- Email: [email protected]
Coseismic horizontal shortening associated with the 2008 Wenchuan Earthquake along the Baishahe segment from high resolution satellite images
Journal of Asian Earth Sciences 50 (2012) 164–170
Contents lists available at SciVerse ScienceDirect
Journal of Asian Earth Sciences journal homepage: www.elsevier.com/locate/jseaes
Coseismic horizontal shortening associated with the 2008 Wenchuan Earthquake along the Baishahe segment from high resolution satellite images Feng Shi, Hong-Lin He ⇑, Zhan-Yu Wei Key Laboratory of Active Tectonics and Volcano, Institute of Geology, China Earthquake Administration, Beijing 100029, China
a r t i c l e
i n f o
Article history: Received 12 May 2010 Received in revised form 11 January 2012 Accepted 17 January 2012 Available online 22 February 2012 Keywords: Wenchuan Earthquake Coseismic horizontal shortening Image contrast
a b s t r a c t Many coseismic deformation data from the Wenchuan Earthquake of 12 May, 2008, have been published; however, most of the data record strike-slip and vertical offsets, and there is little information on the horizontal shortening components. To determine the amount of horizontal shortening, we suggest a new method of measuring the differences in the positions of geometrical markers observed in satellite images before and after the earthquake. We found two roads that run nearly parallel to the earthquake rupture but on opposite sides, and we examined the spacing between them before and after the earthquake. We were able to measure horizontal shortenings that average 6.6 ± 1.84 m, with a maximum of 11.0 m and a minimum of 1.1 m. As an alternative method, we used a GPS RTK field survey system to measure the spacing of the two roads after the earthquake, and then measured the difference between that and the spacing determined from satellite images before the earthquake. This method gave horizontal shortenings that average 7.1 ± 1.3 m, with a maximum of 9.85 m and a minimum of 3.8 m. The shortenings measured, using either of these two methods, are larger than those directly measured in the field in the immediate vicinity of the fault, which suggests that the coseismic horizontal shortening is distributed over a wider area that extends some distance away from the brittle fault scarp itself. Finally, the results support a low-angle fault model for the Wenchuan Earthquake, and they provide useful insights into the seismotectonics of eastern Tibet, especially the building of the Longmenshan range. Ó 2012 Elsevier Ltd. All rights reserved.
1. Introduction The Ms 8.0 Wenchuan Earthquake, which struck Sichuan, China, on 12 May 2008, was accompanied by the rupture of two NWdipping imbricate reverse faults along the Longmenshan fault zone at the eastern margin of the Tibetan Plateau (Burchfiel et al., 2008; Liu-Zeng et al., 2009; Xu et al., 2009). Seismological studies indicate that the main shock was initiated approximately 20 km southwest of the town of Yingxiu, with the rupture propagating northeastwards for about 300 km at an average speed of 2.8 km/ s, and for a period of 110–120 s (Nakamura et al., 2010; Wang et al., 2008, 2009). Field investigations show that the earthquake generated a major rupture about 250 km in length along the Beichuan–Yingxiu Fault, as well as a minor rupture about 72 km in length along the parallel Guanxian–Jiangyou Fault (Li et al., 2010; Liu-Zeng et al., 2009; Xu et al., 2009). The observed co-seismic slips show that faulting on the major rupture zone (the Beichuan–Yingxiu Fault) consisted of upwards thrusting of the northwest hanging wall, together with some component of right-lateral strike-slip. On ⇑ Corresponding author. Address: 19 Beituchengxilu Road, Beijing 100029, China. Tel./fax: +86 10 62009215. E-mail addresses: [email protected] (F. Shi), [email protected] (H.-L. He), [email protected] (Z.-Y. Wei). 1367-9120/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. doi:10.1016/j.jseaes.2012.01.001
the minor rupture zone, the Guanxian–Jiangyou Fault, only reverse slip is observed (Liu-Zeng et al., 2009; Xu et al., 2009). Along the Beichuan–Yingxiu Fault, the observed maximum vertical displacements are 9.5 m (Ran et al., 2010) or 6.5 m (Xu et al., 2010), and the strike slip displacements are 4.8 m (He et al., 2008) or 4.9 m (Xu et al., 2009). On the Guanxian–Jiangyou Fault, the observed maximum vertical displacement is 3.5 m (Liu-Zeng et al., 2009; Xu et al., 2009). Most of the coseismic slip data, collected in the field (e.g., He et al., 2008; Liu-Zeng et al., 2009; Xu et al., 2009), record strike-slip and vertical offsets. The amount of horizontal shortening, which is as important as the strike-slip and vertical offsets, has seldom been measured (Chen et al., 2008, 2009; He et al., 2008; Shi et al., 2009; Wang et al., 2010). The coseismic slip vector has strike slip, vertical slip, and horizontal shortening (or extension) as its three components in space, and a reasonable and reliable estimate of all three components is essential if we are to understand the coseismic surface deformation. Although GPS and InSAR data have provided information on the coseismic displacement field of the Wenchuan Earthquake (Hao et al., 2009; Hashimoto et al., 2010; The Project of Crustal Movement Observation Network of China, 2008), the InSAR data often fail near the fault due to excessive temporal decorrelation of the images and signal saturation (Klinger et al., 2006). Furthermore, SAR data only provide the satellite to ground component
F. Shi et al. / Journal of Asian Earth Sciences 50 (2012) 164–170
165
Fig. 1. Surface ruptures in the Baishahe segment and locations of the reference roads. (a) Active tectonics in China (Deng, 2007); the black rectangle shows the location of (b). (b) Surface rupture formed during the Wenchuan Earthquake shown by red lines (Xu et al., 2009); black lines show other active faults; black rectangle shows the location of (c). (c) Surface ruptures in the Baishahe segment (from He et al. (2008)) and selected roads; black rectangle shows the location of (d). (d) Local roads on QuickBird images preearthquake (blue zone) and post-earthquake (red zone), and the eastern roadside measured in the field (black dotted line). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
of the deformation (Michel and Avouac, 2002). Because there are only a few stations immediately adjacent to the fault, the GPS data only show the deformation field at some distance from the fault (The Project of Crustal Movement Observation Network of China, 2008). The recently developed sub-pixel correlation method provides new information on co-seismic displacements, and it has now been applied to several earthquakes (Dominguez et al., 2003; Klinger et al., 2006; Michel and Avouac, 2002; Pumbroeck et al., 2000). The method is limited mainly by the decorrelation of the images, the accuracy of the digital elevation model (DEM), the aliasing of the images, and the uncertainties concerning the measured roll, pitch, and yaw of the satellite (Pumbroeck et al., 2000). To avoid the error due to the contrast of two optical satellite images, we provide here a new and simpler method of measuring the coseismic shortening by subtracting differences in the positions of geometrical markers observed in satellite images before and after the earthquake. In this paper, the rupture along the Baishahe River is used as a case study to calculate the coseismic shortenings. This rupture is located to the southwest of the major rupture, it extends 14 km along the Baishahe River, and is approximately 15 km northeast of the epicenter of the Wenchuan Earthquake. Its geometry is complicated, with many short sections striking 50° on average and dipping
northwest. The maximum vertical and strike-slip coseismic offsets, measured along the Baishahe River rupture (He et al., 2008), are 6.5 m and 4.8 m, respectively. There are two roads that run nearly parallel to most of the rupture (Fig. 1), one to the northwest of the rupture, the other to the southeast. They provide excellent markers for measuring the horizontal shortening. On satellite images we can measure the distance between the two roads normal to the average strike of the rupture, and observe the changes before and after the Wenchuan Earthquake. These observations can then be used to calculate the coseismic horizontal shortening.
2. Methods and data QuickBird satellite images were used for this study because of their high resolution (0.6 m), arguably the best among the commercial satellite images presently available. The two roads run parallel to the Baishahe River rupture, but on opposite sides, and they are used as geometric markers. The procedures were as follows: Firstly, a field survey was made of the two selected roads using a GPS RTK measuring system. Road segments that were apparently repaired or reconstructed after the earthquake were not included in the survey.
166
F. Shi et al. / Journal of Asian Earth Sciences 50 (2012) 164–170
Secondly, we made an orthorectification of optical satellite images. The pre-earthquake satellite image was rectified using the 20 m pixel DEM, made from 1:50,000 topographic maps published by the National Geomatics Center of China, while the postearthquake satellite image was rectified by Ground Control Points (GCPs) and the 20 m pixel DEM. Thirdly, we extracted the roads in the pre- and post-earthquake ortho-corrected satellite images, and measured the spacings between them along the direction perpendicular to the rupture in both the pre- and post-earthquake images. Fourthly, we measured the differences in the spacings between the roads, before and after the earthquake, for various positions along the strike of the rupture. The changes in spacing before and after the earthquake are measures of the coseismic horizontal shortening. Finally, we carried out an analysis of errors in the results. 3. Results 3.1. Horizontal shortening measured from the pre- and postearthquake satellite images The two roads between Taziping and Gaoyuancun extend along the rupture of the Wenchuan Earthquake for 5 km (Fig. 1). We measured the distance between the two roads by projecting them onto the direction 140°, normal to the average strike of the rupture. Measurements of the spacing between the two roads were made at 3–5 m intervals. Consequently, we obtained 906 measurements of spacing, both before and after the earthquake (Fig. 2a). Comparing the 906 data pairs, we obtained a series of results that represents the horizontal shortenings. The maximum, minimum, and average values are 11.0 m, 1.1 m, and 6.6 m, respectively (Fig. 2b). 3.2. Horizontal shortening measured from the pre-earthquake satellite images and GPS field measurements after the earthquake At 180 days after the earthquake, we made a GPS RTK survey along a 2 km length of the two roads. Taking measurements at 3–5 m intervals along the strike, we obtained 199 measurements of the spacings between the two roads (shown by the black dotted lines in Fig. 2c). Comparing these spacings with those extracted from the satellite images before the earthquake on the same road segments, we obtained a second series of results for the horizontal shortenings. The maximum, minimum, and average values are 9.85 m, 3.8 m, and 7.1 m, respectively (Fig. 2d). 3.3. Analysis of errors The errors in the satellite images depend on five factors: orbit, viewing angle, topography (stereoscopic effect), platform attitude (or in other words, roll, yaw, and pitch), and geometry of the detectors (Pumbroeck et al., 2000). We selected Ortho-Ready QuickBird images with a resolution of 0.6 m. The absolute accuracy RMSE (Root Mean Square Error) of the images is 14 m without any other disposals (DigitalGloble, Inc., 2006), and this is insufficient to measure the coseismic displacements. However, accuracies up to 2 m RMSE can be expected (DigitalGloble, Inc., 2005), when the images are processed further using commercial software, Rational Polynomial Coefficients (RPCs), high-quality DEM (e.g., DTED Level 2: 30 m), and sub-meter GCPs. For even higher accuracy, we selected a higher-quality DEM with a resolution of 20 m, and sub-centimeter GCPs, and as a result, the RMSE of the images is estimated to be 0.7 m. The stereoscopic effect leads to offsets characterized by strong gradients, particularly in areas with steep topography. Those offsets are thus difficult to measure, and the standard methods yield
Fig. 2. (a) Spacing between the two roads in the direction of 140°; red dotted lines and black dotted lines show them pre- and post-earthquake, respectively; the black rectangle shows the field measurement area (see c and d). (b) Coseismic horizontal shortening calculated by the method depicted in Section 3.1, red dots show the maximum and minimum values, the red dashed line show the integrated average, and the black dashed lines show the averages from two measured sections. (c) Spacing between the two roads in the direction of 140°; the red and black dotted lines show the data for before and after the earthquake, respectively. (d) Coseismic horizontal shortening calculated by the method depicted in Section 3.2; red dots show the maximum and minimum values, and the red dashed line shows the integrated average. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
measurements with an accuracy of 0.5 pixel (RMS) (Pumbroeck et al., 2000). Although the topography of the Baishahe section has particularly strong gradients, the topography of the two road markers is not so steep. Considering the topography of the two roads and the resolution of QuickBird images, the errors due to topography are much smaller than 0.3 m. In other words, it is reasonable to take the RMSE of the images to be 0.7 m (e1 = 0.7 m), which includes the errors due to topography. A second significant error can come from the process of extracting the road data from the images. Generally, this error is less than one pixel (0.6 m), so we use here the reasonable figure of 0.6 m (e2 = 0.6 m). The method described in Section 3.1 compares a series of road spacings measured from pre- and post-earthquake satellite images. The method did not involve a direct comparison of the pre- and post-earthquake satellite images, which means that the errors that
F. Shi et al. / Journal of Asian Earth Sciences 50 (2012) 164–170
167
Fig. 3. Distribution of the differences in measured values using two methods. The differences are equal to the results using the method described in Section 3.1 minus the results using the method of Section 3.2.
might have resulted from that are effectively eliminated. Other possible errors, such as those due to rotation caused by rectification, can also be neglected because they are much less important than the two main errors already taken into account. Having identified the two major errors, we can note further that the relative accuracy of field measurements, using the GPS RTK technique, is better than a centimeter (e3 = 0.01 m). Therefore, the integrated error when measuring the horizontal shortening with the method described in Section 3.1 is 1.84 m (E1 from the following Eq. (1)), and it is 1.3 m (E2 from the following Eq. (2)) when using the method described in Section 3.2 (Fig. 3).
E1 ¼
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ðe21 þ e22 Þ þ ðe21 þ e22 Þ þ ðe21 þ e22 Þ þ ðe21 þ e22 Þ 1:84 m
ð1Þ
E2 ¼
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ðe21 þ e22 Þ þ ðe21 þ e22 Þ þ e23 þ e23 1:3 m
ð2Þ
If the two roads were absolutely parallel, any strike-slip component would not influence the calculations, and our calculations have been based on the assumption that no strike-slip component exists. In fact, although dip-slip is dominant along the Baishahe River section, a component of strike-slip can be found in some places. Unfortunately, we could not obtain a strike-slip component for each measurement point along the two roads, and the two roads are not absolutely parallel. Variations in the directions of the roads are within 30°, and the average strike-slip component along the Baishahe section is about 2 m (He et al., 2008). If we take this figure of 2 m, together with a maximum deviation of 30° from parallelism for the two roads, the error would be 1.16 m. This increases the calculated error E1 to a new total of 2.17 m, when using the method of Section 3.1, and increases the calculated E2 to a new total of 1.74 m for the method of Section 3.2. In order to further examine this analysis of errors, we have compared the results in Sections 3.1 and 3.2 point by point (Fig. 3). The differences between the two methods may derive from the calculated errors or the corresponding uncertainties. But the contrasts in values are all smaller than 2 m, and the RMSE is 1.18 m, which is smaller than both E1 and E2. This provides support for the validity and reasonableness of the analysis of errors that we have provided. 4. Discussion The measured amounts of shortening have quite large variations, from 1.1 m to 11.0 m (black dots in Fig. 4), or from 3.8 m to 9.85 m (purple dots in Fig. 4), despite the claimed errors of 1.84 m (red bars in Fig. 4) or 1.3 m (yellow bars in Fig. 4). The minimum value of 1.1 m was measured in Huangjiaping village, and the other small values below 3 m are all from locations near Huangjiaping. The two roads happened to be on the same side of the fault near Huangjiaping village, and this is the main reason for the small horizontal shortenings measured here. But, even if we ignore the smallest and biggest horizontal shortenings, the main horizontal shortenings still exhibit quite large variations, from 4 to 10 m. We suggest that variations in fault geometry are the most important factors in controlling these fluctuations in the values of horizontal shortening. If the slip vector was constant in the
Fig. 4. Distribution of the values of coseismic horizontal shortening along the Baishahe segment. Black dots with short red bars represent coseismic horizontal shortenings and the errors calculated in Section 3.1; purple dots with short yellow bars represent coseismic horizontal shortenings and the errors calculated in Section 3.2; blue vertical lines are the coseismic horizontal shortenings calculated on the basis of restoring deformed man-made constructions (Shi et al., 2009). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Baishahe River section, the amount of horizontal shortening would be determined by the rupture geometry. Although dips are generally difficult to obtain from the rupture fault at the surface, two exact values for the dip of the fault, 76° and 56°, have been obtained from an outcrop and a trench, respectively. These values indicate a large variation in the dip of the fault in the Baishahe River section, consistent with the variation in dip of 38.3° to 64.7°, observed on the surface rupture by He et al. (2011). If we assume that dip slip vector is constant, and that the fault dip variation is only from 56° to 76° along the Baishahe River section, the maximum amount of horizontal shortening (scos 56° = 0.56 s) with a fault dip of 56° is much larger than the minimum amount of horizontal shortening (scos 76° = 0.24 s) with a fault dip of 76°. It indicates that variations in fault dip can explain the apparently wild fluctuations in the amounts of horizontal shortening. Furthermore, the horizontal shortenings are also affected by the surface terrain (He et al., 2008; Ran et al., 2010) and the thickness of the soil (Yuan, 2008). Considering all these factors together, namely the variations in fault dip, the influence of the thickness of soil, and the nature of the surface terrain, the large variations in the measured shortenings are not unexpected. The distribution of measured horizontal shortening values (Fig. 4) shows an abrupt step at 4.5 km, near Xiajiaping village. The horizontal shortenings in the northeastern part of the section are 2.2 m larger on average than those in the southwestern part (Fig. 4). Field investigations found that the geometries of the surface rupture in these two parts of the section are very different from each other. The northeastern part is more complex than the southwestern part (He et al., 2008; Liu-Zeng et al., 2010), and it consists of two sub-parallel strands, dipping in the opposite direction to the general NW-dip, with the block between them extruded due to compression. The southwestern part consists of a relatively simple single fault trace (He et al., 2008). Liu-Zeng et al. (2010) put forward a bold model to explain the locally reversed dip of the surface-rupture planes, and the corresponding sense of vertical motion on the Baishahe River rupture section. It seems clear that the sharp break in the distribution pattern of horizontal shortening measurements is mainly due to these different fault geometries and kinematics in the two parts of the section. Moreover, the horizontal shortenings obtained by the method described in Section 3.2 are almost all from the northeastern part of the section. Contrasting the two series of horizontal shortenings obtained from the two different methods in the same region, we found that the average value (7.4 m), using the method in Section 3.1, is 0.3 m larger than the average value (7.1 m) using the method in Section 3.2. This small difference of 0.3 m may indicate a small
168
F. Shi et al. / Journal of Asian Earth Sciences 50 (2012) 164–170
overestimation of the amount of horizontal shortening when using only the satellite images. Nevertheless, the difference is within the margins of error. For faults generally, the widths of zones with observable deformation are hardly ever more than 30 m, no matter whether the deformation was generated by strike-slip or thrusting (Xu et al., 2002). For the Wenchuan Earthquake, the widths are usually less than 40 m, with half of them between 10 and 30 m (Zhou et al., 2010), and just a few reaching 100 m (Xu et al., 2009). This means that field measurements associated with the Wenchuan Earthquake are derived from a zone less than 100 m wide. Based on field measurements, the maximum horizontal shortening in the Beichuan rupture segment along the Beichuan–Yingxiu Fault is 3.3 m (Li et al., 2009), in the Bailu rupture section, along the Guanxian– Jiangyou Fault, it is 2.83 m (Wang et al., 2010), and these figures are consistent with the maximum horizontal shortening of 3.37 m in the Baishahe River rupture section (Shi et al., 2009). By measuring and recovering information from a number of sources, Shi et al. (2009) obtained eight horizontal shortening measurements with a minimum value of 0.15 m and a maximum of 3.37 m. From these figures, we can see that all directly measured amounts of horizontal shortening in the observable surface rupture zone are less on average than the shortening amounts we determined with the methods described in Sections 3.1 and 3.2. Based on the GPS coseismic vector, we have calculated the coseismic horizontal shortening in the far field, also along the 140° direction. The results show that shortening decreases gradually from 2.1 m (covering a 40 km span across the rupture) to 0.1 m (covering a 600 km span). The coseismic horizontal shortenings for different spans are shown in Fig. 5. The purple dots in Fig. 5 represent the results obtained by Shi et al. (2009) through surface observation, and they show spans that are all less than 100 m. The blue dots in Fig. 5 are the results we obtained from GPS data, and we find that spans across the rupture are all over 10 km. The black and red dots in Fig. 5 represent our results obtained by the methods described in Sections 3.1 and 3.2. Our measured shortenings are in general larger than those derived simply from GPS data, which is consistent with the theoretic model of a coseismic deformation field. This model shows that the greater the distance from the rupture is, the smaller the deformation
Fig. 5. (a) Distribution of the values of coseismic horizontal shortenings for different span lengths across the rupture, and using different methods; purple dots are horizontal shortenings for spans of less than 100 m across the rupture, using direct measurements in the field (Shi et al., 2009), and the average is shown by the purple dashed line; black dots and red dots represent our results for spans of 100– 1000 m across the rupture, calculated using the methods described in Sections 3.1 and 3.2, respectively, and their averages are shown by black and red dashed lines, respectively; blue dots represent the horizontal shortenings for spans greater than 10 km, using GPS data (The Project of Crustal Movement Observation Network of China, 2008), and the average values are shown by blue dashed line. (b) The theoretic model of coseismic displacement (modified from Mavko, 1981). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
is (Fig. 5b). The coseismic deformation is produced by relieving the strain of dynamic faulting, and the model has been verified by the successful application of elastic dislocation models (Chinnery, 1961; Savage and Hastie, 1966). The observed coseismic deformation fields produced by the large Tango earthquake (strike-slip) in Japan, 1927 (Mavko, 1981), and the Nankaido earthquake (thrusting) in Japan, 1946 (Fitch and Scholz, 1971), are also consistent with the theoretic model. However, our results for larger spans are, on average, larger than those for smaller spans, where the measurements were made directly in the field immediately around the rupture, and this seems quite different from the theoretic prediction. A similar phenomenon was observed for some other strong earthquakes, such as the Chi–Chi earthquake of 1999 (Lee et al., 2010), the Landers earthquake of 1992 (Pumbroeck et al., 2000), and the Kunlun earthquake of 2001 (Klinger et al., 2006). All their results indicate that the place of maximum deformation is not immediately within the surface rupture, but distributed over a wider span of hundreds of meters to 1 km across the fault. After the Chi–Chi earthquake of 1999, Lee et al. (2010) found a deformation zone of about 200 m located on the hanging wall next to the fault. Our results imply that there is a similar deformation zone associated with the Wenchuan Earthquake. There could be other possible reasons for our measured amounts of shortening to be larger than amounts measured directly in the field, and these include variations in the dip of the fault with depth (Shi et al., 2009). A comparison of our results with the direct field measurements suggests that the coseismic shortening deformation was actually distributed over a wider area, including some continuous and discontinuous unobservable parts. The fact that the amounts of coseismic horizontal shortening measured by us are larger than either the coseismic vertical offsets observed in the field (He et al., 2008; Liu-Zeng et al., 2010; Xu et al., 2009) or the geophysical inversions (Shen et al., 2009), implies that the rupture is a low-angle (<45°) fault at depth, even though it dips (>45°) steeply at the surface. Based on the seismic reflection data and well logs (Jia et al., 2003; Luo, 1998; Song, 1994), as well as the observed coseismic deformation data, Xu et al. (2009) proposed a low-angle fault model, with a predicted total crustal shortening of 8.5 m, and a total vertical uplift of 7.5 m. Our estimates of the average horizontal shortening support these predictions. With regard to the building of the steep Longmenshan range, there have been three well-known hypotheses proposed: channel flow (Burchfiel et al., 2008; Clark and Royden, 2000; Royden et al., 2008), ‘‘classical’’ tectonic convergence (Hubbard and Shaw, 2009; Liu-Zeng et al., 2009; Xu et al., 2009), and tectonic deformation plus isostatic rebound caused by long-term rapid erosion (Fu et al., 2011). Surface deformation, especially when only caused by an earthquake, cannot be directly used to determine the structure in the lower crust. But the surface deformation may provide useful pointers to the nature of the structure in the lower crust, or provide useful information. For the Wenchuan Earthquake, the coseismic landsliding produced a greater volume of erodible material than the volume added to the orogen by coseismic rock uplift (Ouimet, 2011; Parker et al., 2011). This discrepancy indicates that it is not just rock uplift during earthquakes that is responsible for the high topography of the Longmengshan range, and a more general shortening of the crust may provide the explanation. In this context, our estimates of horizontal shortenings provide essential information for understanding the seismotectonics of eastern Tibet, especially the building of the Longmenshan range.
5. Conclusions We have obtained a series of horizontal shortening measurements by contrasting the spacings between two roads observed
F. Shi et al. / Journal of Asian Earth Sciences 50 (2012) 164–170
on QuickBird satellite images made before and after the Wenchuan Earthquake. The maximum, minimum, and average values are 11.0 m, 1.1 m, and 6.6 m, respectively. The error is estimated to be 1.84 m. By comparing the spacings between the two roads using pre-earthquake images and GPS field measurements after the earthquake, we obtained a second series of horizontal shortening measurements. In this case, the maximum value is 9.85 m, the minimum 3.8 m, and the average 7.1 m. A comparison of our coseismic horizontal shortening measurements with those derived simply from GPS data, shows our results to be much larger on average. This is consistent with the theoretic model for a coseismic deformation field. On the other hand, our results for large spans across the fault are, on average, larger than those obtained from direct measurements in the field, immediately adjacent to the fault for smaller spans, and this differs from the theoretic prediction, suggesting that the coseismic shortening was actually distributed over a wide area, including some unobservable continuous or discontinuous parts. Finally, our estimates of the amount of coseismic horizontal shortening support the low-angle fault model for the Wenchuan Earthquake (Xu et al., 2009). Furthermore, our results help us reach a better understanding of the seismotectonics of eastern Tibet, especially the mountain building of the Longmenshan range. Acknowledgements This research was supported by the National Natural Science Foundation of China (Grant Numbers 40841019 and 40872128) and special funds from the Science and Technology Ministry of China for the Scientific Investigation of the Wenchuan Earthquake Rupture. We thank Dr. Zongqi Yue for his very helpful review that improved the manuscript. Many thanks are also due to anonymous reviewers for their constructive comments, and to the journal editors for their corrections. References Burchfiel, B.C., Royden, L.H., Van der Hilst, R.D., Hager, B.H., Chen, Z., King, R.W., Li, C., Lü, J., Yao, H., Kirby, E., 2008. A geological and geophysical context for the Wenchuan earthquake of 12 May 2008, Sichuan, People’s Republic of China. GSA Today 18, 4–11. Chen, G.H., Xu, X.W., Yu, G.H., An, Y.F., Yuan, R.M., Guo, T.T., Gao, X., Tan, X.B., 2009. Co-seismic slip and slip partitioning of multi-faults during the Ms8.0 2008 Wenchuan earthquake. Chinese Journal of Geophysics 52 (5), 1384–1394 (in Chinese with English abstract). Chen, G.H., Xu, X.W., Zheng, R.Z., Yu, G.H., Li, F., Li, C.X., Wen, X.Z., He, Y.L., Ye, Y.Q., Chen, X.Ch., Wang, Z.C., 2008. Quantitative analysis of the co-seismic surface rupture of the 2008 Wenchuan earthquake, Sichuan, China along the Beichuan– Yingxiu Fault. Seismology and Geology 30 (3), 723–738 (in Chinese with English abstract). Chinnery, M.A., 1961. Deformation of the ground around surface faults. Bulletin of the Seismological Society of America 51, 355–372. Clark, M., Royden, L.H., 2000. Topographic ooze: building the eastern margin of Tibet by lower crustal flow. Geology 28, 703–706. Deng, Q.D., 2007. Map of Active Tectonics in China. Seismological Press, Beijing. DigitalGloble, Inc., 2005. QuickBird Imagery Products – FAQ. DigitalGloble, Inc., 2006. QuickBird Imagery Products – Product Guide. Dominguez, S., Avouac, J.P., Michel, R., 2003. Horizontal coseismic deformation of the 1999 Chi–Chi earthquake measured from SPOT satellite images: implications for the seismic cycle along the western foothills of central Taiwan. Journal of Geophysical Research 108 (B2), 2083. doi:10.1029/ 2001JB000951. Fitch, T.J., Scholz, C.H., 1971. Mechanism of underthrusting in southwest Japan: a model of convergent plate interactions. Journal of Geophysical Research 76, 7260–7292. Fu, B.H., Shi, P.L., Guo, H.D., Okuyama, S., Ninomiya, Y., Wright, Y., 2011. Surface deformation related to the 2008 Wenchuan earthquake, and mountain building of the Longmen Shan, eastern Tibetan Plateau. Journal of Asian Earth Science 40, 805–824. Hao, K.X., Si, H., Fujiwara, H., Ozawa, T., 2009. Coseismic surface-ruptures and crustal deformations of the 2008 Wenchuan earthquake Mw7.9, China. Geophysical Research Letters 36, L11303. doi:10.1029/2009GL03791. Hashimoto, M., Enomoto, M., Fukushima, Y., 2010. Coseismic deformation from the 2008 Wenchuan, China, earthquake derived from ALOS/PALSAR Images. Tectonophysics 491, 59–71.
169
He, H.L., Sun, Z.M., Wei, Z.Y., Dong, Sh.P., Gao, X., Wang, Sh.Y., Wang, J.Q., 2008. Rupture of the MS8.0 Wenchuan earthquake along Baishahe River. Seismology and Geology 30 (3), 658–673 (in Chinese with English abstract). He, H.L., Wei, Zh.Y., Chen, Ch.Y., Shi, F., 2011. On-site determination of slip vectors along the Longmenshan fault during the Mw 7.9 Wenchuan, China, earthquake of May 12, 2008. Journal of Asian Earth Sciences 41, 274–282. Hubbard, J., Shaw, J.H., 2009. Uplift of the Longmen Shan and Tibetan Plateau, and the 2008 Wenchuan (M = 7.9) earthquake. Nature 458, 194–197. Jia, D., Chen, Z., Jia, C., Wei, G., Li, B., Zhang, Q., Wei, D.T., Shen, Y., 2003. Structural features of the Longmen Shan fold and thrust belt and development of the Western Sichuan foreland basin, central China. Geological Journal of China Universities 9 (3), 402–409 (in Chinese with English abstract). Klinger, Y., Michel, R., King, G.C.P., 2006. Evidence for an earthquake barrier model from Mw 7.8 Kokoxili (Tibet) earthquake slip-distribution. Earth and Planetary Science Letters 242, 354–364. Lee, Y.H., Wu, K.C., Rau, R.J., Chen, H.C., Lo, W., Cheng, K.C., 2010. Revealing cosesimic displacements and the deformation zones of the 1999 Chi–Chi Earthquake in the Tsaotung Area, Central Taiwan, using digital cadastral data. Journal of Geophysical Research 115 (B03419). doi:10.1029/2009JB006397. Li, C.Y., Wei, Z.Y., Ye, J.Q., Han, Y.B., Zheng, W.J., 2010. Amounts and styles of coseismic deformation along the northern segment of surface rupture, of the 2008 Wenchuan Mw 7.9 earthquake, China. Tectonophysics 491, 35–58. Li, T., Chen, J., Huang, M.D., Yu, S., 2009. Methods to calculating slip vectors of reverse-fault earthquake surface rupture: a case study in Wenchuan earthquake. Quaternary Sciences 29 (3), 524–534 (in Chinese with English abstract). Liu-Zeng, J., Wen, L., Sun, J., Zhang, Z.H., Hu, G.Y., Xing, X.C., Zeng, L., Xu, Q., 2010. Surficial slip and rupture geometry on the Beichuan Fault near Hongkou during the Mw 7.9 Wenchuan earthquake, China. Bulletin of the Seismological Society of America 100 (5B), 2615–2650. Liu-Zeng, J., Zhang, Z., Wen, L., Tapponnier, P., Sun, J., Xing, X., Hu, G., Xu, Q., Zeng, L., Ding, L., Ji, C., Hudnut, K.W., Van der Woerd, J., 2009. Co-seismic ruptures of the 12 May 2008. Ms8.0 Wenchuan earthquake, Sichuan: east–west crustal shortening on oblique, parallel thrusts along the eastern edge of Tibet. Earth and Planetary Science Letters 286 (3–4), 355–370. Luo, Z.L., 1998. New recognition of basement in Sichuan Basin. Journal of Chengdu University of Technology 25 (2), 191–200 (in Chinese with English abstract). Mavko, G.M., 1981. Mechanics of motion on major faults. Annual Review of Earth and Planetary Sciences 9, 81–111. Michel, R., Avouac, J.P., 2002. Deformation due to the 17 August 1999 Izmit, Turkey, earthquake measured from SPOT images. Journal of Geophysical Research 107 (B4), 2062. doi:10.1029/2001JB000102. Nakamura, T., Tsuboi, S., Kaneda, Y., Yamanaka, Y., 2010. Rupture process of the 2008 Wenchuan China earthquake inferred from teleseismic waveform inversion and forward modeling of broadband seismic waves. Tectonophysics 491, 72–84. Ouimet, W.B., 2011. The hills came tumbling down. Nature Geoscience 4, 424–425. Parker, R.N., Densmore, A.L., Rosser, N.J., Michele, M., Li, Y., Huang, R.Q., Whadcoat, S., Petley, D.N., 2011. Mass wasting triggered by the 2008 Wenchuan earthquake is greater than organic growth. Nature Geoscience 4, 449–452. Pumbroeck, N., Michel, R., Bient, R., Avouac, J.P., Taboury, J., 2000. Measuring earthquakes from optical satellite images. Applied Optics 39 (20), 3486–3494. Ran, Y.K., Shi, X., Wang, H., Chen, L.Ch., Chen, J., Liu, R.Ch., Gong, H.L., 2010. The maximum coseismic vertical surface displacement and surface deformation pattern accompanying the Ms 8.0 Wenchuan earthquake. Chinese Science Bulletin 55 (9), 841–850. Royden, L.H., Burchfiel, B.C., Van der Hilst, R.D., 2008. The geological evolution of the Tibetan Plateau. Science 321, 1054–1058. Savage, J.C., Hastie, L.M., 1966. Surface deformation associated with dip-slip faulting. Journal of Geophysical Research 71, 4897–4904. Shen, Z.-K., Sun, J., Zhang, P., Wan, Y., Wang, M., Bürgmann, R., Zeng, Y., Gan, W., Liao, H., Wang, Q., 2009. Slip maxima at fault junctions and rupturing of barriers during the 2008 Wenchuan earthquake. Nature Geoscience 2, 718–724. Shi, F., He, H.L., Wei, Zh.Y., 2009. Co-seismic horizontal shortening on Baishahe segment of Wenchuan earthquake rupture zone. Quaternary Sciences 29 (3), 546–553 (in Chinese with English abstract). Song, H.B., 1994. The comprehensive interpretation of geological and geophysical data in the organic belt of Longmen Mountains, China. Journal of Chengdu Institute of Technology 21 (2), 79–88 (in Chinese with English abstract). The Project of Crustal Movement Observation Network of China, 2008. The coseismic displacement fields of Wenchuan Ms8.0 earthquake occurrence in 2008 using GPS data. Science in China (Series D – Earth Sciences) 38(10), 1195– 1206 (in Chinese). Wang, H., Ran, Y.K., Chen, L.Ch., Shi, X., Liu, R.Ch., Gomez, F., 2010. Determination of horizontal shortening and amount of reverse-faulting from trenching across the surface rupture of the 2008 MW7.9 Wenchuan earthquake, China. Tectonophysics 491, 10–20. Wang, W.M., Zhao, L.F., Li, J., Yao, Z.X., 2008. Rupture process of the Ms 8.0 Wenchuan earthquake of Sichuan, China. Chinese Journal of Geophysics 51 (5), 1403–1410 (in Chinese with English abstract). Wang, Z., Fukao, Y., Pei, Sh.P., 2009. Structural control of rupturing of the Mw7.9 2008 Wenchuan earthquake, China. Earth and Planetary Science Letters 279, 131–138. Xu, X.W., Chen, G.H., Yu, G.H., Sun, X.Zh., Tan, X.B., Chen, L.C., Sun, J.B., Chen, Y.G., Chen, W.Sh., Zhang, Sh.P., Li, K., 2010. Reevaluation of surface rupture parameters of the 5.12 Wenchuan earthquake and its tectonic implication for
170
F. Shi et al. / Journal of Asian Earth Sciences 50 (2012) 164–170
Tibetan uplift. Chinese Journal of Geophysics 53 (10), 2321–2336 (in Chinese with English abstract). Xu, X.W., Wen, X.Z., Yu, G.H., Chen, G.H., Klinger, Y., Hubbard, J., Shaw, J., 2009. Coseismic reverse- and oblique-slip surface faulting generated by the 2008 Mw7.9 Wenchuan earthquake, China. Geology 37 (6), 515–518. Xu, X.W., Yu, G.H., Ma, W.T., Ran, Y.K., Chen, G.H., Han, Zh.J., Zhang, L.F., You, H.Ch., 2002. Evidence and methods for determining the safety distance from the potential earthquake surface rupture on active fault. Seismology and Geology 24 (4), 470–483 (in Chinese with English abstract).
Yuan, R.M., 2008. Study on the influence of the movement of earthquake fault on the overlaying soil stratum based on the centrifuge modeling test & change law of geofluid pressure during basin evolution. Institute of Geology, China Earthquake Administration, Beijing. Zhou, Q., Xu, X.W., Yu, G.H., Chen, X.Ch., He, H.L., Yin, G.M., 2010. Width distribution of the surface ruptures associated with the Wenchuan earthquake: implication for the setback zone of the seismogenic faults in postquake reconstruction. Bulletin of the Seismological Society of America 100 (5B), 2660–2668.
Contents lists available at SciVerse ScienceDirect
Journal of Asian Earth Sciences journal homepage: www.elsevier.com/locate/jseaes
Coseismic horizontal shortening associated with the 2008 Wenchuan Earthquake along the Baishahe segment from high resolution satellite images Feng Shi, Hong-Lin He ⇑, Zhan-Yu Wei Key Laboratory of Active Tectonics and Volcano, Institute of Geology, China Earthquake Administration, Beijing 100029, China
a r t i c l e
i n f o
Article history: Received 12 May 2010 Received in revised form 11 January 2012 Accepted 17 January 2012 Available online 22 February 2012 Keywords: Wenchuan Earthquake Coseismic horizontal shortening Image contrast
a b s t r a c t Many coseismic deformation data from the Wenchuan Earthquake of 12 May, 2008, have been published; however, most of the data record strike-slip and vertical offsets, and there is little information on the horizontal shortening components. To determine the amount of horizontal shortening, we suggest a new method of measuring the differences in the positions of geometrical markers observed in satellite images before and after the earthquake. We found two roads that run nearly parallel to the earthquake rupture but on opposite sides, and we examined the spacing between them before and after the earthquake. We were able to measure horizontal shortenings that average 6.6 ± 1.84 m, with a maximum of 11.0 m and a minimum of 1.1 m. As an alternative method, we used a GPS RTK field survey system to measure the spacing of the two roads after the earthquake, and then measured the difference between that and the spacing determined from satellite images before the earthquake. This method gave horizontal shortenings that average 7.1 ± 1.3 m, with a maximum of 9.85 m and a minimum of 3.8 m. The shortenings measured, using either of these two methods, are larger than those directly measured in the field in the immediate vicinity of the fault, which suggests that the coseismic horizontal shortening is distributed over a wider area that extends some distance away from the brittle fault scarp itself. Finally, the results support a low-angle fault model for the Wenchuan Earthquake, and they provide useful insights into the seismotectonics of eastern Tibet, especially the building of the Longmenshan range. Ó 2012 Elsevier Ltd. All rights reserved.
1. Introduction The Ms 8.0 Wenchuan Earthquake, which struck Sichuan, China, on 12 May 2008, was accompanied by the rupture of two NWdipping imbricate reverse faults along the Longmenshan fault zone at the eastern margin of the Tibetan Plateau (Burchfiel et al., 2008; Liu-Zeng et al., 2009; Xu et al., 2009). Seismological studies indicate that the main shock was initiated approximately 20 km southwest of the town of Yingxiu, with the rupture propagating northeastwards for about 300 km at an average speed of 2.8 km/ s, and for a period of 110–120 s (Nakamura et al., 2010; Wang et al., 2008, 2009). Field investigations show that the earthquake generated a major rupture about 250 km in length along the Beichuan–Yingxiu Fault, as well as a minor rupture about 72 km in length along the parallel Guanxian–Jiangyou Fault (Li et al., 2010; Liu-Zeng et al., 2009; Xu et al., 2009). The observed co-seismic slips show that faulting on the major rupture zone (the Beichuan–Yingxiu Fault) consisted of upwards thrusting of the northwest hanging wall, together with some component of right-lateral strike-slip. On ⇑ Corresponding author. Address: 19 Beituchengxilu Road, Beijing 100029, China. Tel./fax: +86 10 62009215. E-mail addresses: [email protected] (F. Shi), [email protected] (H.-L. He), [email protected] (Z.-Y. Wei). 1367-9120/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. doi:10.1016/j.jseaes.2012.01.001
the minor rupture zone, the Guanxian–Jiangyou Fault, only reverse slip is observed (Liu-Zeng et al., 2009; Xu et al., 2009). Along the Beichuan–Yingxiu Fault, the observed maximum vertical displacements are 9.5 m (Ran et al., 2010) or 6.5 m (Xu et al., 2010), and the strike slip displacements are 4.8 m (He et al., 2008) or 4.9 m (Xu et al., 2009). On the Guanxian–Jiangyou Fault, the observed maximum vertical displacement is 3.5 m (Liu-Zeng et al., 2009; Xu et al., 2009). Most of the coseismic slip data, collected in the field (e.g., He et al., 2008; Liu-Zeng et al., 2009; Xu et al., 2009), record strike-slip and vertical offsets. The amount of horizontal shortening, which is as important as the strike-slip and vertical offsets, has seldom been measured (Chen et al., 2008, 2009; He et al., 2008; Shi et al., 2009; Wang et al., 2010). The coseismic slip vector has strike slip, vertical slip, and horizontal shortening (or extension) as its three components in space, and a reasonable and reliable estimate of all three components is essential if we are to understand the coseismic surface deformation. Although GPS and InSAR data have provided information on the coseismic displacement field of the Wenchuan Earthquake (Hao et al., 2009; Hashimoto et al., 2010; The Project of Crustal Movement Observation Network of China, 2008), the InSAR data often fail near the fault due to excessive temporal decorrelation of the images and signal saturation (Klinger et al., 2006). Furthermore, SAR data only provide the satellite to ground component
F. Shi et al. / Journal of Asian Earth Sciences 50 (2012) 164–170
165
Fig. 1. Surface ruptures in the Baishahe segment and locations of the reference roads. (a) Active tectonics in China (Deng, 2007); the black rectangle shows the location of (b). (b) Surface rupture formed during the Wenchuan Earthquake shown by red lines (Xu et al., 2009); black lines show other active faults; black rectangle shows the location of (c). (c) Surface ruptures in the Baishahe segment (from He et al. (2008)) and selected roads; black rectangle shows the location of (d). (d) Local roads on QuickBird images preearthquake (blue zone) and post-earthquake (red zone), and the eastern roadside measured in the field (black dotted line). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
of the deformation (Michel and Avouac, 2002). Because there are only a few stations immediately adjacent to the fault, the GPS data only show the deformation field at some distance from the fault (The Project of Crustal Movement Observation Network of China, 2008). The recently developed sub-pixel correlation method provides new information on co-seismic displacements, and it has now been applied to several earthquakes (Dominguez et al., 2003; Klinger et al., 2006; Michel and Avouac, 2002; Pumbroeck et al., 2000). The method is limited mainly by the decorrelation of the images, the accuracy of the digital elevation model (DEM), the aliasing of the images, and the uncertainties concerning the measured roll, pitch, and yaw of the satellite (Pumbroeck et al., 2000). To avoid the error due to the contrast of two optical satellite images, we provide here a new and simpler method of measuring the coseismic shortening by subtracting differences in the positions of geometrical markers observed in satellite images before and after the earthquake. In this paper, the rupture along the Baishahe River is used as a case study to calculate the coseismic shortenings. This rupture is located to the southwest of the major rupture, it extends 14 km along the Baishahe River, and is approximately 15 km northeast of the epicenter of the Wenchuan Earthquake. Its geometry is complicated, with many short sections striking 50° on average and dipping
northwest. The maximum vertical and strike-slip coseismic offsets, measured along the Baishahe River rupture (He et al., 2008), are 6.5 m and 4.8 m, respectively. There are two roads that run nearly parallel to most of the rupture (Fig. 1), one to the northwest of the rupture, the other to the southeast. They provide excellent markers for measuring the horizontal shortening. On satellite images we can measure the distance between the two roads normal to the average strike of the rupture, and observe the changes before and after the Wenchuan Earthquake. These observations can then be used to calculate the coseismic horizontal shortening.
2. Methods and data QuickBird satellite images were used for this study because of their high resolution (0.6 m), arguably the best among the commercial satellite images presently available. The two roads run parallel to the Baishahe River rupture, but on opposite sides, and they are used as geometric markers. The procedures were as follows: Firstly, a field survey was made of the two selected roads using a GPS RTK measuring system. Road segments that were apparently repaired or reconstructed after the earthquake were not included in the survey.
166
F. Shi et al. / Journal of Asian Earth Sciences 50 (2012) 164–170
Secondly, we made an orthorectification of optical satellite images. The pre-earthquake satellite image was rectified using the 20 m pixel DEM, made from 1:50,000 topographic maps published by the National Geomatics Center of China, while the postearthquake satellite image was rectified by Ground Control Points (GCPs) and the 20 m pixel DEM. Thirdly, we extracted the roads in the pre- and post-earthquake ortho-corrected satellite images, and measured the spacings between them along the direction perpendicular to the rupture in both the pre- and post-earthquake images. Fourthly, we measured the differences in the spacings between the roads, before and after the earthquake, for various positions along the strike of the rupture. The changes in spacing before and after the earthquake are measures of the coseismic horizontal shortening. Finally, we carried out an analysis of errors in the results. 3. Results 3.1. Horizontal shortening measured from the pre- and postearthquake satellite images The two roads between Taziping and Gaoyuancun extend along the rupture of the Wenchuan Earthquake for 5 km (Fig. 1). We measured the distance between the two roads by projecting them onto the direction 140°, normal to the average strike of the rupture. Measurements of the spacing between the two roads were made at 3–5 m intervals. Consequently, we obtained 906 measurements of spacing, both before and after the earthquake (Fig. 2a). Comparing the 906 data pairs, we obtained a series of results that represents the horizontal shortenings. The maximum, minimum, and average values are 11.0 m, 1.1 m, and 6.6 m, respectively (Fig. 2b). 3.2. Horizontal shortening measured from the pre-earthquake satellite images and GPS field measurements after the earthquake At 180 days after the earthquake, we made a GPS RTK survey along a 2 km length of the two roads. Taking measurements at 3–5 m intervals along the strike, we obtained 199 measurements of the spacings between the two roads (shown by the black dotted lines in Fig. 2c). Comparing these spacings with those extracted from the satellite images before the earthquake on the same road segments, we obtained a second series of results for the horizontal shortenings. The maximum, minimum, and average values are 9.85 m, 3.8 m, and 7.1 m, respectively (Fig. 2d). 3.3. Analysis of errors The errors in the satellite images depend on five factors: orbit, viewing angle, topography (stereoscopic effect), platform attitude (or in other words, roll, yaw, and pitch), and geometry of the detectors (Pumbroeck et al., 2000). We selected Ortho-Ready QuickBird images with a resolution of 0.6 m. The absolute accuracy RMSE (Root Mean Square Error) of the images is 14 m without any other disposals (DigitalGloble, Inc., 2006), and this is insufficient to measure the coseismic displacements. However, accuracies up to 2 m RMSE can be expected (DigitalGloble, Inc., 2005), when the images are processed further using commercial software, Rational Polynomial Coefficients (RPCs), high-quality DEM (e.g., DTED Level 2: 30 m), and sub-meter GCPs. For even higher accuracy, we selected a higher-quality DEM with a resolution of 20 m, and sub-centimeter GCPs, and as a result, the RMSE of the images is estimated to be 0.7 m. The stereoscopic effect leads to offsets characterized by strong gradients, particularly in areas with steep topography. Those offsets are thus difficult to measure, and the standard methods yield
Fig. 2. (a) Spacing between the two roads in the direction of 140°; red dotted lines and black dotted lines show them pre- and post-earthquake, respectively; the black rectangle shows the field measurement area (see c and d). (b) Coseismic horizontal shortening calculated by the method depicted in Section 3.1, red dots show the maximum and minimum values, the red dashed line show the integrated average, and the black dashed lines show the averages from two measured sections. (c) Spacing between the two roads in the direction of 140°; the red and black dotted lines show the data for before and after the earthquake, respectively. (d) Coseismic horizontal shortening calculated by the method depicted in Section 3.2; red dots show the maximum and minimum values, and the red dashed line shows the integrated average. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
measurements with an accuracy of 0.5 pixel (RMS) (Pumbroeck et al., 2000). Although the topography of the Baishahe section has particularly strong gradients, the topography of the two road markers is not so steep. Considering the topography of the two roads and the resolution of QuickBird images, the errors due to topography are much smaller than 0.3 m. In other words, it is reasonable to take the RMSE of the images to be 0.7 m (e1 = 0.7 m), which includes the errors due to topography. A second significant error can come from the process of extracting the road data from the images. Generally, this error is less than one pixel (0.6 m), so we use here the reasonable figure of 0.6 m (e2 = 0.6 m). The method described in Section 3.1 compares a series of road spacings measured from pre- and post-earthquake satellite images. The method did not involve a direct comparison of the pre- and post-earthquake satellite images, which means that the errors that
F. Shi et al. / Journal of Asian Earth Sciences 50 (2012) 164–170
167
Fig. 3. Distribution of the differences in measured values using two methods. The differences are equal to the results using the method described in Section 3.1 minus the results using the method of Section 3.2.
might have resulted from that are effectively eliminated. Other possible errors, such as those due to rotation caused by rectification, can also be neglected because they are much less important than the two main errors already taken into account. Having identified the two major errors, we can note further that the relative accuracy of field measurements, using the GPS RTK technique, is better than a centimeter (e3 = 0.01 m). Therefore, the integrated error when measuring the horizontal shortening with the method described in Section 3.1 is 1.84 m (E1 from the following Eq. (1)), and it is 1.3 m (E2 from the following Eq. (2)) when using the method described in Section 3.2 (Fig. 3).
E1 ¼
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ðe21 þ e22 Þ þ ðe21 þ e22 Þ þ ðe21 þ e22 Þ þ ðe21 þ e22 Þ 1:84 m
ð1Þ
E2 ¼
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ðe21 þ e22 Þ þ ðe21 þ e22 Þ þ e23 þ e23 1:3 m
ð2Þ
If the two roads were absolutely parallel, any strike-slip component would not influence the calculations, and our calculations have been based on the assumption that no strike-slip component exists. In fact, although dip-slip is dominant along the Baishahe River section, a component of strike-slip can be found in some places. Unfortunately, we could not obtain a strike-slip component for each measurement point along the two roads, and the two roads are not absolutely parallel. Variations in the directions of the roads are within 30°, and the average strike-slip component along the Baishahe section is about 2 m (He et al., 2008). If we take this figure of 2 m, together with a maximum deviation of 30° from parallelism for the two roads, the error would be 1.16 m. This increases the calculated error E1 to a new total of 2.17 m, when using the method of Section 3.1, and increases the calculated E2 to a new total of 1.74 m for the method of Section 3.2. In order to further examine this analysis of errors, we have compared the results in Sections 3.1 and 3.2 point by point (Fig. 3). The differences between the two methods may derive from the calculated errors or the corresponding uncertainties. But the contrasts in values are all smaller than 2 m, and the RMSE is 1.18 m, which is smaller than both E1 and E2. This provides support for the validity and reasonableness of the analysis of errors that we have provided. 4. Discussion The measured amounts of shortening have quite large variations, from 1.1 m to 11.0 m (black dots in Fig. 4), or from 3.8 m to 9.85 m (purple dots in Fig. 4), despite the claimed errors of 1.84 m (red bars in Fig. 4) or 1.3 m (yellow bars in Fig. 4). The minimum value of 1.1 m was measured in Huangjiaping village, and the other small values below 3 m are all from locations near Huangjiaping. The two roads happened to be on the same side of the fault near Huangjiaping village, and this is the main reason for the small horizontal shortenings measured here. But, even if we ignore the smallest and biggest horizontal shortenings, the main horizontal shortenings still exhibit quite large variations, from 4 to 10 m. We suggest that variations in fault geometry are the most important factors in controlling these fluctuations in the values of horizontal shortening. If the slip vector was constant in the
Fig. 4. Distribution of the values of coseismic horizontal shortening along the Baishahe segment. Black dots with short red bars represent coseismic horizontal shortenings and the errors calculated in Section 3.1; purple dots with short yellow bars represent coseismic horizontal shortenings and the errors calculated in Section 3.2; blue vertical lines are the coseismic horizontal shortenings calculated on the basis of restoring deformed man-made constructions (Shi et al., 2009). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Baishahe River section, the amount of horizontal shortening would be determined by the rupture geometry. Although dips are generally difficult to obtain from the rupture fault at the surface, two exact values for the dip of the fault, 76° and 56°, have been obtained from an outcrop and a trench, respectively. These values indicate a large variation in the dip of the fault in the Baishahe River section, consistent with the variation in dip of 38.3° to 64.7°, observed on the surface rupture by He et al. (2011). If we assume that dip slip vector is constant, and that the fault dip variation is only from 56° to 76° along the Baishahe River section, the maximum amount of horizontal shortening (scos 56° = 0.56 s) with a fault dip of 56° is much larger than the minimum amount of horizontal shortening (scos 76° = 0.24 s) with a fault dip of 76°. It indicates that variations in fault dip can explain the apparently wild fluctuations in the amounts of horizontal shortening. Furthermore, the horizontal shortenings are also affected by the surface terrain (He et al., 2008; Ran et al., 2010) and the thickness of the soil (Yuan, 2008). Considering all these factors together, namely the variations in fault dip, the influence of the thickness of soil, and the nature of the surface terrain, the large variations in the measured shortenings are not unexpected. The distribution of measured horizontal shortening values (Fig. 4) shows an abrupt step at 4.5 km, near Xiajiaping village. The horizontal shortenings in the northeastern part of the section are 2.2 m larger on average than those in the southwestern part (Fig. 4). Field investigations found that the geometries of the surface rupture in these two parts of the section are very different from each other. The northeastern part is more complex than the southwestern part (He et al., 2008; Liu-Zeng et al., 2010), and it consists of two sub-parallel strands, dipping in the opposite direction to the general NW-dip, with the block between them extruded due to compression. The southwestern part consists of a relatively simple single fault trace (He et al., 2008). Liu-Zeng et al. (2010) put forward a bold model to explain the locally reversed dip of the surface-rupture planes, and the corresponding sense of vertical motion on the Baishahe River rupture section. It seems clear that the sharp break in the distribution pattern of horizontal shortening measurements is mainly due to these different fault geometries and kinematics in the two parts of the section. Moreover, the horizontal shortenings obtained by the method described in Section 3.2 are almost all from the northeastern part of the section. Contrasting the two series of horizontal shortenings obtained from the two different methods in the same region, we found that the average value (7.4 m), using the method in Section 3.1, is 0.3 m larger than the average value (7.1 m) using the method in Section 3.2. This small difference of 0.3 m may indicate a small
168
F. Shi et al. / Journal of Asian Earth Sciences 50 (2012) 164–170
overestimation of the amount of horizontal shortening when using only the satellite images. Nevertheless, the difference is within the margins of error. For faults generally, the widths of zones with observable deformation are hardly ever more than 30 m, no matter whether the deformation was generated by strike-slip or thrusting (Xu et al., 2002). For the Wenchuan Earthquake, the widths are usually less than 40 m, with half of them between 10 and 30 m (Zhou et al., 2010), and just a few reaching 100 m (Xu et al., 2009). This means that field measurements associated with the Wenchuan Earthquake are derived from a zone less than 100 m wide. Based on field measurements, the maximum horizontal shortening in the Beichuan rupture segment along the Beichuan–Yingxiu Fault is 3.3 m (Li et al., 2009), in the Bailu rupture section, along the Guanxian– Jiangyou Fault, it is 2.83 m (Wang et al., 2010), and these figures are consistent with the maximum horizontal shortening of 3.37 m in the Baishahe River rupture section (Shi et al., 2009). By measuring and recovering information from a number of sources, Shi et al. (2009) obtained eight horizontal shortening measurements with a minimum value of 0.15 m and a maximum of 3.37 m. From these figures, we can see that all directly measured amounts of horizontal shortening in the observable surface rupture zone are less on average than the shortening amounts we determined with the methods described in Sections 3.1 and 3.2. Based on the GPS coseismic vector, we have calculated the coseismic horizontal shortening in the far field, also along the 140° direction. The results show that shortening decreases gradually from 2.1 m (covering a 40 km span across the rupture) to 0.1 m (covering a 600 km span). The coseismic horizontal shortenings for different spans are shown in Fig. 5. The purple dots in Fig. 5 represent the results obtained by Shi et al. (2009) through surface observation, and they show spans that are all less than 100 m. The blue dots in Fig. 5 are the results we obtained from GPS data, and we find that spans across the rupture are all over 10 km. The black and red dots in Fig. 5 represent our results obtained by the methods described in Sections 3.1 and 3.2. Our measured shortenings are in general larger than those derived simply from GPS data, which is consistent with the theoretic model of a coseismic deformation field. This model shows that the greater the distance from the rupture is, the smaller the deformation
Fig. 5. (a) Distribution of the values of coseismic horizontal shortenings for different span lengths across the rupture, and using different methods; purple dots are horizontal shortenings for spans of less than 100 m across the rupture, using direct measurements in the field (Shi et al., 2009), and the average is shown by the purple dashed line; black dots and red dots represent our results for spans of 100– 1000 m across the rupture, calculated using the methods described in Sections 3.1 and 3.2, respectively, and their averages are shown by black and red dashed lines, respectively; blue dots represent the horizontal shortenings for spans greater than 10 km, using GPS data (The Project of Crustal Movement Observation Network of China, 2008), and the average values are shown by blue dashed line. (b) The theoretic model of coseismic displacement (modified from Mavko, 1981). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
is (Fig. 5b). The coseismic deformation is produced by relieving the strain of dynamic faulting, and the model has been verified by the successful application of elastic dislocation models (Chinnery, 1961; Savage and Hastie, 1966). The observed coseismic deformation fields produced by the large Tango earthquake (strike-slip) in Japan, 1927 (Mavko, 1981), and the Nankaido earthquake (thrusting) in Japan, 1946 (Fitch and Scholz, 1971), are also consistent with the theoretic model. However, our results for larger spans are, on average, larger than those for smaller spans, where the measurements were made directly in the field immediately around the rupture, and this seems quite different from the theoretic prediction. A similar phenomenon was observed for some other strong earthquakes, such as the Chi–Chi earthquake of 1999 (Lee et al., 2010), the Landers earthquake of 1992 (Pumbroeck et al., 2000), and the Kunlun earthquake of 2001 (Klinger et al., 2006). All their results indicate that the place of maximum deformation is not immediately within the surface rupture, but distributed over a wider span of hundreds of meters to 1 km across the fault. After the Chi–Chi earthquake of 1999, Lee et al. (2010) found a deformation zone of about 200 m located on the hanging wall next to the fault. Our results imply that there is a similar deformation zone associated with the Wenchuan Earthquake. There could be other possible reasons for our measured amounts of shortening to be larger than amounts measured directly in the field, and these include variations in the dip of the fault with depth (Shi et al., 2009). A comparison of our results with the direct field measurements suggests that the coseismic shortening deformation was actually distributed over a wider area, including some continuous and discontinuous unobservable parts. The fact that the amounts of coseismic horizontal shortening measured by us are larger than either the coseismic vertical offsets observed in the field (He et al., 2008; Liu-Zeng et al., 2010; Xu et al., 2009) or the geophysical inversions (Shen et al., 2009), implies that the rupture is a low-angle (<45°) fault at depth, even though it dips (>45°) steeply at the surface. Based on the seismic reflection data and well logs (Jia et al., 2003; Luo, 1998; Song, 1994), as well as the observed coseismic deformation data, Xu et al. (2009) proposed a low-angle fault model, with a predicted total crustal shortening of 8.5 m, and a total vertical uplift of 7.5 m. Our estimates of the average horizontal shortening support these predictions. With regard to the building of the steep Longmenshan range, there have been three well-known hypotheses proposed: channel flow (Burchfiel et al., 2008; Clark and Royden, 2000; Royden et al., 2008), ‘‘classical’’ tectonic convergence (Hubbard and Shaw, 2009; Liu-Zeng et al., 2009; Xu et al., 2009), and tectonic deformation plus isostatic rebound caused by long-term rapid erosion (Fu et al., 2011). Surface deformation, especially when only caused by an earthquake, cannot be directly used to determine the structure in the lower crust. But the surface deformation may provide useful pointers to the nature of the structure in the lower crust, or provide useful information. For the Wenchuan Earthquake, the coseismic landsliding produced a greater volume of erodible material than the volume added to the orogen by coseismic rock uplift (Ouimet, 2011; Parker et al., 2011). This discrepancy indicates that it is not just rock uplift during earthquakes that is responsible for the high topography of the Longmengshan range, and a more general shortening of the crust may provide the explanation. In this context, our estimates of horizontal shortenings provide essential information for understanding the seismotectonics of eastern Tibet, especially the building of the Longmenshan range.
5. Conclusions We have obtained a series of horizontal shortening measurements by contrasting the spacings between two roads observed
F. Shi et al. / Journal of Asian Earth Sciences 50 (2012) 164–170
on QuickBird satellite images made before and after the Wenchuan Earthquake. The maximum, minimum, and average values are 11.0 m, 1.1 m, and 6.6 m, respectively. The error is estimated to be 1.84 m. By comparing the spacings between the two roads using pre-earthquake images and GPS field measurements after the earthquake, we obtained a second series of horizontal shortening measurements. In this case, the maximum value is 9.85 m, the minimum 3.8 m, and the average 7.1 m. A comparison of our coseismic horizontal shortening measurements with those derived simply from GPS data, shows our results to be much larger on average. This is consistent with the theoretic model for a coseismic deformation field. On the other hand, our results for large spans across the fault are, on average, larger than those obtained from direct measurements in the field, immediately adjacent to the fault for smaller spans, and this differs from the theoretic prediction, suggesting that the coseismic shortening was actually distributed over a wide area, including some unobservable continuous or discontinuous parts. Finally, our estimates of the amount of coseismic horizontal shortening support the low-angle fault model for the Wenchuan Earthquake (Xu et al., 2009). Furthermore, our results help us reach a better understanding of the seismotectonics of eastern Tibet, especially the mountain building of the Longmenshan range. Acknowledgements This research was supported by the National Natural Science Foundation of China (Grant Numbers 40841019 and 40872128) and special funds from the Science and Technology Ministry of China for the Scientific Investigation of the Wenchuan Earthquake Rupture. We thank Dr. Zongqi Yue for his very helpful review that improved the manuscript. Many thanks are also due to anonymous reviewers for their constructive comments, and to the journal editors for their corrections. References Burchfiel, B.C., Royden, L.H., Van der Hilst, R.D., Hager, B.H., Chen, Z., King, R.W., Li, C., Lü, J., Yao, H., Kirby, E., 2008. A geological and geophysical context for the Wenchuan earthquake of 12 May 2008, Sichuan, People’s Republic of China. GSA Today 18, 4–11. Chen, G.H., Xu, X.W., Yu, G.H., An, Y.F., Yuan, R.M., Guo, T.T., Gao, X., Tan, X.B., 2009. Co-seismic slip and slip partitioning of multi-faults during the Ms8.0 2008 Wenchuan earthquake. Chinese Journal of Geophysics 52 (5), 1384–1394 (in Chinese with English abstract). Chen, G.H., Xu, X.W., Zheng, R.Z., Yu, G.H., Li, F., Li, C.X., Wen, X.Z., He, Y.L., Ye, Y.Q., Chen, X.Ch., Wang, Z.C., 2008. Quantitative analysis of the co-seismic surface rupture of the 2008 Wenchuan earthquake, Sichuan, China along the Beichuan– Yingxiu Fault. Seismology and Geology 30 (3), 723–738 (in Chinese with English abstract). Chinnery, M.A., 1961. Deformation of the ground around surface faults. Bulletin of the Seismological Society of America 51, 355–372. Clark, M., Royden, L.H., 2000. Topographic ooze: building the eastern margin of Tibet by lower crustal flow. Geology 28, 703–706. Deng, Q.D., 2007. Map of Active Tectonics in China. Seismological Press, Beijing. DigitalGloble, Inc., 2005. QuickBird Imagery Products – FAQ. DigitalGloble, Inc., 2006. QuickBird Imagery Products – Product Guide. Dominguez, S., Avouac, J.P., Michel, R., 2003. Horizontal coseismic deformation of the 1999 Chi–Chi earthquake measured from SPOT satellite images: implications for the seismic cycle along the western foothills of central Taiwan. Journal of Geophysical Research 108 (B2), 2083. doi:10.1029/ 2001JB000951. Fitch, T.J., Scholz, C.H., 1971. Mechanism of underthrusting in southwest Japan: a model of convergent plate interactions. Journal of Geophysical Research 76, 7260–7292. Fu, B.H., Shi, P.L., Guo, H.D., Okuyama, S., Ninomiya, Y., Wright, Y., 2011. Surface deformation related to the 2008 Wenchuan earthquake, and mountain building of the Longmen Shan, eastern Tibetan Plateau. Journal of Asian Earth Science 40, 805–824. Hao, K.X., Si, H., Fujiwara, H., Ozawa, T., 2009. Coseismic surface-ruptures and crustal deformations of the 2008 Wenchuan earthquake Mw7.9, China. Geophysical Research Letters 36, L11303. doi:10.1029/2009GL03791. Hashimoto, M., Enomoto, M., Fukushima, Y., 2010. Coseismic deformation from the 2008 Wenchuan, China, earthquake derived from ALOS/PALSAR Images. Tectonophysics 491, 59–71.
169
He, H.L., Sun, Z.M., Wei, Z.Y., Dong, Sh.P., Gao, X., Wang, Sh.Y., Wang, J.Q., 2008. Rupture of the MS8.0 Wenchuan earthquake along Baishahe River. Seismology and Geology 30 (3), 658–673 (in Chinese with English abstract). He, H.L., Wei, Zh.Y., Chen, Ch.Y., Shi, F., 2011. On-site determination of slip vectors along the Longmenshan fault during the Mw 7.9 Wenchuan, China, earthquake of May 12, 2008. Journal of Asian Earth Sciences 41, 274–282. Hubbard, J., Shaw, J.H., 2009. Uplift of the Longmen Shan and Tibetan Plateau, and the 2008 Wenchuan (M = 7.9) earthquake. Nature 458, 194–197. Jia, D., Chen, Z., Jia, C., Wei, G., Li, B., Zhang, Q., Wei, D.T., Shen, Y., 2003. Structural features of the Longmen Shan fold and thrust belt and development of the Western Sichuan foreland basin, central China. Geological Journal of China Universities 9 (3), 402–409 (in Chinese with English abstract). Klinger, Y., Michel, R., King, G.C.P., 2006. Evidence for an earthquake barrier model from Mw 7.8 Kokoxili (Tibet) earthquake slip-distribution. Earth and Planetary Science Letters 242, 354–364. Lee, Y.H., Wu, K.C., Rau, R.J., Chen, H.C., Lo, W., Cheng, K.C., 2010. Revealing cosesimic displacements and the deformation zones of the 1999 Chi–Chi Earthquake in the Tsaotung Area, Central Taiwan, using digital cadastral data. Journal of Geophysical Research 115 (B03419). doi:10.1029/2009JB006397. Li, C.Y., Wei, Z.Y., Ye, J.Q., Han, Y.B., Zheng, W.J., 2010. Amounts and styles of coseismic deformation along the northern segment of surface rupture, of the 2008 Wenchuan Mw 7.9 earthquake, China. Tectonophysics 491, 35–58. Li, T., Chen, J., Huang, M.D., Yu, S., 2009. Methods to calculating slip vectors of reverse-fault earthquake surface rupture: a case study in Wenchuan earthquake. Quaternary Sciences 29 (3), 524–534 (in Chinese with English abstract). Liu-Zeng, J., Wen, L., Sun, J., Zhang, Z.H., Hu, G.Y., Xing, X.C., Zeng, L., Xu, Q., 2010. Surficial slip and rupture geometry on the Beichuan Fault near Hongkou during the Mw 7.9 Wenchuan earthquake, China. Bulletin of the Seismological Society of America 100 (5B), 2615–2650. Liu-Zeng, J., Zhang, Z., Wen, L., Tapponnier, P., Sun, J., Xing, X., Hu, G., Xu, Q., Zeng, L., Ding, L., Ji, C., Hudnut, K.W., Van der Woerd, J., 2009. Co-seismic ruptures of the 12 May 2008. Ms8.0 Wenchuan earthquake, Sichuan: east–west crustal shortening on oblique, parallel thrusts along the eastern edge of Tibet. Earth and Planetary Science Letters 286 (3–4), 355–370. Luo, Z.L., 1998. New recognition of basement in Sichuan Basin. Journal of Chengdu University of Technology 25 (2), 191–200 (in Chinese with English abstract). Mavko, G.M., 1981. Mechanics of motion on major faults. Annual Review of Earth and Planetary Sciences 9, 81–111. Michel, R., Avouac, J.P., 2002. Deformation due to the 17 August 1999 Izmit, Turkey, earthquake measured from SPOT images. Journal of Geophysical Research 107 (B4), 2062. doi:10.1029/2001JB000102. Nakamura, T., Tsuboi, S., Kaneda, Y., Yamanaka, Y., 2010. Rupture process of the 2008 Wenchuan China earthquake inferred from teleseismic waveform inversion and forward modeling of broadband seismic waves. Tectonophysics 491, 72–84. Ouimet, W.B., 2011. The hills came tumbling down. Nature Geoscience 4, 424–425. Parker, R.N., Densmore, A.L., Rosser, N.J., Michele, M., Li, Y., Huang, R.Q., Whadcoat, S., Petley, D.N., 2011. Mass wasting triggered by the 2008 Wenchuan earthquake is greater than organic growth. Nature Geoscience 4, 449–452. Pumbroeck, N., Michel, R., Bient, R., Avouac, J.P., Taboury, J., 2000. Measuring earthquakes from optical satellite images. Applied Optics 39 (20), 3486–3494. Ran, Y.K., Shi, X., Wang, H., Chen, L.Ch., Chen, J., Liu, R.Ch., Gong, H.L., 2010. The maximum coseismic vertical surface displacement and surface deformation pattern accompanying the Ms 8.0 Wenchuan earthquake. Chinese Science Bulletin 55 (9), 841–850. Royden, L.H., Burchfiel, B.C., Van der Hilst, R.D., 2008. The geological evolution of the Tibetan Plateau. Science 321, 1054–1058. Savage, J.C., Hastie, L.M., 1966. Surface deformation associated with dip-slip faulting. Journal of Geophysical Research 71, 4897–4904. Shen, Z.-K., Sun, J., Zhang, P., Wan, Y., Wang, M., Bürgmann, R., Zeng, Y., Gan, W., Liao, H., Wang, Q., 2009. Slip maxima at fault junctions and rupturing of barriers during the 2008 Wenchuan earthquake. Nature Geoscience 2, 718–724. Shi, F., He, H.L., Wei, Zh.Y., 2009. Co-seismic horizontal shortening on Baishahe segment of Wenchuan earthquake rupture zone. Quaternary Sciences 29 (3), 546–553 (in Chinese with English abstract). Song, H.B., 1994. The comprehensive interpretation of geological and geophysical data in the organic belt of Longmen Mountains, China. Journal of Chengdu Institute of Technology 21 (2), 79–88 (in Chinese with English abstract). The Project of Crustal Movement Observation Network of China, 2008. The coseismic displacement fields of Wenchuan Ms8.0 earthquake occurrence in 2008 using GPS data. Science in China (Series D – Earth Sciences) 38(10), 1195– 1206 (in Chinese). Wang, H., Ran, Y.K., Chen, L.Ch., Shi, X., Liu, R.Ch., Gomez, F., 2010. Determination of horizontal shortening and amount of reverse-faulting from trenching across the surface rupture of the 2008 MW7.9 Wenchuan earthquake, China. Tectonophysics 491, 10–20. Wang, W.M., Zhao, L.F., Li, J., Yao, Z.X., 2008. Rupture process of the Ms 8.0 Wenchuan earthquake of Sichuan, China. Chinese Journal of Geophysics 51 (5), 1403–1410 (in Chinese with English abstract). Wang, Z., Fukao, Y., Pei, Sh.P., 2009. Structural control of rupturing of the Mw7.9 2008 Wenchuan earthquake, China. Earth and Planetary Science Letters 279, 131–138. Xu, X.W., Chen, G.H., Yu, G.H., Sun, X.Zh., Tan, X.B., Chen, L.C., Sun, J.B., Chen, Y.G., Chen, W.Sh., Zhang, Sh.P., Li, K., 2010. Reevaluation of surface rupture parameters of the 5.12 Wenchuan earthquake and its tectonic implication for
170
F. Shi et al. / Journal of Asian Earth Sciences 50 (2012) 164–170
Tibetan uplift. Chinese Journal of Geophysics 53 (10), 2321–2336 (in Chinese with English abstract). Xu, X.W., Wen, X.Z., Yu, G.H., Chen, G.H., Klinger, Y., Hubbard, J., Shaw, J., 2009. Coseismic reverse- and oblique-slip surface faulting generated by the 2008 Mw7.9 Wenchuan earthquake, China. Geology 37 (6), 515–518. Xu, X.W., Yu, G.H., Ma, W.T., Ran, Y.K., Chen, G.H., Han, Zh.J., Zhang, L.F., You, H.Ch., 2002. Evidence and methods for determining the safety distance from the potential earthquake surface rupture on active fault. Seismology and Geology 24 (4), 470–483 (in Chinese with English abstract).
Yuan, R.M., 2008. Study on the influence of the movement of earthquake fault on the overlaying soil stratum based on the centrifuge modeling test & change law of geofluid pressure during basin evolution. Institute of Geology, China Earthquake Administration, Beijing. Zhou, Q., Xu, X.W., Yu, G.H., Chen, X.Ch., He, H.L., Yin, G.M., 2010. Width distribution of the surface ruptures associated with the Wenchuan earthquake: implication for the setback zone of the seismogenic faults in postquake reconstruction. Bulletin of the Seismological Society of America 100 (5B), 2660–2668.