G33C-0701
Coseismic Strain Steps of the 2008 Wenchuan Earthquake Indicate EW Extension of Tibetan Plateau and Increased Hazard South to Epicenter
The 2008 Wenchuan earthquake (Ms8.0) occurred at the east edge of Tibetan Plateau. It is the biggest seismic disaster in China since the 1976 Tangshan earthquake. To determine the effects of the earthquake on the deformation field of Tibetan Plateau, we collect and analyze continuing strain data of three stations before and after the earthquake in Tibetan Plateau observed by capacitance-type bore-hole strainmeters (Chi, 1985). We collect strain data in NS, EW, NE-SW and NW-NS directions at each borehole. Then we deduce the co-seismic strain steps at time point 14:28 of May 12, 2008 (at this time point the earthquake occurred) with the data before and after the earthquake using the least squares method. Our observation shows that in Tibetan Plateau significant co-seismic strain steps are accompanied with the 2008 Wenchuan earthquake. Extension in EW direction is observed at interior and north Tibetan Plateau which indicates a rapid EW extension of the whole Plateau. Field investigation shows that the 2008 Wenchuan earthquake is a manifestation of eastward growth of the Tibetan Plateau (Dong et al., 2008). Eastwards growth of the Tibetan Plateau results naturally in the extension of the Plateau in EW direction. Our co-seismic strain observation agrees well with the conclusion from surface rupture investigation. The magnitude of co-seismic strain step equals to five times of average year extensional strain rate throughout the plateau interior. Shortening in SE- NW direction is observed at the east edge of the Plateau. As hints that the eastward extension of Tibetan Plateau is resisted by Sichuan rigid basin which increases the potential earthquake hazard around the observation station, manifests the declaration from co-seismic stress changes calculation (Persons et al., 2008). Our observed co-seismic strain steps are in total lager than theoretical calculations of dislocation theories which indicate that magnitude of the great earthquake should be bigger than 7.9. Due to the special structures of Tibetan Plateau, only part of observed co-seismic steps can be explained well by dislocation theories.
G33C-0702 INVITED
Interferograms of Coseismic Deformation from the May 12, 2008, Sichuan, China, Earthquake obtained by ALOS/PALSAR
A devastating earthquake hit southern China on May 12, 2008, and claimed more than 80,000 casualties. This event is the first M8 earthquake in history in and around the Sichuan basin. The epicenter is located near the Sichuan basin, and aftershocks are aligned in the NE-SW direction parallel to the Longmen Shan fault zone, which suggests that this active fault zone bounding the basin and mountain ranges is the source of this event. It is important to reveal the fault movement from the viewpoints of future earthquake hazard evaluation and tectonics of collision zone. Japan's Advanced Land Observing Satellite (hereafter ALOS) has been conducting observation of the earth's surface since its launch on Jan. 24, 2006. ALOS is equipped with an L-band synthetic aperture radar, Phased Array L-band SAR (hereafter PALSAR), which is suitable for the observation of heavily vegetated areas. We analyzed ALOS/PALSAR images along 8 paths acquired before and after the mainshock and revealed entire features of coseismic deformation using the Gamma software and hole-filled SRTM DEM by Jarvis et al.(2008). We observe fringes around the Longmen Shan fault zone. In spite of different period and timing of acquisition of SAR images, several fringes can be traced across the interferograms. These continuous fringes can be recognized in an about 100km wide region surrounding the fault zone and regarded as coseismic deformations. We can count 8-9 fringes, i.e. 90-100cm LOS displacements, between the fault zone and Chengdu located about 50km away from the Pengguan fault, southern most trace of the Longmen Shan fault zone. On the other hand, we could not obtain correlation high enough in the source region due to large displacements. This low correlation zone can be traced nearly 240km along the surface faults. It is worth noting that numbers of fringes are almost the same on the both sides of the source region. This pattern suggests right-lateral motion on a steeply dipping fault plane. On the other hand, 6-7 broad concentric fringes are recognized in the image where the epicenter is located. This pattern corresponds to about 70-80cm LOS displacements away from the satellite around the epicenter and implies a low-angled thrust. We also performed pixel matching to estimate offsets in the region of low correlation. Positive and negative range offsets larger than 2m are found on the north and south sides of the Beichuan fault, central trace of the Longmen Shan fault zone, respectively. Clear difference in offsets cannot be seen along other faults such as Pengguan fault. This observation implies that the coseismic rupture is concentrated on the Beichuan fault. We calculated a theoretical interferogram with three segment fault model with total length of 240km. In this model, southwestern part is assumed to dip toward northwest by 33° with purely thrust motion, while central and northeastern part has a much steeper dip angle and large strike-slip component of slip. Coseismic slip may be concentrated in the shallower part of faults than 15km, especially in central and northeastern parts. The geodetic moment is estimated to be 7.31x1020Nm (Mw=7.84).
G33C-0703
Mapping the earthquake rupture and displacement field using correlation of ALOS PALSAR amplitude images: Application to the Mw 7.9 Sichuan earthquake, 12 May 2008.
This study concerns with the mapping of earthquake rupture in the presence of a high magnitude earthquake using remote sensing techniques. In particular, we focus on the Sichuan 2008 earthquake where large surface deformation field hampers the precise localisation of the earthquake rupture using InSAR techniques, even with L-band radar system. If radar interferometry is able to precisely map the seismic deformation on a wide area, near the rupture where the surface spatial displacement gradient becomes too steep, InSAR technique fails due to loss of signal coherence. To complement the deformation mapping in the close neighbourhood of the surface rupture, we propose to compute the line/column pixel offsets on radar images using sub-pixel correlation techniques along with a novel approach . The Sichuan 2008 earthquake surface displacements is expected to be of several metres. Therefore, sub- pixel correlation of 5-10m pixel sized radar images (such as ALOS/ASAR data)is therefore relevant to study the phenomenon. Two shortcomings could hamper the result: 1 - The offsets in the range direction contains a topographic component particularily in the high relief area of the Sichuan province. 2 - The offsets in the azimuth direction could be affected by ionospheric perturbancies, particularily the L - band radar imagery (such as ALOS/PALSAR data) In this paper, firstly we propose a novel approach based on PCA processing to separate surface deformation from image offset due high relief. Secondly , we show that the ionospheric perturbancies do not sistematically mask all the information in the images offset in the azimuth direction, allowing us to help mapping in certain cases the surface rupture as well as the near field azimuth displacement field. The results provide us with a good localization of the surface rupture as well as an estimate of the surface displacement along the seismogenic fault.
G33C-0704
Geologic Hazards Associated with Longmen Shan Fault zone, During and After the Mw 8.0, May 12, 2008 Earthquake
A magnitude 8.0 earthquake shook the northeastern margin of the Tibetan plateau, on May 12, 2008 along the Longmen Shan orogenic belt that marks the boundary between the Songpan Ganzi terrane and Yangtze block. The Tibetan plateau is expanding eastwards, and GPS observations show that surface motion directions are northeast relative to the Sizhuan basin where the earthquake occurred. This sense of motion of crustal blocks is the reason why the main faults in Longmen Shan are oblique thrust-dextral strike slip faults. There are three main parallel thrust/ dextral-slip faults in Longmen Shan. All three faults strike northeast and dip to northwest. The May 12 rupture extends 270 km along the fault zone, and the epicenter of the magnitude 8.0 earthquake was located in Wenchuan, 90 km WNW of Chengdu, Sichuan, China. The devastating earthquake killed at least 87,652 people and destroyed all the buildings in epicenter. The victims of the earthquake zone want to rebuild their homes immediately, but they need more suggestions about the geologic hazards to help them withstand future possible earthquakes. So after earthquake, we went to disaster areas from July 5th to 10th to get first-hand field data, which include observations of surface ruptures, landslides, features of X joints on the damaged buildings, parameters of the active faults and landslides. If we only depend on the field data in accessible locations, we can only know the information of the ruptures in these positions, and we can't learn more information about the whole area affected by the earthquake. The earthquake zone shows surface rupture features of both thrust and strike-slip fault activities, indicating oblique slip followed by thrusting during the May 12 earthquake. In my talk, I will show the general regional geological disaster information by processing the pro- and post-earthquake satellite data. Then we combine the raw field data and regional geology as the restrictive conditions to determine the deformation, movement directions, the main stress directions and what kind of the hazards (rock falls, landslides, barrier lakes, strong ground shaking, ruptures) happened in different segments along the earthquake zone. If we can exactly know the main stress directions and degrees of deformation in the place where towns and buildings are located, it will help people determine where it is relatively safer to rebuild, and improve the strength of the earthquake-resistant buildings, and further it will be helpful to decrease damage in the next earthquake.
G33C-0705
Ground deformation associated with the 2008 Sichuan Earthquake in China, estimated using a SAR offset-tracking method
Introduction: A catastrophic earthquake struck Chinafs Sichuan area on May 12, 2008, with the moment magnitude of 7.9 (USGS). The hypocenter and their aftershocks are distributed along the western edge of the Sichuan Basin, suggesting that this seismic event occurred at the Longmeng Shan fault zone which is constituted of major three active faults (Wenchuan-Maowen, Beichuan, and Pengguan faults). However, it is unclear whether these faults were directly involved in the mainshock rupture. An interferometry SAR (InSAR) analysis generally has a merit that we can detect ground deformation in a vast region with high precision, however, for the Sichuan event, the surface deformation near the fault zone has not been satisfactorily detected from the InSAR analyses due to a low coherency. An offset-tracking method is less precise but more robust for detecting large ground deformation than the interferometric approach. Our purpose is to detect the detail ground deformation immediately near the faults involved in the Sichuan event with applying the offset-tracking method. Analysis Method: We analyzed ALOS/PALSAR images, which have been taken from Path 471 to 476 of ascending track, acquired before and after the mainshock. We processed SAR data from the level-1.0 product, using a software package from Gamma Remote Sensing. For offset-tracking analysis we adopt intensity tracking method which is performed by cross-correlating samples of backscatter intensity of a master SAR image with samples from the corresponding search area of a slave image in order to estimate range and azimuth offset fields. We reduce stereoscopic effects that produce apparent offsets, using SRTM3 DEM data. Results: We have successfully obtained the surface deformation in range (radar look direction) component, while in azimuth (flight direction) no significant deformation can be detected in some orbits due to gazimuth streaksh that are errors caused by ionospheric effects. Some concluding remarks are as follows: On the Beichuan F. and its northeastward extension, a clear boundary of a motion toward and away from the satellite can be recognized just along the fault, which is almost consistent with a right-lateral fault motion. On the other hand, in the southwestern region from the Beichuan city where the three major faults are running almost parallel, two boundaries of motions can be recognized; On the Beichuan F. there are a clear displacement boundary in range component, while on the Pengguan F. a boundary can be identified in azimuth component rather than in range, suggesting that the seismic ruptures proceeded with different fault motions at each fault. For the Wenchuan-Maowen F., no significant displacement boundary can be recognized. Acknowledgments: PALSAR data are provided from Earthquake Working Group and PIXEL (PALSAR Interferometry Consortium to Study our Evolving Land surface) under a cooperative research contract with JAXA. The ownership of PALSAR data belongs to METI (Ministry of Economy, Trade and Industry) and JAXA.
G33C-0706 INVITED
2008 Wen-Chuan Mw 8.3 Earthquake: Dynamic Deformation Analysis for Seismic Surface Waveforms with 1-Hz GPS PPP Results
Some 1-Hz GPS data in China during Sichuan Wen-Chuan Mw8.3 Earthquake are processed with instantaneous Precise Point Positioning (PPP) technique using software PANDA developed in Wuhan University. Near-field GPS stations are located 30km away but lost data after 10s while several far-field sites are 300km~1200km away from epicenter. Results indicate that plane displacements exceed 40cm within 10seconds in near field while the post-seismic deformation is northwestward at the magnitude of around 30cm. Seismic surface waveforms as LOVE and Rayleigh waves can be seen clearly in the dynamic deformation process of station XANY and CHGO (in KunMing city, Yunnan Province). Love wave's first vibration is in northwestward and then southeastward, after about one cycle, Rayleigh wave comes. Love wave is at the speed of about 4km/s, Rayleigh's is about 2.85km/s. There are two deformation pulses in Rayleigh waveforms during the period of about 120 seconds monitored at station XANY, which is about 600km from epicenter. One pulse is in counter-clockwise movement and the other is clock-wise, which seems to reveal the fact that there are two kinds of rupture process in totally opposite direction. Spectral analysis shows that there are mainly two kinds of periods in the surface waves reflected by the 1-Hz horizontal deformations, one is about 5~10s (6s and 8s) and another is about 20s, corresponding to Rayleigh and Love wave respectively. Rayleigh's wavelength is about 20km, which indicates that the hypocenter is about in the depth of 10km.
G33C-0707
Interpreting and Integrating DInSAR Data of Mountainous Areas: The Case of the 2008 Sichuan Earthquake
ALOS PALSAR data is an excellent sensor for monitoring seismic events due to its long wavelength and very large critical baseline. Initial PALSAR interferometry results for the Sichuan Earthquake of May 2008 have demonstrated this very well and the results continue to be interpreted. However, in this particular case, the areas exhibiting the most significant land surface motion occurred in an extremely mountainous region of China which presents great challenges for the interpretation of the SAR results and its integration with other types of data, such as GPS. The overwhelming problem for the interpretation is that the foreshortened geometry greatly restricts any chance of an even distribution of coherent resolution cells. Furthermore, the extremes of layover and shadow can dominate images of mountainous regions, effectively rendering large areas without any reliable SAR measurements whatsoever. In this paper, we consider the incident geometry of the ALOS PALSAR system and describe the problems associated with the interpretation and geolocation of differential interferometry results using the Sichuan earthquake as a case study. We show the effect that this has on the characterisation of the seismic displacement and also on the ability to monitor associated small-scale events such as mass-movements and landslides. The study is conducted around the SE termination of the May 2008 rupture zone where fault slip and surface deformations were relatively low and topography is relatively high. We attempt a superficial reconciliation of the deformation pattern with structural geology. The interpretation shows that the motion sensed by the interferometry is a combination of effects from a number of earthquakes that occurred within the 46 day interval of the PALSAR image pair.
G33C-0708
Near real-time radar interferometric mapping of the 12 May 2008 Ms 8.0 Wenchuan Earthquake in China
The Wenchuan Earthquake (Ms 8.0, 12 May 2008) ruptured about 400km of faults in the Sichuan Province,
southwest of China. The PALSAR sensor onboard the ALOS satellite imaged the earthquake-stricken area
on seven overflight tracks before and after the quake, while the ASAR sensor onboard the Envisat satellite
provided images from two tracks.
Working closely with the Japanese Earth Remote Sensing Data Analysis Centre and the European Space
Agency, our team generated interferograms for most passes in "near real-time" with the support of off-shore
satellite ground receiving stations, i.e. within 18 hours of the post-quake overflights, and immediately provide
the results to the multiple Chinese authorities to guide the deployment of field survey teams and rescue
teams and to monitor after-shocks and landslides. Estimates of slip along the main rupture are three meters,
in good agreement with slip derived from geophysical modeling.
This study demonstrates that radar interferometry can be used to support post-disaster response and
emergency management. The crucial requirement is that there are sufficient radar satellites in orbit to
minimize the satellite revisit time.
http://www.gmat.unsw.edu.au/LinlinGe
G33C-0709
A fault model of the 2008 Wenchuan earthquake estimated from SAR measurements
A huge Mw7.9 earthquake struck the western Sichuan province, China, on May 12, 2008. The co-seismic
deformation is investigated using Synthetic Aperture Radar (SAR) data. The ALOS "Daichi" satellite data are
used to generate interferograms and pixel offset maps of the co-seismic surface displacement field.
Interferometric SAR data show edges of the rupture area clearly. The estimated length of the rupture area is
about 285km. Interferograms and pixel offset maps show location of surface ruptures. In northeastern part
of hypocentral area, surface rupture runs along the Beichuan fault. In southwestern area, interferograms
show large deformation around the Pengguan fault, which runs south of the Beichuan fault. Pixel offset
images indicate both the Beichuan fault and the Pengguan fault rupture on this area.
We invert the InSAR data for slip distribution of the Wenchuan earthquake rupture. We set geometry of
faults from fringe pattern of InSAR images and results of seismic wave inversion. Dip of northeastern faults
(55°) is steeper than southwestern fault (40°). We estimate amplitude and direction of slip on
each 5x5km sub-fault. Figure 1 shows estimated result. Near main shock area, estimated slip shows mostly
reverse fault motion and slightly right-lateral slip. Large slip is estimated at deep (15-30km) part. On the
other hand, estimated slip of other area shows large right-lateral motion and slip area locates at shallow (0-
10km) part. The maximum slip is 11.4m at near Beichuan, about 150km northeast from the hypocenter. The
seismic moment derived from estimated fault parameters is 9.2x1020 Nm (Mw7.9). These features are
consistent with other results, such as waveform inversion, field survey of surface ruptures.
G33C-0710 INVITED
Postseismic Deformation Monitoring of the 2008 Mw7.9 Wenchuan earthquake using GPS
We synthesize GPS data observed from a continuous array and monitor postseismic deformation of the 2008 Mw7.9 Wenchuan, China earthquake. The array is composed of three networks. Two of the networks, preexisted before the quake, are located in the northwest Sichuan basin and along the central Xianshuihe fault respectively. They are 15 sites in total, and have had continuous postseismic data recordings since the occurrence of the quake. The third network is composed of 25 sites, and was deployed in the vicinity of the seismogenic fault within 1.5 months after the quake. Preliminary analysis of the postseismic data reveals that: (a) Postseismic deformation is significant, up to a couple of centimeters for stations located within tens of kilometers from the seismogenic fault. (b) Postseismic deformation is unevenly distributed across the fault: It is significantly larger and decays much slower in the hanging-wall northwest of the fault than in the footwall southeast of the fault. (c) Deformation was fast for the initial 50 days of the quake, and tapered down slowly afterwards. In the meeting we will give an updated report about the spatio-temporal postseismic deformation pattern, and our progress on inverting for the deformation sources and inferring the deformation mechanisms.
G33C-0711
ALOS Emergency Observations of 2008 Sichuan Earthquake in China
The Advanced Land Observing Satellite (ALOS) carries two high-resolution optical sensors, AVNIR-2 and PRISM, and an L-band synthetic aperture radar, PALSAR. They have several operational modes and are well-suited for disaster monitoring like the 2008 Sichuan earthquake occurred in remote part of China where it is difficult to perform field investigations and there have been few geophysical observations. JAXA had performed emergency observations since right after the main shock, and it was possible to quickly detect surface changes using a large number of data in the areas accumulated since the launch of ALOS on January 26, 2006. The differential interferometric SAR (DInSAR) technique of PALSAR data is a powerful tool for detecting geodetic information associated with seismogenic faulting and the precise geodetic data over a wide area is helpful for understanding earthquake mechanisms. In the case of the Sichuan Earthquake, entire length of the fault is estimated about 300km and areas afflicted by crustal deformation are estimated about 400~400km, so 7 or 8 tracks are needed to cover the study area. All PALSAR data before and after the earthquake were acquired from 34.3E#8249; off-nadir angle in ascending orbit, and the perpendicular baseline distances of image pairs were enough short for DInSAR processing. Despite the fact that the study areas include precipitous mountains area, the interferometric coherence of image pairs is well due to a long wave length of PALSAR (23.6cm). Resultant interferograms reveal significant surface deformations caused by the faulting, although the coherence is significantly reduced by the large displacement and surface collapses in the area of just above the estimated fault. The surface collapses caused by many landslides or debris avalanches in the area were confirmed by the data from optical sensors. Preliminary result from a simple fault modeling with several fault segments can roughly explain the regional deformation patterns, however it can not explain local deformation patterns close to the low-coherence area. To improve the fault modeling, we need more geodetic information in the low-coherence area using not DInSAR but pixel offset method or Multiple Aperture InSAR method, and we should try to estimate a slip distribution of the fault.
G33C-0712
Analysis of ALOS/PALSAR images for the detection of crustal deformations associated with the Wenchuan earthquake
On 12 May 2008, an Mw.7.9 Wenchuan Earthquake occurred in the Longmen Shan fault zone. This earthquake occured by NW-SE trending compressive stress that originates from the collision of the Indian continent to Tibet region. We computed coseismic displacements using SAR-interferometry and offset-tracking. We could not obtain the whole image with only one path of SAR observation, because the earthquake fault is approximately 300km long (swath is 70km for PALSAR). Therefore, we analyzed the SAR-images observed from neighboring paths. We found a few fringes in the far field in the interferograms; therefore we collected images acquired before and after the earthquake as many as possible and evaluated the measurement uncertainties. In the interferograms obtained from preseismic pairs, we found fringes similar to those found in the far field in the coseismic interferograms. This fact implies that coseismic far-field fringes may not be coseismic deformations but errors due to other sources. The dominant factor of the noise is unknown at the moment. In spite of these disturbances, we found coherent fringes in coseismic interferograms. These fringes indicate that the fault slips do not only have dip-slip components revealed by CMT solutions but also noticeable strike-slip components. A set of interferograms shows at least 6 fringes on the northern side of the Longmen Shan fault zone, and 5 fringes on its southern side. We also find some localized concentric fringe patterns around the faults, suggesting a complicated rupture. We can not estimate the displacements using SAR interferometry in the area closest to the faults because of the low coherence, but the offset-tracking results, which calculated the offsets between two SAR images, revealed about 3m of range increase and about 1m of range decrease on the northern side and southern side of the faults, respectively.
G33C-0713
Constraining the 2008 Wenchuan, China Earthquake Rupture Model Using SAR Interferometry Data and an Angular Dislocation Model
The May 12, 2008 Mw7.9 Wenchuan earthquake struck Sichuan, China and claimed ~70000 human lives. Seismological data show that the great earthquake has both thrust and dextral motions on the NE-trending Longmen Shan fault zone (USGS, 2008). Seismological data offered the quickest solution of the fault rupture within a few days of the earthquake (Ji, 2008; Parsons et al., 2008; Yao and Wang, 2008; Chen et al., 2008), but it is difficult to constrain the fault geometry using teleseismic data only, because of the low spatial resolution of the data and the complexity of the ~340 km long seismogenic fault. ESA and JAXA acquired high quality SAR data with earthquake included deformation recorded. For the hanging-wall side of the fault, however, the C-band radar data are much nosier than the L-band PALSAR data because of the rugged topography and strong vegetation. Because the ALOS satellite only has ascending pass acquisitions and the earthquake had both thrust and dextral components, it is still not easy to constrain a fault rupture model using ALOS data only. In order to amend the disadvantage of the PALSAR data, we also use the available ASAR data both in IM mode and WS mode covering the footwall of the fault to constrain the fault model. GPS derived coseismic displacements at sparsely-distributed locations are also used to constrain the frames of InSAR observations. The amplitude offset data in range direction are used to detect the surface location of fault breaks. The field data are used to supplement radar data for identifying the fault location at the southern end. The overall length of surface breaks of the main fault is ~280 km. A second ruptured fault of ~60-km-long, located in parallel with and about 13 km southeast of the main fault, is also identified. We simplify the faults using a model composed of 7 uniform-slip patches and invert for fault dips, widths, and slip components using the Okada dislocation model and a nonlinear global searching method. The simplified model shows consistency with the InSAR observation, and most of the near-field features are recovered. However, the displacement magnitudes and the spatial distributions are not very consistent with the radar observation, which may be related with the simplified fault geometry and the uniform-slip assumption. In order to obtain a distributed-slip solution, we use the angular dislocation model (Thomas, 1993) in the inversion, but the fault geometry is still controlled by the Okada uniform-slip model. The angular dislocation model is superior to the rectangle dislocation model in modeling complex fault geometry. A constrained least-squares solution is obtained for the linear inversion with smoothing operation applied to suppress slip oscillations. Data fitting of our preliminary distributed-slip model is improved dramatically, especially for the near-field measurements. Model update will be reported at the meeting.
G33C-0714 INVITED
Fault Trace and Slip in the 2008 Mw 7.9 Sichuan, China Earthquake From InSAR Observations
The 2008 Mw 7.9 Sichuan, China earthquake was one of the most devastating events in recent times, and claimed more than 80000 lives. Six adjacent pairs of ascending ALOS PALSAR images and three independent pairs of descending ENVISAT ASAR images were processed for both interferograms and SAR pixel offsets. Three additional pairs of ENVISAT ASAR wide-swath (ScanSAR) interferograms were also processed. In the rugged terrain and heavy vegetation of the Longmen Shan, coseismic deformation signals cannot be resolved by the C-band interferograms, although there is some coherence in the flat plains and less steep areas further into the plateau; In contrast, coherence at L-band is generally good and deformation is well resolved. This highlights the main advantage of L-band over C-band, i.e. less temporal decorrelation due to its capability to penetrate more deeply in vegetation. The shallower 38° incidence of PALSAR also helps in the high-relief areas. No useful signals can be resolved in either C-band or L-band interferograms within c.5-30 km along the faults because the large deformations of more than one pixel cause localized misregistration. SAR pixel offsets provide an unambiguous measurement of the surface displacement, and the presence of macroscopic surface features such as ridges, lake shorelines or roads often make the technique successful even in incoherent regions. Both our PALSAR and ASAR offset maps clearly show the fault trace of this large event, and their patterns are nicely consistent with each other, however the PALSAR along-track offsets are heavily contaminated with large wavelike features that are likely due to transient ionospheric disturbances. Based on InSAR observations, we model the large event with four sub-faults: (1) the strikes and lengths of the sub-faults are determined using offset maps; (2) the fault geometries are estimated using interferometric phases and range offsets with the strikes and lengths fixed; (3) fault slip is re- determined with fixed fault geometries. It appears that our preliminary models can explain more than 90% of the observed deformation signals, suggesting the average slips in the northern sub-faults were smaller than those in the southern one. The latter had a significant thrust component and the other segments had predominantly right-lateral slip. Note: (1) COMET: Centre for the Observation and Modelling of Earthquakes and Tectonics; (2) The first three co-authors are in alphabetical order.
G33C-0715
Coseismic Displacement of M8.0 Sichuan Earthquake Derived by ALOS Radar Interferometry
On May 12th 2008, a major (M8.0) earthquake struck Wenchuan County, Sichuan Province in China. The causalities include approximately 70,000 dead and 374,000 injured. The rupture accompanying the events extends over 270 km toward the northeast, and is a result of the convergent tectonic movement associated with Longmen Shan fault. We assemble and process ALOS PALSAR data including 6 ascending tracks (> 72 scenes) to map the coseismic displacement with a full coverage of 400 km by 400 km of the fault zone. Mosaicking the 6 tracks requires careful treatment of the precise orbits as well as a new correction for the ionospheric delay, which is typically 10 cm ramp in the range direction. Severe ionospheric disturbances cause azimuth pixel shifts, which mask the shifts due to the coseismic displacement. We are developing methods to perform ScanSAR interferometry along the descending track in order to obtain two independent components of deformation for analysis. Field observations show the surface rupture occurred on two parallel faults (Beichuan and Pengguan Faults) having significant thrust and right lateral surface displacement. Interferograms show a large area of decorrelation that coincides with the region of landslides and intense ground shaking, mostly on the hanging wall of the thrust faults. We present preliminary results of inversions of the InSAR data for the coseismic slip distribution. The model allows for spatially variable rake angles and illustrates relative contributions of the thrust and strike-slip motion in the deeper part of the fault, as well as the along-strike variability. The inferred slip distribution will be used to drive models of postseismic relaxation, which will be compared to InSAR data collected after the earthquake to constrain the relaxation mechanisms, effective rheology, and provide insights into time-dependent stress transfer and future seismic hazard.
G33C-0716
GPS-derived Coseismic Deformation Associated with the 12 May 2008 Wenchuan Earthquake
The 12 May 2008 Wenchuan earthquake (Mw7.9) that struck the Longmen Shan, eastern Tibet, was the largest seismic event in continental Asia since the advent of modern space and ground based instruments. Mapping the spatial distribution of coseismic deformation of this rare event offers a unprecedented opportunity to understand the rupture mechanism of a great intraplate earthquake and the mechanical behavior of the intensive deformed lithosphere. Here we present a surface displacement field around the rupture zone using near and far-field GPS data recorded at ~400 sites. It shows that the great event caused the Sichuan basin to move coherently westward with the documented maximum displacement of 2.4 m near the rupture. We observe 1-2 cm of westward motion even at GPS sites located 300-400 km east of the epicenter. Sites in Eastern Tibet moved easterly with a maximum magnitude of 2.0 m. The coseismic deformation shows a pattern of anticlockwise rotation in the eastern Tibet, characterized by southeastward movement along the south and middle segments of Longmen Shan thrust, and northeastward displacement along the north part of rupture. We invert the coseismic deformation based on an elastic dislocation model. The best-fit model suggests a length of rupture up to 330 km. The inferred slip on a single fault plane are restricted a upper 15 km of crust with a maximum slip of 8-10 m between Huanwang and Beichuan where the deadly damages were reported. Our preliminary analysis indicates that dip slip may have propagated to >30 km depth on the south part of rupture where the Wenchuan earthquake nucleated, although this result may depend on the assumed fault geometry. The geodetic moment is estimated to ~9.9 ×1020 Nm ( assuming a rigidity of 33 GPa) equivalent to an event of Mw 7.94.