2. Key Laboratory of Marine Disaster Prediction Technology, National Marine Environmental Forecasting Center, State Oceanic Administration, Beijing 100081, China;
3. Key Laboratory of Computational Geodynamics, University of Chinese Academy of Sciences, CAS, Beijing 100049, China
According to the official determination of China Earthquake Networks Center, an MS6.7 earthquake occurred in Akto County, Kizilsu Kirghiz Autonomous Prefecture, Xinjiang(39.27°N, 74.04°E) at 22:24 Beijing Time on November 25, 2016, with a focal depth of 10km, and about 5 minutes before the main shock, an MS4.8 earthquake occurred in the epicentral area. The MS6.7 earthquake was followed by a major aftershock of MS5.0. A total of 732 earthquakes have been recorded as of 10:00 a.m. on November 30, 2016, including 1 earthquake with M5.0, 3 earthquakes with M4.0-4.9 and 46 earthquakes with M3.0-3.9 (http://www.cea.gov.cn/publish/dizhenj/464/515/20161126105536683371966/index.html).
The Akto earthquake region is located at the northwest edge of the collision between the Indian Plate and the Eurasian Plate, which is one of the most seismically active areas and is also one of the few intermediate-depth earthquake source regions in all the continents of the world (Fig. 1). Tectonic and seismic activities are intensive in this region, and many deep and large faults are developed, which exert certain control over the inland orogenic movements(Robinson A.C. et al., 2007) and also become the dominant of regional seismic activities (Qiao Xuejun et al., 2014). There have been many shallow strong earthquakes in the history of this region, including the 1944 south Wuqia M7.0 earthquake in Xinjiang, the 1974 Kashgar M7.3 earthquake, the 2005 Pakistan M7.8 earthquake, the 2008 Wuqia M6.8 earthquake and the 2015 Tajikistan M7.4 earthquake. These earthquakes are mainly distributed in the top of the northern Pamir arc and some areas of the interior of plateau, and the distribution pattern is consistent with the regional active structure (Su Jinrong et al., 2013). Since the late Cenozoic Era, the deformation of Pamir has been characterized by forward thrusting in its frontal area, oblique-thrusting-strike-slip on its sides and extension in its interior (Li Wenqiao, 2013). The distribution of focal mechanism in this area shows mainly strike-slip and thrusting movement. The epicenter of this earthquake is located inside the Pamirs Plateau, and reverse faults in the Pamirs are not very active, indicating that the shortening and thickening of the crust inside the tectonic syntaxis has stopped, and the deformation is mainly controlled by the EW-stretching Kongur normal fault in the east and the EW-stretching kapakon graben in the west (Chen Jie et al., 2011).
According to the result given by Institute of Geology, China Earthquake Administration, the Muji fault is the seismogenic fault of the Akto MS6.7 earthquake (http://www.eq-igl.ac.cn/upload/images/2016/11/28153257503.jpg). The Muji fault, about 80km long, NWW-striking, is a Holocene right-lateral strike-slip fault located at the northern end of the Kongur extensional system in the Pamirs Plateau, which is nearly vertical with both thrusting and normal faulting movement (Li Tao, 2012). This earthquake is one of the largest earthquakes occurring in the Pamirs Plateau in recent years, and it is of great significance to study the seismogenic mechanism of this earthquake to understand the interaction between the Indian Plate and the Eurasina Plate and the characteristics of seismic activities in this region.
Previous research results have explained that the spatial distribution characteristics of aftershock sequences and focal mechanism solutions are helpful to understand the seismogenic faults and extension of fractures (Zhang Guangwei et al., 2014, 2016; Liang Shanshan et al., 2017, 2018). Therefore, in order to further explore the seismogenic mechanism of this earthquake, waveform data and seismic phase data from Xinjiang regional seismic network and waveform data from IRIS(Incorporated Research Institutions for Seismology) shared global seismic stations are utilized to obtain the characteristics of the Akto MS6.7 earthquake sequence by using the HypoDD relocation method(Waldhauser F. and Ellsworth W. L., 2000), and focal mechanism of the MS4.8 foreshock, the MS6.7 main shock and the MS5.0 aftershock are obtained by using gCAP method(Zhu Lupei and Ben-Zion Y., 2013), and thus to discuss the seismogenic pattern of this earthquake and its regional tectonic significance.1 DATA AND METHODS
Seismic phase data and initial source parameters used in this study are from the seismic observation report of Earthquake Agency of Xinjiang Uygur Autonomous Region. We collect the Akto MS6.7 earthquake sequence from November 25, 2016 to November 30, 2016 to obtain a total of 174 M≥2.0 earthquakes, and relocation of the earthquake sequence is carried out. In order to better understand the seismogenic mechanism of this earthquake, waveform data of the Akto MS6.7 main shock, the MS4.8 foreshock and the MS5.0 aftershock are collected in the meantime to obtain focal mechanism solution. As the Akto MS6.7 earthquake sequence occurred in the border areas of Xinjiang (Fig. 2(a)), the distribution of permanent stations in the study area is not satisfactory. For this reason, we also collect waveform data from shared global seismic stations provided by IRIS, which fill in the blank areas of regional fixed stations used alone and improved the azimuth coverage of stations in the study area.
In this study, the double difference location method (HypoDD; Waldhauser F. and Ellsworth W. L., 2000) is used to relocate earthquakes. In this method, relative travel time residuals are used to determine earthquake location. It is mainly based on the fact that the distance between two earthquake focuses is much smaller than the distance between the event and the station, and it is believed that the ray paths of the two events propagating to the station are almost the same, so the influence of the inaccuracy of the velocity model on the positioning results can be reduced. The HypoDD method has been widely used in relocation work (Zhao Bo et al., 2013). Data of seismic phase arrival times incorporate the data from observation reports of fixed stations and data picked up by IRIS shared stations. Arrival times of events which are recorded by at least four stations within 400km from the epicenters are selected. In total, there are 1, 919 pieces of P-wave arrival time data and 897 pieces of S-wave arrival time data. In the positioning process, P- and S-wave weights are set as 1.0 and 0.5 respectively, and the maximum distance of an event pair is set as 10km.
The gCAP method (Zhu Lupei and Ben-Zion Y., 2013) is used to calculate focal mechanism solution. Pnl- and S-wave (or surface wave) are assigned with different weights by this method, and the fitted error function between theoretical and real waveforms is obtained, and the optimal solution with minimum error can be obtained by grid search. Its major advantage is that the weight of Pnl wave is increased, which has a better control over depth; Moreover, by allowing relative time-shift fitting between theoretical waveform and real waveform during each time window, the defects of velocity model and errors of earthquake location can be effectively eliminated. In the inversion of focal mechanism, because of different magnitudes, three earthquakes are filtered in two frequency bands: the filter ranges of Pnl and surface waves of the Akto MS6.7 earthquake are 0.02-0.1Hz and 0.02-0.6Hz, and the filter ranges of Pnl and surface waves of the MS4.8 foreshock and the MS5.0 aftershock are 0.05-0.2Hz and 0.03-0.1Hz, respectively. When calculating the fault plane solution parameters, we take 5° as the search interval of earthquake strike, dip angle and rake angle, and 1km as the search interval of depth. Frequency-wave number method(F-K) is adopted to perform Green's function calculation (Zhu Lupei and Rivera L. A., 2002), which is widely used at present. This method integrates the frequency and wave number respectively, uses propagation matrix to calculate the displacement of the full wave field of the earthquake, and can calculated waveforms of body wave and surface wave at various frequencies, which can be used to invert source parameters. Taking 0.1s as sampling interval, there are altogether 1024 sampling points.
A one-dimensional velocity model for earthquake relocation and focal mechanism inversion is obtained by integrating artificial earthquake sounding and receiver function (Fig. 2(b), Li Qiusheng et al., 2001; Gao Rui et al., 2011; Tang Mingshuai et al., 2014).2 RESULTS AND ANALYSIS 2.1 Relocation Results of the Akto MS6.7 Earthquake Sequence
In this study, the HypoDD relative positioning method is firstly adopted to relocate the Akto MS6.7 earthquake sequence, and relocation results of 138 earthquake events are finally obtained. In order to know the positioning effect before and after relocating, we compare the results before and after positioning (Fig. 3). Before relocating, the earthquake sequence is scattered on the plane and shows distinct stripe distribution in depth, so it is difficult to distinguish the major seismogenic fault plane (Fig. 3(a)). After relocating, the earthquake sequence is distributed more obviously along the SEE direction, and focal depths show a prominent seismic layer of 5-20km, which is in line with the characteristics of frequent shallow earthquakes in this region. Moreover, the focal depth of the Akto MS6.7 main shock is about 14.44km after relocating, which is close to the initial positioning depth of 10km. The focal depth of Akto MS5.0 aftershock is improved from the initial 28km to a more reasonable 11.59km (Fig. 3(b)). All these indicate that the positioning results are better after repositioning.
In order to better display the extension of ruptures of fault planes, in this study, the distribution characteristics of earthquakes are shown along the main spreading direction of aftershocks and the perpendicular direction within 24 hours (Fig. 4) and 160 days(Fig. 5) respectively after the earthquake. As can be seen from the vertical section AA' 24 hours after the earthquake (Fig. 4(a)), the aftershock sequence is mainly distributed in SEE direction starting from the epicenter on the time scale, and the earthquake sequence extends for about 50km. From transverse sections BB', CC' and DD' that cross through the MS4.8, MS6.7 and MS5.0 earthquakes, we can see that the focal depths are almost distributed in vertical direction, thus it can be inferred that the MS4.8 foreshock, the MS6.7 main shock and the MS5.0 aftershock all have steep seismogenic fault planes, which may belong to the same seismic fault.
In order to further explore the rupture range and major seismogenic layer of this earthquake, we provide profiles of distribution of aftershock sequence within 160 days after the earthquake(Fig. 5). The results show that earthquakes extends about 70km along the fault direction AA', which is different from the extension range(about 50km) 24 hours after the earthquake. The spreading of aftershocks is consistent with that 24 hours after the earthquake, indicating that the aftershock activity pattern in a relatively short time after the main shock can basically represent the pattern of fault activities, which can provide important technical support for earthquake relief. This is consistent with the analysis results of Lushan earthquake sequence in Sichuan(Zhang Guangwei and Lei Jianshe, 2013). The major seismogenic layer of the aftershock sequence is at a depth of 5-20km.2.2 Focal Mechanisms of the MS4.8 Foreshock, the Akto MS6.7 Main Shock and the MS5.0 Aftershock
The optimal centroid depth of the main shock is 21km, which is obtained by using gCAP method and inversion of focal mechanisms at different depths. Among them, nodal plane Ⅰ strikes to the direction of 192°, with a dip angle of 82° and a rake angle of -17°, and nodal plane Ⅱ strikes to the direction of 284°, with a dip angle 74° and a rake angle of -171°. Focal mechanisms are consistent at different depths, which indicates the reliability of the inversion results in this study(as shown in Fig. 6(a)). Fig. 6(b) shows the focal mechanism solution of the main shock at the depth of 21km(lower hemisphere projection), and gives the fitting lines of theoretical and real waveforms. The fitting results of earthquake waveforms are good and the source parameters obtained are reliable. We compared the inversion results of this study with focal mechanisms given by different studies(Table 1), in order to evaluate the reliability of the results of this study. According to Table 1, values of dip angles and rake angles of nodal plane Ⅰ and Ⅱ are close in different studies, all show high dip angles, indicating that this earthquake is a high-angle strike-slip event. But there are some deviations in the strikes of nodal plane Ⅰ and Ⅱ given by different researches. Nodal plane Ⅰ and Ⅱ strike to the direction of 192° and 284° respectively according to this study, which differ from the results provided by global moment tensor solutions and China Earthquake Networks Center by 180°, and are close to the strikes of nodal planes given by the Institute of Geophysics, China Earthquake Administration. The main reason for such difference is that when dip angle of nodal plane is close to 90°, the striking direction is easy to be overturned and a deviation of 180° occurs in the numerical value. There is actually a good consistency between the strikes of nodal planes in this study and GCMT. Fig. 7, Fig. 8 and Table 1 give the optimal centroid depths of the MS4.8 and the MS5.0 earthquakes and their fitted nodal plane solutions.
As can be seen from the distribution characteristics of the earthquake sequence(Fig. 9), aftershocks mainly spread along the direction of nearly SE. The trend of the aftershock sequence is more consistent with the strikes of nodal plane Ⅱ according to focal mechanisms of the three earthquakes we obtained, which all show a strike-slip earthquake with a steep dip angle. It is consistent with the results of field geological investigation of the Muji fault(Chen Jie et al., 2016), which proves the reliability of the results of this study.
In this study, IRIS shared waveform data is firstly used to obtain arrival times of seismic phases from shared seismic stations by manually picking up seismic phases, and combined with seismic phase report data from the Earthquake Agency of Xinjiang Uygur Autonomous Region, HypoDD method is adopted to obtain the repositioning results of the Akto MS6.7 earthquake. Then combined with waveform data from the Xinjiang regional stations and shared waveform data, focal mechanism solutions for the Akto MS4.8 foreshock, the MS6.7 main shock and the MS5.0 aftershock are calculated by using gCAP method. The relocating results of this study show that the initial rupture length of the main shock is about 14km, the aftershocks are mainly distributed along the SEE direction, about 70km in length, and earthquake focuses are mostly distributed at the depth of 5-20km. The depth profiles show that the seismogenic fault planes of the three earthquakes are all steep. The focal mechanism results of gCAP show that the three earthquakes are all strike-slip events, and combined with regional structure, the three earthquakes are all right-lateral strike-slip events. Based on the results of earthquake relocating, focal mechanism and field geological survey, it is inferred that the Muji right-lateral strike-slip fault might be the seismogenic fault of the Akto MS6.7 earthquake.
The Akto MS6.7 earthquake occurred at the top of the Pamir arc, which belongs to the north end of the Kongur extensional system inside the Pamirs Plateau. Present deformation in this region shows mainly EW extensional movement, and the extension rate gradually decreases from north to south. Activities and deformation of the late Quaternary and present day are also concentrated here. The Kongur extensional system, NW-SE-striking, W-dipping, is about 250km long, with a dip angle of 20°-45°(Chen Jie et al., 2011). The cause of this earthquake is that the Indian Plate continues to push northward, and during this compression process, the Indian Plate is affected by the clockwise rotation of the Tarim basin, which causes the accumulation of right-lateral action of the Muji fault, resulting in this earthquake. Earthquakes inside the Kongur extensional system are generally moderate-strong earthquakes(Fig. 1), seismic activities in this region is far less active than that in the front edge of the Pamir syntax, and this earthquake is one of the biggest in recent years, which may indicate that seismic activities have begun to increase in this region, especially in the north with higher extension rate, where there will be a higher seismic risk.
Chen Jie, Li Tao, Li Wenqiao, Yuan ZhaodeChen Jie, Li Tao, Li Wenqiao, Yuan Zhaode. Late Cenozoic and present tectonic deformation in the Pamir salient, northwestern China[J]. Seismology and Geology, 2011, 33(2): 241-259 (in Chinese with English abstract).
Chen Jie, Li Tao, Sun Jianbao, Fang Lihua, Yao Yuan, Li Yuehua, Wang Haoran, Fu BoChen Jie, Li Tao, Sun Jianbao, Fang Lihua, Yao Yuan, Li Yuehua, Wang Haoran, Fu Bo. Coseismic surface ruptures and seismogenic Muji fault of the 25 November 2016 Arketao MW6.6 earthquake in northern Pamir[J]. Seismology and Geology, 2016, 38(4): 1160-1174 (in Chinese with English abstract).
Deng Qidong, Zhang Peizhen, Ran Yongkang, Yang Xiaoping, Min Wei, Chu QuanzhiDeng Qidong, Zhang Peizhen, Ran Yongkang, Yang Xiaoping, Min Wei, Chu Quanzhi. Basic characteristics of active tectonics of China[J]. Science in China (Series D), 2003, 46(4): 356-372.
Gao Rui, Guan Ye, He Rizheng, Shi Danian, Li Dexing, Li Qiusheng, Lu Deyuan, Li LijunGao Rui, Guan Ye, He Rizheng, Shi Danian, Li Dexing, Li Qiusheng, Lu Deyuan, Li Lijun. The integrated geophysical observation and research along the Xinjiang (XUAR) geotransect and its surrounding areas[J]. Acta Geoscientia Sinica, 2011, 22(6): 527-533 (in Chinese with English abstract).
Li Qiusheng, Lu Deyuan, Gao Rui, Zhang Zhiying, Liu Wen, Li Yingkang, Li Jingwei, Fan Jingyi, Xiong XianmingLi Qiusheng, Lu Deyuan, Gao Rui, Zhang Zhiying, Liu Wen, Li Yingkang, Li Jingwei, Fan Jingyi, Xiong Xianming. An integrated study of deep seismic sounding profiling along Xinjiang global geosciences transect (Quanshuigou-Dushanzi)[J]. Acta Geoscientia Sinica, 2001, 22(6): 534-540 (in Chinese with English abstract).
Li Tao. Active Thrusting and Folding along the Pamir Frontal Thrust System [D]. Beijing: Institute of Geology, China Earthquake Administrator, 2012 (in Chinese with English abstract).
Li Wenqiao. Active Tectonics and Strong Earthquakes in Tashkurgan Valley, Northeast Pamir Plateau[D] Beijing: Institute of Geology, China Earthquake Administrator, 2013 (in Chinese with English abstract).
Liang Shanshan, Lei Jianshe, Xu Zhiguo, Zou Liye, Liu JingguangLiang Shanshan, Lei Jianshe, Xu Zhiguo, Zou Liye, Liu Jingguang. Relocation of the aftershock sequence and focal mechanism solutions of the 21 January 2016 Menyuan, Qinghai, MS6.4 earthquake[J]. Chinese Journal of Geophysics, 2017, 60(6): 2091-2103 (in Chinese with English abstract).
Liang Shanshan, Lei Jianshe, Xu Zhiguo, Xu Xiwei, Zou Liye, Liu Jingguang, Chen HongfengLiang Shanshan, Lei Jianshe, Xu Zhiguo, Xu Xiwei, Zou Liye, Liu Jingguang, Chen Hongfeng. Relocation of aftershocks of the 2017 Jiuzhaigou, Sichuan, MS7.0 earthquake and inversion for focal mechanism of the mainshock[J]. Chinese Journal of Geophysics, 2018, 61(5): 2163-2175 (in Chinese with English abstract).
Qiao Xuejun, Wang Qi, Yang Shaomin, Li JieQiao Xuejun, Wang Qi, Yang Shaomin, Li Jie. Study on the focal mechanism and deformation characteristics for the 2008 MS6.7 Wuqia earthquake, Xinjiang by InSAR[J]. Chinese Journal of Geophysics, 2014, 57(6): 1805-1813 (in Chinese with English abstract).
Robinson A.C., Yin An, Manning C.E., Harrison T.M., Zhang Shuanhong, Wang XiaofengRobinson A.C., Yin An, Manning C.E., Harrison T.M., Zhang Shuanhong, Wang Xiaofeng. Cenozoic evolution of the eastern Pamir: Implications for strain-accommodation mechanisms at the western end of the Himalayan-Tibetan orogen[J]. GSA Bulletin, 2007, 119(7-8): 882-896. DOI:10.1130/B25981.1
Su Jinrong, Guo ZhiSu Jinrong, Guo Zhi. A study of focal mechanisms of the MS6.7 earthquake of 5 October 2008 in southwestern Tianshan-Pamier[J]. Chinese Journal of Geophysics, 2013, 56(2): 504-512 (in Chinese with English abstract).
Tang Lanlan, Zhao Cuiping, Wang HaitaoTang Lanlan, Zhao Cuiping, Wang Haitao. Study on the source characteristics of the 2008 MS6.8 Wuqia, Xinjiang earthquake sequence and the stress field on the northeastern boundary of Pamir[J]. Chinese Journal of Geophysics, 2012, 55(4): 1228-1239 (in Chinese with English abstract).
Tang Mingshuai, Zheng Yong, Ge Can, Wang Haitao, Ji Zhanbo, Kong Xiangyan, Wei Bin, Li ZhihaiTang Mingshuai, Zheng Yong, Ge Can, Wang Haitao, Ji Zhanbo, Kong Xiangyan, Wei Bin, Li Zhihai. Study on crustal structure in the northeastern Pamir region by P receiver functions[J]. Chinese Journal of Geophysics, 2014, 57(10): 3176-3188 (in Chinese with English abstract).
Waldhauser F., Ellsworth W.L.Waldhauser F., Ellsworth W.L. A double-difference earthquake location algorithm: method and application to the Northern Hayward Fault, California[J]. Bulletin of the Seismological Society of America, 2000, 90(6): 1353-1368. DOI:10.1785/0120000006
Wang Xiaoqiang, Lu Xing, Liu Bin, Li Jie, Liu DaiqinWang Xiaoqiang, Lu Xing, Liu Bin, Li Jie, Liu Daiqin. Study on present-day vertical crust movement and seismic activity in Xinjiang area[J]. Journal of Geodesy and Geodynamics, 2009, 29(6): 22-27, 31 (in Chinese with English abstract).
Yang Jilin, Song Lijun, Tuniyazi·Shawuti, Li Chengri, Suo Xianglin, Zhao Baozong, AIerken·HasimuYang Jilin, Song Lijun, Tuniyazi·Shawuti, Li Chengri, Suo Xianglin, Zhao Baozong, AIerken·Hasimu. Exploration of the macroscopic intensity and discussion on the seismogenic structrue for the Akto M5.9 earthquake on Sep. 2, 2003[J]. Inland Earthquake, 2004, 18(4): 353-358 (in Chinese with English abstract).
Zhang Guangwei, Lei JiansheZhang Guangwei, Lei Jianshe. Relocations of Lushan, Sichuan strong earthquake (MS7.0) and its aftershocks[J]. Chinese Journal of Geophysics, 2013, 56(5): 1764-1771 (in Chinese with English abstract).
Zhang Guangwei, Lei Jianshe, Sun ChangqingZhang Guangwei, Lei Jianshe, Sun Changqing. Relocation of the 12 February 2014 Yutian, Xinjiang, mainshock (MS7.3) and its aftershock sequence[J]. Chinese Journal of Geophysics, 2014, 57(3): 1012-1020 (in Chinese with English abstract).
Zhang Guangwei, Zhang Hongyan, Sun ChangqingZhang Guangwei, Zhang Hongyan, Sun Changqing. Mechanism of the 2015 Pishan, Xinjiang, MS6.5 mainshock and relocation of its aftershock sequences[J]. Seismology and Geology, 2016, 38(3): 711-720 (in Chinese with English abstract).
Zhao Bo, Gao Yuan, Huang Zhibin, Zhao Xu, Li DahuZhao Bo, Gao Yuan, Huang Zhibin, Zhao Xu, Li Dahu. Double difference relocation, focal mechanism and stress inversion of Lushan MS7.0 earthquake[J]. Chinese Journal of Geophysics, 2013, 56(10): 3385-3395 (in Chinese with English abstract).
Zhu Lupei, Helmberger D.V.Zhu Lupei, Helmberger D.V. Advancement in source estimation techniques using broadband regional seismograms[J]. Bulletin of the Seismological Society of America, 1996, 86(5): 1634-1641.
Zhu Lupei, Rivera L.A.Zhu Lupei, Rivera L.A. A note on the dynamic and static displacements from a point source in multilayered media[J]. Geophysical Journal International, 2002, 148(3): 619-627. DOI:10.1046/j.1365-246X.2002.01610.x
Zhu Lupei, Ben-Zion Y.Zhu Lupei, Ben-Zion Y. Parametrization of general seismic potency and moment tensors for source inversion of seismicwaveform data[J]. Geophysical Journal International, 2013, 194(2): 839-843. DOI:10.1093/gji/ggt137