1 Adaptive Short Time Fourier Transform (STFT) Analysis of Seismic Electric Signal (SES): A Comparison of Hamming and Rectangular Window W.Astuti, W.Sediono, A.M. Aibinu, R. Akmeliawati, M. J. E. Salami Department of Mechatronics Engineering International Islamic University Malaysia Gombak, Selangor Darul Ehsan, Malaysia email: winda1977@gmail.com , wsediono@iium.edu.my, rakmelia@iium.edu.my Abstract— Seismic electric signal (SES) is one of features for predicting earthquakes (EQs) because of its significant changes in the amplitude of the signal prior to the earthquake. This paper presents detailed analysis of SES recorded prior to earthquake that occurred in Greece in the period from January 1, 2008 to June 30, 2008. During this period of time 5 earthquakes were recorded with magnitudes greater than 6R. In this analysis STFT involving adaptively sliding window technique is used, in which Hamming and rectangular window functions are applied and compared. The comparison shows that Hamming window gives better results in analyzing the first significantly changes of SES prior to the EQ. The application of Hamming window resulted in less rippled spectrum shape which is more suitable to be used in characterizing the SES. Keywords; seismic electric signal (SES); adaptive STFT; Hamming window; rectangular window I. INTRODUCTION Earthquake is one of natural disasters that results in major distraction. Many efforts have been put in place in order to reduce disastrous effect of the earthquake. Prediction system could be one of the important solutions in order to reduce the number of fatalities and disaster effects [1]. Seismic electric signal (SES) is one of the phenomena that can be used to predict the earthquake, since it shows significant changes prior to the earthquake [2]. SES, superposition of earth’s electric field, is caused by tectonics stress load change of the lithosphere specifically in the seismogenic region. Seismic electric signal used in this work was collected from the database of earth’s electric field [3]. The usage of SES to predict the earthquake, which is known as VAN method, has been extensively implemented in Greece. This method was carried out by monitoring continuously the earth’s electric potential change and their East-West (E-W) and North-South (N-S) polarity gradients. SES, as train-like pulses of 3-5 minutes period, is oscillating electric field of T=1 day due to lithosphere oscillation caused by the K1 tidal component [4]. This signal belongs to very long period electric signal, which is caused by very long term strain rate of seismic region, mainly due to long term lithosphere plate motion [5]. These signals are different and generated during the last phase of the preparation of an earthquake when drastic change take place in seismogenic area in short time before rock fracturing. SES is registered by a pair of electrodes in contact to the ground surface at certain distance to the epicenter area. This SES corresponds to the total field generated in the focal area. Earthquake focal area is an area where seismogenic area is activated [6]. In the moment of seismic area has been activated, the electric potential at focal area is P. The electric potential at certain point far from the focal area (P r ) is a function of distance (r) and the observation time (t). ( ) t r f P r , = (1) The relationship between the earth’s electric potential and the distance from the epicenter area is inversely proportional to the electric potential at focal area P. Electric field consists of three components, which are located at x direction, y direction and z direction. Based on the measurements of ground surface earth’s electric field, it is assumed that x direction and y direction are registered by horizontal dipoles. The z component, which is in vertical direction, contains the same quality of information as component x and y. However, due to technical difficulties z component measurement required a vertical dipole in the ground of 150-200 m depth. Based on this difficulty z component has been ignored (Private communication with Dr. C. Thanassoulas, 2012). The total magnitude (E) of the electric field is obtained as 2 2 y x E E E + = G (2) The SES database is available at www.earthquakeprediction.gr. The measurement data consists of SES per day where each day consists of 1,440 samples of data. The SES database contains measurement data from three different monitoring sites. These monitoring sites are located in certain areas (as shown in Figure 3) which are; • Athens (ATH) monitoring site, operated from May 23, 2003 up to now. 2012 IEEE Symposium on Industrial Electronics and Applications (ISIEA2012), September 23-26, 2012, Bandung, Indonesia 978-1-4673-3005-3/12/$31.00 ©2011 IEEE 372