The crustal shortening that drives the development of a mountain belt is not simply accommodated by uniform thickening of continental crust, nor by subduction of one block underneath the other. Instead, the deformation is typically localized into separate zones in which activity varies over time and space [e.g., Noblet et al., 1996]. A consequence of such variability is the complex structural relationship and superposition of structural styles preserved in all mountain belts, which poses a major problem in the assessment of the importance of individual orogenic processes [Burchfiel et al., 1992; Molnar and Lyon-Caen, 1988]. Because of the accumulation of deformation in mountain building, datasets spanning different timescales will provide distinct and sometimes apparently contradictory information about the tectonic activity in the orogen.
Geodetic measurements and observations from earthquake seismology may be used to determine rates and the spatial differentiation of deformation with high precision over a short time period[ e.g., Abdrakhmatov et al., 1996; Lukk et al., 1995]. However, the historic timescale is too short to adequately span the length of a typical seismic cycle of several hundred to several thousand years, whereas deformation viewed on a time scale spanning the Quaternary period is often sufficiently long enough to activate significant deformation sources, but short enough that they do not evolve or the regime changes markedly. Observations of the Quaternary deformation preserved in offsets of landforms and young deposits therefore allow one to bridge the gap between the limited historic timescales of geodesy and earthquake seismology and the integrated longer timescale of the geologic record of orogeny [e.g., Arrowsmith et al., 1996; Brgmann et al., 1994a; Yeats et al., 1997] . Furthermore, if the contribution to the regional deformation rate by recurrent earthquakes can be determined and compared to geologic estimates of deformation magnitude, the importance of earthquake-related deformation in the absorption of continental collision can be evaluated. Thus, by investigating active deformation in combination with long-term deformation in the Quaternary, space is substituted for time, and the cumulative effect of deformation and its role in longer-term mountain growth can be investigated.
In this study, we will use geologic and geomorphic observations of the Quaternary deformation in the Pamir-Tien Shan convergence zone from the northwestern sector of the India/Eurasia collision zone to illustrate how convergence between two continental plates is absorbed and differentiated, and how mountain fronts at the leading edge of a collision zone are seismotectonically segmented. This will be accomplished by field investigations that extend our preliminary mapping further along strike and to other subparallel zones of deformation, and detailed paleoseismic investigations that will help define fault slip rates, and earthquake magnitude and recurrence along major structures of the convergence zone. The field investigations will be complemented by analyses of remote sensing data, and with morphometric investigations of digital topography of the region (both of which will be used to extend the results of the mapping to broader areas).
Our preliminary results show that the central part of the collision zone is characterized by a thrust belt with slip localized along a narrow zone of faulting. Transitions to the peripheral portions of the range front are accommodated by complex areas in which convergence is kinematically transferred via obliquely slipping transfer faults and thrust systems. This large-scale structural arrangement is also reflected in geomorphic zonation defined by systematic landform responses to the deformation, such as uplifted pediment and terrace surfaces, and areas of important landsliding, and it is further documented in the distribution of historic earthquakes. The preliminary minimum slip rate along the thrust belt (>6 mm/yr) implies that this narrow sector accommodates 10 % or more of the shortening of the India-Eurasia collision. No basement rocks are exposed in the uplifted range, indicating that the high deformation rate has not been associated with significant exhumation and thus may not have been sustained in the area for longer than 1-2 Myr.
These data will be useful in the determination of the deformation budget for the northwestern portion of the India-Eurasia collision zone. Our preliminary observations show the potential for the more detailed investigations that we propose. With more complete geologic and geomorphic mapping and better age control for landforms and deposits, we will define the activity of faults in the center of this zone of continental collision. We expect to compile published and unpublished geologic data for the Alai Valley region, identify the active faults within the region, determine their slip rates to various degrees of precision, and to define paleoseismic rupture timing and extent (Figure 3). Closely related and coordinated studies by Manfred Strecker and German colleagues at the Universities of Potsdam, Wrzburg, and Tbingen will define a Quaternary glacial chronology using cosmogenic and radiocarbon age control and determine the long term exhumation rate for the Trans Alai Range by apatite fission track analysis. The exhumation rate will be determined by both direct sampling of the range up to at least 5000 m, and by investigation of the unroofing sequence preserved in the mollasse deposits in the adjacent basin that have been uplifted by active thrust faulting. Furthermore, our results will complement recently completed [Hamburger et al., 1992; Lukk et al., 1995, Pavlis, 1997 #633], and ongoing [Burbank et al., 1997; Hager et al., 1996; Hamburger et al., 1996; Weldon et al., 1997] work to the northeast and southwest.
Regional tectonic setting
With elevations in excess of 7100 m and Paleozoic to Quaternary rocks documenting ongoing deformation since the Paleogene, the Trans Alai Range and the adjacent Alai Valley of Kyrgyzstan are well suited locations to document the geometry, rate, and spatial development of deformation structures within a young mountain belt (Figures 1 and 3). The Trans Alai Range defines the northern perimeter of the convex-northward Pamir Mountains, where Eurasian continental crust is subducting southward to depths in excess of 200 km. The Pamir Mountains consist of amalgamated Paleozoic and Mesozoic terranes that were sutured to southern Eurasia and further deformed and displaced northward in the course of the Cenozoic collision processes. Truncation of Cretaceous and Paleogene facies boundaries, systematic anomalies in paleomagnetic declination, and the pronounced northward deflection of a Paleozoic suture imply that the outer rim of the Pamir has been overthrust onto southern Eurasia by approximately 300 km [Burtman and Molnar, 1993], Figure 2.
The Trans Alai is located in the updip projection of the southward dipping zone of seismicity and is bounded by northvergent thrust faults (Main Pamir Thrust system; also called the Trans Alai thrust in this area [Burtman and Molnar, 1993]); to the east and west the orogen is bounded by important dextral and sinistral strike-slip faults, respectively (Figures 1 and 2). The Main Pamir Thrust system separates the Trans Alai from the intramontane Alai Valley; to the north it is bounded by the Alai Range of the Tien Shan Mountains, which constitute a Variscan mountain range reactivated and uplifted to elevations higher than 5000 m in the course of the collision process. Burtman and Molnar  inferred that the thrusting along the Trans Alai is part of a thrust system soling into a detachment horizon which also accommodates a major thrust at the northern boundary of the Tien Shan. The 70-km-thick crust of the overriding Pamir indicates that along with subduction of Eurasian crust, the collision has been accommodated by significant crustal shortening. Active deformation quantified by repeated GPS measurements in the Pamir, the adjacent Tien Shan Mountains, and its foreland, shows that shortening in this region is active and may account for about 30 mm/yr convergence between India and Eurasia [Michel et al., 1997], which is two thirds of the relative motion between India and Eurasia [Burtman and Molnar, 1993].
Review of summer 1996 observations in Alai region
In summer, 1996, Arrowsmith and Manfred Strecker worked in the Alai Valley-Trans Alai range front building on reconnaissance work from a 1993 trip by Strecker and colleagues [Strecker et al., 1995]. We present these data to illustrate the productivity of our association, to indicate the utility of the mapping and investigation of Quaternary landforms and deposits, and to demonstrate the potential for our proposed paleoseismic investigations (see the supplemental information section I of this proposal for a letter of intent to collaborate form Strecker and our Kyrgyz colleague, Andrey Korjenkov). A manuscript is about 5 hours of work away from submission to Journal of Geophysical Research, and we have steadily presented the results of that work [Arrowsmith and Strecker, 1996; Arrowsmith and Strecker, 1997]. The preservation of tectonic landforms and Quaternary deposits is extraordinary and the efficiency with which we were able to work without bureaucratic hindrance encourage a thorough documentation of the Quaternary deformation and paleoseismology of structures in the region.
The Trans Alai Range along the northern perimeter of the Pamirs of Kyrgyzstan is located at the leading edge of the India-Eurasia collision zone, where continental crust is subducted southward coupled with important neotectonic deformation (Figures 1-3). The range defines a 300-km- long section of this portion of the plate margin, which is divided into eastern (E of 72 degrees, 55' E; 39 degrees, 30' N), central, and western (W of 72 degrees, 15' E;39 degrees, 28' N) segments straddling the Kyrgyz/Tadjik border. These segments bound semi-independent seismotectonic blocks that record differential absorption of plate convergence (Figure 1-3). The 40-km-long central fault segment (the portion in the center of the geologic map and the Landsat image drape over digital elevation model in Figure 3) has dip-slip thrust fault offsets and is linked to the other segments via NW-striking transfer faults with dextral strike-slip kinematic indicators.
The Trans Alai mountain front is comprised of thrust-faulted Neogene conglomerates and Paleogene gypsiferous mudstones that are separated by an erosional unconformity and are unconformably overlain by faulted Quaternary gravels. Along the central segment, the active mountain front is straight due to repeated offset of a regionally correlative fluvioglacial terrace gravels which grade into late Pleistocene moraine deposits upstream. Deformation of this inferred latest Pleistocene terrace is concentrated in a very narrow zone <10 m wide at the rangefront (see Figures 4-7). Figure 6 shows topographic profiles along fluvial terraces in basins draining the Trans Alai range front. The profiles illustrate that no folding or other distributed deformation is evident, and thus the deformation is localized at the rangefront. Furthermore, the lack of deformation of the terraces indicates that the thrust faults must be planar downdip a similar distance as the extent of the profiles [Arrowsmith et al., 1996]. The relationship between the moraines and the terrace surfaces as well as the lack of distributed deformation are documented by high precision topographic mapping (Figures 4, 6). The vertical offset of the correlated terrace surface is 0 m at the western transfer zone; it rapidly increases to 13 m 4 km E; and reaches a maximum about 8 km E of the transfer, but diminishes again to 0 m at the eastern transfer (see Figure 7). The asymmetry of the offset distribution may be indicate that the central segment interacts mechanically with the segments to the west, but not with those to the east [Willemse et al., 1994a].
The 20 - 40 m deep incision of the modern drainages offers the opportunity to observe the near-surface geometry of the principal mountain bounding fault. The fault zone is usually characterized by a roughly east-west striking and southward dipping (between 35 and 45 degrees) thrust fault (see below). The exposure of the fault zone is often subtle, but reveals folded gravel beds in the hanging and footwalls, as well as small shear zones parallel to the main trace, emphasized by fault -parallel sheared gravels. The well exposed fault zone at the Syrinardjar River is exemplary in this respect as it comprises numerous small lensoid shear bodies and oblique fault strands (Figure 3). The fault at Syrinardjar overrode organic material that is part of a sheared body in the lower portion of the exposed fault, which yields a corrected 14C age of about 4400 years BC. Based upon our precise topographic surveying (using an electronic total station), and inferring that the organic materials were deposited in front of a fault scarp that was active during the accumulation of the Qt3 gravels, the slip rate could be determined. We assume that the deposition of the gravels was rapid. Therefore the date represents a maximum age for the Qt3 surface. The vertical offset of the Qt3 surface is 18 m, and with a dip of the fault zone of 27 degrees, a total slip of 40 m can be obtained, resulting in a minimum Holocene dip-slip rate on the principal thrust fault of 6.3 mm/yr (40 m/6.4 kyr). This slip rate is broadly compatible with the only other determination for the area of a minimum of 2-4 mm/yr, which is based on the overthrusting of Stone age artifacts [Burtman and Molnar, 1993; Nikonov et al., 1983]. These data are intriguing but we believe preliminary. Better determination of the distribution of slip in space and time along the faults in this region will provide important constrains on our understanding of the relationship of fault activity to continental collision.
Although the Quaternary displacements typically rupture coarse terrace gravels, in several places (Komansu, Minjar and Tashkunguey rivers), we noted wedge-shaped deposits of matrix-supported fine silt and sand with coarse pebbles to cobbles that were both massive and stratified. These deposits extend from a sharp wedge boundary about 2 m below the upper portion of the fault scarp upward and grade into the soil profiles of the hanging and footwalls. We interpret these features as colluvial wedges that formed after a major surface rupture which were overridden in the last slip event along the range front. The slip in the last event was at least 1.5 m based upon our observations at Minjar (Figure 5). The preservation of such a relationship encourages further investigation.
Neotectonic deformation in the western transfer occurs in a 20-km-wide zone (see Figure 3 and the dextrally slipping fault segments in the western portion of the area). The transfer faults are characterized by large striated fault surfaces, pressure ridges at stepovers, aligned springs along straight mountain fronts, and beheaded stream channels. The faults terminate in dextrally oblique thrust faults which bound more sinuous mountain fronts and uplifted terrace and pediment surfaces. The westernmost transfer fault records the structural transition to predominantly N vergent thrusting that persists for 10s of km to the W via an asymmetric flower structure with the N vergent low- angle thrusts rooted in a steeply S-dipping dextral shear zone (Figure 3). In contrast, the eastern segment deformation is widely distributed and geomorphically less evident, but the transfer also takes place in a structurally complex zone of transfer faults (Figure 3). The westward asymmetry in the offset distribution along the central segment implies that it may mechanically interact with the structures of the western transfer and western segments and not with those to the east (Figures 3, 7). The distribution of historic ruptures mapped by Russian geologists (1974 Mw 7.1 Markansu and 1978 Mw 6.7 Zaalay earthquakes) and systematic geomorphic zonation of uplifted pediments, fluvioglacial terraces, and areas of important landsliding mimic these structural trends [Nikonov et al., 1983]. These preliminary observations suggest that collisional plate margins are not exclusively comprised of broad zones of diffusely distributed deformation, but may also exhibit discrete seismotectonic zonation reminiscent of ocean/ocean or ocean/continent convergent plate boundaries.
We propose the investigation of the northern portion of the Pamir arc in the region of the Alai Valley, Alai Range (part of the Tien Shan), and the Trans Alai (part of the Pamirs). We expect to extend our previous research by continued geologic and geomorphic mapping and paleoseismic investigations of selected structures in the area. Those field observations will complement and be done in coordination with the efforts of Manfred Strecker and German colleagues to develop a Quaternary glacial chronology and to determine exhumation rates for the Trans Alai range. The field observations will be anticipated by detailed interpretation of remotely sensed imagery, including the development of digital elevation models by softcopy photogrammetry. We will develop our interpretations of the field and remotely sensed observations with 3 dimensional modeling of the faulting and with combined geomorphic and tectonic modeling of the development of select tectonic landforms.
Interpretation of remotely sensed imagery
In order to document features in the field and identify mapping targets and strategy, we expect to interpret remotely sensed imagery of the Pamir-Tien Shan convergence zone. These data include Russian aerial photographs of the region, Declassified Intelligence Satellite Photography (DISP), and Landsat Thematic Mapper imagery. All of the data, including the Landsat imagery, are already in hand by Arrowsmith and Strecker. In fact, Sean McManus, a masters student working with Arrowsmith, has begun a pilot study of these data in the central Trans Alai region Figure 3 inset and [McManus et al., 1997]. The DISP are particularly useful because of their high resolution (~3-5 m) and stereo coverage. These data were acquired by the first generation of U. S. photo reconnaissance satellites in the 1960s and excellent coverage is available because of the strategic importance of central Asia for monitoring Soviet and Chinese military developments at that time [Wheelon, 1997]. The photos were declassified by executive order in 1995 and we have found them to be inexpensive and useful. The DISP stereo coverage will permit us to develop digital elevation models of selected areas soft copy photogrammetry (MIPS and ERDAS Imagine are programs that we will use). This process uses autocorrelation of coreferenced pixels in stereo images and applies the displacement of objects inherent in the parallax to determine an elevation model that best fits the observed data. We expect to manipulate these data using our remote sensing analysis tools (ERMapper), and develop a Geographic Information System in Arc-Info that includes these data, their interpretation, field observations, and fault model results (see below). A Geographic Information System (GIS) is defined as "An organized collection of computer hardware, software, geographic data, and personnel designed to efficiently capture, store, update, manipulate, analyze, and display all forms of geographically referenced information" [(ESRI), 1993]. Any type of field data that is geographically referenced may be entered into the GIS database. These data may then be analyzed and the results displayed in the form of a map. The principle advantage of a GIS over another type of database (i.e. FoxPro, Microsoft Excel, etc.) is that a GIS' information is spatially referenced. Not only can analyses be performed on data attributes, but the data's position may also be taken into consideration when performing the analyses.
Field observations--geologic and geomorphic mapping
During the first field season, we will map landforms and Neogene and younger deposits from where we stopped in summer 1996 at Tyuksu (see Figure 3) to the east for approximately 60 km at a scale of ~1:30,000 (on the DISPs and Russian airphoto bases) with occasional more detailed mapping on topographic bases developed in the field with Total Station surveying equipment (such as the map shown in Figure 4). We mapped about 60 km of rangefront in one month of field work in summer 1996, so this is a reasonable target for the first field season.
We will map selected portions of the southern foothills of the Tien Shan in the second field season. These areas will be identified by interpretation of the remotely sensed imagery. The purpose of this mapping is to evaluate the significance of other structures in the region in the deformation budget. For example, we expect that the main Trans Alai fault system is the most important structure accommodating the convergence in the region. However, we do not know how important, if at all, structures noted in passing in the southern Tien Shan may be. If they are active, we will document their geometry and the offsets of deposits and landforms in order to determine the rate of deformation associated with them. These structures may include both surface rupturing faults as well as buried faults and folds. Preliminary investigations in summer 1997 by Lothar Ratschbacher (University of Wrzburg) and Martina Schwab (University of Tbingen) indicate that the convergence in the Tien Shan may be accommodated by conjugate strike-slip faulting, implying that the stress magnitudes may change from those consistent with predominantly thrusting in the central Trans Alai (maximum compressive principal stress horizontal and the minimum stress vertical) to strike-slip with the maximum and minimum stresses in the horizontal plane. Such an explanation suggests that the magnitudes of the intermediate and least compressive stresses are similar, and thus they can switch positions. We will explore other interpretations of these data and others during the analysis phase of the project.
Field observations--paleoseismic trenching
During the two field seasons, we will spend several weeks (about half of the field time) performing paleoseismic trenching investigations. Target sites such as the Minjar and Komansu scarps shown in Figures 4 and 5 have been identified along the central segment. During our mapping efforts, we will identify other sites. The selection of sites (along with their suitability for preservation of offset dateable materials) will depend on appropriate coverage of the important structures in the region. Therefore, we expect to trench two sites per field season: one site along the central segment of the Trans Alai, one in the western transfer/western segment, one in the eastern segment, and one in the southern Tien Shan. Thus we will spend about two weeks on each site. We will make detailed topographic maps of each site before trenching. Equipment used for maintenance of irrigation systems for wheat farming in the Alai Valley includes bulldozers and backhoes. We will work with our Kyrgyz colleagues to hire this equipment for the excavations. For safety, the trenches will be shored with thick beams and plywood purchased in Fergana. Meter grids will be defined on the trench walls, and materials will be identified and mapped. During summer 1996, we found dateable charcoal and other organic material in natural exposures. The preservations of features such as those shown in Figure 5 are encouraging and we expect to identify previous slip events as well as longer term slip rates along these faults.
The high precision topographic data collected during our field investigations as well as that possibly developed using the DEM autocorrelation of the DISPs will be combined with our dates of selected landforms and offsets in the trench investigations to calibrate models for the morphologic development of the tectonic, glacial, and fluvial landforms in the region [Arrowsmith, 1995; Arrowsmith et al., 1995; Avouac et al., 1993; Avouac, 1993; Hanks, 1997; Hanks et al., 1984; Nash, 1980; Nash, 1981; Nash, 1984; Nash, 1986]. This approach, called morphologic dating, has been used to interpret the development of landforms such as fault scarps, fluvial terraces, and glacial moraines [Anderson and Humphrey, 1989]. By first calibrating the transport rate on features of known age, we can then determine the ages of other features in the region, assuming that the same transport processes operate of those surfaces. While this approach does not have the precision of the paleoseismic investigations, it does provide a broader view of the development of landforms in the region. Thus, we can take topographic data such as that shown in figure 4 and simulate the development of the scarp and infer the slip rate along the thrust fault. Because of the relative ease of this analysis, we will complement the trenching observations with morphologic analyses.
Three dimensional mechanical models of faulting
In our analysis of the data from this project, we will use 2 and 3 dimensional mechanical models of faulting to interpret the slip distributions that we infer based upon variable fault geometries and fault driving stresses inferred from our field observations, as well as those interpreted from regional investigations of seismic data [Bossu and Grasso, 1996; Lukk et al., 1995]. This approach has been followed by ArrowsmithÕs colleagues from Stanford University, where they have investigated slip distributions along a variety of idealized faults [Brgmann et al., 1994b], and in particular have focused on slip along normal faults [Willemse et al., 1994b]. The same approach and tools may be used to interpret slip distributions along thrust and strike-slip faults. We will use a recently developed method applying polygonal displacement discontinuities and allowing for uniform displacement discontinuity and or uniform tractions along each element and a remote stress field [Thomas, 1993]. These techniques calculate the elastic fields (displacements, stresses, strains, and tilts) caused by any amount of strike-slip, dip-slip, and opening displacement discontinuity or traction on any number of arbitrarily striking and dipping fault elements within a linear elastic half-space. Such methods are commonly used for modeling displacements associated with slip events on faults and the stress changes associated with earthquakes[Reasenberg and Simpson, 1992]. George Hilley, a PhD. student working with Arrowsmith at ASU, has completely overhauled the POLY3D code and he and Arrowsmith are confident in the results generated by the method. For example, we have analyzed the slip distribution that might develop along the central segment, western transfer, and western segment of the Trans Alai system. A uniform remote maximum compressive stress striking 140¡ (inferred based upon observations of opening mode fracture orientations in uplifted Neogene gravels in the Trans Alai foothills; we have also used a 90¡ orientation) was resolved onto each of the fault planes. These traction boundary conditions were then applied to the faults to determine the variation of slip along and between the fault segments. This simple model illustrated the slip distribution that might develop in a Mw 7 earthquake that ruptured the Trans Alai fault system. Maximum dip-slip of about 1 m develops in the center of the central segment and that is similar to the offset interpreted at Minjar. This modeling tool will be used to interpret fault interaction and thus rupture scenarios based upon our field observations. While we do not expect to gather all of the data such as stress magnitudes, down dip fault extents, etc. that might be used to constrain such modeling more completely, we do expect to infer what is possible from the data that we collect (such as the shallow downdip fault orientation indicated from the relatively planar upstream terrace profiles along the central segment Figure 6) and test hypotheses for the possible consequences of continental collision on active faults.
Arrowsmith and a graduate student (ASU group) dedicated to this project full time will perform the proposed work. He and the student will travel to Germany on the way to central Asia for the two field seasons and consult with and then work with Strecker and colleagues in the field. Our previous experience in summer 1996 will help us to anticipate logistical aspects of the work. See the attached letters of intent to collaborate from Manfred Strecker at U. Potsdam and from Andrey Korjenkov of the Institute of Seismology, National Academy of Sciences, Kyrgystan. Note that we will also consult during visits by Strecker to the American Geophysical Union meeting. Arrowsmith will work directly with Strecker and the graduate student in the geologic and geomorphic mapping.
Arrowsmith will be the lead in the geologic mapping, paleoseismic trenching, morphologic dating, and mechanical modeling. Both Arrowsmith and Strecker have experience in the interpretation of paleoseismic data from trenches and the effects of earthquake surface rupture [Arrowsmith et al., 1997; Arrowsmith and Rhodes, 1994; Sims et al., 1993; Strecker and Hecker, 1995]. We will present preliminary findings of our work at the Fall Meeting of the American Geophysical Union in 1998, and more complete results at that meeting in 1999. We expect to author at least 1 major paper in a refereed journal such as the Journal of Geophysical Research. We also anticipate that significant results will be highlighted in a paper submitted to Science. Because of the significant mapping that will be accomplished as part of this project, we will publish our maps as part of the major paper or alone with brief text explanations. Our data including Arc-Info coverages will be available to the research community. Furthermore, at Arizona State University, we will manage a World Wide Web site dedicated to quick updates and presentation of these data to the community. Arrowsmith has experience in the development of web sites, and we think that this is an important means of disseminating information. For examples of existing web pages, see http://www.public.asu.edu/~arrows/ (Arrowsmith homepage), http://www.public.asu.edu/~arrows/pamir-intro.html (web pages already developed related to this project), http://www-glg.la.asu.edu/~at/carrizo/ (web pages describing previous active tectonics work by Arrowsmith and colleagues in the Carrizo Plain), http://www-glg.la.asu.edu/~at/scm/index.html (web pages describing ongoing research on the geologic and geomorphic development of the Santa Cruz Mountains of California), and http://www-glg.la.asu.edu/~at/ (web pages established for Active Tectonics, Quantitative Structural Geology, and Geomorphology Research Group at Arizona State University).
As mentioned above and as demonstrated by our productivity in 1996, we expect that the logistics of working the Pamir-Tien Shan region will not be difficult. A summer 1997 field campaign in the regions to collect samples for fission track analysis by Lothar Ratschbacher (University of Wrzburg) and Martina Schwab (University of Tbingen) was successful. We have a good working relationship with Lothar and Martina and will coordinate our research with theirs as well as check with them for up-to-date logistical information.
Preservation of dateable material
In this high elevation area with limited vegetation, preservation of abundant dateable material is a potential difficulty. However, in 1996 during fairly cursory investigations of natural exposures, we were able to find adequate material for three radiocarbon dates, one of which was discussed above. The other two provide results consistent with the slip rate that we determined. We will select sites where as much vegetation upstream is present in order to optimize these efforts. Furthermore, in similar settings, [Michel et al., 1997; Weldon et al., 1997] have used radiocarbon age control to determine earthquake chronologies and fault slip rates.
Regional scale tectonics problems addressed by this research
The work of [Burtman and Molnar, 1993] which concluded that the outer rim of the Pamir has been overthrust onto southern Eurasia by approximately 300 km and that the Trans Alai is located in the updip projection of the southward dipping zone of seismicity provides a valuable and well developed framework within in which we expect to place our research questions and research. Furthermore, the research has implications for the study of the cumulative effects of short and long term deformation:
Accommodation of India-Eurasia relative motion
Burtman and Molnar  suggested that the 44 mm/yr (+/-5 mm/yr) convergence between India and Eurasia at the latitude of the Pamir is partitioned into northward underthrusting of India below the Salt Range of Pakistan, distributed deformation within the Hindu Kush-Pamir, and southward underthrusting of Eurasia below the Pamir. Convergence and underthrusting in the Pamir are accommodated at the surface not only along the Main Pamir Thrust, but also within the Tien Shan and adjacent regions to the north. Michel, et al.  determined that the overall shortening between the central Pamir and the Kazakh platform (Eurasia) is about 3 cm/yr. Southward tilting of the Alai Valley basement and deformation along the southern margin of the Tien Shan as seen on seismic reflection profiles, as well as the geodetically determined velocity distribution [Michel et al., 1997] all indicate that significant shortening may be accommodated within the Tien Shan and adjacent regions. Preliminary estimates of the magnitude of active shortening in the Pamir region relative to the slip rate determined in this study for the central segment of the Trans Alai thrust fault indicate that the Main Pamir Thrust alone accommodates a minimum of 10% of the total India-Eurasia relative motion.
Annihilation of an intermontane basin by collision tectonics
The progressive closure of the Alai Valley by the northward advance of thrusting exemplifies the annihilation of an intermontane basin by collision tectonics. The continuity of the Paleogene sedimentary basin from the Tadjik Depression in the west across the Alai region to the Tarim Basin to the east was ended by north-vergent and northward propagating thrust faulting [Burtman and Molnar, 1993], which caused an approachment between the Tien Shan and Pamir Mountains. Thrust faulting along the central segment of the Trans Alai stepped outward from that fault which places Paleogene over Neogene units to develop the active range bounding thrust which places Neogene over Quaternary. This event transferred a section of the Neogene basin fill to the hanging wall of the Main Pamir thrust system (Figure 3). The structural embayment of these Neogene rocks coincides with the central segment of the Main Pamir Thrust that we have defined based upon Quaternary geologic relationships. Its simple geometry may be consistent with the recency with which it has been established.
Due to high sediment yields of the Alai piedmont rivers, the Kyzilsu is being pushed to the northern margin of the Alai Valley where the elevations vary from about 3000 m in the east to about 2600 m at the west end of the central segment (Figure 2). In the area of the central segment the Alai Valley is about 15-20 km wide. Within the narrow valley west of the confluence with the Tar Ata River, the elevation of the Kyzilsu is 2400 m. Thus the position of thalweg, the elevation, and the broad northward-sloping Trans Alai piedmont indicate that deposition of sediments derived from the Trans Alai Range has created a wedge of sediment that fills the Alai Valley and onlaps the southern side of the Tien Shan. This is seen on N-S seismic reflection profiles, which show a pronounced wedge geometry with facies onlapping northward. The width is reduced to a narrow valley and gorge in the area of the western thrust segment of the Main Pamir Thrust, immediately west of the Tar Ata River. In this narrow valley the thrust front is separated from the Tien Shan by the Kyzilsu River. Active landslides remove much of the material that is uplifted by slip along the thrust and deliver it to the river. [Pavlis et al., 1997] suggested that the Kyzilsu acts as a lateral conveyor belt that removes material delivered to the deformation front by thrusting, and that the river facilitates advance of the thrust front by constantly removing the mass that would otherwise be deposited as mollasse in a foredeep such as the Alai Valley proper. However, further to the west, thrusting has completely closed this intramontane basin, the Kyzilsu (now the Surkob River) flows across the hanging wall, and backthrusts place Tien Shan basement rocks over the Cretaceous-Neogene sequence [Hamburger et al., 1992; Leith and Alvarez, 1985; Pavlis et al., 1997]. Thus the progressive closure of this intermontane basin is documented by the change from completely closed basin to active thrusting and landsliding in a gorge to the asymmetric foredeep. We conclude that the Alai Valley is being closed in a zipper-like fashion whereby the fault splays of the western transfer zone represent the closing zipper in the western Alai Valley.
Recent construction of large mountain ranges in the Pamir-Tien Shan
One of the implications of the high slip rate along the Main Pamir Thrust, yet lack of basement rocks exposed in the Trans Alai is that the range may have begun to uplift only a few Myr ago. If the uplift rate due to slip along the Trans Alai system has been constant and the individual surface rupturing thrust faults accommodate the total offset in turn, it would take about 1 Myr to build the current relief (4 km/4 km/Myr) with no erosion. If we assume that thrusting did not significantly thicken the cover rocks, we can use the correlative and adjacent Tadjik Basin sedimentary cover thickness of about 10 km. The 10 km of cover rocks minus 4 km of relief with no basement provides an estimate of 6 km for the possible (maximum) exhumation in the Trans Alai. The thickness of the post-Paleogene sediments of the Alai Valley is about 3 km at their thickest adjacent to the Trans Alai. Therefore, an estimate of the post-Paleogene vertical offset along the Trans Alai thrust fault is 6 km denudation + 4 km of relief + 3 km of foredeep sediments which is equal to 13 km. Assuming a minimum vertical offset rate of 4 mm/yr, the Trans Alai thrust system and associated relief could have initiated about 4 Myr ago or more recently. This rapid construction of a major mountain range resonates with the general conclusions of [Molnar and Lyon-Caen, 1988] regarding the spatial and temporal variability of deformation in mountain belts and is consistent with the results of [Abdrakhmatov et al., 1996] who extrapolated high historic (geodetically determined) shortening rates to geologic timescales for the eastern Kyrgyz Tien Shan and concluded that they may have developed in the last 10 Myr. They implied that the construction of the mountains was a response to an increased regional northward directed horizontal compression following an abrupt rise of the Tibetan Plateau [Harrison et al., 1992; MacDonald, 1991; Molnar and Lyon-Caen, 1988].
Pages maintained by
Last modified Saturday, August 5, 2000 5:55 PM