Gobi-Altay, Mongolia - GeoScienceWorld

7 downloads 0 Views 445KB Size Report
TECTONIC SETTING. This gigantic paleolandslide occurred in the. Gobi-Altay mountains in Mongolia, along the eastern segment of the Bogd fault, a strike-slip.
Gigantic paleolandslide associated with active faulting along the Bogd fault (Gobi-Altay, Mongolia) Hervé Philip* Jean-François Ritz*

Laboratoire de Géophysique et Tectonique—CNRS, Université de Montpellier II, Place Eugène Bataillon, 34095 Montpellier Cedex 05, France

ABSTRACT On the basis of analyses of satellite imagery, aerial photographs, and field observations, we describe the occurrence of one of the largest paleolandslides (50 km3 ) ever recognized in an intracontinental domain. The slide occurred along the active Bogd fault in the Gobi-Altay mountain range in Mongolia. Morphological and structural analyses of the relationships between the landslide and the area affected by active tectonics suggests that this gigantic mass movement was associated with surface faulting during a strong earthquake. INTRODUCTION Among other geologic phenomena related to tectonically active continental regions, mass movements may affect large areas and cause great damage. Large landslides are commonly observed in mountainous regions and/or under wet climatic conditions. Outstanding examples include the Tsergo Ri landslide (100 km3 ) in the Langthang valley in north-central Nepal, which occurred about 40 ka ago (Ibetsberger, 1996), the Saidmarreh landslide (20–30 km3 ) in southwestern Iran, which occurred in prehistoric times (Harrison and Falcon, 1938), and the Flims landslide (12 km3) in Switzerland, which took place during an interglacial stage (Heim, 1920). Mass movements also are commonly triggered by ground shaking from earthquakes (e.g., Keefer, 1984; Jibson, 1996). Among the modern earthquaketriggered spectacular landslides are the rock avalanche (~0.1 km3 ) that occurred in 1970 after the Mw 7.9 earthquake in the Cordillera Blanca in Peru, (Plafker et al., 1971) and the Bairaman landslide (0.2 km3 ) in New Britain, Papua New Guinea, that was triggered by an Mw 7.1 earthquake on May 11, 1985 (King et al., 1989). Although landslides are commonly triggered by ground shaking from strong earthquakes, it is rare to observe a direct relationship between landsliding and coseismic rupture, as has been proposed, for example, for the Beni Rached landslide (15 km wide) that was associated with the El Asnam fault *E-mail: Philip: [email protected]; Ritz: [email protected].

rupture of the October 10, 1980, M 7.3 earthquake (Philip and Meghraoui, 1983). On the basis of satellite images, aerial photographs, and field observations, we present here the first results of an analysis of one of the largest continental landslides (estimated volume: ~50 km3 ) whose geometry and mechanics appear directly associated with surface faulting during a strong earthquake. Until now, the structure corresponding to the landslide has not been understood as a gravitational phenomenon. TECTONIC SETTING This gigantic paleolandslide occurred in the Gobi-Altay mountains in Mongolia, along the eastern segment of the Bogd fault, a strike-slip fault that ruptured most recently on December 4, 1957, during one of the strongest recorded intracontinental earthquakes (M 8.3) (Florensov and Solonenko, 1963) (Fig. 1). The surface rupture follows the northern edge of the two massifs of Ih Bogd and Baga Bogd, which stand 2500 m above the Valley of Lakes. Surface ruptures were also observed along a set of active thrust-related ridges, the Dalan Türüü and the Gurvan Bulag ridges, situated north and south, respectively, of the Ih Bogd massif, and the Hetsüü ridge situated north of the Baga Bogd massif. Florensov and Solonenko (1963) provided an outstanding description of ground-surface effects related to the seismic event. Kurushin et al. (1997) furnished an updated quantification of coseismic surface breaks, and preliminary results from a long-term slip rate analysis (Ritz et al., 1995) and

from a paleoseismic expedition (Schwartz et al., 1996) showed that the Bogd fault is probably a low-slip-rate fault (~1 mm/yr) capable of generating major earthquakes (M ≥ 8) separated by recurrence intervals of several thousands of years. One of the most impressive structures described by Florensov and Solonenko (1963) is the Bitut landslide (2.5 × 4 km2), located in the central part of the Ih Bogd massif (Fig. 1). No equivalent gravitational feature related to the 1957 event is known from the Baga Bogd massif. A new interpretation of remote images related to a detailed-scale morphologic study of this massif allowed us to show that an even larger landslide occurred in the past. MORPHOLOGICAL AND STRUCTURAL ANALYSIS OF THE BAGA BOGD LANDSLIDE A satellite image (Fig. 2A) allows recognition of the giant landslide as a morphologically distinctive structure; the landslide includes a zone of relief extending over an area of 20 × 15 km2 and standing 150–200 m above alluvial fans deposited north of the Baga Bogd massif. It affects sedimentary rocks (Goshu Formation of Berkey and Morris, 1927) that appear in the ancient alluvial fans located along the northern front of the GobiAltay. The Goshu Formation is composed of homogeneous gray, poorly cemented, coarse- and fine-grained conglomerates and breccias with a rhythmic structure caused by changes in the size of the fragments from one horizon to the next, with alternations of thin horizons of argillaceous sands containing angular gravelly material (Florensov and Solonenko, 1963). The landslide has a wellpreserved morphology and is limited laterally by two parallel and linear north-trending scarps. Two main morphological domains can be distinguished (Figs. 2 and 3): (1) To the north, half of the slide mass is deformed into east-trending topographic ridges that curve laterally on both sides. This pattern can be interpreted as frontal compres-

zz  ,, yy y,{{ ||  zz ,,y  Geology; March 1999; v. 27; no. 3; p. 211–214; 5 figures.

Figure 1. Sketch map of 1957 Gobi-Altay earthquake ruptures and location of Baga Bogd paleolandslide.

211

sional structures that rotate at each lateral edge due to friction at the boundaries of a northward mass movement. In front of the toe of the landslide, a 2-km-wide rim of chaotic deposits, now mostly covered by modern sand, may correspond to soft material pushed and overthrust by the frontal part of the sliding mass. (2) To the south, the topography appears less deformed and consists of large allochthonous blocks separated by depressions filled by young, fine-grained lake sediments. The largest block, which forms the southeastern quarter of the surface of the landslide, has a flat surface cut by two major east-trending grabens 0.5–1 km wide. At its southeastern corner, the slide mass is affected by a pattern of northwest-, east-, and northeast-trending normal faults (altitude point: 1533 m; Fig. 2B). This set of small grabens parallels a remnant set of faults at the northeastern edge of the in situ Goshu Formation. There, paleoscarps limited by steep northeast-trending normal faults and an east-trending fault are located north and south, respectively, of an east-trending row of remnant hills situated just at the border of the in situ Goshu Formation (altitude point: 1808 m; Fig. 2B). We interpret these graben structures to be related to initial normal faulting before landsliding, and we consider the remnant hills as the edge of a backward-tilted block. Blocks in the southwestern part of the landslide are smaller and separated from one another by small valleys and fans. They are characterized by planar surfaces that dip to the south. In the field, we observed the southward general dip of alluvial bedding. The strike of the bedding is N140°E and the dip is 20°S at the western edge of the main slide block (Ih Hetsüü structure of Kurushin et al., 1997) (Fig. 2B). Analysis of aerial photographs of the surface of the Ih Hetsüü block shows paleodrainage flowing to the north (Fig. 4). This feature indicates the backward rotation of the block, consistent with the south-dipping bedding of the alluvial deposits. We also observed small remnants of the surface that did not slide and are still attached to the northern slope of the Baga Bogd massif (south and southwest of altitude 1805 m in Fig. 2B). These observations suggest that this large block is part of the Goshu alluvial surface that was deposited along the foothills of the Baga Bogd massif before sliding. The well-preserved surface at the southeastern part of the landslide belongs also to the Goshu Formation (Florensov and Soloneko, 1963). As mentioned above, this surface correlates with the undisplaced and deeply incised part of the in situ Goshu Formation lying at the foothills of the Baga Bogd massif (Fig. 2). The analogy of the southern contours of the slide blocks with the northern contours of the border of the in situ Goshu Formation allows crude estimation of the displacement. A 3 km minimum displacement is measured between B and B′ (Fig. 2B), corresponding to the displacement of the Ih Hetsüü block, and a 4.5 km maximum displacement is measured between C and C′ along

cross-section line A–A′ (Fig. 2B), corresponding to the displacement of the main slide mass. The Goshu Formation also crops out west and east of the landslide area (Fig. 2), as remnants of alluvial fans that were deposited along the Baga Bogd massif. There, the erosional overprint is much weaker than in the landslide area. The stronger erosion of the in situ Goshu Formation southward of the landslide is explained by the sudden drop of the drainage base level due to the gap that opened behind the landslide and allowed a strong and localized backward erosion of the formation (Fig. 2). The gap is now partly filled by modern alluvial material that postdates the sliding

(Figs. 2 and 3). To the west, small depressions between the dislocated blocks are filled with finegrained lacustrine sediments and are re-incised by recent drainage. This alluvial pattern shows that the drainage was dammed by the massive landslide; the lakes so created emptied eventually. AGE OF THE LANDSLIDE The sediments that are affected by the landslide belong to the Goshu Formation inferred to be of early Quaternary age (Berkey and Morris, 1927). The top surface of the main slide block shows few traces of incision, which suggests that the landslide occurred not long after the deposition of the Goshu

Figure 2. A: SPOT2 panchromatic image of area affected by landslide (KJ 247-259/0; March 10, 1996, used by permission of Centre National d’Etudes Spatiales 1996—Distribution SPOT IMAGE). B: Morphological and structural map drawn from satellite images, aerial photographs, topographic maps, and field observations.

,

,, ,, ,,

,,,,,,,,,,,,,, ,,,,,,,,,,,,,, ,,,,,,,,,,,,,, ,,,, ,,,,,,,,,,,,,, ,,,, ,,,,,,,,,,,,,, ,, ,,,, , ,,,,,,,,,,,,,, ,, ,,, ,,,, ,,,, , ,,,, ,, ,, ,, ,,,,,,,,,,,,,, ,,,, ,,,,, ,,,, ,,,, ,, ,,, , ,,, ,,, ,, ,,,, ,,,,, ,,,,, ,,, ,,,, ,,,, ,,,,, ,,,,, ,,, ,,,,, ,,, ,,,, ,,,,, ,,,,, ,,, ,,, ,, ,, ,,,, ,,,,, ,,,,, ,,, ,,, ,, ,, ,,,, ,,,,, ,,,,, ,,, ,,, ,, ,, ,,,, ,,,,, ,,,,, ,,, ,,, ,,,, ,,,,, ,,,,,

,, ,,,,,, ,,,,,,,,,, ,, ,,,,,,,, ,,,,,,      ,    ,  ,  ,

212

,    ,, ,    ,,,, ,  ,,,,  ,  ,    ,,,, 

GEOLOGY, March 1999

Formation. The surface is not covered by recent alluvial fans and shows, on the dislocated and tilted blocks, only weak indications of an ancient hydrographic network that drained the surface before it slid (Fig. 4). On the other hand, the erosion and incision of the in situ Goshu Formation and the periphery of the landslide suggest that a significant period of time has passed since the landslide. After the landslide occurred, the gap that separates the landslide mass from its source was partly filled by alluvial sediments. Two main fan sequences are distinguished (Fig. 2B). 10Be dates of uplifted alluvial surfaces along the Gurvan Bulag ridge, south of Ih Bogdo (see Fig. 1), suggested that the last two main alluvial pulses occurred at the beginning of the last interglacial (~130 ka) and postglacial intervals (~15 ka), respectively (Ritz et al., 1997). By comparison with this Gurvan Bulag alluvial chronology, the Baga Bogd postlandslide alluvial setting suggests that the landslide is older than the late Quaternary period. STRUCTURAL RELATIONSHIPS BETWEEN PALEOLANDSLIDE AND ACTIVE FAULT The paleolandslide is located at the northern flank of the Baga Bogd massif, which is bounded by the active Bogd fault. In 1957, this fault was reactivated and broke the ground surface. The geometry and kinematics of ruptures along the Baga Bogd massif correspond mainly to thrust

faulting associated with left-lateral movement (Kurushin et al., 1997). Nevertheless, along the segment that limits the in situ Goshu Formation, between altitude points 1805 and 2504 m (Fig. 2B), analysis of 1:10000 and 1:25000 aerial photographs of the fault zone revealed steep dipslip movement toward the north (see intersections of the fault trace with stream channels indicated in Fig. 2B). This observation suggests that there was a slight northward collapse of the in situ Goshu Formation during the 1957 event. During the 1957 event, the Hetsüü ridge also ruptured with dip-slip thrust movement and broke the ground surface. This north-northwest– trending thrust fault ends abruptly to the east against the western border of the landslide. There, the thrust component is absorbed by rightlateral strike-slip movement along a north-trending fault that cuts through the westernmost slide block (Fig. 2B). At the base of the Hetsüü overthrusting ridge, corresponding to uplifted hills consisting of Goshu Formation deposits (Florensov and Soloneko, 1963), the curvature of the 1957 scarp defines a low-dip-angle thrust fault, along which we observed sheared gypsum deposits. These deposits belong to the gypsiferous clays of the Khung-Kure Formation (Berkey and Morris, 1927) (Fig. 2B). We hypothesize that the basal shear surface beneath the Baga Bogd landslide was also localized within these waterimpregnated deposits.

Figure 3. Panoramic view of landslide taken from altitude point 1808 m, at border of in situ Goshu Formation, southeast of landslide and topographic cross profile A–A′ (location in Fig. 2B). a: In situ Goshu Formation. b: Gap opened behind landslide and filled by postsliding fans. c: Graben structures. d: Main well-preserved flat surface. e: Compressional ridges. Note backward-rotated Ih Hetsüü block in background between a and b.

,,,,,, ,,,, ,,,,

,

Figure 4. Morphological sketch of Ih Hetsüü tilted block (location in Fig. 2B). GEOLOGY, March 1999

TRIGGERING MECHANISM Having not performed a geotechnical analysis (i.e., static and dynamic slope-stability analysis) (e.g., Jibson and Keefer, 1993; Jibson, 1996), we cannot conclude definitely that the landslide was triggered by seismic activity. Nevertheless, three of the six criteria defined by Crozier (1992) support a seismic origin: (1) the ongoing seismic activity of the Bogd fault system, (2) the spatial relationships with the active fault, (3) the gigantic size of the landslide. Criteria 2 and 3 would be sufficient to argue for a seismic origin from Solonenko’s (1977) description of earthquaketriggered landslides. Moreover, the blockier appearance, coupled with the well-limited depositional area of the landslide, suggests a seismically induced slide (e.g., Perrin and Hancox, 1992). The evidence of a slight collapse of the in situ Goshu Formation during the 1957 earthquake (see above) also supports the idea of a seismic origin for the paleolandslide, but to explain the amplitude of the paleolandslide, an additional mechanism is required. The fact that the alluvial-fan material that fills the gap, behind the landslide, is also found cutting through the Hetsüü ridge (Fig. 2B) shows that the Hetsüü ridge predates the landsliding. Because the surface breaks associated with the 1957 faulting event stopped against the western border of the paleolandslide, we infer that the pre-landslide geometry of the Hetsüü ridge was similar to that during the 1957 seismic event. A lateral surface fault probably already existed at the eastern termination of the Hetsüü ridge, which cut through the alluvial Goshu Formation. We propose that this preexisting lateral fracturing within the Goshu Formation favored the initiation of the landslide. A reconstruction of the geometry of the landslide just before and just after sliding (Fig. 5) shows that the landslide can be defined as a soil block slide according to Varnes’s (1978) classification. The Goshu Formation is thicker upstream than downstream, and the displacement decreases from the head to the toe, where the alluvial cover is stacked in numerous tilted slices. The basal slip surface appears very flat. Analysis of topography and satellite images shows that the landslide dammed the Valley of Lakes, isolating several depressions (Fig. 2). In these ponded depressions, we observed in the field that the deposits were perfectly horizontal from the base to the top. This shows that the sliding occurred during one event. Keefer (1984) reported that solid-block slides induced by earthquakes can occur with frontal slopes as gentle as 5° in areas with high water tables, probably involving saturated basal shear surfaces. In the Baga Bogd landslide area, the average slope of the alluvial surfaces is about 3°. For this mass of material (~100 billion tons) to move several kilometers down a 3° slope in a single event requires some unusual basal shear conditions. Taking into account the sliding distance, the landslide, once triggered, must have had sufficient momentum (mass × velocity) to 213

|||| {{{ zzz yyyy    ,,,,  } € „… ˆ   



 ~ ƒ ‚ ‡ † y z { } ~  y z { | } ~  € y z { | } ~  € y z { | } ~  € y z { | } ~  € y z { | } ~  € y z { | } ~  € y z { | } ~  € y z { | } ~  € y z { | } ~  € y z { | } ~  € ,      ,        ,        ,        ,        ,        ,        ,        ,        ,        ,        ˆˆˆˆ ‡‡‡ ††† ………… €€€€ „„„„ }}}} ~~~   ‚‚‚ ƒƒƒ |||| {{{ zzz  yyyy 

    



   ,,,,  ‚ ƒ … † ‡  ‚ ƒ „ … † ‡ ˆ  ‚ ƒ „ … † ‡ ˆ  ‚ ƒ „ … † ‡ ˆ  ‚ ƒ „ … † ‡ ˆ  ‚ ƒ „ … † ‡ ˆ  ‚ ƒ „ … † ‡ ˆ  ‚ ƒ „ … † ‡ ˆ  ‚ ƒ „ … † ‡ ˆ  ‚ ƒ „ … † ‡ ˆ  ‚ ƒ „ … † ‡ ˆ 



 y z { } ~  



  y z { | } ~  € 



  y z { | } ~  € 



  y z { | } ~  € 



  y z { | } ~  € 



  y z { | } ~  € 



  y z { | } ~  € 



  y z { | } ~  € 



  y z { | } ~  € 



  y z { | } ~  € 



  y z { | } ~  € ,      ,        ,        ,        ,        ,        ,        ,        ,        ,        ,        ˆˆˆˆ ‡‡‡ ††† ………… €€€€ „„„„ }}}} ~~~   ‚‚‚ ƒƒƒ |||| {{{ zzz  yyyy 

    



   ,,,,  ‚ ƒ … † ‡  ‚ ƒ „ … † ‡ ˆ  ‚ ƒ „ … † ‡ ˆ  ‚ ƒ „ … † ‡ ˆ  ‚ ƒ „ … † ‡ ˆ  ‚ ƒ „ … † ‡ ˆ  ‚ ƒ „ … † ‡ ˆ  ‚ ƒ „ … † ‡ ˆ  ‚ ƒ „ … † ‡ ˆ  ‚ ƒ „ … † ‡ ˆ  ‚ ƒ „ … † ‡ ˆ 



 



  



  



  



  



  



  



  



  



  



  ˆˆˆˆ ‡‡‡ ††† …………  ,,,,       |||| 



yyyy zzz {{{ }}}} ~~~ €€€€ „„„„   ‚‚‚ ƒƒƒ  

,   y }  ,        

y z { | } ~  €  ‚ ƒ „ ,        

y z { | } ~  €  ‚ ƒ „  ,,,,        |||| 





 yyyy zzz {{{ }}}} ~~~ €€€€ „„„„ ˆˆˆˆ   ‚‚‚ ƒƒƒ ………… ††† ‡‡‡ ,        y z { | } ~  € ,        y z { | } ~  € …

  … † ‡ ˆ

  … † ‡ ˆ      





 }}}} ~~~ €€€€ „„„„ ˆˆˆˆ   ‚‚‚ ƒƒƒ ………… ††† ‡‡‡ ,,,,,, ,,,,,, ,,,,,, ,,,,,, ,,,,,, , ,,,,

,,,, ,,,,

,,,,,,,,,,,,,,,,,,,,,,,,,,,,, ,,,,,,,,,,,,,,,,,,,,,,,, ,,,,,,,,,

Figure 5. Interpretative cross section of landslide just before and just after sliding.

keep moving after the strong shaking stopped. Such gigantic mass movement probably required a condition of no frictional or cohesive strength along its basal shear plane, possibly achieved by fluidization in the basal sediments (i.e., the gypsiferous clays of the Khung-Kure Formation). Reconstruction of the initial position of the soil block slide (Fig. 5) shows that the front of the future landslide was located within the flood plain, in the Valley of Lakes. Therefore, drainage conditions at the base of the alluvial Goshu Formation were closely dependent on groundwater conditions in the flood plain. It is possible that the landslide occurred in wet climatic conditions during which the water table was much higher than at present.

CONCLUSIONS The Baga Bogd landslide appears to be one of the largest landslides ever described from a continental interior. Its morphological, structural, and lithological aspects allow us to define the Baga Bogd paleolandslide as a gigantic soil block slide after Varnes’s (1978) classification. Its relationship with the active Bogd fault system suggests that the landslide was earthquake induced. Preexisting surface fracturing, strong shaking, peculiar stratigraphic conditions (coherent alluvial cover underlain by gypsiferous clays), and perhaps a wet climate combined to produce this extraordinary mass movement. ACKNOWLEDGMENTS We thank J. Jackson and A. Bayasgalan for fruitful discussions and help to support this work; R. Jibson and R. Bürgmann for helpful reviews; K. Berryman, M. Mattauer, D. Schwartz, A. Taboada, H. Wust, and R. Yeats for their comments; P. Molnar, A. Wendt, and C. Wibberley for helping to improve the manuscript, and A. Delplanque for the drawings. Field study was mainly supported by the French Embassy in Mongolia and the Laboratory of Geophysics and Tectonics— CNRS in Montpellier.

214

REFERENCES CITED Berkey, C. P., and Morris, F. K., 1927, Geology of Mongolia: U.S. Museum of Natural History, Natural History of Central Asia, v. 11. Crozier, M. J., 1992, Determination of paleoseismicity from landslides, in Bell D. H., ed., Landslides (Glissements de terrain) (Proceedings, International Symposium, Christchurch, New Zealand): Rotterdam, Netherlands, A. A. Balkema, v. 2, p. 1173–1180. Florensov, N. A., and Solonenko, V. P., editors, 1963, The Gobi-Altai earthquake: Moscow, Akademiya Nauk USSR, 391 p. [in Russian] (English translation by Israel Program for Scientific Translations) Jerusalem, 1965, 424 p. Harrison, J. V., and Falcon, N. L., 1938, An ancient landslip at Saidmarreh in southwestern Iran: Journal of Geology, v. 66, p. 296–309. Heim, A., 1920, Geologie der Schweiz: Basel, p. 399– 401. Ibetsberger, H. J., 1996, The Tsergo Ri Landslide: An uncommon area of high morphological activity in the Langthang valley, Nepal: Tectonophysics, v. 260, p. 85–93. Jibson, R. W., 1996, Use of landslide for paleoseismic analysis: Engineering Geology, v. 43, p. 291–323. Jibson, R. W., and Keefer, D. K., 1993, Analysis of the seismic origin of landslides—Examples from the New Madrid seismic zone: Geological Society of America Bulletin, v. 105, p. 421–436. Keefer, D. K., 1984, Landslides caused by earthquakes: Geological Society of America Bulletin, v. 95, p. 406–421. King, J., Loveday, I., and Schuster, R. L., 1989, The 1985 Bairaman landslide dam and resulting debris flow, Papua New Guinea: Quaternary Journal of Engineering Geology, v. 22, p. 257–270. Kurushin, R. A., Bayasgalan, A., Ölzybat, M., Enhtuvshin, B., Molnar, P., Bayarsayhan, C., Hudnut, K. W., and Lin, J., 1997, The surface rupture of the 1957 Gobi-Altay, Mongolia, earthquake: Geological Society of America Special Paper 320, 143 p. Perrin, N. D., and Hancox, G. T., 1992, Landslidedammed lakes in New Zealand—Preliminary studies on their distribution, causes and effects, in Bell, D. H., ed., Landslides (Glissements de terrain) (Proceedings, International Symposium,

Printed in U.S.A.

Christchurch, New Zealand): Rotterdam, Netherlands, A. A. Balkema, v. 2, p. 1457–1466. Philip, H., and Meghraoui, M., 1983, Structural analysis and interpretation of the surface deformations of the El Asnam earthquake of October 10, 1980: Tectonics, v. 2, p. 17–49. Plafker, G., Ericksen, G. E., and Fernandez, C. J., 1971, Geological aspects of the May 31, 1970 Peru earthquake: Seismological Society of America Bulletin, v. 61, p. 543–578. Ritz, J.-F., Brown, E. T., Bourlès, D. L., Philip, H., Schlupp, A., Raisbeck, G. M., Yiou, F., and Enkhtuvshin, B., 1995, Slip rates along active faults estimated with cosmic-ray–exposure dates: Application to the Bogd fault, Gobi-Altaï, Mongolia: Geology, v. 23, p. 1019–1022. Ritz, J.-F., Brown, E. T., Bourlès, D., Raisbeck, G., Yiou, F., Hanks, T., Kendrick, K., Finkel, R., Carretier, S., Chery, J., Philip, H., Enktuvshin, B., Galsan, P., and Schlupp, A., 1997, Comparison of uplift rates on 20 and 100 ky timescale along the Gurvan Bulag Thrust Fault (Gobi-Altay, Mongolia) using cosmic-ray exposure dates: Eos (Transactions, American Geophysical Union), v. 78. Schwartz, D., Hanks, T., Prentice, C., Kendrick, K., DePetris, A., Bayasgalan, A., Rockwell, T., Thorup, K., Ritz, J.-F., Lund, W., Hanson, K., and Ruzchitch, V., 1996, 1996 Gobi-Altay, Mongolia, Paleoseismology expedition: Initial results: Eos (Transactions, American Geophysical Union), v. 77, p. 462. Solonenko, V. P., 1977, Landslides and collapses in seismic zones and their prediction: International Association of Engineering Geologists Bulletin, v. 15, p. 4–8. Varnes, D. J., 1978, Slope movement types and processes, in Schuster, R. L., and Krizek R. J., eds., Landslides—Analysis and control: National Academy of Science, Transportation Research Board, Special Report 176, p. 11–33. Manuscript received July 9, 1998 Revised manuscript received November 5, 1998 Manuscript accepted November 17, 1998

GEOLOGY, March 1999