T32B-01
Subduction and accretion across the Ganges-Brahmaputra Delta: is it seismogenic?
Huge earthquakes generally contribute a major portion of the strain at subduction zones, but are rare, and may thus be overlooked in hazard analysis. The India-Sunda boundary is increasingly oblique to convergence from northern Sumatra to the eastern Himalayan Syntaxis and had not experienced a well-documented great subduction earthquake prior to the Mw9.3 1,300km rupture of 26 Dec 2004. The 500km long Arakan segment north of the 2004 rupture may have ruptured in 1762, when a regional earthquake caused changes in coastal elevation and ejection of sand and water from anticlinal ridges. Further north, we found some evidence for a 1548 event which may have ruptured the 300km long Tripura segment from Chittagong to the Shillong Plateau. In this segment, the subduction zone encounters the huge clastic wedge of the Ganges-Brahmaputra delta, which is prograding onto oceanic or transitional crust. This wedge is being accreted into a wide active foldbelt that extends beyond its topographic expression into the subsurface of the delta as far as the main estuary (Megna). This 50 km-wide blind area of foldbelt shows pedological evidence (the Tippera Surface) of Holocene uplift and is higher than the active part of the delta to the west by 3-4m, an amount that may be expected in a one or more large subduction ruptures. Interseismic tectonic deformation is evident in both drainage changes and GPS data. An eastward shift of the major channel of the Ganges into the Bay of Bengal from the Hooghly (at Calcutta) to the present Meghna Estuary since the 1600s may reflect subsidence and tilting of the delta. Preliminary results from a 6-station continuous GPS network in Bangladesh show rapid regional subsidence, that may reflect a combination of compaction and lithospheric flexure in response to tectonic and sediment load, and north-south shortening, consistent with the Shillong Plateau moving south relative to India. E-W motions are inconsistent with aseismic accretion along the Tripura segment and the network's resolution in this direction is expected to improve soon. In light of the 1762 and 2004 examples, the lack of a well-known historical precedent, high pore-fluid pressure and oblique convergence at the Tripura segment do not rule out large subduction earthquakes across the delta, which need to be included in Hazard estimates. Their effects are difficult to predict given the unusual conditions of a >20 km foredeep and rapid sedimentation, but they are likely to be catastrophic for the 200M people living in this area.
T32B-02 INVITED
Unprecedented Massive earthquakes in the Himalaya driven by Elastic Strain stored within the Tibetan Plateau?
In view of the absence of historical precedent for the 1600 km long Sumatra/Andaman rupture of the eastern edge of the Indian plate in 2004, we question whether similar earthquakes without historical precedent could also occur in the Himalaya. A seismic slip deficit corresponding to a single Mw=8.7 earthquake currently exists, and although no great earthquake has ruptured the Himalayan frontal thrusts in the past 300 years, evidence that larger earthquakes have done so between 900 and 1505 is now available in the form of trench investigations of slip in the range 5-21 m, over rupture lengths of 200-300 km. The magnitude of these and possibly larger earthquakes is unknown. We investigate possibly magnitudes for Himalayan earthquakes by developing a synthetic scaling law relating renewal time to rupture length. We first emulate the accumulation of present-day elastic strain near the southern edge of the Tibet plateau using 3D boundary element methods, and then, after incrementing this strain by a time interval of many hundreds of years, we investigate the amount of slip that corresponds to different rupture lengths. Our interseismic model uses the inferred subsurface geometry of the Indian plate (20-40 km depth) and the observed GPS velocity field to 1000 km north of the Himalaya as constraints, with minor adjustments to the shallow geometry beneath the Greater Himalaya as our only variable. Assuming that frictionless aseismic slip occurs beneath the plateau, we find that its rate decays to zero 25 km south of what we have previously considered the locking line, with dip increasing to 15 deg. at 12-18 km depth, consistent with focal mechanisms in this region. An unexpected result is that vertical deformation is twice the width (70 km) & half the amplitude (3 mm/yr) of that assoicated with uniform-slip elastic models, though still consistent with the lower error envelope of leveling data. To account for the 14-21 mm/yr long term advance of the Himalaya over India we find that elastic strain stored more than 300 km north of the edge of the Tibetan plateau must participate in driving the largest Himalayan earthquakes (8.2<Mw<8.7). We find that a renewal time of &~ 500 years is consistent with the length/Mw scaling of recent 6.5<Mw<8.2 earthquakes, but that a doubling of this interval must occur to account for the largest slip events. We find also that, just as in the Andaman/Nicobar islands, regions of the Himalaya that have experienced recent Mw=7.8 earthquakes are vulnerable to much larger earthquakes sooner than would be anticipated by seismic gap theory.
T32B-03
Evidence For An Earthquake Barrier Model From Mw7.8 Kokoxili (Tibet) Earthquake Slip-Distribution
The slip distribution of the Mw~7.8 Kokoxili (Tibet, 2001) earthquake has been measured at high resolution using optical correlation of Spot satellite images, which provides both the parallel and perpendicular components of the horizontal co-seismic slip. This reveals a variation of the horizontal slip at a scale of ~20km along-strike. Anti-correlation of slip parallel and perpendicular to the fault indicates transfer of slip from the horizontal to the vertical component at the ends of segments. These features suggest a rupture model with segments separated by strong persistent geometric barriers. The unexpected ending of the rupture south of the main fault can be explain by such a structure, which bears important implications for the initiation and rupture directivity of the next earthquake.
T32B-04
Paleoseismic Investigation of the Ranong and Khlong Marui faults, Chumphon Province, Southern Thailand
The Ranong and Khlong Marui faults are northeast-southwest trending structures in the Isthmus of Kra, southern Thailand, that apparently link the extensional regimes of the Mergui Basin in the Andaman Sea and the Gulf of Thailand. These faults are depicted commonly as strike-slip faults, acting as conjugate structures to the dominant northwest-southeast trending strike-slip faults, in Southeast Asia. These faults are parallel to the predominant structural grain in the Carboniferous rocks of peninsular Thailand. In addition, they appear to be bounding structures for several Tertiary basins, including the onshore parts of the Surat Thani basin and the offshore Chumphon basin. Initial remote sensing studies showed that both faults have relatively subdued geomorphic expressions. Field reconnaissance investigations indicated a lack of youthful tectonic geomorphology along the Khlong Marui fault and ambiguous evidence for recent movement along the Ranong fault. Fault exposures along both fault trends and on minor parallel faults in the region indicated that, rather than predominantly strike-slip motion, these faults have experienced up-to-the-west reverse movement. Because of its more youthful geomorphic expression, several sites along the Ranong fault were chosen for paleoseismic trenching. Initial trench exposures indicate an absence of Holocene movement. Some exposures indicate the possibility of Late Tertiary-Early Holocene vertical movement. These investigations are currently ongoing and we hope to report our conclusions at the Fall Meeting.
T32B-05
Coseismic Strike-Slip Offsets of Surface Rupture Produced by 2001 Mw 7.8 Kunlun (Tibet) Earthquake, Measured From High-Resolution IKONOS Imagery and Field
The magnitude Mw 7.8 (Ms 8.1) Central Kunlun, earthquake occurred on 14 November 2001, in the Kunlun mountain area, north Tibet, and produced extensive surface rupturing over a distance of >400 km along the pre-existing Kunlun fault zone (Lin et al., 2002, 2003). The deformation characteristics and strike-slip displacement distribution of the coseismic surface rupture zone revealed by field investigation, seismic data, and satellite images have been described in a series of previous papers by ourselves and other groups. However, there are some conflicts on the amounts of strike-slip offsets reported by these groups. We demonstrated that large coseismic strike-slip offsets of >10 m and up to 16.3 m occurred in several locations by our filed observations (Lin et al., 2002, 2003), but other groups (e.g. Xu et al., 2002) showed no offset amounts over 10 m. Such differences in the field-measured offset amounts might be caused by the different measurement methods and the lake of field data due to the remote high mountain region and bed weather in the study area where is too difficult to access and to work long time in site. In this study, we focus our attention on the coseismic strike-slip offsets of the 2001 surface rupture by using high-resolution IKONOS and QUICKBIRD images and compare the analytical results from the images with the field data. Analyses and interpretations of 1m-resolution IKONOS, and 0.6m-resolution QUICKBIRD images reveal that the 2001 coseismic strike-slip offsets measured from IKONOS images range from 2 m up to 16.7 m, generally 3-8 m, which are generally consistent with those measured immediately in the field after the earthquake. The coseismic surface ruptures along which offsets were observed are mainly composed of a numerous of en echelon shear faults and cracks which are concentrated in a deformation zone ranging from a few meters up to ~500 m in width. The offsets measured along individual shear faults or cracks are typically 2-7 m, but up to >10 m in several locations. Our results show that high resolution remote sensing imagery provides a powerful tool for measuring coseismic strike-slip offsets and detecting ground deformation produced by large earthquakes in the remote and high mountain Tibet region.
T32B-06 INVITED
Airborne laser swath mapping of the metropolitan Taipei area
Surface faulting commonly occurred in the tectonically active Taiwan region. Investigating and characterizing crustal surface faults become urgent because damaging earthquakes usually resulted from crustal surface faulting. In recent years, remote sensing technology has provided mapping tools that can pinpoint active structures more accurately. Among the mapping tools, airborne laser swath mapping, also known as airborne LIDAR (Light Detection and Ranging), has received much attention for its ability of collecting reliable and high-density elevation data. Also among its practical ability are filtering tree tops and buildings to provide earth ground models. Such digital elevation models are invaluable for better characterizing and evaluating urban crustal faulting and tectonic activities. Because of earthquake threats for the metropolitan Taipei area, we have organized a mapping and research team which uses airborne LIDAR to produce meter-resolution digital elevation models. These derived LIDAR models will subject to field verification and then be applied for active fault and geologic hazards research. In the Taipei region, two major earthquake sources were previously recognized as the Chinshan fault and the Sanchiao fault. The Chinshan fault is considered as a frontal thrust fault during the earlier episode of the Taiwan mountain collision. Around 0.5 Ma the stress field of northern Taiwan gradually turned into regional extension, normal faulting started to dominate the Taipei area. The Sanchiao fault, a normal fault, is a product of the crustal extension which contributes to the formation of the Taipei basin. However, the precise locations and the northern extension of the Sanchiao fault are still unclear and debatable. It is hoped that the LIDAR mapping project can unveil urban surface faults in a precise manner. We expect the project will produce reliable and high-precision digital elevation models for applications including evaluating tectonic and volcanic geomorphology in the populated metropolitan Taipei area. We have started reevaluating fault traces of active faults and will show preliminary results of the surficial expression of the Sanchiao fault as well as the Chinshan fault in the metropolitan Taipei area.
T32B-07
Maturity and Activity of Faults in a Backarc Region of Southwest Japan Inferred From Integration of Topography, Paleoseismology, Coseismic Slip and Fault Rocks
Back arc region of southwest Japan (San_fin district) is characterized by relatively low strain rate and sparse distribution of major active faults with high slip rate. The San_fin district, however, has been frequently hit by _gmedium-sized_h damaging earthquakes since the historical age, including the recent 2000 western Tottori earthquake (Mw 6.6). Our integrated study of topography, paleoseismology, coseismic slip and fault rocks revealed that the 2000 western Tottori earthquake was produced by a fault system with low maturity and low activity, but long history at least since the Miocene. The low maturity is indicated by narrow shear zones (several mm to cm wide) and poor traceability (several ten to hundred meters long) of their topographic expressions. We think that the low maturity of the fault system caused sporadic occurrence of coseismic surface ruptures only within a 6-km-long area, which is far smaller than the seismologically estimated ca. 20-km-long source fault. Trenching survey identified a preceding rupture event around 28 ka, and post-earthquake geodetic survey revealed 70 cm to 1 m coseismic slip. Based on the data, average slip rate of the fault system responsible for the 2000 earthquake is estimated to be less than 0.1 mm/y. Fault rock study has showed that NW-striking mesoscopic faults in Paleogene granite are distributed more densely in the aftershock area than the outside area. Cataclasites are characteristically distributed within the aftershock area, and a large number of Miocene andesite dikes have intruded into the granite in the fault-controlling NW direction. The facts indicate that the fault system responsible for the 2000 western Tottori earthquake has a long formation history at least since Miocene time. Inoue et al. (2002) have revealed that a 4-km-long parallel fault located about 2 km west of the aftershock zone ruptured between AD 770 and 1260, suggesting the relation of this event to the AD 880 historical earthquake. Our integrated survey also has identified two parallel fault systems about 10 km long in the vicinity of the 2000 aftershock area, 5 km southwest and 9 km northeast, respectively. Maturity of the latter fault system inferred from shear zone characteristics and topographic expressions is as low as the fault system that caused the 2000 earthquake. Excavation survey indicates that the fault system has repeatedly ruptured at intervals of more than several ten thousand years. Rather dense distribution of low-maturity and low-activity faults might cause frequent _gmedium-sized_h damaging earthquakes in the San_fin district.