The declustering of an earthquake catalog identify the background events (seismic events generated by regular earth movements), which leads to unbiased estimation of seismic activities in a region. The ergodicity of a seismic region represents ensemble average of events in time and space. The ergodicity of a seismic catalog is represented by a Thirumalai-Mountain (TM) metric. If the inverse TM metric becomes linear with time then the catalog is assumed to be declustered (it contains only the background events). But the original catalog normally contains backgrounds as well as seismically triggered events (Foreshocks and Aftershocks). The objective here is to optimally remove the triggered events from the catalog with an optimization algorithm so that the remaining catalog contains only the backgrounds. Here a parallel Grey Wolf Optimization (P-GWO) is introduced to perform the optimization task. Compared to the original GWO here the new updated positions of wolves are computed in parallel which reduces the computational complexity of the algorithm keeping the same accuracy level. The analysis is carried out on South California catalog and the results obtained are superior than that achieved by Cho et al. using PSO in 2010. Comparative results also demonstrate better performance over three benchmark statistical de-clustering methods by Gardner-Knopoff, Uhrhammer and Reseanberg.
