- Email: [email protected]
ULF geomagnetic anomalies of possible seismogenic origin observed at Teoloyucan station, México, in 1999–2001: Intermediate and short-time analysis
Tectonophysics 431 (2007) 249 – 262 www.elsevier.com/locate/tecto
ULF geomagnetic anomalies of possible seismogenic origin observed at Teoloyucan station, México, in 1999–2001: Intermediate and short-time analysis A. Kotsarenko a,⁎, R. Pérez Enríquez a , J.A. López Cruz-Abeyro a , S. Koshevaya b , V. Grimalsky c , V. Yutsis d , I. Kremenetsky e a
d
Centro de Geociencias (CGEO), Juriquilla, UNAM, Apdo Postal 1-742, Centro Querétaro, Querétaro, México, C.P. 76001, Mexico b Universidad Nacional Autónoma de Morelos (UAEM), CIICAp, Cuernavaca, Morelos, Mexico c Instituto Nacional de Astrofísica, Óptica y Electrónica (INAOE), Tonantzintla, P.O.51 and 216, C.P. 72000, Puebla, Pue., México Universidad Autónoma de Nuevo León, Facultad de Ciencias de la Tierra, Car. A Cerro Prieto Km. 8 C.P. 67700, Linares, Nuevo León, México e Institute of Space Research (ISR), Nat.Acad.Sci (NASU), and Nat.Space Agency (NSAU), Ukraine Received 14 January 2006; accepted 25 May 2006 Available online 4 December 2006
Abstract The analysis of ULF geomagnetic field measured at Teoloyucan station (Central Mexico, 11′35.735″ W, 19 44′45.100″ N, 2280 m height) is presented in an intermediate (±15 days) and short time scale (the day of the EQ occurrence) in relation to 7 major earthquakes occurred in Mexico in 1999–2001. Local changes in the fractal dynamics of the magnetic field are found to be important: a pronounced fall of the fractal index is frequently observed prior to the main shock. The study of the ULF resonant structure recently discovered in the frequencies fR1 = 10.2−11.1 mHz and fR2 = 13.6−14.5 mHz reveals changes in their character probably related to the processes of the earthquakes preparation. The success of the observation of the mentioned anomalies (specially the fractal index decrease) strongly depends on how close is the station from the epicenter, and what is the magnitude of the earthquake. © 2006 Published by Elsevier B.V. Keywords: Geomagnetic field; Seismogenic phenomena; Natural electromagnetic resonance; Self-Organized Criticality (SOC); Fractal properties
1. Introduction Continuous monitoring in the areas with high tectonic activity reveals anomalies in a variety of geophysical parameters of quite different nature. Geochemical ⁎ Corresponding author. Tel.: +52 442 328 11 16 (+ 123); fax: +52 442 328 11 00. E-mail address: [email protected] (A. Kotsarenko). 0040-1951/$ - see front matter © 2006 Published by Elsevier B.V. doi:10.1016/j.tecto.2006.05.036
and geo-electrical changes of the subterranean water, such as a sharp pre-seismic depletion by dissolved isotopes of helium (3He, 4He); gradual increase of the water quantity and synchronous drop of its electrical conductivity observed in the different monitoring spots for moderate and strong earthquakes (EQs), such as those occurring in Armenia and Turkey in 1983–1999, as discussed in (Balasanyan, 2005). A wide family of atmospheric and thermal phenomena has been discovered by mean of both ground based
250
A. Kotsarenko et al. / Tectonophysics 431 (2007) 249–262
and satellite observations: anomalous changes in the coupled air parameters Humidity–Temperature (Dunajecka and Pulinets, 2005) and anomalous variations of the surface latent heat flux (SLHF, Dey and Singh, 2003) for the Colima 2003 earthquake; temperature increase up to 5 °C prior to earthquakes that occurred in Italy, Japan and China (Tramutoli et al., 2001; Tronin et al., 2002); and increase of land surface temperature and decrease of sea surface temperature close to the active tectonic fault (Ouzounov and Freund, 2004) before Gujarat earthquake (M7.7, January, 26, 2001). Electromagnetic (EM) phenomena observed at the time of strong seismic events are extensively studied worldwide in long-term investigations, registering a variety of phenomena detected by numerous passive and active measuring techniques. A number of results are accumulated in the comprehensive monographs edited by Hayakawa and Fujinawa (1994), Hayakawa (1999), and Hayakawa and Molchanov (2002). The most characteristic phenomena from ground-based observation on these subjects are probably: the anomaly subionospheric VLF wave propagation (Molchanov et al., 1998; Molchanov and Hayakawa, 1998), seismoelectrical anomaly signals (SES, or VAN method) discovered and studied in Greece by the group leaded by Varostos (Varotsos et al., 1994; Varotsos and Lazardou, 1999; Varotsos, 2001; Varotsos and Sarlis, 2002), anomaly TEC (total electron content) and foF2 critical frequency variation acquired by GPS measurements and ionosphere sounding (Pulinets and Boyarchuk, 2004). A wide spectrum of the EQ related EM phenomena observed during remote sensing satellite observations are summarized in the review by Pulinets (2006). Finally, geomagnetic phenomena focused on the interest of the scientific community since the STAR Lab team, Stanford Univ. (Fraser-Smith et al., 1990) and the team from St. Petersburg Dept. of IZMIRAN (Kopytenko et al., 1993) almost simultaneously reported
evidence for the anomaly ULF (Ultra Low Frequency) EM emission observed before two destructive earthquakes: Loma Prieta (Ms 7.1, October 17, 1989, USA) and Spitak (Ms 6.9, December 8, 1988, Armenia, USSR). The single-meaning theoretical explanation of the seismogenic electromagnetic phenomena is not definitely solved, but there are different models suggesting plausible mechanisms (Molchanov and Hayakawa, 1995; Vallianatos and Tzanis, 1999; Surkov et al., 2003; Pulinets and Boyarchuk, 2004), and mostlikely, all of them could occur together. The key feature in all the experimental studies on the geomagnetic phenomena is to distinguish between signals originating from tectonic activity, natural EM background and artificial EM noise. A variety of advanced methods aimed to resolve this problem have been recently proposed. The most successful efforts are probably: the discrimination signal from the noise using the transfer functions (Harada et al., 2004), the Principal Component Analysis (PCA) (Gotoh et al., 2002), and the location of the area of seismogenic geomagnetic disturbances (Ismaguilov et al., 2001; Kopytenko et al., 2001). Recently we reported some anomaly changes in the geomagnetic field character registered in 1999–2001 at the Teoloyucan station, Mexico, possibly related to the strongest EQs occurring in Mexico during this period. We performed the long-time analysis for the continuous part of the geomagnetic spectra by 2 methods: a multifrequency band study of the spectral values SH,D,Z and spectral ratio SZ/SH as a part of the traditional analysis, and a study of the spectral exponent (fractal index) β for the fractal analysis (Kotsarenko et al., 2004). We proved the existence of local geomagnetic pulsations, in the line spectrum structure, possibly generated by a crustal source (Kotsarenko et al., 2004), and noticeable longtime changes in recently discovered ULF resonant structure for the EQ times (Kotsarenko et al., 2005). In the present paper we complement our study through a
Table 1 The 7 earthquakes during 1999–2001 chosen for analysis Event
Year
Month
Day
Hour
Min
Longitude
Latitude
Magnitude, Ms
Depth, km
Distance, km
ρ, km
1 2 3 4 5 6 7
1999 1999 1999 2000 2001 2001 2001
6 6 9 8 5 5 10
15 21 30 9 19 20 7
20 17 16 11 23 4 21
42 43 31 41 21 21 39
− 97.51 − 101.72 − 97.03 − 102.66 − 105.72 − 105.12 − 100.16
18.18 17.99 15.95 17.99 18.27 18.64 16.98
7 5.8 7.5 6.5 6.5 6.0 6.1
69 54 16 16 20 12 10
262 343 481 428 736 663 322
1023 312 1678 624 624 380 420
Year/month/day/hour/min are: the exact time of the EQ; Latitude and Longitude: the geographic coordinates of the epicenter, magnitude and depth: magnitude and depth of the EQ, distance: the distance between epicenter and Teoloyucan station, ρ: is the radius of the EQ preparation zone estimated by Dobrovolsky formula. The EQ magnitudes are given in bold.
A. Kotsarenko et al. / Tectonophysics 431 (2007) 249–262
251
detailed analysis of the mentioned seismic events in the intermediate and short-time scale. 2. Physical conception and the methodology of the study The basic conception, applied for our experimental study, is the following. Geodynamical processes in the EQ preparation zone can produce current systems of different configuration (Molchanov and Hayakawa, 1995; Surkov et al., 2003). The ULF part of electromagnetic emission due to negligible skin depth (Kopytenko et al., 2001) can be recorded at the Earth's surface by modern magnetometers without significant attenuation. Higher frequencies would have higher skin depths and, therefore, greater attenuation before reaching the Earth's surface. Another source of EM emission can be generated at ionospheric heights due to the variability of the ionosphere. The evidence of the last proposition is supported by experimental observations (Tramutoli et al., 2001) and explained within different (but not contradictive) models: the conversion of atmospheric gravity waves (AGW) in the lower ionosphere (Molchanov, 2004) and the ionosphere modification caused by increased radon emanation before strong EQs (Pulinets and Boyarchuk, 2004). In order to select seismic events which can produce observable contribution to the geomagnetic signal we have used criteria satisfying both mentioned theoretical conceptions. It is assumed that the geomagnetic station should be inside or close to the EQ preparation zones: ρ = 100.43M km, where ρ is the radius of the preparation zone (Dobrovolsky et al., 1979). We have chosen 7 earthquakes for our analysis (see Table 1 for the EQ data, estimations on the EQ preparation zone radius ρ, and distances D between Teoloyucan (TEO) station and EQ epicenters, and Fig. 1 for the map of the EQ epicenters and location of TEO station). Recently, studies by Kotsarenko et al. (2004, 2005) performed for long-time intervals (year scale) challenged the necessity to perform a more detailed analysis for the mentioned EQs focusing on intermediate (± 15 days) and short time scales and concentrating on 2 subject: the dynamics of the fractal index and the behavior of the ULF RSR (Resonant Spectral Structure). According to the phenomenological model proposed for strong earthquakes as a Self-Organized Criticality (SOC) process (Kopytenko et al., 2001), the state of the medium during the EQ preparation passed through different stages: initial stage (random chaos) → subcritical → critical → supercritical stage. As a critical process, the EQ output parameters manifest fractal
Fig. 1. Geomagnetic field emphasize fractal properties: its power spectra fits the inverse power law equation S( f ) ∼f − β. FFT (Fast Fourier Transform) method: best linear fit approximation for presentation log(S)vs-log( f ), fractal index β is calculated as the absolute value of the line declination coefficient.
properties both in time and space scale. One of them, the geomagnetic field spectra, estimated by inverse power law as S( f ) ∼f −β, is indicative of the dynamical processes contributing to geomagnetic field changes during the time of the EQ in terms of the EQ preparation and the postseismic relaxation. Generally, the fractal index dynamics reveals an expressed 27-day variation, connected with a rotational period of the Sun (Hayakawa et al., 1999). The gradual decrease of the fractal index before the EQ and the intense depression of the solar period variation observed in 2 post-seismic months were reported (Hayakawa et al., 1999; Kopytenko et al., 2001) in relation to Guam EQ, 1993, Ms = 8.0. The calculation of the fractal index can be made by different methods, described by Gotoh et al. (2004). We used the FFT (Fast Fourier Transform) method and calculated β as a best linear fit approximation log(S)-vs-log( f ), where S is a spectral density of the signal and f is the frequency (Fig. 1). Another subject of interest, the ULF structure, has been recently discovered in the Teoloyucan magnetic station data (Kotsarenko et al., 2005). Two resonant lines were observed in the H-component (S–N direction, linear polarization) in the frequency bands fR1 = 10.2−11.1 mHz and fR2 = 13.6−14.5 mHz, sometimes accompanied by sporadically generated harmonics (Fig. 2). Following our provisional considerations, the source of the observed resonances is located in the ionosphere and the mechanism of its generation is connected with a geomagnetic
252
A. Kotsarenko et al. / Tectonophysics 431 (2007) 249–262
Fig. 2. ULF geomagnetic resonances observed in the H-component (linearly polarized) in Teoloyucan station, March 30–April 26, 1999. Top panel (a): Equatorial DST index of the geomagnetic activity. Panel (c): SRS (Spectral Resonant Structure) dynamics in the Teoloyucan (TEO) station. fR1, fR2 are 2 dominating resonant frequency bands, fSH-sporadically generated (lower) harmonic. Panels (b) and (d): Referring spectra for Boulder (BLD) and Jicamarca (JIC) stations accordingly. The coordinates of stations are given in the Table 2.
location of the station (geomagnetic latitude λ = 29.1°) close to the latitude of the equatorial electrojet system. The H-component of the magnetic field is chosen for analysis in the presented study, where spectrograms are calculated with NFFT = 213 (Number of points for FFT transform, corresponding time domain T = 2 h 16 min 32 s) to achieve the optimal resolution in the analyzed interval of frequencies f = 0–30 mHz (considered most promising for observations of the seismogenic ULF geomagnetic emission, see Molchanov and Hayakawa (1995), Ismagulov et al. (2001), Kopytenko et al. (2001) for explanation/argumentation). The fractal index β is calculated every hour (NFFT = 3600, time domain T = 1 h). We use also a 6-hours running average for tracing daily variations and 24-hours running average (trend) to see the general behavior of the β.
The analyzed data set is the same as the one used for the recent analysis by Kotsarenko et al. (2004, 2005): the geomagnetic data were recorded at Teoloyucan station (Central Mexico, geographic coordinates: 99 11′ 35.735″ W, 19 44′45.100″ N, 2280 m height). This station was equipped with a 3-component fluxgate magnetometer designed at UCLA, operating at 1 Hz sampling rate frequency, with a GPS system for data synchronization. In our study we used the DST index obtained at Kyoto World Data Center for Geomagnetism (web page) discarding events occurring during high geomagnetic activity from our analysis. In order to distinguish the local character of the phenomena from the global one compared our results with those ones calculated for reference stations from the CANOPUS and Mid-continent
A. Kotsarenko et al. / Tectonophysics 431 (2007) 249–262
253
Fig. 3. Location of Teoloyucan and referring stations (see details in Table 2).
Magneto-seismic Chain (McMAC) network (Chi et al., 2005) equipped with the same type of instruments. 3. Description and discussion of the obtained results Anticipating the analysis of the mentioned 7 seismic events in terms of the dynamics of fractal index and behavior of ULF geomagnetic resonances we should add some technical preface for the validation of the results obtained in our study. The mentioned SRS (Resonance Spectral Structure) was observed for the first time in 2005 in the dominant frequency bands f R1 = 10.2−11.1 and f R2 = 13.6− 14.5 mHz only at Teoloyucan geomagnetic station
(Kotsarenko et al., 2005) and was not detected at other stations in the American continent (Fig. 2, panel c). However, a very similar resonance structure was observed at Beijing (BJI) geomagnetic station (almost the same geomagnetic latitude as Teoloyucan) but was absent even at the nearest Chinese station (situated at about 150 km from Beijing) so leading to the conclusion that it corresponds to a very special localization (channeling character) of the mentioned phenomena, generated with a high probability by the equatorial electrojet system (Kotsarenko et al., 2005). The preliminary qualitative estimations show a good correlation of the intensities of the resonant structure with an Sq variation and its moderate dependence on the solar
Table 2 Geographic and geomagnetic coordinates for Teoloyucan and referring stations Station name
Abbreviation
Latitude Geography
Longitude Geography
Latitude Magnitude
Longitude Magnitude
L value
Athabasca Boulder Teoloyucan Jicamarca
ATH BLD TEO JIC
54.7 40.1 19.7 −12.0
246.7 254.8 260.8 283.1
62.3 49.1 29.1 0.0
305.6 319.6 330.1 354.2
4.64 2.34 1.31 1.00
254
A. Kotsarenko et al. / Tectonophysics 431 (2007) 249–262
Fig. 4. Dynamics of fractal index β for the period of the EQ 1 (Ms 7.0, June 15, 1999). Panel (c): Teoloyucan station, Panels (a, b and d): referring stations: Athabasca (ATH), Boulder (BLD) and Jicamarca (JIC) stations accordingly. Bottom panel (e): Equatorial DST index of the geomagnetic activity. The coordinates of stations are given in the Table 2.
Fig. 5. Map of Mexico. Position of the Teoloyucan magnetic station is marked by triangle. The epicenters of 7 seismic events are displayed as hexagonal stars. Description of the events is given in Table 1.
A. Kotsarenko et al. / Tectonophysics 431 (2007) 249–262
activity (Dst index) which is good in terms of the natural character of the phenomena. However, the principal obstacles (difficulties) for the quantitative analysis are still the insufficient sample rate of the data, very low signal/noise ratio and imperfection of our techniques. Therefore, in this paper we present only qualitative analysis of the changes in the SRS in possible relation to 7 strong and moderate seismic events that occurred in Mexico during 1999–2001. In them, the referring behavior of the SRS (given by Fig. 2, panel c) was calculated for the analyzed interval March
255
30–April 26, 1999 which was “seismically quiet” (1 week out of the any seismic shock with magnitude higher the Ms = 5). The behavior of both resonant modes show that during the strongest geomagnetic storm the SR structure is depressed (Fig. 2c, area marked by circle 1) and saturated by the perturbed spectra (dark areas) while at the time of the smaller geomagnetic storms the depression and saturation are much smaller (Fig. 2c, circles 1 and 2). Another subject of our study, the fractal property of geomagnetic field (confirmed by already mentioned
Fig. 6. EQ 1 (Ms 7.0, June 15, 1999) and EQ 2 (Ms 5.8, June 21, 1999), Teoloyucan station results. Panel (a): Geomagnetic field, H-component; Panel (b): Power spectra sonogram for the H-component; Panel (c): Equatorial DST index of the geomagnetic activity; Panel (d): Temporal distribution of the foreand aftershock Magnitudes (Ms); Pentagonal stars: shocks in the 1st EQ zone; Asterisks: shocks in the 2nd EQ zone; Main shocks are marked by hexagonal stars; Panel (e): Temporal distribution of the fore- and aftershock depths symbols are the same as in (d); Panel (f): Fractal index β temporal dynamics.
256
A. Kotsarenko et al. / Tectonophysics 431 (2007) 249–262
numerous studies) is analyzed in quantitative form: the power spectra (or fractal) tendency S(f) ∼ f −β is approximated by a best linear fit and fractal index β is calculated as the absolute value of the line inclination coefficient (Fig. 2). In order to distinguish the local character of the fractal index β dynamics from the global changes, we compare our results with a reference β calculated for the stations located at different geomagnetic latitudes (see the map in Fig. 3 and Table 2 for details). Of course, the fractal dynamics of the geomagnetic field will be different for different stations because of the difference of physical processes in the corresponding magneto-
sphere cavities (the L-values are quite different for every station of the station set). For EQ 1, we found a significant drop of the fractal index in the Teoloyucan β close to the time of the event (Fig. 4) compared to the corresponding monotonous changes of referent stations series thus indicating that it is a local anomaly. The observed phenomena may be interpreted as due to a high contribution of the extended ionosphere perturbed by the EQ preparation process throw the lithosphere– ionosphere–magnetosphere coupling. The global reduction of the fractal index at other days may be caused by a sharp change in the ring current system: the Dst value (both direction and intensity) for the equatorial and low-
Fig. 7. EQ 3 (Ms 7.5, September 30, 1999). The table of symbols is the same as in Fig. 6.
A. Kotsarenko et al. / Tectonophysics 431 (2007) 249–262
latitudinal stations (for the high-latitudinal stations such as Athabasca (ATH) the polar currents (AE index) will be also important). In our analysis and in the presented figures we combined 1st and 2nd EQs as they occurred in a close time interval (6 days). We did the same for the 5th and 6th events (one day difference). 3.1. Tehuacan Ms 7.0 (June, 15, 1999) and Guerrero Ms 5.8 (June, 21, 1999) earthquakes The June 15, 1999, magnitude 7.0 earthquake occurred in the central region of México (EQ1, Fig. 5), at 15:42 h (20:42 GMT), killing 16 person and injuring more then 200. The epicenter of the earthquake was located near the border of the states of Puebla and Oaxaca, about 20 km southeast of Tehuacan, and about 230 km east-southeast of Mexico City. A second strong event (EQ2 at Fig. 5) with magnitude 5.8 (Mw = 6.3 according USGS) occurred on June 21 at 12:43 h (17:43 GMT) in the coast of Guerrero state. Both earthquakes, especially the 1st one, damaged a number of buildings, walls, bridges and communications, disabled electricity and telephone services, the first EQ even gained a name
257
of “church” earthquake for the destruction of temples in Puebla state. The 1st EQ was preceded by a 2 small foreshocks, one at a small depth and one almost superficial; several aftershocks were clustered in 2 very short time intervals: some hours just after the main shocks and about 1 day after it (Fig. 6, panels d, e). The geomagnetic conditions are very favorable for the analyzed interval (Fig. 6, panel c): the DST index was mostly positive never dropping lower than − 40 gammas for more than 13 days (+ 13 days) after the EQ. Our principal results obtained for the EQ 1 are the following. We observe periodical post-seismic depression of both resonant modes especially clear at the period +2 to + 9 days after the 1st event (Fig. 6b, area marked by an ellipse). Due to lack of data (2 days of the record lost starting some hours after the main shock) we cannot say exactly about the time when these depressions started. The character of the fractal index shows a noticeable fall 2 days before the 1st event (Fig. 6f, area marked by circle), whereas the geomagnetic activity (Fig. 6c) display a very weak character. Neither can we estimate the character of its recovery during the days + 1–2 days.
Fig. 8. Spectral Resonant Structure observed on 21–22 of September 1999 in the H- and D-components in the Teoloyucan station 8 days before the EQ 3 (Ms 7.5, September 30, 1999). Panels (a–c): H-, D-, and Z-components. Panel (d): Equatorial DST index of the geomagnetic activity.
258
A. Kotsarenko et al. / Tectonophysics 431 (2007) 249–262
The observed anomaly drop has a local character: as it was mentioned before (explanations for the fractal dynamics analysis presented in Fig. 4) no evident changes were observed for the referring geomagnetic stations. In relation to the second EQ we cannot emphasize any expressed feature directly associated with this event in the intermediate scale. 3.2. Oaxaca Ms 7.5 (September, 30, 1999) earthquake A large earthquake struck Oaxaca State, Mexico earthquake at 11:31 Mexico time (16:31 GMT) at Mexican
beach resort of Puerto Escondido (about 60 km from Puerto Angel, 440 km southern México City, see EQ 3 at Fig. 5), killing 33 people, injuring 215, destroying and making damages to thousands of buildings, disrupting utilities and blocking roads by landslides in the state of Oaxaca. The EQ was preceded by a dozen foreshocks and followed by numerous aftershocks which also revealed a tendency to group into short-time clusters (Fig. 7d, e): 3 foreshocks at − 2 days (occurred during about 4 hours time interval), 6 events just after the main shock (12 h), 5 shocks at + 1 day (6 h), 5 events at +2 days (6 h) and 3 events at + 14 days (6 h). Another group of 7 events occurred at +3 to +4 in the interval of 27 h.
Fig. 9. EQ 4 (Ms 6.5, August 9, 2000). The table of symbols is the same as in Fig. 6.
A. Kotsarenko et al. / Tectonophysics 431 (2007) 249–262
The Sq-variation decreased and almost disapeared during various long periods; some of them, like a preseismic depression − 15–14 days and + 10–15 days can be explained by a perturbed geomagnetic activity (small storms not lower than − 50–60 gammas), another postseismic depression +1–7 days possess the anomaly character (Fig. 7a, marked by ellipse). The fractal index displays a tendency to decrease 2 days before the EQ (Fig. 7g, marked by circle 1) which is a regional change. Both resonant modes at the analyzed period are too week among the noise level so we cannot postulate any reliable changes in them. But another curious event occurred 8 days before the main shock (Fig. 7b, marked
259
by ellipse) and is presented in detail in Fig. 8. An SRS (Spectral Resonant Structure) polarized in H- and Dcomponents (Fig. 8, panels a and b, area marked by ellipse) was generated approximately at 18 UT h., September 21, 1999 during quiet geomagnetic conditions (monotonously lowered down to the peak − 40 nT at time 2 UT, September 22, Fig. 8d) and disappeared at 13 UT, September 22 just at the time when the Dst index went drastically from zero level to + 20 nT. Due to the luck of the preceding 2-day interval of the data we cannot postulate now the reliability of the related drop of the regional fractal index (Fig. 8f, marked by circle 2). A more detailed analysis for β behavior with a precise
Fig. 10. EQ 5 (Ms 6.5, May 19, 2001) and EQ 6 (Ms 6.0, May 20, 2001). The table of symbols is the same as in Fig. 6.
260
A. Kotsarenko et al. / Tectonophysics 431 (2007) 249–262
reconstruction of the mentioned fragment is expected in the near future. 3.3. Lazaro Cardenas Ms 6.5 (August 9, 2000) earthquake The earthquake centered near Mexico's Pacific Ocean coast shook western and central Mexico early Wednesday at 6:41 a.m. (11:41 GTM, EQ 4 at Fig. 5), cracking walls, sending people racing out of their homes and injuring at least one person when parts of a wall collapsed in a hotel under repair in the coastal city of Lazaro Cardenas, about 30 mi (48 km) from the quake's epicenter.
This EQ occurred with no foreshocks (Fig. 9d, e), but several aftershocks occurred in 2 very short time clusters (dispersal 1–2 h) at the day of the main event, and the main part of the aftershock activity was detected during +3–5 days almost uniformly in time. For this earthquake most of the changes in the fractal dynamics use to be global (no unusual features in a regional scale). Both the 1st and 2nd resonant modes tend to decrease slightly at the pre- and post-seismic interval − 2 to + 2d (Fig. 9b, marked by a circle), with generation of a sporadic lower harmonic detected at the frequencies Δf ∼ 0.6–9 mHz. The + 2d and the following interval is discarded from the analysis due
Fig. 11. EQ 7 (Ms 6.1, October 7, 2001). The table of symbols is the same as in Fig. 6.
A. Kotsarenko et al. / Tectonophysics 431 (2007) 249–262
to the occurrence of strong geomagnetic storms (− 100 nT and − 150 nT, Fig. 9c). 3.4. Jalisco Ms 6.5 (May 19, 2001) and Ms 6.0 (May 20, 2001) earthquakes Both analyzed events occurred in the undersea tectonic zone, with no damage to human beings nor to buildings being reported (see EQ 5 and 6 at Fig. 5). The second event, taken from the SSN (National Seismological Service) database, can be erroneous with a high probability as no reports confirming its occurrence was discovered in international data bases. The EQs were preceded by several foreshocks, several aftershocks occurred after the main shock, and some shocks took place at the post-seismic 15-day interval. Nothing anomalous is detected for these events in geomagnetic field, neither fractal dynamics nor resonant modes behavior. The changes in the H-component and a week depression of both resonances at the period − 12–3 days should be caused by geomagnetic activity, but both resonant modes show smearing (thickening) tendency at the post-seismic period 0 to +4d (Fig. 10b, area marked by circle). In other words, we observe the decrease of the quality Q factor for both resonant modes. 3.5. Guerrero Ms 6.1 (October 7, 2001) earthquake The earthquake occurred in Guerrero State (EQ 7 at Fig. 5), no damages or human losses being reported. The USGS catalog gives a quite different magnitude Ms = 5.5 and depth D = 33 km for the mentioned EQ. The most interesting feature of that EQ was that that in spite of not being of high magnitude it triggered an intensive seismic activity not only in the regional seismic zone but all over Mexico and Guatemala (Fig. 11d, e). The interval − 14–2 days has been discarded from the analysis because of the high geomagnetic activity present. We observe a small regional fall of the fractal index in the near-earthquake time interval ±1 day (Fig. 10 f, area marked by circle), the rest of the relative changes found to be global and caused by geomagnetic storms. Both resonant modes (fR2, higher mode) are almost depressed among the noise for most of the analyzed period (especially the higher mode fR2). 4. Conclusions The results obtained for the 7 analyzed earthquakes can be summarized as follows:
261
1. A decrease of the regional part of fractal index of the geomagnetic field β is observed for events (EQs) 1, 3 and 7 starting about 1–3 days before the major seismic events, and was not detected (“failed”) for the event 4 (in spite of the fact that also satisfies the Dobrovolsky criteria ρ N D) as well as for events 2, 5 and 6 (when the station was out of the EQ preparation zone estimated by Dobrovoslky formula). For the rest of the observed anomalies (EQs 1, 3 and 7) the “magnitude” of the respective fall of the fractal index (in terms of the relation of the value for the observed near-seismic decrease to levels of the “natural” daily and monthly variation) decreases with the ρ/D relation (more attenuated). Actually, the most pronounced drop of the fractal index was observed for the case of the Tehuacan EQ (EQ 1, Ms 7.0, June 15, 1999) with an estimated EQ preparation zone ρ = 1023 km and distanced D = 262 km (ρ/D = 3.9), that can be considered as a are very favorable condition for the observation (lowest attenuation). 2. The ULF resonant modes of the geomagnetic field also undergo pre-seismic (event 3), near-seismic (event 4) and post-seismic depressions (events 1 and 4) and smearing of the resonance structure (the decrease of the SRS quality Q factor, events 5 and 6) possibly related to the mentioned earthquakes. “Possible” because the level of confidence of the observed phenomena is lower and future studies are required for the validation of the anomaly character of the seismo-related changes of the SRS. 3. The unusual multi-frequency SRS (fine structure) was observed in the H-and D-components 8 days prior to the strongest of the analyzed events (EQ 3, Ms 7.5, September 30, 1999). The mentioned SRS was generated under relatively quiet (Dst N − 40 nT) geomagnetic conditions and terminated just at the time of the change of the direction (and intensity) of the equatorial ring current system (positive Dst value). Acknowledgements We are grateful to two anonymous referees for their constructive criticisms, useful corrections and comments, which allowed us to improve the quality of this paper and stimulates our understanding of the reported phenomena. The authors (a) also grateful to the internal UNAM foundation DGAPA and Mexican government foundation CONACyT for the partial support of this work by the projects PAPIIT IN117106 and CONACyT 47662, respectively. We also thank Peter Chi for providing us with the referring McMAC data.
262
A. Kotsarenko et al. / Tectonophysics 431 (2007) 249–262
References Balasanyan, S.Yu., 2005. High-sensitivity Energo-active Points of the Earth (phenomenon HEP): a key to the future understanding of the physics of the earthquakes? Geol. Geophys. 46, 83–99 (in Russian). Chi, P.J., Engebretson, M.J., Moldwin, M.B., Russell, C.T., Mann, I.R., Samson, J.C., López Cruz-Abeyro, J.A., Yumoto, K., Lee, D.-H., 2005. Mid-continent Magnetoseismic Chain (McMAC): A Meridional Magnetometer Chain for Magnetospheric Sounding. Proceedings of the Enviroment Modeling Workshop, June 17–22, 2005, Snowmass, Colorado, USA. Dey, S., Singh, R.P., 2003. Surface latent heat flux as an earthquake precursor. Nat. Haz. Earth Syst. Sci. (NHESS) 3, 749–755. Dobrovolsky, I.R., Zubkov, S.I., Myachkin, V.I., 1979. Estimation of the size of earthquake preparation zones. PAGEOPH 117, 1025–1044. Dunajecka, M.A., Pulinets, S.A., 2005. Atmospheric and thermal anomalies observed around the time of strong earthquakes in Mexico. Atmósfera 18 (4), 235–247. Fraser-Smith, A.C., Bernardy, A., McGill, P.R., et al., 1990. Low frequency magnetic field measurements near the epicenter of the Loma Prieta earthquake. Geophys. Res. Lett. 17, 1465–1468. Gotoh, K., Akinaga, Y., Hayakawa, M., Hattori, K., 2002. Principal component analysis of ULF geomagnetic data for Izu islands earthquakes in July 2000. J. Atmos. Electr. 22, 1–12. Gotoh, K., Hayakawa, M., Smirnova, N., Hattori, K., 2004. Fractal analysis of seismogenic ULF emissions. Phys. Chem. Earth 29, 419–424. Harada, M., Hattori, K., Isezaki, N., 2004. Transfer function approach to signal discrimination of ULF geomagnetic data. Phys. Chem. Earth 29, 409–417. Hayakawa, M. (Ed.), 1999a. Atmospheric and Ionospheric Electromagnetic Phenomena Associated with Earthquakes. TERAPPUB, Tokyo. Hayakawa, M., Fujinawa, Y. (Eds.), 1994. Electromagnetic Phenomena Related to Earthquake Prediction. Terra Sci. Pub. Co, Tokyo. Hayakawa, M., Molchanov, O.A. (Eds.), 2002. Seismo Electromagnetics: Lithosphere–Atmosphere–Ionosphere Coupling. TERRAPUB, Tokyo. Hayakawa, M., Ito, T., Smirnova, N., 1999b. Fractal analysis of ULF geomagnetic data associated with the Guam earthquake on 8 August 1993. Geophys. Res. Lett. 26, 2797–2800. Ismaguilov, V.S., Kopytenko, Yu.A., Hattori, K., Voronov, P.M., Molchanov, O.A., Hayakawa, M., 2001. ULF magnetic emissions connected with under sea bottom Earthquakes. Nat. Hazards Earth Syst. Sci. (NHESS) 1, 23–31. Kopytenko, Yu.A., Matiashvili, T.G., Voronov, P.M., Kopytenko, E.A., Molchanov, O.A., 1993. Detection of ultra-low-frequency emissions connected with the Spitak earthquake and its aftershock activity, based on geomagnetic pulsations data at Dusheti and Vardzia observatories. Phys. Earth Planet. Inter. 77 (1–2), 85–95. Kopytenko, Yu., Ismaguilov, V., Hayakawa, M., Smirnova, N., Troyan, V., Peterson, Th., 2001. Investigation of the ULF electromagnetic phenomena related to earthquakes: contemporary achievements and the perspectives. Ann. Geofis. 44 (2), 325–334. Kotsarenko, A., Pérez Enríquez, R., López Cruz-Abeyro, J.A., Koshova, S., Grimalsky, V., 2004. Analysis of the ULF
electromagnetic emission related to seismic activity, Teoloyucan geomagnetic station, 1998–2001. Nat. Hazards Earth Syst. Sci. (NHESS) 4, 679–684. Kotsarenko, A., Pérez Enríquez, R., López Cruz-Abeyro, J.A., Koshevaya, S., Grimalsky, V., Kremenetsky, I., 2005. Possible seismogenic origin of the changes in the ULF EM resonant structure observed at Teoloyucan geomagnetic station, Mexico, 1999–2001. Nat. Hazards Earth Syst. Sci. (NHESS) 4, 679–684. Molchanov, O.A., 2004. On the origin of low- and middle-latitude ionospheric turbulence. Phys. Chem. Earth 29, 559–556. Molchanov, O.A., Hayakawa, M., 1995. Generation of ULF electromagnetic emissions by microfracturing. Geophys. Res. Lett. 22, 3091–3094. Molchanov, O.A., Hayakawa, M., 1998. Subionospheric VLF signal perturbations, possibly related to earthquakes. J. Geophys. Res. 103, 17489–17504. Molchanov, O.A., Hayakawa, M., Ondoh, T., Kawai, E., 1998. Precursory effects in the subionospheric VLF signals for the Kobe earthquake. Phys. Earth Planet. Inter. 105, 239–248. Ouzounov, D., Freund, F., 2004. Mid-infrared emission prior to strong earthquakes analyzed by remote sensing data. Adv. Space Res. 33, 268–273. Pulinets, S.A., 2006. Space technologies for short-term earthquake warning. Adv. Space Res. 37 (4), 643–652. Pulinets, S.A., Boyarchuk, K.A., 2004. Ionospheric Precursors of Earthquakes. Springer, Berlin. 316 pp. Surkov, V.V., Molchanov, O.A., Hayakawa, M., 2003. Preearthquake ULF electromagnetic perturbations as a result of inductive seismomagnetic phenomena during microfracturing. J. Atmos. Sol.-Terr. Phys. 65, 31–46. Tronin, A.A., Hayakawa, M., Molchanov, O.A., 2002. Thermal IR satellite data application for earthquake research in Japan and China. J. Geodyn. 33, 519–534. Tramutoli, V., Di Bello, G., Pergola, N., Piscitelli, S., 2001. Robust satellite techniques for remote sensing of seismically active areas. Ann. Geofis. 44, 295–312. Vallianatos, F., Tzanis, A., 1999. A model for the generation of precursory electric and magnetic fields associated with the deformation rate of the earthquake focus. In: Hayakawa, M. (Ed.), Atmospheric and Ionospheric Electromagnetic Phenomena Related with Earthquakes. TERRAPUB, Tokyo, pp. 287–305. Varotsos, P., 2001. A review and analysis of electromagnetic precursory phenomena. Acta Geophys. Pol. 49, 1–42. Varotsos, P., Lazardou, M., 1999. A review of the VAN method. In: Hayakawa, M. (Ed.), Atmospheric and Ionospheric Electromagnetic Phenomena Associated with Earthquakes. TERRAPUB, Tokyo, pp. 47–53. Varostos, P., Sarlis, N., 2002. A review of recent VAN efforts: the explanations of SES physical properties. In: Hayakawa, M., Molchanov, O.A. (Eds.), Seismo Electromagnetics: Lithosphere – Atmosphere – Ionosphere Coupling. TERRAPUB, Tokyo, pp. 131–140. Varotsos, P., Ueda, S., Alexopoulos, K., Nagao, T., Lazaridou, M., 1994. Prediction of recent destructive seismic activities in Greece based on seismic electric signals. In: Hayakawa, M., Fujinawa, Y. (Eds.), Electromagnetic Phenomena Related to Earthquake Prediction. TERRAPUB, Tokyo, pp. 13–25.
ULF geomagnetic anomalies of possible seismogenic origin observed at Teoloyucan station, México, in 1999–2001: Intermediate and short-time analysis A. Kotsarenko a,⁎, R. Pérez Enríquez a , J.A. López Cruz-Abeyro a , S. Koshevaya b , V. Grimalsky c , V. Yutsis d , I. Kremenetsky e a
d
Centro de Geociencias (CGEO), Juriquilla, UNAM, Apdo Postal 1-742, Centro Querétaro, Querétaro, México, C.P. 76001, Mexico b Universidad Nacional Autónoma de Morelos (UAEM), CIICAp, Cuernavaca, Morelos, Mexico c Instituto Nacional de Astrofísica, Óptica y Electrónica (INAOE), Tonantzintla, P.O.51 and 216, C.P. 72000, Puebla, Pue., México Universidad Autónoma de Nuevo León, Facultad de Ciencias de la Tierra, Car. A Cerro Prieto Km. 8 C.P. 67700, Linares, Nuevo León, México e Institute of Space Research (ISR), Nat.Acad.Sci (NASU), and Nat.Space Agency (NSAU), Ukraine Received 14 January 2006; accepted 25 May 2006 Available online 4 December 2006
Abstract The analysis of ULF geomagnetic field measured at Teoloyucan station (Central Mexico, 11′35.735″ W, 19 44′45.100″ N, 2280 m height) is presented in an intermediate (±15 days) and short time scale (the day of the EQ occurrence) in relation to 7 major earthquakes occurred in Mexico in 1999–2001. Local changes in the fractal dynamics of the magnetic field are found to be important: a pronounced fall of the fractal index is frequently observed prior to the main shock. The study of the ULF resonant structure recently discovered in the frequencies fR1 = 10.2−11.1 mHz and fR2 = 13.6−14.5 mHz reveals changes in their character probably related to the processes of the earthquakes preparation. The success of the observation of the mentioned anomalies (specially the fractal index decrease) strongly depends on how close is the station from the epicenter, and what is the magnitude of the earthquake. © 2006 Published by Elsevier B.V. Keywords: Geomagnetic field; Seismogenic phenomena; Natural electromagnetic resonance; Self-Organized Criticality (SOC); Fractal properties
1. Introduction Continuous monitoring in the areas with high tectonic activity reveals anomalies in a variety of geophysical parameters of quite different nature. Geochemical ⁎ Corresponding author. Tel.: +52 442 328 11 16 (+ 123); fax: +52 442 328 11 00. E-mail address: [email protected] (A. Kotsarenko). 0040-1951/$ - see front matter © 2006 Published by Elsevier B.V. doi:10.1016/j.tecto.2006.05.036
and geo-electrical changes of the subterranean water, such as a sharp pre-seismic depletion by dissolved isotopes of helium (3He, 4He); gradual increase of the water quantity and synchronous drop of its electrical conductivity observed in the different monitoring spots for moderate and strong earthquakes (EQs), such as those occurring in Armenia and Turkey in 1983–1999, as discussed in (Balasanyan, 2005). A wide family of atmospheric and thermal phenomena has been discovered by mean of both ground based
250
A. Kotsarenko et al. / Tectonophysics 431 (2007) 249–262
and satellite observations: anomalous changes in the coupled air parameters Humidity–Temperature (Dunajecka and Pulinets, 2005) and anomalous variations of the surface latent heat flux (SLHF, Dey and Singh, 2003) for the Colima 2003 earthquake; temperature increase up to 5 °C prior to earthquakes that occurred in Italy, Japan and China (Tramutoli et al., 2001; Tronin et al., 2002); and increase of land surface temperature and decrease of sea surface temperature close to the active tectonic fault (Ouzounov and Freund, 2004) before Gujarat earthquake (M7.7, January, 26, 2001). Electromagnetic (EM) phenomena observed at the time of strong seismic events are extensively studied worldwide in long-term investigations, registering a variety of phenomena detected by numerous passive and active measuring techniques. A number of results are accumulated in the comprehensive monographs edited by Hayakawa and Fujinawa (1994), Hayakawa (1999), and Hayakawa and Molchanov (2002). The most characteristic phenomena from ground-based observation on these subjects are probably: the anomaly subionospheric VLF wave propagation (Molchanov et al., 1998; Molchanov and Hayakawa, 1998), seismoelectrical anomaly signals (SES, or VAN method) discovered and studied in Greece by the group leaded by Varostos (Varotsos et al., 1994; Varotsos and Lazardou, 1999; Varotsos, 2001; Varotsos and Sarlis, 2002), anomaly TEC (total electron content) and foF2 critical frequency variation acquired by GPS measurements and ionosphere sounding (Pulinets and Boyarchuk, 2004). A wide spectrum of the EQ related EM phenomena observed during remote sensing satellite observations are summarized in the review by Pulinets (2006). Finally, geomagnetic phenomena focused on the interest of the scientific community since the STAR Lab team, Stanford Univ. (Fraser-Smith et al., 1990) and the team from St. Petersburg Dept. of IZMIRAN (Kopytenko et al., 1993) almost simultaneously reported
evidence for the anomaly ULF (Ultra Low Frequency) EM emission observed before two destructive earthquakes: Loma Prieta (Ms 7.1, October 17, 1989, USA) and Spitak (Ms 6.9, December 8, 1988, Armenia, USSR). The single-meaning theoretical explanation of the seismogenic electromagnetic phenomena is not definitely solved, but there are different models suggesting plausible mechanisms (Molchanov and Hayakawa, 1995; Vallianatos and Tzanis, 1999; Surkov et al., 2003; Pulinets and Boyarchuk, 2004), and mostlikely, all of them could occur together. The key feature in all the experimental studies on the geomagnetic phenomena is to distinguish between signals originating from tectonic activity, natural EM background and artificial EM noise. A variety of advanced methods aimed to resolve this problem have been recently proposed. The most successful efforts are probably: the discrimination signal from the noise using the transfer functions (Harada et al., 2004), the Principal Component Analysis (PCA) (Gotoh et al., 2002), and the location of the area of seismogenic geomagnetic disturbances (Ismaguilov et al., 2001; Kopytenko et al., 2001). Recently we reported some anomaly changes in the geomagnetic field character registered in 1999–2001 at the Teoloyucan station, Mexico, possibly related to the strongest EQs occurring in Mexico during this period. We performed the long-time analysis for the continuous part of the geomagnetic spectra by 2 methods: a multifrequency band study of the spectral values SH,D,Z and spectral ratio SZ/SH as a part of the traditional analysis, and a study of the spectral exponent (fractal index) β for the fractal analysis (Kotsarenko et al., 2004). We proved the existence of local geomagnetic pulsations, in the line spectrum structure, possibly generated by a crustal source (Kotsarenko et al., 2004), and noticeable longtime changes in recently discovered ULF resonant structure for the EQ times (Kotsarenko et al., 2005). In the present paper we complement our study through a
Table 1 The 7 earthquakes during 1999–2001 chosen for analysis Event
Year
Month
Day
Hour
Min
Longitude
Latitude
Magnitude, Ms
Depth, km
Distance, km
ρ, km
1 2 3 4 5 6 7
1999 1999 1999 2000 2001 2001 2001
6 6 9 8 5 5 10
15 21 30 9 19 20 7
20 17 16 11 23 4 21
42 43 31 41 21 21 39
− 97.51 − 101.72 − 97.03 − 102.66 − 105.72 − 105.12 − 100.16
18.18 17.99 15.95 17.99 18.27 18.64 16.98
7 5.8 7.5 6.5 6.5 6.0 6.1
69 54 16 16 20 12 10
262 343 481 428 736 663 322
1023 312 1678 624 624 380 420
Year/month/day/hour/min are: the exact time of the EQ; Latitude and Longitude: the geographic coordinates of the epicenter, magnitude and depth: magnitude and depth of the EQ, distance: the distance between epicenter and Teoloyucan station, ρ: is the radius of the EQ preparation zone estimated by Dobrovolsky formula. The EQ magnitudes are given in bold.
A. Kotsarenko et al. / Tectonophysics 431 (2007) 249–262
251
detailed analysis of the mentioned seismic events in the intermediate and short-time scale. 2. Physical conception and the methodology of the study The basic conception, applied for our experimental study, is the following. Geodynamical processes in the EQ preparation zone can produce current systems of different configuration (Molchanov and Hayakawa, 1995; Surkov et al., 2003). The ULF part of electromagnetic emission due to negligible skin depth (Kopytenko et al., 2001) can be recorded at the Earth's surface by modern magnetometers without significant attenuation. Higher frequencies would have higher skin depths and, therefore, greater attenuation before reaching the Earth's surface. Another source of EM emission can be generated at ionospheric heights due to the variability of the ionosphere. The evidence of the last proposition is supported by experimental observations (Tramutoli et al., 2001) and explained within different (but not contradictive) models: the conversion of atmospheric gravity waves (AGW) in the lower ionosphere (Molchanov, 2004) and the ionosphere modification caused by increased radon emanation before strong EQs (Pulinets and Boyarchuk, 2004). In order to select seismic events which can produce observable contribution to the geomagnetic signal we have used criteria satisfying both mentioned theoretical conceptions. It is assumed that the geomagnetic station should be inside or close to the EQ preparation zones: ρ = 100.43M km, where ρ is the radius of the preparation zone (Dobrovolsky et al., 1979). We have chosen 7 earthquakes for our analysis (see Table 1 for the EQ data, estimations on the EQ preparation zone radius ρ, and distances D between Teoloyucan (TEO) station and EQ epicenters, and Fig. 1 for the map of the EQ epicenters and location of TEO station). Recently, studies by Kotsarenko et al. (2004, 2005) performed for long-time intervals (year scale) challenged the necessity to perform a more detailed analysis for the mentioned EQs focusing on intermediate (± 15 days) and short time scales and concentrating on 2 subject: the dynamics of the fractal index and the behavior of the ULF RSR (Resonant Spectral Structure). According to the phenomenological model proposed for strong earthquakes as a Self-Organized Criticality (SOC) process (Kopytenko et al., 2001), the state of the medium during the EQ preparation passed through different stages: initial stage (random chaos) → subcritical → critical → supercritical stage. As a critical process, the EQ output parameters manifest fractal
Fig. 1. Geomagnetic field emphasize fractal properties: its power spectra fits the inverse power law equation S( f ) ∼f − β. FFT (Fast Fourier Transform) method: best linear fit approximation for presentation log(S)vs-log( f ), fractal index β is calculated as the absolute value of the line declination coefficient.
properties both in time and space scale. One of them, the geomagnetic field spectra, estimated by inverse power law as S( f ) ∼f −β, is indicative of the dynamical processes contributing to geomagnetic field changes during the time of the EQ in terms of the EQ preparation and the postseismic relaxation. Generally, the fractal index dynamics reveals an expressed 27-day variation, connected with a rotational period of the Sun (Hayakawa et al., 1999). The gradual decrease of the fractal index before the EQ and the intense depression of the solar period variation observed in 2 post-seismic months were reported (Hayakawa et al., 1999; Kopytenko et al., 2001) in relation to Guam EQ, 1993, Ms = 8.0. The calculation of the fractal index can be made by different methods, described by Gotoh et al. (2004). We used the FFT (Fast Fourier Transform) method and calculated β as a best linear fit approximation log(S)-vs-log( f ), where S is a spectral density of the signal and f is the frequency (Fig. 1). Another subject of interest, the ULF structure, has been recently discovered in the Teoloyucan magnetic station data (Kotsarenko et al., 2005). Two resonant lines were observed in the H-component (S–N direction, linear polarization) in the frequency bands fR1 = 10.2−11.1 mHz and fR2 = 13.6−14.5 mHz, sometimes accompanied by sporadically generated harmonics (Fig. 2). Following our provisional considerations, the source of the observed resonances is located in the ionosphere and the mechanism of its generation is connected with a geomagnetic
252
A. Kotsarenko et al. / Tectonophysics 431 (2007) 249–262
Fig. 2. ULF geomagnetic resonances observed in the H-component (linearly polarized) in Teoloyucan station, March 30–April 26, 1999. Top panel (a): Equatorial DST index of the geomagnetic activity. Panel (c): SRS (Spectral Resonant Structure) dynamics in the Teoloyucan (TEO) station. fR1, fR2 are 2 dominating resonant frequency bands, fSH-sporadically generated (lower) harmonic. Panels (b) and (d): Referring spectra for Boulder (BLD) and Jicamarca (JIC) stations accordingly. The coordinates of stations are given in the Table 2.
location of the station (geomagnetic latitude λ = 29.1°) close to the latitude of the equatorial electrojet system. The H-component of the magnetic field is chosen for analysis in the presented study, where spectrograms are calculated with NFFT = 213 (Number of points for FFT transform, corresponding time domain T = 2 h 16 min 32 s) to achieve the optimal resolution in the analyzed interval of frequencies f = 0–30 mHz (considered most promising for observations of the seismogenic ULF geomagnetic emission, see Molchanov and Hayakawa (1995), Ismagulov et al. (2001), Kopytenko et al. (2001) for explanation/argumentation). The fractal index β is calculated every hour (NFFT = 3600, time domain T = 1 h). We use also a 6-hours running average for tracing daily variations and 24-hours running average (trend) to see the general behavior of the β.
The analyzed data set is the same as the one used for the recent analysis by Kotsarenko et al. (2004, 2005): the geomagnetic data were recorded at Teoloyucan station (Central Mexico, geographic coordinates: 99 11′ 35.735″ W, 19 44′45.100″ N, 2280 m height). This station was equipped with a 3-component fluxgate magnetometer designed at UCLA, operating at 1 Hz sampling rate frequency, with a GPS system for data synchronization. In our study we used the DST index obtained at Kyoto World Data Center for Geomagnetism (web page) discarding events occurring during high geomagnetic activity from our analysis. In order to distinguish the local character of the phenomena from the global one compared our results with those ones calculated for reference stations from the CANOPUS and Mid-continent
A. Kotsarenko et al. / Tectonophysics 431 (2007) 249–262
253
Fig. 3. Location of Teoloyucan and referring stations (see details in Table 2).
Magneto-seismic Chain (McMAC) network (Chi et al., 2005) equipped with the same type of instruments. 3. Description and discussion of the obtained results Anticipating the analysis of the mentioned 7 seismic events in terms of the dynamics of fractal index and behavior of ULF geomagnetic resonances we should add some technical preface for the validation of the results obtained in our study. The mentioned SRS (Resonance Spectral Structure) was observed for the first time in 2005 in the dominant frequency bands f R1 = 10.2−11.1 and f R2 = 13.6− 14.5 mHz only at Teoloyucan geomagnetic station
(Kotsarenko et al., 2005) and was not detected at other stations in the American continent (Fig. 2, panel c). However, a very similar resonance structure was observed at Beijing (BJI) geomagnetic station (almost the same geomagnetic latitude as Teoloyucan) but was absent even at the nearest Chinese station (situated at about 150 km from Beijing) so leading to the conclusion that it corresponds to a very special localization (channeling character) of the mentioned phenomena, generated with a high probability by the equatorial electrojet system (Kotsarenko et al., 2005). The preliminary qualitative estimations show a good correlation of the intensities of the resonant structure with an Sq variation and its moderate dependence on the solar
Table 2 Geographic and geomagnetic coordinates for Teoloyucan and referring stations Station name
Abbreviation
Latitude Geography
Longitude Geography
Latitude Magnitude
Longitude Magnitude
L value
Athabasca Boulder Teoloyucan Jicamarca
ATH BLD TEO JIC
54.7 40.1 19.7 −12.0
246.7 254.8 260.8 283.1
62.3 49.1 29.1 0.0
305.6 319.6 330.1 354.2
4.64 2.34 1.31 1.00
254
A. Kotsarenko et al. / Tectonophysics 431 (2007) 249–262
Fig. 4. Dynamics of fractal index β for the period of the EQ 1 (Ms 7.0, June 15, 1999). Panel (c): Teoloyucan station, Panels (a, b and d): referring stations: Athabasca (ATH), Boulder (BLD) and Jicamarca (JIC) stations accordingly. Bottom panel (e): Equatorial DST index of the geomagnetic activity. The coordinates of stations are given in the Table 2.
Fig. 5. Map of Mexico. Position of the Teoloyucan magnetic station is marked by triangle. The epicenters of 7 seismic events are displayed as hexagonal stars. Description of the events is given in Table 1.
A. Kotsarenko et al. / Tectonophysics 431 (2007) 249–262
activity (Dst index) which is good in terms of the natural character of the phenomena. However, the principal obstacles (difficulties) for the quantitative analysis are still the insufficient sample rate of the data, very low signal/noise ratio and imperfection of our techniques. Therefore, in this paper we present only qualitative analysis of the changes in the SRS in possible relation to 7 strong and moderate seismic events that occurred in Mexico during 1999–2001. In them, the referring behavior of the SRS (given by Fig. 2, panel c) was calculated for the analyzed interval March
255
30–April 26, 1999 which was “seismically quiet” (1 week out of the any seismic shock with magnitude higher the Ms = 5). The behavior of both resonant modes show that during the strongest geomagnetic storm the SR structure is depressed (Fig. 2c, area marked by circle 1) and saturated by the perturbed spectra (dark areas) while at the time of the smaller geomagnetic storms the depression and saturation are much smaller (Fig. 2c, circles 1 and 2). Another subject of our study, the fractal property of geomagnetic field (confirmed by already mentioned
Fig. 6. EQ 1 (Ms 7.0, June 15, 1999) and EQ 2 (Ms 5.8, June 21, 1999), Teoloyucan station results. Panel (a): Geomagnetic field, H-component; Panel (b): Power spectra sonogram for the H-component; Panel (c): Equatorial DST index of the geomagnetic activity; Panel (d): Temporal distribution of the foreand aftershock Magnitudes (Ms); Pentagonal stars: shocks in the 1st EQ zone; Asterisks: shocks in the 2nd EQ zone; Main shocks are marked by hexagonal stars; Panel (e): Temporal distribution of the fore- and aftershock depths symbols are the same as in (d); Panel (f): Fractal index β temporal dynamics.
256
A. Kotsarenko et al. / Tectonophysics 431 (2007) 249–262
numerous studies) is analyzed in quantitative form: the power spectra (or fractal) tendency S(f) ∼ f −β is approximated by a best linear fit and fractal index β is calculated as the absolute value of the line inclination coefficient (Fig. 2). In order to distinguish the local character of the fractal index β dynamics from the global changes, we compare our results with a reference β calculated for the stations located at different geomagnetic latitudes (see the map in Fig. 3 and Table 2 for details). Of course, the fractal dynamics of the geomagnetic field will be different for different stations because of the difference of physical processes in the corresponding magneto-
sphere cavities (the L-values are quite different for every station of the station set). For EQ 1, we found a significant drop of the fractal index in the Teoloyucan β close to the time of the event (Fig. 4) compared to the corresponding monotonous changes of referent stations series thus indicating that it is a local anomaly. The observed phenomena may be interpreted as due to a high contribution of the extended ionosphere perturbed by the EQ preparation process throw the lithosphere– ionosphere–magnetosphere coupling. The global reduction of the fractal index at other days may be caused by a sharp change in the ring current system: the Dst value (both direction and intensity) for the equatorial and low-
Fig. 7. EQ 3 (Ms 7.5, September 30, 1999). The table of symbols is the same as in Fig. 6.
A. Kotsarenko et al. / Tectonophysics 431 (2007) 249–262
latitudinal stations (for the high-latitudinal stations such as Athabasca (ATH) the polar currents (AE index) will be also important). In our analysis and in the presented figures we combined 1st and 2nd EQs as they occurred in a close time interval (6 days). We did the same for the 5th and 6th events (one day difference). 3.1. Tehuacan Ms 7.0 (June, 15, 1999) and Guerrero Ms 5.8 (June, 21, 1999) earthquakes The June 15, 1999, magnitude 7.0 earthquake occurred in the central region of México (EQ1, Fig. 5), at 15:42 h (20:42 GMT), killing 16 person and injuring more then 200. The epicenter of the earthquake was located near the border of the states of Puebla and Oaxaca, about 20 km southeast of Tehuacan, and about 230 km east-southeast of Mexico City. A second strong event (EQ2 at Fig. 5) with magnitude 5.8 (Mw = 6.3 according USGS) occurred on June 21 at 12:43 h (17:43 GMT) in the coast of Guerrero state. Both earthquakes, especially the 1st one, damaged a number of buildings, walls, bridges and communications, disabled electricity and telephone services, the first EQ even gained a name
257
of “church” earthquake for the destruction of temples in Puebla state. The 1st EQ was preceded by a 2 small foreshocks, one at a small depth and one almost superficial; several aftershocks were clustered in 2 very short time intervals: some hours just after the main shocks and about 1 day after it (Fig. 6, panels d, e). The geomagnetic conditions are very favorable for the analyzed interval (Fig. 6, panel c): the DST index was mostly positive never dropping lower than − 40 gammas for more than 13 days (+ 13 days) after the EQ. Our principal results obtained for the EQ 1 are the following. We observe periodical post-seismic depression of both resonant modes especially clear at the period +2 to + 9 days after the 1st event (Fig. 6b, area marked by an ellipse). Due to lack of data (2 days of the record lost starting some hours after the main shock) we cannot say exactly about the time when these depressions started. The character of the fractal index shows a noticeable fall 2 days before the 1st event (Fig. 6f, area marked by circle), whereas the geomagnetic activity (Fig. 6c) display a very weak character. Neither can we estimate the character of its recovery during the days + 1–2 days.
Fig. 8. Spectral Resonant Structure observed on 21–22 of September 1999 in the H- and D-components in the Teoloyucan station 8 days before the EQ 3 (Ms 7.5, September 30, 1999). Panels (a–c): H-, D-, and Z-components. Panel (d): Equatorial DST index of the geomagnetic activity.
258
A. Kotsarenko et al. / Tectonophysics 431 (2007) 249–262
The observed anomaly drop has a local character: as it was mentioned before (explanations for the fractal dynamics analysis presented in Fig. 4) no evident changes were observed for the referring geomagnetic stations. In relation to the second EQ we cannot emphasize any expressed feature directly associated with this event in the intermediate scale. 3.2. Oaxaca Ms 7.5 (September, 30, 1999) earthquake A large earthquake struck Oaxaca State, Mexico earthquake at 11:31 Mexico time (16:31 GMT) at Mexican
beach resort of Puerto Escondido (about 60 km from Puerto Angel, 440 km southern México City, see EQ 3 at Fig. 5), killing 33 people, injuring 215, destroying and making damages to thousands of buildings, disrupting utilities and blocking roads by landslides in the state of Oaxaca. The EQ was preceded by a dozen foreshocks and followed by numerous aftershocks which also revealed a tendency to group into short-time clusters (Fig. 7d, e): 3 foreshocks at − 2 days (occurred during about 4 hours time interval), 6 events just after the main shock (12 h), 5 shocks at + 1 day (6 h), 5 events at +2 days (6 h) and 3 events at + 14 days (6 h). Another group of 7 events occurred at +3 to +4 in the interval of 27 h.
Fig. 9. EQ 4 (Ms 6.5, August 9, 2000). The table of symbols is the same as in Fig. 6.
A. Kotsarenko et al. / Tectonophysics 431 (2007) 249–262
The Sq-variation decreased and almost disapeared during various long periods; some of them, like a preseismic depression − 15–14 days and + 10–15 days can be explained by a perturbed geomagnetic activity (small storms not lower than − 50–60 gammas), another postseismic depression +1–7 days possess the anomaly character (Fig. 7a, marked by ellipse). The fractal index displays a tendency to decrease 2 days before the EQ (Fig. 7g, marked by circle 1) which is a regional change. Both resonant modes at the analyzed period are too week among the noise level so we cannot postulate any reliable changes in them. But another curious event occurred 8 days before the main shock (Fig. 7b, marked
259
by ellipse) and is presented in detail in Fig. 8. An SRS (Spectral Resonant Structure) polarized in H- and Dcomponents (Fig. 8, panels a and b, area marked by ellipse) was generated approximately at 18 UT h., September 21, 1999 during quiet geomagnetic conditions (monotonously lowered down to the peak − 40 nT at time 2 UT, September 22, Fig. 8d) and disappeared at 13 UT, September 22 just at the time when the Dst index went drastically from zero level to + 20 nT. Due to the luck of the preceding 2-day interval of the data we cannot postulate now the reliability of the related drop of the regional fractal index (Fig. 8f, marked by circle 2). A more detailed analysis for β behavior with a precise
Fig. 10. EQ 5 (Ms 6.5, May 19, 2001) and EQ 6 (Ms 6.0, May 20, 2001). The table of symbols is the same as in Fig. 6.
260
A. Kotsarenko et al. / Tectonophysics 431 (2007) 249–262
reconstruction of the mentioned fragment is expected in the near future. 3.3. Lazaro Cardenas Ms 6.5 (August 9, 2000) earthquake The earthquake centered near Mexico's Pacific Ocean coast shook western and central Mexico early Wednesday at 6:41 a.m. (11:41 GTM, EQ 4 at Fig. 5), cracking walls, sending people racing out of their homes and injuring at least one person when parts of a wall collapsed in a hotel under repair in the coastal city of Lazaro Cardenas, about 30 mi (48 km) from the quake's epicenter.
This EQ occurred with no foreshocks (Fig. 9d, e), but several aftershocks occurred in 2 very short time clusters (dispersal 1–2 h) at the day of the main event, and the main part of the aftershock activity was detected during +3–5 days almost uniformly in time. For this earthquake most of the changes in the fractal dynamics use to be global (no unusual features in a regional scale). Both the 1st and 2nd resonant modes tend to decrease slightly at the pre- and post-seismic interval − 2 to + 2d (Fig. 9b, marked by a circle), with generation of a sporadic lower harmonic detected at the frequencies Δf ∼ 0.6–9 mHz. The + 2d and the following interval is discarded from the analysis due
Fig. 11. EQ 7 (Ms 6.1, October 7, 2001). The table of symbols is the same as in Fig. 6.
A. Kotsarenko et al. / Tectonophysics 431 (2007) 249–262
to the occurrence of strong geomagnetic storms (− 100 nT and − 150 nT, Fig. 9c). 3.4. Jalisco Ms 6.5 (May 19, 2001) and Ms 6.0 (May 20, 2001) earthquakes Both analyzed events occurred in the undersea tectonic zone, with no damage to human beings nor to buildings being reported (see EQ 5 and 6 at Fig. 5). The second event, taken from the SSN (National Seismological Service) database, can be erroneous with a high probability as no reports confirming its occurrence was discovered in international data bases. The EQs were preceded by several foreshocks, several aftershocks occurred after the main shock, and some shocks took place at the post-seismic 15-day interval. Nothing anomalous is detected for these events in geomagnetic field, neither fractal dynamics nor resonant modes behavior. The changes in the H-component and a week depression of both resonances at the period − 12–3 days should be caused by geomagnetic activity, but both resonant modes show smearing (thickening) tendency at the post-seismic period 0 to +4d (Fig. 10b, area marked by circle). In other words, we observe the decrease of the quality Q factor for both resonant modes. 3.5. Guerrero Ms 6.1 (October 7, 2001) earthquake The earthquake occurred in Guerrero State (EQ 7 at Fig. 5), no damages or human losses being reported. The USGS catalog gives a quite different magnitude Ms = 5.5 and depth D = 33 km for the mentioned EQ. The most interesting feature of that EQ was that that in spite of not being of high magnitude it triggered an intensive seismic activity not only in the regional seismic zone but all over Mexico and Guatemala (Fig. 11d, e). The interval − 14–2 days has been discarded from the analysis because of the high geomagnetic activity present. We observe a small regional fall of the fractal index in the near-earthquake time interval ±1 day (Fig. 10 f, area marked by circle), the rest of the relative changes found to be global and caused by geomagnetic storms. Both resonant modes (fR2, higher mode) are almost depressed among the noise for most of the analyzed period (especially the higher mode fR2). 4. Conclusions The results obtained for the 7 analyzed earthquakes can be summarized as follows:
261
1. A decrease of the regional part of fractal index of the geomagnetic field β is observed for events (EQs) 1, 3 and 7 starting about 1–3 days before the major seismic events, and was not detected (“failed”) for the event 4 (in spite of the fact that also satisfies the Dobrovolsky criteria ρ N D) as well as for events 2, 5 and 6 (when the station was out of the EQ preparation zone estimated by Dobrovoslky formula). For the rest of the observed anomalies (EQs 1, 3 and 7) the “magnitude” of the respective fall of the fractal index (in terms of the relation of the value for the observed near-seismic decrease to levels of the “natural” daily and monthly variation) decreases with the ρ/D relation (more attenuated). Actually, the most pronounced drop of the fractal index was observed for the case of the Tehuacan EQ (EQ 1, Ms 7.0, June 15, 1999) with an estimated EQ preparation zone ρ = 1023 km and distanced D = 262 km (ρ/D = 3.9), that can be considered as a are very favorable condition for the observation (lowest attenuation). 2. The ULF resonant modes of the geomagnetic field also undergo pre-seismic (event 3), near-seismic (event 4) and post-seismic depressions (events 1 and 4) and smearing of the resonance structure (the decrease of the SRS quality Q factor, events 5 and 6) possibly related to the mentioned earthquakes. “Possible” because the level of confidence of the observed phenomena is lower and future studies are required for the validation of the anomaly character of the seismo-related changes of the SRS. 3. The unusual multi-frequency SRS (fine structure) was observed in the H-and D-components 8 days prior to the strongest of the analyzed events (EQ 3, Ms 7.5, September 30, 1999). The mentioned SRS was generated under relatively quiet (Dst N − 40 nT) geomagnetic conditions and terminated just at the time of the change of the direction (and intensity) of the equatorial ring current system (positive Dst value). Acknowledgements We are grateful to two anonymous referees for their constructive criticisms, useful corrections and comments, which allowed us to improve the quality of this paper and stimulates our understanding of the reported phenomena. The authors (a) also grateful to the internal UNAM foundation DGAPA and Mexican government foundation CONACyT for the partial support of this work by the projects PAPIIT IN117106 and CONACyT 47662, respectively. We also thank Peter Chi for providing us with the referring McMAC data.
262
A. Kotsarenko et al. / Tectonophysics 431 (2007) 249–262
References Balasanyan, S.Yu., 2005. High-sensitivity Energo-active Points of the Earth (phenomenon HEP): a key to the future understanding of the physics of the earthquakes? Geol. Geophys. 46, 83–99 (in Russian). Chi, P.J., Engebretson, M.J., Moldwin, M.B., Russell, C.T., Mann, I.R., Samson, J.C., López Cruz-Abeyro, J.A., Yumoto, K., Lee, D.-H., 2005. Mid-continent Magnetoseismic Chain (McMAC): A Meridional Magnetometer Chain for Magnetospheric Sounding. Proceedings of the Enviroment Modeling Workshop, June 17–22, 2005, Snowmass, Colorado, USA. Dey, S., Singh, R.P., 2003. Surface latent heat flux as an earthquake precursor. Nat. Haz. Earth Syst. Sci. (NHESS) 3, 749–755. Dobrovolsky, I.R., Zubkov, S.I., Myachkin, V.I., 1979. Estimation of the size of earthquake preparation zones. PAGEOPH 117, 1025–1044. Dunajecka, M.A., Pulinets, S.A., 2005. Atmospheric and thermal anomalies observed around the time of strong earthquakes in Mexico. Atmósfera 18 (4), 235–247. Fraser-Smith, A.C., Bernardy, A., McGill, P.R., et al., 1990. Low frequency magnetic field measurements near the epicenter of the Loma Prieta earthquake. Geophys. Res. Lett. 17, 1465–1468. Gotoh, K., Akinaga, Y., Hayakawa, M., Hattori, K., 2002. Principal component analysis of ULF geomagnetic data for Izu islands earthquakes in July 2000. J. Atmos. Electr. 22, 1–12. Gotoh, K., Hayakawa, M., Smirnova, N., Hattori, K., 2004. Fractal analysis of seismogenic ULF emissions. Phys. Chem. Earth 29, 419–424. Harada, M., Hattori, K., Isezaki, N., 2004. Transfer function approach to signal discrimination of ULF geomagnetic data. Phys. Chem. Earth 29, 409–417. Hayakawa, M. (Ed.), 1999a. Atmospheric and Ionospheric Electromagnetic Phenomena Associated with Earthquakes. TERAPPUB, Tokyo. Hayakawa, M., Fujinawa, Y. (Eds.), 1994. Electromagnetic Phenomena Related to Earthquake Prediction. Terra Sci. Pub. Co, Tokyo. Hayakawa, M., Molchanov, O.A. (Eds.), 2002. Seismo Electromagnetics: Lithosphere–Atmosphere–Ionosphere Coupling. TERRAPUB, Tokyo. Hayakawa, M., Ito, T., Smirnova, N., 1999b. Fractal analysis of ULF geomagnetic data associated with the Guam earthquake on 8 August 1993. Geophys. Res. Lett. 26, 2797–2800. Ismaguilov, V.S., Kopytenko, Yu.A., Hattori, K., Voronov, P.M., Molchanov, O.A., Hayakawa, M., 2001. ULF magnetic emissions connected with under sea bottom Earthquakes. Nat. Hazards Earth Syst. Sci. (NHESS) 1, 23–31. Kopytenko, Yu.A., Matiashvili, T.G., Voronov, P.M., Kopytenko, E.A., Molchanov, O.A., 1993. Detection of ultra-low-frequency emissions connected with the Spitak earthquake and its aftershock activity, based on geomagnetic pulsations data at Dusheti and Vardzia observatories. Phys. Earth Planet. Inter. 77 (1–2), 85–95. Kopytenko, Yu., Ismaguilov, V., Hayakawa, M., Smirnova, N., Troyan, V., Peterson, Th., 2001. Investigation of the ULF electromagnetic phenomena related to earthquakes: contemporary achievements and the perspectives. Ann. Geofis. 44 (2), 325–334. Kotsarenko, A., Pérez Enríquez, R., López Cruz-Abeyro, J.A., Koshova, S., Grimalsky, V., 2004. Analysis of the ULF
electromagnetic emission related to seismic activity, Teoloyucan geomagnetic station, 1998–2001. Nat. Hazards Earth Syst. Sci. (NHESS) 4, 679–684. Kotsarenko, A., Pérez Enríquez, R., López Cruz-Abeyro, J.A., Koshevaya, S., Grimalsky, V., Kremenetsky, I., 2005. Possible seismogenic origin of the changes in the ULF EM resonant structure observed at Teoloyucan geomagnetic station, Mexico, 1999–2001. Nat. Hazards Earth Syst. Sci. (NHESS) 4, 679–684. Molchanov, O.A., 2004. On the origin of low- and middle-latitude ionospheric turbulence. Phys. Chem. Earth 29, 559–556. Molchanov, O.A., Hayakawa, M., 1995. Generation of ULF electromagnetic emissions by microfracturing. Geophys. Res. Lett. 22, 3091–3094. Molchanov, O.A., Hayakawa, M., 1998. Subionospheric VLF signal perturbations, possibly related to earthquakes. J. Geophys. Res. 103, 17489–17504. Molchanov, O.A., Hayakawa, M., Ondoh, T., Kawai, E., 1998. Precursory effects in the subionospheric VLF signals for the Kobe earthquake. Phys. Earth Planet. Inter. 105, 239–248. Ouzounov, D., Freund, F., 2004. Mid-infrared emission prior to strong earthquakes analyzed by remote sensing data. Adv. Space Res. 33, 268–273. Pulinets, S.A., 2006. Space technologies for short-term earthquake warning. Adv. Space Res. 37 (4), 643–652. Pulinets, S.A., Boyarchuk, K.A., 2004. Ionospheric Precursors of Earthquakes. Springer, Berlin. 316 pp. Surkov, V.V., Molchanov, O.A., Hayakawa, M., 2003. Preearthquake ULF electromagnetic perturbations as a result of inductive seismomagnetic phenomena during microfracturing. J. Atmos. Sol.-Terr. Phys. 65, 31–46. Tronin, A.A., Hayakawa, M., Molchanov, O.A., 2002. Thermal IR satellite data application for earthquake research in Japan and China. J. Geodyn. 33, 519–534. Tramutoli, V., Di Bello, G., Pergola, N., Piscitelli, S., 2001. Robust satellite techniques for remote sensing of seismically active areas. Ann. Geofis. 44, 295–312. Vallianatos, F., Tzanis, A., 1999. A model for the generation of precursory electric and magnetic fields associated with the deformation rate of the earthquake focus. In: Hayakawa, M. (Ed.), Atmospheric and Ionospheric Electromagnetic Phenomena Related with Earthquakes. TERRAPUB, Tokyo, pp. 287–305. Varotsos, P., 2001. A review and analysis of electromagnetic precursory phenomena. Acta Geophys. Pol. 49, 1–42. Varotsos, P., Lazardou, M., 1999. A review of the VAN method. In: Hayakawa, M. (Ed.), Atmospheric and Ionospheric Electromagnetic Phenomena Associated with Earthquakes. TERRAPUB, Tokyo, pp. 47–53. Varostos, P., Sarlis, N., 2002. A review of recent VAN efforts: the explanations of SES physical properties. In: Hayakawa, M., Molchanov, O.A. (Eds.), Seismo Electromagnetics: Lithosphere – Atmosphere – Ionosphere Coupling. TERRAPUB, Tokyo, pp. 131–140. Varotsos, P., Ueda, S., Alexopoulos, K., Nagao, T., Lazaridou, M., 1994. Prediction of recent destructive seismic activities in Greece based on seismic electric signals. In: Hayakawa, M., Fujinawa, Y. (Eds.), Electromagnetic Phenomena Related to Earthquake Prediction. TERRAPUB, Tokyo, pp. 13–25.