William E. Holt

Dynamics of the India‐Eurasia collision zone

We present simple new dynamic calculations of a vertically averaged deviatoric stress field (over a depth average of 100 km) for Asia from geodetic, geologic, topographic, and seismic data. A first estimate of the minimum absolute magnitudes and directions of vertically averaged deviatoric stress is obtained by solving force balance equations for deviatoric stresses associated with gravitational potential energy differences within the lithosphere plus a first-order contribution of deviatoric stresses associated with stress boundary conditions. This initial estimate of the vertically averaged deviatoric stress field is obtained independent of assumptions about the rheology of the lithosphere. Absolute magnitudes of vertically averaged deviatoric stresses vary between 5 and 40 MPa. Assuming bulk viscous behavior for the lithosphere, the magnitudes of deviatoric stresses, together with the magnitudes of strain rates inferred from Quaternary fault slip rate and GPS data, yield vertically averaged effective viscosities for Tibet of 0.5–5×1022 Pa s, compared with 1–2.5×1023 Pa s in more rigid areas elsewhere in the region. A forward modeling method that solves force balance equations using velocity boundary conditions allows us to refine our estimates of the vertically averaged effective viscosity distribution and deviatoric stress field. The total vertically averaged deviatoric stress and effective viscosity field are consistent with a weak lower crust in Tibet; they are consistent with some eastward motion of Tibet and south China lithosphere relative to Eurasia; and they confirm that gravitational potential energy differences have a profound effect on the spatially varying style and magnitude of strain rate around the Tibetan Plateau. Our results for the vertically averaged deviatoric stress argue for a large portion of the strength of the lithosphere to reside within the seismogenic upper crust to get deviatoric stress magnitudes there to be as high as 100–300 MPa (in accord with laboratory and theoretical friction experiments indicating that stress drops in earthquakes are small fractions of the total deviatoric stress).

Velocity field in Asia inferred from Quaternary fault slip rates and Global Positioning System observations

We perform a joint inversion of Quaternary strain rates and 238 Global Positioning System (GPS) velocities in Asia for a self-consistent velocity field. The reference frames for all geodetic velocity observations are determined in our inversion procedure. India (IN) moves relative to Eurasia (EU) about a pole of rotation at (29.78°N, 7.51°E, 0.353° Myr−1), which yields a velocity along the Himalaya within India that is ∼73–76% of the magnitude of the IN-EU NUVEL-1A velocity and a vector azimuth that is 8–10° clockwise of NUVEL-1A IN-EU vector azimuth. Relative to Eurasia, south China moves at 9–11 mm/yr in the direction 110–120° with a pole position (64.84°N, 156.74°E, 0.12° Myr−1). Amurian block motion has a pole position in a similar location but at a slower rate (64.61°N, 158.23°E, 0.077° Myr−1) and most of the Amurian-Eurasia motion is accommodated by extension across Lake Baikal. Tarim Basin moves relative to Eurasia about a pole of rotation at (39.24°N, 98.2°E, −0.539° Myr−1) and ∼16–18 mm/yr of shortening is accommodated across the west central Tien Shan. There is distributed E-W extension throughout both southern and north central Tibet. Within southern Tibet, between the longitudes of 77°E to 92°E, the deformation field accommodates ∼16–19 mm/yr of E-W extension. We compare predicted seismic moment rates with those observed in this century in Asia. Total observed seismic moment rates within the entire area of central and east Asia (2.2×107 km2) in this century are 2.26±0.7×1020 N m yr−1 as compared with a predicted total rate of 2.03±0.066×1020 N m yr−1. Comparisons between observed and predicted moment rates within 42 subregions reflect the generally unstable process of inferring long-term seismic moment rates from a catalog of limited duration (94 years). An observation period of ∼10,000 years would be required to reduce uncertainties in observed seismic moment rate to the same size as the uncertainties in model tectonic moment rates, inferred from the joint inversion of GPS and Quaternary rates of strain. We show that in general, a better correlation with model tectonic moment rate is inferred from the seismicity catalog by considering the numbers of earthquakes above a cutoff magnitude (mb ≥ 5.0, for the period January 1, 1965, to January 1, 1999).

The accommodation of Arabia-Eurasia Plate convergence in Iran

Continental convergence between Arabia and Eurasia is taken up by distributed deformation in Iran. At wavelengths large compared with the thickness of the lithosphere this deformation is best described by a continuous velocity field. The only quantitative source of information on the spatial distribution of strain rates within Iran is the record of earthquakes. We find that we can reproduce the style of deformation observed in the seismicity by simply minimizing the rate of work in a continuous viscous medium that has to accommodate the Arabia-Eurasia plate motion between the defined shapes of Irans rigid borders. When, in addition, we specify central Iran, Azerbaijan, and the southern Caspian basin to be relatively rigid blocks within the deforming zone, then the fit to the style of the observed strain rate distribution is even better. We conclude that much of the pattern of deformation in Iran is predetermined by the shape of its rigid borders and by the disposition of relatively rigid blocks within it. This is likely also to have been a common occurrence in older orogenic belts. We confirm earlier suggestions that earthquakes between 1909 and 1992 can account for only a small part (∼10–20%) of the total deformation required by the convergence between the Arabia and Eurasia plates. We then show that the whole plate motion can be accommodated by a velocity field with the same orientations and relative magnitudes of principal strain rates seen in the earthquakes but with larger absolute magnitudes. There is therefore no requirement that the large proportion of aseismic deformation in Iran is substantially different in style, orientation, or distribution from that released seismically in the earthquakes.

The active tectonics of the eastern Himalayan syntaxis and surrounding regions

Source parameters of 53 moderate-sized earthquakes, obtained from the joint inversion of regional and teleseismic distance long-period body waves, provide the data set for an analysis of the style of deformation and kinematics in the region of the Eastern Himalayan Syntaxis. Focal mechanisms of Eastern Himalayan events show oblique thrust, consistent with the N-NE directed movement of the Indian plate as it underthrusts a boundary that strikes at an oblique angle to the direction of convergence. Earthquakes near the Sagaing fault show strike-slip mechanisms with right-lateral slip. Earthquakes on its northern splays, however, indicate predominant thrusting, evidence that the dextral motion on the Sagaing fault, which accommodates a portion of the lateral motion between India and southeast Asia, terminates in a zone of thrust faulting at the Eastern Himalayan Syntaxis. Remaining motion between India and southeast Asia is accommodated in a zone of distributed shear in east Burma and Yunnan, manifested by strike-slip and oblique normal faulting, east-west extension, crustal thinning, and clockwise rotation of crustal blocks. We determined strain rates throughout the region with a moment tensor summation using 25 years (modern) and 85 years (modern and historic) of earthquake data. We matched the observed strains with a fifth-order polynomial function, and from this we determined both the velocity field and rotations with respect to a specified region. Velocities calculated relative to south China stationary show that the entire area, extending from 20°N–36°N, within deforming Asia (Yunnan, western Sichuan, and east Tibet), constitutes a distributed dextral shear zone with clockwise rotations up to 1.7°/m.y., maximum in the region of the Eastern Syntaxis proper. Integrated strains across this zone, relative to south China stationary, show 38 mm/yr ± 12mm/yr of north-directed motion at the Himalaya. Remaining plate motion, relative to south China fixed, must be taken up by the underthrusting of India beneath the lesser Himalaya, strike-slip motion on the Sagaing fault, and intraplate NE directed shortening within NE India as well as NE directed shortening within the Eastern Syntaxis proper. 10 mm/yr ± 2 mm/yr of relative right-lateral motion between India and southeast Asia is absorbed in the region between the Sagaing and Red River faults (94°E–100°E). It is the clockwise vorticity (relative to south China) associated with the deformation in Yunnan, east Tibet, and western Sichuan that provides the relative north-directed motion of 38 ± 12 mm/yr at the Himalaya. Not all of the deformation is accommodated in right-lateral shear between India and south China and between east Tibet and south China; velocity gradients exist that are parallel to the trend of the shear zone. Relative to a point within western Sichuan (32°N, 100°E), the velocity field shows that the Yunnan crust is moving S-SE at rates of 8–10 mm/yr. Relative to south China, there is no eastward expulsion of crustal material beyond the eastern margin of the Tibetan plateau.

Accretionary tectonics of Burma and the three-dimensional geometry of the Burma subduction zone

The geometry of the Burma Wadati-Benioff zone (WBZ) has been determined by fitting a trend surface parameterized with eight effective degrees of freedom to 184 well-located hypocenters. The dip of this surface, which passes through the middle of the WBZ, varies from about 50° in the north near the eastern Himalayan syntaxis to about 30° in the Bay of Bengal area. The eastern edge of the Indo-Burman ranges closely follows the map projection of the 60 km depth contour of the WBZ. The curvature of the Indo-Burman ranges is controlled by the geometry of the interface between the more steeply dipping part of the Indian plate and the leading edge of the overriding Burma platelet. Shallow earthquakes beneath the Indo-Burman ranges are primarily confined to the underthrusting Indian plate. Their focal mechanisms indicate strike-slip faulting and north-south shortening parallel to the eastern margin of the Indian plate.

Correlated crust and mantle strain fields in Tibet

Vertical to subvertical planes of shear within the active crustal deformation field in Tibet align with fast directions of shear-wave polarization. This observation suggests that the present-day velocity gradient tensor field within the crust of Tibet correlates with the lithospheric mantle velocity gradient tensor field beneath Tibet. This inference requires the following. (1) The [100] axes of olivine are aligned within established zones of lithospheric mantle fabric or anisotropy, built up through finite strain. (2) Present-day shear in the mantle lithosphere is occurring within these established zones of fabric, parallel to the direction of olivine [100] axis alignment. (3) The zones of mantle lithospheric fabric, or zones of mantle shear, are aligned with zones of crustal shear (faults). The correlation of crustal and mantle strain fields most simply results from the fact that both crust and mantle lithosphere are under the influence of similar velocity boundary conditions. Furthermore, the observations confirm the distributed nature of lithospheric mantle deformation beneath east-central Tibet and suggest that left-lateral shear has been the dominant component of finite shear there.

Active deformation in eastern Indonesia and the Philippines from GPS and seismicity data

In this study we combine Global Positioning System (GPS) velocities with information on the style of regional seismicity to obtain a self-consistent model velocity and strain rate field for the entire eastern Indonesia and Philippines region. In the process of interpolating 93 previously published GPS velocities, the style and direction of the seismic strain rate field, inferred from earthquakes with M0 < 1 × 1020 N m (from the Harvard centroid moment tensor catalog), are used as constraints on the style and direction of model strain rates within the plate boundary zones. The style and direction of the seismic strain rate field are found to be self-similar for earthquakes up to M0 = 1 × 1020 N m (equivalent to Mw < 7.3). Our inversion result shows the following: The Java Trench, which lacks any significant (historic) seismicity, delineates the Australian plate (AU) - Sunda block (Sunda) plate boundary west of the island of Sumba. East of Sumba, convergence is distributed over the back arc and Banda Sea, and there is no subduction at the Timor Trough, suggesting that the northern boundary of the AU plate runs north of this part of the Banda arc through the Banda Sea. In New Guinea most motion is taken up as strike-slip deformation in the northern part of the island, delineating the Pacific plate (PA) - AU boundary. However, some trench-normal convergence is occurring at the New Guinea Trench, evidence that the strain is partitioned in order to accommodate oblique PA-AU motion. PA-AU motion is consistent with NUVEL-1A direction, but ∼ 8 mm yr−1 slower than the NUVEL-1A estimate for PA-AU motion. The Sulawesi Trench and Molucca Sea delineate zones of high strain rates, consistent with high levels of active seismicity. The Sulawesi Trench may take up some of the AU-Sunda motion. Philippine Sea plate motion is in a direction slightly northward of the NUVEL-IA estimate and is partitioned in some strike-slip strain rates along the Philippine Fault and relatively larger trench-normal convergence along the Philippine Trench and on the Philippine mainland in the southern Philippines and along the Manila Trench in the northern Philippine islands. The high level of strain rate along the Manila Trench is not released by any significant (historic) seismic activity. For the entire eastern Indonesia-Philippines region, seismicity since 1963 has taken up ∼40% of the total moment rate inferred from our model.

The horizontal velocity field in the deforming Aegean Sea region determined from the moment tensors of earthquakes

We use the spatial distribution of seismic moment tensors of earthquakes in the Aegean region over the time interval 1909–1983 to recover a continuous horizontal velocity field that describes the overall deformation of the lithosphere at large length scales. The calculated velocity field is dominated by two effects: (1) an E-W right-lateral shear of the eastern Aegean, related to motion on the North Anatolian fault becoming distributed as it enters the Aegean; and (2) a N-S extension, probably related to the sinking of the slab in the Hellenic Trench. The southern part of the central Aegean is found to be moving in a SW direction relative to Europe at a rate of about 30 mm/yr (probably a lower bound, with an error of around ± 10 mm/yr) and rotating clockwise. In the seismogenic upper crust this velocity field is accommodated by right-lateral strike-slip faults in the eastern Aegean and by normal faults that rotate clockwise in central Greece. A comparison of paleomagnetic declination data with the expected rotation of rigid elongate inclusions in the velocity field shows in most places an agreement in sense and approximate agreement in rate of rotation. Expected rotation rates of line elements are sometimes too low: probably because our derived velocity field is smoothed and unable to match locally high strain rates. There is only one part of the region where line elements are predicted to rotate in either clockwise or counterclockwise directions, depending on their orientation; this is in western Turkey, which, coincident ally, is the only place where paleomagnetic rotations in both directions have been observed. This coincidence in particular suggests to us that the analogy of rigid elongate inclusions in the velocity field, responding to forces on their bases, may be useful in predicting the senses and approximate rates of rotation of crustal blocks in deforming continental regions. The velocity field we obtain preserves the strike directions of the major faults as directions of zero length change, in spite of considerable smoothing. We use this observation to speculate that the interaction between the upper crust and the rest of the lithosphere beneath it may involve an interplay of effects. On one hand the variation of strength with direction in the crust may control the strike directions of faults that form or become reactivated and may also limit the velocity fields that are allowable. On the other hand, the fault bounded blocks may rotate in the velocity field in response to forces on their bases, which would require the velocity field to change with time if the directions of zero length change are fixed to the blocks.

The kinematics of northern South Island, New Zealand, determined from geologic strain rates

Relative motions within the distributed plate boundary zone of northern South Island, New Zealand, are determined through an inversion of geologic strain rate estimates. The Quaternary fault slip rate estimates define the shear strain rates, and rock uplift rates provide information on the horizontal divergence rates. An erosion rate to rock uplift rate ratio along with a crustal compensation factor is estimated in order to convert rock uplift rates to horizontal divergence rates. Because of the uncertainty in erosion rates, horizontal divergence rates σ are given a large standard error of ±σ. The three horizontal strain rate components obtained from these data completely define the horizontal velocity gradient tensor. Strain rate distributions are matched with spline polynomial functions, which can be constrained to behave rigidly within specified regions, such as the Pacific or Australian plates. Inversion of the strain rate distribution, assuming uniform erosion rates across the northern South Island, yields a velocity field that has small differences in both magnitude (10% larger) and direction with the NUVEL-1A plate motion model between Pacific and Australian plates. A revised strain rate data set, obtained from a variable erosion model in which erosion rates are a linear function of the log of the average annual rainfall magnitudes, yields a velocity field with expected directions that are indistinguishable from the NUVEL-1A plate motion model between Pacific and Australian plates, but velocity magnitudes are still 10- 15% higher than the plate motion model. Therefore the average values of slip rate on strike-slip faults in Marlborough, required by the NUVEL-1A plate motion model, are typically close to the low end of the published range of slip values for those structures. The major strike-slip structures within the Marlborough region are accommodating 80 - 100% of the total plate motion between Australia and Pacific plates on northern South Island ; as much as 20% of the relative motion could be accommodated by shear generated by folding or thrust faulting with a single orientation between the major structures. The magnitude of plate motion contributed by the major faults indicates that the distributed geodetic shear strains measured across northern South Island, and discussed by Bibby [1981] and Walcott [1984], are not a form of irrecoverable strain but rather a feature of the strain field that will eventually be released primarily as concentrated slip on the strike-slip structures within the region. Strain modeling indicates that the earthquake moment release over the last 150 years within the northwest Nelson province, which is west of the Marlborough region, is at least half an order of magnitude higher than the long-term rate of moment release. Overall, the northwest Nelson region is playing only a minor role in taking up the long-term plate motion.

Velocity fields in deforming Asia from the inversion of earthquake-released strains

Average strain rates in sectors of deforming Asia are matched by a fifth-order polynomial function, and that function integrated, to obtain the relative velocities and rotations occurring within east Tibet, western Sichuan, Yunnan, and south China. The method was applied to strain rates obtained from moment tensor summation of both modern and historic earthquakes but can be applied as well to strains obtained from Quaternary slip rates on major faults. If south China has negligible motion relative to Siberia, then the velocity results indicate that nearly all of the expected motion between India and the south China portion of Eurasia has, in the last 85 years, been accommodated by distributed intraplate deformation in east Burma, Yunnan, western Sichuan, and east Tibet. Calculations indicate that these regions constitute a zone of distributed right-lateral shear that accommodates an overall north-south sense of relative motion between east Tibet and south China and India and south China. Line elements parallel to both right-lateral and left-lateral faults in east Burma and western Yunnan are rotating clockwise relative to south China, with the line elements parallel to left-lateral faults rotating most rapidly (2.0 ± 0.5°/m.y.). In eastern Tibet and the Gansu-Ningxia, NW-SE trending left-lateral faults give rotation clockwise relative to south China (1–2.5°/m.y.). In central Tibet and western Sichuan, right-lateral faults give slight counterclockwise rotation rates relative to south China (0.5–0.75°/m.y.) Instantaneous rotation rates within the deforming region, extrapolated over a 20–40 m.y. time period, are in rough agreement with the paleomagnetic rotations measured in Cretaceous-aged rocks.