Improving the Locations of Local/Regional Earthquakes Using the Circle Method with Fuzzy Logic

In this study, the hypocenter location method based on the “circle method,” developed with a fuzzy logic approach, was used to resolve local/regional earthquake locations. To evaluate the performance and usefulness of the method in various situations, it was initially tested on synthetic models. After gaining experience, it was tested on real earthquake data. Three artificial seismic networks with different station distribution geometries were selected to investigate the effect of network shape on the solution. Additionally, all networks took into account the earthquake's proximity to the networks, whether the earthquake was inside, near, or far away from them. The hypocenter location was determined in three ways within the same method. One approach involved defuzzifying the output of the intersection process on three fuzzy logic output matrices. The second method applied defuzzification only to the grid points with the highest fuzzy output values. The last method used defuzzification on values of 1 obtained through a normalization process based on the highest membership value among the fuzzy outputs. While the results for earthquakes occurring away from the network were considered slightly more successful, in these cases, the volumes of the fuzzy logic output matrices being processed significantly increased, leading to much longer computation times. The proposed method was used to estimate the hypocenter locations of 122 shallow earthquakes that occurred in the eastern Black Sea region and was found to produce results comparable to those obtained by other seismic agencies. Compared to classical methods, most earthquakes show an average total spatial difference of approximately 5 km. This method is considered highly effective for local earthquakes within the station network, as well as for regional earthquakes occurring outside the network, particularly when the hypocenter-station distance is significantly greater than the distances between stations, resulting in substantial azimuthal gaps.