Optimal Model of Scattered Seismicity in Tasks of Seismic Hazard Assessment

The construction of an earthquake source zones (ESZ) model is an important step in seismic hazard assessment. Earthquakes occur mainly on pre-existing faults; therefore, the strongest earthquakes are generally considered to be confined to extended linear structures. However, a significant number of earthquakes occur outside such structures and are classified as so-called scattered seismicity. Historically, in ESZ models, scattered seismicity has been represented by a uniform distribution of epicenters within areas considered relatively homogeneous, known as “domains.” The domain boundaries are defined by experts. This approach was developed in the late 1960s and was justified by the short period of instrumental observations and the insufficiently high quality of earthquake parameter determination in catalogs. In modern conditions, high-quality data are quite sufficient for more detailed modeling of scattered seismicity in the form of local estimates on a regular grid. Various approaches exist for such modeling. The aim of this work is to select or develop an optimal method for spatially local estimates of the number of earthquakes at or above the magnitude of completeness. Obviously, the goal of an optimal model is to predict the locations of epicenters (hypocenters) of future earthquakes. Therefore, the main criterion for selecting the optimal method is the best fit of the model to real data that were not used in constructing the model or estimating its parameters. The maximum likelihood method is used to determine the degree of correspondence between the model and the data. Another selection criterion is the minimization of subjective assumptions in model construction, ensuring full reproducibility of the model using only observed data.