服部 克巳

A tomographic approach is used to investigate the fine structure of electron density in the ionosphere. In the present paper, the Residual Minimization Training Neural Network (RMTNN) method is selected as the ionospheric tomography with which to investigate the detailed structure that may be associated with earthquakes. The 2007 Southern Sumatra earthquake (M = 8.5) was selected because significant decreases in the Total Electron Content (TEC) have been confirmed by GPS and global ionosphere map (GIM) analyses. The results of the RMTNN approach are consistent with those of TEC approaches. With respect to the analyzed earthquake, we observed significant decreases at heights of 250-400 km, especially at 330 km. However, the height that yields the maximum electron density does not change. In the obtained structures, the regions of decrease are located on the southwest and southeast sides of the Integrated Electron Content (IEC) (altitudes in the range of 400-550 km) and on the southern side of the IEC (altitudes in the range of 250-400 km). The global tendency is that the decreased region expands to the east with increasing altitude and concentrates in the Southern hemisphere over the epicenter. These results indicate that the RMTNN method is applicable to the estimation of ionospheric electron density. © 2011 Author(s).

Three-dimensional ionospheric tomography is effective for investigations of the dynamics of ionospheric phenomena. However, it is an ill-posed problem in the context of sparse data, and accurate electron density reconstruction is difficult. The Residual Minimization Training Neural Network (RMTNN) tomographic approach, a multilayer neural network trained by minimizing an objective function, allows reconstruction of sparse data. In this study, we validate the reconstruction performance of RMTNN using numerical simulations based on both sufficiently sampled and sparse data. First, we use a simple plasma-bubble model representing the disturbed ionosphere and evaluate the reconstruction performance based on 40 GPS receivers in Japan. We subsequently apply our approach to a sparse data set obtained from 24 receivers in Indonesia. The reconstructed images from the disturbed and sparse data are consistent with the model data, except below 200 km altitude. To improve this performance and limit any discrepancies, we used information on the electron density in the lower ionosphere. The results suggest the restricted RMTNN-tomography-assisted approach is very promising for investigations of ionospheric electron density distributions, including studies of irregular structures in different regions. In particular, RMTNN constrained by low-Earth-orbit satellite data is effective in improving the reconstruction accuracy. Copyright © 2011 by the American Geophysical Union.

We report seismo-ionospheric precursors of anomalous decreases in the total electron content (TEC) appearing day 5 prior to an M9.3 earthquake, the largest one in the last five decades, which occurred in Sumatra-Andaman, Indonesia on 26 December 2004. Sequences of global ionosphere maps of the TEC derived from worldwide ground-based receivers of the global positioning system (GPS) are used to statistically study the temporal and spatial precursors of the earthquake. It was found that the temporal precursor of the GPS TEC around the epicenter was significantly reduced during the afternoon period on d 5 before the earthquake. The spatial precursors prominently, persistently, and simultaneously appear around the epicenter and its conjugate areas of the SumatraAndaman earthquake.

Many researchers studied the relationships between appearances of geomagnetic anomalies and their distances to earthquake epicenters or faults. Yet, occasionally some magnetometer stations located nearby earthquake epicenters and/or faults do not observe geomagnetic anomalies. In this paper, a new hybrid system which simultaneously takes the hypocenter and fault plane solution into account is constructed to examine 38 earthquakes interpreted to be associated with geomagnetic anomalies during the period 1988-2001 in Taiwan. The Surface Magnetic Anomaly Reference Tip (SMART) of the new system is used instead of the epicenter or the fault to investigate statistically the distance relationship between the anomalies and the earthquake parameters. Results show that the anomalies gather along the fault and in the belt zone to the SMART. Possible mechanisms causing the anomalies in the two zones are proposed and discussed. Characteristics of the anomaly might shed some light on locations of faults before earthquake occurrences. © 2010 Elsevier B.V.

In this paper, the total electron content (TEC) of Global Ionosphere Maps (GIMs) is normalized and employed to study the seismo-ionospheric anomalies at the time of the 12 May 2008 M 8.0 Wenchuan Earthquake. The space weather conditions are taken into account. It is found that remarkable reductions appear locally around the epicentre and their conjugate points during the daytime of 29 April and 6-10 May 2008. A global study and a strict criterion are applied to detect anomalies. Results show that the anomalies on 29 April and 6 and 7 May 2008, which are respectively days 13, 6 and 5 before, are possibly related to the earthquake. The conjugate signature implies the seismo-generated electric field is essential. © 2010 Taylor & Francis.

Despite its extreme importance and years of efforts, practical short-term earthquake prediction still remains to be achieved in future. However, earthquake-related electromagnetic phenomena are recently considered as a promising candidate for short-term earthquake prediction. There have been accumulated a lot of evidences of precursory signatures in a wide frequency range. In this paper, the recent studies on seismo-electromagnetics such as pre-seismic phenomena such as ULF geomagnetic anomalous changes and ionospheric disturbances associated the 2004 Sumatra-Andaman earthquake are presented.

Scientists analyzed temporal variations of local data recorded in Taiwan and reported anomalous reductions in the ionospheric total electron content (TEC) on days 4 and 3 before (17 and 18 September) the Chi-Chi earthquake (Mw 7.6, ML 7.3), and days 3 and 1 prior to (19 and 21 October) the Chia-Yi earthquake (ML 6.4) in Taiwan. In this paper, we propose a spatial analysis which compares the data recorded inside and outside of the earthquake area around Taiwan to discriminate whether the anomalies are local (earthquake related) or global (non-earthquake related) phenomena. Results suggest that the anomalies appearing day 3 before the Chi-Chi earthquake, and days 3 and 1 before the Chia-Yi earthquake are local (earthquake related) phenomena and the anomaly appearing on 17 September 1999 (day 4 before) might be the result of both global phenomena (i.e., a geomagnetic storm) and the local effect of the Chi-Chi earthquake.

Despite its extreme importance and years of efforts, practical short-term earthquake prediction still remains to be achieved in future. However, earthquake-related electromagnetic phenomena are recently considered as a promising candidate for short-term earthquake prediction. There have been accumulated a lot of evidences of precursory signatures in a wide frequency range. This paper deals with various electromagnetic phenomena associated with the 2004 Sumatra-Andaman earthquake (M9) in order to show the state of the art of seismo-electromagnetics for prediction research. Both post- and/or co-seismic and pre-seismic phenomena such as ULF geomagnetic and ionospheric disturbances are presented.

The global ionospheric map (GIM) is used to observe variations in the total electron content (TEC) of the global positioning system (GPS) associated with 35 M ≥ 6.0 earthquakes that occurred in China during the 10-year period of 1 May 1998 to 30 April 2008. The statistical result indicates that the GPS TEC above the epicenter often pronouncedly decreases on day 3-5 before 17 M ≥ 6.3 earthquakes. The GPS TEC of the GIM and electron density profiles probed by six microsatellites of FORMOSAT3/ COSMIC (F3/C) are further employed to simultaneously observe seismoionospheric anomalies during an Mw7.9 earthquake near Wenchuan, China, on 12 May 2008. It is found that GPS TEC above the forthcoming epicenter anomalously decreases in the afternoon period of day 6-4 and in the late evening period of day 3 before the earthquake, but enhances in the afternoon of day 3 before the earthquake. The spatial distributions of the anomalous and extreme reductions and enhancements indicate that the earthquake preparation area is about 1650 km and 2850 km from the epicenter in the latitudinal and longitudinal directions, respectively. The F3/C results further show that the ionospheric F2 peak electron density, N nF2, and height, hmF2, significantly decreases approximately 40% and descends about 50-80 km, respectively, when the GPS TEC anomalously reduces. Copyright 2009 by the American Geophysical Union.

Significant geoelectric and magnetic field anomalous changes have been observed prior to and during the Mw7.9 of June 4, 2000 Bengkulu earthquake in the southern Sumatran plate boundary, Indonesia. The mainshock occurred in the intra-oceanic slab and triggered the aftershocks that mainly occurred in the subduction interface of Sumatra Island at a distance of about 220 km from our monitoring sites. We obtained a set of electric and magnetic data for a couple of hours before, during and after the mainshock. Changes of electric field of about 5-10 mV/100 m and magnetic field of about 40 pT were observed at all monitoring sites at about 90 min before the mainshock. These anomalous signals occurred at the same time as the global geomagnetic storm that was clearly observed at Biak, Alice Springs and Kakadu Geomagnetic Observatories. A monitoring instrument located at the so-called sensitive site recorded a remarkable change of geoelectric potential of up to 140 mV/100 m at about 40 s after the origin time of the mainshock. These co-seismic anomalous signals occurred coincidently with the arrival of seismic P-wave. We interpret these signal changes in terms of the electrokinetic effect associated with seismic wave propagation. Crown Copyright © 2008.

The Space Environment Research Center of Kyushu University has obtained geomagnetic data in the Circum-pan Pacific Magnetometer Network (CPMN) region for over 10 years, and has recently deployed a new real-time Magnetic Data Acquisition System (MAGDAS) in the CPMN region and an FM-CW radar network along the 210° magnetic meridian (MM) for space weather research and applications. This project intends to get the MAGDAS network fully operational and provide data for studies on space and lithosphere weather. In connection with this project, we propose a new ultra-low frequency (ULF) wave analysis method to study ULF anomalies associated with large earthquakes using magnetic data. From a case study of the 1999/05/12 Kushiro earthquake with magnitude M = 6.4, we found a peculiar increase of H-component power ratio AR/AM of Pc 3 magnetic pulsations a few weeks before the earthquake, where AR is the power obtained at Rikubetsu station (r = 61 km) near the epicenter and AM is the power obtained at a remote reference station, Moshiri (r = 205 km). It is also found that the H-component power ratio AD/AY of Pc 3 increased three times just a few weeks before the earthquake and after one week decreased to the normal level, where AD is one-day power at Rikubetsu station and AY is the one-year-average power. © 2008 Elsevier Ltd. All rights reserved.

Anomalous ULF geomagnetic field change is considered to be one of the most convincing and promising phenomena among earthquake-related electromagnetic effects, because we expect emissions from the crust of the source region. There has been a great deal of accumulated and convincing evidence of ULF magnetic signatures before large earthquakes as reported in the previous studies. In order to verify these phenomena preceding large earthquakes and to clarify the relationship between electromagnetic phenomena and possible physical mechanism, we have investigated the data of ULF geomagnetic observations at Kototabang and Biak stations associated with the 2004/2005 Sumatra earthquakes. A case study is carried out in this work to investigate the pre-earthquake ULF geomagnetic anomalies during the Sumatra-Andaman earthquake on December 26, 2004 (magnitude Mw = 9.0 and depth = 30 km from USGS catalog), and Sumatra-Nias earthquake on March 28, 2005 (Mw = 8.7 and depth = 30 km). For the aim of finding any precursors, we have adopted the spectral density ratio analysis and transfer functions analysis based on wavelet transform method. Results of the spectral density analysis indicate similar variations to those of amplitude for the induction arrow in transfer function analysis. Both of these variations at Kototabang exhibit strange or anomalous changes from a few weeks before the Sumatra-Andaman earthquake to July 2005, while there are no apparent changes at remote station of Biak. To make these results more convincing, the fractal analysis based on a Higuchi method has been applied to the same observed data, which also show a significant change (or decrease) in mono-fractal dimension a few weeks before the earthquakes. This result would lend a further support to those by the spectral density ratio and transfer function analyses. So, we can conclude that the anomalous change as observed simultaneously by the three methods, might be a possible signature related with the earthquake preparation phase of Sumatra earthquakes. © 2008 Elsevier Ltd. All rights reserved.

In this paper, 3-component geomagnetic fields recorded by a fluxgate magnetometer are employed in examining 181 ML≧ 4.0 earthquakes around the Taiwan area from 2002 to 2005. To search anomalies, the singular value decomposition (SVD) is used to extract the declinations and inclinations of the principal components from the geomagnetic field every 900 s for each 5-day period and the entire four years. Six reference angles of surface magnetic anomaly reference tip (SMART) are employed to readjust the azimuth of anomalous polarizations for unifying earthquakes. A bootstrap method is further applied to find whether the anomalies are significant or not. Meanwhile, the odds proportions between the inward and outward geomagnetic field on the earth's surface reveal the associated anomalous inclinations. With the anomalous inclinations and declinations observed 10 days before earthquakes, SMART and the polarizations of the anomalous field can be located and described, respectively. © 2008 Elsevier Ltd. All rights reserved.

The microwave emission at 22 GHz, 2GHz and 300 MHz was experimentally found in association with rock crash in laboratory. This phenomenon could be applied to detect earthquakes or volcanic eruption phenomena which are accompanied with the rock crash. In order to confirm the microwave emission in natural phenomena, we planned a field test at the volcano of Miyake-jima. This paper describes the estimation of the signal to noise ratio of the test, the test site and the measuring system. The obtained data and their analysis showed successful detection of signals at 300 MHz. ©2009 The Institute of Electrical Engineers of Japan.

Indonesia is one of the most seismically active regions in the world, containing numerous active volcanoes and subject to frequent earthquakes with epicenters distributed along the same regions as volcanoes. In this paper, a case study is carried out to investigate pre-earthquake ionospheric anomalies in total electron content (TEC) during the Sulawesi earthquakes of 1993 - 2002, and the Sumatra-Andaman earthquake of 26 December 2004, the largest earthquake in the world since 1964. It is found that the ionospheric TECs remarkably decrease within 2 - 7 days before the earthquakes, and for the very powerful Sumatra-Andaman earthquake, the anomalies extend up to about 1600 km from the epicenter.

The scaling properties of the time dynamics of 2000-2001 ULF geomagnetic data (in the frequency range between 1 mHz and 10 Hz) observed at Izu Peninsula in Japan were investigated. On the basis of detrended fluctuation analysis (DFA), which is a powerful method to detect scaling in non-stationary time series, deviations from uniform power-law scaling were identified and quantified. Our results suggest that a significant non-uniform scaling behavior in ULF geomagnetic data could be related to the occurrence of intense seismic clusters. These first results could be useful in the framework of short-term earthquake prediction. © 2008 Elsevier B.V. All rights reserved.

Earthquake-related electromagnetic phenomena have recently been considered a promising candidate for short-term earthquake prediction. There has accumulated a lot of evidence of precursory signatures in a wide frequency range (DC to VHF), but the ULF geomagnetic change is one of the most promising phenomena. We review our results on seismogenic ULF emissions; and characteristics of those emissions and methodologies to detect abnormal changes in ULF electromagnetic field associated with earthquakes. © 2007 Wiley Periodicals, Inc.

A well-developed cumulonimbus and/or a mesoscale convective system (MCS) causes hazardous weather phenomena, including tornados, strong gust wind, lightning discharges, and torrential rain. These phenomena might lead to disasters and serious damages to our life. So we need to observe them in close range, thereby enabling us to elucidate their fine structure and characteristics. In order to achieve these purposes, more compact weather radars are developed. Their spatial and temporal resolution, and various functions are improved. In this paper, we introduce the recent development in weather radars for research. We additionally introduce new experiments for surveillances and detection for hazardous atmospheric phenomena with these radars. © 2008 The Institute of Electrical Engineers of Japan.

Investigating highly non-stationary time series, which typically exhibit long-range correlations, is a classic problem in physics. Here, we analyze the scaling properties of the dynamics of ultra-low-frequency (ULF) geomagnetic data (in the frequency range between 1 mHz and 10 Hz) observed at Izu Peninsula in Japan. On the basis of the detrended fluctuation analysis (DFA), deviations from uniform power-law scaling were quantified. Our findings point out to a significant non-uniform scaling behavior in ULF geomagnetic data in relationship with the occurrence of intense seismic clusters. © 2007 Elsevier B.V. All rights reserved.

ULF (ultra-low-frequency) electromagnetic emission is recently recognized as one of the most promising candidates for short-term earthquake prediction. This paper reviews previous convincing evidence on the presence of ULF emissions before a few large earthquakes. Then, we present our network of ULF monitoring in the Tokyo area by describing our ULF magnetic sensors and we finally present a few, latest results on seismogenic electromagnetic emissions for recent large earthquakes with the use of sophisticated signal processings. © 2007 by MDPI.

Despite its extreme importance and years of efforts, practical short-term earthquake prediction still remains to be achieved in future. However, earthquake- related electromagnetic phenomena are recently considered as a promising candidate for short-term earthquake prediction. There have been accumulated a lot of evidences of precursory signatures in a wide frequency range (DC-VHF). In this paper, the state of the art and research progress of seismo-electromagnetics are summarized.

The variations of geoelectrical potential differences in Hakuba area, Nagano Prefecture, Japan have been investigated. The noises originated from the DC-driven trains were found to contaminate the natural geoelectrical potential data. The most intense influence of trains occurred when the train was running nearby measuring dipoles. The gradient of the potential was deflected towards the railways and/or the position of the train, exhibiting a certain correlation between the power supply data at substation and the geoelectrical potential data at measuring sites. Extracting the high correlation part (r > 0.7), "idealized" train noise can be computed by the least square method. The reduction of train noise by more than 60% was achieved by subtracting the idealized noise from observed data.

Earthquake- related electromagnetic phenomena are recently considered as a promising candidate for short-term earthquake prediction. There have been accumulated a lot of evidences of precursory signatures in a wide frequency range (DC-VHF), but the ULF geomagnetic change is one of the most promising phenomena. We review our latest results on seismogenic ULF emissions; characteristics of those emissions and methodologies to detect abnormal changes in ULF electromagnetic field associated with earthquakes.

Despite its extreme importance and decades of efforts, practical short-term earthquake prediction still remains to be achieved in future. However, the electromagnetic research has been demonstrating some promises. This paper briefly reviews the recent progress of what we call “seismo-electromagnetics”, mainly referring to Japanese studies by the observational point of view. We demonstrate some results of observations, in varied frequency ranges, on the anomalous telluric current, ULF geomagnetic transient change, VLF-HF natural emissions and anomalous transmission of VLF and VHF band radio waves. We also summarize proposed physical mechanisms of these phenomena, including the notion called “Lithosphere-Atmosphere-Ionosphere (LAI) Coupling”. We believe that electromagnetic studies will play an important role in not only earthquake prediction but also in understanding physical processes of earthquake generation.

Results of studying the behavior of the vectors of gradients and phase velocities of ULF geomagnetic disturbances (F < 1 Hz) in the Japan seismic zone are presented. The gradient and phase velocity vectors along the Earth's surface were determined using data of the group of three high-sensitivity three-component magnetovariation stations (MVC-3DS) located at triangle vertices at a small (∼5 km) distance from one another (magnetic gradiometer). Two such groups of stations were installed in 1999 southwest and southeast of Tokyo at a distance of ∼150 km from each other. It has been indicated that, several months before strong earthquakes (M > 5), the values of gradients and phase velocities start anomalously changing, and directions toward sources of impending strong earthquakes appear in the distribution of gradient vector directions. Directions from sources of impending earthquakes appear in the distribution of phase velocity vector directions. It is proposed to use gradients and phase velocities of ULF and ELF geomagnetic disturbances as one of the factors in a short-term prediction of strong earthquakes. © Pleiades Publishing, Inc. 2006.

In order to extract any ULF signature associated with earthquakes, the singular spectral analysis (SSA) and the principal component analysis (PCA) have been performed to investigate a possibility of discrimination of signals from different sources (geomagnetic variation, artificial noise, and the other sources (earthquake-related ULF emissions)). We apply SSA to the time series data observed at closely spaced stations; Seikoshi (SKS), Mochikoshi (MCK), and Kamo (KAM) stations. Then, PCA is applied to the time series data sets filtered at 0.01 Hz of NS component at three stations. In order to remove the most intense signal like the first principal component, we make the differential data sets of SKS-MCK and MCK-KAM for the above data. The major findings are summarized as follows. (1) It is important to apply simultaneously SSA and PCA, because SSA gives the structure of signals and the number of sensors for PCA is estimated. This makes the results more convincing. (2) There is a significant advantage using PCA with differential data sets of filtered (0.01 Hz band) signal between SKS-KAM and MCK-KAM in NS component for removing the most intense signal like global variation (solar-terrestrial interaction). This yields that the anomalous changes in the second principal component were detected more clearly, and the contribution of the second principal component is found to be 20-40%. It is enough to prove mathematical accuracy of the signal. Further application is required to accumulate events. These facts demonstrate a possibility of monitoring the crustal activity with using the SSA and PCA. © 2006 Elsevier Ltd. All rights reserved.

Behavior of gradient and phase velocity vectors of ULF (F < 1 Hz) geomagnetic disturbances was investigated for two earthquake (EQ) events in Japan in 2000 (M > 6) and 2003 (M = 5.8) years. The gradient and phase velocity vectors were determined using the data of three high-sensitive three-component magnetic stations situated at a distance ∼5 km from each other (magnetic gradientometer). Two gradientometers were installed southeast and southwest of Tokyo at a distance ∼150 km from each other. It was found that the gradient and phase velocity values had anomaly changes 3-4 months before the strong EQs. New directions of the gradient vectors took place in the same period - the directions just to the forthcoming EQ epicenter. The directions from the forthcoming EQ epicenter arouse for the phase velocity vectors. We propose to use the gradient and phase velocity vectors of the ULF geomagnetic disturbances as important factors of the short-term prediction of strong earthquakes. © 2006.

The work is devoted to a numerical analysis specifically adapted for multi-point detection of geomagnetic events, which precede or accompany earthquakes. Both numerically simulated data and geomagnetic perturbations during real earthquakes were investigated. The objective of this paper is to study the correlation properties of the ULF fields on small distances in seismoactive zone of Japan. By applying a combination of the Fourier, Hilbert and wavelet transforms to a single component and to their combinations we developed five correlation techniques and analyzed the geomagnetic variations at five observation points for two-year permanent observations. Our study was in two main directions: correlation properties of the field before an earthquake, and a coherency of geomagnetic pulsations launched by the earthquake already taken place. Five different coherency criteria have been examined: (1) the cross correlation function; (2) semblance; (3) in-phase; (4) σ; and (5) frequency stability criterion. We found that the signal detection times determined by the cross correlation, semblance and in-phase criteria are much more stable compared to that found with the help of the sigma correlation technique which allows detecting when the signal to noise ratio is about ∼3. The study revealed anomaly signals in the period range from 15 min to 1.5 h independently at all observatories. The signature of the anomaly signal is a wave train usually consisting of 2-3 of 0.5-1 h pulses which appears 3-4 h and 8 h prior to the earthquake time if we study correlations within 90 km radius. Intensity of the signals is around 0.5-1 nT. © 2006 Elsevier Ltd. All rights reserved.

In this paper, a relationship between M ≥ 5.0 earthquakes and diurnal variations of the total geomagnetic field recorded at eight magnetometers in Taiwan during 1988-2001 are examined. One magnetometer station was setup in a seismic quiet area as a reference, while the others were located in areas of high seismicity or crustal activity observing earthquake effects. We compute the distribution of diurnal range ratios between the reference and each observation station for the entire thirteen years as a background and compare it with the monitored distributions during five different time periods before and after an M ≥ 5.0 earthquake occurring within a distance of 50 km from the observation station. Three specific earthquakes with different magnitudes, including the M = 7.3 Chi-Chi earthquake show that the monitored distributions one month before and during the month of the earthquakes significantly depart from the associated background. It is found that changes of underground conductivities and currents around the forthcoming epicenter significantly affect the near-by geomagnetic field on the ground during the earthquake preparation period. The statistical results demonstrate that the monitored distributions of geomagnetic anomalies are highly related to the focal mechanism. © 2006.

We study preseismic behavior of the ULF geomagnetic variations at Karymshino (E158.13, N52.83), Russia, and Matsukawa (E140.94, N39.88), Japan. A depression of ULF power around local midnight is registered several days before strong isolated earthquakes. The relative depression, i.e. δ D = (frac(1, G) - 〈 frac(1, G) 〉) / 〈 frac(1, G) 〉 has been analyzed, where the absolute depression is defined as a sum of inversed values of spectral power densities of horizontal components (where frac(1, G) = 1 / Phh and Pdd are the mean spectral densities of H and D components) in a sliding window ±(10-15) days. At Karymshino the most evident effect is found 3 days before an earthquake for the 0.5 h vicinity of local midnight in the frequency range 0.02-0.05 Hz. The data from Matsukawa are used to estimate the spatial scale of the effect and its local nature has been confirmed. Furthermore there is a definite correlation between the depression value and index of seismic activity. Thus, the analysis supports a hypothesis on seismic origin of the observed ULF geomagnetic field depression preceding an earthquake. © 2006.

In order to study earthquake-related ULF geomagnetic field changes, it is important to discriminate the signal from the other noises such as magnetic pulsations originated from solar activity and artificial noises from the DC driven train and factories. For this goal, the interstation transfer function method with wavelet transform has been applied to data observed in the ULF electromagnetic sensor array at Boso Peninsula, Japan. The results indicate that the proposed method has the capacity to eliminate the global signal. © 2005 Wiley Periodicals, Inc.

Geoelectrical and geomagnetic fluctuations are the end product of several geophysical phenomena. In particular signals measured in seismically active areas can be attributed to stress and strain changes associated with earthquakes. The complexity of the phenomena requires the development of advanced signal discrimintion and statistical methods. In this paper, we analyzed the short-term variability of geoelectrical potential and geomagnetic fields measured at Kiyosumi, Uchiura, and Fudago stations located in the southern part of Boso Peninsula, which is one of the most seismically active areas in Japan. The directions of signal arrival have been investigated with the use of local mid-night geoelectrical potential data, when artificial noise is low. Reduction of intense signals with external origin has also been made with the use of the interstation transfer function method. These analyses have shown the capability to discriminate the signal sources. Especially, direction finding analysis of time series data has shown to be effective in source separation. To identify the waveform and the direction of arrival of signals is important for monitoring the ULF electromagnetic environment in seismic areas and for understanding the preparation process of crustal activity. © 2005, The Institute of Electrical Engineers of Japan. All rights reserved.

Preseismic intensification of fracturing has been investigated from occurrence analysis of seismo-acoustic pulses (SA foreshocks) and ULF magnetic pulses (ULF foreshocks) observed in Karimshino station in addition to seismic foreshocks. Such analysis is produced for about 40 rather strong and nearby isolated earthquakes during 2 years of recording. It is found that occurrence rate of SA foreshocks increases in the interval (-12, 0 h) before main shock with 3-times exceeding of background level in the interval (-6, -3 h), and occurrence probability of SA foreshocks (pA∼75%) is higher than probability of seismic foreshocks (ps∼30%) in the same time interval. ULF foreshocks are masked by regular ULF activity at local morning and daytime, nevertheless we have discovered an essential ULF intensity increase in the interval (-3, +1 h) at the frequency range 0.05-0.3 Hz. Estimated occurrence probability of ULF foreshocks is about 40%. After theoretical consideration we conclude: 1) Taking into account the number rate of SA foreshocks, their amplitude and frequency range, they emit due to opening of fractures with size of L=70-200 m (M=1-2); 2) The electro-kinetic effect is the most promising mechanism of ULF foreshocks, but it is efficient only if two special conditions are fulfilled: a) origin of fractures near fluid-saturated places or liquid reservoirs (aquifers); b) appearance of open porosity or initiation of percolation instability; 3) Both SA and ULF magnetic field pulses are related to near-distant fractures (r<20-30 km); 4) Taking into account number rate and activation period of seismic, SA and ULF foreshocks, it is rather probable that opening of fractures and rupture of fluid reservoirs occur in the large preparation area with horizontal size about 100-200 km. © 2005 Author(s). This work is licensed under a Creative Commons License.

In order to extract any ULF signature associated with earthquakes, the principal component analysis (PCA) and singular spectral analysis (SSA) have been performed to investigate the possibility of discrimination of signals from different sources (geomagnetic variation, artificial noise, and the other sources (earthquake-related ULF emissions)). We adopt PCA to the time series data observed at closely separated stations, Seikoshi (SKS), Mochikoshi (MCK), and Kamo (KAM). In order to remove the most intense signal like the first principal component, we make the differential data sets of filtered 0.01Hz SKS-KAM and MCK-KAM in NS component and 0.01 Hz band. The major findings are as follows. (1) It is important to apply principal component analysis and singular spectral analysis simultaneously. SSA gives the structure of signals and the number of sensors for PCA is estimated. This makes the results convincing. (2) There is a significant advantage using PCA with differential data sets of filtered (0.01 Hz band) SKS-KAM and MCKKAM in NS component for removing the most intense signal like global variation (solar-terrestrial interaction). This provides that the anomalous changes in the second principal component appeared more sharply. And the contribution of the second principal component is 20-40%. It is large enough to prove mathematical accuracy of the signal. Further application is required to accumulate events. These facts demonstrate the possibility of monitoring the crustal activity by using the PCA and SSA. © 2005 Author(s). This work is licensed under a Creative Commons License.

Atmospheric electric field (AEF) observations and other concurrent measurements have been carried out in Taiwan over a three-year period in the context of seismo-electromagnetic studies. Generally, AEF variations appear to be due to meteorological effects. However, there seems to have been a correlation between AEF and the largest M6.8 earthquake on 31 March, 2002.

Despite its extreme importance and years of effort, practical short-term earthquake prediction still remains to be seen. However, research in earthquake-related electromagnetic phenomena have recently shown that such phenomena make for a promising candidate for short-term earthquake prediction. There is a good deal of accumulated evidence of precursory signatures in a wide frequency range (DC-VHF). ULF geomagnetic change is one of the most promising phenomena and it suggests that short-term prediction is realizable. This report first reviews earlier observational facts and presents the latest results in detection of ULF emissions by means of sophisticated signal processing. Finally, we discuss the possible generation and propagation mechanism of earthquake-related ULF signals.

An effective method is developed to identify and eliminate signals originated from external magnetic variations in the frequency range of 0.001 to 1 Hz in the geoelectric potential data. It is based on the interstation transfer function (ISTF) approach with the use of wavelet transform. Good performance has been found for the data at Kiyosumi station using Kakioka station as a reference, which means that almost whole of the inductive field could be eliminated successfully and other signals could be recognized with keeping their waveform. The detectability of simulated seismic electric signals (SES) has also been examined, and the result shows the successful discrimination of SES. These experimental facts indicate the effectiveness of our proposed signal reduction method. © 2004, The Institute of Electrical Engineers of Japan. All rights reserved.

Short-term earthquake prediction is still an urgent and important subject for seismo-active countries like Japan. There have been recently found a lot of convincing evidences on the presence of electromagnetic phenomena associated with earthquakes. One of the most promising candidates for short-term earthquake prediction is ultra-low-frequency (frequency less than 10Hz) (abbreviated as ULF) electromagnetic emission. We first review earlier observational facts, and then show our latest results, especially on the existence of such ULF emissions by means of sophisticated signal processing. Finally, we discuss the generation and propagation mechanisms of those seismo-ULF emissions. © 2004, The Institute of Electrical Engineers of Japan. All rights reserved.

We review the new findings on natural electromagnetic phenomena in the near-Earth environment and will show the importance of electromagnetic analyses in elucidating the essential points of these phenomena. The topics include (1) atmospheric phenomena related to lightning (e.g. mesospheric optical emissions); (2) seismo-electromagnetic phenomena (electromagnetic phenomena associated with earthquakes and/or volcano eruptions); (3) plasma and wave phenomena in the Earth's ionosphere and magnetosphere; and (4) electromagnetic or electrodynamic coupling among different regions. We pay our greatest attention to the unsolved essential problems for each subject, and suggest how electromagnetics would contribute to a solution to those problems. © 2004, The Institute of Electrical Engineers of Japan. All rights reserved.

We present evidence that the Hurst exponent of the ultra low frequency component of the geomagnetic field varies similarly to the energy release of the local swarm of earthquakes observed in Japan in 2000. Similarity is demonstrated between the energy release by the shocks and variations of the Hurst exponent derived for the three component records of the geomagnetic field. © 2004 Elsevier Ltd. All rights reserved.

Fractal analysis has been performed on ultra-low-frequency (ULF) geomagnetic field data observed at the Izu peninsula. The first attempt of using fractal analysis for ULF geomagnetic field data during the large Guam and Biak earthquakes has been based on FFT. In order to quantitatively estimate the FFT-based fractal analysis, we first study a few fractal analyses, including the Burlaga-Klein method and the Higuchi method, as compared with the former FFT-based method. The accuracies of these three methods have been evaluated by applying them to the test data-sets of fractional Brownian motion with white noise. We conclude that the Higuchi method is superior to the FFT-based method and the Burlaga-Klein method. The S/N effect was also discussed. Then, we have applied the Higuchi method to the ULF geomagnetic field data during a big earthquake swarm (during June-August 2000) in the Izu peninsula. It is found that the fractal dimension exhibits a significant increase just before earthquakes with a magnitude Mj>6.0 associated with the Izu islands swarm with any of these three fractal analyses. This experimental finding will lend further convincing support to the presence of precursory ULF emissions. © 2004 Elsevier Ltd. All rights reserved.

Electromagnetic phenomena associated with large earthquakes have been investigated to establish any method to monitor the crustal activity such as earthquakes and volcanic eruptions. The ULF geomagnetic approach is considered to be one of the most promising methods for the imminent prediction. In this paper, a new method of principal component analysis (PCA) has been applied to the ULF geomagnetic data associated with 2000 Izu Islands earthquake swarm. This swarm activity started on June 26, 2000 and terminated in September, 2000, and during this activity there were observed 5 large earthquakes whose magnitude was grater than 6 (July 1, 8, 15, 30, and August 18). During this period our ULF stations were fortunately in operation, and there are three stations closely distributed (about 5 km distances). The epicentral distances are about 80-100 km. The PCA has been applied to the ULF horizontal NS component, because we can distinguish a few noises sources based on the orthogonal decomposition in the PCA. We investigate the temporary variations of eigenvalues and eigenvectors of PCA results, and the results are summarized as follows. The first principal component is found to be the signal originated in the solar-terrestrial effects such as geomagnetic pulsation. The variation of eigenvector for the first principal component suggests that this signal is very stable over the whole analyzed period. As for the second principal component, the local artificial noise is included. As for the smallest third component in the local midnight at the Izu Peninsula, it indicated an apparent increase in the third eigenvalue a few days before the large earthquakes. Also about three months before the beginning of swarm activity, the level of the third eigenvalue was slightly enhanced. Correspondingly, the pattern of eigenvector direction in the signal subspace is changed simultaneously and it recovered to the original position after the swarm activity. These features are likely to be correlated with large earthquakes. Finally we want to emphasize that PCA approach is promising for monitoring the crustal activity. © 2004 Elsevier Ltd. All rights reserved.

Previous studies on Spitak, Loma Prieta, and Guam earthquakes indicate that large earthquakes were accompanied by preceding magnetic anomalies. In order to confirm these facts and to investigate ULF phenomena in details, a network of ULF magnetometers has been installed in Japan. Network observations including small arrays have been carried out. Also, the geomagnetic data observed at Matsukawa station associated with Iwateken Nairiku Hokubu earthquake (M6.1, September 3, 1998, depth 10 km) are presented. 4.5 years data have been analyzed and the obtained result is discussed. The variation of spectral density ratio between the horizontal and vertical components (polarization) exhibits an anomalous behavior two weeks before the earthquake. This is a unique change discovered from the rather long-term analysis, which suggests that this anomalous change might be a possible signal associated with the earthquake preparation phase. © 2004 Elsevier Ltd. All rights reserved.

In order to study earthquake-related ULF geomagnetic field changes, it is important to discriminate the noises such as magnetic pulsations originated from solar-terrestrial interactions and artificial noises from DC-driven trains and factories. For this aim, the interstation transfer functions and wavelet transform method have been proposed and applied to the data obtained at the ULF electromagnetic sensor array at the Boso Peninsula, Japan. It is concluded that this interstation transfer function approach has a capacity for signal discrimination; that is, it shows effectiveness to eliminate noises originated from the external ionospheric sources and its secondary effects. © 2004 Elsevier Ltd. All rights reserved.