Estimation of Hypocentral Parameters of Regional Earthquakes Using a Fuzzy Logic Approach

This study presents improvements on the hypocentral location method using the fuzzy logic approach proposed by Lin and Sanford (Bulletin of the Seismological Society of America 91:82–93, ) in resolving regional earthquakes location. In addition to trying to determine epicentral parameters of the earthquake, focal depths are also attempted utilizing this method in a three-dimensional (3D) space. In this study, instead of P- and S- wave models, a trapezoidal membership function constructed by the direct use of travel time difference between the two stations for P- and S- arrival times was utilized. In this procedure, it is not necessary to precisely know the velocity structure of the studied area in order to map the uncertainties in arrival-times into the fuzzy logic space. The location determination process was performed with the defuzzification of only the grid points having the membership value of 1 obtained by normalizing all the maximum fuzzy output values of the highest values which gives more reliable epicentral locations for both real and synthetic earthquakes. The results show that this method for epicentral estimation works well if the depth of the earthquake is sufficiently shallow and according to the distance between the event and seismic network. In this method of hypocenter determination, firstly, epicenter coordinates are determined, then the hypocenter location is found by considering only the epicentral area in 3D. In the case of very shallow events (e.g., h < 10 km), the estimation of the focal depth using this method seems to fail, causing faulty results in the hypocenter coordinates; however, the estimation of epicentral coordinates is successful in giving a more reasonable epicentral location. Otherwise, the obtained results would be reversed. Consequently, the deeper the focal depth, the better the hypocenter location, but a worse epicenter determination. This method is only valid in the hypocentral location of the earthquake occurring in the upper crust because only primarily the arrivals of Pg and Sg on the seismograms and a half-space velocity model are used. The hypocenter locations of some selected shallow earthquakes that occurred in the south east of the Aegean Sea have been determined by this method with results that are highly comparable with those obtained by other seismic agencies.