Armstrong F. Sompotan, Nanang T. Puspito, Endra Joelianto, Katsumi Hattori
Analysis of ionospheric precursor is not easy because the ionosphere is very dynamic as well as the earthquake phenomena. If the analysis method is the same dynamics with the earthquake phenomena, the estimation of earthquake parameters is possible to be realized. Neural network is an adaptive system that changes its structure to solve the problem during a learning phase. Therefore, the neural network is potentially to estimate the parameter of earthquake based on ionospheric precursor. A preliminary attempt was made to construct the neural network that can estimate the epicenter area. The GIM-TEC star method is useful to determine ionospheric anomalies associated with large earthquakes as ionospheric precursor data. The Kriging method is good to interpolate GIM-TEC star data as input of neural networks to estimate the epicenter area. The conclusion of five models of the ionosphere anomalies due to seismic activity show that the epicenter is at the edge of the less developed anomalies, whereas, for the growing anomalies, the epicenter is always located near the boundary of high and low density of TEC anomalies. The boundary is projection of the boundary of the unstressed and stressed rock area below the earth’s surface. © 2015 Academic Journals Inc.
Department of Earth Sciences, Institute of Technology Bandung, Jl. Ganesha 10, Bandung, 40132, Indonesia; Department of Physics, Manado State University, Tondano, 95618, Indonesia; Global Geophysics Research Group, Institute of Technology Bandung, Jl. Ganesha 10, Bandung, 40132, Indonesia; Instrumentation and Control Research Group, Institute of Technology Bandung, Jl. Ganesha 10, Bandung, 40132, Indonesia; Department of Earth Sciences, Chiba University, 1-33 Yayoi-cho, Inage, Chiba, 263-8522, Japan