Relocation of the aftershocks of Mohammad Abad Rigan Earthquake December 20,2010, (Mn = 6.5) using a nonlinear method

Major earthquakes are often associated with large earthquakes which have a magnitude smaller than the main shock known as aftershocks. The occurrence of aftershocks with different magnitudes and times is a random process and therefore in the area affected by the main shock, it can cause greater damage than the main shock, and hence they are very important. Study of aftershocks can be useful to get information from tectonic activites and causative faults. Many studies have considered the aftershocks of large earthquakes, such as Omori (1894), Otsu (1961) and kisslinger (1996). Among the aftershock studies, the exact relocation of the main earthquake and its aftershocks help us find the causative fault and the releasing energy associated with that fault. Many studies have used relocation methods to examine the aftershocks. Some of these methods are Hong et al (2008), Hugh et al (2009) and Zhao et al (2011). Due to the complexity of the earth sub-layers and the three-dimensional structure of the crustal velocity and also the seismic wave path from the source to stations, there is a nonlinear relationship between the arrival time of seismic waves at the stations and the hypocenter of the earthquake. In order to simplify the earthquake location problem solving, most methods and programs use linearized relationships. Most of these methods and algorithms are based on the Geiger’s principles (Geiger, 1912). Using the linearized relationships reduces the accuracy of earthquake location due to losing the higher terms of Taylor series. It may also lead to failure in determining the location of earthquakes using a suboptimal network, e.g. where the earthquake is located outside the seismic network. Thurber (1985) showed that when the depth of an earthquake was smaller than the closest distance to the station, determining the focal depth was not possible in linearized methods. Furthermore, using higher terms of Taylor series is required to calculate higher degree derivatives, which are very complex and sometimes impossible, using a three-dimensional velocity model. In order to avoid calculating the partial derivatives, Tarantola and Valette (1982) presented a method that determines the location of earthquakes with fully non-linear relationships with no need to calculate the partial derivatives. The basic theory of nonlinear probabilistic method to determine the location of the earthquakes was introduced by Tarantola and Valette (1982) and Tarantola (1987). In this study, we used a nonlinear probabilistic method based on Tarantula and Valette theory and NonLinLoc program (Lomax et al, 2000) to relocate the earthquakes. The Rigan earthquake with Mn = 6.5 occurred on Dec 20, 2010 in the Southeastern region of Iran. After this earthquake, a lot of aftershocks occurred in this area which in some cases the magnitude of aftershocks was in order of the main shock. The largest aftershock with a magnitude Mn = 6.0 occurred after 37 days which itself included a lot of aftershocks. To improve the quality of data, in this study we combined the arrival time data from the Iranian Seismological Center (IRSC) stations and the data from the International Institute of Seismology and Earthquake Engineering (IIEES). Due to the lack of proper station coverage in the southeastern region of Mohammad Abad Rigan, we added IIEES stations data in this area which greatly helped us increase the station coverage. Regarding the lack of a proper regional velocity model in the Eastern and the Southeastern regions of Iran, we used Tatar et al (2003) local velocity model and determined the depth of Moho based on Dehghani and Makris (1983) study in an order of 55 km. In this study, we used Omori’s law to specify the energy release in the media and occurrence of aftershocks chronologically.We found that a large number of aftershocks have occurred in two different time windows near the two large earthquakes; in this regard, we divided the data based on these two time windows. The first time window contained the main shock with Mn = 6.5 and aftershocks until the occurrence of second earthquake with Mn = 6.0. The second time window contained the second earthquake Mn = 6.0 and its aftershocks.In order to get good results, we considered those earthquakes recorded at least by five stations. Finally, we could relocate 222 aftershocks out of 296 aftershocks associated with Rigan area. The relocation results of the earthquakes showed that the two main earthquakes and their aftershocks were distributed in the epicenter and the focal depth separated completely. They also showed two different fault trends. Relocated aftershocks in the first time window showed a fault trend parallel to Kahurak Fault, and aftershocks with Mn > 4 in the second time window showed a fault trend parallel to Kahurak fault.