An M_S4.9 earthquake occurred at 08:17 on the December 27, 2016 in Rongchang District, Chongqing, and the epicenter is located in the north central section of Huayingshan basement fault system on the eastern margin of Sichuan Basin. The seismicity shown in the historical earthquake catalogue was originally very weak in this area. Since the late 1980s, due to the impact of waste water reinjection in the natural gas field, earthquakes of magnitude ≥4.0 occurred frequently and 14 earthquakes with M_S≥4.0 have occurred, the largest of which was Rongchang M_S5.0 earthquake in 1997. In this paper, the fine three-dimensional P-wave velocity structures and relocation results of seismic events in Rongchang and its surrounding areas are inversed by double difference tomography method, based on the P-wave and S-wave arrival time data of 1786 seismic events recorded by Chongqing regional fixed network, mobile network and Zigong local network from January 2008 to June 2020.
The results show that: 1)The distribution of high-velocity and low-velocity zones within 4km depth is significantly different from that below 7~13km depth. The P-wave high-velocity zone at 4km depth is mainly distributed in Renyi-Rongchang area, where there are four water injection wells, a major concentration area of continuous water injection in Rongchang since 2008. The range of Renyi-Rongchang high velocity zone significantly gets narrowed at the 7km depth and is obviously different from that at the shallow surface. The velocity structures on the east and west sides of Huayingshan basement fault vary greatly from 7 to 13km. The P-wave velocity structures of different sections across Huayingshan basement fault all indicate that the depth of the interface between the sedimentary cover and crystalline basement is 12km in Rongchang area, which is basically consistent with the previous research results and the characteristics of seismic reflection profiles in Rongchang area. The inversed velocity structures well mirror the shape of Luoguanshan fold, and further confirm the reliability of our results. 2)The lateral difference of P-wave velocity structure in the shallow layer of Rongchang area varies greatly. There is a high-velocity zone near the Luo2^# water injection well at the axis of Luoguanshan anticline and the depth distribution is 3~7km. The hidden fault in the north wing of Luoguanshan anticline with buried depth of 1.7km is developed near well Luo2^#, and the high velocity zone is distributed along the dip of the hidden fault, which may indicate that the hidden fault may be the main channel for wastewater infiltration. The depth of wastewater infiltration is up to 7km, resulting in a large velocity difference between the two sides of the fault. The M_S4.9 earthquake on December 27, 2016 and the M_S4.0 earthquake on December 28, 2016 are just distributed in the velocity transition zone. Obvious high-velocity body was not found below 3km in Luo4^# water injection well, which may be related to the cessation of water injection in Luo4^# well in February 2001. 3)The results of seismic relocation indicate that earthquakes are mainly distributed in the axis of the strongly deformed Luoguanshan anticline, showing obvious stripe distribution in NE direction, and the focal dominant depth is 0~6km. Based on the focal mechanism solution and the regional seismotectonic environment, it is believed that the seismogenic fault of earthquakes above M_S4.0 on the south side of Guangshun transverse fault should be the hidden fault on the south wing of Luoguanshan, while the seismicity on the north side of Guangshun transverse fault may be related to the hidden fault on the north wing of Luoguanshan.

The Wulong M_S5.0 earthquake on 23 November 2017, located in the Wolong sap between Wenfu, Furong and Mawu faults, is the biggest instrumentally recorded earthquake in the southeastern Chongqing. It occurred unexpectedly in a weak earthquake background with no knowledge of dramatically active faults. The complete earthquake sequences offered a significant source information example for focal mechanism solution, seismotectonics and seismogenic mechanism, which is helpful for the estimation of potential seismic sources and level of the future seismic risk in the region. In this study, we firstly calculated the focal mechanism solutions of the main shock using CAP waveform inversion method and then relocated the main shock and aftershocks by the method of double-difference algorithm. Secondly, we determined the seismogenic fault responsible for the M_S5.0 Wulong earthquake based on these calculated results. Finally, we explored the seismogenic mechanism of the Wulong earthquake and future potential seismic risk level of the region.
The results show the parameters of the focal mechanism solution, which are:strike24°, dip 16°, and rake -108° for the nodal plane Ⅰ, and strike223°, dip 75°, and rake -85° for the nodal plane Ⅱ. The calculations are supported by the results of different agencies and other methods. Additionally, the relocated results show that the Wulong M_S5.0 earthquake sequence is within a rectangular strip with 4.7km in length and 2.4km in width, which is approximately consistent with the scales by empirical relationship of Wells and Coppersmith(1994). Most of the relocated aftershocks are distributed in the southwest of the mainshock. The NW-SE cross sections show that the predominant focal depth is 5~8km. The earthquake sequences suggest the occurrence features of the fault that dips northwest with dip angle of 63° by the least square method, which is largely consistent with nodal planeⅡof the focal mechanism solution. Coincidentally, the field outcrop survey results show that the Wenfu Fault is a normal fault striking southwest and dipping 60°~73° by previous studies. According to the above data, we infer that the Wenfu Fault is the seismogenic structure responsible for Wulong M_S5.0 earthquake.
We also propose two preliminary genetic mechanisms of "local stress adjustment" and "fluid activation effect". The "local stress adjustment" model is that several strong earthquakes in Sichuan, such as M8.0 Wenchuan earthquake, M7.0 Luzhou earthquake and M7.0 Jiuzhaigou earthquake, have changed the stress regime of the eastern margin of the Sichuan Basin by stress transference. Within the changed stress regime, a minor local stress adjustment has the possibility of making a notable earthquake event. In contract, the "fluid activation effect" model is mainly supported by the three evidences as follows:1)the maximum principle stress axial azimuth is against the regional stress field, which reflects NWW-SEE direction thrusting type; 2)the Wujiang River crosscuts the pre-existing Wenfu normal fault and offers the fluid source; and 3)fractures along the Wenfu Fault formed by karst dissolution offer the important fluid flow channels.

Based on the seismic data collected from regional permanent stations and 6 temporal stations, we analyzed the seismic activity from October 2008 to July 2011 in Rongchang area. On the basis of HypoDD relocated results, we used Match&Locate method to detect and located the micro-earthquakes. We obtained the focal mechanism solutions of some earthquakes with M_L ≥ 3.5 by using CAP method. Then we analyzed the temporal-spatial distribution of earthquakes and discussed the characteristics of micro-seismicity before the M_L5.1 earthquake occurring on September 10, 2010. We totally detected 3 354 micro-earthquake events, which are nearly 5 times of the earthquakes in the seismic catalog issued by China Earthquake Networks Center. The magnitude of the detected events is mostly from M_L-1 to 1, and the focal depth is from 2 to 4km. The magnitude-frequency analysis shows that the catalog completeness is obviously improved after adding the detected earthquakes, with the lowest magnitude decreasing from M_L1.0 to 0.3. The earthquakes hypocenters are mainly clustered along faults or buried faults and in a dominant depth range consistent with the depth of injection wells, and also show a tendency of lateral extension from injection wells. The focal mechanism solutions of 9 earthquakes of M_L ≥ 3.5 presented reverse faulting, as the same as the preexisting faults, indicating that earthquakes were surely related to reactivation of the faults. The strike, dip and rate of the causative faults separated in wide ranges, which indicates not only obvious changes in structure and strike of preexisting faults but also the effect of increasing pore pressure on the local stress field. Before the M_L5.1 earthquake on September 10 of 2010, seismicity firstly showed clustering in time and covered the most part of the seismogenic fault in space. Then an obvious seismic quiescence occurred and lasted about 3 months. The phenomenon is consistent with the mechanism of creep sliding and resistance-uniformization along the fault zone, suggested on the basis of laboratory experiments, and it may be one of patterns of sub-instability along fault zone. However, such explanation needs to be further confirmed.

Stress changes due to the co-seismic slip on the source fault of the 2008 M_W7.9 Whenchuan earthquake and delayed response of inelastic deformation in the lower crust and upper mantle have an important role in the seismicity in Longmenshan area. After the Wenchuan earthquake,seismicity shows progressively increasing in a wide region. However,the south segment of the Longmenshan Fault did not show any significant change in seismicity,where positive Coulomb failure stress change(ΔCFS)was estimated under the elastic half-space model. Under such a background,the 2013 M_W6.6 Lushan earthquake occurred. This paper presents some preliminary results based on seismicity analysis and stress analysis using lithology models in which the lower crust and the upper mantle are suggested to be viscoelastic. The Wenchuan earthquake resulted in a miner negative coseismic ΔCFS in the hypocenter region of the Lushan earthquake. As a result of inelastic response the estimated ΔCFS reached the order of 0.2～0.4bar,a value sufficient to trigger earthquakes in critically loaded faults. We thus conclude that the Lushan earthquake provides a case of inelastic triggering of the Wenchuan earthquake. The 1970 M6.2 Dayi earthquake caused an obvious Coulomb stress shadow in its source area,which partly overlaps to the seismic gap between the ruptures of the Lushan and Wenchuan earthquakes. The stress shadow still exists although the area has been loaded by both the Wenchuan and Lushan earthquakes. We thus suggest that it is less likely that a great earthquake,which ruptures the entire gap,may occur in the near future if there are no other unknown factors.

Based on data collected from a temporary seismic network,in addition to some nearby permanent stations,we investigate velocity structure and seismicity in Rongchang gas field,where significant injection-induced seismicity has been identified. First,we use receiver functions from distant earthquakes to invert for detailed 1-D velocity structures beneath typical stations. Then,we use the double-difference hypocenter location method to relocate earthquakes of the 2010 M_L5.1 earthquake sequence occurring at the region. The relocated hypocenters show that the 2010 M_L 5.1 earthquake sequence was distributed in a small area surrounding major injection wells and clustered mostly along pre-existing faults. Major earthquakes show a focal depth less than 5km with a dominant depth of～2km,a depth of major reservoirs and injection wells. We thus conclude that the 2010 M_L 5.1 earthquake sequence might be induced by deep well injection of unwanted water at a depth～3km in Rongchang gas field.

Rongchang area had exhibited low levels of natural seismicity,and there was no record of earthquake with M_L>5 in the history.However,following the injection of unwanted water from gas production,seismicity has increased dramatically and showed progressive increase of magnitude since July 1988,and an earthquake of M5.2 occurred in 1997.Rongchang area is thus an ideal site for studying seismicity induced by deep well injection.Unfortunately,there was only one seismic station in the area,and the research was limited by the poor detectability and hypocenter location accuracy.In order to make a thorough investigation on the injection-induced seismicity in the area,a temporal seismic network was installed in July,2008 under the cooperation of the State Key Laboratory of Earthquake Dynamics,Geological Survey of Japan and Chongqing Earthquake Administration.The seismic network consists of 6 stations,by which waveforms are continuously recorded.As a result,both the detectability and location accuracy are improved greatly.This paper presents a brief summary of the cooperative project and some preliminary results of recent seismicity in the area.

Based on teleseimic data for the period of 2007 to 2010 acquired from the Three Gorges reservoir(Chongqing section)seismic network and Chongqing regional seismic network,we obtained the crustal thickness and Possion's ratio of the area through receiver function method.The results show that the crustal thickness ranges from 38.9km to 50.9km.Station CHK in northeastern Chongqing has the largest thickness which is about 50.9km;the thinnest crust is 38.9km under the station ROC in the west of Chongqing,and Poisson's ratio is about 0.27.The results show that the maximum Posson's ratio is beneath station JIZ in the east of the Chongqing section of Three Gorges Reservoir area,and the minimum Possion's ratio is beneath the station WUL,which is 0.228.In comparison with the Poisson medium(0.25),its maximum deviates up to 20.8%,and the minimum deviates by-8.80%.The Bouguer gravity anomaly often reflects subsurface density and crustal thickness variation.In this paper,the two agree with each other.The largest negative Bouguer anomaly is between CHK(50.9km)and WUX(49.7km),where the crust is the thickest.The smallest negative Bouguer anomaly is between ROC(38.9km)and YUM(41km),where the crust is the thinnest.