Application of maximum likelihood method to event association

Event association in seismic monitoring systems has been known to be one of the most important procedures to identify the occurrence of earthquakes, but also one of the most difficult to understand and to improve. Most local seismic monitoring systems, including earthquake early warning systems, have suffered from the occurrence of large regional or teleseismic events because they are often associated as a significant local event. In this study, we apply a maximum likelihood method to associate seismic event in the Earthworm platform. It has been shown that the maximum likelihood earthquake location (MAXEL) method could locate an event very well even, with a small number of triggers including outliers (Sheen, 2015; Sheen et al., 2016). We modified the original Earthworm event association module (binder_ew) to use the maximum likelihood method by keeping the original algorithm. It has been tested for data from seismic networks in South Korea and Ohio. It is found that regional or teleseismic events mislocated to a local event at the original module were located at the boundary of the monitoring grid or not associated as an event.