Gary L. Pavlis

The generalized earthquake-location (GENLOC) package: an earthquake-location library

We describe a library and associated set of applications for locating seismic events. The library is called the GENeralized LOCation (GENLOC) library because it is a general library that implements most methods commonly used for single event locations. The library has a flexible implementation of the standard Gauss-Newton method with many options for weighting schemes, inversion methods, and algorithms for choosing an initial location estimate. GENLOC also has a grid-search algorithm that makes no assumptions about the geometry of the grid it is searching returning only the point with a best fit solution for the specified residual norm. GENLOC supports both arrival time and array slowness vector measurements. A unique feature is the strong separation between the travel time/earth model problem and the location estimations. GENLOC can utilize data from any seismic phase for which the user can supply an earth model and method to compute theoretical travel times and/or slowness values. The GENLOC library has been used in five different working applications: (1) a simple command line program, (2) an interactive graphical user interface version used in an analyst information system, (3) a database-driven relocation program, (4) a recent implementation of the progressive multiple event location method, and (5) a real-time location program. We ran a validation test against LOCSAT and found reasonable consistency in estimated locations. We attribute observed differences in the solutions to roundoff errors in different calculators used by the two programs.

Erosional processes as a control on the structural evolution of an actively deforming fold and thrust belt: An example from the Pamir-Tien Shan region, central Asia

The foreland of the Pamir Mountains displays an unusual orogenic topography in which an active north vergent, thin-skinned fold-and-thrust belt shows no topographic foredeep in advance of the thrust sheets. Instead, the thrusts are impinging on a topographic barrier formed by the southern edge of the Tien Shan. The deformation front of this thrust belt is coincident with a deeply incised river valley occupied by a large river, the Surkhob. We suggest that this river played a key role in the evolution of this orogenic system by acting as a lateral conveyor belt for the transport of mass delivered to the deformation front by thrusting. On the basis of reconnaissance geologic mapping it is clear that the present system developed in the Late Miocene or Pliocene, when the Tien Shan were uplifted within an existing foreland basin. As the Tien Shan highland developed, a longitudinal river within the foreland basin was trapped between the north directed thrusts and the Tien Shan highland. With continued deformation the river was squeezed into a narrow canyon and since that time the system has been in an unusual equilibrium where thrusting acts to push the deformation front northward toward the Tien Shan, but fluvial erosion and transport act to keep it in a constant position. Geologic relationships indicate that the Vakhsh thrust, the frontal thrust of the north directed system, is now emerging on an erosion surface beneath a thin Quaternary cover throughout the canyon, suggesting that the river has been unable to keep pace with the tectonic mass flux, and the thrust front, as well as the river, have been pushed northward across the beveled edge of what was originally the southern edge of the Tien Shan. We suggest that the original deformation front of the Tien Shan now lies ∼10 km south of the present-day topographic front of the Tien Shan, beneath the north vergent Vakhsh thrust, as indicated by: (1) a near-vertical zone of intense basement seismicity (the “Petrovsky” seismic zone); (2) an en echelon array of backthrusts along the south flank of the Peter the First Range; and (3) footwall cutoff relationships of the Vakhsh thrust system. Order-of-magnitude, mass balance considerations based on sediment budgets support this conclusion but suggest that the mass imbalance may be small. Thus large volumes of material may have passed through the thrust fron and been carried out of the system by erosion, a relationship that may account for the narrow (30–50 km wide) band of deformed sedimentary rock between the Pamir salient and the Tien Shan. The present near equilibrium is maintained by long-term tectonic transport rates and fluvial transport rates, the latter closely linked to climate, particularly precipitation. Thus, although there is apparently a mass “pile up” within this system, other systems of this type might show grossly different structural and stratigraphic relationships depending on the amount of rainfall and corresponding variations in rates of fluvial sediment transport. Similar settings can be recognized in the active thick-skinned thrust systems of the Tien Shan and ancient systems like the Laramide belt of western North America.

The Pamir‐Hindu Kush seismic zone as a strain marker for flow in the upper mantle

We introduce a new hypothesis to explain the unusual pattern of intermediate-depth earthquakes in the Pamir-Hindu Kush region of central Asia. In our model the seismicity of the zone is defined by a remnant piece of oceanic crust that became neutrally buoyant and now hangs in the mantle underneath the center of the active mountain belt. We interpret an abrupt cutoff in seismicity shallower than 70 km under the Pamir as evidence that the slab has become decoupled from the surface deformation, and we suggest that the remnant slab has acted as a strain marker for flow in the upper mantle. The contortion of the slab can be explained by superimposed vertical and lateral simple shear related to the western edge of the Indian Plate. This model requires relatively low strain rates of the order of 10−15 s−1, consistent with strain rates estimated from earthquake moment release in Benioff zones. In contrast, the double-facing subduction zone model applied by a number of previous authors requires strain rates of the order of 10−13 s−1, which are much higher than inferred strain rates in subducting slabs. The results are important for understanding mantle dynamics, providing a unique way to measure strain in the upper mantle. The implied flow field also has significant implications for the kinematics of the entire collisional plate boundary in Tibet and the Tien Shan.

Convolutional quelling in seismic tomography

This paper introduces convolutional quelling as a technique to improve imaging of seismic tomography data. We show the result amounts to a special type of damped, weighted, least‐squares solution. This insight allows us to implement the technique in a practical manner using a sparse matrix, conjugate gradient equation solver. We applied the algorithm to synthetic data using an eight nearest‐neighbor smoothing filter for the quelling. The results were found to be superior to a simple, least‐squares solution because convolutional quelling suppresses side bands in the resolving function that lead to imaging artifacts.

Evolution of the Southern Caribbean Plate Boundary

It is generally accepted that the cores of the continents, called cratons, formed by the accretion of island arcs into proto-continents and then by proto-continental agglomeration to form the large continental masses. Mantle-wedge processes, combined with higher melting temperatures during the Archean (2.5–3.8 billion years ago) and possibly thrust stacking of highly depleted Archean oceanic lithosphere, produced a strong, buoyant, upper mantle chemical boundary layer. This stabilizing mantle layer, known as the tectosphere, has shielded the Archean cratons from most subsequent tectonic disruption and is highly depleted in iron, providing the positive buoyancy that is required to ‘float’ the continents more than four kilometers above the surrounding ocean basins.

Evidence for a high-velocity slab associated with the Hindu Kush seismic zone

We used teleseismic travel time residuals to determine lateral velocity variations of the crust and upper mantle in the Pamir-Hindu Kush region in Tadjikistan and Afghanistan. Data from 29 analog seismic stations in Tadjikistan and northern Afghanistan were used to determine travel time residuals for 210 teleseismic events ranging in distance from 28° to 87° and covering a broad range of azimuths. We inverted for velocity perturbations over a rectangular grid with a block size of 99 × 99 km. The model extended to a depth of 350 km with a 50-km-thick first layer and two 150-km-thick deeper layers. The results show a strong and well-resolved zone of high velocities in the upper mantle at depths greater than 200 km, coincident with the location of the Hindu Kush seismic zone. No clear velocity perturbations are associated with the Pamir seismic zone. Above 200 km little correlation is observed with the seismic zone, but indications of thicker crust under the Pamir and thinner crust under the Tadjik Depression are seen. The high velocities are most likely caused by the presence of oceanic lithosphere at depth.

Imaging P‐to‐S conversions with multichannel receiver functions

We present a new methodology in the direct imaging of P-to-S converted phases recorded on broadband seismic arrays. Our approach is based on conventional three-component array processing and receiver function techniques with the key addition of a weighted stack based on an aerial smoothing function. This creates synthetic arrays with a specified aperture whose image points vary continuously across the array. With this approach, it is possible to interpolate data from an array of broadband stations onto an arbitrarily fine grid. We have applied this technique to a single deep event recorded by the Lodore broadband array, located in northern Colorado and southern Wyoming. The resulting images show distinct differences in crustal structure across the array, and also image major upper mantle discontinuities.

Cooperation among tectonic and surface processes in the St. Elias Range, Earth's highest coastal mountains

Investigations of tectonic and surface processes have shown a clear relationship between climate-influenced erosion and long-term exhumation of rocks. Numerical models suggest that most orogens are in a transient state, but observational evidence of a spatial shift in mountain building processes due to tectonic-climate interaction is missing. New thermochronology data synthesized with geophysical and surface process data elucidate the evolving interplay of erosion and tectonics of the colliding Yakutat microplate with North America. Focused deformation and rock exhumation occurred in the apex of the colliding plate corner from > 4 to 2 Ma and shifted southward after the 2.6 Ma climate change. The present exhumation maximum coincides with the largest modern shortening rates, highest concentration of seismicity, and the greatest erosive potential. We infer that the high sedimentation caused rheological modification and the emergence of the southern St. Elias, intercepting orographic precipitation and shifting focused erosion and exhumation to the south.

AFTERSHOCK SEQUENCES OF INTERMEDIATE-DEPTH EARTHQUAKES IN THE PAMIR-HINDU KUSH SEISMIC ZONE

We examine the space-time seismicity patterns associated with the 40 largest (M≥5.6), intermediate-depth earthquakes that occurred in the Pamir-Hindu Kush zone between 1962 and 1984 to determine if any of these events were followed by aftershock sequences. To address this issue we examined data from the Soviet Central Asia regional earthquake catalog with two complementary techniques: (1) a subjective technique based on inspection of epicentral maps, and (2) a quantitative method which assumed that the background seismicity was a Poisson process and which detected aftershocks as significant departures, as quantified by the Poisson model, from background rates. Of the 40 events we examined, only three were identified by both methods as having clearly defined aftershock sequences. These three events are also remarkably similar in several other ways: (1) they occurred at almost the same point in space, (2) they have indistinguishable focal mechanisms, (3) they have similar magnitudes, and (4) they define two nearly equal time intervals (9.37 and 9.41 years). No other events in the Pamir-Hindu Kush zone exhibit comparable aftershock behavior. Six other events show rate anomalies, but the subjective method rated them negative or at least ambiguous. Four of these six we argue are caused by coincidental or man-made rate increases at the time of the mainshock. The remaining two are noteworthy because they are the two largest events to have occurred beneath the central Pamir. Each of these events shows a small, short burst of seismicity around the time period of the largest event in the sequence. We suggest two alternative hypotheses to explain our observations: (1) more events may have aftershock sequences, but the aftershocks are too small to be detected; or (2) the three events that had clear aftershocks are unique, and trigger a burst of aftershocks only because they have a special relationship to the the most active nest of activity in the Hindu Kush zone.

Velocity variations in the crust and upper mantle beneath the Tien Shan inferred from Rayleigh wave dispersion: Implications for tectonic and dynamic processes

We used data from nuclear explosions and earthquakes recorded by the Kyrgyzstan telemetered broadband seismic network to examine crustal structure in the Tien Shan region of central Asia. Seventy-nine measurements of fundamental and first higher mode Rayleigh waves were analyzed using the single-station multiple filtering technique to determine group velocity dispersions. We found that there was a distinct difference in Rayleigh wave group velocity values between different regions, but the data could be combined in three groups: (1) paths from the west and south across the Turan platform and western Tarim Basin, (2) paths across the eastern Tarim Basin, and (3) nuclear explosions from the Lop Nor test site. Group velocity values were inverted to derive one-dimensional shear wave velocity models for the Tien Shan region. Results reveal crustal thickness varying from 50 to 60 km with thickness increasing toward the eastern Tien Shan region in agreement with published crustal thickness maps for the region. Array analysis indicates that paths along the eastern Tien Shan and Tarim-Tien Shan are affected by lateral refractions and multipathing, but we argue that these have a small effect on the measured dispersion curves. Our results indicate that upper mantle velocities in the eastern Tarim and Tien Shan are significantly lower than those beneath the western Tien Shan and Turan platform. Our results suggest that the boundary between these two regimes lies somewhere west of the depocenter of the Tarim Basin and east of Kashgar, China.