Comprehensive geomagnetic signal processing for sucessful earthquake prediction

Previous studies suggested that seismo-electromagnetic signals were generated during earthquake preparation phase. While the anomalous signals in Japan, Russia, and Greece are rather clear; observational results in some regions including America and Indonesian have not conclusive yet, whether the signals are really exist or normal magnetic storm phenomena only. Some problems in measuring instrumentations might also result in anomalous signals. This paper reviews further possibilities by processing a nearly continuous geomagnetic data of INTERMAGNET observatories in three regions (Japan, America, and Indonesia) during the last twelve years (2000-2011). MATLAB ® with digital signal processing and statistical toolboxes is used for night-time data filtering thereafter processing methods of differentiation and moving average are applied. To get relationship between these anomalous signals and natural processes, tectonic settings of the regions and global geomagnetic conditions are also considered in the analysis. Fast decreasing of geomagnetic intensities signals are showed prior to earthquakes in Japanese regions, while some little increases of differentiation signals between two adjacent observatories were seen from their moving average values. The anomaly was 10.08 nT, during 8 days started from 32 days prior to the earthquake. Decreasing of geomagnetic intensities prior to earthquakes in American regions occurred rather longer and formerly (during 40 days started from 137 days prior to the earthquake), but the magnitudes are not so high. In Indonesian regions, fast decreasing of geomagnetic signals and increasing of differentiation signals are very clear, especially prior to the giant M9.1 Sumatra earthquake. The anomaly reached extreme magnitude of more than 200 nT during several days prior to the earthquakes. All observational results are back to the normal (base line) values after earthquake occurrences. The distances from the observatories to the epicent- rs and the magnitudes of earthquakes very affect the results. Elevations of observatories and the tectonic settings of regions may also affect the magnitude of anomalous signals. Some missing data during earthquake days are in fact very valuable to complete the analysis. After all, the geomagnetic signal processing is very valuable in searching appropriate precursors for successful earthquake predictions.