S31B-1902
Comparing surface rupture with fault mapping in the Longmen Shan region, May 12, 2008 Wenchuan earthquake from field observations, SRTM elevation model and remote sensing data
In many regions, especially mountainous areas with heavy rainfall and long interseismic periods, the earthquake hazard may be unknown or underestimated on active structures that do not traverse significant expanses of Quaternary deposits. The region surrounding the Beichuan fault zone, which ruptured as a result of the May 12th, 2008, M 7.9 Wenchuan earthquake, is characterized by high erosion rates, high relief, and deep stream incision with constant and widespread landsliding. We mapped the potential active fault during the disaster response phase of the event between May 14th and May 23rd, using public domain SRTM elevation data and remote sensing imagery. The May 12th event provides an opportunity to compare surface rupture from a large seismic event with Quaternary and(or)active faults that can be mapped from their geomorphic expression in digital data sets. There is a spatial correlation between aftershocks, focal mechanisms and geomorphically distinct sections of the Beichuan fault zone. It appears from both aftershocks and surface ruptures that the seismicity cascaded from the main trace onto adjacent structures or splays from the dominant fault strand within moments of the main event. Although many sections of the Beichuan fault that ruptured have strong expression in the landscape, important deviations also occurred. We use the geomorphic expression of the fault and hanging-wall to consider the long-term effects on landscape evolution from repeated motion on this fault. This study provides an independent method to evaluate active fault mapping in mountainous terrain which can be applied to areas where the hazard has not been assessed or recognized.
S31B-1903
Structural heterogeneities in the source area of the Mw 7.9 2008 Wenchuan Earthquake, China
The Mw 7.9 Wenchuan Earthquake occurred on May 12, 2008 (06:28:01 UTC) in the Longmen-Shan fault zone at the eastern margin of Tibet and adjacent to the Sichuan foreland basin, where as much as 9 m of coseismic slip was observed. This is the most significant earthquake to have struck China since the 1976 Tangshan Earthquake (Mw 7.6). Chinese authorities estimated more than 69,000 people were killed and 374,176 injured. About 1,485,000 people were forced into temporary shelters (http://www.gov.cn). The coseismic faulting zone coincides roughly with this aftershock distribution. The extent of the great damage may have resulted from the variation of crustal structures along the tectonic thrust faulting and strike slipping zone, causing significant coseismic displacement and acceleration on a regional scale. For a better understanding of what may have triggered this earthquake and how the rupture proceeded after the initiation, we conducted an investigation of the seismic structure in the Wenchuan earthquake source area. A large number of arrival time data of P and Pn, S and Sn phases from local earthquakes were collected and inverted jointly for the three-dimensional P- and S-wave velocity (Vp and Vs) models. The Poisson's ratio model was then calculated from these velocity models. The overall patterns of Vp and Vs anomalies are similar to each other at any of these three depths. The Sichuan Basin is a distinct entity in tomographic images, anomalously slow at 13 km depth and anomalously fast at 20 km depth, in sharp contrast to the anomalies on the mountain side across the Longmen-Shan fault zone. The area of the fault zone is characterized in general by high Vp and Vs anomalies at a depth of 13 km, although a low Vp and Vs anomaly patch is present in the middle of a total fault length of 300 km. At depths around 20 km, low Vp and Vs anomalies dominate on the mountain side including its eastern margin, where the shallower (¡13 km) and deeper (¡30 km) depths are overwhelmed by high Vp and Vs anomalies. The low Vp and Vs zone with high Poisson's ratio immediately below the hypocenter may be associated with fluid-filled, fractured rock because widespread existence of fluids, such as partial melt or aqueous fluids, has been imaged as a zone of high electrical conductance in the middle-lower crust of the eastern margin of Tibet by the magnetotelluric studies. Continuation of high Poisson's ratio anomaly from the lower crust to the upper crust in the vicinity of the hypocenter suggests fluid ascent from the lower crust into the relatively permeable fault belt to be maintained there at high pressures. The high Poisson's ratio anomaly in the upper crust near the hypocenter is not associated with low Vp and Vs anomalies but with weakly positive Vp and Vs anomalies¡¦suggesting the presence of water rather than melt. Intrusion of water into a shallow brittle zone through thin pores (aspect ratio ¡0.01) would reduce positive Vp and Vs anomalies characteristic to this brittle zone but would increase its Poisson's ratio. We suggest that the vicinity of the hypocenter near the SW end of the fault belt was intruded from below by aqueous fluids originally in the middle to lower crust, reducing the strength of source rocks and thereby bringing the system into brittle failure.
S31B-1904
Source mechanism and rupture process of Wenchuan great earthquake
On May 12, 2008, occurred a disastrous earthquake, whose magnitude was later modified to be Ms8.0. Focal mechanism and rupture process of the event were determined by inverting long period waveforms from the world-wide stations. The focal mechanism was obtained by a technique of moment tensor inversion considering source time complexity, indicating that the earthquake was a thrust-dominant event with quite right-lateral component. Based on one of the inverted fault planes, which was of strike 230¡ã/dip35¡ã/rake123¡ã, we selected an area with a length of 510 km in strike direction and a width of 50 km in dip direction and divided it into 255 subfaults of 10 km¡Á10km for inversion of rupture process. The inverted results showed that the source time function was consisted of at least 3 bigger sub-events with a total duration time of 90s, and the maximum slip value is about 7.3 m and the average slip value is 2.4 m, corresponding the maximum stress drop of 53 MPa and the average stress drop of 18 MPa, respectively. From the static slip dislocation distribution, there were mainly 2 slip-concentrated areas located in Wenchuan and Beichuan, respectively. The rupture broke the ground surface and resulted in the maximum slip of 6.1 m in Wenchuan and 5.6 m in Beichuan, respectively. From the slip-rate snapshots, there existed an evident phenomenon that the rupture strongly extended to SW in 15 s-30 s after the earthquake started, which implied that the rupture have ever come back for some time in the whole rupture process of propagating to NE direction. Also, we calculated the near-fault co-seismic displacement field based on the inverted static slip distribution, and the calculated displacement field well featured the observed intensity contours.
S31B-1905
Rupture Imaging of the 12 May 2008 Wenchuan Earthquake Using Teleseismic Arrays
In this work we constrain the source properties of the 7.9 Mw Wenchuan earthquake by back-projecting teleseismic P energy from virtual seismic arrays. Using the IRIS DMC, we collected 255 broadband P waves at 30-90 degrees from the epicenter. We applied a multi-channel cross-correlation algorithm and found that at periods > 5 s nearly all of the station averaged correlation coefficients were above 0.9. We used a genetic algorithm to search for sets of station weights that yielded array response functions (ARFs) with symmetric, narrow mainlobes and small sidelobes at a period of 5 s; however, none of them were significantly better than those derived from a simple downsampling heuristic in which we randomly selected a single station from each 1000 km by 1000 km patch on the Earth's surface. We also considered three regional subarrays in Alaska, Australia, and Europe with apertures less than 30 degrees. Coherence observations implied that these arrays would be effective to periods of 1-2 s. Individual ARFs for the regional subarrays were skewed towards the seismometers, however the linear sum of the regional subarray beams at 1 s produced a symmetric ARF similar to that of the groomed global subarray at 5 s. We considered three measures of beam power in constructing the rupture images: linear, Nth root, and F stacking. We found that F stacking was best for detecting the beginning and duration of the rupture, but that it produced the most biased relative amplitudes among patches of energy release. Linear stacking showed relative amplitudes that agreed with the finite fault results, but it gave the poorest spatial resolution. The Nth root stacks (with N=4) worked the best overall, showing consistent relative amplitudes with high spatial resolution. The spatial distribution of energy release imaged by the global subarray at 5 s is slightly different than that imaged by the combination of three regional subarrays at 1 s, which may indicate some frequency dependence to the energy release. Nevertheless, for both configurations we find the same rupture direction, rupture time, and rupture velocity. We find that the Wenchuan earthquake had three main energy peaks at 0, 23, and 57 s after the origin time, with the 23-s peak being largest, and that it ruptured unilaterally to the northeast for about 280 km and 100 s, with an average speed of 2.9 km/s.
S31B-1906
Rupture Characters of the 2008 Mw=7.9 Wenchuan Earthquake Revealed From Empirical Green's Function Deconvolution
On May 12, 2008, the Mw=7.9 Wenchuan earthquake struck the eastern margin of the Tibetan Plateau in the vicinity of the Sichuan Basin and caused a surface rupture extended laterally about 300 km. To retrieve the rupture characters of the 2008 Wenchuan earthquake, we used the empirical Green's function deconvolution analysis of teleseismic waveforms to avoid the effect from complicated 3D velocity structure. The Longmen Shan region is composed of not only the typical thrust faults but also active dextral-slip structures. To validate the rupture characters, we selected three aftershocks within the mainshock rupture region as the empirical Green's functions, with magnitudes of 5.7, 5.7, and 6.0, respectively. Two selected aftershocks were thrust events, one is located in the southern and the other located in the northern portion of the fault, similar to that of the mainshock from Harvard CMT solution. One event in the northern portion of the fault is strike-slip event. Nine IRIS stations with well recorded seismograms and good azimuthal coverage were investigated for the empirical Green's function deconvolution. The deconvolved source characters (relative source time function, RSTF) from the aftershocks with thrust events show consistent features with two distinct asperities at the first 40 sec. The 2nd asperity in the time window of 15-40sec has the largest amplitude and longer duration. The rupture directivities were observed clearly for the events toward rupture front with smaller pulse width. Careful examination also found the character involved with dipping directivity. The RSTFs after the 2nd asperity have smaller amplitude and are less consistent from the thrust aftershocks, suggesting slip character rather than thrust after the time around 40-60 sec. The RSTF obtained from the strike-slip aftershock show larger amplitudes after 40 sec compared to the RSTF from thrust aftershocks. These results confirmed the dominant rupture characters of thrust slip in the first 40-60 sec, then, became strike-slip in the rest of fault rupture. Using some simple assumptions, the RSTFs suggest that the thrust slip with maximum of 8.4m near Wenchuan, and the strike-slip motion with maximum of 3.6m located between Beichuan to Qingchuan.
S31B-1907
Extent, Duration and Speed of the 2008 Wenchuan Ms 8.0 Earthquake Directly Imaged by the Alaska Regional Network
On 12th of May, 2008, an Ms 8.0 earthquake occurred in Wenchuan county, Sichuan province of China. We took advantage of the broadband waveform data from a regional seismic network in Alaska, USA, and imaged the temporal and spatial characteristic of the high-frequency energy sources by means of the generalized seismic array technique. The imaged results indicated that the earthquake fault was extending about 300 km to northeast from the epicenter; the most energy released within the first 80 seconds of the whole process. The average rupture speed was about 3.1km/s. Keywords: Wenchuan earthquake, generalized seismic array technique, rupture extent, rupture spread speed
S31B-1908
Rupture Propagation of the 2008/05/12 Ms8.0 Wenchuan Earthquake Using Multiple Teleseismic Arrays
We study the rupture propagation of the 2008/05/12 Ms8.0 Wenchuan Earthquake. We apply array
techniques such as semblance vespagram analysis to P and S body waves recorded at seismic broadband
station within 30-100° epicentral distance. By combination of multiple large aperture station groups
spatial and temporal resolution is enhanced and problems due source directivity and source mechanism are
avoided.
We find that seismic energy was released for at least 125 s. Propagating unilaterally at sub-shear rupture
velocity of about 2.5 km/s in NE direction, the earthquake reaches a lateral extent of more than 300 km.
Whereas high semblance during within 70 s from rupture start indicates simple propagation more complex
source processes are indicated thereafter by decreases coherency in seismograms.
We emphasize that first result of our computations where obtain within 30 minutes after source time by using
an atomized algorithm. This procedure has been routinely and globally applied to major earthquakes. Results
are made public through internet.
http://www.geo.uni-
potsdam.de/arbeitsgruppen/Geophysik_Seismologie/forschung/ruptrack/index.php
S31B-1909
Studies on aftershocks of the 12 May 2008 Wenchuan Earthquake
The 12 May, 2008 Wenchuan earthquake between the eastern Tibetan Plateau and the western Sichuan basin in central China strongly activated the ~300 km-long LongMenShan Fault. The region around the mainshock is highly seismic active and has occurred thousands of aftershocks since then. In order to understand the initiation and rupture dynamics of the Wenchuan earthquake by monitoring its aftershocks, we deployed 16 sets of short-band (SENSL-22E) seismometers around the mainshock epicenter after July of 2008. We processed one month of the recorded data and retrieved more than two thousands of aftershocks among which ~600 micro-earthquakes have magnitude (ML) greater than 2.0. The aftershocks clearly delineate the fault plane of the mainshock. The main fault is characterized by a dip of ~45 degree. Besides, another vertical fault is well determined by the aftershocks and converges with the main fault at about 15 km depth. This interesting feature indicates that the strong Wenchuan earthquake occurred in the crossing region of two faults. Using the highly dense aftershocks, we carried out local body-wave travel-time tomography. The aftershocks are consistently observed in low velocity weak zone.
S31B-1910
The slip history of the 2008 Wenchuan Sichuan earthquake inverted by empirical Green's function
The Ms 8.0 Wenchuan earthquake of 12 May 2008 was an extraordinary event that produced a 300-km-long surface rupture and was followed by moderate-size many aftershocks in the following days. Due to complex geological conditions around Longmen Shan fault, we use several aftershocks as empirical Green's functions (EGFs) to retrieve the source time functions (STFs) and further image the rupture velocity and slip on the mainshock. First, we try to just use one EGF to inverse, we get some information of slip history of Wenchuan earthquake. However, for such a complicate earthquake which has long surface rupture, we find that just using one EGF is not enough. Finally, according to focal mechanisms of aftershocks occurred on the same fault, we choose one for northeast part and another for southwest part as EGFs and then use the simulated annealing method to retrieve the slip history of Wenchuan earthquake.
S31B-1911
Focal Depth of the WenChuan Earthquake Aftershocks from modeling of Seismic Depth Phases
After the 05/12/2008 great WenChuan earthquake in Sichuan Province of China, tens of thousands earthquakes occurred with hundreds of them stronger than M4. Those aftershocks provide valuable information about seismotectonics and rupture processes for the mainshock, particularly accurate spatial distribution of aftershocks is very informational for determining rupture fault planes. However focal depth can not be well resolved just with first arrivals recorded by relatively sparse network in Sichuan Province, therefore 3D seismicity distribution is difficult to obtain though horizontal location can be located with accuracy of 5km. Instead local/regional depth phases such as sPmP, sPn, sPL and teleseismic pP,sP are very sensitive to depth, and be readily modeled to determine depth with accuracy of 2km. With reference 1D velocity structure resolved from receiver functions and seismic refraction studies, local/regional depth phases such as sPmP, sPn and sPL are identified by comparing observed waveform with synthetic seismograms by generalized ray theory and reflectivity methods. For teleseismic depth phases well observed for M5.5 and stronger events, we developed an algorithm in inverting both depth and focal mechanism from P and SH waveforms. Also we employed the Cut and Paste (CAP) method developed by Zhao and Helmberger in modeling mechanism and depth with local waveforms, which constrains depth by fitting Pnl waveforms and the relative weight between surface wave and Pnl. After modeling all the depth phases for hundreds of events , we find that most of the M4 earthquakes occur between 2-18km depth, with aftershocks depth ranging 4-12km in the southern half of Longmenshan fault while aftershocks in the northern half featuring large depth range up to 18km. Therefore seismogenic zone in the northern segment is deeper as compared to the southern segment. All the aftershocks occur in upper crust, given that the Moho is deeper than 40km, or even 60km west of the Longmenshan fault. Absence of mid-lower crustal shocks supports the model of lower crustal flow beneath eastern Tibetan plateau, which is probably responsible for Longmenshan uplifting and hence the Wenchuan earthquake.
S31B-1912
12 May 2008 M=7.9 Wenchuan, China, Earthquake Calculated to Increase Failure Stress and Seismicity Rate on Three Major Fault Systems
The Wenchuan earthquake on the Longmen Shan fault zone devastated cities of Sichuan, claiming at least 69,000 lives. We calculate that the earthquake also brought the Xianshuihe, Kunlun and Min Jiang faults 150- 400 km from the mainshock rupture in the eastern Tibetan Plateau 0.2-0.5 bars closer to Coulomb failure. We use the calculated stress changes and the observed M≥3.2 background seismicity during the preceding decade to forecast the rate and distribution of M≥3.2 and damaging shocks during the next decade. The forecast rate of M≥3.2 shocks provides a testable hypothesis for the future distribution and rate of earthquakes. An increased rate of such small shocks on the stressed parts of the Xianshuihe, Kunlun, and Min Jiang faults, and a decreased rate in the calculated stress shadows, would provide support for the hypothesis that large shocks are now also more likely where the static stress imparted by the mainshock has risen. Because some portions of these stressed faults have not ruptured in more than a century, the Wenchuan earthquake could promote additional M>7 earthquakes, potentially subjecting regions from Kangding to Daofu and Maqin to Rangtag to strong shaking. We calculate the earthquake probability in the region to be 57-71% for M≥6 shocks, and 8-12% for M≥7 shocks, during the next decade. These are up to twice the probabilities for the decade before the Wenchuan earthquake struck, on the basis of the observed M≥3.2 seismicity rate and an assumed b-value of 1.0. The Wenchuan earthquake thus poses a continuing seismic hazard, and also provides a test of earthquake promotion.
S31B-1913
Coulomb stress Changes on faults Caused by the 2008 Mw7.9 Wenchuan earthquake, China
The 12 May 2008 Mw 7.9 Wenchuan earthquake has changed the regional tectonic stress field significantly. It is important to know how such a change affects the tectonic loading process of the faults and therefore the seismic potential of the region. Utilizing slip distribution of the Wenchuan earthquake as the driving source we compute the Coulomb stress changes on the neighborhood faults. Our result shows that the Coulomb stresses have increased on the southernmost Xianshuihe, northern and southern ends of the Longmen Shan, East Kunlun, Longxian-Baoji, Ela Shan, Baiyu, southernmost Riyuan Shan, southern segment of the Mabian-Yanjin, western segment of the Bangong-Jiali, and Zemuhe faults. The faults with most significant increases of the Coulomb stresses are: 0.2 and 0.4 bars at the southern and northern ends of the Longmen Shan, 0.09 bars at the southeast end of the East Kunlun, 0.07 and 0.04 bars at the southernmost and central Xianshuihe, 0.04 bars at the Longxian-Baoji, and 0.03 bars at the western segment of the Qinling Southern Frontal faults. The quake also relaxed the Coulomb stresses on the Longriba, Nujiang, western segment of the Qinling Northern Frontal, Zhuanglanghe, Haiyuan, and Minjiang faults. Large relaxations occurred on the ruptured part of the Longmen Shan, Minjiang, and Luhuo segment of the Xianshuihe fault, with the Coulomb stress reductions ranging 0.4~7, 0.02~1.9, and 0.11 bars respectively.
S31B-1914
Source Process and Ground Motions of the 2008 Wenchuan, China, Earthquake
The 2008 Wenchuan, China, earthquake occurred on May 12 (local time), and ground motions from this earthquake propagated around the world. We obtained the ground motion records observed at stations of FDSN from IRIS DMC. According to surface fault investigations (this study; Hao and Si, 2008) and the aftershock distribution by USGS, we defined a two-segment fault plane. The strike and dip angles for the southern and northern segments are respectively given to be (228, 35) and (232, 65) based on the results of point source analyses. We then carried out a finite source inversion of the first 140 s of 43 P and 8 SH teleseismic waves using the Green's functions of Kikuchi and Kanamori (1991) and the inverse algorithm of Yoshida et al. (1996) with modifications. We also used the location of the hypocenter determined by USGS (103.33E, 30.99N, depth 12 km) as a rupture initiation point. The resultant slip distribution indicates the first asperity with the largest reverse-faulting slip of about 9 m and the second asperity with a strike slip of about 4 m to be located 45 and 170 km northeast of the hypocenter, respectively. Major aftershocks are located surrounding these two asperities. The total seismic moment is 1.0 x 10**21 Nm, which corresponds to a moment magnitude (Mw) of 7.9, and the rupture duration is estimated to be 100 to 120 s. Significant slips appear in a 250 km long region (10,000 km**2) of the source fault, and these length and area are close to averages for an Mw 7.9 low-angle reverse-faulting earthquake. The area is ten times larger than that of the 1995 Kobe, Japan, earthquake. This fault area and the maximum slip three times larger result in energy release about thirty times larger than that of the 1995 Kobe earthquake (Mw 6.9). This difference almost coincides with the difference in the number of fatalities between the 2008 Sichuan and 1995 Kobe earthquakes. Heavily damaged towns such as Wenchuan and Beichuan are located within the surface projection of the source fault, so that the damage should have been caused by strong ground motions in the source region. In particular, the zones of seismic intensity XI are assumed just above the two asperities by the China Earthquake Administration. Strong motion records observed in and around the source region will be provided by the Institute of Engineering Mechanics, the China Earthquake Administration. They will then be inverted for a detailed image of the source process. We also note that long-period ground motions were felt in Beijing, Shanghai, and Taipei, which are 1,500 km or more away from the hypocenter.
S31B-1915
Numerical Simulation of Strong Ground Motion for the Ms 8.0 Wenchuan Earthquake of 12 May 2008 With Realistic Surface Topography
The Wenchuan earthquake of 12 May 2008 is the most destructive earthquake in China in the past 30 years in terms of property damage and human losses. In order to understand the earthquake process and the geo- morphological factors affecting the seismic hazard, we simulated the strong ground motion caused by the earthquake, incorporating three-dimensional earth structure, finite-fault rupture, and realistic surface topography. The simulated ground motions reveal that the fault rupture and basin structure control the overall pattern of the peak ground shaking. Large peak ground velocity (PGV) is distributed in two narrow areas: The one with the largest PGV values is above the hanging wall of the fault and attributed to the locations of fault asperities and rupture directivity, the other is along the northwestern margin of the Sichuan Basin and caused by both the directivity of fault rupture and the amplification in the thick sediment basin. Rough topography above the rupture fault causes wave scattering, resulting in significantly larger peak ground motion on the apex of topographic relief than in the valley. Topography and scattering also reduce the wave energy in the forward direction of fault rupture but increase the PGV in other parts of the basin. These results suggest the need for a localized hazard assessment in places of rough topography that takes into account of the topographic effects. Finally, had the earthquake started at the northeast end of the fault zone and ruptured to the southwest, Chengdu would have suffered a much stronger shaking than it experienced on 12 May, 2008.
S31B-1916
Fault Frictional Motions of the 2008 Mw7.9 Wenchuan Earthquake and the 2001 Mw7.8 Kunlun Earthquake: An Implication to Their Aftershock Triggering Processes
We studied the source frictional processes of two recent large continental earthquakes, the 2008 Wenchuan event, the 2001 Kunlun event, associated with strike-slip and thrust faulting, from a macroscopic point of view. We also compared the inferred frictional mechanisms and observed aftershock distributions following Wenchuan and Kunlun main shocks. The 2008 Sichuan Wenchuan earthquake (Mw7.9) ruptured about 300 km along a northeast-striking, west- dipping Longmen Shan thrust fault belt located on the margin of the eastern Tibetan Plateau. The 2001 Kunlun earthquake (Mw7.8) ruptured about 500 km of the western Kunlun fault in high northern Tibet. The Kunlun earthquake appears to have ruptured the left-lateral Kunlun fault which extends about east-west for about 1600 km in north Tibet. Geologically, both earthquakes occurred in the same tectonic regime in which the Tibetan Plateau is deformed as a result of the collision between two tectonic plates, the India plate and he Eurasia plate. Using current available earthquake data set from the GCMT, USGS/NEIC, and Chinese Earthquake Network Center (CENC), first, we calculated the static stress drops, apparent stresses of these two main events. According to the Savage and Wood inequality, our results showed that, for the 2008 Wenchuan earthquake, the dynamic frictional motion on the fault underwent an overshoot process, in contrast, for the 2001 Kunlun earthquake, the dynamic frictional motion on the fault underwent an undershoot process or abrupt locking model. Related to the frictional overshoot/undershoot, a uniform representation used in the calculation of radiation efficiency is derived based on a simple slip-weakening friction model. The primary results showed that, for Wenchuan and Kunlun events, the corresponding radiation efficiencies are about 0.2 and 0.9 respectively. We discussed physical based frictional overshoot/undershoot mechanisms in which the energies dissipated by frictional undershoot and overshoot correspond to the faulting re-strength work (radiated friction) and visco-elastic relaxation work spent at rupture arrest, respectively. In addition, we also showed that the triggered aftershocks following these two main events exhibited obviously different characteristics both in the magnitude-frequency distribution and in the spatial distribution, which could belong to the different triggering mechanisms. For Wenchuan earthquake, the most of aftershocks are focused on the Longmen Shan main fault belt with a linear distribution pattern along the fault, implying the triggered events are related to the stress-relaxation process caused by frictional overshoot, in contrast, for Kunlun earthquake, the triggered aftershocks usually are scattered around main fault, implying the triggered events could be caused by the passage of high frequency radiation due to frictional undershoot and produces high stress over off-fault activation We conclude that, from currently available data, it appears that no single triggering model appears capable of accounting for the wide variation observed in the nature of triggered activity.
S31B-1917
Groundwater level changes in the digital monitoring network in Mainland of China induced by the 2008 Wenchuan earthquake
We observed water level changes in groundwater wells following the great 2008 Wenchuan earthquake by using the ground water monitoring network in Mainland of China to aim at earthquake prediction. There are about 104 step-like changes found from 376 wells of the Digital Monitoring Network. The hydrogeological background of the monitoring network shows that the water level changes induced by earthquakes fit the poroelastic theory under undrained condition. Based on the poroelastic theory, we calculated the strain changes from water level steps by using the tidal information recorded in the same well and named it as equivalent strain changes. Then we analyzed the spatial distribution of the equivalent strain changes. The results showed that the amplitude of the equivalent strain changes did not decrease with epicentral distance of wells, except the epicenter, wells with large amplitude distributed along active tectonic belt or fault zones. The results above imply that the coseismic changes of well water level can be used as an approach to active tectonic investigation and earthquake prediction.
S31B-1918
Coseismic velocity variation of Wenchuan aftershocks measured from active source monitoring
M8.0 Wenchuan earthquake ruptured about 300km northeast ward with an aftershock series migrating in the same direction. To monitor the stress changes in the north edge of Wenchuan earthquake fault zone, we conducted a field experiment to measure the subsurface velocity variation. In the experiment, an electric hammer was used as a repeatable seismic source and the seismic signals were recorded by eight short period seismometers composed of Guralp 40T sensor and Reftek 130B digitizer. Seismometers were deployed with epicenter distances ranging from several meters to 200 meters, forming a cross-type array. The digitizer recorded the data with sampling rate of 500 samples per second. The experiment was repeated once an hour. In each round, electric hammer generated several (usually 12) seismic pulses. The experiment began from 12:00, 7, June 2008, and lasted for three weeks. Two moderate aftershocks with magnitude 4.5 and 4.4 occurred during the period of active source experiment. To estimate the velocity change of subsurface media in the experiment site, we estimated the absolute arrival time of the direct arrival and surface waves. The arrival time of the direct waves, recorded by the station closest to the source, were used as reference time to calculate the travel time of different phases. The absolute arrival time of certain phase at one station, was measured by cross-correlating the waveform with the reference waveform, and then the cross-correlation function was interpolated with a cosine function to obtain sub-sample precision. The digitizers were timing continuously with GPS, and clock errors for most of digitizers were found to be less than 5 microseconds, which can be corrected by using Reftek utilities. Thus our absolute travel time measurements have precision to 10-5 to 10-6. We average the travel time for different shot in each round, and the standard errors are estimated to be ~ 10-4 and 10-5 for direct arrival and surface wave, respectively. The travel time variation of direct arrival and surface waves were measured for different stations, and results from different stations show the similar tendency. Both direct arrival and surface wave velocities show a general dependency on barometric pressure. But surface wave velocity appears to be negatively correlated with barometric pressure, while dependency of direct wave on barometric appears to be positive. We found two prominent drops by ~ 1% in the surface wave velocity measurement. They occurred right after the two moderate aftershocks. Surface wave velocity recovered in several hours. But these coseismic changes were not obviously shown in the direct arrival time. We speculate the coseismic surface velocity drops may be caused by fluid migration in the environment layer in responding to the dynamic stress changes.
S31B-1919
The Great Sichuan Earthquake of May 12, 2008 (Mw7.9) - An Unpredictable Earthquake
The great Sichuan earthquake (Mw7.9, Ms8.0) occurred on the Longmenshan Rupture Zone (LRZ) triplet of parallel faults, the boundary fault rupture zone that separates the eastern Tibetan Plateau from the Sichuan Basin. No really scientific, successful prediction was made for this event in terms of place or size, or time (even on long–time scale) with an acceptable precision. That is, the 2008 event was an unpredicted earthquake, a failure of prediction. The 2008 Sichuan earthquake, with a fault rupture length L about 270 km and displacements up to 9-12 m, produced relief changes and landslides, massive damages, and 69,000+ deaths. On the other hand, the Tibet region is known for large/great earthquake occurrences (e.g. in 1920, 1976 and 2001). In view of these, we analyzed different studies if the 2008 earthquake was actually predictable or unpredictable. We found in these researches that: (1) no great earthquake from the same magnitude order on the same fault is known to us, both in historical and paleoearthquake data to have occurred at the place of the 2008 Sichuan earthquake, until the occurrence of the earthquake, (2) only one earthquake close to Sichuan earthquake is known: the Aug. 25, 1933 Diexi earthquake (M=7.3, NEIC), but this event has ruptured a different fault, related or connected with the Min Jiang fault (Cheng et al. 1994), (3) the great 2001 Kun Lun earthquake Mw8.1, L=450 km, generated no coseismic Coulomb stress transfer as to trigger the nucleation of the 2008 earthquake on the actual fault rupture, (4) a detailed study (Papadimitriou et al., 2004) of Coulomb stress transfer from large earthquakes on the Xianshuihe show clearly no forecast of the possible place of a large or great earthquake along the LRZ or its segments of the Sichuan earthquake, (5) low dip-slip rate of, about 1-2 mm/year for LRZ, and the displacements from the source process (Ji, 2008; Nisimura and Yagi, 2008) suggest a mean repeat time for such an earthquake class longer than at least 1000 years or more, and (6) our long time observations and rationale led us to the fact that it is fundamental for conceiving the predictability of phenomena and patterns in nature when they provide the existence of their repeatability in space and time, more or less regularly, or irregularly. Under this view the 2008 Sichuan earthquake was an unpredictable earthquake, relative to place and size, and time. The predictability of this earthquake, after its occurrence can, on the other hand, be discussed at the philosophical level. Future paleoseismological researches should constrain the LRZ repeat-times for M8 earthquakes.