of the Thulean flood basalt province, which was centred on Iceland and had a radius of more than 1000 km, is attributed to the impingement of the Iceland plume on the Norwegian–Greenland Sea rift (Ziegler, 1988; Morton and Parson, 1988; Larsen et al., 1999; Skogseid et al., 2000). Tomographic data image a mantle plume beneath Iceland rising up from near the core–mantle boundary (Bijwaard and Spakman, 1999) (Figure 53). Correspondingly, the Paleocene extrusion and intrusion of large volumes of basaltic rocks on the Mid-Norway margin is generally attributed to a plume-related temperature increase of the asthenosphere. However, the interplay between extension and magmatism during continental breakup is still debated and recent numerical modeling studies suggest that the volumes of melts extruded at volcanic margins may also be generated by ‘standard’ thermal conditions, provided high extension rates can be implied (Figure 54) (Van Wijk et al., 2001). 6.11.4.5 Postrift Inversion, Borderland Uplift, and Denudation Following Early Eocene crustal separation, the MidNorway margin was partly inverted during the Late
Tectonic Models for the Evolution of Sedimentary Basins
(a)
545
Surface elevation (m) (c)
100
Time (My)
80 60 40 20 0
500
0
1000 (km)
Elevation (m)
1100 835 565 300 35 –230 –500 –765
1000 0 1000 (km)
–1000 100
500
m Ti
50
(b)
0
e y)
(M
1000
0
0 (m) –1000 20 60 100 (My) Figure 51 (a) Evolution of relative surface topography for a migrating rift, vext 6 mm yr1. Gray lines indicate the positions of the corresponding synthetic subsidence curves derived from this panel and shown in (b). (b) Synthetic subsidence curves for three locations indicated in (a): in the first-stage basin (right panel), in the second stage basin (left panel) and in the transition zone separating the two basins (middle panel). (c) Perspective view of the relative surface topography shown in (a). Modified from Van Wijk and Cloetingh 2002.
1012 N m–1
100
7.2 6.8
80 Time (My)
6.4 6.0
60
5.6 5.2
40
4.8 4.4
20
0
500
1000 (km)
Figure 52 Lithosphere strength evolution for a migrating rift, vext 6 mm yr1. Modified from Van Wijk JW and Cloetingh S (2002) Basin migration caused by slow lithospheric extension. Earth and Planetary Science Letters 198: 275–288.
Eocene–Early Oligocene and Miocene in the prolongation of the Iceland and Jan Mayen fracture zones (Mosar et al., 2002; Dore´ and Lundin, 1996). Mechanisms controlling the development of these inversion structures remain enigmatic as theoretical
models predict the inversion of passive margins in response to the buildup of compressional ridge-push forces only a few tens of million years after crustal separation. Most of the shortening on the MidNorway margin was accommodated along preexisting major fault zones (Pascal and Gabrielsen, 2001; Gabrielsen et al., 1999). Compressional structures, such as long-wavelength arches and domes, strongly modified the architecture of the deep Cretaceous basins and controlled sedimentation patterns during Cenozoic (Bukovics and Ziegler, 1985; Va˚gnes et al., 1998). From the Oligocene onward, the near-shore parts of the Mid-Norway margin were uplifted and deeply truncated (Dore´ and Jensen, 1996; Holtedahl, 1953). Mechanisms controlling the observed broad uplift of the inner shelf and the adjacent on-shore areas, as evident all around Norway also remain enigmatic. However, the long-wavelength of the uplifted area is suggestive for mantle processes (Rohrman and Van der Beek, 1996; Rohrman et al., 2002). Olesen et al., (2002) interpreted the long-wavelength component of the gravity field in terms of both Moho topography and large-scale intrabasement density contrasts.
546
Tectonic Models for the Evolution of Sedimentary Basins
5
10
15
20
25
30
35
250
0
0 km) th ( 0 50 Dep 100 0 0 0 15 200
0
–0.5%
a
+0.5%
Figure 53 Tomographic cross section showing a large plume-shaped anomaly in the mantle below Iceland. A 0.15% contour is indicated in black for clarity. Dashed lines indicate the 410 and 660 km discontinuities. Modified from Bijwaard H and Spakman W (1999) Fast Kinematic ray tracing of first- and later-arriving global seismic phases. Geophysical Journal International 139: 359–369.
Comparing the Bouguer gravity field (Figure 40) to the gravity responses from Airy roots at different depths for the northern Scandinavia mountains shows that the compensating masses are located at relatively shallow depths in the upper crust. Consequently, the gravity field of the northern Scandinavian mountains must originate from intracrustal low-density rocks in addition to Moho depth variations. On the other hand, the mountains of southern Norway appear to be supported by lowdensity rocks within the mantle (see Cloetingh et al., 2005). Southwestern Norway was uplifted by as much as to 2 km during Neogene times (Figure 42) (Rohrman et al., 1995), as evidenced by the progradation of clastic wedges into the North Sea Basin ( Jordt et al., 1995). However, the uplift patterns and timing of southwestern Norway and the northern Scandes differ (Hendriks and Andriessen, 2002). By Late Tertiary times, cold climatic conditions prevailed.
The seaward facing side of uplifted land areas was affected by strong glacial erosion, which in turn enhanced their uplift and increased sedimentation and subsidence rates in the flanking basins (Figure 41) (Cloetingh et al., 2005). FT analyses along major on-shore lineaments of southern Norway show that preexisting major normal faults, dating back to the Late Palaeozoic and Mesozoic, played a significant role in the Cenozoic uplift pattern (Hendriks and Andriessen, 2002; Cloetingh et al., 2005; Redfield et al., 2005). Under the presently prevailing northwest-directed compressional stress field the Mid-Norwegian margin and its adjacent highlands are seismically active (Gru¨nthal, 1999) with some faults showing evidence for recent movement (Mo¨rner, 2004). Uplift of the South Norwegian highland continues, whilst the North Sea Basin experiences a phase of accelerated subsidence that began during the Pliocene and that is attributed to stress-induced deflection of the lithosphere (Figure 10) (Van Wees and Cloetingh, 1996).
6.11.5 Black Sea Basin: Compressional Reactivation of an Extensional Basin The Black Sea (Figure 55), in which water depths range up to 2.2 km, is underlain by a larger western and a smaller eastern sub-basin that are separated by the Andrusov Ridge. The western basin is floored by oceanic and transitional crust and contains up to 19 km of Cretaceous to recent sediments. The eastern basin is floored by strongly thinned continental crust and contains up to 12 km of Cretaceous and younger sediments. The Andrusov Ridge is buried beneath 5–6 km thick sediments and is upheld by attenuated continental crust. Significantly, the sedimentary fill of the Black Sea basin system is characterized by nearly horizontal layering that is only disturbed along its flanks bordering the orogenic systems of the Balkanides and Pontides in the south, and the Great Caucasus and Crimea in the north and northeast (Figure 56). The Black Sea basin system is thought to have evolved by Aptian–Albian back-arc rifting that progressed in the western sub-basin to crustal separation and Cenomanian–Coniacian seafloor spreading. During the Late Senonian and Paleocene the Black Sea was subjected to regional compression in conjunction with the evolution of
Tectonic Models for the Evolution of Sedimentary Basins
MPa 400
Depth (km)
Depth (km)
Depth (km)
Horizontal deviatoric stress field 0
200 100 0
2 My
0
0
1000 (km) 350 200
100
8 My 0
0
1000 (km)
Depth (km) Depth (km)
0 300 150
80
14 My
0 0
Depth (km)
547
1000 (km) °C 1200
Temperature field
0
800 100 0
400
2 My
0
0
1000 (km) 1200 800
100 0
400
8 My 0
1000 (km)
0 1200 800
80
400
14 My 0
1000 (km)
0
Figure 54 Dynamic numerical model for extension of lithosphere and passive continental margin formation, simulating the evolution of the northern Atlantic volcanic rifted margin. Upper three panels show deviatoric stress field at three subsequent time steps. Lower three panels show temperature field at these time steps. Note the development of melts, usually attributed to mantle plumes, as a direct consequence of lithospheric extension and breakup. Modified from Van Wijk JW, Huismans RS, Ter Voorde M, and Cloetingh S (2001). Melt generation at volcanic continental margins: No need for a mantle plume? Geophysical Research Letters 28: 3995–3998.
its flanking orogenic belts. During the Early Eocene major rifting and volcanism affected the eastern Black Sea Basin and the eastward adjacent Acharat–Trialeti Basin. During the Late Eocene and Oligocene, the Pontides thrust belt developed along the southern margin of the Black Sea and inversion of the Great Caucasus Trough and the Acharat–Trialeti rift commenced. The present stress regime of the Black Sea area as deduced from earthquake focal mechanisms, structural, and GPS data is compression dominated, reflecting continued collisional interaction of the Arabian
and the Eurasian plates that controls ongoing crustal shortening in the Great Caucasus. In the absence of intrabasinal deformations, the Pliocene and Quaternary accelerated subsidence of the Black Sea Basin is attributed to stress-induced downward deflection of its lithosphere (Nikishin et al., 2001, 2003; Cloetingh et al., 2003). Although there is general agreement that the Black Sea evolved in response to Late Cretaceous and Eocene back-arc extension, the exact timing and kinematics of opening of its western and eastern subbasins is still being debated (e.g., Robinson et al., 1995;
548
Tectonic Models for the Evolution of Sedimentary Basins
25E 48N
Russian platform B′
Karkinit trough
Do
br
og
Indo
lo-Ku
ea
A′
Tua p
se
An dr us ov
Western Black Sea Line 202
A
ate
C'
auc
asu
h
Sh
rid
ats
0
ge
B
tides
rn pon Weste
Basin
rC
Trou g
04
Bulgaria
ban
Gre
ky
e8
an thi r pa nd a C ela for
N.
Russia
ov High Mid-Az
Lin
Romania
Ukraine
Rid
s
ge
Eastern Black Sea
C Eastern pontides
Turkey
40N 43E
Areas of major tertiary compressional deformation Areas flored by oceanic or highly extended continental crust
Limit of major compression Limit of major extension Major strike-slip faults
0
300 km
Other boundaries
Figure 55 Regional tectonic map of the Black Sea. Modified from Cloetingh S, Spadini G, van Wees JD, and Beekamn F (2003) Thermo-mechanical modelling of Black Sea Basin (de)formation. Sedimentary Geology 156: 169–184.
Nikishin et al., 2001, 2003; Cloetingh et al., 2003). This applies particularly to the exact opening timing of the eastern Black Sea for which different interpretations have been advanced varying from Middle to Late Cretaceous (Finetti et al., 1988) to Early Eocene (Robinson et al., 1995) or a combination thereof (Nikishin et al., 2003). Gravity data show an important difference in the mode of flexural compensation between the western and eastern Black Sea (Spadini et al., 1997). The western Black Sea appears to be in a state of isostatic undercompensation and upward flexure, consistent with a deep level of lithospheric necking. By contrast, for the eastern Black Sea gravity data point toward isostatic overcompensation and a downward state of flexure, compatible with a shallow necking level (Figure 57). This is thought to reflect differences in the prerift mechanical properties of the lithosphere of the western and eastern Black Sea sub-basins (Spadini et al., 1996; Cloetingh et al., 1995b). Below we address the relationship between the prerift finite strength of the lithosphere and geometry of extensional basins and discuss the effects of differences in prerift rheology on the Mesozoic–Cenozoic
stratigraphy of the Black Sea basin system. These findings raise important questions on postrift tectonics and on intraplate stress transmission into the Black Sea Basin from its margins. We discuss the results of thermomechanical modeling of the Black Sea Basin along a number of regional cross sections through its western and eastern parts (Figure 55), that are constrained by a large integrated geological and geophysical database (see Spadini, 1996; Spadini et al., 1996, 1997). 6.11.5.1 Rheology and Sedimentary Basin Formation Inferred differences in the mode of basin formation between the western and eastern Black Sea (Figure 58), basins can be largely explained in terms of palaeorheologies. The prerift lithospheric strength of the western Black Sea (Figure 58) appears to be primarily controlled by the combined mechanical response of a strong upper crust and strong upper mantle (Spadini et al., 1996). The shallow necking level in the eastern Black Sea is compatible with a prerift strength controlled by a strong upper crust decoupled from the weak hot underlying mantle
Tectonic Models for the Evolution of Sedimentary Basins
Depth (km)
0
A SSE
549
A′ NNW
Western Black Sea
2
5
10 35
10
Prerift basement
65 97
15 50
0
100
C SSW Sinop Trough
150
200 250 Distance (km)
Mid Black Sea High
350
Shatsky Ridge
Eastern Black Sea
? Depth (km)
300
400
C′ NNE
2
5
10 35 39
10
15
Prerift basement
50
100
150 Distance (km)
200
250
300
Figure 56 Observed first-order geometry of basement configuration and sediment fill of the Western and Eastern Black Sea. Numbers refer to stratigraphic ages (Ma). Note the substantial thickness of Quarternary sediments. For location of sections A-A9 and C-C9 see Figure 55. Modified from Cloetingh S, Spadini G, van Wees JD, and Beekamn F (2003) Thermomechanical modelling of Black Sea Basin (de)formation. Sedimentary Geology 156: 169–184.
(Figure 58). These differences point to important differences in the thermotectonic age of the lithosphere of the two sub-basins (Cloetingh and Burov, 1996). The inferred lateral variations between the western and eastern Black Sea suggest thermal stabilization of the western Black Sea prior to rifting whilst the lithosphere of the eastern Black Sea was apparently already thinned and thermally destabilized by the time of Eocene rifting. The inferred differences in necking level and in the timing of rifting between the western and eastern Black Sea suggest an earlier and more pronounced development of rift shoulders in the western Black Sea Basin as compared to the eastern Black Sea (Robinson et al., 1995). Figures 56 and 59 show observed and modeled stratigraphies along the two selected profiles through the western and eastern Black Sea, respectively (see also Figure 58). Figure 60 illustrates the evolution of basin subsidence and water loaded tectonic subsidence in time (Steckler and Watts, 1978; Bond and Kominz, 1984), calculated for both the Odin (1994) and Harland et al., (1990) time scales. Subsidence
curves are displayed for locations at the center and the margin of western and eastern Black Sea subbasins, respectively. In the western Black Sea rifting began during the Late Barremian–Aptian and progressed to crustal separation at the transition to the Cenomanian with seafloor spreading ending in the Coniacian. In this deep marine basin up to 12.5 km thick sediments accumulated prior to the late Middle Miocene Sarmatian sea level fall when it was converted into a relatively small up to 800 m deep lake. Late Miocene to recent sediments attains thicknesses of up to 2.5 km (Figure 59). The eastern Black Sea basin may have undergone an Aptian–Turonian and a Campanian–Maastrichtian rifting stage prior its Paleocene–Early Eocene rift-related main subsidence and deepening that was accompanied by little flank uplift and erosion. During the Late Eocene, sediment supply from the compressionally active Pontides and Greater Caucasus belts increased and led in the basin center to the deposition of an up to 5 km thick sediments (Figure 59). Also the eastern Black Sea remained a deep marine basin until the late
Tectonic Models for the Evolution of Sedimentary Basins
C′
C Overcompensation
Undercompensation
Observed Isostatic Best-fit model
Depth (km) 30 20 10
0
FAA (mGal) –50 0 50
Residual (mGal) –50 0 50
550
Sediments Isostatic Moho Refraction Moho Best-fit Moho
50
100
150 Distance (km)
200
250
Figure 57 Results of gravity modeling for the Eastern Black Sea, demonstrating isostatic flexural overcompensation in the center of the basin. For location see Figure 55 line C-C9. Modified from Cloetingh S, Spadini G, van Wees JD, and Beekamn F (2003) Thermo-mechanical modelling of Black Sea Basin (de)formation. Sedimentary Geology 156: 169–184.
Middle Miocene Sarmatian when it was converted into a lake. When sea level returned to normal in the Late Miocene, water depth increased dramatically to 2800 m in both the western and eastern Black Sea Basin, presumably in response to the loading effect of the water (Spadini et al., 1996). By the Quaternary, increased sediment supply led to significant subsidence and sediment accumulation, with a modest decrease in water depth to the present-day value of 2200 m. Overall uplift of the margins of the Black Sea commenced in Middle Miocene times (Nikishin et al., 2003). Although the reconstructions by Spadini et al., (1997) and Nikishin et al., (2003) differ in the assumed maximum depth of the basin, its palaeo-bathymetry and sea-level fluctuations during its evolution, the Pliocene–Quaternary subsidence acceleration appears to be a robust (Spadini et al., 1997; Robinson et al., 1995; Nikishin et al., (2003). 6.11.5.2
Role of Intraplate Stresses
Constraints on the present-day stress regime are lacking for the central parts of the Black Sea Basin.
However, structural geological field studies and earthquake focal mechanisms in areas bordering the Black Sea (see Nikishin et al., 2001), as well as GPS data (Reilinger et al., 1997) demonstrate that in the collisional setting of the European and Arabian plate the area is subjected to compression. Field studies of kinematic indicators and numerical modeling of present-day and palaeo-stress fields in selected areas (e.g., Go¨lke and Coblentz, 1996; Bada et al., 1998, 2001) have yielded new constraints on the causes and the expression of intraplate stress fields in the lithosphere. Ziegler et al., (1998) have discussed the key role of mechanical controls on collision related compressional intraplate deformation. These authors discuss the buildup of intraplate stresses in relation to mechanical coupling between an orogenic wedge and its fore- and hinterlands, as well as the implications to the understanding of a number of first-order features in crustal and lithospheric deformation. Temporal and spatial variations in the level and magnitude of these stresses have a strong impact on the record of vertical motions in sedimentary basins (Cloetingh et al., 1985, 1990; Cloetingh and Kooi,
Tectonic Models for the Evolution of Sedimentary Basins
SSE
NNW Western Black Sea
‘COLD’ prerift lithosphere
A′
0
A
551
Depth (km) 20 10
Sediments (observed)
Models Neck 35 Neck 25 Neck 15
30
Crust (observed)
50
100
150
200
250
300
350
400
SSE
NNW ‘COLD’ prerift lithosphere
Western Black Sea
B′
Depth (km) 20 10
0
B
30
Neck 35 Neck 25 Neck 15
50
150
‘WARM’ prerift lithosphere
200
250
300
350
400
450
500
NNE C′
Eastern Black Sea
Neck 35 Neck 25 Neck 15
30
Depth (km) 20 10
0
SSW C
100
50
100
150 Distance (km)
200
250
Figure 58 Crustal scale models for extensional basin formation for the Western and Eastern Black Sea. For location of cross sections see Figure 55. A comparison of predicted and observed Moho depths provides constraints on levels of necking and thermal regime of the prerift lithosphere. The models support the presence of cold prerift lithosphere compatible with a deep necking level of 25 km in the Western Black Sea. For the Eastern Black Sea, the models suggest the presence of a warm prerift lithosphere with a necking level of 15 km. Modified from Cloetingh S, Spadini G, van Wees JD, and Beekamn F (2003) Thermo-mechanical modelling of Black Sea Basin (de)formation. Sedimentary Geology 156: 169–184.
1992; Zoback et al., 1993; Van Balen et al., 1998). Stresses at a level close to lithospheric strength propagating from the margins of the Black Sea Basin into its interior parts had not only a strong effect on their stratigraphic record, but presumably induced by
lithospheric folding the observed late-stage subsidence acceleration (Cloetingh et al., 1999), similar to what is recognized in the Pannonian Basin and the North Sea Basin (Horva´th and Cloetingh, 1996; Van Wees and Cloetingh, 1996). Over the last few years,
552
Tectonic Models for the Evolution of Sedimentary Basins
SSE
Akcakoca-Delfin Profile – Western Black Sea (A-A′)
Depth (km) 10 5
0 Ma 2 10 35 65
15
97 50
100
SSW
150
200 250 Distance (km)
300
350
Badut Profile – Eastern Black Sea (C-C′)
400
NNE
0 2 10 35 39
Depth (km) 10 5 15
NNW
50
100
150 Distance (km)
200
250
Figure 59 Stratigraphy and basement topography of the Western and Eastern Black Sea modeled along profiles A-A9 and C-C9 (for location see Figure 55). Adopted necking levels, constrained by gravity modeling, are 25 km and 15 km for the Western and Eastern Black Sea, respectively. Modified from Cloetingh S, Spadini G, van Wees JD, and Beekamn F (2003). Thermo-mechanical modelling of Black Sea Basin (de)formation. Sedimentary Geology 156: 169–184.
increasing attention has been directed to this topic, advancing our understanding of the relationship between changes in plate motions, plate interaction, and the evolution of rifted basins ( Janssen et al., 1995; Dore´ et al., 1997) and foreland areas (Ziegler et al., 1995, 1998, 2001). A continuous spectrum of stress-induced vertical motions can be expected in the sedimentary record, varying from subtle faulting effects (e.g., Figure 16) (Ter Voorde and Cloetingh, 1996; Ter Voorde et al., 1997) and basin inversion (Brun and Nalpas, 1996; Ziegler et al., 1998) to the enhancement of flexural effects and to lithospheric folding induced by high levels of stress approaching lithospheric strengths (Stephenson and Cloetingh, 1991; Nikishin et al., 1993; Burov et al., 1993; Cloetingh and Burov, 1996; Bonnet et al., 1998; Cloetingh et al., 1999). Crustal and lithospheric folding can be an important mode of basin formation on plates involved in continental collision (Cobbold et al., 1993; Ziegler et al., 1995, 1998; Cloetingh et al., 1999). Numerical models have been developed for simulating the interplay of faulting and folding during intraplate compressional deformation (Beekman et al., 1996; Gerbault et al., 1998; Cloetingh et al., 1999). Models
have also been developed to investigate the effects of faulting on stress-induced intraplate deformation in rifted margin settings (Van Balen et al., 1998). The collisional Caucasus orogeny commenced during the Late Eocene and culminated during Oligocene–Quaternary times. On the other hand, the North Pontides thrust belt was activated during the Late Eocene and remained active until the end of the Oligocene (Nikishin et al., 2001). Correspondingly, the Late Eocene accelerated subsidence of the Black Sea Basin can be attributed to the buildup of a regional compressional stress field (Robinson et al., 1995). The Late Eocene–Quaternary Caucasus orogeny, overprinting back-arc extension in the Black Sea, was controlled by the collisional interaction of the Arabian plate with the southern margin of the East-European craton (Nikishin et al., 2001). 6.11.5.3 Strength Evolution and Neotectonic Reactivation at the Basin Margins during the Postrift Phase Automated backstripping analyses and comparison of results with forward models of lithospheric stretching
Tectonic Models for the Evolution of Sedimentary Basins
(Van Wees et al., 1998) provide estimates of the integrated lithospheric strength at various syn- and postrift stages. Adopted modeling parameters are listed in Tables 5 and 6. Figure 61 shows a comparison of observed and forward modeled tectonic subsidence curves for the center of the Western Black Sea Basin. Automated backstripping yields a stretching factor of 6. Modeling fails, however, to predict the pronounced Late Neogene subsidence acceleration, documented by the stratigraphic record that may be attributed to late stage compression. As postrift cooling of the lithosphere leads in time to a significant increase in its integrated strength, its early postrift deformation is favoured. Present-day lithospheric strength profiles calculated for the center and margin of western Black Sea show a pronounced difference. The presence of relatively strong lithosphere in the basin center and weaker lithosphere at the basin margins favours deformation of the latter during late-stage compression. This may explain why observed compressional structures appear predominantly at the edges of the Black Sea Basin and not in its interior (Figure 55). In Figure 62 the observed and forward modeled tectonic subsidence curves for the center of eastern Black Sea Basin are compared, adopting for modeling a stretching factor of 2.3 that is compatible with the subsidence data and consistent with geophysical constraints. During the first 10 My of postrift evolution, integrated strengths are low but subsequently increase rapidly owing to cooling of the lithosphere. During the first 10 My after rifting, and in the presence of very weak lithosphere, the strength of which was primarily controlled by the rift-inherited mechanical properties of its upper parts, we expect that this area would be prone to early postrift compressional deformation. It should be noted, however, that with a cooling-related progressive increase of the integrated lithospheric strength, with time increasingly higher stress levels are required to cause large-scale deformation (Figure 62). Important in this context is also that, due to substantial crustal thinning, a strong upper mantle layer is present in the central part of the basin at relatively shallow depths. Based on the present thermomechanical configuration (Figure 62) with relatively strong lithosphere in the basin center and relatively weak lithosphere at the basin margins, we predict that a substantial amount of late-stage shortening induced by orogenic activity in the surrounding areas will be taken up along the basin margins, with only minor
553
deformation occurring in the relatively stiff central parts of the basin. The relative difference in rheological strength of the marginal and central parts of the basin is more pronounced in the eastern than in the western Black Sea. These predictions have to be validated by new data focusing on the neotectonics of the Black Sea. High-resolution shallow seismic profiles and acquisition of stress-indicator data could provide the necessary constraints for such future modeling. Figure 63 gives predictions for basement and surface heat flow in the eastern and western Black Sea and shows markedly different patterns in timing of the rift-related heat flow maximum. The predicted present-day heat flow is considerably lower for the western than for the eastern sub-basin. This is attributed to the presence of more heat producing crustal material in the eastern than in the western Black Sea that is partly floored by oceanic crust. In heat flow modeling studies the effects of sedimentary blanketing were taken into account (Van Wees and Beekman, 2000). Heat flow values vary between 30 mW m2 in the center of the basins up to 70 mW m2 at the Crimean and Caucasus margins (Nikishin et al., 2003). Note the pronounced effect of thermal blanketing in the western Black Sea that contains up to 15 km thick sediments (Figure 58). As a result its present-day integrated strength is not that much higher than its initial strength. By contrast, the integrated strength of the eastern Black Sea is much higher than its initial strength as the blanketing effect of its up to 12 km thick sedimentary fill is less pronounced and as water depths are greater. Figure 32 shows a comparison of theoretical predictions for lithosphere folding of rheologically coupled and decoupled lithosphere, as a function of its thermomechanical age with estimates of folding wavelengths documented in continental lithosphere for various representative areas of the globe (see Cloetingh et al., 1999). The western Black Sea center is marked by a thermomechanical age of around 100 My with rheological modeling indicating mechanical decoupling of the crust and lithospheric mantle (see Figure 61). These models imply an EET of at least 40 km (Burov and Diament, 1995) and folding wavelengths of 100–200 km for the mantle and 50–100 km for the upper crust (Cloetingh et al., 1999). For the eastern Black Sea, a probably significantly younger thermomechanical lithospheric age of 55 My implies an EET of no more than 25 km, and indicates mantle folding at wavelengths of 100–150 km and a crustal folding wavelength
Tectonic Models for the Evolution of Sedimentary Basins Western Black Sea Basin margin
Basin center Age (Ma)
Age (Ma) 150
125
100
75
50
25
150
0
125
100
75
50
25
0 0
0
1000
Depth (m)
2000
1000
3000 4000
1500
PWD (Harland) PWD (Odin)
Depth (m)
500
PWD (Harland) 5000 PWD (Odin)
2000
6000
100
Age (Ma) 75 50
Age (Ma) 25
150 0
0 0
500
2000
1000
4000
0
1500 2000 BS (Harland) BS (Odin) WL TS (Harland) WL TS (Odin)
125
100
75
50
25
Depth (m)
125
Depth (m)
150
6000 8000
2500
10 000
BS (Harland) BS (Odin) WL TS (Harland) WL TS (Odin)
3000
12 000 14 000
3500
Eastern Black Sea
40
30
20
10
0
70
60
50
40
30
20
10
0
0
0
500
1000
1000 1500
2000 3000
2000 PWD (Harland) PWD (Odin)
70
60
50
Age (Ma) 40 30
4000 PWD (Harland) PWD (Odin)
2500
20
10
0
70
60
50
Age (Ma) 40 30
20
10
0
1000
2000 BS (Harland) BS (Odin) WL TS (Harland) WL TS (Odin)
2500 3000
0 0
BS (Harland) BS (Odin) WL TS (Harland) WL TS (Odin)
500
1500
5000
2000 4000 6000 8000
3500 4000
10000
Depth (m)
50
Age (Ma)
Depth (m)
60
Basin center
Age (Ma)
Depth (m)
70
Basin margin
Depth (m)
554
Tectonic Models for the Evolution of Sedimentary Basins Table 5 Model parameters used to calculate the tectonic subsidence in the rheological models Symbol
Model parameter
Value
A
Initial lithosphere thickness Initial crustal thickness Asthenospheric temperature Thermal diffusivity Surface crustal density Surface mantle density Water density Thermal expansion coeff. Crustal stretching factor Subcrustal stretching factor
120 km (wb), 80 km (eb) 35 km 1333 C
C Tm K c m w
1 106 m2 s1 2800 kg m3 3400 kg m3 1030 kg m3 3.2 105 K1 6 (wb), 2.3 (eb) 6 (wb), 2.3 (eb)
The (eb) and (wb) refer to eastern and western Black Sea, respectively.
Table 6 Default rheological and thermal properties of crust and lithosphere Heat production (106 W m3)
Layer
Rheology
Conductivity (W m1 K1)
Sediments Upper crust Lower crust Upper mantle
Quartzite (d) Quartzite (d)
1.5 2.9
0.5 2
Diorite (w)
2.9
0.5
Olivine (d)
2.9
0
The (d) and (w) refer respectively to dry or wet rock samples that contain little or variable amounts of structural water. For more details on the rheological rock properties, see Van Wees and Beekman (2000).
similar to the western Black Sea. A comparison of estimated folding wavelengths with theoretical predictions shows a systematic deviation to larger values. This is characteristic for ‘a-typical’ folding where the large dimension of the preexisting rift basin causes during the late-stage compressional phase a
555
pronounced increase in the wavelength of the stressinduced down warp (Cloetingh et al., 1999). This effect has been observed in the North Sea Basin (Van Wees and Cloetingh, 1996) and the Pannonian Basin (Horva´th and Cloetingh, 1996), both of which are characterized by large sediment loads and a wide rift basin. Such a neotectonic compressional reactivation provides an alternative to previous explanations for recent differential motions in the northern Black Sea Basin (Smolyaninova et al., 1996) that were attributed to convective mantle flow. In view of the recent evidence for crustal shortening in the Black Sea region as a consequence of the Arabian– Eurasian plate interaction (Reilinger et al., 1997), an interpretation in terms of an increased Late Neogene compressional stress level appears to be more likely. According to modeling results, the eastern Black Sea Basin is much weaker than the western Black Sea. As compared to the western Black Sea, the eastern Black Sea is relatively stronger in the center than at its margins. The margins of the eastern Black Sea appears to be more prone to lithospheric folding, whereas the western Black Sea as a whole is more prone to stress transmission.
6.11.6 Modes of Basin (De)formation, Lithospheric Strength, and Vertical Motions in the Pannonian–Carpathian Basin System The Pannonian–Carpathian system in Central and Eastern Europe (Figure 64) has been the focus of considerable research efforts to integrate a wide range of geophysical and geological data, making it a key area for quantitative basin studies (see Cloetingh et al., 2006; Horva´th et al., 2006 for recent reviews). A vast geophysical and geological database has been established during the last decades as a result of a major international research collaboration in this area, largely carried out in the framework of European programs such as the EU Integrated Basin
Figure 60 Results of backstripping analyses for locations at the margins and center of the Western and Eastern Black Sea, respectively. Top panels show palaeo water depth (PWD), bottom panels show basement subsidence (BS) and water loaded tectonic subsidence (WLTS). Each curve is calculated for two different timescales (Odin, 1994; Harland et al., 1990) to illustrate sensitivities. Modified from Cloetingh S, Spadini G, van Wees JD, and Beekamn F (2003) Thermo-mechanical modelling of Black Sea Basin (de)formation. Sedimentary Geology 156: 169–184.
Tectonic Models for the Evolution of Sedimentary Basins
Western Black Sea basin analysis
100
80
60
40
20
0
0 1000
Forward
2000
Observed
3000 4000 5000 6000
35 30
Compressional strength
25 20 15 10 5
Integrated strength (TN m–1)
40
Extensional strength
0 100
50
0
Centre strength (MPa T –1)
Margin strength (MPa T –1) 0 500 1000 1500
0
500
1000
1500
2000
0
0 Temperature
10 000
10 000
Compressional strength
Moho 20 000
30 000 40 000 50 000
Moho
Depth (m)
20 000 Depth (m)
556
30 000 40 000 50 000 Temperature
60 000
60 000
70 000
70 000
Compressional strength
Tectonic Models for the Evolution of Sedimentary Basins
Studies project (Cloetingh et al., 1995b; Durand et al., 1999), the ILP-ALCAPA (Cloetingh et al., 1993; Neubauer et al., 1997) and EUROPROBEPANCARDI (Decker et al., 1998) programs, and the Peri-Tethys program (Ziegler and Horva´th, 1996; Brunet and Cloetingh, 2003), partly funded in the context of petroleum exploration. These studies, building on previous land-marking compilations (Royden and Horva´th, 1988), marked a major advance in applying basin analysis concepts to the Pannonian–Carpathians system. An important asset of this natural laboratory is the existence of high-quality constraints on basin evolution obtained through the systematic acquisition of seismic, gravity, heat flow, and magnetotelluric data by various research groups (see Royden and Horva´th, 1988; Posgay et al., 1995; Szafia´n et al., 1997; Tari et al., 1999; Wenzel et al., 1999; Hauser et al., 2001). Extensive industrial reflection-seismic coverage and well data acquired in the context of petroleum exploration and surface studies permitted to construct a high-resolution stratigraphic framework for this area (e.g., Vakarcs et al., 1994; Sacchi et al., 1999; Vasiliev et al., 2004). At the same time, the Carpathian fold and thrust belt has been the focus of concerted efforts, highlighting the connection between lateral variations in structural style, basement characteristics and foreland flexure development in different segments of the Carpathian orogen (Sa˘ndulescu, 1988; Roure et al., 1993; Mat¸enco et al., 1997a, 1997b, 2003; Sanders et al., 1999; Tari et al., 1997; Zoetemeijer et al., 1999) and its hinterland, the Transylvanian Basin (e.g., Ciulavu et al., 2002). The Pannonian–Carpathian system, therefore, permits to test models for the development of sedimentary basins and their subsequent deformation, and for ongoing continental collision. The lithosphere of the Pannonian Basin is a particularly sensitive recorder of changes in lithospheric stress induced by near-field intraplate and far-field plate
557
boundary processes (Bada et al., 2001). High-quality constraints are available on regional (palaeo)stress fields (Fodor et al., 1999; Gerner et al., 1999) owing to earthquake focal mechanism studies, analyses of borehole breakouts and studies on kinematic field indicators. A close relationship has been established between the timing and nature of stress changes in extensional basins and structural episodes in the surrounding thrust belts, pointing to mechanical coupling between the orogen and its back-arc basin. In parallel, significant efforts have been devoted to reconstruct the spatial and temporal variations in thrusting along the Carpathian orogen (Roure et al., 1993; Schmid et al., 1998; Mat¸enco and Bertotti, 2000) and its relationship to foredeep depocenters (Meulenkamp et al., 1996; Mat¸enco et al., 2003; Ta˘ra˘poanca˘ et al., 2003), changes in foreland basin geometry and lateral variations in the flexural rigidity of the foreland lithosphere. A general feature of flexural modeling studies carried out on the Carpathian system (e.g., Zoetemeijer et al., 1999; Mat¸enco et al., 1997b) is the inferred low rigidity of the foreland platform lithosphere that dips beneath the SE Carpathians. Studies constrained by gravity data also point to an important role of the flexural response of the foreland lithosphere to erosional unroofing of the Carpathian mountain chain (Szafia´n et al., 1997). Below we present an overview of the tectonic evolution of the Pannonian–Carpathian system in terms of various thermomechanical models. We present a lithospheric strength map of the system that highlights rheological constraints for basin analysis studies and for the reconstruction of the deformation history of this area. We then focus on the development of the system as inferred from subsidence analyses and stretching models of the Pannonian Basin, and the flexural behaviour of the Carpathian foreland lithosphere. The neotectonic reactivation of the region is described in terms of anomalous late-
Figure 61 A comparison of observed and forward modeled tectonic subsidence for the Western Black Sea center. Automated backstripping yields an estimated stretching factor of 6 (top panel). A pronounced Late Neogene subsidence acceleration (see also Figure 60) documented in the stratigraphic record could be an indication of late stage compression. Postrift cooling leads in time to a significant increase in the predicted integrated strength for both compressional and extensional regimes (middle panel; 1 TN m1 ¼ 1012 N m1). Present-day lithospheric compressional strength profiles calculated for the center and margin of western Black Sea show with depth a pronounced difference (bottom panels). Temperature profiles (in C) and Moho depth are given for reference. Note the important role of the actual Moho position in terms of mechanical decoupling of the upper crust and mantle parts of the Black Sea lithosphere. Parameters used for modeling of the tectonic subsidence and for rheological strength calculations are listed in Tables 5 and 6, respectively. Modified from Cloetingh S, Spadini G, van Wees JD, and Beekamn F (2003) Thermo-mechanical modelling of Black Sea Basin (de)formation. Sedimentary Geology 156: 169–184.
558
Tectonic Models for the Evolution of Sedimentary Basins
100
Eastern Black Sea centre basin analysis 80 60 40
20
0
0 Forward
1000
Observed 2000 3000 4000 5000
12
Compressional strength
10 8 6 4 2
Integrated strength (TN m–1)
14
Extensional strength
0 100
80
60
40
20
0
Center strength (MPa T –1)
Margin strength (MPa T –1) 0 500 1000 1500
0
0
500
1000
1500
0 Temperature
10 000
10 000
Compressional strength
30 000 40 000
20 000
Moho
Depth (m)
Depth (m)
20 000
Moho
30 000 40 000
50 000
50 000
60 000
60 000
70 000
70 000
Temperature Compressional strength
Figure 62 Comparison of observed and forward modeled tectonic subsidence for the Eastern Black Sea center. Automated backstripping yields an estimated stretching factor of 2.3 (upper panel). Postrift cooling leads in time to a significant increase in the predicted integrated strength for both compressional and extensional regimes (middle panel; 1 TN m1 ¼ 1012 N m1). Present-day lithospheric strength profiles calculated for the center and margin of the Eastern Black Sea show with depth a pronounced difference (bottom panels). Modified from Cloetingh S, Spadini G, van Wees JD, and Beekamn F (2003) Thermo-mechanical modelling of Black Sea Basin (de)formation. Sedimentary Geology 156: 169–184.
Tectonic Models for the Evolution of Sedimentary Basins
175
125 100 75 50
WBS (surface) EBS (surface) WBS (Bsmt) EBS (Bsmt)
Heat flow (mW m–2)
150
25 0
140
120
100
80 60 Age (Ma)
40
20
0
Figure 63 Predictions for basement (‘Bsmt’) and surface heat flow in the Eastern (triangles) and Western (circles) Black Sea show markedly different patterns in the timing of the heat flow maximum, related to the timing of initial rifting. The predicted present-day heat flow is considerably lower in the Western Black as compared to the Eastern Black Sea. See text for implications for differences in the strength evolution between the Western and Eastern Black Sea. Modified from Cloetingh S, Spadini G, van Wees JD, and Beekamn F (2003). Thermo-mechanical modelling of Black Sea Basin (de)formation. Sedimentary Geology 156: 169–184.
stage vertical movements, that is, accelerated subsidence in the center of the Pannonian Basin and fast uplift of the Carpathians orogen due to isostatic rebound in the aftermath of continental convergence and slab detachment. We conclude with a discussion on the thermomechanical aspects of basin inversion, lithospheric folding and related temporal and spatial variations of continental topography in the Pannonian–Carpathian system.
6.11.6.1 Lithospheric Strength in the Pannonian–Carpathian System The Pannonian–Carpathian system shows remarkable variations in crustal thickness (Figure 65) and thermomechanical properties of the lithosphere. Lithospheric rigidity varies in space and time, giving rise to important differences in the tectonic behavior of different parts of the system. As rheology controls the response of the lithosphere to stresses, and thus the formation and deformation of basins and orogens, the characterization of rheological properties and their temporal changes has been a major challenge to constrain and quantify tectonic models and scenarios. This is particularly valid for the Pannonian– Carpathian region where tectonic units with a different history and rheological properties are in close contact.
559
Figure 66(a) displays three strength envelopes for the western, central and eastern part of the Pannonian lithosphere that were constructed on the basis of extrapolated rock mechanic data, incorporating constraints on crustal and lithospheric structure, and present-day heat flow along the modeled rheological section. These strength profiles show that the average strength of the Pannonian lithosphere is very low (see also Lankreijer, 1998), which is mainly due to high heat flow related to upwelling of the asthenosphere beneath the basin system. The Pannonian Basin, the hottest basin of continental Europe, has an extremely low rigidity lithosphere that renders it prone to repeated tectonic reactivation. This is the result of Cretaceous and Paleogene orogenic phases involving nappe emplacement and crustal accretion, thickening and loading. In this process, the strength of the different Pannonian lithospheric segments gradually decreased, allowing for their tensional collapse under high-level strain localization that lead to the development of the Pannonian Basin. Another essential feature is the present-day lack of lithospheric strength in the lithospheric mantle of the Pannonian Basin. Strength appears to be concentrated in the crustal upper 7–12 km of the lithosphere. This finding is in very good agreement with the depth distribution of seismicity. Earthquake hypocenters are restricted to the uppermost crustal levels, suggesting that brittle deformation of the lithosphere is limited to depth of 5–15 km (To´th et al., 2002). Figure 66(b) shows estimates of the total integrated strength (TIS) of the Pannonian–Carpathian lithosphere along section A-A’. Rheology calculations suggest major differences in the mechanical properties of different tectonic units within the system (Lankreijer et al., 1997, 1999). In general, there is a gradual increase of TIS away from the basin center towards the basin flanks in the peripheral areas (see also Figure 66(c)). The center of the Pannonian Basin and the Carpathian foreland are the weakest and strongest parts of the system, respectively. The presence of a relatively strong lithosphere in the Transylvanian Basin is due to the absence of largescale Tertiary extension that prevails in the Pannonian Basin (Ciulavu et al., 2002). The Carpathian arc, particularly its western parts, shows a high level of rigidity apart from the southeastern bend zone where a striking decrease in lithospheric strength is noticed. Calculations for the seismically active Vrancea area indicate the presence of a very weak crust and mantle lithosphere, suggestive of mechanical decoupling between the Transylvanian Basin and the Carpathian Orogen. The pronounced
560
Tectonic Models for the Evolution of Sedimentary Basins
18° E
22° E
re ei ss ey -T i st qu rn To
14° E
26° E
East European Platform
Bohemian Massif
W. Carpathians
49° N
ne zo
R an
ge
Vienna Basin
ia n
E. Carpathians 47° N
Tr
an
sd
an
ub
E. Alps
Pannonian Basin
Transylvanian Basin S. Carpathians 45° N
Dinarides Moesian Platform
1
2
3
4
5
6
7
8
9
10
0
100
200
300 km
Figure 64 Late Neogene structural pattern in the Pannonian basin system and its vicinity. 1: foreland (molasse) basins; 2: flysch nappes; 3: Neogene vulcanites; 4: pre-Tertiary units on the surface; 5: Variscan basement of the European plate; 6: Dinaric and Vardar ophiolites; 7: tectonic windows in the Eastern Alps; 8: normal and low-angle normal fault; 9: thrust, anticline; 10: strike-slip fault. Modified from Horva´th F (1993) Towards a mechanical model for the formation of the Pannonian basin. Tectonophysics 226: 333–357.
contrast in TIS between the Pannonian Basin (characterized by TIS <2.0 1012 N m1) and the Carpathian Orogen and its foreland (characterized by TIS >3.0 1012 N m1) indicates that recent lithospheric deformation is more likely concentrated in the hot and hence weak Pannonian lithosphere rather than in the surrounding Carpathians. By conversion of strength predictions to EET values at a regional scale, Lankreijer (1998) mapped the EET distribution for the entire Pannonian– Carpathian system (Figure 67). Calculated EET values are largely consistent with the spatial variation of lithospheric strength of the system. Lower values are characteristic for the weak central part of the Pannonian Basin (5–10 km), whereas EET increases toward the Dinarides and Alps (15– 30 km) and particularly toward the Bohemian Massif and Moesian Platform (25–40 km). This trend is in good agreement with EET estimates
obtained from flexural studies and forward modeling of extensional basin formation. Systematic differences, however, can occur and may be the consequence of significant horizontal intraplate stresses (e.g., Cloetingh and Burov, 1996) or of mechanical decoupling of the upper crust and uppermost mantle that can lead to a considerable reduction of EET values. The range of calculated EET values reflects the distinct mechanical characteristics and response of the different domains forming part of the Pannonian–Carpathian system to the present-day stress field. These characteristics can be mainly attributed to the memory of the lithosphere. In this respect it must be kept in mind that the tectonic and thermal evolution of these domains differed considerably during the Cretaceous–Neogene Alpine development of both the outer and intraCarpathians units and the Neogene extension of the
Tectonic Models for the Evolution of Sedimentary Basins
51°
200 km
55
45
A
55
30 .5 32
100 km 32.5
0 km
45
35
45
50
55
.5 37
35
50
.5
45 50
.5 57
32 37.5
40
55
32.5 30
48°
25
35
B C 30
.5
25
27
40
35
11°
35 40
40
42
45°
A‘
37.5
32 .5
35
32.5
27.5 42
30
.5
30
30
30
32.5
45
25
35
A
35
35
32.5
32
.5
30
60
35
40
55 50 45
25
40 45
50
561
30
.5
28° 20°
1
2
3
4
5
6
Figure 65 Crustal thickness in the Pannonian Basin and the surrounding mountains. Values are given in km. Letters A, B, and C indicate the location of strength envelopes shown in Figure 66(a). Regional strength profile A-A9 is shown in Figure 66(b). 1: foreland (molasse) foredeep; 2: flysch nappes; 3: pre-Tertiary units on the surface; 4: Penninic windows; 5: Pieniny Klippen Belt; 6: trend of abrupt change in crustal thickness. Modified from Horva´th F, Bada G, Szafia´n P, Tari G, A´da´m A, and Cloetingh S (2006) Formation and deofrmation of the Pannonian basin: Constraints from observational data. In: Gee D and Stephenson R (eds.) Geological Society, London, Memoirs, 32: European Lithosphere Dynamics, pp. 191–206. London: Geological Society, London.
Pannonian Basin, resulting in a wide spectrum of lithospheric strengths. These, in turn, exert a strong control on the complex present-day pattern of ongoing tectonic activity. 6.11.6.2 Neogene Development and Evolution of the Pannonian Basin 6.11.6.2.1 Dynamic models of basin formation
Following closure of oceanic basins in the Dinarides– Pannonian–Carpathian domain during the Cretaceous and Paleogene continental convergence phases of the Alpine orogeny, the style of tectonic deformation changed fundamentally in the areas of the Pannonian Basin. Whilst in the Carpathian arc large-scale tectonic transport of flysch nappes continued during the Miocene, crustal elements forming its internal parts were disrupted in response to strike–slip motions, extension and rigid body rotations, controlling the early development stages of the
Pannonian Basin (e.g., Balla, 1986; Sa˘ndulescu, 1988; Csontos et al., 1992; Roure et al., 1993; Kova´cˇ et al., 1994; Fodor et al., 1999). Several models have been proposed to explain the dynamics of Neogene rifting in the Pannonian Basin. An active versus passive mode of rifting has been a matter of continued debate (see Bada and Horva´th, 2001) resulting in the proposal of various dynamic models (Figure 68) that take into account such prominent features of the Pannonian–Carpathian system as thinned and hot versus thickened and colder lithosphere in its central and peripheral sectors, respectively. For instance, Sza´deczky-Kardoss (1967) and, at least in his early works, Stegena (1967) argued for the presence of a mantle diapir beneath the Intra-Carpathian area (Figure 68(a)). In this model upwelling of the asthenosphere beneath the Pannonian area caused thinning of the lithosphere and by active rifting subsidence of its central parts, whereas the nappe structure and the roots of surrounding orogens formed above the descending
Tectonic Models for the Evolution of Sedimentary Basins
(a)
A : Western Pannonian Basin
0
20 Crust Mantle
30 40
10
20 Crust Mantle
30 40
50
50 0
2000 1000 Strength (MPa)
C : Eastern Pannonian Basin
0
10 Depth (km)
10 Depth (km)
B : Central Pannonian Basin
0
Depth (km)
562
20 Crust Mantle
30 40
0
50
2000 1000 Strength (MPa)
0
2000 1000 Strength (MPa)
red ee p Fo
ns y Ba lvani sin an E Ca aste rpa rn thi an s
Tra
ts. en iM
Pannonian Basin
us
1.0
Ap
W Ca este rpa rn thi an s
red ee p
1.2
Fo
Integrated strength (×1013 N m–1)
(b)
0.8 0.6 0.4 0.2 0.0 0
100
(c)
200
300
m
Subsidence in basin center
c/δ m/β
c Moho
Mantle
Lithosphere
Crust
Basin flank
400 500 600 Distance along profile (km)
700
800
900
1000
Basin flank
Brittle T=1200° C
Ductile Brittle Ductile
Asthenosphere updoming T=1200° C
Figure 66 (a) Typical strength envelopes for the western (A), central (B), and eastern (C) parts of the Pannonian Basin. For locations see Figure 65. For calculation numerous constraints on the lithospheric structure and petrography, heat flow, strain rate, and stress regime have been adopted (for details, see Lankreijer et al., (1997), Sachsenhofer et al., (1997), Lenkey et al., (2002). Note the nearly complete absence of lithospheric mantle strength predicted by the model. (b) Total integrated lithospheric strength (TIS, in 1013 N m) along a regional profile through the Pannonian–Carpathian system (Lankreijer, 1998). (c) Schematic cross section showing nonuniform stretching of the Pannonian lithosphere and its effect on depth-dependent rheology. In the basin center, the thickness of the crust (c) and lithospheric mantle (m) has been reduced by stretching factors and , respectively. The ascending asthenosphere heats up the system, the isotherms become significantly elevated. As a result, the thinned and hot Pannonian lithosphere became extremely weak and, thus, prone to tectonic reactivation. Modified from Cloetingh S, Bada G, Mat¸enco L, Lankreijer A, Horva´th F, and Dinu C (2006) Thermo-mechanical modelling of the Pannonian-Carpathian system: Modes of tectonic deformation, lithospheric strength and vertical motions. In: Gee D and Stephenson R (eds.) Geological Society, London, Memoirs, 32: European Lithosphere Dynamics, pp. 207–221. London: Geological Society, London.
Tectonic Models for the Evolution of Sedimentary Basins
16°E
24°E
20°E
563
Alpine-Carpathian foredeep
15
Flysch nappes
BM
15–23
40
48°N
24
WC
Basement units on surface Neoegen volcanites EET (km)
2
–2
20
18
20
5–7
29
5–10
Alps
20 EC
7–9
PB 12–20
37
25 10
SC 45°N
Adriatic sea
20
10 30
Dinarides
MP
Figure 67 Effective elastic thickness (EET – in km) of the lithosphere in and around the Pannonian Basin predicted from rheological calculations. BM: Bohemian Massif; MP: Moesian Platform; PB: Pannonian basin; EC, SC, WC: Eastern, Southern and Western Carpathians, respectively. Modified from Lankreijer (1998) Rheology and basement control on extensional basin evolution in Central and Eastern Europe: Variscan and Alpine-Carpathian-Pannonian tectonics. PhD thesis, Vrieje Universiteit, Amsterdam, 158p.
branches of this convection cell. Whilst Horva´th et al., (1975) and Stegena et al., (1975) also proposed rifting as the driving mechanism for the evolution of the Pannonian Basin, their model applied plate tectonic concepts to the Pannonian region and considered basin subsidence, intense Neogene–Quaternary volcanic activity, extremely high heat flow, and the presence of an anomalous upper mantle and thinned lithosphere as closely related phenomena. These features were attributed to thermal thinning of the Pannonian lithosphere in response to upwelling of the mantle above the subducted European and Adriatic lithospheric slabs that dip beneath the Pannonian domain (Figure 68(b)). Other models stress the back-arc position of the Pannonian domain with respect to the Carpathian arc and postulate that gravity-driven passive roll-back of the subducting European lithospheric slab is the driving force for the tensional subsidence of the Pannonian Basin (Figure 68(c)) (e.g., Royden and Karner, 1984; Royden and Horva´th, 1988; Csontos et al., 1992; Horva´th, 1993; Csontos, 1995; Linzer, 1996). In a modification of this model, eastward mantle flow is thought to control roll back of the subducted slab and related retreat of its hinge line
(Figure 68(d)) (Doglioni, 1993). Both models account for passive rifting in the Pannonian Basin with tension being exerted on the overriding plate at its contact with the subducting plate by trench suction forces. Based on thermomechanical finite element modeling Huismans et al., (2001b) were able to simulated temporal changes in rifting style in the Pannonian Basin, suggesting a two-phase evolutionary scheme for the system. According to their model, the initial early Middle Miocene basin subsidence was driven mainly by passive rifting in response to roll-back of the subducted Carpathian slab, involving gravitational collapse of the thickened prerift Pannonian lithosphere. This triggered small-scale convective upwelling of the asthenosphere that has favoured the late-stage thermal subsidence (Horvath, 1993) as well as the compressional inversion of the Pannonian domain during Late Miocene–Pliocene times (e.g., Bada et al., 1999; Fodor et al., 2005). 6.11.6.2.2 Stretching models and subsidence analysis
Efforts to quantify the evolution of the Pannonian Basin started in the early 1980s with the application of classical basin analysis techniques. Due to the
564
Tectonic Models for the Evolution of Sedimentary Basins
(a)
(b)
Active rifting: mantle upwell (c)
Active rifting: melting of slab (d)
Passive rifting: slab retreat
Passive rifting: mantle flow
Figure 68 Dynamic models proposed for the evolution of the Pannonian Basin system. (a) Asthenospheric doming results in active rifting of the lithosphere above the central axis of the dome, whereas shortening is taking place in the peripheral areas. (b) Active rifting may also be caused by a subduction generated mantle diapir. (c) Hinge retreat of the subducting European margin driven by the negative buoyancy of the slab induces passive rifting in the overriding plate. (d) The same hinge retreat may be sustained by an eastward mantle flow pushing against the down going slab. Modified from Bada and Horva´th (2001).
availability of excellent geological and geophysical constraints, this basin has been a key area for testing stretching models. At the same time, the main characteristics of the Pannonian basin system, such as its extremely high heat flow, the presence of an anomalously thin lithosphere and its position within Alpine orogenic belts, made it particularly suitable and challenging for basin analysis. Research on the Pannonian Basin was triggered by its hydrocarbon potential and addressed local tectonics and regional correlations, as well as studies on its crustal configuration, magmatic activity and related mantle processes. Sclater et al., (1980) were the first to apply the stretching model of McKenzie (1978) to the intraCarpathian basins. They found that the development of peripheral basins could be fairly well simulated by the uniform pure shear extension concept with a stretching factor of about 2 ( ¼ 2). For the more centrally located basins, however, their considerable thermal subsidence and high heat flow suggested unrealistically high stretching factors up to 5. Thus, they postulated differential extension of the Pannonian lithosphere with moderate crustal stretching ( factor) and larger stretching of the lithospheric
mantle ( factor). Building on this and using a wealth of well data, Royden et al., (1983) introduced the nonuniform stretching concept according to which the magnitude of lithospheric thinning is depthdependent. This concept accounts for a combination of uniform mechanical extension of the lithosphere and thermal attenuation of the lithospheric mantle (Ziegler, 1992, 1996b; Ziegler and Cloetingh, 2004). This is compatible with the subsidence pattern and thermal history of major parts of the Pannonian Basin that suggest a greater attenuation of the lithospheric mantle as compared to the finite extension of the crust. Horva´th et al., (1988) further improved this concept by considering radioactive heat generation in the crust, and the thermal blanketing effect of basin-scale sedimentation. By reconstructing the subsidence and thermal history, and by calculating the thermal maturation of organic matter in the central region of the Pannonian Basin (Great Hungarian Plain), a major step forward was made in the field of hydrocarbon prospecting by means of basin analysis techniques. These studies highlighted difficulties met in explaining basin subsidence and crustal thinning in terms of uniform extension, and point toward the
Tectonic Models for the Evolution of Sedimentary Basins
applicability of anomalous subcrustal mantle thinning. This issue was central to subsequent investigations involving quantitative subsidence analyses (backstripping) of an extended set of Pannonian Basin wells and cross sections and their forward modeling (Lankreijer et al., 1995; Sachsenhofer et al., 1997; Juha´sz et al., 1999; Lenkey, 1999). Kinematic modeling, incorporating the concept of necking depth and finite strength of the lithosphere during and after rifting (Van Balen et al., 1999), as well as dynamic modeling studies (Huismans et al., 2001b),
suggested that the transition from passive to active rifting was controlled by the onset of subcrustal flow and small-scale convection in the asthenosphere. In order to quantify basin-scale lithospheric deformation, Lenkey (1999) carried out forward modeling applying the concept of nonuniform lithospheric stretching and taking into account the effects of lateral heat flow, flexure, and necking of the lithosphere. Calculated crustal thinning factors () indicate large lateral variation of crustal extension in the Pannonian Basin (Figure 69). This is consistent with the areal 20°E
16°E
565
24°E
WC ES Vi
48°N Da EA
Já De
TR St
Za
Bé AM Ma
Sa
Dr SC
45°N
DIN
Alpine-Carpathian foredeep Neogene volcanites
Flysch nappes Internal basement units
1.0
Crustal thinning factors, δ 1.2 1.4 1.6 1.8 2.0
Figure 69 Crustal thinning factors calculated by forward modeling for the Pannonian Basin applying the nonuniform stretching concept and taking the effects of lateral heat flow and flexure of the lithosphere into account (after Lenkey, 1999). Note the pronounced lateral variations in crustal extension controlling the development of deep sub-basins separated by areas of limited deformation. AM: Apuseni Mts.; DIN: Dinarides; EA: Eastern Alps; TR: Transdanubian Range; SC, WC: Southern and Western Carpathians, respectively. Local depressions of the Pannonian Basin system: Be´: Be´ke´s; Da: Danube; De: Derecske; Dr: Drava; ES: East Slovakian; Ja´: Ja´szsa´g; Ma: Mako´; Sa: Sava; St: Styrian; Vi: Vienna; Za: Zala. Modified from Cloetingh S, Bada G, Mat¸enco L, Lankreijer A, Horva´th F, and Dinu C (2006) Thermo-mechanical modelling of the Pannonian-Carpathian system: Modes of tectonic deformation, lithospheric strength and vertical motions. In: Gee D and Stephenson R (eds.) Geological Society, London, Memoirs, 32: European Lithosphere Dynamics, pp. 207–221. London: Geological Society, London.
566
Tectonic Models for the Evolution of Sedimentary Basins
pattern of the depth to the pre-Neogene basement (Horva´th et al., 2006). The indicated range of crustal thinning factors of 10–100% crustal extension is in good agreement with the prerift palinspastic reconstruction of the Pannonian Basin, and the amount of cumulative shortening in the Carpathian orogen (e.g., Roure et al., 1993; Fodor et al., 1999). As a major outcome of basin analysis studies, Royden et al., (1983) provided a two-stage subdivision for the evolution of the Pannonian Basin with a synrift (tectonic) phase during Early to Middle Miocene times, and a postrift thermal subsidence phase during the Late Miocene–Pliocene. Further development of the stratigraphic database, however, demonstrated the need to refine this scenario. According to Tari et al., (1999), the regional Middle Badenian unconformity, marking the termination of the syn-rift stage, is followed by a postrift phase that is characterized by only minor tectonic activity. Nevertheless, the subsidence history of the Pannonian Basin can be subdivided into three main phases that are reflected in the subsidence curves of selected sub-basins (Figure 70). The initial syn-rift phase is characterized by rapid tectonic subsidence, starting synchronously at about 20 Ma in the entire Pannonian Basin. This phase of pronounced crustal extension is recorded everywhere in the basin system and was mostly limited to relatively narrow, fault bounded grabens or sub-basins. During the subsequent postrift phase much broader areas began to subside, reflecting general down warping of the lithosphere in response to its thermal subsidence. This is particularly evident in the central parts of the Pannonian Basin, suggesting that in this area syn-rift thermal attenuation of the lithospheric mantle played a greater role than in the marginal areas (e.g., Sclater et al., 1980; Royden and Do¨ve´nyi, 1988). The third and final phase of basin evolution is characterized by the gradual structural inversion of the Pannonian Basin system during the Late Pliocene–Quaternary. During these times intraplate compressional stresses gradually built up and caused basin-scale buckling of the Pannonian lithosphere that was associated with latestage subsidence anomalies and differential vertical motions (Horva´th and Cloetingh, 1996). As evident from subsidence curves (Figure 70), accelerated latestage subsidence characterized the central depressions of the Little and Great Hungarian Plains (Figure 70(b)–(c)), whereas the peripheral Styrian and East Slovakian sub-basins were uplifted by a few hundred meters after mid-Miocene times (Figure 70(d)–(e)) and the Zala Basin during the Pliocene–Quaternary (Figure 70(f)). The importance
of tectonic stresses, both during the rifting (extension) and subsequent inversion phase (compression), is highlighted by this late-stage tectonic reactivation, as well as by other episodic inversion events in the Pannonian Basin (Horva´th, 1995; Fodor et al., 1999). For the Carpathian foreland, modeling curves (Figure 70(h)–(j)) indicate an important Late Miocene (Sarmatian) phase of basin subsidence that relates to tectonic loading of the Eastern and Southern Carpathian foreland by intra-Carpathian terranes. This phase is coeval with the end of synrift subsidence of the Pannonian Basin (Horva´th and Cloetingh, 1996). Subsidence curves for the Transylvanian Basin (Figure 70(g)) indicate for the Badenian–Pannonian a subsidence pattern similar to that of the Carpathian foreland, and for the Pliocene– Quaternary a phase of uplifting that correlated with the inversion of the Pannonian Basin. 6.11.6.3 Neogene Evolution of the Carpathians System During the last few years, research on the Carpathian system focused on spatial and temporal variations in thrusting, lateral changes in the flexural response of the foreland lithosphere and the development of an unusual foredeep geometry in the Focs¸ani Depression. Reconstruction of orogenic uplift and erosion (e.g., Sanders et al., 1999), coupled with foreland subsidence modeling (Mat¸enco et al., 2003) elucidated the complex interplay between syn-orogenic deflection of the foreland lithosphere and its lateral variability, subsequent slab-detachment-related isostatic rebound of the orogenic wedge, and increased subsidence in the SE Carpathians corner (Mat¸enco et al., 1997b; Zoetemeijer et al., 1999; Bertotti et al., 2003; Ta˘ra˘poanca˘ et al., 2003, 2004b; Cloetingh et al., 2004). Tertiary subsidence patterns recorded in the Carpathian foreland units reflect significant vertical motions (e.g., Mat¸enco et al., 2003; Bertotti et al., 2003 – Figure 70(h)–(j)). In the foreland of the Southern Carpathians, corresponding to the Getic Depression and western Moesian Platform, up to 5 km thick Early Miocene sediments accumulated in a transtensional basin, whilst adjacent platform areas were characterized by nondeposition and/or erosion. Starting in the Middle to Late Miocene, the entier Carpathian foreland subsided flexurally in response to thrust loading. In the evolving foreland basin a major depocenter developed in the Carpathian Bend Zone, corresponding to the Focs¸ani Depression
Tectonic Models for the Evolution of Sedimentary Basins
(Figure 71), in which Late Miocene sedimentation rates reached 1500–3000 m My1. During the Pliocene–Pleistocene, when movements on the Carpathian thrusts had essentially ceased, subsidence of the Focs¸ani Depression continued at sedimentation rates averaging, 200–300 m My1. Regional subsidence analyses on the Carpathian foreland basin demonstrate that comparable kinematically related episodes of vertical motions occurred simultaneously also in the frontal parts of the Eastern and Southern Carpathians (Mat¸enco et al., 2003; Bertotti et al., 2003), with limited to no evidence for depocenter migration in the East-Carpathian foreland, contrasting with previous inferences (e.g., Meulenkamp et al., 1996).
6.11.6.3.1 Role of the 3-D distributions of load and lithospheric strength in the Carpathian foredeep
The Focs¸ani Depression, that contains almost 13 km of Middle Miocene to Recent sediments, is located in front of the Carpathian Bend Zone (Figure 71), next to the Vrancea seismogenic area. This basin is of great interest insofar as its partly pronounced postthrusting subsidence and the position of its depocentre in front of the thrust-belt are anomalous in terms of flexural foreland basin development. To explain these phenomena, several models invoked a lithospheric slab that sinks beneath the Focs¸ani Depression into the asthenosphere and accounts for the seismicity of the Vrancea area (Figure 72). This slab is interpreted as having formed either by SEward roll-back and detachment of the subducted oceanic part of the lower plate (Royden, 1993; Linzer, 1996; Wortel and Spakman, 2000), associated with its tearing along the Trotus¸ fault (Mat¸enco and Bertotti, 2000), or by postcollisional delamination of the continental lithospheric mantle from the foreland (Giˆrbacea and Frisch, 1998; Chalot-Prat and Giˆrbacea, 2000). The so-called hidden loads have been inferred to affect the SE Carpathians and their foredeep basins (e.g., Royden, 1993). As modeling studies showed that the topographic load of the Carpathians is insufficient to account for the observed foredeep geometry, an extra vertical force had to be applied to the deep end of the subducting plate (Royden and Karner, 1984). Moreover, these models inferred large lateral variations in the rigidity of the Carpathian foreland lithosphere (Mat¸enco et al., 1997a), possibly inherited from precompressional stages.
567
The deepest part of the Focs¸ani Depression (Figure 73, see also Figure 82) is located immediately in front of the Carpathian Bend zone and is limited to the W and NW by the Carpathian thrust front. This basin shallows out rapidly towards the N and SW along the front of the Eastern and Southern Carpathians and SE-ward towards the foreland platform.The 3-D architecture of the Focs¸ani Depression and its evolution through time is well constrained by an extensive reflection-seismic survey. These data show that this basin developed in two stages, namely during the Middle Miocene (Badenian) in response to NE–SW directed extension, and from the Late Miocene onward in response to regional tilting and subsidence that was not accompanied by faulting (Ta˘ra˘poanca˘ et al., 2003). Modeling of the Focs¸ani Depression aimed at assessing how much of its total subsidence can be attributed to syn-rift tectonics and postrift thermal subsidence and whether the remaining subsidence can be explained by foreland flexure in response to 3-D loads in the presence of lateral changes in the flexural strength of the foreland lithosphere (Ta˘ra˘poanca˘ et al., 2004). 6.11.6.3.1.(i)
Preorogenic extensional basin
Reflection-seismic data define a set of NW–SE trending normal faults along the eastern margin of the Focs¸ani Depression, outlining half-grabens that contain up to 4 km thick Badenian syn-rift sediments and that are bounded by ENE sticking transfer faults (Figures 73 and 74). The thickness of Badenian sediments decreases generally toward the eastern shoulder of this graben system that is superimposed on the North Dobrogean orogen. As to the NW and W the Badenian graben system is overridden by the Carpathian nappes, its full extent cannot be reconstructed. The Badenian syn-rift sediments are capped by a nearly regional unconformity that corresponds to the Badenian/Sarmatian boundary and marks the syn- to postrift transition (Ta˘ra˘poanca˘ et al., 2004). Badenian extension in the Carpathians Bend foreland was modeled (Ta˘ra˘poanca˘ et al., 2004) using a forward 2-D numerical modeling code that simulates the kinematic and thermal behavior of extending lithosphere (Kooi, 1991). In this model, prerift lithospheric and crustal thicknesses are an input parameter. As in the area of the Carpathian Bend zone presentday lithospheric and crustal thicknesses range between 170 and 190 km, and 30 to 40 km, respectively (Ra˘dulescu, 1988; Horva´th, 1993; Nemcok et al., 1998), a prerift lithospheric thickness of 180 m was adopted. Differential stretching factors were then
Tectonic Models for the Evolution of Sedimentary Basins
16°E
(a)
20°E
(b)
24°E
0
1 Basement subsidence Great Hungarian Plain
1
3
48°N
EA
Depth (km)
2
WC
EC
2 4
7
PanBas 1
5
AM
6
SC
45°N
3 4 5 6
8
(c)
Tectonic subsidence 2 Little Hungarian Plain
(d)
1.0
1.5
Synrift OK
18
BAD
16
14
Postrift SA
12
0.0
PAN
10
Inversion
PO
8
6
PL
16
14
Postrift SA
12
PAN
10
Inversion
PO
8
6
Age (Ma)
0.0
PL
4
Q
2
0
Tectonic subsidence East Slovak basin
3
4
2
Age (Ma) Tectonic subsidence 4 Styrian basin
1.0
1.5
Synrift
2.0
Q
0
OK
18
(f)
BAD
16
14
Postrift SA
PAN
12
10
Inversion
PO
8
6
PL
4
Q
2
Age (Ma) Basement subsidence 5 Zala basin
0 1
0.5
0
2
Depth (km)
Depth (km)
BAD
0.5
Depth (km)
Depth (km)
0.5
(e)
OK
18
0.0
2.0
Synrift
7
9
DIN
1.0
1.5
3 4 5 6
2.0
Synrift OK
18
(g)
Postrift
BAD
SA
16 14
12
0.0
PAN
10
Inversion
PO
8
6
PL
4
2
0
Age (Ma) Basement subsidence 6 Transylvanian basin
18
(h)
Depth (km)
Depth (km)
1.5
Pre- and postcollision OK
18
BAD
16
14
SA
12
Postcollision
BAD
16
14
Postrift SA
12
0.0
PAN
10
Inversion
PO
8
6
PL
4
Q
2
0
Age (Ma) Basement subsidence 7 E. Carpathian foreland
PAN
10
PO
8
6
1.0
1.5
Pre- and postcollision
Inversion PL
4
2.0
Q
2
0
Age (Ma) Basement subsidence 8 Bend Zone foreland
0.5
1.0
1.5
OK
18
BAD
16
14
Postcollision
SA
PAN
12
10
PO
8
6
PL
4
Q
2
0
Age (Ma) Basement subsidence 9 S. Carpathian foreland
(j) 0.0
Depth (km)
(i) 0.0
2.0
OK
0.5
1.0
2.0
Synrift
7
Q
0.5
Depth (km)
568
0.5
1.0 1.5
Pre- and postcollision O K
18
16
BAD
14
SA
12
Postcollision PAN
10
PO
8
Age (Ma)
6
Inversion PL
4
2.0
Q
2
0
Pre- and postcollision O K
18
BAD
16
14
Postcollision
SA
PAN
12
10
PO
8
Age (Ma)
6
PL
4
Q
2
0
Tectonic Models for the Evolution of Sedimentary Basins
20 E
16 E 200 km
24 E
569
External moldavides Internal moldavides Outer dacides Middle dacides Internal dacides Dinarides Transylvanides Neogene volcanics Paleogene volcanics
50 N
Wien Bratislava
East European platform Budapest
Debrecan Cluj Napoca
Scy
thia
46 N zagreb
n pl
FD IMF
Beograd
fig.2 outline
BF rm
atfo
No
Sibiu
rth TF 46 N Do b PCF roge a
Bucharest
Moesian platform
BLACK SEA 42 N 16 E
20 E
Sofia 24 E
42 N 28 E
Figure 71 General map of the Carpathian–Pannonian system and location of study area. FD approximate location of t Focs¸ani Depression. BF, IMF, PCF, and TF are Bistria, Intramoesian, Peceneaga–Camena, and Trotus¸ faults, respectively, forming the boundaries between different lithospheric domains (platforms) of the Carpathians foreland plate. Modified from Sa˘ ndulescu M (1984) Geotectonics of Romania (in Romanian). Bucharest: Editrons Tehnica.
assigned to the crust and lithospheric mantle to simulate depth-dependent extension whilst basin subsidence was computed for the center of each box. The cross section chosen for the 2-D modeling is given in Figure 73. By flattening the top Badenian marker the configuration of the rifted basin prior to thrusting and flexural loading was restored (Figures 74 and 75). For modeling purposes, it was assumed that this basin was filled with sediments up to sea level. Moreover, its fill was decompacted, assuming a silty sand lithology. In order to fit the basin geometry,
a range of stretching factors was tested until a good fit between observed and modeled geometries was achieved. The duration of rifting was assumed to be 3 My, corresponding to Badenian times. Further input values are the EET and necking depth of the lithosphere with the former defined either as a constant or being associated with a specific isotherm. Modeling parameters used are summarized in Table 7 (Ta˘ra˘poanca˘ et al., 2004). The modeled profile extends beyond the width of the basin to avoid boundary effects. The best-fit model is shown
Figure 70 Subsidence curves for selected sub-basins of the Pannonian Basin and the Carpathian foreland. Note that after a rapid phase of general subsidence throughout the entire Pannonian Basin, the sub-basins show distinct subsidence histories from Middle Miocene times onward. Arrows indicate generalized vertical movements. Timing of the syn- and postrift phases t after Royden et al., (1983). Timing of Carpathians collision after Maenco et al., (2003). For timescale the central Paratethys stages are used. O, K: Ottnangian and Karpatian, respectively (Early Miocene); BAD, SA: Badenian and Sarmatian, respectively (Middle Miocene); PAN, PO: Pannonian and Pontian, respectively (Late Miocene); PL: Pliocene; Q: Quaternary. AM: Apuseni Mts.; DIN: Dinarides; EA: Eastern Alps; PanBas: Pannonian Basin; SC, WC: Southern and Western Carpathians foreland basins, respectively. Modified from Cloetingh S, Bada G, Mat¸enco L, Lankreijer A, Horva´th F, and Dinu C (2006) Thermo-mechanical modelling of the Pannonian-Carpathian system: Modes of tectonic deformation, lithospheric strength and vertical motions. In: Gee D and Stephenson R (eds.) Geological Society, London, Memoirs, 32: European Lithosphere Dynamics, pp. 207–221. London: Geological Society, London.
570
Tectonic Models for the Evolution of Sedimentary Basins
0
East Carpathians nappes 150 200
Transylvania Basin 100 0
250
Focsani Basin Foreland 300 km 0
0 20 40 60
Transylvanides
Middle Dacides (upper plate)
Crustal earthquakes < 20 km > 20 km
Moesia and westward units (lower plate)
Dobrogea
20 40
Restored, Late Miocene (11Ma) collisional Moho
80
Present-day Moho
60 80
100
100 P-wave velocity anomalies
120
+1.0% + 0.0% –1.0%
140 160
120 140 160 180
200
200
220
220
240
km
km
180
240
Figure 72 Simplified cross-section through the Romanian Carpathians showing the main tectonic units of the upper and lower tectonic plates and the thin-skinned thrustbelt that developed at their contact. Present-day Moho after Diehl et al., (2005). The Moho surface reconstructed for the moment of collision (11 Ma, gray dashed line) was constructed by retrodeforming the cumulative postcollisional (latest Miocene–Quaternary) vertical movements. Earthquakes from the SE Carpathians were projected into the cross section as a function of depth and magnitude. Crustal seismicity was projected perpendicular to the cross section (i.e., along the strike of Quaternary folds of the Focs¸ani Basin). Mantle earthquakes were projected along the strike of the NE–SW oriented intermediate Vrancea mantle slab, oblique to the trace of the cross section. The ternary diagram represents the types of focal mechanisms for the crustal earthquakes that apparently do not display any preferred type of fault plane solution. Mantle P-wave velocity anomalies after the regional seismic tomography of Wortel and Spakman (2000). Note that the location and geometry of these anomalies are only qualitative due to the large SE Carpathians zoom. For complete sections at the regional Carpathians scale, see Wortel and Spakman (2000). Modified from Mat¸enco et al., (2006); see also Schmid et al., (2006).
in Figure 75. This model overestimates, however, the uplift of the eastern rift shoulder as the top Badenian marker horizon does not extend across it. Apart from this, a good fit is obtained for most of the profile with minor differences in its westernmost part probably being related to post-Badenian movements along distributed faults (Ta˘ra˘poanca˘ et al., 2004). The best-fit model was obtained using an EET of 35 km, a necking depth of 25 km, a prerift crustal thickness of 32 km and assuming uniform extension of the crust and lithospheric mantle. Stretching factors are small and even in the deepest central basin do not exceed 1.1 (Figure 75). The rift-induced thermal anomaly is very small due to the low stretching factors and lateral heat dissipation. Correspondingly, the post-Badenian postrift subsidence is less than 100 m or nil (Figure 76).
Sensitivity studies showed that changing the prerift crustal thickness from 30 to 40 km had little effect on modeling results. By contrast, changing in the EET and particularly the necking depth caused important changes in results. Assuming the same distribution of stretching factors as in Figure 75 and a crustal thickness of 32 km, basin geometries for four necking depths and for four EET values, respectively, are shown in Figure 77. Due to the small stretching factors, changes in these two parameters affect particularly the depth of the central and central-western part of the basin. As in essence the same basin and the same amount of stretching are obtained, using a lower EET and a shallower necking depth and vice versa, no unique solution can be derived from the basin shape alone.
Tectonic Models for the Evolution of Sedimentary Basins
SW 0
571
NE 30
60
90
120
150 km 0
Peceneaga-Camena. f.
Cross section 1
M
O
E
S
I
A
North Dobrogea Orogen
2
4 Pliocenequaternary Uppermost Miocene (Pontian) Upper Miocene (Sarmatian-Meotian) Middle Miocene (Badenian) Pre-Tertiary
6
8 km
Figure 73 Cross section through southern flank of Focs¸ani Depression showing Badenian extensional basins. Based on maps derived from the interpretation of industrial seismic lines. Modified from Ta˘ rapoa˘ nca˘ M, Bertotti G, Mat¸enco L, Dinu C, and Cloetingh S (2003) Architeccture of the Focsani depression: A 13 km deep basin in the Carpathians bend zone (Romania). Tectonics 22: 1074 (doi:10.1029/2002TC001486).
WSW
ENE
2s
Top Meotian
2s
Top Sarmatian
3s
Top Badenian
4s
Pre-Tertiary
3s
4s 0
1
2 km
Figure 74 Example of an interpreted seismic line across Badenian fault blocks in the Focs¸ani Depression. Modified after Ta˘ rapoa˘nca˘ M, Dinu C, and Ciulavu D (2004b) Neogene kinematics of the northeastern sector of the Moesian platform (Romania). American Association of Petroleum Geologists Bulletin in press.
0
30
60 predicted
90
120
150
180 km
observed
0
2 km
0
Znecking = 25 km EET = 35 km
Lithospheric thickness = 180 km Crustal thickness = 32 km 3 km
1.1
1.1 1.05
1.05 1.0
2 km
Thinning factor
1.0
Figure 75 Extensional modeling results for Badenian basins. Upper panel: dark gray field shows configuration of Badenian basins derived from Figure 73 by flattening the top Badenian marker. Bold line shows best-fit modeled cross section adopting stretching factors given in lower panel. Lower panel: inferred stretching factors (same for crust and lithospheric mantle) plotted for constant 3 km wide steps. Modified from Ta˘ rapoa˘ nca˘ et al., (2004a).
572
Tectonic Models for the Evolution of Sedimentary Basins
Nevertheless, as the model predicts almost no postrift subsidence, only a very small part of the observed Sarmatian–Quaternary subsidence can be attributed to Badenian rifting. Moreover, owing to the small amount of stretching, the rift-induced strength reduction of the Moesian Platform is presumably also small.
Table 7 Parameters used as input data in the extensional modeling of the Focs¸ani depression Temperature Density at surface conditions
Surface Asthenosphere Crust
0 C 1333 C 2800 kg m3
Lithospheric mantle
3330 kg m3 2660 kg m3
Sediment grain density Thermal expansion coefficient Thermal diffusivity
6.11.6.3.1.(ii) Flexural modeling of the foredeep basin The foredeep basin of the
3.4 105 K1 7.8 107 m2 s1
180
210
240
270
300
Carpathian Bend Zone is in so far peculiar as its depocenter does not lie beneath this thrust belt, as predicted by simple flexural models. Shallowing of 330
360
450 km
390
Postrift 0
0 Synrift
1
1
2 km
Lithospheric thickness = 180 km Crustal thickness = 32 km
2 km
Znecking = 25 km EET = 35 km
Figure 76 Syn- and postrift subsidence obtained for the best-fit model (see Figure 75). Modified from Ta˘ rapoa˘ nca˘ M, Garcia-Castellanos D, Bertotti G, Mat¸enco L, Cloetingh S, and Dinu C (2004a) Role of 3-D distributions of load and lithospheric strength in orogenic arcs: polystage subsidence in the Carpathians foredeep. Earth and Planetary Science Letters 221: 163–180.
(a) 180
210
240
270
300
330
360
390
420
0
0
2 km
450 km
Lithospheric thick. = 180 km Crustal thick. = 32 km EET = 35 km
2 km
Zneck. = 15 km Zneck. = 20 km Zneck. = 25 km Zneck. = 32 km
(b) 180
210
240
2 km
270
300
330
360
390
420
450 km 0
0 Lithospheric thick. = 180 km Crustal thick. = 32 km Znecking = 25 km EET = 15 km EET = 25 km EET = 35 km EET = 50 km
2 km
Figure 77 Sensitivity analysis on basin depth for best-fit model (Figure 75). Effect of (a) necking depth variations, (b) elastic thickness variations. Modified from Ta˘ rapoa˘ nca˘ M, Garcia-Castellanos D, Bertotti G, Mat¸enco L, Cloetingh S, and Dinu C (2004a) Role of 3-D distributions of load and lithospheric strength in orogenic arcs: polystage subsidence in the Carpathians foredeep. Earth and Planetary Science Letters 221: 163–180.
Tectonic Models for the Evolution of Sedimentary Basins
this foredeep towards the thrust belt appears to be associated with the presence of lithospheric blocks characterized by significantly different strengths within the Carpathian domain. Following an initial cross-section experiment, a planform modeling approach was taken by Ta˘ra˘poanca˘ et al., (2004) to test: (1) the effects of a 3-D load distribution and lateral variations in lithospheric strength on the position of the basin, and (2) whether the observed very large subsidence is the effect of 3-D loading rather than of a hidden load. In the Carpathian Bend zone, flexural subsidence of the foreland in response to nappe emplacement commenced during the early Sarmatian (Sa˘ndulescu, 1984, 1988; Mat¸enco et al., 2003). The thickness of Sarmatian deposits gradually increases from the foreland to about 2.1 km near the present-day thrust front (Figure 78). By contrast, the thickness of Meotian deposits decreases from the axial parts of the Focs¸ani Depression towards the Carpathian deformation front (Figure 78(a)), indicative for the Late Miocene progressive advance of the latter. During the Early Pliocene, emplacement of the frontal (a) WSW 0
30
60
ENE 90 km
573
Carpathian thrust elements at their present position involved the development of a typical triangle zone. This resulted in uplift and eastward tilting of the western margin of the Focs¸ani Depression whereas its central parts continued to subside (Mat¸enco et al., 2003; see their Figure 5 cross section B). In the deepest parts of the Focs¸ani Depression, which are located in front of the Carpathian thrust front, Pliocene and Quaternary sediments attain a thickness of over 4 km and rest on 4 km thick Late Miocene deposits (Figure 78(a)). The eastern flank of the Focs¸ani Depression dips gently towards the basin center whilst its western flank dips steeply eastward with outcropping Late Miocene strata displaying nearly vertical dips (Dumitrescu et al., 1970). Flattening the sedimentary fill of the Focs¸ani Depression at the base-Pliocene level indicates that that this basin was considerably broader during the Late Miocene than at present. To the S of the Focs¸ani Depression, the Late Miocene (Sarmatian) base reaches a depth of 3.7 km in front of the Carpathians belt and shallows southward with the Pliocene–Quaternary sequence showing roughly
(b) NW 0
30
SE 90 km
60
Cross section 3
2 2 Pliocenequaternary
4
Uppermost Miocene (Pontian)
4
Upper Miocene (Sarmatian-Meotian) Middle Miocene (Badenian)
6
Pre-Tertiary
6 km 8
(c) WSW 0
ENE 30
60
90 km
PlioceneQuaternary
10
0
Uppermost Miocene (Pontian) Upper Miocene (Sarmatian-Meotian) Middle Miocene (Badenian)
12
km
Upper Miocene (Sarmatian and only locally Meotian) Middle Miocene (Badenian)
Pre-Tertiary
Cros section 2
Cross section 4
2
Pre-Tertiary
4 km
Figure 78 Cross sections through Focs¸ani Depression: (a), central part, (b) southern flank, and (c) northern flank. Modified from Ta˘ rapoa˘ nca˘ M, Garcia-Castellanos D, Bertotti G, Mat¸enco L, Cloetingh S, and Dinu C (2004a) Role of 3-D distributions of load and lithospheric strength in orogenic arcs: polystage subsidence in the Carpathians foredeep. Earth and Planetary Science Letters 221: 163–180.
574
Tectonic Models for the Evolution of Sedimentary Basins
the same thickness as the Late Miocene one (Figure 78(b)). The East-Carpathian orogenic wedge has overridden lithospheric blocks with very different characteristics, namely the East-European Platform, the Dobrogean Orogen and the Moesian Platform, the boundaries of which are sharp and fault controlled (e.g., the Trotus¸ fault) (Figure 73).
To illustrate the effect of lateral strength changes of a loaded plate on the geometry of a flexural basin, the deflection of a 2-D thin elastic plate (e.g., Turcotte and Schubert, 1982) under the weight of a rectangular load was calculated for different EET distributions (Figure 79). In these calculations, topography after lithospheric deflection was set to 2 km in the orogen and to zero elsewhere. In the case of a
(a) EET (km)
30 20 10 0 –100
–50
0 x (km)
50
100
3
Elevationn (km)
2 1 2700 kg m–3
0 –1
2200 kg m–3
–2 3200 kg m–3
–3 –4 –5 –6 –100
–50
0 x (km)
50
100
(b) 7 20 15 10 5 0 –5 –10
5 4 3 2
Position of depocenter (km) relative to the outer margin of the load (+ beneath load; –out of load)
Maximum deflection (km)
EET ‘foreland’ = 5 km 6
Basin in front of the load
1 5
10
15
20
25 30 35 EET (km) ‘loaded plate’
40
45
50
Figure 79 (a) Flexural deflection profiles for two different EET distributions applying the same topographic load. With a uniform EET the maximum deflection occurs beneath the load (dotted line). With a sharp decrease in EET at the right edge of the load the maximum deflection shifts toward the weaker foreland (bold line). (b) Deflection maximum as a function of the strength of the loaded plate (black line) and position of the depocenter relative to the load (gray line). When the EET of the lithosphere is significantly lower in the foreland than beneath the orogen, the deepest part of a foredeep shifts away from the orogen. Modified from Ta˘ rapoa˘ nca˘ M, Garcia-Castellanos D, Bertotti G, Mat¸enco L, Cloetingh S, and Dinu C (2004a) Role of 3-D distributions of load and lithospheric strength in orogenic arcs: polystage subsidence in the Carpathians foredeep. Earth and Planetary Science Letters 221: 163–180.
Tectonic Models for the Evolution of Sedimentary Basins
constant EET, the zone of maximum deflection is located directly beneath the imposed load. On the other hand, with a laterally changing EET of the foreland plate and the load being imposed on its stronger part, the maximum deflection tends to shift away from directly beneath the load towards the weaker part of the plate, and, depending on the conditions, can even be separated from the load to the end that the foredeep basin shallows toward the load, as seen in the Focs¸ani Depression. This suggests that the subsidence of the unusually deep SE Carpathians foredeep basin may be related to a reduction of the EET of its substrate, possibly caused by preorogenic extension. To further investigate the subsidence mechanism of the Focs¸ani Depression, Ta˘ra˘poanca˘ et al., (2004) used a flexural model to calculate the deflection of a foreland lithosphere that consists of various blocks with different EET (see, e.g., Van Wees and Cloetingh, 1994 for governing equations) in response to a laterally variable load (see Garcia-Castellanos et al., 2002 for methodology). The load was calculated from the observed topography of the Carpathians region with the evolving flexural basin being filled with constant density material. For the post-Badenian subsidence of the Focs¸ani Depression, a flexural model was used, taking into account the plan view distribution of topographic loads and lateral variations in the lithospheric strength. Assigning realistic strengths to the different domains of the Carpathian region, a basin is predicted in front of the Carpathians Bend Zone where the deflection of the foreland lithosphere is localized owing to its preexisting extensional weakening and to Quaternary crustal faulting at the transition between the rheologically weak Moesia and the northern more stable blocks (East-Europe/Scythia/North Dobrogea) (Mat¸enco et al., 2006). As the deepest modeled basin is only 3.3 km deep (40% of the observed value), either an additional load is required in the Carpathians Bend Zone or intraplate stresses have to be invoked to explain the geometry and depth of the Focs¸ani Depression. Additional loads applied to the foreland lithosphere may be attributed to the weight of the steeply dipping Vrancia slab that is located beneath the Focs¸ani Depression (Figure 72) or to an unrealistic greater topographic load. Modeling of the effects of intraplate stresses on the geometry of the Carpathian foredeep (Ta˘ra˘poanca˘ et al., 2004), applying an NW–SE directed compressional stress field during postcollisional evolution
575
of the Carpathians, that is, the latest MioceneQuaternary (see Mat¸enco and Bertotti, 2000) indicates that with a 500 m topographic load the depth of the Focs¸ani Depression increases to 6.5 km, corresponding to 75–80% of its post-Badenian subsidence (i.e., postextensional Carpathians thrust loading and postcollisional inversion), whilst with a topographic load of 800 m, its actual depth and depocenter location can be simulated (Figure 80). Modifying the magnitude of the intraplate stresses causes in the depocenter depth variations in the order of few hundred meters only. However, similar as models without intraplate stresses, the predicted basin width is larger than the observed one. Modeling indicates that the previously neglected extensional weakening of the eastern Moesian foreland, combined with compressional intraplate stresses, as well as the 3-D distribution of the topographic load and lithospheric strength variations may play an important role in the subsidence pattern of the SE Carpathian foreland. As such, it suggests a potential alternative to models that exclusively invoke topographic loading and slab pull forces to account for the observed subsidence of this foredeep, and specifically of the Focs¸ani Depression. 6.11.6.4 Deformation of the Pannonian– Carpathian System The present-day deformation pattern and related topography development in the Pannonian– Carpathian system is characterized by pronounced spatial and temporal variations in the stress and strain fields (Figure 81) (see, e.g., Cloetingh et al., 2006). Horva´th and Cloetingh (1996) established the importance of Late Pliocene and Quaternary compressional deformation of the Pannonian Basin that explains its anomalous uplift and subsidence, as well as intraplate seismicity. Based on the case study of the Pannonian–Carpathian system, these authors established a novel conceptual model for the structural reactivation of back-arc basins within orogens. At present, the Pannonian Basin has reached an advanced evolutionary stage as compared to other Mediterranean back-arc basins in so far as it has been partially inverted during the last few million years. Inversion of the Pannonian Basin can be related to temporal changes in the regional stress field, from one of tension that controlled its Miocene extensional subsidence, to one of Pliocene–Quaternary compression resulting in deformation, contraction, and flexure of the lithosphere associated with differential vertical motions.
Tectonic Models for the Evolution of Sedimentary Basins
(a)
(b)
700000
700000
600000
600000
–2
–2 00 0
29 km
00
0
37 km
500000
500000
–3
00
400000
300000
300000
200000
200000
0
26 km
CS 4
–2
–10
00
400000
–20 00
–40
00
l lo na
100000
Ad d
itio
–1
0
00
0
CSmax.
–2000
CS 3
–1000
0
CS 2
–3000
ad
0
1000
00
100000
10 km
30 km
–100000
–100000 –300000 –200000 –100000
0
–300000 –200000 –100000
100000 200000 300000 400000 500000
(c)
0
30
60
90 km
0
30
60
90 km
Sarmatian base
Predicted deflection (CS max.) 0m 50
80
0
m
2
Sarmatian base
4
500
i he
gh
mh
igh
m 00
he
8
t
Cross section 3
km Topography
0 0
6
2 km
8 km
t
ht
eig
ht ig he
4
Predicted deflection
Topography
0
2
0 100000 200000 300000 400000 500000
0
Topography
30
60
Sarmatian base
500 m
heigh
90 km
Predicted deflection
height 800 m
t
Cross section 2
Cross section 4
(d)
0
30
60
90 km
0
700000
Sarmatian base –2
600000 –2
00
0
Topography 00
0
37 km
29 km
2
500000
0 –100
–4 00
–1
0
0
ht ig he
6 CSmax.
00
0
0
CS 2
00
–3000
–3
–1000
00
200000
100000
50
–20
300000
he
80
4
0m
0m
26 km
400000
ht
ig
Predicted deflection (CS max.)
–3 00 0
–200 0
576
–2000
–1000
10 km
30 km
8 km
–100000 –300000 –200000 –100000
0
100000 200000 300000 400000 500000
N–S intraplate force = 2E 12 Pa m
Cross section 2
Tectonic Models for the Evolution of Sedimentary Basins
20°E
15°E
25°E
3500 1600 1400
50°N
1000 750
BM
– EA
+
–
D
+ 45°N
–
200
+
100
150 10
+ +
TR
+
WC
–
EC
GHP
+ TB
–
–
0?
SC
S DI
+
+ F –
MP
+ +
BA
Figure 81 Topography of the Pannonian–Carpathian system showing present-day maximum horizontal stress (SHmax) trajectories (after Bada et al., 2001). ‘þ’ and ‘’ symbols denote areas of Quaternary uplift and subsidence, respectively. BA: Balkanides; BM: Bohemian Massif; D: Drava Trough; DI: Dinarides; EA: Eastern Alps; EC: Eastern Carpathians; F: Focs¸ani Depression; MP: Moesian Platform; PB: Pannonian Basin; S: Sava Trough; SC: Southern Carpathians; TB: Transylvanian Basin; TR: Transdanubian Range; WC: Western Carpathians. Modified from Cloetingh S, Bada G, Mat¸enco L, Lankreijer A, Horva´th F, and Dinu C (2006). Thermomechanical modelling of the Pannonian-Carpathian system: Modes of tectonic deformation, lithospheric strength and vertical motions. In: Gee D and Stephenson R (eds.) Geological Society, London, Memoirs, 32: European Lithosphere Dynamics, pp. 207–221. London: Geological Society, London.
Therefore, the spatial distribution of uplifting and subsiding areas within the Pannonian Basin can be interpreted as resulting from the buildup of intraplate compressional stresses, causing large-scale positive and negative deflection of the lithosphere at various scales. This includes basin-scale positive reactivation of Miocene normal faults, and large-scale folding of the system leading to differential uplift and subsidence of anticlinal and synclinal segments of the
577
Pannonian crust and lithosphere. Model calculations are in good agreement with the overall topography of the system (Figure 81). Several flat-lying, low-elevation areas (e.g., Great Hungarian Plain, Sava and Drava troughs) subsided continuously since the Early Miocene beginning of basin development and contain 300–1000 m thick Quaternary alluvial sequences. By contrast, the periphery of this basin system, as well as the Transdanubian Range, the Transylvanian Basin and the adjacent Carpathian orogen were uplifted and considerably eroded from Late Miocene–Pliocene times onward (see Figures 70 and 82). Quantitative subsidence analyses confirm that late-stage compressional stresses caused accelerated subsidence of the central parts of the Pannonian Basin (Van Balen et al., (1999) whilst the Styrian Basin (Sachsenhofer et al., 1997), the Vienna and East Slovak Basins (Lankreijer et al., 1995), and the Transylvanian Basin (Ciulavu et al., 2002) were uplifted by several hundred meters starting in Late Mio–Pliocene times (Figure 70). The mode and degree of coupling of the Carpathians with their foreland controls the Pliocene to Quaternary deformation patterns in their hinterland, and particularly interesting, in the Transylvanian Basin (Ciulavu et al., 2002). During their evolution, the Western and Easter Carpathians were intermittently mechanically coupled with the strong European foreland lithosphere, as evidenced by coeval deformations in both the upper and lower plate (Krzywiec, 2001; Mat¸enco and Bertotti, 2000; Oszczypko, 2006). In terms of coupled deformation, the Carpathian Bend Zone had two distinct periods during its Tertiary evolution, that is, Early–Middle Miocene and Late Miocene–Quaternary. During the first period, the orogen was decoupled from its Moesian lower plate during the Middle Miocene (Badenian) as evidenced by contraction in the upper plate (e.g., Hippolyte et al., 1999) and extensional collapse of the western Moesian Platform (Ta˘ra˘poanca˘ et al., 2003). During the Late Miocene collisional coupling
Figure 80 (a) Location of the additional load added to the topography (for discussion see text). (b) Deflection predicted for actual topographic load plus the additional 500 m load. Underlined numbers give the EET for each lithospheric block. Deflection isolines in metres. (c) Cross-sections showing actual and predicted basin depths (for location see (b)). Solid line: observed basin depth; gray line: deflection for 500 m additional topographic load; dashed line: deflection for 800 m additional topographic load. (d) Deflection predicted (map view and cross section) for additional 500 m topographic load combined with an N–S directed intraplate compressional stress of 2 1012 N m1. Solid, gray, and dashed lines as in (c). Modified from Ta˘ rapoa˘ nca˘ M, Garcia-Castellanos D, Bertotti G, Mat¸enco L, Cloetingh S, and Dinu C (2004a) Role of 3-D distributions of load and lithospheric strength in orogenic arcs: polystage subsidence in the Carpathians foredeep. Earth and Planetary Science Letters 221: 163–180.
578
Tectonic Models for the Evolution of Sedimentary Basins
48° N 12 My 0
11 My
50
100 Km
12 My Tro tu
s fa
46° N
ult
Focsani depression
4 My 12 My
11 My
11 My
44° N 24° E Eroded column
26° E 5500 m
0
1000
2000
5000 m
4000 m 3000
28° E
4000
5000
1000 m
3000 m 6000
7000
8000
9000 m
Foredeep thickness Figure 82 Contours of amount of erosion (km) inferred from fission track analyses in the Romanian Carpathians and Apuseni Mts. (Sanders et al., 1999) and isopachs of sediment thickness in the foreland basin (km). Numbers in elliptic boxes indicate timing of erosion onset in Ma. Numbers in square boxes indicate timing of main subsidence. Note the pronounced lateral differences in uplift ages along the arc, while the main subsidence period is essentially coeval. Modified from Cloetingh S, Bada G, Mat¸enco L, Lankreijer A, Horva´th F, and Dinu C (2006). Thermo-mechanical modelling of the PannonianCarpathian system: Modes of tectonic deformation, lithospheric strength and vertical motions. In: Gee D and Stephenson R (eds.) Geological Society, London, Memoirs, 32: European Lithosphere Dynamics, pp. 207–221. London: Geological Society, London.
between the orogenic wedge and the foreland increased and persisted to the present (e.g., Bala et al., 2003). Fission track studies in the Romanian Carpathians demonstrate up to 5 km of erosion that migrated since 12 Ma systematically from their northwestern and southwestern parts towards the Bend area where uplift and erosion was initiated around 4 Ma ago (Sanders et al., 1999) (Figure 82). This region coincides with the actively deforming Vrancea zone that
is associated with considerable seismic activity at crustal levels and in the mantle (Figure 72). These findings can be related to the results of seismic tomography that highlight upwelling of hot mantle material under the Pannonian Basin and progressive detachment of the subducted lithospheric slab that is still ongoing in the Vrancea area (Wortel and Spakman, 2000; Wenzel et al., 2002). Moreover, in the internal parts of the Carpathians, magmatic activity related to slab detachment decreases
Tectonic Models for the Evolution of Sedimentary Basins
systematically in age from 16–14 Ma in their northern parts to 4–0 Ma in the Bend Zone (Nemcok et al., 1998). As such, it tracks the uplift history of the Carpathians that can be related to isostatic rebound of the lower plate upon slab detachment. These rapid differential motions along the rim of the Pannonian Basin and in the adjacent Carpathians had important implications for the sediment supply to depocenters, as well as for the hydrocarbon habitat (Dicea, 1996, Tari et al., 1997, Horva´th and Tari, 1999). In summary, results of forward basin modeling show that an increase in the level of compressional tectonic stress during Pliocene–Quaternary times can explain the first-order features of the observed pattern of accelerated subsidence in the center of the Pannonian Basin and uplift of basins in peripheral areas. Therefore, both observations (see Horva´th et al., 2006) and modeling results lead to the conclusion that compressional stresses can cause considerable differential vertical motions in the Pannonian–Carpathian back-arc basin–orogen system. In the context of basin inversion, the sources of compression were investigated by means of finite element modeling (Bada et al., 1998, 2001). Results suggest that the present stress state of the Pannonian– Carpathian system (Figure 81), and particularly of its western part, is controlled by the interplay between plate boundary and intraplate forces. The former include the counterclockwise rotational northward motion of the Adriatic microplate and its indentation into the Alpine-Dinaridic orogen, whereas intraplate buoyancy forces are associated with the elevated topography and related crustal thickness variation of the Alpine–Carpathian–Dinarides belt (Figure 65). Model predictions indicate that uplifted regions surrounding the Pannonian basin system can exert compression of about 40–60 MPa on its thinned lithosphere, comparable to values calculated for farfield tectonic stresses (Bada et al., 2001). The analysis of tectonic and gravitational stress sources permitted to estimate the magnitude of maximum horizontal compression, amounting to as much as 100 MPa. These significant compressional stresses are concentrated in the elastic core of the lithosphere, consistent with the ongoing structural inversion of the Pannonian Basin. Such high-level stresses are close to the integrated strength of the system, which may lead to whole lithospheric failure in the form of large-scale folding and related differential vertical motions, and intense brittle deformation in the form of seismo-active faulting.
579
6.11.7 The Iberia Microcontinent: Compressional Basins within the Africa–Europe Collision Zone By mid-Cretaceous times, the Iberian microcontinent, including Corsica-Sardinia, was separated from Europe owing to opening of the Bay of Biscay–Valaisan Ocean. By this time, Iberia was flanked to the southeast by the Alpine-Tethys that had opened during the Middle Jurassic, and to the west by the North Atlantic that had started to open during the Early Cretaceous (Stampfli et al., 1998, 2001, 2002). With the Late Cretaceous onset of Africa–Europe convergence (Rosenbaum et al., 2002), the Pyrenean and Betic–Balearic orogens evolved along the northern and southeastern margins of Iberia, respectively. Closure of the Pyrenean rift commenced during the Late Senonian and involved northward subduction of the continental Iberian lithosphere beneath Europe and southward subduction of the oceanic Bay of Biscay beneath Iberia. Evolution of the Pyrenean–Cantabrian orogen, in which crustal shortening persisted until endOligocene times, was accompanied by a gentle clockwise rotation of Iberia, causing reactivation of fault systems along its Atlantic margin (Munoz, 1992; Verge´s et al., 1998; Verge´s and Garcia-Senez, 2001). Northwestward subduction of the oceanic AlpineTethys beneath Iberia was initiated during the latest Cretaceous–Paleocene. During the Late Oligocene– Early Miocene, roll-back of the Alpine-Tethys slab commenced, giving rise to back-arc extension in the Gulf of Lions and the domain of the Valencia Trough, culminating in Burdigalian separation of Corsica-Sardinia from Iberia and the opening of the oceanic Ligurian–Provenc¸al Basin (Roca, 2001). At the same time, separation of the Kabylian block from the Balearic promontory resulted in opening of the oceanic Algerian Basin. Crustal shortening in the evolving Betic–Balearic Orogen persisted until midMiocene times, when the Kabylian–Alboran terrane collided with the African margin (Frizon de Lamotte et al., 2000). Minor late Miocene to Pleistocene extensional reactivation of the Valencia Trough (Roca, 2001) was accompanied by the extrusion of mantlederived partial melts in NE Iberia (Olot-Gerona-La Selva), on the Colombretes Island and in the Calatrava province (Wilson and Bianchini, 1999). During the late Eocene and Oligocene, the cratonic parts of Iberia were subjected to intraplate compressional stresses, originating at the Pyrenean
580
Tectonic Models for the Evolution of Sedimentary Basins
and Betic–Balearic collision zones, causing inversion of the Mesozoic rifted Catalan Coastal Ranges and the Central Iberian basins and up-thrusting of the Central Spanish basement block. During the Late Oligocene and Early Miocene extensional stresses controlled subsidence of the Valencia Trough that was associated with thermal thinning of the lithosphere (Banda and Santanach, 1992). Regarding the latter, it is noteworthy that the NE Atlantic mantle plume was activated during the Campanian– Maastrichtian, as evidenced by magmatic activity in southern Portugal (Hoernle et al., 1995; Tavares Martins, 1998). Similarly, Late Miocene to Pleistocene magmatic activity along the eastern margin of Iberia may be plume related. Seismic tomography indicates that the subducted Alpine-Tethys slab is still attached to the African lithosphere in the area of the Rif fold belt but that a slab window gradually opened eastward (Wortel and Spakman, 2000). This is compatible with Late Miocene slab detachment-related magmatic activity in the Maghrebian domain (Wilson and Bianchini, 1999). Subduction activity apparently ceased in the Pyrenean and Betic collision zones at end Oligocene and mid-Miocene times, respectively, with remnant deep seismicity being restricted to the southeastern margin of Iberia (Blanco and Spakman, 1993). Nevertheless, continued convergence of Africa with
Europe is held responsible for still ongoing intraplate deformation of Iberia that is manifested, amongst others, by earthquake activity concentrated on the Pyrenees, the Betic Orogen, the Central Iberian and Catalan Coastal Ranges, and the Atlantic coastal domain. Controlling stresses are related to collisional coupling between the African and European plates and the intervening Iberian microplate and to Atlantic ridge-push forces (Figures 83 and 84). Figure 85 shows the main Neogene structural features of Iberia and the magnitude of PlioPleistocene uplift of its different parts. Although mechanisms underlying the observed vertical motions are not fully resolved, processes such as slab detachment and/or lithospheric delamination appear to play a minor role (Janssen et al., 1993; Docherty and Banda, 1995; Seber et al., 1996), whereas mantle plume-related thermal thinning of the lithosphere cannot be excluded, for example, NE Iberia. This is compatible with available geophysical data on the upper mantle structure (Blanco and Spakman, 1993; Seber et al., 1996). On the other hand, along the eastern margin of Iberia, rift shoulder uplift, related to the Late Neogene extensional reactivation of the Valencia Trough (Janssen et al., 1993) and the Alboran Sea (Docherty and Banda, 1995; Cloetingh et al., 1992) may be a contributing factor. However, recent analyses of the stress field,
Pyrenean or ‘Iberian’ phase Paleogene–Early Neogene N–S to NE–SW compression
Intraplate record of stress field (changes): Ridge push from the Atlantic (M. -L. Cretaceous till present day)
Polyphase fault reactivation under different stress fields Erosion/sedimentation/tectonics Tertiary basin formation and deformation Significant intraplate deformation
Valencian phase E.Miocene–Pliocene WNW-ESE extension
Vertical motions
Betic phase M. Miocene–recent NNW-SSE compression
Figure 83 Plate-tectonic setting of Iberia and timing of Alpine to recent plate boundary reorganizations, their impact on intraplate deformation of the Iberian microcontinent and related basin (de)formation processes within Iberian and along its Atlantic and West Mediterranean margins. Modified from Cloetingh S, Burov E, Beekman F, Andeweg B, Andriessen PAM, GarciaCastellanos D, de Vince G, and Vegas R (2002) Lithospheric folding in Iberia. Tectonics 21(5): 1041 (doi:10.1029/2001TC901031).
Tectonic Models for the Evolution of Sedimentary Basins
10° W
5° W
0°
581
5° E
45° N
40° N
35° N
-
4 mm yr–1 4 mm yr–1
5 mm yr–1
Figure 84 Present-day stress trajectories (thin blue lines) in Iberia and the western Mediterranean, based on fault slip data, borehole breakout data, and focal mechanisms (black and white arrows). Thin redlines: major tectonic structures and lineaments. The spatial orientation (‘fanning’) of the first-order present-day stress regime suggests a strong control by plate boundary processes (Africa/Eurasia collision, Atlantic ridge push) and large-scale weakness zones, such as formerly active plate boundaries (Pyrenees, Betics). Modified from Andeweg B, De Vincente G, Cloetingh S, Giner J, and Mun˜oz Martin A (1999) Local stress fields and intraplate deformation of Iberia: variations in spatial and temporal interplay of regional stress sources. Tectonophysics 305: 153–164.
topography evolution, and gravity data indicate that folding of the continental lithosphere of Iberia played an important, if not the dominant, role in its Late Neogene and Quaternary deformation. Indeed, recent studies provide increasing evidence for a strong contribution of crustal-scale kinematics on the record of vertical motions of the Iberian Peninsula (e.g., Friend and Dabrio, 1996; CasasSainz et al., 2000). In the following, novel concepts are presented on the control of lithospheric deformation on the development of drainage patterns, crustal topography, and intracontinental basins that are based on an integrated approach, linking structural field studies, thermogeochronology and modeling of lithospheric and surface processes (see Cloetingh et al., 2002, 2005). 6.11.7.1
Constraints on Vertical Motions
Palaeogeographic constraints indicate that during Late Cretaceous times much of Iberia was located
close to sea level with the West Iberian Massif and the area of the future Ebro Basin forming low-relief highs (Stampfli et al., 2001, 2002). However, today large parts of cratonic Iberia are located at elevations of 750–1000 m, with areas affected by Late Eocene to Oligocene intraplate compression forming up to 1500 m high mountain chains (Figure 86). In order to determine the timing of uplift of cratonic Iberia to its present elevation, apatite fission track studies were carried out on its Mediterranean (Stapel et al., 1996) and Atlantic margins (Stapel, 1999), on the Spanish Central System (De Bruijne and Andriessen, 2000, 2002), on the Betics (Zeck et al., 1992), and on the Pyrenees (Fitzgerald et al., 1999). These studies evidence a rapid post-Miocene cooling phase (uplift and erosion) for topographically high areas along the western and eastern margin of Iberia (Figure 87), as well as for the Spanish Central System (Figure 88). For the latter, results indicate that an initial Late Eocene to Early Oligocene uplift phase, associated with compressional deformations,
582
Tectonic Models for the Evolution of Sedimentary Basins
France 50–60
Cantabria
Galicia
~500 Pyrenees
Eb
Duero basin
ro b
>150
asi
n
~500 Madrid basin
iv
lqu
da
a Gu
2
tro
h
40° N
len
Va
150–200
125–1000
sin
a ir b
42° N
ug
cia
~1000
500–700
44° N
1
38° N
Betic cordillera
ran
o Alb
sea 36° N Northern Africa
12° W
8° W
4° W
0°
4° E
Figure 85 Tectonic map of Iberia, showing main structural features and sedimentary basins, its northern [1] and southern [2] plate boundaries, and Pliocene–Quaternary uplift patterns (numbers in boxes give estimated magnitude of uplift in meters). Modified from Cloetingh S, Burov E, Beekman F, Andeweg B, Andriessen PAM, Garcia-Castellanos D, de Vince G, and Vegas R (2002). Lithospheric folding in Iberia. Tectonics 21(5): 1041 (doi:10.1029/2001TC901031).
44° C
B
42° tem
A Sp
sys
is
an
basin anian Lusit
40°
al
ntr
e hc
da rra Se trela s E
38° Sierra Morena
36° –10°
–8° 0
Sorbas Tabernas basins Hinojar
–6° 100
200
–4° 300
–2°
0°
2°
4°
400 500 750 1000 1250 1500 3000 Mean elevation (m)
Figure 86 Digital elevation model of Iberia (data from GTOPO30). Yellow boxes refer to location of sites sampled for quantitative subsidence analyses and fission-track analyses for exhumation quantification (see also Figures 87 and 88). A, B, and C mark the location of profiles given in Figure 90. Modified from Cloetingh S, Burov E, Beekman F, Andeweg B, Andriessen PAM, GarciaCastellanos D, de Vince G, and Vegas R (2002) Lithospheric folding in Iberia. Tectonics 21(5): 1041 (doi:10.1029/2001TC901031).
Tectonic Models for the Evolution of Sedimentary Basins
240
180
Time (Ma) 120
60
0 0
No. tracks (%)
20 40 60
30 20
PAZ
10
110
0 0.5 5.5 10.5 15.5 Length (μm)
Serr da Estrela SE3: 875 m - West 66.2 ± 7.4 Ma 12.6 ± 2.0 μm
T (°C)
Lusitanian Basin Lus41 176 ± 12 Ma 11.6 ± 1.6 μm
583
240
180
Time (Ma) 120
60
0 0
30
60
T (°C)
No. tracks (%)
20 40 PAZ
20 10
110
0 0.5 5.5 10.5 15.5 Length (μm)
Sierra Morena GS80 184 ± 13 Ma 12.0 ± 1.4 μm
240
Time (Ma) 180 120
60
0 0
30
60 PAZ
20
T (°C)
No. tracks (%)
20 40
10 110
0 0.5 5.5 10.5 15.5 Length (μm)
Figure 87 Constraints on timing and magnitude of Late Neogene vertical motions derived from quantitative subsidence and fission-track analyses for samples from western Iberia (LB Lusitanian Basin, SE Serra da Estrela) and from southern–western Iberia (SM Sierra Morena). Plotted is a range of cooling histories that fit the fission-track age and track length distribution. Also shown are Monte-Carlo boxes based on information from fission-track data and geologic observations. The gray-shaded area is the partial annealing zone (PAZ). Modified from Stapel G (1999) The Nature of Isostasy in Western Iberia. PhD thesis, Vrije Universiteit, Amsterdam, 148p.
was followed by significant cooling during Middle Miocene times (15 Ma) and a pronounced cooling acceleration from the Early Pliocene (5 Ma) onward (De Bruijne and Andriessen, 2000). In the eastern part of the Spanish Central System, the Sierra de Guadarrama, Pliocene uplift and erosion amounted to up to 6 km (Ter Voorde et al, 2004; De Bruijne and Andriessen, 2000). Results of precision levelling and VLBI laser ranging (Rutigliano et al., 2000) indicate
present vertical uplift rates of about 1 mm yr1 for the Spanish Central System. For the Betics, thermal modeling of fission track data supports a scenario of enhanced uplift and erosion from Pliocene times onward until the present (Ter Voorde et al., 2004). This is further supported by the results of backstripping analyses and forward modeling of the sedimentary record of Neogene pull-apart basins that are superimposed on the Betic
584
Tectonic Models for the Evolution of Sedimentary Basins
0
45 ± 5 Ma h 2105 m MTL 12.4 μm σ 1.5 μm N 150
T (°C)
SCS 19 60
110 80 0
40 t (Ma)
12
16
33 ± 5 Ma h 900 m MTL 12.3 μm σ 1.7 μm N 34*
SCS 18
T (°C)
8
0
60
110 160
120
80 t (Ma)
40
T (°C)
0
T (°C)
8
0
12
16
66 ± 8 Ma h 970 m MTL 11.5 μm σ 1.8 μm N 230
SCS 12
60
110 80
40 t (Ma)
0
8
T (°C)
T (°C)
0
12
16
11 ± 1 Ma h 1080 m MTL 11.6 μm σ 2.2 μm N 43
SCS 11
60
110 80
40 t (Ma)
0
8
12
16
208 ± 21 Ma h 800 m MTL 12.8 μm σ 1.6 μm N 150
SCS 55 T (°C)
0
60
110 280
240
200
160
120 t (Ma)
80
40
0
8
12
16
Figure 88 Constraints on timing and magnitude of Late Neogene vertical motions derived from quantitative subsidence and fission-track analyses. .Fission-track ages, track length distributions, and thermal histories giving the best-fit data for five samples from the Spanish Central System (SCS). Thermal histories were obtained with the Monte Trax program and are shown as black lines within the 100-best-fit envelopes (gray shaded). The modeled track length distributions are projected on the measured track length distribution. Modified from De Bruijne and Andriessen (2000, 2002).
Cordillera (Figure 89), revealing that they were uplifted and eroded during the Pliocene– Quaternary by some 500 m (Cloetingh et al., 2002; Janssen et al., 1993; Docherty and Banda, 1995). Moreover, on the base of Late Miocene marine sediments that were uplifted by more than 1200 m in
intramontane basins of the Betic Cordillera, PlioPleistocene shortening and related uplift is documented (Andeweg and Cloetingh, 2001). Initial results of GPS surveys point toward a consistently northwest directed horizontal motion of Iberia at rates of about 5 mm yr1 (Fernandes et al.,
Tectonic Models for the Evolution of Sedimentary Basins
6.11.7.2 Present-Day Stress Regime and Topography
Sorbas Basin
Subsidence (km)
0.0 0.5
β = 1; δ = 1.10
1.0
β = 1; δ = 1.15
1.5
β = 1; δ = 1.20
2.0 10
8
6 4 Age (Ma)
2
0
Tabernas Basin
Subsidence (km)
0.0 0.5
β = 1; δ = 1.1
1.0
β = 1; δ = 1.2
1.5
β = 1; δ = 1.3
2.0 10
8
6 4 Age (Ma)
2
0
2
0
Hinojar Basin
Subsidence (km)
0.0 β = 1; δ = 1.10
0.5
β = 1; δ = 1.15
1.0
β = 1; δ = 1.20
1.5 2.0 10
8
6 4 Age (Ma)
585
Figure 89 Tectonic subsidence curves (bold lines) for three Late Neogene pull-apart basins in the internal zone of the Betic Cordilleras (SE Spain). Thin lines are theoretical subsidence curves calculated for differential stretching of the crust ( 6¼ 1) and in the absence of mantle stretching ( ¼ 1). Note that all basins show accelerated Late Plio-Quaternary uplift that strikingly deviates from prediction of thermal basin subsidence models. Modified from Cloetingh S, Ziegler PA, Beekman F, Andriessen PAM, Mat¸enco L, Bada G, GarciaCastellanos D, Hardebol N, De´zes P, and Sokoutis D (2005) Lithospheric memory, state of stress and rheology: Neotectonic controls on Europe’s intraplate continental topography. Quaternary Science Reviews 24: 241–304.
2000). This raises the question whether the observed Plio-Quaternary vertical motions are related to folding of the Iberian lithosphere (Andeweg and Cloetingh, 2001), thus differentially amplifying its topography.
The present-day stress map of Iberia, given in Figure 84, is based on borehole break-out data, earthquake focal mechanisms and microtectonic stress indicators (Zoback, 1992; Ribeiro et al., 1996; De Vicente et al., 1996; Andeweg et al., 1999). Iberia is characterized by a consistent horizontal compressional stress field that is dominated by northwestdirected stress trajectories fanning out in Portugal into a more westerly direction and in northeastern Spain into a northerly direction. This stress field reflects a combination of forces related to collisional coupling between Africa, Iberia and continental Europe, and Atlantic ridge-push. Northwestward movement of Africa at rates of 4–5 mm yr1 is apparently compensated by crustal shortening in the seismically active Maghrebian, Betic, and Pyrenean zones, as well as by deformation of cratonic Iberia. The topography of cratonic Iberia is characterized by a succession of roughly NE–SW trending highs and intervening lows (Figure 90). From northwest to southeast, these are the river Min˜o, the mountains of Cantabria/Leon, the Duero-Douro river basin, the Spanish Central System, the Tajo river basin, the Toledo Mountains, the river Guadiana, the Sierra Morena, and the Guadalquivir Basin. These topographic highs and lows trend normal to the presentday intraplate compressional stress trajectories, and essentially run parallel to similar trending Bouguer gravity anomalies (Cloetingh et al., 2002). The magnitude of Plio-Pleistocene vertical motions and the results of precision leveling suggest that processes controlling topography development are still ongoing and exert a first-order control on the present topographic configuration of Iberia. The observed Bouguer gravity anomalies reflect longwavelength undulations of deep intralithospheric density interfaces, such as the crust–mantle boundary (Cloetingh and Burov, 1996), and thus mirror deformation of the entire lithosphere. Whereas a correlation between long-wavelength Bouguer gravity and topographic anomalies (Figure 90) speaks for coupled crustal and lithospheric mantle folding, the lack of such a correlation speaks for their decoupled folding (Cloetingh et al., 1999). As so far there is no conclusive evidence from seismic tomography for a thermally perturbed upper mantle underlying most of Iberia (Wortel and Spakman, 2000), apart from its NE region (Spakman and Wortel, 2004; Sibuet et al., 2004), the question remains open whether its high
586
Tectonic Models for the Evolution of Sedimentary Basins
Profile C Cantabrian mnts
Elevation (m)
NW 1500
50
Central system
SE Toledo mnts
Duero river catchment
1000
0 –50
500 Guadiana river
Tajo river
0 700
600
500 400 300 200 Distance along profile (km)
100
–100
Δg (mGal)
2000
–150
0
Profile B NW
SE
Central system
Toledo mnts
Leon mnts
1000
Sierra morena
75 50 25 0
500
–25
0
Mino river
700
600
Duero river
–50
Guadiana river
Tajo river
500 400 300 200 Distance along profile (km)
100
Δg (mGal)
Elevation (m)
1500
–75 0
Profile A NW
Sierra morena
Sierra estrela
750
SE
Toledo mnts
500
0 Tajo river
0 700
600
50 25
Mondego river
250
75
–25 –50
Guadiana river
500 400 300 200 Distance along profile (km)
100
Δg (mGal)
Elevation (m)
1000
–75 0
Figure 90 Topographic profiles across Iberia (black curves) compared to Bouguer gravity anomalies (red curves). For location see Figure 86. Modified from Cloetingh S, Burov E, Beekman F, Andeweg B, Andriessen PAM, Garcia-Castellanos D, de Vicente G, and Vegas R (2002) Lithospheric folding in Iberia. Tectonics 21(5): 1041(doi:10.1029/2001TC901031).
average elevation is caused by a large-scale asthenospheric thermal anomaly, giving rise to a gravity signal with a wavelength of over 500 km. 6.11.7.3 Pattern
Lithospheric Folding and Drainage
For central Iberia, thermal modeling of fission track data favours a Plio-Pleistocene phase of uplift and erosion rather than a Miocene uplift phase followed by erosion (Ter Voorde et al., 2004). In response to lithospheric folding, elevated areas will be subjected to erosion whereas sediments will accumulate in topographic lows. Whereas erosional unroofing of structural highs causes their isostatic uplift, sedimentary loading causes flexural subsidence of structural lows. However, the drainage pattern of Iberia removes
all erosion products and deposits them on the Atlantic and Mediterranean continental shelves. Numerical modeling of the evolution of drainage networks (Garcia-Castellanos, 2002; Garcia-Castellanos et al., 2003) shows that surface transport processes can effectively enhance the tectonically induced large-scale continental topography. Despite the intrinsic nonlinear nature of drainage networks, moderate vertical movements appear also to be able to organize drainage patterns in relatively flat areas where drainage is not well-organized or incised (Garcia-Castellanos, 2002). Under the temperate climatic conditions, which characterized Iberia during most of the Cenozoic, rivers accounted for most of the surface mass transport. River catchments have typically strong 3-D asymmetries since their drainage is hierarchically organized to collect the waters and sediment from a
Tectonic Models for the Evolution of Sedimentary Basins
widespread area and channel them into a single outlet. This implies that whereas erosion affects large headwater catchment areas, deposition is localized along the lower riverbeds, around the river mouths in the Atlantic Ocean and Mediterranean and in possible intermediate lakes (Figure 86). Modeling also permits to analyze possible feedback mechanisms between this asymmetric surface transport and lithospheric folding (Cloetingh et al., 2002, 2005). To further evaluate the interplay between surface transport and lithospheric folding, fluvial transport can be simulated via a drainage network in which runoff water flows along the maximum slope with a sediment transport capacity proportional to the slope and water discharge (Garcia-Castellanos et al., 2003). Implementation of this approach, which intrinsically predicts planform hierarchical organization of drainage networks, is based on the relationships established by Beaumont et al., (1992) that were adopted in several subsequent studies (e.g., Kooi and Beaumont, 1996; Braun and Sambridge, 1997; Van der Beek and Braun, 1998). Analyses focused on the firstorder features of the interplay between fluvial transport and folding, rather than on the properties of fluvial transport. Furthermore, the applied model incorporated sediment deposition in topographic minima (lakes) and in the Atlantic Ocean, thus allowing for modeling of closed transport systems in which virtually all eroded material is deposited within the model (for details see Garcia-Castellanos, 2002). Lithospheric folding was calculated as the response of a thin homogeneous viscoelastic 2-D (plan form) thin-plate to tectonic loading (horizontal load) and surface mass redistribution (vertical load). The equation governing deformation of this plate was derived by applying the principle of correspondence between elasticity and viscoelasticity (Lambeck, 1983) to the equivalent elastic equation (e.g., Van Wees and Cloetingh, 1994). The 3D response of this plate to lithosphere shortening was calculated, adopting an EET of 30 km. This value corresponds to a mechanically coupled crust and lithospheric mantle with a thermotectonic age of 350 Ma that is consistent with results of a flexural analysis on the Ebro Basin of northeastern Spain (Gaspar-Escribano et al., 2001). A viscous relaxation time of 1.2 My was adopted from a recent flexural analysis of the Guadalquivir Basin (southern margin of the Iberian Massif) (GarciaCastellanos et al., 2002). The initial configuration of the synthetic models is a flat, square continent elevated to 400 m above sea level with a random perturbation between 10 and þ10 m.
587
Figure 91(a) shows the drainage network that developed after a period of 12 My of N–S directed horizontal compression (see Appendix 2 for a detailed description of the modeling approach). The main axes of topographic linear highs and lows predicted by the plan-view synthetic models are perpendicular to the main axes of shortening. This corresponds with the actually observed relationship between topography and stress field, with large-scale topographic linear highs and lows perpendicular to the main axis of present-day intraplate compression in Iberia (Figure 84). In this figure the smooth patterns of vertical motions indicate that river sediment transport plays a key role in defining the location of subsequent folding vertical motions. Note that these models illustrate the conceptual link between intraplate compression, lithospheric folding, and surface processes; an NW–SE oriented main axis of shortening (Figure 84) would rotate the main axis of drainage and erosion and sedimentation into an NE–SW orientation, more closely corresponding to the topography of Iberia (Figure 86). The drainage pattern is clearly controlled by the 350 km wavelength of lithospheric folds. A reduction of the EET to 18 km, closer to the values found in the Guadalquivir Basin (Garcia-Castellanos et al., 2002), reduces this characteristic wavelength to 220 km. During the initial stages of compression, the peneplain geometry limits erosion and deposition to the vicinity of the shoreline, whereas the interior of the continent remains almost unperturbed. Interestingly, vertical isostatic movements associated with this initial mass transport near the shoreline play a role as perturbations that trigger folding. Consequently, maximum uplift occurs along the N and S shores, as they trend normal to the compression axis. Inversely, the organization of drainage system has also important effects on vertical motions. This is illustrated by a comparison of Figures 91(a) and 91(b) with an identical model in which transport is diffusive (Figure 91(c)), adopting a value of ke ¼ 33.0 m2 yr1. Although the chosen diffusive constant to produce a total transport of sediments is similar to the previous model (71.9 103 km3), the spatial organization of the drainage pattern strongly influenced the spatial mass redistribution. Deposition is localized near the main river mouths and in phase with vertical motions, since drainage is forced along the E–W trending folding-induced depressions (Figure 91(a)). This accentuates the folding pattern off the W shore, at x<420 km. By contrast, a diffusive approach to surface transport does not introduce
588
Tectonic Models for the Evolution of Sedimentary Basins
relevant asymmetries in the folding pattern (Figure 91(c)), apart from some overprinting along the western, Atlantic margin of Iberia. 6.11.7.4 Interplay between Tectonics, Climate, and Fluvial Transport during the Cenozoic Evolution of the Ebro Basin (NE Iberia) Endorheic drainage basins, also referred to as landlocked, internally drained or closed drainage basins, are essential for understanding the evolution of sedimentary basins because they do not fit the notion that
erosion products of orogens are par force carried to the oceans. Particularly in intraorogenic settings, such as the Altiplano or the Tibetan Plateau, endorheic basins can trap important sediment volumes at high elevations above sea level (e.g., Sobel et al., in press). Endorheic basins occupy, 20% of the Earth’s land surface but collect only about 2% of the global river runoff, reflecting that they develop mostly under arid conditions. The Ebro Basin is a well-documented example of a long-lasting intraorogenic endorheic basin, the deposits of which are presently located at more than 1000 m above sea level. Because the sedimentary fill of the Ebro Basin has been heavily incised and exposed
(a) 400
400
y (km)
U 200
200
0
0
–200
–200
D
D
U
D
D U
–400
–400 Topography + drainage
–600
–400
–200
0
Erosion rate (shading) + subsidence rate (m My–1)
200
–600
–400
–200
0
200
(b) 400
400
D
y (km)
U 200
200
0
0
–200
–200
–400
–400
D
D
Erosion rate (shading) + subsidence rate (m My–1)
Topography + drainage
–600
–400
Figure 91 (Continued )
–200
0
200
–600
–400
–200
0
200
Tectonic Models for the Evolution of Sedimentary Basins
589
y (km)
(c) 400
400
200
200
0
0
–200
–200
–400
–400 Erosion rate (shading) + subsidence rate (m My–1)
Topography + drainage
–600
–400
–200 x (km)
0
200
–600
–400
–200 x (km)
0
200
Total erosion/sedimentation (m) –1000
–300 –30
0
30
300
1000
Figure 91 Synthetic models (map view) for the interplay between lithospheric folding, surface erosion, and sedimentation. An initially flat, square continent with dimensions representative for Iberia was adopted. The thickness of rivers is plotted proportional to their water discharge. Contours indicate vertical motions (m) induced by N–S directed compression during 12 My and the resulting isostatic response of the model to surface mass transport. Upper panels show topography and drainage networks. Lower panels show cumulative erosion/deposition (shading) and vertical motions (contours in m). For details and model parameter specifications, see Appendix 2. (a) Model incorporating superimposed lithospheric folding and fluvial transport through a drainage network adopting an EET of 30 km. (b) Same as (a) but adopting an EET of 18 km. (c) Same as (b) but adopting a diffusive model for mass redistribution. Modified from Cloetingh S, Ziegler PA, Beekman F, Andriessen PAM, Mat¸enco L, Bada G, Garcia-Castellanos D, De`zes P, and Sokoutis D (2005) Lithospheric memory, state of stress and rheology: Neotectonic controls on Europe’s intraplate continental topography. Quaternary Science Reviews 24: 241–304.
owing to late-stage opening of its drainage toward the Mediterranean (Figure 92), it serves as a natural laboratory to study the interplay between tectonics, climate, and sediment transport (e.g., Arenas et al., 2001). The Ebro Basin is rimmed by three Alpine mountain ranges (Figure 92), namely by the collisional Pyrenees to the N and the Catalan Coastal Ranges and the Iberian Range to the SE and SW, respectively, that represent inverted Mesozoic rifted basins. Development of the Ebro Basin began during the Paleocene in response to its flexural subsidence in the prowedge foreland of the evolving Pyrenees (Verge´s et al., 1998). Inversion of the Catalan Coastal Ranges and the Iberian Range commenced during the early Middle Eocene and persisted until the Late Oligocene (Roca et al., 1999). Uplift of these ranges resulted in earliest Late Eocene closure of connections between oceanic domains and the Ebro Basin. With this, its endorheic stage commenced that lasted through Oligocene and most of Miocene times (Riba et al., 1983). During this
stage, syn- and postorogenic conglomerate fans prograded from the Iberian Range, the Catalan Coastal Ranges and the rapidly rising Pyrenees into the tectonically silled Ebro Basin, burying the frontal Pyrenean thrust elements, and giving way to distal lacustrine sediments (Coney et al., 1996). During the latest Oligocene and Early Miocene rifting phase of the Valencia Trough, tensional reactivation of reverse faults of the Catalan Coastal Ranges resulted in their gradual breakdown. By late Middle Miocene times the deltaic complex of the Castello´n Group (Serravallian–early Messinian) began to prograde into the Valencia Trough, reflecting the gradual development of the modern Ebro drainage system (Bartrina et al., 1992; Banda and Santanach, 1992; Verge´s and Sa`bat, 1999; Roca, 2001). The internal drainage of the Ebro Basin was open to the Mediterranean in the course of the Late Miocene in response to progressive incision and backcutting of the evolving river Ebro, exposing
590
Tectonic Models for the Evolution of Sedimentary Basins
W Europe 4°
3°
2°
1°
0°
1°
Atlantic Ocean Iberia
Mediterranean
Aquitanian Basin
San Sebastian
Africa Western pyrenees 43°
Centr al and Jaca
Easte
rn Py
renee
s
PL
Ebro
Ib er
NL
San Caprasio
PL
ia
PL PL
n n ra ge
Lacustrine teriary Ecocene-present Ecocene-Oligocene Late cretaceous-paleocene Early cretaceous Paleozoic basement
Thrust Blind thrust Normal fault
Llobregat
Basin
tal oa s nC a l ra Ca
ng Ra
Ter
42°
es Barcelona
València Trough Mediterranean Sea 0
50
41°
100 km
Figure 92 Geological map of the Ebro Basin and the bounding Pyrenees, Catalan Coastal Range and Iberian Range, showing the present river network and the approximate extent of Neogene lacustrine deposits (PL Paleogene, NL, Neogene). Modified from Garcia-Castellanos D, Verge´s J, Gaspar-Escribano JM, and Cloetingh S (2003) Interplay between tectonics, climate and fluvial transport during the Cenozoic evolution of the Ebro Basin (NE Iberia). Journal of Geophysical Research 108: 2347.
impressive compressional structures in syn-tectonic Eocene sediments at the boundary between the Catalan Coastal Ranges and the Ebro Basin. With this, clastic influx into the Valencia Trough increased, as evidenced by the rapid progradation of the Castello´n Group delta and the modern postMessinian Ebro Group delta (Dan˜obeitia et al., 1990; Ziegler, 1988). 6.11.7.4.1 Ebro Basin evolution: a modeling approach
In order to gain a better understanding of processes underlying the drainage evolution of the Ebro Basin, a numerical model was applied that integrates a surface transport model with quantitative approaches to the tectonic processes of thrusting and isostasy (for details see Garcia-Castellanos et al., 2003). Quantitative studies on the interplay between lithosphere dynamics and drainage networks in sedimentary basins are scarce. Since the early models of foreland basin development in the 1980s, the flexural response of the lithosphere to orogenic thrust loading
has been accepted as the key process generating accommodation space and sediment accumulation in front of orogens (e.g., Beaumont, 1981; Flemings and Jordan, 1989; Sinclair and Allen, 1992; Ford et al., 1999; Garcia-Castellanos et al., 2002). However, these studies used simplistic approaches to surface mass transport, neglecting the dynamics and 3-D nature of fluvial networks. Later numerical experiments showed that the spatial and temporal distribution of sediment facies is strongly influenced by the 3-D character of fluvial transport (Johnson and Beaumont, 1995), and that the coupled tectonic–fluvial network system may control the evolution of sediment supply from orogens (Tucker and Slingerland, 1996). In turn, drainage networks in foreland basins can be under certain conditions controlled by the flexural behaviour of the lithosphere (Burbank, 1992; Garcia-Castellanos, 2002). The numerical model applied to study the drainage and sedimentary evolution of the Ebro Basin was based on the approach developed by Garcia-Castellanos (2002) and incorporates planform numerical solutions
Tectonic Models for the Evolution of Sedimentary Basins
(a) Atlantic Ocean
Up
Iberian range
pe Lo r cru s we rc t Lith rus osp t he ric ma ntle As the no sp he re
(b)
es Pyrene
Basin
g Thrustin
Lithospheric flexure
)
(Fluid
Precipitation Evaporation
Py Sediment infill
Ibe
heic Endorsed) (clo e lak
ria
nr an
ge
ren ee
s Catalan coastal range
Lake level change
ion
ent eros
Escarpm
Flexural uplift
ia Valenc
Crustal extension
trough
Figure 93 (a) Lithospheric-scale block diagram showing tectonic setting of the flexural Ebro Basin in the pro-wedge foreland of the Pyrenees and the Iberian Range bounding it to the SW. White arrows indicate tectonic transport direction. (b) Endorheic stage of the Ebro Basin prior to capture of its closed lake in response to erosional breakdown of the Catalan Coastal Range. Modified from Garcia-Castellanos et al., (2003).
Tectonic deformation n tio ma for
ing
de of on
Lo ad
ati
sis
Lo
ne
ca
ge
liz
n tio
ro
no
Flexural isostasy
a er
lo
tro
on
Ebro basin evolution
en
lc
na
fg lie Re
io os
Er
to the following processes (Figures 93 and 94): (1) tectonic deformation is mainly driven by upper crustal thrust stacking and normal faulting, (2) surface transport is driven by the fluvial network, and (3) mass redistribution resulting from (1) and (2) is compensated by regional isostasy (lithospheric flexure). The interplay between lithospheric and surface processes is explicit in two ways. First, surface processes determine the way mass is redistributed in space and time, thus inducing isostatic vertical movements. Second, crustal and lithospheric-scale tectonics control and organize the nonlinear, chaotic nature of drainage networks. In zones affected by folding and thrusting, the long-term drainage evolution is controlled by the deformation kinematics (e.g., Tucker and Slingerland, 1996; Ku¨hni and Pfiffner, 2001), but the second-order role of lithospheric flexure can also become relevant in undeformed areas such as the distal margin of a foreland basin (Burbank, 1992; Garcia-Castellanos, 2002) or during posttectonic erosion-induced isostatic rebound of an escarpment (Tucker and Slingerland, 1994). Tectonic deformation was simulated with units (or blocks) moving relative to the foreland. These units preserve their vertical thickness during movement (vertical shear approach) simulating a noninstantaneous tectonic deformation. As modeling did not intend to reproduce the internal geometry of the orogenic wedge and the kinematic details of its evolution (Beaumont et al., 2000), and for the sake of simplicity, crustal shortening in each mountain range was represented in the model by the least number of blocks necessary and at constant shortening rates. Isostatic subsidence and uplift were calculated assuming that the lithosphere behaves as a 2-D thin elastic plate (e.g., Van Wees and Cloetingh, 1994) that is loaded by the large-scale mass redistribution and rests on a fluid asthenosphere. Erosion and sedimentation were calculated (see Appendix 2) by defining the drainage network on top of a time-dependent topography resulting from tectonic deformation and isostasy. To account for the endorheic stage of the Ebro Basin, and as evaporation can eliminate excess water in a lake reducing its level below the outlet, thus causing closure of the basin, evaporation was incorporated in the model as an improvement on the algorithm of Garcia-Castellanos (2002). (See Figure 95.) The amount of lake evaporation was calculated by multiplying the lake surface with a constant evaporation rate and by subtracting results from the lake discharge.
591
Flexural effects on river network
Climate and fluvial transport
Mass redistribution Figure 94 Interaction of processes controlling drainage evolution and large-scale sediment transport in the Ebro Basin and surrounding mountain ranges. Solid arrows: interactions addressed in the text. Stippled arrows: interactions not explicitly addressed in models presented. Modified from Garcia-Castellanos D, Verge´s J, Gaspar-Escribano JM, and Cloetingh S (2003). Interplay between tectonics, climate and fluvial transport during the Cenozoic evolution of the Ebro Basin (NE Iberia). Journal of Geophysical Research 108: 2347.
592
Tectonic Models for the Evolution of Sedimentary Basins
(a)
6.11.7.4.1.(i) Endorheic period: interplay between tectonics and climate Modeling
Rain Evaporation
Input discharge Input seds. Eros./sed.
Output discharge, Qw Output seds, q
(b)
Lake evaporation
Water and sed. discharge
Figure 95 Cartoon illustrating the numerical surface processes model. (a) Water and sediment input and output at each cell of the model. (b) Rivers follow the maximum slope of the discretised topography, taking into account evaporation at lakes forming in local topographic minima. Modified from Garcia-Castellanos D, Verge´s J, GasparEscribano JM, and Cloetingh S (2003) Interplay between tectonics, climate and fluvial transport during the Cenozoic evolution of the Ebro Basin (NE Iberia). Journal of Geophysical Research 108: 2347.
Owing to the coupled response of these processes, the numerical model simulated the 3-D evolution of the geometry of the orogen/basin system, including topography, drainage networks, sediment horizons, erosion patterns, and vertical isostatic movements. Results demonstrate the importance of the interplay between climatic, fluvial, and tectonic processes in shaping long-term landscape evolution and mass transport from mountain ranges to sedimentary basins. Most coupled tectonic-landscape evolution models have disregarded the role of lakes and their climate-controlled hydrographic balance (e.g., their endorheic/closed or exorheic/open character) in the evolution of continental topography and surface sediment transport. This aspect is of central importance for the Ebro Basin, in which the presence of a closed Oligocene–Miocene lake appears to have played a major role in the landscape evolution and sediment budget of the south Pyrenean area.
shows that tectonic closure of the Ebro Basin transformed it into a trap for debris derived from the surrounding mountain ranges with increasing sedimentation rates giving it its singular architecture (Garcia-Castellanos et al., 2003). Models validate the hypothesis formulated by Coney et al., (1996) that thick syn- and posttectonic conglomerates buried the frontal, presently outcropping Pyrenean thrusts during the endorheic stage of basin evolution. Vitrinite reflectance studies in the eastern Ebro Basin suggest burial depths of the presently outcropping sediments ranging between 2750 250 m near the Pyrenean front and 950 150 m near the Catalan Coastal Ranges (Clavell, 1992; Waltham et al., 2000). Model predictions agree with these observations, indicating that the removed sediment cover was up to 1500 m thick and that the maximum basin fill prior to drainage opening amounted to about 140 000 km3. Moreover, the models explain the presence of a single lake body with a very asymmetric sediment distribution, both in facies and thickness, as inferred by geological field studies (Arenas and Pardo, 1999). Modeling provides a quantitative framework for the understanding of these asymmetries that resulted from differences in the tectonic evolution of the surrounding mountain ranges: The greater tectonic shortening and catchment areas in the Pyrenees and the Iberian Chain as compared to the Catalan Coastal Ranges underlay the increased sediment supply to the basin from the N and SW, slowly shifting the lake and depocentre toward the SE. 6.11.7.4.1.(ii) Opening and incision of the Ebro Basin: interplay of lithospheric and surface processes Models predict that reworking
of at least 30 000 km3 of the Cenozoic Ebro Basin fill requires that opening of its drainage toward the Mediterranean has occurred long before the Messinian, probably between 13 and 8.5 Ma (middle Serravallian to middle Tortonian). This is compatible with the occurrence of the large-scale preMessinian prograding Castello´n Group clastic wedge that advanced into the Valencia Trough (Ziegler, 1988; Roca and Desegaulx, 1992; Roca, 2001). Progradation of these clastics, starting in the middle Serravallian (Martı´nez del Olmo, 1996) can be interpreted as reflecting the opening of the endorheic Ebro paleo-lake. This is also consistent with an important increase in sedimentation rate in the Valencia Trough (Dan˜obeitia et al., 1990; Martı´nez
Tectonic Models for the Evolution of Sedimentary Basins
del Olmo, 1996) that may be associated with a lake overflow during a coeval transition to a wetter climate (Alonso-Zarza and Calvo, 2000; Sanz de Siria Catalan, 1993). In addition to the classical concept of opening of the Ebro Basin by escarpment erosion driven by Neogene Mediterranean streams, modeling suggests that four additional large-scale processes played a contributing and possibly major role (Figure 93(b)): (1) extensional partial breakdown of the Catalan Coastal Ranges topographic barrier; (2) extensioninduced flexural flank uplift of the Catalan Coastal Range; (3) sediment overfill of the Ebro lake; and (4) lake level rise related to a long-term climate change toward wetter conditions. Development of the Valencia Trough had two opposing effects on the opening of the Ebro Basin drainage system. First, it facilitated its opening by extensionally disrupting and narrowing the topographic barrier presented by the Eocene intraplate Catalan Coastal Ranges and by increasing the topographic gradient and erosion capacity of short streams that developed on the new Mediterranean margin. Second, it delayed its opening by flexural uplift of the flanks of the Catalan Coastal Ranges (a process similar to that described by Tucker and Slingerland, 1994). This uplift, and related exhumation, is compatible with numerical modeling (Gaspar-Escribano et al., 2004) and thermogeochronological results (Juez-Larre´ and Andriessen, 2002, 2006), indicating that the Catalan Coastal Ranges were exhumed during the Neogene by as much as 2 km. The combination of well-dated episodes of basin evolution, long-term denudation rates, and sediment delivery rates, together with numerical modeling provides a process-based approach to determine how tectonics and river dynamics shaped the present landscape of the NE parts of the Iberian Peninsula. Foreland basins are generally regarded as sedimentary accumulations whose evolution depends mainly on the tectonic history of the adjacent orogen and the mechanical response of the foreland lithosphere. Studies carried out on the Ebro Basin demonstrate that the drainage history and the organization of the fluvial–lacustrine network are additional major factors controlling its evolution. The interaction between surface processes (climate, erosion, and transport) and lithospheric tectonic deformation (e.g., Avouac and Burov, 1996; Willet, 1999; Beaumont et al., 2000; Garcia-Castellanos, 2002; Cloetingh et al., 2002; see also Avouac, this volume (Chapter 6.09)) apparently controlled the end of the 25 My long period of
593
endorheic drainage and sediment entrapment in the Ebro Basin. Aridity, probably enhanced by an intramontane orographic position, appears to have prolonged sediment entrapment and the closure of this basin. This feedback between aridity and basin closure appears to have been preconditioned by a favourable and unique tectonic setting. In the case of the Ebro Basin, the role of isostasy is enhanced as by the time of lake capture, a dramatic drainage change affected a catchment area of nearly 85 000 km2 that was very sensitive to vertical movements in the domain of its eastern drainage divide.
6.11.8 Conclusions and Future Perspectives Sedimentary basin systems are sensitive recorders of dynamic processes controlling the deformation of the lithosphere and its interaction with the deep mantle and surface processes. The thermomechanical structure of the lithosphere exerts a prime control on its response to basin-forming mechanisms, such as rifting and its deflection under vertical loads, as well as its compressional deformation. Polyphase deformation is a common feature of some of Europe’s bestdocumented sedimentary basin systems. The relatively weak lithosphere of intraplate Europe renders many of its sedimentary basins prone to neotectonic reactivation. Tectonic processes operating during basin formation and during the subsequent deformation of basins can generate significant differential topography in basin systems. As there is a close link between erosion of topographic highs and sedimentation in subsiding areas, constraints are needed on uplift and coeval subsidence to validate quantitative process-oriented models for the evolution of sedimentary basins. Integration of analog and numerical modeling provides a novel approach to assess the feedback mechanisms between deep mantle, lithospheric, and surface processes. The analysis of European basin systems demonstrates the importance of preorogenic extension of the lithosphere in the evolution of flexural foreland basins. Furthermore, compressional reactivation of extensional basins during their postrift phase appears to be a common feature of Europe’s intraplate rifts and passive-margins, reflecting temporal and spatial changes in the orientation and magnitude of the intraplate stress regime. Moreover, in the intraplate domain of continental Europe, that was thermally perturbed by
594
Tectonic Models for the Evolution of Sedimentary Basins
Cenozoic upper mantle plume activity, lithospheric folding plays an important role and strongly affects the pattern of vertical motions, both in terms of basin subsidence and the uplift of broad arches. The development of sedimentary basins is, therefore, the intrinsic result of the interplay between lithospheric stresses and rheology and thermal perturbations of the lower lithosphere–upper mantle. The elucidation of palaeo-stresses, palaeo-temperatures and palaeo-rheologies of the lithosphere is vital for the reconstruction of dynamically supported palaeo-topography. Rift shoulder topography and flexural topography in compressional systems are directly linked to the thermomechanical properties and rheological stratification of the underlying lithosphere. The temporal evolution of the strength of continents and the spatial variations in stress and strength at continental margins, rifts, and orogenic belts govern the mechanics of basin development and destruction in time and space. Structural discontinuities and preexisting weakness zones are prone to reactivation in response to the buildup of extensional stresses and thus play an important role in the localization of rifts and related transfer zones. Similarly, their reactivation in response to the buildup of compressional stresses propagating from plate boundaries into continental platform areas controls the inversion of extensional basins and the upthrusting of basement blocks and contributes to the localization of lithospheric folding. The timing and nature of underlying changes in the controlling stress field and the resulting deformation processes can be unravelled by quantitative analyses of the stratigraphic record and the architecture of such complex polyphase sedimentary basins. This permits to assess systematic differences in the timing of the transition from rifting to intraplate compression and the development of foreland fold–thrust belts and lithospheric folds and vice versa. The study of sedimentary basins demands added focus on surrounding highs that act as sediment sources as it is recognized that also uplifted/uplifting areas contain valuable information concerning, for instance, underlying deep-seated processes. Stateof-the art geothermochronology and the study of erosion and river pattern evolution provide major constraints on uplift histories. To elucidate the contribution of internal lithospheric processes and external forcing toward rates of erosion and sedimentation presents, therefore, a major challenge for sedimentary basin studies. This is particularly so as the sedimentary cover of the
lithosphere provides a high-resolution record of the changing environment and of deformation and mass transfer at the surface and at different depth levels in the crust, lithosphere, and mantle system. Monitoring lithospheric deformation and its sedimentary record provides constraints on past and present-day structures and on deformation rates. Whereas in sedimentary basin analyses tectonics, eustasy, and sediment supply have been commonly treated as separate factors, an integrated approach constrained by fully 3D quantitative subsidence and uplift history analyses in carefully selected natural laboratories is now possible and ought to be further pursued. Recent deformation, involving tectonic reactivation, has strongly affected the structure and fill of many sedimentary basins. The long-lasting memory of the lithosphere appears to play a far more important role in basin reactivation than hitherto assumed. A better understanding of the 3D linkage between basin formation and basin deformation is, therefore, an essential step in research aiming at linking lithospheric forcing and upper-mantle dynamics to crustal uplift and erosion and their effects on the dynamics of sedimentary systems. Structural analysis of the basin architecture, including palaeo-stress assessment, provides important constraints on the transient nature of intraplate stress fields and their effects on the evolution of sedimentary basins. Reconstruction of the basin history is a prerequisite for identifying transient processes having a bearing on basin (de)formation. A full 3D reconstruction, including the use of sophisticated 3D visualization and geometric construction techniques, is required for sedimentary basin with a faulted architecture. 3D backstripping, including the effects of flexural isostasy and faulting, allows a thorough assessment of sedimentation and faulting rates, as well as changing facies and geometries through time. The established transient architecture of the preserved sedimentary record serves as key input for process quantification. We have presented results of several studies integrating geothermochronology, analyses of material properties of the lithosphere and the reconstruction of the geological past from the sedimentary record. As such, we have trespassed traditional boundaries between endogene and exogene geology. During the last decades, basin analysis has been at the forefront of integrating the sedimentary and lithosphere components of geology and geophysics. In this context, it is essential to link neotectonics, surface processes and
Tectonic Models for the Evolution of Sedimentary Basins
lithospheric dynamics in reconstructions of the palaeo-topography of sedimentary basins and their surrounding areas. Combining dynamic topography and sedimentary basin dynamics is also important, especially when considering the key societal role sedimentary basins play as hosts of major resources. Moreover, most of the human population is concentrated in sedimentary basins, often close to the coastal zone and deltas that are vulnerable to geological hazards inherent to activity of the Earth system.
Acknowledgments
–Q RT
½2
where A, n, and Q are experimentally derived material constants, n is the power law exponent, Q is activation energy, R is the gas constant, and T is temperature. The state of stress is constrained by the force balance: ½3
where g is gravity and is density. In this model it is assumed that mass is conserved and the material is incompressible. The continuity equation following from the principle of mass conservation for an incompressible medium is ½4
In the model, the density is dependent on the temperature following a linear equation of state:
To study lithosphere extension, a 2-D finite element model is used. The program is based on a Lagrangian formulation, which makes it possible to track (material) boundaries, like the Moho, in time and space. A drawback of the Lagrangian method is that it is not suitable for solving very large grid deformation problems. This is a problem in analyzing of extension of the lithosphere, which is often accompanied by large deformations. The elements might become too deformed to yield accurate or stable solutions. In order to overcome this problem the finite element grid is periodically remeshed (see Van Wijk and Cloetingh (2002) and references therein). In the numerical model the base of the lithosphere is defined by the 1300 C isotherm. Under such conditions, approximately the upper half of the thermal lithosphere behaves elastically on geological time scale, while in the lower half stresses are relaxed by viscous deformation. This viscoelastic behaviour is well described by a Maxwell body, resulting in the following constitutive equation for a Maxwell viscoelastic material: 1 1 d þ 2 E dt
"_ ¼ An exp
rv ¼ 0
Appendix 1: Dynamic Model for Slow Lithospheric Extension
"_ ¼
fluid the dynamic viscosity is constant. In the lithosphere, however, nonlinear creep processes prevail and the relation between stress and strain rate can be described by
r ? þ g ¼ 0
The authors thank the Netherlands Research Centre for Integrated Solid Earth Sciences (ISES) for funding the integrated numerical/analog tectonic modeling laboratory (NUMLAB/TECLAB).
595
½1
in which "_ is strain rate, is dynamic viscosity, is stress, and E is Young’s modulus. For a Newtonian
¼ 0 ð1 – T Þ
½5
where 0 is the density at the surface, is the thermal expansion coefficient, and T is temperature. Besides viscoelastic behaviour, processes of fracture and plastic flow play an important role in deformation of the lithosphere. This deformation mechanism is active when deviatoric stresses reach a critical level. Here the Mohr–Coulomb criterion is used as a yield criterion to define this critical stress level. The Mohr–Coulomb strength criterion is defined as jtn j c – n ? tan #
½6
where tn is the shear stress component, n is the normal stress component, c is the cohesion of the material, and # is the angle of internal friction. Stresses are adjusted at each time step at which this criterion is reached. Frictional sliding and fault movement are not explicitly described by this criterion. The displacement field is obtained by solving eqns [1]–[6]. A total of 2560 straight-sided seven-node triangular elements were used with a 13-point Gaussian integration scheme. As the time discretization schemes used are fully implicit, the system is unconditionally stable. However, as the accuracy of the solution remains dependent on the dimension of the time steps, the Courant criterion was implemented.
596
Tectonic Models for the Evolution of Sedimentary Basins
Processes like sedimentation and erosion are not incorporated in the modeling, although they affect the evolution of a rift basin and rift shoulders, and can change the strength of the lithosphere. The temperature field in the lithosphere is calculated for each time step using the heat flow equation: cp
dT ¼ qj kqj T þ H dt
½7
where the density is defined by eqn [5], cp is specific heat, k is conductivity, and H is crustal heat production. Temperatures are calculated on the same grid as the velocity field, and advection of heat is accounted for by the nodal displacements.
Appendix 2: Surface Transport Model Following the formulation by Beaumont et al., (1992) and Kooi and Beaumont (1994), the surface transport numerical model adopted here assumes that the main agent of basin-scale incision and transport is the fluvial network. Although there is an ongoing discussion on the optimal empirical relationships governing these processes (e.g., Willgoose et al., 1991; Howard et al., 1994; Whipple and Tucker, 1999), this is of secondary importance here, as our analysis focuses on the large-scale, first-order features of the interplay between fluvial transport and lithospheric deformation, rather than on the properties of fluvial transport itself. According to the approach by Beaumont et al., (1992), the equilibrium transport capacity qeq of a river (defined as the amount of mass transported by a river producing no net erosion or sedimentation) is proportional to the mean water discharge Qw and the slope S along the river profile: qeq ðx ; y ; t Þ ¼ Kf S ðx; y ; t ÞQw ðx ; y ; t Þ
½8
where Kf is the fluvial transport coefficient, for which we adopt a standard value of 60 kg m3 (e.g., Kooi and Beaumont, 1996; van der Beek and Braun, 1999). For comparison with these works, note that here the sediment load has units of [mass]/[time] and Kf has units of [mass]/[volume]. In general, rivers are out of equilibrium such that the amount of material dq eroded/deposited along a river segment of length dl is proportional to the difference between the actual transported sediment q and qeq following the relation dq ðx; y ; t Þ 1 ¼ q ðx; y ; t Þ – qeq ðx ; y ; t Þ dl lf
½9
where lf is the length scale of erosion/deposition. Under these conditions a river can change from incision to deposition by a reduction in qeq, that is, by a decrease in discharge and/or slope. As water flows down the river, fluvial transport typically evolves from supply-limited to transport-limited. The main difference between this and previous models is the explicit treatment of lakes forming in local topographic minima and their implied water losses by evaporation (Figure 95). When a river reaches a lake or the sea, transported sediments are distributed in all directions from the river mouth, assuming zero transport capacity in eqn [9]. This implies exponentially decreasing deposition rates with increasing distance from the river mouth. This approach overlooks other processes affecting the pattern of deposition in deltas and must be regarded as the simplest possible approach that eventually produces lake/basin overfilling by localizing deposition in the vicinity of the river mouth. Lake overfilling by sediment accumulation, together with erosion at the lake outlet (decreasing the water level of the lake), ensures that lakes behave in the model as transitory or ephemeral phenomena that tend to disappear in the absence of tectonic relief generation.
References Aichroth B, Prodehl C, and Thybo H (1992) Crustal structure along the central segment of the EGT from seismic-refraction studies. Tectonophysics 207: 43–64. Alonso-Zarza AM and Calvo JP (2000) Palustrine sedimentation in an episodically subsiding basin: The Miocene of the northern Teruel Graben (Spain). Palaeogeography, Palaeoclimatology, Palaeoecology 160: 1–21. Anderson DL, Zang Y-S, and Tanimoto T (1992) Plume heads, continental lithosphere, flood basalts and tomography. In: Storey BC, Alabaster T, and Pankhurst RJ (eds.) Geological Society, London, Special Publications, 68: Magmatism and the Causes of Continental Break-up, pp. 99–124. London: Geological Society, London. Andeweg B and Cloetingh S (1998) Flexure and ‘unflexure’ of the North Alpine German–Austrian Molasse Basin: Constraints from forward tectonic modelling. In: Mascle A, Puigdefa`bregas C, Luterbacher HP, and Fernandez M (eds.) Geological Society, London, Special Publications, 134: Cenozoic Foreland Basins of Western Europe, pp. 403–422. London: Geological Society, London. Andeweg B and Cloetingh S (2001) Evidence for an active sinistral shear zone in the Western Alboran region. Terra Nova 13: 44–50. Andeweg B, De Vicente G, Cloetingh S, Giner J, and Mun˜oz Martin A (1999) Local stress fields and intraplate deformation of Iberia: variations in spatial and temporal interplay of regional stress sources. Tectonophysics 305: 153–164. Arenas C and Pardo G (1999) Latest Oligocene–late Miocene lacustrine systems of the north-central part of the Ebro Basin (Spain): Sedimentary facies model and palaeogeographic
Tectonic Models for the Evolution of Sedimentary Basins synthesis. Palaeogeography, Palaeoclimatology, Palaeoecology 151: 127–148. Arenas C, Milla´n H, Pardo G, and Pocovı´ A (2001) Ebro Basin continental sedimentation associated with late compressional Pyrenean tectonics (north-eastern Iberia): Controls on basin margin fans and fluvial system. Basin Research 13: 65–89. Argus DF and Heflin MB (1995) Plate motion and crustal deformation estimated with geodetic data from the Global Positioning System. Geophysical Research Letter 22: 1973–1976. Artemieva IM (2006) Global 10 10 thermal model TC1 for the continental lithosphere: Implications for lithosphere secular evolution. Tectonophysics 416: 245–277. Artemieva IM and Mooney WD (2001) Thermal thickness and evolution of Precambrian lithosphere. A global study. Journal of Geophysical Research 106: 16387–16414. Artemieva IM and Mooney WD (2002) On the relationship between cratonic lithosphere thickness, plate motions, and basal drag. Tectonophysics 358: 211–231. Artemieva IM, Thybo H, and Kaban MK (2006) Deep Europe today: Geophysical synthesis of upper mantle structure and lithospheric processes over 3.5 Ga. In: Gee DG and Stephenson RA (eds.) London: Geological Society, London. 32: European Lithosphere Dynamics, pp. 11–42. Arthaud F and Matte P (1977) Late Paleozoic strike-slip faulting in Southern Europe and North Africa: results of a right-lateral shear between the Appalachians and Urals. Geological Society of America Bulletin 88: 1305–1320. Artyushkov EV and Baer MA (1990) Formation of hydrocarbon basins: Subsidence without stretching in West Siberia. In: Pinet B and Bois C (eds.) The Potential of Deep Seismic Reflection Profiling for Hydrocarbon Exploration, pp. 45–61. Paris: Technip. Avouac JP and Burov EB (1996) Erosion as a driving mechanism of intracontinental mountain growth. Journal of Geophysical Research 101: 17747–17769. Babuska V and Plomerova J (1992) The lithosphere in central Europe – seismological and petrological aspects. Tectonophysics 207: 141–163. Babuska V and Plomerova J (1993) Lithospheric thickness and velocity anisotropy – seismological and geothermal aspects. Tectonophysics 225: 79–89. Babuska V and Plomerova J (2001) Subcrustal lithosphere around the Saxothuringian–Moldanubian Suture Zone – a model derived from anisotropy of seismic wave velocities. Tectonophysics 332: 185–199. Bada G and Horva´th F (2001) On the structure and tectonic evolution of the Pannonian basin and surrounding orogens. Acta Geologica Hungarica 44: 301–327. Bada G, Gerner P, Cloetingh S, and Horva´th F (1998) Sources of recent tectonic stress in the Pannonian region: inferences from finite element modelling. Geophysical Journal International 134: 87–102. Bada G, Horvath F, Gerner P, and Fejes I (1999) Review of the present-day geodynamics of the Pannonian Basin; progress and problems. Journal of Geodynamics 27: 501–527. Bada G, Horva´th F, Cloetingh S, Coblentz DD, and To´th T (2001) The role of topography induced gravitational stresses in basin inversion: The case study of the Pannonian basin. Tectonics 20: 343–363. Bala A, Radulian M, and Popescu E (2003) Earthquakes distribution and their focal mechanism in correlation with the active tectonic zones of Romania. Journal of Geodynamics 36: 129–145. Balla Z (1986) Paleotectonic reconstruction of the central Alpine-Mediterranean belt for the Neogene. Tectonophysics 127: 213–243. Bally AW and Snelson S (1980) Realms of subsidence. In: Miall AD (ed.) Bulletin Du Centre De Recherches Exploration Production
597
Elf Aquitaine, 6: Facts and Principles of World Petroleum Occurrence. Mem. Can. Soc. Pet. Geol. 6, pp. 9–94. Banda E and Santanach P (1992) The Valencia trough (western Mediterranean): an overview. Tectonophysics 208: 183–202. Barbier F, Duverge´ J, and le Pichon X (1986) Structure profonde de la marge Nord-Gascogne. Implications sur le mechanism de rifting et la formation de la marge continentale. Bulletin Du Centre De Recherches Exploration Production Elf Aquitaine 10: 105–121. Barton P and Wood R (1984) Tectonic evolution of the North Sea basin: crustal stretching and subsidence. Geophysical Journal of the Royal Astronomical Society 79: 987–1022. Bartrina MT, Cabrera L, Jurado MJ, Guimera´ J, and Roca E (1992) Evolution of the central Catalan margin of the Valencia Trough (western Mediterranean). Tectonophysics 203: 219–247. Bassi G (1995) Relative importance of strain rate and rheology for the mode of continental extension. Earth and Planetary Science Letters 122: 195–210. Bayer U, Scheck M, and Rabbel W (1999) An integrated study of the NE German Basin. Tectonophysics 314: 285–307. Beaumont C (1981) Foreland basins. Geophysical Journal of the Royal Astronomical Society 65: 291–329. Beaumont C, Keen CE, and Boutillier R (1982) On the evolution of rifted continental margins: comparison of models and observations for the Nova Scotian Margin. Geophysical Journal of the Royal Astronomical Society 70: 667–715. Beaumont C, Fullsack Ph, and Hamilton W (1992) Erosional control of active compressional orogens. In: McClay KR (ed.) Thrust Tectonics, pp. 1–18. London: Chapman & Hall. Beaumont C, Mun˜oz JA, Hamilton J, and Fullsack P (2000) Factors controlling the Alpine evolution of the central Pyrenees inferred from a comparison of observations and geodynamical models. Journal of Geophysical Research 105: 8121–8145. Beekman F, Bull JM, Cloetingh S, and Scrutton RA (1996) Crustal fault reactivation as initiator of lithospheric folding in the Central Indian Ocean. Geological Society, London, Special Publications 99: 251–263. Behr HJ and Heinrichs T (1987) Geological interpretation of DEKORP 2 –S: A deep seismic reflection profile across the Saxothuringian and possible implications for late Variscan structural evolution of Central Europe. Tectonophysics 142: 173–202. Benek R, Kramer W, McCann T, Scheck M, Negendank JFW, Kronich D, Huebscher H-D, and Bayer U (1996) PermoCarboniferous magmatism of the Northeast German Basin. Tectonophysics 266: 379–404. Bergerat F (1987) Stress fields in the European platform at the time of Africa–Eurasia collision. Tectonics 6: 99–132. Bertotti G and Ter Voorde M (1994) Thermal effects of normal faulting during rifted basin formation 2. The Lugano–Val Grande normal fault and the role of preexisting thermal anomalies. Tectonophysics 240: 145–157. Bertotti G, ter Voorde M, Cloetingh S, and Picotti V (1997) Thermomechanical evolution of the South Alpine rifted margin (North Italy): constraints on the strength of passive continental margins. Earth and Planetary Science Letters 146: 181–193. Bertotti G, Podlachikov Y, and Daehler A (2000) Dynamic link between the level of ductile crustal flow and style of normal faulting of brittle crust. Tectonophysics 320: 195–218. Bertotti G, Mat¸enco L, and Cloetingh S (2003) Vertical movements in and around the SE Carpathian foredeep: Lithospheric memory and stress field control. Terra Nova 15: 299–305. Bijwaard H and Spakman W (1999a) Fast kinematic ray tracing of first- and later-arriving global seismic phases. Geophysical Journal International 139: 359–369. Bijwaard H and Spakman W (1999b) Tomographic evidence for a narrow whole mantle plume below Iceland. Earth and Planetary Science Letters 166: 121–126.
598
Tectonic Models for the Evolution of Sedimentary Basins
Blanco M-J and Spakman W (1993) The P-wave velocity structure of the mantle below the Iberian Peninsula; evidence for subducted lithosphere below southern Spain. Tectonophysics 221: 13–34. Blundell D, Freeman R, and Mueller S (eds.) (1992) A Continent Revealed, The European Geotraverse, 275p. Cambridge: Cambridge University Press. Bond GC and Kominz M (1984) Construction of tectonic subsidence curves for the early Paleozoic miogeocline southern Canadian Rocky Mountains: Implications for subsidence mechanisms age of break up and crustal thinning. Geological Society of America Bulletin 95: 155–173. Bonin B (1990) From orogenic to anorogenic settings: evolution of granitoids suits after a major orogenesis. Geological Journal 25: 260–270. Bonin B, Bra¨ndlin P, Bussy F, Desmons J, Eggenberger U, Finger F, Graf K, Marro C, Mercolli L, Oberha¨nsli R, Ploquin A, von Quadt A, von Raumer J, Schaltegger U, and Steyer HP (1993) Late Variscan magmatic evolution of the Alpine basement. In: von Raumer JF and Neubauer F (eds.) Pre-Mesozoic Geology of the Alps, pp. 171–201. New York: Springer. Bonjer KP (1997) Seismicity pattern and style of seismic faulting at the eastern borderfault of the southern Rhine Graben. Tectonophysics 275: 41–69. Bonnet S, Guillocheau F, and Brun J-P (1998) Relative uplift measured using river incision: The case of the Armorican basement (France). Comptes Rendus Academie des Sciences, Earth and Planetary Sciences 327: 245–251. Bonnet S, Guillocheau F, Brun J-P, and Van den Driessche J (2000) Large-scale relief development related to Quaternary tectonic uplift of a Proterozoic–Paleozoic basement: The Armorican Massif, NW France. Journal of Geophysical Research 105: 19273–19288. Bousquet R, Goffe´ B, Henry P, and Chopin Ch (1997) Kinematic, thermal and petrological model of the Central Alps: Lepontine metamorphism in the upper crust and eclogitisation of the lower crust. Tectonophysics 273: 105–127. Bott MHP (1992) Modelling the loading stresses associated with continental rift systems. Tectonophysics 215: 99–115. Bott MHP (1993) Modelling of plate-driving mechanisms. Journal of the Geological Society 150: 941–951. Bott MHP and Kusznir NJ (1979) Stress distribution associated with compensated plateau uplift structures with application to the continental splitting mechanism. Geophysical Journal of the Royal Astronomical Society 56: 451–459. Braathen A, Osmundsen P-T, Nordgulen O, Roberts D, and Meyer GB (2002) Orogen-parallel extension of the Caledonides in northern central Norway: An overview. Norwegian Journal of Geology 82: 225–242. Braun J and Beaumont C (1989) A physical explanation of the relationship between flank uplifts and the breakup unconformity at rifted continental margins. Geology 17: 760–764. Braun J and Sambridge M (1997) Modelling landscape evolution on geological time scales: A new method based on irregular spatial discretization. Basin Research 9: 27–52. Breitkreuz C and Kennedy A (1999) Magmatic flare-up at the Carboniferous/Permian boundary in the NE German basin revealed by SHRIMP zircon ages. Tectonophysics 302: 307–326. Brun JP and Nalpas T (1996) Graben inversion in nature and experiments. Tectonics 15: 677–687. Brunet M-F and Cloetingh S (eds.) (2003) Sedimentary Geology, 156: Integrated Peri-Tethyan Basins Studies (Peri-Tethys Programme), 288p. Buck WR (1991) Modes of continental lithospheric extension. Journal of Geophysical Research 96: 20161–20178.
Bukovics C and Ziegler PA (1985) Tectonic development of the Mid-Norway continental margin. Marine and Petroleum Geology 2: 2–22. Burbank D (1992) Causes for recent uplift deduced from deposited patterns in the Ganges basin. Nature 357: 48–50. Burg J-P, van den Driesschen J, and Brun J-P (1994) Syn- to post-thickening extension in the Variscan Belt of Western Europe: Modes and structural consequences. Ge´ologie de la France 3: 33–51. Burov EB and Cloetingh S (1997) Erosion and rift dynamics: new thermo-mechanical aspects of postrift evolution of extensional basins. Earth and Planetary Science Letters 150: 7–26. Burov E and Diament M (1995) The effective elastic thickness of continental lithosphere: What does it really mean? (constraints from rheology, topography and gravity). Journal of Geophysical Research 100: 3905–3927. Burov EB and Molnar P (1998) Gravity Anomalies over the Ferghana Valley (central Asia) and intracontinental Deformation. Journal of Geophysical Research 103: 18137–18152. Burov EB, Nikishin AM, Cloetingh S, and Lobkovsky LI (1993) Continental lithosphere folding in central Asia (Part II): Constraints from gravity and tectonic modelling. Tectonophysics 226: 73–87. Burton R, Kendall CGStC, and Lerche I (1987) Out of our depth: on the impossibility of fathoming eustasy from the stratigraphic record. Earth Science Reviews 24: 237–277. Calcagnile G and Panza GF (1987) Properties of the lithosphere–asthenosphere system in Europe with a view toward Earth conductivity. Pure and Applied Geophysics 125: 241–254. Carswell DA (1990) Eclogite and eclogite facies: definitions and classification. In: Carswell DA (ed.) Eclogite Facies Rocks, pp. 1–13. New York: Blackie. Carter NL and Tsenn MC (1987) Flow properties of continental lithosphere. Tectonophysics 136: 27–63. Casas-Sainz AM, Corte´s-Gracia AL, and Maestro-Gonza´les A (2000) Intraplate deformation and basin formation during the Tertiary within the northern Iberian plate: origin and evolution of the Almaza´n Basin. Tectonics 19: 258–289. Chalot-Prat F and Giˆrbacea R (2000) Partial delamination of continental mantle-lithosphere, uplift-related crust-mantle decoupling, volcanism and basin formation: a new model for the Pliocene-Quaternary evolution of the southern EastCarpathians, Romania. Tectonophysics 327: 83–107. Chang HK, Kowsmann RO, Figueiredo AMF, and Bender AA (1992) Tectonics and Stratigraphy of the East Brazil Rift System: an overview. Tectonophysics 213: 97–138. Ciulavu D, Dinu C, and Cloetingh S (2002) Late Cenozoic tectonic evolution of the Transylvanian Basin and northeastern part of the Pannonian Basin (Romania): Constraints from seismic profiling and numerical modelling. In: Cloetingh S, Horva´th F, Bada G, and Lankreijer A (eds.) EGU St. Mueller Special Publication Series, 3: Neotectonics and Surface Processes: The Pannonian Basin and Alpine/ Carpathian System, pp. 105–120. EGU. Clavell E (1992) Geologia del petroli de les conques terciaries de Catalunya. Ph.D. thesis, University of Barcelona, Barcelona. Cloetingh S (1988) Intra-plate stress: a new element in basin analysis. In: Kleinspehn KL and Paola C (eds.) Frontiers in Sedimentary Geology – New Perspectives in Basin Analysis, pp. 205–230. New York: Springer Verlag. Cloetingh S and Kooi H (1992) Intraplate stresses and dynamical aspects of rift basins. Tectonophysics 215: 167–185. Cloetingh S and Burov E (1996) Thermomechanical structure of European continental lithosphere: constraints from rheological profiles and EET estimates. Geophysical Journal International 124: 695–723.
Tectonic Models for the Evolution of Sedimentary Basins Cloetingh S and Wortel R (1986) Stress in the Indo-Australian plate. Tectonophysics 132: 49–67. Cloetingh S, McQueen H, and Lambeck K (1985) On a tectonic mechanism for regional sea level fluctuations. Earth and Planetary Science Letters 75: 157–166. Cloetingh S, Lambeck K, and McQueen H (1987) Apparent sealevel fluctuations and a paleo-stress field for the North Sea region. In: Brooks J and Glennie K (eds.) Petroleum Geology of North West Europe, pp. 49–55. London: Graham and Trotman. Cloetingh S, Wortel R, and Vlaar NJ (1989) On the initiation of subduction zones. Pure and Applied Geophysics 129: 7–25. Cloetingh S, Gradstein F, Kooi H, Grant A, and Kaminski M (1990) Plate reorganization: a cause of rapid late Neogene subsidence and sedimentation around the North Atlantic? Journal of the Geological Society 147: 495–506. Cloetingh S, Van der Beek PA, Van Rees D, Roep TB, Biermann C, and Stephenson RA (1992) Flexural interaction and the dynamics of Neogene extensional basin formation in the Alboran–Betic region. Geo-Marine Letters 12: 66–75. Cloetingh S, Sassi W, and Horva´th F (eds.) (1993) The origin of sedimentary basins; inferences from quantitative modelling and basin analyses. Tectonophysics 226, 518p. Cloetingh S, Sassi W, and Task Force Team (1994) The origin of sedimentary basins: a status report from the task force of the International Lithosphere Program. Marine and Petroleum Geology 11: 659–683. Cloetingh S, d’Argenio B, Catalano R, Horvath F, and Sassi W, (eds.) (1995a) Interplay of extension and compression in basin formation. Tectonophysics 252: pp. 1–484. Cloetingh S, Van Wees JD, Van der Beek PA, and Spadini G (1995b) Role of pre-rift rheology in kinematics of basin formation: constrains from thermo-mechanical modelling of Mediterranean basins and intracratonic rifts. Marine and Petroleum Geology 12: 793–808. Cloetingh S, Durand B, and Puigdefabregas C (eds.) (1995c) Special Issue on Integrated Basin Studies (IBS) – A European Commission (DGXII) project. Marine and Petroleum Geology, 12(8): 787–963. Cloetingh S, Ben-Avraham Z, Sassi W, and Horva´th F, (eds.) (1996) Dynamics of strike slip tectonics and basin formation. Tectonophysics 266: 1–523. Cloetingh S, Van Balen RT, Ter Voorde M, Zoetemeijer BP, and Den Bezemer T (1997) Mechanical aspects of sedimentary basin formation: development of integrated models for lithospheric and surface processes. International Journal of Earth Sciences 86: 226–240. Cloetingh S, Burov E, and Poliakov A (1999) Lithosphere folding: primary response to compression? (from Central Asia to Paris Basin). Tectonics 18: 1064–1083. Cloetingh S, Burov E, Beekman F, Andeweg B, Andriessen PAM, Garcia-Castellanos D, de Vicente G, and Vegas R (2002) Lithospheric folding in Iberia. Tectonics 21(5): 1041 (doi:10.1029/2001TC901031). Cloetingh S, Spadini G, van Wees JD, and Beekman F (2003) Thermo-mechanical modelling of Black Sea Basin (de)formation. Sedimentary Geology 156: 169–184. Cloetingh S, Burov E, Mat¸enco L, Toussaint G, Bertotti G, Andriessen P, Wortel MJR, and Spakman W (2004) Thermomechanical controls on the mode of continental collision in the SE Carpathians (Romania). Earth and Planetary Science Letters 218: 57–76. Cloetingh S, Ziegler PA, Beekman F, Andriessen PAM, Mat¸enco L, Bada G, Garcia-Castellanos D, Hardebol N, De`zes P, and Sokoutis D (2005) Lithospheric memory, state of stress and rheology: Neotectonic controls on Europe’s intraplate continental topography. Quaternary Science Reviews 24: 241–304. Cloetingh S, Bada G, Mat¸enco L, Lankreijer A, Horva´th F, and Dinu C (2006) Thermo-mechanical modelling of the
599
Pannonian-Carpathian system: Modes of tectonic deformation, lithospheric strength and vertical motions. In: Gee D and Stephenson R (eds.) Geological Society, London, Memoirs, 32: European Lithosphere Dynamics, pp. 207–221. London: Geological Society, London. Cobbold PR, Davy P, Gapais EA, Rossello EA, Sadybasakov E, Thomas JC, Tondji Biyo JJ, and De Urreiztieta M (1993) Sedimentary basins and crustal thickening. Sedimentary Geology 86: 77–89. Coney PJ, Mun˜oz JA, McKlay KR, and Evenchick CA (1996) Syntectonic burial and post-tectonic exhumation of the Southern Pyrenees foreland fold-thrust belt. Journal of the Geological Society 153: 9–16. Cortesogno L, Dallagiovanna A, Gaggero L, Oggiano G, Ronchi A, Seno S, and Vanossi M (1998) The Variscan postcollisional volcanism in Late Carboniferous-Permian sequences of Ligurian Alps, Southern Alps and Sardinia (Italy): a synthesis. Lithos 45: 305–328. Coward MP (1993) The effects of Late Caledonian and Variscan escape tectonics on basement structure, Paleozoic basin kinematics and subsequent Mesozoic basin development in NW Europe. In: Parker J (ed.) Petroleum Geology of Northwest Europe; Proceedings of the 4th Conference, pp. 1095–1108. London: Geological Society. Csontos L (1995) Tertiary tectonic evolution of the IntraCarpathian area: A review. Acta Vulcanologica 7: 1–13. Csontos L, Nagymarosy A, Horva´th F, and Kova´cˇ M (1992) Tertiary evolution of the Intra-Carpathian area: a model. Tectonophysics 208: 221–241. Curray JR and Moore DG (1971) Growth of the Bengal deep-sea fan and denudation of the Himalayas. Geological Society of America Bulletin 82: 563–572. Dalmayrac P and Molnar B (1981) Parallel thrust and normal faulting in Peru and constraints on the state of stress. Earth and Planetary Science Letters 55: 473–481. Dan˜obeitia J, Alonso B, and Maldonado A (1990) Geological framework of the Ebro continental margin and surrounding areas. Marine Geology 95: 265–287. Davies JH and von Blanckenburg F (1996) Slab breakoff: a model of lithosphere detachment and its test in the magmatism and deformation of collisional orogens. Earth and Planetary Science Letters 129: 85–102. De Bruijne CH and Andriessen PAM (2000) Interplay of intraplate tectonics and surface processes in the Sierra de Guadarrama (central Spain), assessed by apatite fission track analysis. Physics and Chemistry of the Earth (A) 25: 555–563. De Bruijne CH and Andriessen PAM (2002) Far field effects of Alpine plate tectonism in the Iberian microplate recorded by fault-related denudation in the Spanish Central System (central Spain). Tectonophysics 349: 161–184. De Vicente G, Giner JL, Mun˜oz Martin A, Gonzalez-Casado JM, and Lindo R (1996) Determination of present-day stress tensor and neotectonic interval in the Spanish Central System and Madrid Basin, central Spain. Tectonophysics 266: 405–424. Decker K, Lillie B, and Tomek Cˇ, (eds.) p. 293 (1998) PANCARDI: The lithospheric structure and evolution of the Pannonian/ Carpathian/Dinarides region. Tectonophysics 297: 293p. Desegaulx P, Kooi H, and Cloetingh S (1991) Consequences of foreland basin development on thinned continental lithosphere: application to the Aquitaine basin (SW France). Earth and Planetary Science Letters 106: 116–132. Dewey JF (1988) Extensional collapse of orogens. Tectonics 7: 1123–1139. Dewey JF and Burke K (1975) Hot-spots and continental breakup. Geology 2: 57–60. De`zes P, Schmid SM, and Ziegler PA (2004) Evolution of the European Cenozoic Rift System: Interaction of the Alpine and Pyrenean orogens with their foreland lithosphere. Tectonophysics 389: 1–33.
600
Tectonic Models for the Evolution of Sedimentary Basins
De`zes P and Ziegler PA (2004) Moho depth map of western and central Europe. EUCOR-URGENT homepage: http:// www.unibas.ch/eucor-urgent (acessed Jul 2007). De`zes P, Schmid SM, and Ziegler PA (2005) Reply to comments on ‘Evolution of the European Cenozoic Rift System: Interaction of the Alpine and Pyrenean orogens with their foreland lithosphere. Tectonophysics 401: 257–262. Dicea O (1996) Tectonic setting and hydrocarbon habitat of the Romanian external Carpathians. In: Ziegler PA and Horva´th F (eds.) Me´moires du Museum National d’Histoire Naturelle, 170, Peri-Tethys Memoir 2: Structure and Prospects of Alpine Basins and Forelands, pp. 403–425. Paris: Commission for the Geological Map of the World. Diehl T, Ritter JRR, and the CALIXTO group, (2005) The crustal structure beneath SE Romania from teleseismic receiver functions. Geophysical Journal International 163: 238–251. Docherty C and Banda E (1995) Evidence for the eastward migration of the Alboran Sea based on regional subsidence analysis; a case for basin delamination of the subcrustal lithosphere? Tectonics 14: 804–814. Doglioni C (1993) Comparison of subduction zones versus the global tectonic pattern: A possible explanation for the Alps-Carpathians system. Geophysical Transaction 37: 253–264. Dore´ AG and Jensen LN (1996) The impact of late Cenozoic uplift and erosion on hydrocarbon exploration: offshore Norway and some other uplifted basins. Global and Planetary Change 12: 415–436. Dore´ AG and Lundin ER (1996) Cenozoic compressional structures on the NE Atlantic margin: nature, origin and potential significance for hydrocarbon exploration. Petroleum Geosciences 2: 299–311. Dore´ A, Augustson JH, Hermanrud C, Steward DJ, and Sylta O, (eds.) (1993) Basin modelling: Advances and applications. Norwegian Petroleum Society, Special Publications 3: 1–675. Dore´ AG, Lundin ER, Birkeland O, Eliassen PE, and Jensen LN (1997) The NE Atlantic margin: implications of late Mesozoic and Cenozoic events for hydrocarbon prospectivity. Petroleum Geoscience 3: 117–131. Doust H and Omatsola E (1989) Niger Delta. In: Edwards JD and Santogrossi PA (eds.) American Association of Petroleum Geologists Memoirs, 48: Divergent Passive Margin Basins, pp. 201–238. Tulsa: American Association of Petroleum Geologists. Du ZJ, Michelini A, and Panza GF (1998) EurID, a regionalized 3D seismological model of Europe. Physics of the Earth and Planetary Interiors 105: 31–62. Dumitrescu I, Sa˘ ndulescu M, and Bandrabur T (1970) Geological map scale 1:200,000, Sheet 29 Covasna. Bucharest: Geological Institute of Romania. Dunbar JA and Sawyer DS (1989) How preexisting weaknesses control the style of rifting. Journal of Geophysical Research 94: 7278–7292. Durand B, Jolivet L, Horva´th F, and Se´ranne M (eds.) (1999) Geological Society, London, Special Publications, 156: The Mediterranean basins: Tertiary extension within the Alpine orogen, 570p. London: Geological Society, London. Eide EA, Osmundsen PT, Meyer GB, Kendrick MA, and Corfu F (2002) The Nesna Shear zone, north-central Norway: An 40 Ar/39Ar record of Early Devonian–Early Carboniferous ductile extension and unroofing. Norwegian Journal of Geology 82: 317–339. Eisbacher GH, Lu¨schen E, and Wickert F (1989) Crustal-scale thrusting and extension in the Hercynian Schwarzwald and Vosges, Central Europe. Tectonics 8: 1–21. Eldholm O, Thiede J, and Taylor E (1989) Evolution of the Vøring volcanic margin. In: Eldholm O, Thiede J, and Taylor E (eds.) Proceedings of the Ocean Drilling Program, Scientific
Results, 104, pp. 1033–1065. College Station: Ocean Drilling Program. Elfrink NM (2001) Quaternary groundwater avulsions: Evidence for large-scale midcontinent folding? Association of Engineering Geologists News 44: 60. England P (1983) Constraints on extension of continental lithosphere. Journal of Geophysical Research 88: 1145–1152. England P (1986) Comment on ‘Brittle failure in the upper mantle during extension of the continental lithosphere’ by DW Sawyer. Journal of Geophysical Research 91: 10487–10490. Favre P and Stampfli GM (1992) From rifting to passive margin: the Red Sea, Central Atlantic and Alpine Tethys. Tectonophysics 215: 69–97. Fernandes R, Ambrosius B, Noomen R, Basos L, and Davila J (2000) Analysis of a permanent GPS Iberian network (GIN). 10th General Assembly Wegener Project, Extended Abstract. Finetti I, Bricchi G, Del Ben A, Pipan M, and Xuan Z (1988) Geophysical study of the Black Sea basin. Bollettino di Geofisica Teorica ed Applicata 30: 197–324. Fitzgerald PG, Munoz JA, Coney PJ, and Baldwin SL (1999) Asymmetric exhumation across the Pyrenean orogen: Implications for the tectonic evolution of a collisional orogen. Earth and Planetary Science Letters 173: 157–170. Fleitout L and Froidevaux C (1982) Tectonics and topography for lithosphere containing density heterogeneities. Tectonics 1: 21–56. Flemings PB and Jordan TE (1989) A synthetic stratigraphic model of foreland basins development. Journal of Geophysical Research 94: 3851–3866. Fodor L, Csontos L, Bada G, Benkovics L, and Gyo¨rfi I (1999) Tertiary tectonic evolution of the Carpatho-Pannonian region: A new synthesis of paleostress data. In: Durand B, Jolivet L, Horva´th F, and Se´ranne M (eds.) Geological Society, London, Special Publications, 156: The Mediterranean basins: Tertiary extension within the Alpine orogen, pp. 295–334. London: Geological Society, London. Fodor L, Bada G, Csillag G, et al. (2005) An outline of neotectonic structures and morphotectonics of the western and central Pannonian Basin. Tectonophyiscs 410: 15–41. Ford M, Lockorish WH, and Kusznir NJ (1999) Tertiary foreland sedimentation in the Southern Subalpine Chains, SE France: A geodynamic appraisal. Basin Research 11: 315–336. Forsyth D and Uyeda S (1975) On the relative importance of the driving forces of plate motions. Geophysical Journal of the Royal Astronomical Society 43: 163–200. Franke W (2000) The Mid-European segment of the Variscides: tectonostratigraphic units, terrane boundaries and plate tectonic evolution. In: Franke W, Haak V, Oncken O, and Tanner D (eds.) Geological Society, London, Special Publications, 179: Orogenic processes: Quantification and modelling in the Variscan Belt, pp. 35–62. London: Geological Society, London. Friend PF and Dabrio CJ, (eds.) (1996) Tertiary Basins of Spain. The stratigraphic record of crustal kinematics. In: World and Regional Geology, vol. 6, 418p. Cambridge: Cambridge University Press. Frizon de Lamotte D, Mercier E, Saint Bezar B, and Brace`ne R (2000) Two step Atlas building and geodynamics of the West Mediterranean. Tectonics 19: 740–761. Gabrielsen RH, Odinsen T, and Grunnaleite I (1999) Structuring of the Northern Viking Graben and the Møre Basin; the influence of basement structural grain, and the particular role of the Møre-Trøndelag Fault Complex. Marine and Petroleum Geology 16: 443–465. Garcia-Castellanos D (2002) Interplay between lithospheric flexure and river transport in foreland basins. Basin Research 14: 89–104.
Tectonic Models for the Evolution of Sedimentary Basins Garcia-Castellanos D, Cloetingh SAPL, and Van Balen RT (2000) Modeling the Middle Pleistocene uplift in the Ardennes–Rhenish Massif: thermo-mechanical weakening under the Eifel? Global and Planetary Change 27: 39–52. Garcia-Castellanos D, Ferna`ndez M, and Torne´ M (2002) Modeling the evolution of the Guadalquivir foreland basin (southern Spain). Tectonics 21(3): 1018 (doi:10.1029/ 2001TC001339). Garcia-Castellanos D, Verge´s J, Gaspar-Escribano JM, and Cloetingh S (2003) Interplay between tectonics, climate and fluvial transport during the Cenozoic evolution of the Ebro Basin (NE Iberia). Journal of Geophysical Research 108: 2347. Gaspar-Escribano JM, Van Wees J-D, Ter Voorde M, et al. (2001) 3D flexural modeling of the Ebro Basin (NE Iberia). Geophysical Journal International 145: 349–368. Gaspar-Escribano JM, Garcia-Castellanos D, Roca D, and Cloetingh S (2004) Cenozoic vertical motions of the Catalan Coastal Ranges (NE Spain): The role of tectonics, isostasy, and surface transport. Tectonics 23: doi:10.1029/ 2003TC001511. Genser J, Van Wees JD, Cloetingh S, and Neubauer F (1996) Eastern Alpine tectonometamorphic evolution: Constraints from two-dimensional P–T–t modeling. Tectonics 15: 584–604. Gerbault M, Burov E, Poliakov ANB, and Daignieres M (1998) Do faults trigger folding in the lithosphere? Geophysical Research Letter 26: 271–274. Gerner P, Bada G, Do¨ve´nyi P, Mu¨ller B, Oncescu MC, Cloetingh S, and Horva´th F (1999) Recent tectonic stress and crustal deformation in and around the Pannonian basin: Data and models. In: Durand B, Jolivet L, Horva´th F, and Se´ranne M (eds.) Geological Society, London, Special Publications, 156: The Mediterranean basins: Tertiary extension within the Alpine orogen, pp. 269–294. London: Geological Society, London. Giˆrbacea R and Frisch W (1998) Slab in the wrong place Lower lithospheric mantle delamination in the last stage of the Eastern Carpathian subduction retreat. Geology 26: 611–614. Goes S, Spakman W, and Bijwaard H (1999) A lower mantle source for Central European volcanism. Science 286: 1928–1930. Goes S, Govers R, and Vacher P (2000a) Shallow upper mantle temperatures under Europe from P and S wave tomography. Journal of Geophysical Research 105: 11153–11169. Goes S, Loohuis JJP, Wortel MJR, and Govers R (2000b) The effect of plate stresses and shallow mantle temperatures on tectonics of northwestern Europe. Global and Planetary Change 27: 23–39. Go¨lke M and Coblentz DD (1996) Origin of the European regional stress field. Tectonophysics 266: 11–24. Go¨lke M, Cloetingh S, and Coblentz DD (1996) Patterns of stress at the mid-Norwegian margin 62–68 N. Tectonophysics 266: 48–62. Govers R and Wortel MJR (1993) Initiation of assymetric extension in continental lithosphere. Tectonophysics 223: 75–96. Gradstein FM and Ogg JG (1996) A Phanerozoic time-scale. Episodes 19: 3–5. Granet M, Wilson M, and Achauer U (1995) Imaging a mantle plume beneath the French Massif Central. Earth and Planetary Science Letters 136: 281–296. Griffin WL, O’Reilly SY, and Pearson NY (1990) Eclogite stability near the crust–mantle boundary. In: Carswell DA (ed.) Eclogite Facies Rocks, pp. 291–314. New York: Blackie. Gru¨nthal G (1999) Seismic hazard assessment for Central, North and Northwest Europe: GSHAP Region 3. Annali di Geophisica 42: 999–1011. Guggisberg B, Kaminski W, and Prodehl C (1991) Crustal structure of the Fennoscandian Shield: A traveltime interpretation of the long-range Fennolora seismic refraction profile. Tectonophysics 195: 105–137.
601
Harland WB, Armstrong RB, Cox AV, Craig LE, Smith AG, and Smith DG (1990) A Geologic Time Scale, 263p. Cambridge: Cambridge University Press. Hauser F, Raileanu V, Fielitz W, Bala A, Prodehl C, Polonic G, and Schulze A (2001) VRANCEA99 – The crustal structure beneath the southeastern Carpathians and the Moesian Platform from a seismic refraction profile in Romania. Tectonophysics 340(3/4): 233–256. Hegner F, Chen F, and Hann HP (2001) Chronology of basin closure and thrusting in the internal zone of the Variscan belt in the Schwarzwald, Germany: Evidence from zircon ages, trace element geochemistry, and Nd isotopic data. Tectonophysics 332: 169–184. Hellinger SJ and Sclater JG (1983) Some comments on twolayer extensional models for sedimentary basins. Journal of Geophysical Research B 88: 8251–8269. Hendriks BWH and Andriessen PAM (2002) Pattern and Timing of the Post-Caledonian Denudation of Northern Scandinavia Constrained by Apatite Fission Track Thermochronology. Geological Society, London, Special Publication 196: 117–137. Henk A (1999) Did the Variscides collapse or were they torn apart? a quantitative evaluation of the driving forces for postconvergent extension in central Europe. Tectonics 18: 774–792. Henk A, Von Blanckenburg F, Finger F, Schaltegger U, and Zulauf G (2000) Syn-convergent high-temperature metamorphism and magmatism in the Variscides: A discussion of potential heat sources. In: Franke W, Haak V, Oncken O, and Tanner D (eds.) Geological Society, London, Special Publications, 179: Orogenic processes: Quantification and modelling in the Variscan belt, pp. 387–399. London: Geological Society, London. Hippolyte J-C, Badescu D, and Constantin P (1999) Evolution of the transport direction of the Carpathian belt during its collision with the east European Platform. Tectonics 18(6): 1120–1138. Hoernle K, Zhang Y, and Graham D (1995) Seismic and geochemical evidence for large-scale mantle upwelling beneath the eastern Atlantic and western and central Europe. Nature 374: 34–39. Holtedahl O (1953) On the oblique uplift of some northern lands. Norges Geologiske Undersøkelse Bulletin T14: 132–139. Horva´th F (1993) Towards a mechanical model for the formation of the Pannonian basin. Tectonophysics 226: 333–357. Horva´th F (1995) Phases of compression during the evolution of the Pannonian basin and its bearing on hydrocarbon exploration. Marine and Petroleum Geology 12: 837–844. Horva´th F and Cloetingh S (1996) Stress-induced late-stage subsidence anomalies in the Pannonian basin. Tectonophysics 266: 287–300. Horva´th F and Tari G (1999) IBS Pannonian basin project: a review of the main results and their bearings on hydrocarbon exploration. In: Durand B, Jolivet L, Horva´th F, and Se´ranne M (eds.) Geological Society, London, Special Publications, 156: The Mediterranean basins: Tertiary extension within the Alpine orogen, pp. 195–213. London: Geological Society, London. Horva´th F, Stegena L, and Ge´czy B (1975) Ensimatic and ensialic interarc basins. Journal of Geophysical Research 80: 281–283. Horva´th F, Szalay A´, and Royden LH (1988) Subsidence, thermal and maturation history of the Great Hungarian Plain. In: Royden LH and Horva´th F (eds.) American Association of Petroleum Geologists Memoirs, 45: The Pannonian Basin: A case study in basin evolution, pp. 355–372. Tulsa: American Association of Petroleum Geologists. Horva´th F, Bada G, Szafia´n P, Tari G, A´da´m A, and Cloetingh S (2006) Formation and deformation of the Pannonian basin:
602
Tectonic Models for the Evolution of Sedimentary Basins
Constraints from observational data. In: Gee D and Stephenson R (eds.) Geological Society, London, Memoirs, 32: European Lithosphere Dynamics, pp. 191–206. London: Geological Society, London. Houseman G and England P (1986) A dynamical model for lithosphere extension and sedimentary basin formation. Journal of Geophysical Research 91: 719–729. Howard AD, Dietrich EW, and Seidl AM (1994) Modeling fluvial erosion on regional to continental scales. Journal of Geophysical Research 99: 13971–13986. Huismans RS, Podlachikov YY, and Cloetingh S (2001a) The transition from passive to active rifting: relative importance of asthenospheric doming and passive extension of the lithosphere. Journal of Geophysical Research 106: 11271–11291. Huismans R, Podladchikov Y, and Cloetingh S (2001b) Dynamic modeling of the transition from passive to active rifting, application to the Pannonian basin. Tectonics 20: 1021–1039. Izotova TS and Popadyuk, IV (1996) Oiland gas accumulations in the Late Jurassicreefal complex of the West Ukrainian Carpathian fore deep. In: Ziegler PA and Horva`th F (eds.) Me´moires du Museum National d’Histoire Naturelle, 170, Peri-Tethy Memoir 2: Structure and Prospects of Alpine Basins and Forelands, pp. 375–390. Paris: Commission for the Geological Map of the World. Janssen ME (1996) Intraplate Deformation in Africa as a Consequence of Plate Boundary Changes. PhD thesis, Vrije Universiteit, Amsterdam, 161p. Janssen ME, Torne M, Cloetingh S, and Banda E (1993) Pliocene uplift of the eastern Iberian margin: inferences from quantitative modelling of the Valencia Trough. Earth and Planetary Science Letters 119: 585–597. Janssen ME, Stephenson RA, and Cloetingh S (1995) Temporal and spatial correlations between changes in plate motions and the evolution of rifted basins in Africa. Geological Society of America Bulletin 107: 1317–1332. Jarvis GT and McKenzie DP (1980) Sedimentary basin formation with finite extension rates. Earth and Planetary Science Letters 48: 42–52. Johnson DD and Beaumont C (1995) Preliminary results from a planform kinematic model of orogen evolution, surface processes and the development of clastic foreland basin stratigraphy. In: Dorobek SL and Ross GM (eds.) SEPM Special Publications, 52: Stratigraphic Evolution of Foreland Basins, pp. 1–24. Tulsa: SEPM. Jolivet L, Huchon P, and Rangin C (1989) Tectonic setting of Western Pacific marginal basins. Tectonophysics 160: 23–47. Jones CH, Wernicke BP, Farmer LG, Coleman DS, McKenna LW, and Perry FV (1992) Variations across and along a major continental rift: an interdisciplinary study of the Basin and Range Province, Western USA. Tectonophysics 213: 57–96. Jordt H, Faleide JI, Bjorlykke K, and Ibrahim MT (1995) Cenozoic sequence stratigraphy in the Central and Northern North Sea Basin: Tectonic development, sediment distribution and provenance areas. Marine and Petroleum Geology 12: 845–880. Judenherc S, Granet M, Brun J-P, Poupinet G, Plomerova J, Mocquet A, and Achauer U (2002) Images of lithospheric heterogeneities in the Armorican segment of the Hercynian Range in France. Tectonophysics 358: 121–134. Juez-Larre´ J and Andriessen PAM (2002) Post Late Paleozoic tectonism in the southern Catalan Coastal Ranges (NE Spain), assessed by apatite fission-track analysis. Tectonophysics 349: 367–368. Juez-Larre´ J and Andriessen PAM (2006) Tectonothermal evolution of the northeastern margin of Iberia since the
break-up of Pangea to present, revealed by low-temperature fission-track and (U-Th)/He thermochronology. A case history of the Catalan Coastal Ranges. Earth and Planetary Science Letters 243: 159–180. Juha´sz E, Phillips L, Mu¨ller P, et al. (1999) Late Neogene sedimentary facies and sequences in the Pannonian Basin, Hungary. In: Durand B, Jolivet L, Horva´th F, and Se´ranne M (eds.) Geological Society, London, Special Publications, 156: The Mediterranean basins: Tertiary extension within the Alpine orogen, pp. 335–356. London: Geological Society, London. Kaikkonen P, Moiso K, and Heeremans M (2000) Thermomechanical lithospheric structure of the central Fennoscandian Shield. Physics of the Earth and Planetary Interiors 119: 209–235. Karner GD, Egan SS, and Weissel JK (1992) Modelling the tectonic development of the Tucano and Sergipe-Alagoas rift basins, Brazil. Tectonophysics 215: 133–160. Kazmin VG (1991) The position of continental flood basalts in rift zones and its bearing on models of rifting. Tectonophysics 199: 375–387. Keen CE and Boutilier R (1990) Geodynamic modelling of rift basins: the syn-rift evolution of a simple half-graben. In: Pinet B and Bois C (eds.) Collection Colloques et Se´minaires, 24: The Potential of Deep Seismic Profiling for Hydrocarbon Exploration, pp. 23–33. Paris: Editions Technip. Kooi H (1991) Tectonic modelling of extensional basins: the role of lithospheric flexure, intraplate stresses and relative sealevel change. PhD Thesis, Vrije Universiteit, Amsterdam, 183p. Kooi H and Beaumont C (1994) Escarpment evolution on highelevation rifted margins: Insights derived from a surface processes model that combines diffusion, advection, and reaction. Journal of Geophysical Research 99: 12191–12209. Kooi H and Beaumont C (1996) Large-scale geomorphology; classical concepts reconciled and integrated with contemporary ideas via a surface processes model. Journal of Geophysical Research 101: 3361–3386. Kooi H and Cloetingh S (1989) Intraplate stresses and the tectono-stratigraphic evolution of the Central North Sea. In: Tankard AJ and Balkwill HR (ed.) American Association of Petroleum Geologists Memoirs, 46: Extensional Tectonics and Stratigraphy of the North Atlantic Margins, pp. 541–558. Tulsa: American Association of Petroleum Geologists. Kooi H, Hettema M, and Cloetingh S (1991) Lithospheric dynamics and the rapid Pliocene-Quaternary subsidence in the North Sea basin. Tectonophysics 192: 245–259. Kooi H, Cloetingh S, and Burrus J (1992) Lithospheric necking and regional isostasy at extensional basins: Part 1. Subsidence and gravity modelling with an application to the Gulf of Lions margin (SE France). Journal of Geophysical Research 97: 17553–17571. Kova´cˇ M, Kra´l J, Ma´rton E, Plasˇienka D, and Uher P (1994) Alpine uplift history of the Central Western Carpathians: geochronological, paleomagnetic, sedimentary and structural data. Geologica Carpathica 45: 83–96. Krzywiec P (2001) Contrasting tectonic and sedimentary history of the central and eastern parts of the Polish Carpathian Foredeep basin; results of seismic data interpretation. Marine and Petroleum Geology 18: 13–38. Ku¨hni A and Pfiffner OA (2001) Drainage patterns and tectonic forcing: A model study for the Swiss Alps. Basin Research 13: 169–197. Kusznir NJ and Park RG (1987) The Extensional Strength of the Continental Lithosphere; its Dependence on Geothermal Gradient, and Crustal Composition and Thickness. Geological Society, London, Special Publication 28: 35–52. Kusznir NJ and Ziegler PA (1992) Mechanics of continental extension and sedimentary basin formation: A simple-shear/
Tectonic Models for the Evolution of Sedimentary Basins pure-shear flexural cantilever model. Tectonophysics 215: 117–131. Kusznir NJ, Marsden G, and Egan SS (1991) A flexural-cantilever simple-shear/pure-shear model of continental lithosphere extension: Application to the Jeanne d’Arc Basin, Grand Banks and Viking Graben, North Sea. In: Roberts AM, Yielding G, and Freeman B (eds.) Geological Society, London, Special Publications, 56: The Geometry of Normal Faults, pp. 41–60. London: Geological Society, London. Kusznir NJ, Stovba SM, Stephenson RA, and Poplavskii KN (1996) The formation of the northwestern Dniepr-Donets Basin: 2-D forward and reverse syn-rift and postrift modelling. Tectonophysics 268: 237–255. Lagarde J-L, Baize S, Amorese D, Delcaillau B, and Font M (2000) Active tectonics, seismicity and geomorphology with special reference to Normandy (France). Journal of Quaternary Science 15: 745–758. Lambeck K (1983) The role of compressive forces in intracratonic basin formation and mid-plate orogenies. Geophysical Research Letters 10: 845–848. Lankreijer A (1998) Rheology and basement control on extensional basin evolution in Central and Eastern Europe: Variscan and Alpine-Carpathian-Pannonian tectonics. Ph.D thesis, Vrije Universiteit, Amsterdam, 158p. Lankreijer A, Kovacˇ M, Cloetingh S, Pitonˇa´k P, Hloˆsˇka M, and Biermann C (1995) Quantitative subsidence analysis and forward modelling of the Vienna and Danube basins: thinskinned versus thick-skinned extension. Tectonophysics 252: 433–451. Lankreijer A, Mocanu V, and Cloetingh S (1997) Lateral variations in lithospheric strength in the Romanian Carpathians, constraints on basin evolution. Tectonophysics 272: 433–451. Lankreijer A, Bielik M, Cloetingh S, and Majcin D (1999) Rheology predictions across the western Carpathians, Bohemian massif, and the Pannonian basin: Implications for tectonic scenarios. Tectonics 18: 1139–1153. Lardeaux JM, Ledru P, Daniel I, and Duche`ne S (2001) The Variscan French Massif Central – A new addition to the ultrahigh pressure metamorphic ‘club’: exhumation processes and geodynamic consequences. Tectonophysics 332: 143–162. Larsen RM, Brekke H, Larsen BT, and Talleraas E (eds.) (1992)Structural and tectonic modelling and its application to petroleum geology. Norwegian Petroleum Society, Special Publications, 1: pp. 1–549. Larsen TB, Yuen DA, and Storey M (1999) Ultrafast mantle plume and implications for flood basalt volcanism in the North Atlantic region. Tectonophysics 311: 31–43. Laske G and Masters G (1997) A global digital map of sediment thickness. EOS Transactions AGU 78: F483. Lawrence DT, Doyle M, and Aigner T (1990) Stratigraphic simulation of sedimentary basins: Concepts and calibration. American Association of Petroleum Geologists Bulletin 74: 273–295. Le Pichon X, Henry P, and Goffe´ B (1997) Uplift of Tibet: from eclogite to granulite – implications for the Andean Plateau and the Variscan belt. Tectonophysics 273: 57–76. Lefort J-P and Agarwal P (1996) Gravity evidence for an Alpine buckling of the crust beneath the Paris Basin. Tectonophysics 258: 1–14. Lenkey L (1999) Geothermics of the Pannonian basin and its bearing on the tectonics of basin evolution. PhD thesis, Vrije Universiteit, Amsterdam, 215p. Lenkey L, Do¨ve´nyi P, Horva´th F, and Cloetingh S (2002) Geothermics of the Pannonian basin and its bearing on the neotectonics. In: Cloetingh S, Horva´th F, Bada G, and Lankreijer A (eds.) EGU St. Mueller Special Publication Series, 3: Neotectonics and surface processes: The Pannonian basin and Alpine/Carpathian system pp. 29–40. EGU.
603
Lenoˆtre N, Thierry P, Blanchin R, and Brochard G (1999) Current vertical movement demonstrated by comparative leveling in Brittany (France). Tectonophysics 301: 333–344. Letouzey J (1986) Cenozoic paleo-stress pattern in the Alpine foreland and structural interpretation in a platform basin. Tectonophysics 132: 215–231. Letouzey J, Werner P, and Marty A (1991) Fault reactivation and structural inversion. Backarc intraplate compressive deformations. Examples of the eastern Sunda shelf (Indonesia). Tectonophysics 183: 341–362. Linzer HG (1996) Kinematics of retreating subduction along the Carpathian arc, Romania. Geology 24: 167–170. Lobkovsky LI, Cloetingh S, Nikishin AM, et al. (1996) Extensional basins of the former Soviet Union – structure, basin formation mechanisms and subsidence history. Tectonophysics 266: 251–285. Lorenz V and Nicholls IA (1984) Plate and intraplate processes of Hercynian Europe during the late paleozoic. Tectonophysics 107: 25–56. Loup B and Wildi W (1994) Subsidence analysis in the Paris Basin: a key to Northwest European intracontinental basins? Basin Research 6: 45–62. Lundin ER and Dore´ AG (1997) A tectonic model for the Norwegian passive margin with implications for the NE Atlantic: Early Cretaceous to break-up. Journal of the Geological Society 154: 545–550. Manatschal G and Bernoulli D (1999) Architecture and tectonic evolution of nonvolcanic margins: present-day Galicia and ancient Adria. Tectonics 18: 1099–1119. Mareschal J-C and Gliko A (1991) Lithospheric thinning, uplift, and heat flow preceding rifting. Tectonophysics 197: 117–126. Marotta AM, Bayer U, and Thybo H (2000) The legacy of the NE German Basin – Reactivation by compressional buckling. Terra Nova 12: 132–140. Martı´nez del Olmo W (1996) Depositional sequences in the Gulf of Valencia Tertiary Basin. In: Friend PF and Dabiro CJ (eds.) Tertiary Basins of Spain, the Stratigraphic Record of Crustal Kinematics: World and Regional Geology, pp. 55–67. New York: Cambridge University Press. Marx J, Huebscher H-D, Hoth K, Korich D, and Kramer W (1995) Vulkanostratigraphie und Geochemie der Eruptivekomplexe. In: Plein E (ed.) Courier Forschungsinstitut Senkenberg, 183: Norddeutsches Rotliegendbecken, pp. 54–83. Devon: NHBS. Mascle A, Puigdefa`bregas C, Luterbacher HP, and Ferna`andez M (eds.) (1998)Cenozoic Foreland Basins of Western Europe. Geological Society, London, Special Publications, 134: 427p. Mat¸enco L and Bertotti G (2000) Tertiary tectonic evolution of the external East Carpathians (Romania). Tectonophysics 316: 255–286. Mat¸enco L, Bertotti G, Dinu C, and Cloetingh S (1997a) Tertiary tectonic evolution of the external South Carpathians and the adjacent Moesian platform (Romania). Tectonics 16: 896–911. Mat¸enco L, Zoetemeijer R, Cloetingh S, and Dinu C (1997b) Lateral variations in mechanical properties of the Romanian external Carpathians: inferences of flexure and gravity modelling. Tectonophysics 282: 147–166. Mat¸enco L, Bertotti G, Cloetingh S, and Dinu C (2003) Subsidence analysis and tectonic evolution of the external Carpathian–Moesian Platform region during Neogene times. Sedimentary Geology 156: 71–94. Mat¸enco L, Bertotti G, Leever K, et al. (2007) Large scale deformation in a locked collisional boundary: interplay between subsidence and uplift, intraplate stress and inherited lithospheric structure in the late stage of the SE Carpathians evolution. Tectonics submitted.
604
Tectonic Models for the Evolution of Sedimentary Basins
McHargue TR, Heidrick TL, and Livingston J (1992) Episodic structural development of the Central African Rift in Sudan. Tectonophysics 213: 187–202. McKenzie DP (1978) Some remarks on the development of sedimentary basins. Earth and Planetary Science Letters 40: 25–32. Meissner R and Bortfeld RK (eds.) (1990) DEKORP-Atlas, Results of Deutsches Kontinentales Reflexionsseismisches Programm 18p. and 80 plates, Berlin: Springer. Meissner R and Rabbel W (1999) Nature of crustal reflectivity along the DEKORP profiles in Germany in comparison with reflection patterns from different tectonic units worldwide: A review. Earth and Planetary Science Letters 156: 7–28. Mengel K, Sachs PM, Stosch HG, Wo¨rner G, and Look G (1991) Crustal xenoliths from Cenozoic volcanic fields of West Germany: implication for structure and composition of the continental crust. Tectonophysics 195: 271–289. Menning M (1995) A numerical time scale for the Permian and Triassic periods: an integrated time analysis. In: Scholle PA, Peryth TM, and Ulmer-Scholle DS (eds.) The Permian of Northern Pangea, vol. 1, pp. 77–97. Berlin: Springer. Menning M, Weyer D, Drozodzewski G, van Ameron HWJ, and Wendt I (2000) A Carboniferous time scale, 2000: discussion and use of geological parameters as time indicators from Central and Western Europe. Geologisches Jahrbuch A 156: 3–44. Mercier JL, Sebrier M, Lavenu A, et al. (1992) Changes in the tectonic regime above a subduction zone of Andean type: the Andes of Peru and Bolivia during the Pliocene– Pleistocene. Journal of Geophysical Research 97: 11945–11982. Meulenkamp JE, Kova´cˇ M, and Cicha I (1996) On Late Oligocene to Pliocene depocentre migration and the evolution of the Carpathian–Pannonian system. Tectonophysics 266: 301–317. Meyer W and Stets J (1998) Junge Tektonik in Rheinischen Schiefergebirge und ihre Quantifizierung. Zeitschrift der Deutsches Geologisches Gesellschaft 149: 359–379. Millan H, Den Bezemer T, Verges J, et al. (1995) Paleo-elevation and EET evolution at mountain ranges: inferences from flexural modelling in the eastern Pyrenees and Ebro basin. Marine and Petroleum Geology 12: 917–928. Mohr P (1992) Nature of the crust beneath magmatically active continental rifts. Tectonophysics 213: 269–284. Moisio K, Kaikkonen P, and Beekman F (2000) Rheological structure and dynamical response of the DSS profile BALTIC in the SE Fennoscandian shield. Tectonophysics 320: 175–194. Morgan P and Ramberg IB (1987) Physical changes in the lithosphere associated with thermal relaxation after rifting. Tectonophysics 143: 1–11. Mo¨rner N-A (2004) Active faults and palaeoseismicity in Fennoscandia, especially Sweden. Primary structures and secondary effects. Tectonophysics 380: 139–157. Morton AC and Parson LM (1988) Early Tertiary volcanism and the opening of the NE Atlantic. Geological Society, London, Special Publications 39: 477p. Mosar J (2003) Scandinavia’s North Atlantic passive margin. Journal of Geophysical Research 108: 2630. Mosar J, Lewis G, and Torsvik TH (2002) North Atlantic sea-floor spreading rates: implications for the Tertiary development of inversion structures of the Norwegian–Greenland Sea. Journal of the Geological Society 159: 503–515. Munoz JA (1992) Evolution of a continental collision belt: ECORS-Pyrenees crustal balanced cross-section. In: McClay K (ed.) Thrust Tectonics, pp. 235–246. London: Chapman & Hall. Nadin PA and Kusznir NJ (1995) Palaeocene uplift and Eocene subsidence in the northern North Sea basin from 2D forward
and reverse stratigraphic modelling. Journal of the Geological Society 152: 833–848. Nemcok M, Pospisil L, Lexa J, and Donelick RA (1998) Tertiary subduction and slab break-off model of the Carpathian– Pannonian region. Tectonophysics 295: 307–340. Nemes F, Neubauer F, Cloetingh S, and Genser J (1997) The Klagenfurt Basin in the Eastern Alps: An intra-orogenic decoupled flexural basin? Tectonophysics 282: 189–203. Neubauer F, Cloetingh S, Dinu C, and Mocanu V, (eds.) (1997) Tectonics of the Alpine-Carpathian–Pannonian region. tectonophysics tectonophysics, 272: 93–315. Neumann E-R, Olsen KH, and Baldridge S (1995) The Oslo Rift. In: Olsen K H (ed.) Continental Rifts: Evolution, Structure, Tectonics, pp. 345–373. Amsterdam: Elsevier. Neumann E-R, Wilson M, Heeremans M, et al. (2004) Carboniferous-Permian rifting and magmatism in southern Scandinavia, the North Sea and northern Germany: A review. In: Wilson M, Neumann E-R, Davies GR, Timmerman MJ, Heeremans M, and Larsen BT (eds.) Geological Society, London, Special Publications, 223: Permo-Carboniferous Magmatism and Rifting in Europe, pp. 11–40. London: Geological Society, London. Nikishin AM, Cloetingh S, Lobkovsky L, and Burov EB (1993) Continental lithosphere folding in Central Asia (Part I): Constraints from geological observations. Tectonophysics 226: 59–72. Nikishin AM, Ziegler PA, Panov DI, et al. (2001) Mesozoic and Cainozoic evolution of the Scythian Platform–Black Sea– Caucasus domain. In: Ziegler PA, Cavazza W, Robertson AHF, and Crasquin-Soleau S (eds.) Peri-Tethys Mem. 6: PeriTethyan Rift/Wrench Basins and Passive Margins, Me´moires du Museum National d’Histoire Naturelle, 186, pp. 295–346. Paris: Commission for the Geological Map of the World. Nikishin AM, Ziegler PA, Abbott D, Brunet M-F, and Cloetingh S (2002) Permo-Triassic intraplate magmatism and rifting in Eurasia: implications for mantle plumes and mantle dynamics. Tectonophysics 351: 3–39. Nikishin AM, Korotaev MV, Ershov AV, and Brunet M-F (2003) The Black Sea basin: tectonic history and NeogeneQuaternary rapid subsidence modelling. Sedimentary Geology 156: 1–10. Nivie`re B and Winter T (2000) Pleistocene northwards fold propagation of the Jura within the southern Rhine Graben: seismotectonic implications. Global and Planetary Change 27: 263–288. Nottvedt A, Gabrielsen RH, and Steel RJ (1995) Tectonostratigraphy and sedimentary architecture of rift basins with reference to the northern North Sea. Marine and Petroleum Geology 12: 881–901. Odin GS (1994) Geologic time scale. Comptes Rendus de l’Acade´mie des Sciences 318: 59–71. Okaya N, Cloetingh S, and Mueller S (1996) A lithospheric cross section through the Swiss Alps (part II): Constraints on the mechanical structure of a continent–continent collision zone. Geophysical Journal International 127: 399–414. Olesen O, Lundin ER, Nordgulen Ø, et al. (2002) Bridging the gap between the onshore and offshore geology in the Nordland area, northern Norway. Norwegian Journal of Geology 82: 243–262. Olsen KH and Morgan P (1995) Introduction: progress in understanding continental rifts. In: Olsen KH (ed.) Developments in Geotectonics, 25: Continental Rifts: Evolution, Structure, Tectonics, pp. 3–26. Amsterdam: Elsevier. Osmundsen PT, Sommaruga A, Skilbrei JR, and Olesen O (2002) Deep structure of the Norwegian Sea area, North Atlantic margin. Norwegian Journal of Geology 82: 205–224. Oszczypko N (2006) Late Jurassic–Miocene evolution of the Outer Carpathian fold-and-thrust belt and its foredeep basin
Tectonic Models for the Evolution of Sedimentary Basins (Western Carpathians, Poland). Geological Quarterly 50: 168–194. Panza GF (1983) Lateral variations in the European lithosphere and seismic activity. Physics of the Earth and Planetary Interiors 33: 194–197. Parnell J (ed.) (1994)Geofluids: origin migration and evolution of fluids in sedimentary basins. Geological Society, London, Special Publications, 78: 1–372. Parsons T (1995) The Basin and Range Province. In: Olsen K H (ed.) Developments in Geotectonics, 25: Continental Rifts: Evolution, Structure, Tectonics, pp. 227–324. Amsterdam: Elsevier. Pascal C and Gabrielsen RH (2001) Numerical modelling of Cenozoic stress patterns in the mid-Norwegian margin and the northern North Sea. Tectonics 20: 585–599. Pavoni N (1993) Pattern of mantle convection and Pangea break-up, as revealed by the evolution of the African plate. Journal of the Geological Society 150: 953–964. Peper T, Van Balen RT, and Cloetingh S (1994) Implications of orogenics wedge growth intraplates stress variations and sea level change for foreland basin stratigraphy: inferences from numerical modeling. In: Dorobek S and Ross G (eds.) SEPM Special Publications, 52: Stratigraphic development in foreland basins, pp. 25–35. Tulsa: SEPM. Pe´rez-Gussinye´ M and Watts AB (2005) The long-term strength of Europe and its implications for plate forming processes. Nature 436: doi:10.1038/nature03854. Philip H (1987) Plio-Quaternary evolution of the stress field in Mediterranean zones of subduction and collision. Annales Geophysicae 5B: 301–320. Plomerova J, Kouba D, and Babuska V (2002) Mapping the lithosphere-asthenosphere boundary through changes in surface-wave anisotropy. Tectonophysics 358: 175–185. Posgay K, Bodoky T, Hegedu¨s E, et al. (1995) Asthenospheric structure beneath a Neogene basin in southeast Hungary. Tectonophysics 252: 467–484. Price RA (1973) Large scale gravitational flow of supracrustal rocks, southern Canadian Rockies. In: de Jong K and Scholten RA (eds.) Gravity and tectonics, pp. 491–502. New York: Wiley. Prijac C, Doin MP, Gaulier JM, and Guillaucheau F (2000) Subsidence of the Paris Basin and its bearing on the late Variscan lithosphere evolution: A comparison between the Plate and Chablis models. Tectonophysics 323: 1–38. Quinlan G and Beaumont C (1984) Appalachian thrusting lithospheric flexure and the Paleozoic stratigraphy of the eastern interior of North America. Canadian Journal of Earth Sciences 21: 973–996. Ra˘ dulescu F (1988) Seismic models of the crustal structure in Romania. Revue Roumaine de Ge´ologie, Ge´ophysique et Ge´ographie – Se´rie de Ge´ophysique 32: 13–17. Ranalli G (1995) Rheology of the Earth, 2nd Ed. 413p. London: Chapman and Hall. Ranalli G and Murphy DC (1987) Rheological stratification of the lithosphere. Tectonophysics 132: 281–295. Redfield TF, Osmundsen PT, and Hendriks BWH (2005) The role of fault reactivation and growth in the uplift of western Fennoscandia. Journal of the Geological Society 162: 1013–1030. Reemst P (1995) Tectonic modeling of rifted continental margins; Basin evolution and tectono-magmatic development of the Norwegian and NW Australian margin. PhD thesis, Vrije Universiteit, Amsterdam, 163p. Reemst P and Cloetingh S (2000) Polyphase rift evolution of the Vøring margin (mid-Norway): constraints from forward tectonostratigraphic modeling. Tectonics 19: 225–240. Reilinger RE, McClusky SC, Oral MB, et al. (1997) Global positioning system measurements of present-day crustal
605
movements in the Arabia-Africa-Eurasia plate collision zone. Journal of Geophysical Research 102: 9983–9999. Reston TJ (1990) The lower crust and the extension of the continental lithosphere; kinematic analysis of BIRPS deep seismic data. Tectonics 9: 1235–1248. Riba O, Reguant S, and Villena J (1983) Ensayo de sı´ntesis estratigra´fica y evolutiva de la cuenca terciaria del Ebro. Geologı´a de Espan˜a, Libro Jubilar J. M. Rios, vol. II, pp. 131–159. Madrid: Instituto Geolo´gico y Minero de Espan˜a. Ribeiro A, Baptista JC, and Matias L (1996) Tectonic stress pattern in Portugal mainland and the adjacent Atlantic region (West Iberia). Tectonics 15: 641–659. Richardson RM (1992) Ridge forces, absolute plate motions and the intra-plate stress field. Journal of Geophysical Research 97: 11739–11748. Richter F and McKenzie D (1978) Simple plate models of mantle convection. Journal of Geophysics 44: 441–478. Ritter JRR, Achauer U, and Christensen UR (2000) The teleseismic tomography experiment in the Eifel region, central Europe: design and first results. Seismological Research Letters 71: 437–443. Ritter JRR, Jordan M, Christensen UR, and Achauer U (2001) A mantle plume below the Eifel volcanic fields, Germany. Earth and Planetary Science Letters 186: 7–14. Roberts AM, Yielding G, Kusznir NJ, Walker I, and Don-Lopez D (1993) Mesozoic extension in the North Sea: constraints from flexural backstripping, forward modelling and fault populations. In: Parker JR (ed.) Petroleum Geology of Northwest Europe, Proceedings of the 4th Conference, pp. 1123–1136. London: Geological Society, London. Robin C, Allemand P, Burov E, et al. (2003) Vertical movements of the Paris Basin (Triassic- Pleistocene): From 3D stratigraphic database to numerical models. In: Nieuwland DA (ed.) Geological Society, London, Special Publications, 212: New Insights in Structural Interpretation and Modelling, pp. 225–250. London: Geological Society, London. Robinson A, Spadini G, Cloetingh S, and Rudat J (1995) Stratigraphic evolution of the Black Sea: Inferences from basin modelling. Marine and Petroleum Geology 12: 821–836. Roca E (2001) The northwest Mediterranean Basin (Valencia Trough, Gulf of Lions and Liguro-Provencal basins): Structure and geodynamic evolution. In: Ziegler PA, Cavazza W, Robertson AHF, and Crasquin-Soleau S (eds.) Peri-Tethys Memoir 6: Peri-Tethyan Rift/Wrench Basins and Passive Margins, Me´moires du Museum National d’Histoire Naturelle, 186, pp. 671–706. Paris: Commission for the Geological Map of the World. Roca E and Desegaulx P (1992) Analysis of the geological evolution and vertical movements in the Valencia Trough (western Mediterranean). Marine and Petroleum Geology 9: 167–185. Roca E, Sans M, Cabrera L, and Marzo M (1999) Oligocene to middle Miocene evolution of the central Catalan margin (northwestern Mediterranean). Tectonophysics 315: 209–233. Rohrman M and Van der Beek PA (1996) Cenozoic postrift domal uplift of North Atlantic margins; an asthenospheric diapirism model. Geology 24: 901–904. Rohrman M, Van der Beek PA, Andriessen PAM, and Cloetingh S (1995) Meso-Cenozoic morphotectonic evolution of Southern Norway: Neogene domal uplift inferred from apatite fissiontrack thermochronology. Tectonics 14: 704–718. Rohrman M, Van der Beek PA, Van der Hilst RD, and Reemst P (2002) Timing and Mechanism s of North Atlantic Cenozoic Uplift: Evidence for mantle upwelling. Geological Society, London, Special Publications 196: 27–43.
606
Tectonic Models for the Evolution of Sedimentary Basins
Rosenbaum G, Lister GS, and Duboz C (2002) Relative motion of Africa, Iberia and Europe during Alpine orogeny. Tectonophysics 359: 117–129. Roure F and Sassi W (1995) Kinematics of deformation and petroleum system appraisal in Neogene foreland fold-andthrust belts. Petroleum Geoscience 1: 253–269. Roure F, Roca E, and Sassi W (1993) The Neogene evolution of the outer Carpathian flysch units (Poland, Ukraine and Romania): Kinematics of a foreland/fold-and-thrust belt system. Sedimentary Geology 86: 177–201. Roure F, Brun J-P, Colletta B, and Vially R (1994) Multiphase extensional structures fold-reactivation and petroleum plays in the Alpine foreland basin of southeastern France. In: Mascle A (ed.) Special Publication of the European Association of Petroleum Geoscientists 4: Exploration and Petroleum Geology of France, pp. 237–260. Berlin: Springer. Roure F, Choukroune P, and Polino R (1996a) Deep seismic reflection data and new insights on the bulk geometry of mountain ranges. Comptes Rendus de l’Acade´mie des Sciences 322(IIa): 345–359. Roure F, Shein VS, Ellouz N, and Skvortsov L (eds.) (1996b) Geodynamic Evolution of Sedimentary Basins, pp. 1–453. Paris: Editions Technip. Rowley DB and Sahagian D (1986) Depth-dependent stretching: a different approach. Geology 14: 32–35. Royden L (1988) Flexural behaviour of the continental lithosphere in Italy: constraints imposed by gravity and deflection data. Journal of Geophysical Research 93: 7747–7766. Royden LH (1993) The tectonic expression of the slab pull at continental convergent boundaries. Tectonics 12: 303–325. Royden LH and Do¨ve´nyi P (1988) Variations in extensional styles at depth across the Pannonian basin system. In: Royden LH and Horva´th F (eds.) 45: The Pannonian Basin: A Case Study in Basin Evolution, pp. 235–255. Tulsa: American Association of Petroleum Geologists. Royden LH and Horva´th F (eds.) (1988)The Pannonian Basin, A Study in Basin Evolution. American Association of Petroleum Geologists Memoirs, 45: 394p. Royden LH and Karner GD (1984) Flexure of lithosphere beneath the Apennine and Carpathian foredeep basins: Evidence for an insufficient topographic load. Am Assoc Petrol Geol Buln 68: 704–712. Royden L and Keen CE (1980) Rifting process and thermal evolution of the continental margin of eastern Canada determined from subsidence curves. Earth and Planetary Science Letters 51: 343–361. Royden L, Sclater JG, and Herzen RP (1980) Continental margin subsidence and heat flow; important parameters in formation of petroleum hydrocarbons. American Association of Petroleum Geologists Bulletin 62: 173–187. Royden LH, Horva´th F, Nagymarosy A, and Stegena L (1983) Evolution of the Pannonian Basin System. 2. Subsidence and thermal history. Tectonics 2: 91–137. Rutigliano P and VLBI Network Team (2000) Vertical motions in the western Mediterranean area from geodetic and geological data. Boll. Roa, 3, 10th General Assembly Wegener Project, Extended Abstract Sacchi M, Horva´th F, and Magyari O (1999) Role of unconformity-bounded units in the stratigraphy of the continental record: A case study from the Late Miocene of the western Pannonian Basin, Hungary. In: Durand B, Jolivet L, Horva´th F, and Se´ranne M (eds.) Geological Society, London, Special Publications, 156: The Mediterranean Basins: Tertiary Extension within the Alpine Orogen, pp. 357–390. London: Geological Society, London. Sachsenhofer RF, Lankreijer A, Cloetingh S, and Ebner F (1997) Subsidence analysis and quantitative basin modelling in the Styrian basin (Pannonian Basin System, Austria). Tectonophysics 272: 175–196.
Salveson JO (1976) Variations in the oil and gas geology of rift basins. Egyptian General Petroleum Corp, 5th Explor Sem, Cairo, Egypt, 15–17 November, 1976 Sanders CAE, Andriessen PAM, and Cloetingh SAPL (1999) Life cycle of the East Carpathian orogen: Erosion history of a doubly vergent critical wedge assessed by fission track thermochronology. Journal of Geophysical Research 104: 29095–29112. Sa˘ ndulescu M (1984) Geotectonics of Romania (in Romanian) Bucharest: Editions Tehnica. Sa˘ ndulescu M (1988) Cenozoic tectonic history of the Carpathians. In: Royden LH and Horvath F (eds.) American Association of Petroleum Geologists Memoirs, 45: The Pannonian Basin, A Study in Basin Evolution, pp. 17–25. Tulsa: American Association of Petroleum Geologists. Sanz de Siria Catalan A (1993) Datos sobre la paleoclimatologia y paleoecologia del Neogeno del Valles-Penedes segun las macrofloras halladas en la cuenca y zonas proximas. Paleontologia i Evolucio´ 26–27: 281–289. Sassi W, Colletta B, Bale P, and Paquereau T (1993) Modeling of structural complexity in sedimentary basins: The role of preexisting faults in thrust tectonics. Tectonophysics 226: 97–112. Sawyer DS and Harry DL (1991) Dynamic modeling of divergent margin formation: application to the US Atlantic margin. In: Meyer AW, Davies TA, and Wise SW (eds.) Marine Geology 102: Evolution of Mesozoic and Cenozoic Continental Margins, pp. 29–42. Amsterdam: Elsevier. Schmid SM, Pfiffner OA, Froitzheim N, Scho¨nborn G, and Kissling E (1996) Geophysical-geological transect and tectonic evolution of the Swiss–Italian Alps. Tectonics 12: 1036–1064. Schmid SM, Berza T, Diaconescu V, Froitzheim N, and Fuegenschuh B (1998) Orogen-parallel extension in the South Carpathians during the Paleogene. Tectonophysics 297: 209–228. Schmid SM, Fu¨genschuh B, Kissling E, and Schuster R (2004) Tectonic map and overall architecture of the Alpine orogen. Eclogae Geologicae Helvetiae 97: 93–117. Schmid SM and Mat¸enco L (2007) An integrated cross section through the Carpathian system. Tectonics submitted. Schlumberger (1991) World oil reserves – charting the future. Middle East Well Evaluation Review 10: 7–15. Sclater JJG and Christie PAF (1980) Continental stretching: an explanation for the post mid-Cretaceous subsidence of the central North Sea basin. Journal of Geophysical Research 85: 3711–3739. Sclater J, Royden L, Horva´th F, Burchfiel B, Semken S, and Stegena L (1980) The formation of the intra-Carpathian basins as determined from subsidence data. Earth and Planetary Science Letters 51: 139–162. Seber D, Barazangi M, Ibenbrahim A, and Demnati A (1996) Geophysical evidence for lithospheric delamination beneath the Alboran Sea and Rif-Betic mountains. Nature 379: 785–790. Sengo¨r AMC and Burke K (1978) Relative timing of rifting and volcanism on Earth and its tectonic implications. Geophysical Research Letter 5: 419–421. Seyfert M and Henk A (2000) Deformation, metamorphism and exhumation: quantitative models for a collision zone in the Variscides. In: Franke W, Haak V, Oncken O, and Tanner D (eds.) Orogenic Processes: Quantification and Modelling in the Variscan Belt, Special Publication, 179, 217–230. London: Geological Society. Shudofsky GN, Cloetingh S, Stein S, and Wortel MJR (1987) Unusually deep earthquakes in east Africa: constraints on the thermo-mechanical structure of a continental rift system. Geophysical Research Letters 14: 741–744. Sibuet J-C, Srivastava SP, and Spakman W (2004) Pyrenean orogeny and plate kinematics. Journal of Geophysical Research 109: B08104 (doi:10.1029/2003JB002514).
Tectonic Models for the Evolution of Sedimentary Basins Sinclair HD and Allen PA (1992) Vertical vs. horizontal motions in the Alpine orogenic wedge: Stratigraphic response in the foreland basin. Basin Research 4: 215–232. Skogseid J and Eldholm O (1995) Rifted continental margin off mid-Norway. In: Banda E, Talwani E, and Torne´ M (eds.) Rifted Ocean-Continent Boundaries, pp. 147–153. Dordrecht: Kluwer Academic. Skogseid J, Pedersen T, and Larsen VB (1992) Vøring Basin: subsidence and tectonic evolution. In: Larsen RM, Brekke H, Larsen BT, and Talleraas E (eds.) Norwegian Petroleum Society, Special Publications 1: Structural and Tectonic Modelling and Its Application to Petroleum Geology, pp. 55–82. Amsterdam: Elsevier. Skogseid J, Planke S, Faleide JI, Pedersen T, Eldholm O, and Neverda Fl (2000) NE Atlantic continental rifting and volcanic margin formation. Geological Society, London, Special Publications 167: 295–326. Sleep NH (1971) Thermal effects of the formation of Atlantic continental margins by continental break up. Geophysical Journal of the Royal Astronomical Society 24: 325–350. Sleep NH (1973) Crustal thinning on Atlantic coastal margins: evidence from old margins. In: Tarling DH and Runcorn SK (eds.) Implications of Continental Drift to Earth Sciences, Part 6, vol. 2, pp. 685–692. London: Academic Press. Smolyaninova EI, Mikhailov VO, and Lyakhovsky VA (1996) Numerical modelling of regional neotectonic movements in the northern Black Sea. Tectonophysics 266: 221–231. Sobel ER, Hilley GE, and Strecker MR (in press) Formation of internally drained contractional basins by aridity-limited bedrock incision. Journal of Geophysical Research 108(B7), 2344, doi:10.1029/2002JB001883. Sobolev SV, Zeyen H, Granet M, et al. (1997) Upper mantle temperatures and lithosphere-asthenosphere system beneath the French Massif Central constrained by seismic, gravity, petrologic and thermal observations. Tectonophysics 275: 143–164. Solheim A, Riis F, Elverhoi A, Faleide JJ, Jensen LN, and Cloetingh S, (eds.) (1996) Impact of glaciations on basin evolution and models for the Norwegian margin and adjacent areas. Global Planet Change Global Planet Change, 12: 1–450. Sonder LJ and England PC (1989) Effects of a temperaturedependent rheology on large-scale continental extension. Journal of Geophysical Research 94: 7603–7619. Spadini G (1996) Lithospheric deformation and vertical motions in back-arc Mediterranean basins: The Black Sea and the Tyrrhenian Sea. PhD Thesis, Vrije Universiteit, Amsterdam, 152p. Spadini G and Podladchikov Y (1996) Spacing of consecutive normal faulting in the lithosphere: a dynamic model for rift axis migration. Earth and Planetary Science Letters 144: 21–34. Spadini G, Bertotti G, and Cloetingh S (1995a) Tectonostratigraphic modelling of the Sardinian margin of the Tyrrhenian Sea. Tectonophysics 252: 253–268. Spadini G, Cloetingh S, and Bertotti G (1995b) Thermomechanical modeling of the Tyrrhenian Sea: Lithospheric necking and kinematics of rifting. Tectonics 14: 629–644. Spadini G, Robinson A, and Cloetingh S (1996) Western versus eastern Black Sea tectonic evolution: Pre-rift lithospheric controls on basin formation. Tectonophysics 266: 139–154. Spadini G, Robinson A, and Cloetingh S (1997) Thermomechanical modelling of Black Sea basin formation, subsidence and sedimentation. In: Robinson A (ed.) American Association of Petroleum Geologists Memoirs, 68: Regional and Petroleum Geology of the Black Sea and Surrounding Areas, pp. 19–38. Tulsa: American Association of Petroleum Geologists. Spakman W and Wortel R (2004) A tomographoic view of the Western Mediterranean Geodynamics. In: Cavvaza W, Roure F, Spakman W, Stampfli GM, and Ziegler PA (eds.)
607
The TRANSMED Atlas – The Mediterranean Region from Crust to Mantle and CD-ROM. pp. 31–52. Berlin: Springer. Stampfli GM and Borel GD (2004) The TRANSMED transects in space and time: constraints on the paleotectonic evolution of the Mediterranean domain. In: Cavazza W, Roure F, Spakman W, Stampfli GM, and Ziegler PA (eds.) The TRANSMED Atlas – The Mediterranean Region from Crust to Mantle, pp. 53–80. Berlin: Springer. Stampfli GM, Mosar J, Marquer D, Marchant R, Baudin T, and Borell G (1998) Subduction and obduction processes in the Swiss Alps. Tectonophysics 296: 159–204. Stampfli GM, Mosar J, Favre P, Pillevuit A, and Vannay J-C (2001) Permo-Mesozoic evolution of the western Tethys realm: the Neo-Tethys East-Mediterranean connection. In: Ziegler PA, Cavazza W, Robertson AHF, and CrasquinSoleau S (eds.) Memoires du Museum National d’Histoire Naturelle 186: Peri-Tethys Memoir 6: Peri-Tethyan Rift/ Wrench Basins and Passive Margins, pp. 51–108. Paris: Commission for the Geological Map of the World. Stampfli GM, Borel GD, Marchant R, and Mosar J (2002) Western Alps geological constraints on western Tethyan reconstructions. Journal of the Virtual Explorer 8: 77–106. Stapel G (1999) The Nature of Isostasy in Western Iberia. PhD thesis, Vrije Universiteit, Amsterdam, 148p. Stapel G, Cloetingh S, and Pronk B (1996) Quantitative subsidence analysis of the Mesozoic evolution of the Lusitanian Basin (Western Iberian margin). Tectonophysics 266: 493–507. Steckler MS and Ten Brink US (1986) Lithospheric strength variations as a control on new plate boundaries: Examples from the northern Red Sea region. Earth and Planetary Science Letters 79: 120–132. Steckler MS and Watts AB (1978) Subsidence of the Atlantictype continental margin off New York. Earth and Planetary Science Letters 41: 1–13. Steckler MS and Watts AB (1982) Subsidence history and tectonic evolution of Atlantic-type continental margins. In: Scrutton RA (ed.) Geodynamics Series 6: Dynamics of Passive Margins, pp. 184–196. Washington, DC: Ameican Geophysical Union. Stegena L (1967) The formation of the Hungarian basin. Fo¨ldtani Ko¨zlo¨ny 97: 278–285 (in Hungarian). Stegena L, Ge´czy B, and Horva´th F (1975) Late Cenozoic evolution of the Pannonian basin. Fo¨ldtani Ko¨zlo¨ny 105: 101–123 (in Hungarian). Stel H, Cloetingh S, Heeremans M, and van Beek P (1993) Anorogenic granites, magmatic underplating and the origin of intracratonic basins in non-extensional settoing. Tectonophysics 226: 285–299. Stephenson RA (1989) Beyond first-order thermal subsidence models for sedimentary basins? In: Cross TA (ed.) Quantitative Dynamic Stratigraphy, pp. 113–125. Englewood Cliffs, NJ: Prentice-Hall. Stephenson RA and Cloetingh S (1991) Some examples and mechanical aspects of continental lithospheric folding. Tectonophysics 188: 27–37. Stephenson RA, Stovba S, and Starostenko V (2001) PripyatDniepr-Donets basin: implications for dynamics of rifting and tectonic history of the northern Peri-Tethyan platform. In: Ziegler PA, Cavazza W, Robertson AHF, and CrasquinSoleau S (eds.) Me´moires du Museum National d’Histoire Naturelle 186: Peri-Tethys Memoir 6, Peri-Tethyan Rift/ Wrench Basins and Passive Margins, pp. 369–406. Paris: Commission for the Geological Map of the World. Stockmal GS, Beaumont C, and Boutillier R (1986) Geodynamic models of convergent margin tectonics: transition from rifted margin to overthrust belt and consequences for forelandbasin development. American Association of Petroleum Geologists Bulletin 70: 181–190.
608
Tectonic Models for the Evolution of Sedimentary Basins
Straume A˚K and Austrheim H (1999) Importance of fracturing during retro-metamorphism of eclogites. Journal of Metamorphic Geology 7: 637–652. Suhadolc P and Panza GF (1989) Physical properties of the lithosphere–asthenosphere system in Europe from geophysical data. In: Boriani A, Bonafede M, Piccardo GB, and Vai GB (eds.) The Lithosphere in Italy – Advances in Earth Science Research, vol 80, pp. 15–40. Rome: Accademia dei Lincei. Suvorov VD, Mishenkina ZM, Petrick G, Sheludko IF, Seleznev VS, and Solovyov VM (2002) Structure of the crust in the Baikal rift zone and adjacent areas from Deep Seismic Sounding data. Tectonophysics 151: 61–74. Sza´deczky-Kardoss E (1967) Latest results of geological research in Hungary and its prospect in the light of international development. Geolo´gia e´s Ba´nya´szat 1: 5–25 (in Hungarian). Szafia´n P, Horva´th F, and Cloetingh S (1997) Gravity constraints on the crustal structure and slab evolution along a transcarpathian transect. Tectonophysics 272: 233–247. Tait JA, Bachtadse V, Franke W, and Soffel HC (1997) Geodynamic evolution of the European Variscan fold belt; paleomagnetic and geological constraints. Geologische Rundschau 86: 585–598. Ta˘ ra˘ poanca˘ M, Bertotti G, Mat¸enco L, Dinu C, and Cloetingh S (2003) Architecture of the Focsani depression: A 13 km deep basin in the Carpathians bend zone (Romania). Tectonics 22: 1074 (doi:10.1029/2002TC001486). Ta˘ ra˘ poanca˘ M, Garcia-Castellanos D, Bertotti G, Mat¸enco L, Cloetingh S, and Dinu C (2004a) Role of 3-D distributions of load and lithospheric strength in orogenic arcs: polystage subsidence in the Carpathians foredeep. Earth and Planetary Science Letters 221: 163–180. Tari G, Dicea O, Faulkerson J, Georgiev G, Popov S, Stefanescu M, and Weir G (1997) Cimmerian and Alpine stratigraphy and structural evolution of the Moesian platform (Romania/Bulgaria). In: Robinson AG (ed.) Regional and petroleum geology of the Black Sea and surrounding regions, pp. 63–90. Tulsa: Am Assoc Petrol Geol. Tari GD, o¨ve´nyi P, Dunkl I, et al. (1999) Lithospheric structure of the Pannonian basin derived from seismic, gravity and geothermal data. In: Durand B, Jolivet L, Horva´th F, and Se´ranne M (eds.) Geological Society London, Special Publications, 156: The Mediterranean Basins: Tertiary Extension within the Alpine Orogen, pp. 215–250. London: Geological Society, London. Tavares Martins L (1998) Continental tholeiitic magmatism of Algarve, southern Portugal; an examples of in situ crustal contamination. Comunicacoes do Instituto Geologico e Mineiro, Lisbon, Portugal 85: 99–116. Ter Voorde M and Bertotti G (1994) Thermal effects of normal faulting during rifted basin formation, 1. A finite difference model. Tectonophysics 240: 133–144. Ter Voorde M and Cloetingh S (1996) Numerical modelling of extension in faulted crust: effects of localized and regional deformation on basin stratigraphy. Geological Society, London, Special Publications 99: 283–296. Ter Voorde M, Revnas ER, Faerseth R, and Cloetingh S (1997) Tectonic modelling of the Middle Jurasicsyn-rift stratigraphy in the Oseberg-Brage area, Southern Viking Graben. Basin Research 9: 133–150. Ter Voorde M, Van Balen RT, Bertotti G, and Cloetingh SAPL (1998) The influence of a stratified rheology on the flexural response of the lithosphere to (un-)loading by extensional faulting. Geophysical Journal International 134: 721–735. Ter Voorde M, de Bruijne K, Andriessen P, and Cloetingh S (2004) Thermal consequences of thrust faulting: simultaneous versus successive fault activation and exhumation. Earth and Planetary Science Letters 223(3–4): 395–413. Torne´ M, Fernandez M, Wheeler W, and Karpuz R (2003) Three dimensional crustal structure of the Voring Margin (NE
Atlantic): A combined seismic and gravity image. Journal of Geophysical Research 108: 2115. Torsvik TH, Van der Voo R, Meert JG, Mosar J, and Walderhaug HJ (2001) Reconstructions of continents around the North Atlantic at about the 60th parallel. Earth and Planetary Science Letters 187: 55–69. To´th L, Mo´nus P, Zsı´ros T, and Kiszely M (2002) Seismicity in the Pannonian Region – earthquake data. In: Cloetingh S, Horva´th F, Bada G, and Lankreijer A (eds.) EGU St. Mueller Special Publication Series, 3: Neotectonics and Surface Processes: The Pannonian Basin and Alpine/Carpathian System pp. 9–28. EGU. Tucker GE and Slingerland RL (1994) Erosional dynamics, flexural isostasy, and long-lived escarpments: A numerical modeling study. Journal of Geophysical Research 99: 12229–12243. Tucker GE and Slingerland RL (1996) Predicting sediment flux from fold and thrust belts. Basin Research 8: 329–349. Turcotte DL and Emermann SH (1983) Mechanisms of active and passive rifting. Tectonophysics 94: 39–50. Turcotte DL and Schubert G (1982) Geodynamics. New York: John Wiley. Underhill JR and Partington MA (1993) Jurassic thermal doming and deflation in the North Sea: implications of the sequence stratigraphic evidence. In: Parker JR (ed.) Petroleum Geology of Northwest Europe. Proceedings of the 4th Conference, pp. 337–345. London: Geological Society. Uyeda S and McCabe R (1983) A possible mechanism of episodic spreading of the Philippine Sea. In: Hashimoto M and Uyeda S (eds.) Accretion Tectonics in the Circum-Pacific Regions, Dordrecht: D. Reidel Publication, pp. 291–306. Tokyo: Terra Scientific Publication. Va˚gnes E, Gabrielsen RH, and Haremo P (1998) Late Cretaceous-Cenozoic intraplate contractional deformation at the Norwegian continental shelf: timing, magnitude and regional implications. Tectonophysics 300: 29–46. Vakarcs G, Vail PR, Tari G, Poga´csa´s Gy, Mattick RE, and Szabo´ A (1994) Third-order Miocene-Pliocene depositional sequences in the prograding delta complex of the Pannonian basin. Tectonophysics 240: 81–106. Van Balen RT and Cloetingh S (1993) Intraplate stresses and fluid flow in extensional basins. American Association of Petroleum Geologists Studies in Geology 34: 87–98. Van Balen RT and Cloetingh S (1994) Tectonic control of the sedimentary record and stress-induced fluid flow: constraints from basin modeling. Geological Society, London, Special Publications 78: 9–26. Van Balen RT and Cloetingh S (1995) Neural network analyses of stress-induced overpressures in the Pannonian basin. Geophysical Journal International 121: 532–544. Van Balen RT, Van der Beek PA, and Cloetingh S (1995) The effect of rift shoulder erosion on stratal patterns at passive margins: implications for sequence stratigraphy. Earth and Planetary Science Letters 134: 527–544. Van Balen RT, Podladchikov Y, and Cloetingh S (1998) A new multi-layered model for intraplate stress-induced differential subsidence of faulted lithosphere, applied to rifted basins. Tectonics 17: 938–954. Van Balen R, Lenkey L, Horva´th F, and Cloetingh S (1999) Two-dimensional modelling of stratigraphy and compaction-driven fluid flow in the Pannonian basin. In: Durand B, Jolivet L, Horva´th F, and Se´ranne M (eds.) Geological Society, London, Special Publications, 156: The Mediterranean Basins: Tertiary Extension within the Alpine Orogen, pp. 391–414. London: Geological Society, London. Van Balen RT, Houtgast RF, Van der Wateren FM, Vandenberghe J, and Bogaart PW (2000) Sediment budget
Tectonic Models for the Evolution of Sedimentary Basins and tectonic evolution of the Meuse catchment in the Ardennes and the Roer Valley Rift System. Global and Planetary Change 27: 113–129. Van der Beek PA and Braun J (1998) Numerical modelling of landscape evolution on geological time-scales; a parameter analysis and comparison with the south-eastern highlands of Australia. Basin Research 10: 49–68. Van der Beek P and Braun J (1999) Controls on post-MidCretaceous landscape evolution in the southeastern highlands of Australia; insights from numerical surface process models. Journal of Geophysical Research 104: 4945–4966. Van der Beek PA and Cloetingh S (1992) Lithospheric flexure and the tectonic evolution of the Betic Cordillers. Tectonophysics 203: 325–344. Van der Beek PA, Cloetingh S, and Andriessen PAM (1994) Extensional basin formation mechanisms and vertical motion of rift flanks: Constraints from tectonic modelling and fissiontrack thermochronology. Earth and Planetary Science Letters 121: 417–433. Van der Beek PA, Andriessen P, and Cloetingh S (1995) Morphotectonic evolution of rifted continental margins: inferences from a coupled tectonic-surface processes model and fission-track thermochronology. Tectonics 14: 406–421. Van Vliet-Lanoe¨ B, Laurent M, Everaerts M, Mansy J-L, and Manby G (2000) Evolution Neogene et quarternaire de la Somme, une flexuration tectonique active. Comptes Rendus Academie des Sciences, Earth and Planetary Sciences 331: 151–158. Van Wees J-D (1994) Tectonic Modelling of Basin Deformation and Inversion Dynamics: The Role of Pre-existing Faults and Continental Lithosphere Rheology in Basin Evolution. PhD Thesis, Vrije Universiteit, Amsterdam, 164p. Van Wees JD and Beekman F (2000) Lithosphere rheology during intraplate basin extension and inversion: Inferences from automated modelling of four basins in western Europe. Tectonophysics 320: 219–242. Van Wees J-D and Cloetingh S (1994) A finite-difference technique to incorporate spatial variations in rigidity and planar faults into 3-D models for lithospheric flexure. Geophysical Journal International 117: 179–195. Van Wees J-D and Cloetingh S (1996) 3D flexure and intraplate compression in the North Sea basin. Tectonophysics 266: 343–359. Van Wees J-D and Stephenson RA (1995) Quantitative modelling of basin and rheological evolution of the Iberian Basin (Central Spain): Implications for lithosphere dynamics of intraplate extension and inversion. Tectonophysics 252: 163–178. Van Wees J-D, de Jong K, and Cloetingh S (1992) Twodimensional P-T-t modelling and dynamics of extension and inversion in the Betic Zone (SE Spain). Tectonophysics 203: 305–324. Van Wees J-D, Stephenson RA, Stovba SM, and Shymanovsky VA (1996) Tectonic variation in the Dniepr– Donets Basin from automated modelling and backstripped subsidence curves. Tectonophysics 268: 257–280. Van Wees JD, Arche A, Beijdorff CG, Lopez-Gomez J, and Cloetingh S (1998) Temporal and spatial variations in tectonic subsidence in the Iberian basin (eastern Spain): Inferences from automated forward modelling of highresolution stratigraphy. (Permian–Mesozoic). Tectonophysics 300: 285–310. Van Wees J-D, Stephenson RS, Ziegler PA, et al. (2000) On the origin of the Southern Permian Basin, Central Europe. Marine and Petroleum Geology 17: 43–59. Van Wijk JW and Cloetingh S (2002) Basin migration caused by slow lithospheric extension. Earth and Planetary Science Letters 198: 275–288.
609
Van Wijk JW, Huismans RS, Ter Voorde M, and Cloetingh S (2001) Melt generation at volcanic continental margins: no need for a mantle plume? Geophysical Research Letters 28: 3995–3998. Vanderhaeghe O and Teyssier C (2001) Partial melting and flow of orogens. Tectonophysics 342: 451–472. Vasiliev I, Krijgsman W, Langereis CG, Panaiotu CE, Mat¸enco L, and Bertotti G (2004) Magnetostratigraphyc dating and astronomical forcing of the Mio-Pliocene sedimentary sequences of the Focsani basin (Eastern Paratethys – Romania). Earth and Planetary Science Letters 227: 231–247. Vauchez A, Tommasi A, and Barruol G (1998) Rheological heterogeneity, mechanical anisotropy and deformation of the continental lithosphere. Tectonophysics 296: 61–86. Verall P (1989) Speculations on the Mesozoic–Cenozoic tectonic history of the Western United States. In: Tankard AJ and Balkwill HR (eds.) American Association of Petroleum Geologists Memoirs, 46: Extensional Tectonics and Stratigraphy of the North Atlantic Margins, pp. 615–631. Tulsa: American Association of Petroleum Geologists. Verge´s JM and Garcia-Senez J (2001) Mesozoic evolution and Cainozoic inversion of the Pyrenean rift. In: Ziegler PA, Cavazza W, Robertson AHF, and Crasquin-Soleau S (eds.) Me´moires du Museum National d’Histoire Naturelle 186: Peri-Tethys Memoir 6, Peri-Tethyan Rift/Wrench Basins and Passive Margins, pp. 187–212. Paris: Commission for the Geological Map of the World. Verge´s J and Sa`bat F (1999) Constraints on the western Mediterranean kinematics evolution along a 1000-km transect from Iberia to Africa. In: Durand B, Jolivet L, Horva´th F, and Seranne M (eds.) Geological Society, London, Special Publications, 156: The Mediterranean Basins: Tertiary Extension within the Alpine Orogen, pp. 63–80. London: Geological Society, London. Verge´s J, Millan H, Roca E, et al. (1995) Evolution of a collisional orogen: eastern Pyrenees transect and petroleum potential. Marine and Petroleum Geology 12: 903–916. Verge´s JM, Marzo M, Santaeularia T, Serra-Kiel J, Burbank DW, Munoz JA, and Gimenez-Montsant J (1998) Quantified vertical motions and tectonic evolution of the SE Pyrenean foreland basin. Geological Society, London, Special Publications 134: 107–134. Vigneresse JL (1999) Intrusion level of granitic massifs along the Hercynian belt: balancing the eroded crust. Tectonophysics 307: 277–295. Vyssotski AV, Vyssotski NV, and Nezhdanov AA (2006) Evolution of the West Siberian Basin. Marine and Petroleum Geology 23: 93–126. Waltham D, Docherty C, and Taberner C (2000) Decoupled flexure in the South Pyrenean foreland. Journal of Geophysical Research 105: 16329–16340. Watcharanantakul R and Morley CK (2000) Syn-rift and postrift modelling of the Pattani basin, Thailand: evidence for rampflat detachment. Marine and Petroleum Geology 17: 937–958. Watts AB (2001) Isostasy and Flexure of the Lithosphere, 458p. Cambridge: Cambridge University Press. Watts AB and Burov EB (2003) Lithospheric strength and its relationship to the elastic and seismogenic layer thickness. Earth and Planetary Science Letters 213: 113–131. Watts AB and Fairhead JD (1997) Gravity anomalies and magmatism along the western continental margin of the British Isles. Journal of the Geological Society 154: 523–529. Watts AB, Karner GD, and Steckler MS (1982) Lithospheric flexure and the evolution of sedimentary basins. Kent P, Bott MHP, McKenzie DP, and Williams CA (eds.) The Evolution of Sedimentary Basins. Philosophical Transactions of The Royal Society A, 305, pp. 249–281.
610
Tectonic Models for the Evolution of Sedimentary Basins
Watts AB, Platt J, and Buhl P (1993) Tectonic evolution of the Alboran Sea basin. Basin Research 5: 153–177. Wenzel F, Lorenz F, Sperner B, and Oncescu MC (1999) Seismotectonics of the Romanian Vrancea area. In: Wenzel F, Lungu D, and Novak O (eds.) Vrancea Earthquakes: Tectonics, Hazard and Risk Mitigation, pp. 15–26. Dordrecht: Kluwer Academic. Wenzel F, Sperner B, Lorenz F, and Mocanu V (2002) Geodynamics, tomographic images and seismicity of the Vrancea region (SE-Carpathians, Romania). EGU Stephan Mueller Special Publication Series 3: 95–104. Wernicke B (1985) Uniform-sense normal simple shear of the continental lithosphere. Canadian Journal of Earth Sciences 22: 108–125. Wernicke B (1990) The fluid crustal layer and its implication for continental dynamics. In: Salisbury MH and Fountein DM (eds.) Exposed Cross-Sections of the Continental Crust, pp. 509–544. Dordrecht: Kluwer Academic. Whipple KX and Tucker GE (1999) Dynamics of the streampower river incision model; implications for height limits of mountain ranges, landscape response timescales, and research needs. Journal of Geophysical Research 104: 17661–17674. White N and McKenzie D (1988) Formation of the ‘steer’s head’ geometry of sedimentary basins by differential stretching of the crust and mantle. Geology 16: 250–253. White RS and McKenzie DP (1989) Volcanism at rifts. Scientific American 201: 44–59. Whittaker A, Bott MHP, and Waghorn GD (1992) Stresses and plate boundary forces associated with subduction plate margins. Journal of Geophysical Research 97: 11933–11944. Willet SD (1999) Orogeny and orography: The effects of erosion on the structure of mountain belts. Journal of Geophysical Research 104: 28957–28981. Willgoose G, Bras RL, and Rodriguez-Iturbe I (1991) A physically based coupled channel network growth and hillslope evolution model (3 sections). Water Resources Research 27: 1671–1702. Wilson M (1993a) Geochemical signature of oceanic and continental basalts: a key to mantle dynamics. Journal of the Geological Society 150: 977–990. Wilson M (1993b) Magmatism and the geodynamics of basin formation. Sedimentary Geology 86: 5–29. Wilson M (1997) Thermal evolution of the Central Atlantic passive margins: continental break-up above a Mesozoic superplume. Journal of the Geological Society 154: 491–495. Wilson M and Bianchini G (1999) Tertiary–Quaternary magmatism within the Mediterranean and surrounding regions. In: Durand B, Jolivet L, Horva´th F, and Se´ranne M (eds.) Geological Society, London, Special Publication, 156: The Mediterranean Basin: Tertiary Extension within the Alpine Orogen, pp. 141–168. London: Geological Society, London. Wilson M and Patterson R (2001) Intraplate magmatism related to short-wavelength convective instabilities in the upper mantle: Evidence from the Tertiary-Quaternary volcanic province of western and central Europe. Geological Society of America Special Paper 352: 37–58. Wilson M, Rosenbaum JM, and Dunworth EA (1995) Melilites: partial melts of the thermal boundary layer? Journal of Petrology 32: 181–196. Wittenberg A, Vellmer C, Kern H, and Mengel K (2000) The Variscan lower continental crust: evidence for crustal delamination from geochemical and petrological investigations. In: Franke W, Haak V, Oncken O, and Tanner D (eds.) Geological Society, London, Special Publication, 179: Orogenic Processes: Quantification and Modelling in the Variscan Belt, pp. 401–414. London: Geological Society, London.
Worrall DM and Snelson S (1989) Evolution of the northern Gulf of Mexico, with emphasis on Cenozoic growth faulting and the role of salt. In: Bally AW and Palmer AR (eds.) The Geology of North America – An Overview, The Geology of North America, A, pp. 97–138. Boulder, CO: Geological Society of America. Wortel R and Spakman W (2000) Subduction and slab detachment in the Mediterranean–Carpathian region. Science 290: 1910–1917. Yegorova TP, Stephenson RA, Kozlenko VG, Starostenko VI, and Legostaeve OV (1999) 3-D gravity analysis of the Dniepr–Donets Basin and Donbas Foldbelt, Ukraine. Tectonophysics 313: 41–58. Zeck HP, Monie P, Villa IM, and Hansen BT (1992) Very high rates of cooling and uplift in the Alpine belt of the Betic Cordilleras, southern Spain. Geology 20: 79–82. Zeyen H, Volker F, Wehrle V, Fuchs K, Sobolev SV, and Altherr R (1997) Styles of continental rifting; crust-mantle detachment and mantle plumes. Tectonophysics 278: 329–352. Ziegler PA (1983) Crustal thinning and subsidence in the North Sea. Nature 304: 561p. Ziegler PA (1987) Compressional intra-plate deformations in the Alpine foreland – An introduction. Tectonophysics 137: 1–5. Ziegler PA 1988 Evolution of the Arctic-orth Atlantic and the western Tethys, American Association of Petroleum Geologists Memoir, 43: 198p. Ziegler PA (1989a) Evolution of the North Atlantic; an overview. American Association of Petroleum Geologists Memoir 46: 111–129. Ziegler PA (1989b) Evolution of Laurussia. A study in Late Palaeozic Plate Tectonics, 102p. Dordrecht: Kluver Academic. Ziegler PA (1990a) Collision related intraplate compression deformations in western and central Europe. Journal of Geodynamics 11: 357–388. Ziegler PA (1990b) Geological Atlas of Western and Central Europe, 2nd edn., 239p. Bath: Shell Internationale Petroleum Maatschappij BV. Ziegler PA (1992) Geodynamics of rifting and implications for hydrocarbon habitat. Tectonophysics 215: 221–253. Ziegler PA (1993) Plate moving mechanisms: their relative importance. Journal of the Geological Society 150: 927–940. Ziegler PA (1994) Cenozoic rift system of Western and Central Europe: An overview. Geologie en Mijnbouw 73: 99–127. Ziegler PA (1996a) Hydrocarbon habitat in rifted basins. In: Roure F, Ellouz N, Shein VS, and Skvortsov I (eds.) Geodynamic Evolution of Sedimentary Basins, pp. 85–94. Paris: Editions Technip. Ziegler PA (1996b) Geodynamic processes governing development of rifted basins. In: Roure F, Ellouz N, Shein VS, and Skvortsov L (eds.) Geodynamic Evolution of Sedimentary Basins, pp. 19–67. Paris: Editions Technip. Ziegler PA and Cloetingh S (2004) Dynamic processes controlling evolution of rifted basins. Earth-Science Reviews 64: 1–50. Ziegler PA and De`zes P (2005) Evolution of the lithosphere in the area of the Rhine Rift System. Behrmann JH, Granet M, Schmid S, and Ziegler PA (eds.) EUCOR-URGENT Special Issue. International Journal of Earth Sciences, 94, pp. 594–614. Ziegler PA and De`zes P (2006) Crustal evolution of Western and Central Europe. In: Gee DG and Stephenson RA (eds.) Geological Society, London, Memoirs, 32: European Lithosphere Dynamics, pp. 43–56. London: Geological Society, London. Ziegler PA and De`zes P 2007. Cenozoic uplift of Variscan Massifs in the Alpine foreland: Timing and controlling mechanisms. Global and Planetary Change, 58(1-4), 237–269.
Tectonic Models for the Evolution of Sedimentary Basins Ziegler PA and Horva´th F (eds.) (1996) Me´moires du Museum National d’Histoire Naturelle, 170: Peri-Tethys Memoir 2: Structure and Prospects of Alpine Basins and Forelands, 547p. Ziegler PA and Roure F (eds.) (1996) Architecture and petroleum systems of the Alpine orogen and associated basins. Me´moires du Museum National d’Histoire Naturelle, vol 170 Peri–Tethys Mem, vol 2. Paris: Commission for the Geological Map of the World. Ziegler PA and Stampfli GM (2001) Late Palaeozoic–Early Mesozoic plate boundary reorganization: collapse of the Variscan orogen and opening of Neotethys. In: Cassinis G (ed.) Natura Bresciana, Monografia 25: Permian continental deposits of Europe and other areas. Regional reports and Correlations, pp. 17–34. Brescia: Museo Civico di Scienze Naturali di Brescia. Ziegler PA, Cloetingh S, and van Wees J-D (1995) Dynamics of intraplate compressional deformation: the Alpine foreland and other examples. Tectonophysics 252: 7–59. Ziegler PA, Van Wees J-D, and Cloetingh S (1998) Mechanical controls on collision-related compressional intraplate deformation. Tectonophysics 300: 103–129. Ziegler PA, Cloetingh S, Guiraud R, and Stampfli GM (2001) Peri-Tethyan platforms: constraints on dynamics of rifting and basin inversion. In: Ziegler PA, Cavazza W, Robertson AHF, and Crasquin-Soleau S (eds.) Peri-Tethys Memoir 6: Peri-Tethyan Rift/Wrench Basins and Passive Margins, Me´moires du Museum National d’Histoire Naturelle, 186, pp. 9–49. Paris: Commission for the Geological Map of the World. Ziegler PA, Bertotti G, and Cloetingh S (2002) Dynamic processes controlling foreland development—the role of mechanical (de)coupling of orogenic wedges and forelands.
611
In: Bertotti G, Schulmann K, and Cloetingh SAPL (eds.) EGU St. Muller Special Publication Series, 1: Continental Collision and the Tectono-Sedimentary Evolution of Forelands, pp. 17–56. EGU. Ziegler PA, Schumacher ME, De`zes P, van Wees J-D, and Cloetingh S (2004) Post-Variscan evolution of the lithosphere in the Rhine Graben area: constraints from subsidence modelling. In: Wilson M, Neumann E-R, Davies GR, Timmerman MJ, Heeremans M, and Larsen BT (eds.) Geological Society, London, Special Publications, 223: Permo-Carboniferous Magmatism and Rifting in Europe, pp. 289–317. London: Geological Society, London. Ziegler PA, Schumacher ME, De´zes P, van Wees J-D, and Cloetingh S (2006) Post-Variscan evolution of the lithosphere in the area of the European Cenozoic Rift System. In: Gee DG and Stephenson RA (eds.) European Lithosphere Dynamics, 32, pp. 97–112. London, Mem: Geol Soc. Zoback MD, Stephenson RA, Cloetingh S, et al. (1993) Stresses in the lithosphere and sedimentary basin formation. Tectonophysics 226: 1–13. Zoback ML (1992) First and second order patterns of stress in the lithosphere: the World Stress Map Project. Journal of Geophysical Research 97: 11703–11728. Zoetemeijer R, Desegaulx P, Cloetingh S, Roure F, and Morett I (1990) Lithospheric dynamics and tectonic–stratigraphic evolution of the Ebro basin. Journal of Geophysical Research 95: 2701–2711. Zoetemeijer R, Cloetingh S, Sassi W, and Roure F (1993) Modelling of piggy-back basin stratigraphy: record of tectonic evolution. Tectonophysics 226: 253–269. Zoetemeijer R, Tomek C, and Cloetingh S (1999) Flexural expression of European continental lithosphere under the western outer Carpathians. Tectonics 18: 843–861.