Estimation of Hypocentral Parameters of Local/Regional Earthquakes Using Grid Search Methods with A Fuzzy Logic Approach

This study presents improvements to the hypocentral method for identifying regional/local earthquake locations based on the grid search technique with a fuzzy logic approach. For this study, the newly proposed method was tested on synthetic models to evaluate its effectiveness in various situations. Subsequently, the method was tested on actual earthquake data. To investigate the effect of network shape on location identification, the technique was assessed using three artificial seismic networks with different station distribution geometries. Additionally, for all networks, three scenarios concerning the earthquake’s proximity to the networks were examined: Inside, just outside, and further away. Four types of norms were featured in this study: L₂ and L₁ for the P-wave and L₂ and L₁ for the S-wave. These four norms were transformed into fuzzy logic space using a half trapezoidal membership function that constructed minimum and maximum RMS values for all the norms. The location determination process was performed in two ways: Defuzzification of the output of the intersection process on four fuzzy logic output matrices and defuzzification of only grid points with maximum fuzzy output values.

The results show that this method of epicentral estimation is effective when an earthquake is shallow enough according to the distance between the event and the seismic network and has a special advantage when the location of an event is far from the network. This method is applicable only to the hypocentral location of an earthquake occurring in the upper crust, as only Pg and Sg arrivals on seismograms and a half-space velocity model are used; therefore, its validity is somewhat limited. However, by using this method, we were able to estimate the hypocenter locations of 151 shallow earthquakes that occurred in the eastern Black Sea and found that they closely resemble the locations identified by other seismic agencies. The average total difference between most earthquakes is approximately 5 km. This method appears to be highly effective for local earthquakes occurring within a network, for regional earthquakes occurring outside a network, and for hypocenter-station distances significantly greater than the distance between stations, that is, those with a great azimuthal gap.