Gregory A. Houseman

CRUSTAL THICKENING VERSUS LATERAL EXPULSION IN THE INDIAN-ASIAN CONTINENTAL COLLISION

Since the beginning of the continental collision between India and Asia there has been about 2500 km of convergence, and the northward movement of India has been accommodated by major internal deformation of the Asian lithosphere. The crustal thickening in and around the Tibetan Plateau is clearly a direct consequence of this collision, but there is considerable debate as to whether a large fraction of the indentation has been accommodated by eastward motion of the lithospheric blocks of southeastern Asia and southern China. Numerical experiments described here test this hypothesis for a range of indentation geometries and rheological models of the lithosphere. We employ a thin viscous sheet model of the lithosphere with a depth-averaged nonlinear viscous rheology described by a stress-strain rate exponent n and including gravitational buoyancy forces scaled by the dimensionless Argand number Ar. The eastern boundary for the collision region is described as a lithostatic boundary; the precollision normal stress is determined by static balance, and that constant stress is applied throughout the collision. The experiments show that during collision the eastern boundary is smoothly displaced to the east at a rate about 1/4 of the indentation rate, with only minor variation due to geometry or rheology. Crustal thickening rates in the region of the plateau are reduced by between 10% and 25% of the corresponding rates determined for similar experiments with a rigid eastern boundary. To generalize from the results of these experiments, the total north-south shortening strain produced by the collision is partitioned between crustal thickening and eastward displacement in the ratio of at least 3:1 and more probably 4:1.

Crustal structure across Longmenshan fault belt from passive source seismic profiling

[1] We analyse receiver functions from 29 broad-band seismographs along a 380-km profile across the Longmenshan (LMS) fault belt to determine crustal structure beneath the east Tibetan margin and Sichuan basin. The Moho deepens from about 50 km under Songpan-Ganzi in east Tibet to about 60 km beneath the LMS and then shallows to about 35 km under the western Sichuan basin. The average crustal Vp/Vs ratios vary in the range 1.75-1.88 under Songpan-Ganzi in east Tibet, 1.8-2.0 under the LMS, and decrease systematically across the NW part of the Sichuan basin to less than 1.70. A negative phase arrival above the Moho under Songpan-Ganzi and Sichuan basin is interpreted as a PS conversion from the top of a low-velocity layer in the lower crust. The very high crustal Vp/Vs ratio and negative polarity PS conversion at the top of lower crust in east Tibet are inferred to be seismic signatures of a low-viscosity channel in the eastern margin of the Tibetan plateau. The lateral variation of Moho topography, crustal Vp/Vs ratio and negative polarity PS conversion at the top of the lower crust along the profile seem consistent with a model of lower crust flow or tectonic escape.

Geodynamics of the Tarim Basin and the Tian Shan in central Asia

The India-Asia collision has caused crustal thickening in Tibet by at least a factor of 2. In the last 20–30 Myr of this collision the Tian Shan mountain range has also been reactivated. The Tarim Basin, however, shows little internal deformation. We describe a series of numerical experiments which constrain the effective lithospheric strength parameters of the Tarim Basin and the Tian Shan in the context of a thin viscous sheet model. We use a finite element representation of the thin viscous sheet model to approximate the deformation field in the India-Asia collision and compare model crustal thickness distributions with that inferred from a local isostatic model of topography. The experiments show that a strong Tarim Basin, while undergoing little internal deformation, transfers strain to the Tian Shan, producing significant crustal thickening in the Tian Shan region. Based only on diagnostic parameters such as the maximum thickening in the Tian Shan and the minimum in the Tarim Basin, the principal features of the topography can be approximately reproduced using models in which either the Tarim Basin is strong, or the Tian Shan is weak, or both. For a rheological model in which the stress versus strain-rate exponent is n = 3 the strength coefficient for the strong Tarim Basin model, VTarim is between about 1.7 and 2. For the weak Tian Shan model the relative strength coefficient VTian is between about 0.65 and 0.75. If both strong Tarim and weak Tian Shan are included, there is a trade-off between the required values of VTarim and VTian and the shape of the predicted crustal thickness profiles better matches the observed profiles. The steep topographic slope on the southern margin of the Tarim Basin requires that it is anomalously strong, while the rapid decrease of topographic height to the north of the Tian Shan requires that it is anomalously weak. Similar conclusions are obtained with a rheological model based on n = 10. Simplified rheological models of the lithosphere show that the variations in lithospheric strength may be explained by changes to the Moho temperature of the order of 10° to 30°C.

Lithospheric instability beneath the Transverse Ranges of California

Recent high-resolution seismic experiments reveal that the crust beneath the San Gabriel Mountains portion of the Transverse Ranges thickens by 10–15 km (contrary to earlier studies). Associated with the Transverse Ranges, there is an anomalous ridge of seismically fast upper mantle material extending at least 200 km into the mantle. This high-velocity anomaly has previously been interpreted as a lithospheric downwelling. Both lithospheric downwelling and crustal thickening are associated with the oblique convergence of Pacific and North America plates across the San Andreas Fault, though it seems likely that the lithospheric downwelling is driven at least partly by gravitational instability of the cold lithospheric mantle. We show by means of numerical experiment that the balance between buoyancy forces that drive deformation and viscous stresses that resist deformation determines the geometry of crustal thickening and mantle downwelling. We use a simple two-layered lithospheric model in which dense lithospheric mantle overlies relatively inviscid and less dense asthenosphere and is overlain by buoyant crust. External plate motion drives convergence, which is constrained by boundary conditions to occur within a central convergent zone of specified width. A fundamental transition in the geometry of downwelling is revealed by our experiments. For slow convergence, or low crustal viscosity, downwelling occurs as multiple sheets on the margins of the convergent zone. For fast convergence or crust that is stronger than mantle lithosphere a single downwelling occurs beneath the center of the convergent zone. This complexity in the evolution of the system is attributed to the interaction of crustal buoyancy with the evolving gravitational instability. In order for a narrow downwelling slab to have formed beneath the Transverse Ranges within the last 5 Myr, the effective lithospheric viscosity of the convergent region is at most about 10^20 Pa s.

Mixing in vigorous, time‐dependent three‐dimensional convection and application to Earth's mantle

An understanding of the mechanism of mixing in highly viscous convecting fluids is of crucial importance in explaining the observed geochemically heterogeneous nature of Earths mantle. Using constant viscosity numerical experiments, we describe the mixing mechanism of time-dependent Rayleigh-Benard convection with an infinite Prandtl number in a three-dimensional (3-D) rectangular container. Mixing is observed by following the positions of passive tracers advected by the flow. The major mixing mechanisms may be described in terms of the within-cell mixing and the cross-cell mixing. The flow structure in which tracers move on toroidal surfaces, that was previously observed in steady state 3-D convection systems is perturbed by boundary layer instabilities in the time-dependent experiments. This flow structure allows a very efficient exchange of mass between the boundary layers and the core of the convection cell even in the absence of time dependence. We compare this result with calculations carried out in two spatial dimensions. In similar two-dimensional (2-D) experiments, exchange of mass between boundary layers and core of the convection cell is entirely effected by the boundary layer instabilities. Mixing between neighboring cells appears much slower in three dimensions than in similar 2-D experiments, perhaps because the 3-D cell structure is more stable relative to the boundary layer instabilities. The inferred mixing rates are observed to be relatively insensitive to initial tracer location, but the timescale for mixing, tm, decreases with increasing Rayleigh number (tm goes approximately as Ra(−3/2)). The timescale of mixing is an important constraint on the large scale structure of Earth, because large-scale geochemical heterogeneities persist to the present day, implying that the mantle is not well mixed.

3D convection at infinite Prandtl number in Cartesian geometry — a benchmark comparison

Abstract We describe the results of a benchmark study of numerical codes designed to treat problems of high Prandtl number convection in three-dimensional Cartesian geometry. In addition, quantitative results from laboratory convection experiments are compared with numerical data. The cases of bimodal convection at constant viscosity and of square cell convection for temperature-dependent viscosity have been selected.

Lithospheric scale gravitational flow: the impact of body forces on orogenic processes from Archaean to Phanerozoic

Abstract In the Archaean, the combination of warmer continental geotherm with a lighter sub-continental lithospheric mantle suggests that gravitational forces played a more significant role in continental lithospheric deformation. To test this hypothesis, we compare the evolution of the deformation and the regional state of stress in ‘Archaean-like’ and ‘Phanerozoic-like’ lithospheres submitted to the same boundary conditions in a triaxial stress-field with imposed convergence in one direction. For plausible physical parameters, thickening of normal to cold Phanerozoic lithospheres produces relatively weak buoyancy forces, either extensional or compressional. In contrast, for Archaean continental lithospheres, or for anomalously warm Phanerozoic lithospheres, lateral gravitationally-driven flow prevents significant thickening. This conclusion is broadly consistent with: (1) the relative homogeneity of the erosional level now exposed at the surface of Archaean cratons, (2) the sub-aerial conditions that prevailed during the emplacement of up to 20 km of greenstone cover, (3) the relatively rare occurrence in the Archaean record of voluminous detrital sediments, (4) the near absence of significant tectonic, metamorphic and magmatic age gradients across Archaean cratons, (5) the relative homogeneity of strain across large areas, and (6) the ubiquitous presence of crustal-scale strike slip faults in many Late Archaean cratons.

Intra-orogenic extension driven by gravitational instability: Carpathian-Pannonian orogeny

The extensional Pannonian Basin was formed in a few million years during Miocene time synchronously with contraction in the surrounding Alpine and Carpathian orogens. This system is characteristic of a class of extensional basins that form in the midst of active orogenic (mountain-forming) belts. The mechanism that causes this type of geological event is enigmatic but usually has been associated with subduction. We examine a new hypothesis for intra-orogenic extensional basin formation in which gravitational spreading of previously thickened crust triggers gravitational instability of the mantle lithosphere. A basin is formed by lithospheric extension as shortening and lithospheric downwelling occur in the surrounding mountain belts. This mechanism provides a mechanically self-consistent explanation for all of the main structural features of the Pannonian-Carpathian system and presents a plausible alternative to the popular view that subduction and slab rollback have driven the development of this basin.

Comparison of shear wave splitting and finite strain from the India‐Asia collision zone

We investigate whether observations of shear wave splitting in Asia may be related to distributed strain of the lithosphere or to strain caused by motion of the Asian lithosphere over the deep mantle. The HS2-NUVEL angular velocity for Eurasia, and that of Minster and Jordan [1978], predict “absolute” plate motions that are a factor of at least 4 slower than the NNE relative motion of India with respect to stable Eurasia. Shear of the upper mantle beneath the south Asian lithosphere is therefore likely to be dominated by the velocity field caused by the internal deformation of Asia, rather than by plate motion. If that shear causes SKS splitting, fast polarization directions in tectonic Asia should be aligned approximately in the direction of the motion of Asian lithosphere relative to the deep mantle, which for southern Asia is NNE. In general, however, fast polarization directions are not aligned with this direction, and in some cases are nearly orthogonal to it. We calculate finite deformation of the Asian lithosphere, in response to its indentation by India and, from the deformation, predict the orientations of fast polarization directions aligned with the principal axes of elongation. These predictions agree with the gross features of the anisotropy in Tibet and the Tien Shan. Orientations of anisotropy in the Baikal region of Siberia are consistent with the orientations of the local principal axes of active strain. Orientations of anisotropy in Mongolia, however, are inconsistent both with the azimuth of the velocity of this region with respect to the deeper mantle and with the orientation of the calculated maximum elongation of the region. We conclude that SKS splitting observations from Asia agree better with those predicted by calculations of internal lithospheric deformation due to indentation by India than with strain predicted from motion of the lithosphere over the deeper mantle. These results place constraints on the depth range where shear wave splitting occurs, which we propose lies predominantly within the lithospheric mantle.

Weak ductile shear zone beneath a major strike‐slip fault: Inferences from earthquake cycle model constrained by geodetic observations of the western North Anatolian Fault Zone

GPS data before and after the 1999 Izmit/Duzce earthquakes on the North Anatolian Fault Zone (Turkey) reveal a preseismic strain localization within about 25 km of the fault and a rapid postseismic transient. Using 3-D finite element calculations of the earthquake cycle in an idealized model of the crust, comprising elastic above Maxwell viscoelastic layers, we show that spatially varying viscosity in the crust can explain these observations. Depth-dependent viscosity without lateral variations can reproduce some of the observations but cannot explain the proximity to the fault of maximum postseismic velocities. A localized weak zone beneath the faulted elastic lid satisfactorily explains the observations if the weak zone extends down to midcrustal depths, and the ratio of relaxation time to earthquake repeat time ranges from ~0.005 to ~0.01 (for weak-zone widths of ~24 and 40 km, respectively) in the weakened domain and greater than ~1.0 elsewhere, corresponding to viscosities of ~1018 ± 0.3 Pa s and greater than ~1020 Pa s. Models with sharp weak-zone boundaries fit the data better than those with a smooth viscosity increase away from the fault, implying that the weak zone may be bounded by a relatively abrupt change in material properties. Such a change might result from lithological contrast, grain size reduction, fabric development, or water content, in addition to any effects from shear heating. Our models also imply that viscosities inferred from postseismic studies primarily reflect the rheology of the weak zone and should not be used to infer the mechanical properties of normal crust.