- Email: [email protected]
Ionospheric anomalies possibly associated with M ⩾ 6.0 earthquakes in the Japan area during 1998–2010: Case studies and statistical study
Journal of Asian Earth Sciences 41 (2011) 410–420
Contents lists available at ScienceDirect
Journal of Asian Earth Sciences journal homepage: www.elsevier.com/locate/jseaes
Ionospheric anomalies possibly associated with M P 6.0 earthquakes in the Japan area during 1998–2010: Case studies and statistical study Shimpei Kon, Masahide Nishihashi, Katsumi Hattori ⇑ Graduate School of Science, Chiba University, 1-33, Yayoi, Inage, Chiba 263-8522, Japan
a r t i c l e
i n f o
Article history: Available online 31 October 2010 Keywords: Ionospheric anomaly Total electron content Earthquake-related ionospheric anomaly Statistical analysis Case studies M P 6.0 earthquakes around Japan
a b s t r a c t In this paper, we examine pre-earthquake ionospheric anomalies in time series and perform a statistical test by using total electron content (TEC) derived from global ionosphere maps (GIM) around the Japan area for the first time. The normalized GIM-TEC (GIM-TEC), which is computed based on 15 days backward running mean of GIM-TEC, have been investigated for minimizing possible confounding effects of consecutive earthquakes and identify the abnormal signals. Meanwhile, to reduce the effect of strong geomagnetic activities such as geomagnetic storms, the criterion for removing the GIM-TEC data have been adapted; that is when Dst index exceeds 60 nT. Temporal variations of GIM-TEC for large and destructive earthquakes in Japan have been studied; which are the 2004 mid-Niigata Prefecture Earthquake (M6.8), its aftershock (M6.1), the 2007 offshore mid-Niigata Earthquake (M6.8), and the 2008 Iwate–Miyagi Nairiku Earthquake (M7.2). Although there are some positive and negative TEC anomalies before and after the four earthquakes, there is a tendency that positive TEC anomalies appear 1–5 days before all the above earthquakes even during the quiet geomagnetic condition. Superposed epoch analyses have been performed for the statistical analysis of TEC anomalies associated with M P 6.0 earthquakes during the 12-year period of May 1998–May 2010. The statistical result indicates the significance of the positive TEC anomalies 1–5 days before earthquakes within 1000 km from the epicenter around Japan. Ó 2010 Elsevier Ltd. All rights reserved.
1. Introduction A large number of papers have been reported on anomalous electromagnetic phenomena possibly associated with large earthquakes (e.g., Hayakawa and Fujinawa, 1994; Hayakawa, 1999; Hayakawa and Molchanov, 2002; Pulinets and Boyarchuk, 2004; Molchanov and Hayakawa, 2008). One of the most frequent studies among them is to investigate the relationship between ionospheric anomalies and earthquake activities. Scientists have observed anomalous variations appearing a few days before some large earthquakes in the ionospheric plasma frequency foF2 (or F2-peak electron density NmF2) recorded by ionosondes (Pulinets et al., 1994; Pulinets, 1998; Liu et al., 2000, 2004a,b, 2006, 2008; Chuo et al., 2001, 2002; Chen et al., 2004; Hobara and Parrot, 2005; Liperovskaya et al., 2006). Besides foF2 and NmF2, total electron content (TEC) derived by ground and satellite based receivers of the global positioning system (GPS) is also used to observe ionospheric condition (e.g., Sardon et al., 1994; Liu et al., 1996). The NmF2 and GPS-TEC are, in general, highly correlated (Liu et al., 2001, 2004b; Nishihashi et al., 2009). This fact suggests that the
⇑ Corresponding author. Tel.: +81 43 290 2801; fax: +81 43 290 2859. E-mail address: [email protected] (K. Hattori). 1367-9120/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.jseaes.2010.10.005
GPS-TEC can be employed to monitor the ionospheric sciences including preseismic perturbations. Recently, scientists have conducted not only a case study but also a statistical study of the preseismic TEC anomalies (e.g., Nishihashi et al., 2009; Liu et al., 2001, 2004a,b, 2008, 2009; Pulinets et al., 2004, 2007; Dautermann et al., 2007; Zhao et al., 2008). Liu et al. (2004a) reported that the GPS-TEC pronouncedly decreases within 1–5 days prior to 20M P 6.0 earthquakes in Taiwan during September 1999–December 2002. Moreover, Liu et al. (2009) reported that the anomalous TEC reductions appear 3–5 days before 17M P 6.3 earthquakes in China region from May 1, 1998 to April 30, 2008 and the result is consistent with the case study of the 2008 Wenchuan earthquake (M7.9). In addition, an extreme increase of TEC for all the analyzed period is recorded in the afternoon of 3 days before the earthquake. Japan is recognized to be one of the most seismically active regions in the world and a dense GPS network has been established. Although previous studies suggested that significant TEC anomalies would accompany prior to larger earthquakes around Japan (Zakharenkova et al., 2007), statistical studies have not been reported. Therefore, TEC variation around Japan is investigated in this paper. At first, we investigate the TEC variation derived from the global ionospheric maps (GIM) before and after some large earthquakes; which are the 2004 mid-Niigata Prefecture Earthquake
S. Kon et al. / Journal of Asian Earth Sciences 41 (2011) 410–420
(M6.8), its aftershock (M6.1), the 2007 offshore mid-Niigata Earthquake (M6.8), and the 2008 Iwate–Miyagi Nairiku Earthquake (M7.2). Then, we perform a statistical analysis of earthquake-related TEC anomalies around Japan for 12-year from May 1998 to May 2010.
411
(Kelley, 1989; Davies, 1990). To reduce these effects, a criterion for the ionospheric disturbed period is introduced. That is the 2 days after Dst index exceed 60 nT. Therefore, we remove the 2 days TEC data after storm onsets in this paper. Dst index derived from four geomagnetic observations is used to represents the geomagnetic disturbances at the lower geomagnetic latitudes.
2. Data and normalized GIM-TEC (GIM-TEC) Recently, many scientists utilize the global ionosphere maps (GIM) containing grid data of the vertical TEC to investigate ionospheric phenomena (Mendillo et al., 2005; Afraimovich et al., 2008; Hocke, 2008). Therefore, the GIM is also considered a useful tool to analyze earthquake-related TEC variations (Nishihashi et al., 2009; Zakharenkova et al., 2006, 2008; Afraimovich and Astafyeva, 2008; Zhao et al., 2008; Liu et al., 2009; Yu et al., 2009). In this paper, the GIM derived by the Center for Orbit Determination in Europe (CODE; ftp://ftp.unibe.ch/aiub/CODE/) is selected. The vertical TEC is modeled in a solar-geomagnetic reference frame using a spherical harmonics expansion. In 2009, the degree and order of the spherical harmonics expansion were 15 and 15, respectively. Piece-wise linear functions are used for representation in the time domain. The spatial resolution is 2.5° in latitude, and 5° in longitude and the temporal resolution is 2 h. Schaer (1999) describes the computational algorithm in detail. In this paper, we linearly interpolate yielding a 1-h resolution at a grid point. Remember that the TEC derived from GIM (hereafter GIM-TEC) is obtained through the spherical harmonics expansion and the linear interpolation. Fig. 1 shows the relationship between the GIM-TEC at the grid of 140°E–37.5°N and GPS-TEC observed at the 0920 GPS station (140.73°E, 37.82°N) of GPS Earth Observation Network (GEONET) in 2007. Their correlation is found to be about 0.73.To minimize possible confounding effects of consecutive earthquakes and properly identify the abnormal signals, we computed the mean GIM-TEC (GIM—TEC) values for the previous 15 days, and the corresponding standard deviation (r) as a reference at specific times. Then, we derived the normalized GIM-TEC (GIM-TEC) values by the following equation,
GIM TEC ðtÞ ¼
3. Case studies In this section, we would like to demonstrate some case studies on GIM-TEC temporal variation. We select the intense, destructive, and significant earthquakes during the last decade; which are the 2004 mid-Niigata Prefecture Earthquake (M6.8), its large aftershock (M6.1), the 2007 offshore mid-Niigata Earthquake (M6.8), and the 2008 Iwate–Miyagi Nairiku Earthquake (M7.2). Fig. 2 draws the epicenters with magnitude and detailed information is described in Table 1 (Nos. 26, 27, 43, and 45).
3.1. The 2004 mid-Niigata Prefecture Earthquake The 2004 mid-Niigata Prefecture Earthquake (M6.8) occurred at 08:56 UT, October 23, 2004. The epicenter is located at 37.29 N° and 138.87 E° and the depth is 23 km. The source mechanism is reverse faulting. It was one of the most damaging earthquakes in Japan in last 10 years. Seventeen aftershocks with M P 5.0 occurred and the large aftershock (M6.1) occurred at 01:40 UT, October 27, 2004. Fifty-one people were killed, 4805 people were injured, and 16,900 houses were damaged.
GIM TECðtÞ GIM TECðtÞ : rðtÞ
Meanwhile, strong geomagnetic activities such as geomagnetic storm can cause ionospheric disturbances and TEC could be perturbed about a few hours to 2 days after geomagnetic storm onsets
Fig. 1. The correlation between the GIM-TEC at the grid of 140°E–37.5°N and GPSTEC observed at 0920 GPS station of GEONET in 2007. The correlation coefficient is about 0.73.
Fig. 2. The map of epicenters. (a) The 2004 mid-Niigata Prefecture Earthquake, (b) its aftershock, (c) the 2007 offshore mid-Niigata Earthquake, and (d) the 2008 Iwate–Miyagi Nairiku Earthquake.
412
S. Kon et al. / Journal of Asian Earth Sciences 41 (2011) 410–420
Table 1 List of the 52 earthquakes M P 6.0 and depth 640 km that occurred in region (A) (0 6 R < 1000 km) from May 1998–May 2010. No
Year
Month
Day
Hour
Minute
Latitude (°N)
Longitude (°E)
Depth (km)
M
Distance (km)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52
1998 1998 2000 2000 2000 2000 2000 2000 2001 2001 2001 2002 2002 2003 2003 2003 2003 2003 2003 2003 2004 2004 2004 2004 2004 2004 2004 2004 2004 2004 2005 2005 2005 2005 2005 2005 2005 2005 2006 2006 2007 2007 2007 2008 2008 2008 2008 2008 2008 2009 2009 2010
5 5 7 7 7 10 10 11 3 4 8 8 11 7 9 9 9 10 10 12 4 5 9 9 9 10 10 11 11 12 1 3 8 8 8 10 11 12 10 10 2 3 7 5 6 7 7 9 12 6 8 3
14 30 1 15 30 3 6 13 23 14 13 20 3 25 25 26 29 8 31 29 3 29 5 6 8 23 27 11 28 6 19 20 16 24 30 19 14 2 10 23 17 25 16 7 13 19 21 11 20 5 10 14
18 18 7 1 12 4 4 15 11 23 20 10 3 22 19 20 2 9 1 1 23 20 14 23 14 8 1 10 18 14 6 1 2 10 18 11 21 13 23 21 0 0 1 16 23 2 11 0 10 3 20 8
56 18 1 30 25 13 30 57 30 27 11 59 37 13 50 38 36 6 6 30 2 56 57 29 58 56 40 2 32 15 11 53 46 15 10 44 38 13 58 17 2 41 13 45 43 39 30 20 29 30 7 8
40.25 39.03 34.22 34.32 33.9 40.28 35.46 42.49 44.07 30.09 41.05 30.99 38.89 38.42 41.81 41.99 42.45 42.65 37.81 42.42 36.43 34.25 33.18 33.21 33.14 37.23 37.28 42.14 43.01 42.9 34.06 33.81 38.28 38.56 38.48 36.4 38.11 38.09 37.2 29.35 41.79 37.34 37.53 36.16 39.03 37.55 37.19 41.89 36.54 41.82 34.74 37.74
143.25 143.44 139.13 139.26 139.38 143.12 133.13 144.76 148.05 141.77 142.31 141.97 141.98 141 143.91 144.58 144.38 144.57 142.62 144.61 141.01 141.41 137.07 137.23 137.2 138.78 138.88 144.34 145.12 145.23 141.49 130.13 142.04 142.99 143.18 140.84 144.9 142.12 142.66 140.27 143.55 136.59 138.45 141.53 140.88 142.21 142.05 143.75 142.43 143.45 138.26 141.59
33 33 10 10 10 33 10 33 33 10 38 9 39 6 27 33 25 32 10 33 31 16 10 10 21 16 14 32 39 35 27 10 36 10 21 32 11 29 9 11 31 8 12 27 7 22 22 25 19 29 40 32
6 6.2 6.1 6.1 6.5 6.3 6.7 6 6 6 6.4 6.3 6.4 6.1 8.3 6 6.5 6.7 7 6.1 6 6.5 7.4 6.6 6.1 6.6 6 6.1 7 6.8 6.6 6.6 7.2 6.1 6.1 6.3 7 6.5 6 6.4 6 6.7 6.6 6.9 6.9 7 6 6.8 6.3 6.4 6.1 6.5
415 345 371 358 403 410 655 686 995 837 440 744 231 134 584 633 664 691 233 673 148 381 547 538 546 112 101 634 750 746 404 981 198 287 299 142 436 197 238 903 565 302 137 201 186 195 184 583 240 563 343 142
Fig. 3 illustrates the variation of GIM-TEC at a certain pixel above the epicenter together with that of Kp index, Dst index, and F10.7 solar flux. Kp index derived from 13 geomagnetic observations in subauroral region is generally used to evaluate geomagnetic disturbances in high and middle geomagnetic latitudes. The F10.7 index reveals the solar activity level based on the solar radio flux at a wavelength of 10.7 cm. Two vertical lines show the origin time of earthquakes; the first one is the 2004 mid-Niigata Prefecture Earthquake with M6.8 and the latter is its aftershock with M6.1. The period colored by gray is geomagnetically disturbed period according to the criterion of Dst less than 60 nT. Positive GIM-TEC anomalies are found 2–3 days before the 2004 mid-Niigata Prefecture Earthquake and its aftershock even in the relatively quiet geomagnetic condition. Their duration exceeding +2r are about 10 and 18 h, respectively. Meanwhile, there are negative anomalies with 12 h on September 24 and some
impulsive negative anomalies on October 4–6 and October 16–17. For positive anomalies, there are one with 7 h on September 28, some impulsive ones on October 10, and one with 5 h on October 29. 3.2. The 2007 offshore mid-Niigata Earthquake The 2007 offshore mid-Niigata Earthquake (M6.8) occurred at 01:13 UT, July 16, 2007. The epicenter is located at 37.56°N and 138.61°E and the depth is 17 km. The source mechanism is reverse faulting. The largest aftershock occurred at 06:37 UT on the same day. Its magnitude and depth are 5.8 and 23 km, respectively. The number of causality is 15 and 2415 people were injured. It was a serious shock that a nuclear power station at Kashiwazaki was damaged by the earthquake severely and the operation was automatically stopped. That is the first time for human kind. The routine operation restarts partly in 2009 at last.
S. Kon et al. / Journal of Asian Earth Sciences 41 (2011) 410–420
413
Fig. 3. The variation of GIM-TEC at the study pixel together with that of Kp index, Dst index, and F10.7 solar flux. Two vertical lines show the origin time of earthquakes; the first one is the 2004 mid-Niigata Prefecture Earthquake (M6.8) and the latter is its aftershock (M6.1). The period colored by gray is geomagnetically disturbed period according to the criterion of Dst less than 60 nT.
Fig. 4. The variation of GIM-TEC 30 days before and after the 2007 offshore mid-Niigata Earthquake.
Fig. 4 shows the variation of GIM-TEC associated with the 2007 offshore mid-Niigata Earthquake in the same manner with Fig. 3. There are some negative anomalies on June 16, June 26, July 5, and July 14–15 and positive anomalies on June 30, July 4, July 11,
and July 26. Among of them, it is found that strong positive GIMTEC anomalies 5 days before the earthquake. Their total duration exceeding +2r is about 10 h and the largest value reaches +4r. Furthermore, strong reductions appear 3–4 days before the
414
S. Kon et al. / Journal of Asian Earth Sciences 41 (2011) 410–420
Fig. 5. The variation of GIM-TEC 30 days before and after the 2008 Iwate–Miyagi Nairiku Earthquake.
earthquakes. Their duration exceeding 2r is over 18 h and maximum reduction is 3.9r even in a geomagnetically quiet condition. 3.3. The 2008 Iwate–Miyagi Nairiku Earthquake
Hocke, 2008). It has been utilized very profitably for solar–terrestrial, space and atmospheric physics (e.g., Chree, 1912; Haung and Lee, 1975; Taylor et al., 1994; Singh and Badruddin, 2006). For investigation of TEC, the SEA is employed for the analysis of
The 2008 Iwate–Miyagi Nairiku Earthquake (M7.2) occurred at 23:43 UT, June 13, 2008. The epicenter is located at 39.03°N and 140.88°E and the depth is 8 km. The source mechanism is reverse faulting. The earthquake triggered landslide. Several micro-earthquakes (M < 1.4) occurred about 40 min before the earthquake near the epicenter and three aftershocks with M > 5 occurred on June 14, 2008 (M5.7 and M5.2) and June 16, 2008 (M5.3). Over 10 people died and about 450 people were seriously injured. Fig. 5 indicates the variation of GIM-TEC associated as shown in the 2008 Iwate–Miyagi Nairiku Earthquake in the same manner as shown in Figs. 3 and 4. It is found that the large TEC enhancement of +4.5r appears 1 day before the earthquake even though during a small geomagnetic activity. Since the geomagnetic condition 1 day after the earthquake is disturbed (Kp = 6), the maximum TEC enhancement is registered on June 14. Meanwhile, there are also negative anomalies on June 1 and impulsive positive ones on May 15–21 and July 11. 4. Statistical study From the results of the case studies in the previous section, it is highly suggestive that possible positive and negative GIM-TEC anomalies before and after large earthquakes occur. Therefore, it seems to be required to investigate GIM-TEC anomalies statistically. 4.1. Superposed epoch analysis Superposed epoch analysis (SEA) is a statistical method to highlight a weak but a significant signal from noisy data. This technique can reveal typical characteristics, precursors, time schedules, periodicities, and consequences of a special event (Adams et al., 2003;
Fig. 6. The distribution of the earthquakes with M P 5.5 and focal depth 640 km around Japan during May 1998–May 2010. Region (A) is given by the circle centered at (37.5°N, 140°E) with radius 1000 km. Region (B) is outside of region (A) with distance 1000–2224 km from the central point. The location of GIM pixel centered at (37.5°N, 140°E) is shown by the blue box. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
415
S. Kon et al. / Journal of Asian Earth Sciences 41 (2011) 410–420
global mean TEC response to irradiance of solar extreme ultra violet (Hocke, 2008). In this paper, SEA is applied to investigate the existence of earthquake-related ionospheric anomalies around the Japan area based on GIM-TEC. The center of study area is set at (37.5°N, 140°E), which includes the epicenters of the 2007 offshore mid-Niigata Earthquake and some significant inland earthquakes in Japan from May 13, 1998 to May 13, 2010. Fig. 6 illustrates the location of the study area by a blue box and the distribution of the earthquakes with M P 5.5 and focal depth 640 km, selected from an earthquake catalog provided by the US Geological Survey (USGS). In order to compare the earthquake occurrences and the observed anomalies, definitive criteria for extracting the anomaly should be specified. Molchanov and Hayakawa (1998) defined the anomaly as the value in excess of two times of monthly standard deviation. In order to avoid the influence of the main shock,
and the aftershocks on monthly standard deviation, Liu et al. (2000) defined the anomaly as the value in excess of the upper or lower bounds of the previous 15-days running inter-quartile ranges. In this study, we define the anomaly as the value in excess of 2r of GIM-TEC (the previous 15-days running standard deviation). If more than 10 h of the positive or negative anomalies appear in one day, it is considered the positive or negative anomalous day detected. Furthermore, to investigate the dependence of epicentral distance in earthquake-related GIM-TEC anomalies, if any, two areas with the 52M P 6.0 earthquakes are studied; one is the region (A) given by the circle centered at 37.5°N, 140°E with radius 1000 km, and the other is region (B) with distance 1000–2224 km from the central point, which is outside of the region (A), respectively. The number of the earthquakes in region (B) is adjusted to that in region (A). The selected 52 earthquakes in regions (A) and (B) are given in Tables 1 and 2, respectively.
Table 2 List of the 52 earthquakes M P 6.0 and depth 640 km that occurred in region (B) (1000 6 R < 2224 km) from May 1998–May 2010. No
Year
Month
Day
Hour
Minute
Latitude (°N)
Longitude (°E)
Depth (km)
M
Distance (km)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52
1998 1999 1999 1999 1999 2000 2000 2000 2000 2001 2001 2001 2001 2001 2002 2002 2002 2004 2004 2004 2005 2005 2006 2006 2006 2006 2006 2006 2006 2006 2006 2007 2007 2007 2007 2007 2007 2007 2008 2008 2008 2008 2009 2009 2009 2009 2009 2009 2009 2010 2010 2010
11 1 7 10 11 6 6 6 8 5 6 6 12 12 1 3 10 7 9 12 2 11 2 8 9 10 11 11 11 12 12 1 1 4 8 8 10 10 2 3 3 7 1 4 4 8 8 9 10 2 2 2
19 24 7 24 11 6 15 25 4 25 15 24 8 18 28 26 19 22 13 18 9 28 14 20 30 1 15 16 17 7 26 13 30 20 2 7 6 25 27 3 14 24 15 7 18 5 17 10 30 6 7 26
15 0 18 4 2 14 11 6 21 0 6 13 20 4 13 3 12 9 3 6 18 16 15 3 17 9 11 6 18 19 15 4 21 1 2 0 0 13 6 9 22 1 17 4 19 0 0 2 7 4 6 20
39 37 52 21 41 57 10 34 13 40 17 18 29 2 50 45 9 45 0 46 46 41 27 1 50 6 14 20 3 10 19 23 37 45 37 2 38 50 54 31 32 43 49 23 17 17 5 46 3 44 10 31
22.6 30.62 49.23 44.61 49.31 29.42 29.37 31.18 48.79 44.27 18.83 44.19 28.25 23.95 49.38 23.35 44.3 26.49 44 48.84 26.09 20.29 20.82 49.82 46.35 46.47 46.59 46.36 28.59 46.15 48.32 46.24 20.98 25.71 47.12 27.29 18.73 46.01 26.82 46.41 26.99 50.97 46.86 46.05 46.01 24.23 23.5 48.32 29.22 46.84 23.49 25.93
125.78 131.09 155.56 149.44 155.63 131.42 132.08 131.21 142.25 148.39 146.98 148.51 129.57 122.73 155.59 124.09 149.96 128.89 151.41 156.31 144 146.05 146.18 156.41 153.17 153.24 153.27 154.47 129.9 154.39 154.84 154.52 144.71 125.11 141.8 126.84 147.15 154.23 142.44 153.18 142.6 157.58 155.15 151.55 151.43 125.1 123.5 154.19 129.78 152.73 123.61 128.43
10 33 33 33 33 33 10 10 10 33 33 33 33 14 33 33 33 20 8 11 24 11 40 26 11 19 10 9 22 16 10 10 20 9 5 18 20 20 15 10 10 27 36 31 35 25 20 36 34 30 21 25
6.3 6.4 6.1 6 6.1 6.4 6.1 6 6.8 6.7 6 6 6.2 6.8 6.1 6.4 6.3 6.1 6.1 6.2 6.3 6 6.3 6 6.6 6.5 8.3 6 6.2 6.4 6 8.1 6.6 6.3 6.2 6 6.1 6.1 6.2 6.5 6 6.2 7.4 6.9 6.6 6.2 6.7 6 6.8 6 6.3 7
2138 1120 1806 1117 1816 1197 1162 1068 1266 1030 2175 1031 1412 2224 1819 2182 1126 1606 1201 1821 1319 1993 1940 1896 1465 1478 1487 1547 1363 1529 1700 1543 1884 1918 1077 1673 2191 1510 1205 1469 1189 2041 1621 1347 1337 2041 2208 1663 1319 1473 2201 1683
416
S. Kon et al. / Journal of Asian Earth Sciences 41 (2011) 410–420
We make up a dataset of GIM-TEC for 30 days before and after the selected earthquake for regions (A) and (B), respectively. The length of one dataset is totally 61 days and the middle of the dataset is the day of the earthquake. Then we investigate anomalous TEC days defined above. If the day exists, we count one. We repeat this procedure for all the selected earthquakes and we sum the count for all datasets. Then we obtain the SEA result in terms of the GIM-TEC anomaly. Moreover, in order to evaluate the statistical significance, we pick up 52 random days instead of 52 earthquake days in the whole analyzed period, and then perform the same procedure described above, that is the random-SEA for GIM-TEC, at the study pixel. We iterate the random-SEA test by 10,000 times and compute the mean (hereafter random-mean) and the corresponding standard deviation (r). In addition, the magnitude dependence of GIM-TEC anomalies are investigated with using the above SEA procedure. The M P 5.5 earthquakes are studied in this analysis. 4.2. The result 4.2.1. The dependence of the epicentral distance Fig. 7 illustrates the results of SEA for negative GIM-TEC anomalies. The blue, red and green lines show the counts of negative
GIM-TEC anomalies, random-mean, and random-mean +2r values, respectively. The gray bars show the counts for every 5 days. The upper graph is the SEA result for the 52M P 6.0 earthquakes in region (A) (0 6 R < 1000 km) and the lower one is that for the earthquakes in region (B) (1000 6 R < 2224 km). It is found although the counts of negative GIM-TEC anomalies in region (A) increase 29, 21, 19, 18, and 3 days before and 11, 17, 18, 26 days after the earthquake, they do not exceed the threshold of random-mean +2r. The similar result is obtained in the case of region (B). It is safe to say that there is no remarkable negative anomaly in ionospheric disturbance prior to the earthquakes. Fig. 8 shows the results of SEA for positive GIM-TEC anomalies. In region (A), it is found that the counts of positive GIM-TEC anomalies exceed the threshold of random-mean +2r 1 day before the earthquake. Although the counts of 2 days before the earthquake do not exceed the threshold, it is the second largest value in 61 days. Furthermore, 5-days counts of 1–5 days before the earthquake reaches 18 counts and this value is very high. Although the counts of positive GIM-TEC anomalies in region (B) increase 26, 25, and 3 days before and 21 days after the earthquake, they do not exceed random-mean +2r. 5-days counts for region (B) show no specific enhancement. These statistical results using SEA suggest the significance of positive GIM-TEC anomaly 1–5 days
Fig. 7. The result of superposed epoch analysis (SEA) for negative GIM-TEC anomalies. The blue, red and green lines show the counts of negative GIM-TEC anomalies, random-mean, and random-mean ±2r values. The gray bars show 5-days counts. The upper graph is the SEA result for the 52M P 6.0 earthquakes in region (A) (0 6 R < 1000 km) and the lower one is that for the earthquakes in region (B) (1000 6 R < 2224 km). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
S. Kon et al. / Journal of Asian Earth Sciences 41 (2011) 410–420
417
Fig. 8. The result of superposed epoch analysis (SEA) for positive GIM-TEC anomalies. The upper graph is the SEA result for the 52M P 6.0 earthquakes in region (A) (0 6 R < 1000 km) and the lower one is that for the earthquakes in region (B) (1000 6 R < 2224 km).
before M P 6.0 earthquakes within 1000 km from the epicenter in Japan. 4.2.2. The dependence of the magnitude In the Section 4.2.1, positive anomalies in region (A) are found to be important. Therefore, magnitude dependence of SEA in region (A) is examined. The step of magnitude investigation is set 0.1 from 5.5. Fig. 9 shows the result. The same SEA procedure is performed as shown in Figs. 7 and 8. With increasing the magnitude, there is a tendency of the increase of the counts 1–5 days before the earthquake. This result indicates the clear magnitude dependence of the GIM-TEC anomalies; that is, there is a significant enhancement 1–5 days prior to the 79M P 5.8 (especially 52M P 6.0) out of the 140M P 5.5 earthquakes. 5. Discussion and conclusion In this paper, ionospheric anomalies associated with earthquakes in the Japan area have been investigated with using normalized GIM-TEC (GIM-TEC) obtained from GIM. The analyzed period is from May 13, 1998 to May 13, 2010 and the study area is centered at (37.5°N, 140°E). A superposed epoch analysis (SEA) has been applied to statistically investigate the existence of
ionospheric anomalies associated with the M P 6.0 earthquakes with focal depth 640 km around Japan. The result of statistical study indicates that positive GIM-TEC anomalies appear 1–5 days before (especially 1 day before with significance) the selected earthquakes within 1000 km of the epicenter, as seen in Fig. 8. In this paper, the statistical significance of the GIM-TEC anomalies related to the earthquakes around Japan has been presented for the first time. Temporal variations of GIM-TEC for the four specific earthquakes have been examined. Although there are some positive and negative anomalies before and after the earthquakes, one can find out that positive GIM-TEC anomalies appear 1–5 days before all the earthquakes even during relatively quiet geomagnetic condition. These results of the case studies are consistent with that of the statistical study. The obtained statistical result in this paper looks like different from the previous statistical studies, which show significance of negative TEC anomalies before large earthquakes in Taiwan and China (Liu et al., 2004a,b, 2009). The difference may be caused by the geomagnetic latitude dependence. At the lower latitude, equatorial ionospheric anomaly (EIA) controls the ionospheric phenomena strongly. Although the generating mechanism is not understood, the perpendicular component of the upward electric field and the Earth’s magnetic field would be able to produce a
418
S. Kon et al. / Journal of Asian Earth Sciences 41 (2011) 410–420
Fig. 9. The result of superposed epoch analysis (SEA) for positive GIM-TEC anomalies in region (A) (0 6 R < 1000 km). (a) for 140M P 5.5 earthquakes, (b) for 119M P 5.6 earthquakes, (c) for 98M P 5.7 earthquakes, (d) for 79M P 5.8 earthquakes, (e) for 66M P 5.9 earthquakes, (f) for 52M P 6.0 earthquakes.
westward plasma E B drift (Kelley, 1989) which in turn results in the reductions of the GPS-TEC prior to the large earthquakes (Liu et al., 2009). However, the mid-latitudes such as the Japan area the ionospheric variation is different. As for earthquake-related ionospheric phenomena in midlatitude, Zakharenkova et al. (2007) reports the extreme TEC
enhancement 1–5 days before the 2003 Hokkaido earthquake (September 25, 2003, M8.3, 41.81°N, 143.91°E). Moreover, Zhao et al. (2008) and Liu et al. (2009) observe the extreme and localized TEC enhancement in afternoon period 3 days before the 2008 Wenchuan earthquake (May 12, 2008, M7.9, 31.0°N, 103.4°E) in China region. For the Wenchuan earthquake, anomalous enhancement of
S. Kon et al. / Journal of Asian Earth Sciences 41 (2011) 410–420
the electron density is also observed by ionosondes (Zhao et al., 2008). These results of case studies in mid-latitude seem to be consistent with our result. The magnitude dependence of the positive GIM-TEC anomalies has been detected. This result is quite similar in the magnitude dependence of TEC reduction in the China region (Liu et al., 2009). They show the significant increase of TEC reduction 3–5 days before the earthquake. Furthermore, Figs. 7 and 8 show the epicentral distance dependence. The existence of these dependences is highly suggestive of the earthquake-related ionospheric anomalies. This paper has demonstrated the existence of the positive TEC anomaly 1–5 days prior to the large earthquakes in mid-latitude, especially around Japan. This tendency is different from that in equatorial region. Since it is reasonable to consider a physical mechanism of lithosphere–atmosphere–ionosphere coupling is unique, further statistical tests in other mid-latitude regions are essential. If similar results are obtained, positive TEC anomaly could be dominant in mid-latitude. Acknowledgements The authors wish to thank Prof. J.Y. Liu in Institute of Space Science, National Central University for useful comments and discussions. The GIM data were isolated from the Center for Orbit Determination in Europe (CODE). The GPS data observed by ground-based GPS receivers of GEONET were provided by the Geographical Survey Institute of Japan. The GAMIT software developed at the Massachusetts Institute of Technology (MIT) and the Scripps Institution of Oceanography (SIO) was used to extract the GPS-TEC from GPS data. The Dst index and the Kp index were supplied by World Data Center (WDC) for Geomagnetism, Kyoto. The F10.7 solar flux data were provided by the Solar Radio Monitoring Program operated jointly by the National Research Council and the Canadian Space Agency. Also we thank the United States Geological Survey for providing the earthquake data. This research was partly supported by a Grand-in-Aid for Scientific Research of Japan Society for Promotion of Science (19403002) and National Institute of Information and Communication Technology (R&D promotion funding International Joint Research). References ~ o-like Adams, J.B., Mann, M.E., Ammann, C.M., 2003. Proxy evidence for an el Nin response to volcanic forcing. Nature 426 (6964), 274–278. Afraimovich, E.L., Astafyeva, E.I., 2008. TEC anomalies – local TEC changes prior to earthquakes or TEC response to solar and geomagnetic activity changes? Earth, Planets and Space 60, 961–966. Afraimovich, E.L., Voeykov, S.V., Perevalova, N.P., Ratovsky, K.G., 2008. Large-scale traveling ionospheric disturbances of auroral origin according to the data of the GPS network and ionosondes. Advances in Space Research 42, 1213–1217. Chen, Y.I., Liu, J.Y., Tsai, Y.B., Chen, C.S., 2004. Statistical tests for pre-earthquake ionospheric anomaly. Terrestrial, Atmospheric and Oceanic Sciences 15, 385–396. Chree, C., 1912. Some phenomena of sunspots and of terrestrial magnetism, at Kew observatory. Philosophical Transactions Royal Society London Series A 212, 75. Chuo, Y.J., Chen, Y.I., Liu, J.Y., Pulinets, S.A., 2001. Ionospheric foF2 variations prior to strong earthquakes in Taiwan area. Advances in Space Research 27, 1305–1310. doi:10.1016/S0273-1177(01)00209-5. Chuo, Y.J., Liu, J.Y., Pulinets, S.A., Chen, Y.I., 2002. The ionospheric perturbations prior to the Chi-Chi and Chia-Yi earthquakes. Journal of Geodynamics 33, 509– 517. doi:10.1016/S0264-3707(02)00011-X. Dautermann, T., Calais, E., Haase, J., Garrison, J., 2007. Investigation of ionospheric electron content variations before earthquakes in southern California, 2003– 2004. Journal of Geophysical Research 112, B02106. doi:10.1029/2006JB004447. Davies, K., 1990. Ionospheric Radio. Peter Peregrinus Ltd., London, p. 580. Haung, Y., Lee, Y., 1975. On the variation of solar wind velocity following solar flares. Journal of Geophysical Research 80, 2863–2865. Hayakawa, M. (Ed.), 1999. Atmospheric and Ionospheric Electromagnetic Phenomena with Earthquakes. Terra Sci. Pub. Co., Tokyo, p. 996. Hayakawa, M., Fujinawa, Y. (Eds.), 1994. Electromagnetic Phenomena Related to Earthquake Predication. Terra Sci. Pub. Co., Tokyo, p. 667. Hayakawa, M., Molchanov, O.A. (Eds.), 2002. Seismo Electromagnetics: Lithosphere–Atmosphere–Ionosphere Coupling. TERRAPUB, Tokyo, p. 477.
419
Hobara, Y., Parrot, M., 2005. Ionospheric perturbations linked to a very powerful seismic event. Journal of Atmospheric and Solar–Terrestrial Physics 67, 677– 685. doi:10.1016/j.jastp. 2005.02.006. Hocke, K., 2008. Oscillations of global mean TEC. Journal of Geophysical Research A: Space Physics 113, A04302. Kelley, M.C., 1989. The Earth’s Ionosphere. Academic Press, San Diego, p. 487. Liperovskaya, E.V., Parrot, M., Bogdanov, V.V., Meister, C.-V., Rodkin, M.V., Liperovsky, V.A., 2006. On variations of foF2 and F-spread before strong earthquakes in Japan. Natural Hazards and Earth System Sciences 6, 735–739.
420
S. Kon et al. / Journal of Asian Earth Sciences 41 (2011) 410–420
Zakharenkova, I.E., Shagimuratov, I.I., Tepenitzina, N.Y., Krankowski, A., 2008. Anomalous modification of the ionospheric total electron content prior to the 26 September 2005 Peru earthquake. Journal of Atmospheric and Solar– Terrestrial Physics 70, 1919–1928.
Zhao, B., Wang, M., Yu, T., Wan, W., Lei, J., Liu, L., Ning, B., 2008. Is an unusual large enhancement of ionospheric electron density linked with the 2008 great Wenchuan earthquake? Journal of Geophysical Research A: Space Physics 113, A11304. doi:10.1029/2008JA013613.
Contents lists available at ScienceDirect
Journal of Asian Earth Sciences journal homepage: www.elsevier.com/locate/jseaes
Ionospheric anomalies possibly associated with M P 6.0 earthquakes in the Japan area during 1998–2010: Case studies and statistical study Shimpei Kon, Masahide Nishihashi, Katsumi Hattori ⇑ Graduate School of Science, Chiba University, 1-33, Yayoi, Inage, Chiba 263-8522, Japan
a r t i c l e
i n f o
Article history: Available online 31 October 2010 Keywords: Ionospheric anomaly Total electron content Earthquake-related ionospheric anomaly Statistical analysis Case studies M P 6.0 earthquakes around Japan
a b s t r a c t In this paper, we examine pre-earthquake ionospheric anomalies in time series and perform a statistical test by using total electron content (TEC) derived from global ionosphere maps (GIM) around the Japan area for the first time. The normalized GIM-TEC (GIM-TEC), which is computed based on 15 days backward running mean of GIM-TEC, have been investigated for minimizing possible confounding effects of consecutive earthquakes and identify the abnormal signals. Meanwhile, to reduce the effect of strong geomagnetic activities such as geomagnetic storms, the criterion for removing the GIM-TEC data have been adapted; that is when Dst index exceeds 60 nT. Temporal variations of GIM-TEC for large and destructive earthquakes in Japan have been studied; which are the 2004 mid-Niigata Prefecture Earthquake (M6.8), its aftershock (M6.1), the 2007 offshore mid-Niigata Earthquake (M6.8), and the 2008 Iwate–Miyagi Nairiku Earthquake (M7.2). Although there are some positive and negative TEC anomalies before and after the four earthquakes, there is a tendency that positive TEC anomalies appear 1–5 days before all the above earthquakes even during the quiet geomagnetic condition. Superposed epoch analyses have been performed for the statistical analysis of TEC anomalies associated with M P 6.0 earthquakes during the 12-year period of May 1998–May 2010. The statistical result indicates the significance of the positive TEC anomalies 1–5 days before earthquakes within 1000 km from the epicenter around Japan. Ó 2010 Elsevier Ltd. All rights reserved.
1. Introduction A large number of papers have been reported on anomalous electromagnetic phenomena possibly associated with large earthquakes (e.g., Hayakawa and Fujinawa, 1994; Hayakawa, 1999; Hayakawa and Molchanov, 2002; Pulinets and Boyarchuk, 2004; Molchanov and Hayakawa, 2008). One of the most frequent studies among them is to investigate the relationship between ionospheric anomalies and earthquake activities. Scientists have observed anomalous variations appearing a few days before some large earthquakes in the ionospheric plasma frequency foF2 (or F2-peak electron density NmF2) recorded by ionosondes (Pulinets et al., 1994; Pulinets, 1998; Liu et al., 2000, 2004a,b, 2006, 2008; Chuo et al., 2001, 2002; Chen et al., 2004; Hobara and Parrot, 2005; Liperovskaya et al., 2006). Besides foF2 and NmF2, total electron content (TEC) derived by ground and satellite based receivers of the global positioning system (GPS) is also used to observe ionospheric condition (e.g., Sardon et al., 1994; Liu et al., 1996). The NmF2 and GPS-TEC are, in general, highly correlated (Liu et al., 2001, 2004b; Nishihashi et al., 2009). This fact suggests that the
⇑ Corresponding author. Tel.: +81 43 290 2801; fax: +81 43 290 2859. E-mail address: [email protected] (K. Hattori). 1367-9120/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.jseaes.2010.10.005
GPS-TEC can be employed to monitor the ionospheric sciences including preseismic perturbations. Recently, scientists have conducted not only a case study but also a statistical study of the preseismic TEC anomalies (e.g., Nishihashi et al., 2009; Liu et al., 2001, 2004a,b, 2008, 2009; Pulinets et al., 2004, 2007; Dautermann et al., 2007; Zhao et al., 2008). Liu et al. (2004a) reported that the GPS-TEC pronouncedly decreases within 1–5 days prior to 20M P 6.0 earthquakes in Taiwan during September 1999–December 2002. Moreover, Liu et al. (2009) reported that the anomalous TEC reductions appear 3–5 days before 17M P 6.3 earthquakes in China region from May 1, 1998 to April 30, 2008 and the result is consistent with the case study of the 2008 Wenchuan earthquake (M7.9). In addition, an extreme increase of TEC for all the analyzed period is recorded in the afternoon of 3 days before the earthquake. Japan is recognized to be one of the most seismically active regions in the world and a dense GPS network has been established. Although previous studies suggested that significant TEC anomalies would accompany prior to larger earthquakes around Japan (Zakharenkova et al., 2007), statistical studies have not been reported. Therefore, TEC variation around Japan is investigated in this paper. At first, we investigate the TEC variation derived from the global ionospheric maps (GIM) before and after some large earthquakes; which are the 2004 mid-Niigata Prefecture Earthquake
S. Kon et al. / Journal of Asian Earth Sciences 41 (2011) 410–420
(M6.8), its aftershock (M6.1), the 2007 offshore mid-Niigata Earthquake (M6.8), and the 2008 Iwate–Miyagi Nairiku Earthquake (M7.2). Then, we perform a statistical analysis of earthquake-related TEC anomalies around Japan for 12-year from May 1998 to May 2010.
411
(Kelley, 1989; Davies, 1990). To reduce these effects, a criterion for the ionospheric disturbed period is introduced. That is the 2 days after Dst index exceed 60 nT. Therefore, we remove the 2 days TEC data after storm onsets in this paper. Dst index derived from four geomagnetic observations is used to represents the geomagnetic disturbances at the lower geomagnetic latitudes.
2. Data and normalized GIM-TEC (GIM-TEC) Recently, many scientists utilize the global ionosphere maps (GIM) containing grid data of the vertical TEC to investigate ionospheric phenomena (Mendillo et al., 2005; Afraimovich et al., 2008; Hocke, 2008). Therefore, the GIM is also considered a useful tool to analyze earthquake-related TEC variations (Nishihashi et al., 2009; Zakharenkova et al., 2006, 2008; Afraimovich and Astafyeva, 2008; Zhao et al., 2008; Liu et al., 2009; Yu et al., 2009). In this paper, the GIM derived by the Center for Orbit Determination in Europe (CODE; ftp://ftp.unibe.ch/aiub/CODE/) is selected. The vertical TEC is modeled in a solar-geomagnetic reference frame using a spherical harmonics expansion. In 2009, the degree and order of the spherical harmonics expansion were 15 and 15, respectively. Piece-wise linear functions are used for representation in the time domain. The spatial resolution is 2.5° in latitude, and 5° in longitude and the temporal resolution is 2 h. Schaer (1999) describes the computational algorithm in detail. In this paper, we linearly interpolate yielding a 1-h resolution at a grid point. Remember that the TEC derived from GIM (hereafter GIM-TEC) is obtained through the spherical harmonics expansion and the linear interpolation. Fig. 1 shows the relationship between the GIM-TEC at the grid of 140°E–37.5°N and GPS-TEC observed at the 0920 GPS station (140.73°E, 37.82°N) of GPS Earth Observation Network (GEONET) in 2007. Their correlation is found to be about 0.73.To minimize possible confounding effects of consecutive earthquakes and properly identify the abnormal signals, we computed the mean GIM-TEC (GIM—TEC) values for the previous 15 days, and the corresponding standard deviation (r) as a reference at specific times. Then, we derived the normalized GIM-TEC (GIM-TEC) values by the following equation,
GIM TEC ðtÞ ¼
3. Case studies In this section, we would like to demonstrate some case studies on GIM-TEC temporal variation. We select the intense, destructive, and significant earthquakes during the last decade; which are the 2004 mid-Niigata Prefecture Earthquake (M6.8), its large aftershock (M6.1), the 2007 offshore mid-Niigata Earthquake (M6.8), and the 2008 Iwate–Miyagi Nairiku Earthquake (M7.2). Fig. 2 draws the epicenters with magnitude and detailed information is described in Table 1 (Nos. 26, 27, 43, and 45).
3.1. The 2004 mid-Niigata Prefecture Earthquake The 2004 mid-Niigata Prefecture Earthquake (M6.8) occurred at 08:56 UT, October 23, 2004. The epicenter is located at 37.29 N° and 138.87 E° and the depth is 23 km. The source mechanism is reverse faulting. It was one of the most damaging earthquakes in Japan in last 10 years. Seventeen aftershocks with M P 5.0 occurred and the large aftershock (M6.1) occurred at 01:40 UT, October 27, 2004. Fifty-one people were killed, 4805 people were injured, and 16,900 houses were damaged.
GIM TECðtÞ GIM TECðtÞ : rðtÞ
Meanwhile, strong geomagnetic activities such as geomagnetic storm can cause ionospheric disturbances and TEC could be perturbed about a few hours to 2 days after geomagnetic storm onsets
Fig. 1. The correlation between the GIM-TEC at the grid of 140°E–37.5°N and GPSTEC observed at 0920 GPS station of GEONET in 2007. The correlation coefficient is about 0.73.
Fig. 2. The map of epicenters. (a) The 2004 mid-Niigata Prefecture Earthquake, (b) its aftershock, (c) the 2007 offshore mid-Niigata Earthquake, and (d) the 2008 Iwate–Miyagi Nairiku Earthquake.
412
S. Kon et al. / Journal of Asian Earth Sciences 41 (2011) 410–420
Table 1 List of the 52 earthquakes M P 6.0 and depth 640 km that occurred in region (A) (0 6 R < 1000 km) from May 1998–May 2010. No
Year
Month
Day
Hour
Minute
Latitude (°N)
Longitude (°E)
Depth (km)
M
Distance (km)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52
1998 1998 2000 2000 2000 2000 2000 2000 2001 2001 2001 2002 2002 2003 2003 2003 2003 2003 2003 2003 2004 2004 2004 2004 2004 2004 2004 2004 2004 2004 2005 2005 2005 2005 2005 2005 2005 2005 2006 2006 2007 2007 2007 2008 2008 2008 2008 2008 2008 2009 2009 2010
5 5 7 7 7 10 10 11 3 4 8 8 11 7 9 9 9 10 10 12 4 5 9 9 9 10 10 11 11 12 1 3 8 8 8 10 11 12 10 10 2 3 7 5 6 7 7 9 12 6 8 3
14 30 1 15 30 3 6 13 23 14 13 20 3 25 25 26 29 8 31 29 3 29 5 6 8 23 27 11 28 6 19 20 16 24 30 19 14 2 10 23 17 25 16 7 13 19 21 11 20 5 10 14
18 18 7 1 12 4 4 15 11 23 20 10 3 22 19 20 2 9 1 1 23 20 14 23 14 8 1 10 18 14 6 1 2 10 18 11 21 13 23 21 0 0 1 16 23 2 11 0 10 3 20 8
56 18 1 30 25 13 30 57 30 27 11 59 37 13 50 38 36 6 6 30 2 56 57 29 58 56 40 2 32 15 11 53 46 15 10 44 38 13 58 17 2 41 13 45 43 39 30 20 29 30 7 8
40.25 39.03 34.22 34.32 33.9 40.28 35.46 42.49 44.07 30.09 41.05 30.99 38.89 38.42 41.81 41.99 42.45 42.65 37.81 42.42 36.43 34.25 33.18 33.21 33.14 37.23 37.28 42.14 43.01 42.9 34.06 33.81 38.28 38.56 38.48 36.4 38.11 38.09 37.2 29.35 41.79 37.34 37.53 36.16 39.03 37.55 37.19 41.89 36.54 41.82 34.74 37.74
143.25 143.44 139.13 139.26 139.38 143.12 133.13 144.76 148.05 141.77 142.31 141.97 141.98 141 143.91 144.58 144.38 144.57 142.62 144.61 141.01 141.41 137.07 137.23 137.2 138.78 138.88 144.34 145.12 145.23 141.49 130.13 142.04 142.99 143.18 140.84 144.9 142.12 142.66 140.27 143.55 136.59 138.45 141.53 140.88 142.21 142.05 143.75 142.43 143.45 138.26 141.59
33 33 10 10 10 33 10 33 33 10 38 9 39 6 27 33 25 32 10 33 31 16 10 10 21 16 14 32 39 35 27 10 36 10 21 32 11 29 9 11 31 8 12 27 7 22 22 25 19 29 40 32
6 6.2 6.1 6.1 6.5 6.3 6.7 6 6 6 6.4 6.3 6.4 6.1 8.3 6 6.5 6.7 7 6.1 6 6.5 7.4 6.6 6.1 6.6 6 6.1 7 6.8 6.6 6.6 7.2 6.1 6.1 6.3 7 6.5 6 6.4 6 6.7 6.6 6.9 6.9 7 6 6.8 6.3 6.4 6.1 6.5
415 345 371 358 403 410 655 686 995 837 440 744 231 134 584 633 664 691 233 673 148 381 547 538 546 112 101 634 750 746 404 981 198 287 299 142 436 197 238 903 565 302 137 201 186 195 184 583 240 563 343 142
Fig. 3 illustrates the variation of GIM-TEC at a certain pixel above the epicenter together with that of Kp index, Dst index, and F10.7 solar flux. Kp index derived from 13 geomagnetic observations in subauroral region is generally used to evaluate geomagnetic disturbances in high and middle geomagnetic latitudes. The F10.7 index reveals the solar activity level based on the solar radio flux at a wavelength of 10.7 cm. Two vertical lines show the origin time of earthquakes; the first one is the 2004 mid-Niigata Prefecture Earthquake with M6.8 and the latter is its aftershock with M6.1. The period colored by gray is geomagnetically disturbed period according to the criterion of Dst less than 60 nT. Positive GIM-TEC anomalies are found 2–3 days before the 2004 mid-Niigata Prefecture Earthquake and its aftershock even in the relatively quiet geomagnetic condition. Their duration exceeding +2r are about 10 and 18 h, respectively. Meanwhile, there are negative anomalies with 12 h on September 24 and some
impulsive negative anomalies on October 4–6 and October 16–17. For positive anomalies, there are one with 7 h on September 28, some impulsive ones on October 10, and one with 5 h on October 29. 3.2. The 2007 offshore mid-Niigata Earthquake The 2007 offshore mid-Niigata Earthquake (M6.8) occurred at 01:13 UT, July 16, 2007. The epicenter is located at 37.56°N and 138.61°E and the depth is 17 km. The source mechanism is reverse faulting. The largest aftershock occurred at 06:37 UT on the same day. Its magnitude and depth are 5.8 and 23 km, respectively. The number of causality is 15 and 2415 people were injured. It was a serious shock that a nuclear power station at Kashiwazaki was damaged by the earthquake severely and the operation was automatically stopped. That is the first time for human kind. The routine operation restarts partly in 2009 at last.
S. Kon et al. / Journal of Asian Earth Sciences 41 (2011) 410–420
413
Fig. 3. The variation of GIM-TEC at the study pixel together with that of Kp index, Dst index, and F10.7 solar flux. Two vertical lines show the origin time of earthquakes; the first one is the 2004 mid-Niigata Prefecture Earthquake (M6.8) and the latter is its aftershock (M6.1). The period colored by gray is geomagnetically disturbed period according to the criterion of Dst less than 60 nT.
Fig. 4. The variation of GIM-TEC 30 days before and after the 2007 offshore mid-Niigata Earthquake.
Fig. 4 shows the variation of GIM-TEC associated with the 2007 offshore mid-Niigata Earthquake in the same manner with Fig. 3. There are some negative anomalies on June 16, June 26, July 5, and July 14–15 and positive anomalies on June 30, July 4, July 11,
and July 26. Among of them, it is found that strong positive GIMTEC anomalies 5 days before the earthquake. Their total duration exceeding +2r is about 10 h and the largest value reaches +4r. Furthermore, strong reductions appear 3–4 days before the
414
S. Kon et al. / Journal of Asian Earth Sciences 41 (2011) 410–420
Fig. 5. The variation of GIM-TEC 30 days before and after the 2008 Iwate–Miyagi Nairiku Earthquake.
earthquakes. Their duration exceeding 2r is over 18 h and maximum reduction is 3.9r even in a geomagnetically quiet condition. 3.3. The 2008 Iwate–Miyagi Nairiku Earthquake
Hocke, 2008). It has been utilized very profitably for solar–terrestrial, space and atmospheric physics (e.g., Chree, 1912; Haung and Lee, 1975; Taylor et al., 1994; Singh and Badruddin, 2006). For investigation of TEC, the SEA is employed for the analysis of
The 2008 Iwate–Miyagi Nairiku Earthquake (M7.2) occurred at 23:43 UT, June 13, 2008. The epicenter is located at 39.03°N and 140.88°E and the depth is 8 km. The source mechanism is reverse faulting. The earthquake triggered landslide. Several micro-earthquakes (M < 1.4) occurred about 40 min before the earthquake near the epicenter and three aftershocks with M > 5 occurred on June 14, 2008 (M5.7 and M5.2) and June 16, 2008 (M5.3). Over 10 people died and about 450 people were seriously injured. Fig. 5 indicates the variation of GIM-TEC associated as shown in the 2008 Iwate–Miyagi Nairiku Earthquake in the same manner as shown in Figs. 3 and 4. It is found that the large TEC enhancement of +4.5r appears 1 day before the earthquake even though during a small geomagnetic activity. Since the geomagnetic condition 1 day after the earthquake is disturbed (Kp = 6), the maximum TEC enhancement is registered on June 14. Meanwhile, there are also negative anomalies on June 1 and impulsive positive ones on May 15–21 and July 11. 4. Statistical study From the results of the case studies in the previous section, it is highly suggestive that possible positive and negative GIM-TEC anomalies before and after large earthquakes occur. Therefore, it seems to be required to investigate GIM-TEC anomalies statistically. 4.1. Superposed epoch analysis Superposed epoch analysis (SEA) is a statistical method to highlight a weak but a significant signal from noisy data. This technique can reveal typical characteristics, precursors, time schedules, periodicities, and consequences of a special event (Adams et al., 2003;
Fig. 6. The distribution of the earthquakes with M P 5.5 and focal depth 640 km around Japan during May 1998–May 2010. Region (A) is given by the circle centered at (37.5°N, 140°E) with radius 1000 km. Region (B) is outside of region (A) with distance 1000–2224 km from the central point. The location of GIM pixel centered at (37.5°N, 140°E) is shown by the blue box. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
415
S. Kon et al. / Journal of Asian Earth Sciences 41 (2011) 410–420
global mean TEC response to irradiance of solar extreme ultra violet (Hocke, 2008). In this paper, SEA is applied to investigate the existence of earthquake-related ionospheric anomalies around the Japan area based on GIM-TEC. The center of study area is set at (37.5°N, 140°E), which includes the epicenters of the 2007 offshore mid-Niigata Earthquake and some significant inland earthquakes in Japan from May 13, 1998 to May 13, 2010. Fig. 6 illustrates the location of the study area by a blue box and the distribution of the earthquakes with M P 5.5 and focal depth 640 km, selected from an earthquake catalog provided by the US Geological Survey (USGS). In order to compare the earthquake occurrences and the observed anomalies, definitive criteria for extracting the anomaly should be specified. Molchanov and Hayakawa (1998) defined the anomaly as the value in excess of two times of monthly standard deviation. In order to avoid the influence of the main shock,
and the aftershocks on monthly standard deviation, Liu et al. (2000) defined the anomaly as the value in excess of the upper or lower bounds of the previous 15-days running inter-quartile ranges. In this study, we define the anomaly as the value in excess of 2r of GIM-TEC (the previous 15-days running standard deviation). If more than 10 h of the positive or negative anomalies appear in one day, it is considered the positive or negative anomalous day detected. Furthermore, to investigate the dependence of epicentral distance in earthquake-related GIM-TEC anomalies, if any, two areas with the 52M P 6.0 earthquakes are studied; one is the region (A) given by the circle centered at 37.5°N, 140°E with radius 1000 km, and the other is region (B) with distance 1000–2224 km from the central point, which is outside of the region (A), respectively. The number of the earthquakes in region (B) is adjusted to that in region (A). The selected 52 earthquakes in regions (A) and (B) are given in Tables 1 and 2, respectively.
Table 2 List of the 52 earthquakes M P 6.0 and depth 640 km that occurred in region (B) (1000 6 R < 2224 km) from May 1998–May 2010. No
Year
Month
Day
Hour
Minute
Latitude (°N)
Longitude (°E)
Depth (km)
M
Distance (km)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52
1998 1999 1999 1999 1999 2000 2000 2000 2000 2001 2001 2001 2001 2001 2002 2002 2002 2004 2004 2004 2005 2005 2006 2006 2006 2006 2006 2006 2006 2006 2006 2007 2007 2007 2007 2007 2007 2007 2008 2008 2008 2008 2009 2009 2009 2009 2009 2009 2009 2010 2010 2010
11 1 7 10 11 6 6 6 8 5 6 6 12 12 1 3 10 7 9 12 2 11 2 8 9 10 11 11 11 12 12 1 1 4 8 8 10 10 2 3 3 7 1 4 4 8 8 9 10 2 2 2
19 24 7 24 11 6 15 25 4 25 15 24 8 18 28 26 19 22 13 18 9 28 14 20 30 1 15 16 17 7 26 13 30 20 2 7 6 25 27 3 14 24 15 7 18 5 17 10 30 6 7 26
15 0 18 4 2 14 11 6 21 0 6 13 20 4 13 3 12 9 3 6 18 16 15 3 17 9 11 6 18 19 15 4 21 1 2 0 0 13 6 9 22 1 17 4 19 0 0 2 7 4 6 20
39 37 52 21 41 57 10 34 13 40 17 18 29 2 50 45 9 45 0 46 46 41 27 1 50 6 14 20 3 10 19 23 37 45 37 2 38 50 54 31 32 43 49 23 17 17 5 46 3 44 10 31
22.6 30.62 49.23 44.61 49.31 29.42 29.37 31.18 48.79 44.27 18.83 44.19 28.25 23.95 49.38 23.35 44.3 26.49 44 48.84 26.09 20.29 20.82 49.82 46.35 46.47 46.59 46.36 28.59 46.15 48.32 46.24 20.98 25.71 47.12 27.29 18.73 46.01 26.82 46.41 26.99 50.97 46.86 46.05 46.01 24.23 23.5 48.32 29.22 46.84 23.49 25.93
125.78 131.09 155.56 149.44 155.63 131.42 132.08 131.21 142.25 148.39 146.98 148.51 129.57 122.73 155.59 124.09 149.96 128.89 151.41 156.31 144 146.05 146.18 156.41 153.17 153.24 153.27 154.47 129.9 154.39 154.84 154.52 144.71 125.11 141.8 126.84 147.15 154.23 142.44 153.18 142.6 157.58 155.15 151.55 151.43 125.1 123.5 154.19 129.78 152.73 123.61 128.43
10 33 33 33 33 33 10 10 10 33 33 33 33 14 33 33 33 20 8 11 24 11 40 26 11 19 10 9 22 16 10 10 20 9 5 18 20 20 15 10 10 27 36 31 35 25 20 36 34 30 21 25
6.3 6.4 6.1 6 6.1 6.4 6.1 6 6.8 6.7 6 6 6.2 6.8 6.1 6.4 6.3 6.1 6.1 6.2 6.3 6 6.3 6 6.6 6.5 8.3 6 6.2 6.4 6 8.1 6.6 6.3 6.2 6 6.1 6.1 6.2 6.5 6 6.2 7.4 6.9 6.6 6.2 6.7 6 6.8 6 6.3 7
2138 1120 1806 1117 1816 1197 1162 1068 1266 1030 2175 1031 1412 2224 1819 2182 1126 1606 1201 1821 1319 1993 1940 1896 1465 1478 1487 1547 1363 1529 1700 1543 1884 1918 1077 1673 2191 1510 1205 1469 1189 2041 1621 1347 1337 2041 2208 1663 1319 1473 2201 1683
416
S. Kon et al. / Journal of Asian Earth Sciences 41 (2011) 410–420
We make up a dataset of GIM-TEC for 30 days before and after the selected earthquake for regions (A) and (B), respectively. The length of one dataset is totally 61 days and the middle of the dataset is the day of the earthquake. Then we investigate anomalous TEC days defined above. If the day exists, we count one. We repeat this procedure for all the selected earthquakes and we sum the count for all datasets. Then we obtain the SEA result in terms of the GIM-TEC anomaly. Moreover, in order to evaluate the statistical significance, we pick up 52 random days instead of 52 earthquake days in the whole analyzed period, and then perform the same procedure described above, that is the random-SEA for GIM-TEC, at the study pixel. We iterate the random-SEA test by 10,000 times and compute the mean (hereafter random-mean) and the corresponding standard deviation (r). In addition, the magnitude dependence of GIM-TEC anomalies are investigated with using the above SEA procedure. The M P 5.5 earthquakes are studied in this analysis. 4.2. The result 4.2.1. The dependence of the epicentral distance Fig. 7 illustrates the results of SEA for negative GIM-TEC anomalies. The blue, red and green lines show the counts of negative
GIM-TEC anomalies, random-mean, and random-mean +2r values, respectively. The gray bars show the counts for every 5 days. The upper graph is the SEA result for the 52M P 6.0 earthquakes in region (A) (0 6 R < 1000 km) and the lower one is that for the earthquakes in region (B) (1000 6 R < 2224 km). It is found although the counts of negative GIM-TEC anomalies in region (A) increase 29, 21, 19, 18, and 3 days before and 11, 17, 18, 26 days after the earthquake, they do not exceed the threshold of random-mean +2r. The similar result is obtained in the case of region (B). It is safe to say that there is no remarkable negative anomaly in ionospheric disturbance prior to the earthquakes. Fig. 8 shows the results of SEA for positive GIM-TEC anomalies. In region (A), it is found that the counts of positive GIM-TEC anomalies exceed the threshold of random-mean +2r 1 day before the earthquake. Although the counts of 2 days before the earthquake do not exceed the threshold, it is the second largest value in 61 days. Furthermore, 5-days counts of 1–5 days before the earthquake reaches 18 counts and this value is very high. Although the counts of positive GIM-TEC anomalies in region (B) increase 26, 25, and 3 days before and 21 days after the earthquake, they do not exceed random-mean +2r. 5-days counts for region (B) show no specific enhancement. These statistical results using SEA suggest the significance of positive GIM-TEC anomaly 1–5 days
Fig. 7. The result of superposed epoch analysis (SEA) for negative GIM-TEC anomalies. The blue, red and green lines show the counts of negative GIM-TEC anomalies, random-mean, and random-mean ±2r values. The gray bars show 5-days counts. The upper graph is the SEA result for the 52M P 6.0 earthquakes in region (A) (0 6 R < 1000 km) and the lower one is that for the earthquakes in region (B) (1000 6 R < 2224 km). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
S. Kon et al. / Journal of Asian Earth Sciences 41 (2011) 410–420
417
Fig. 8. The result of superposed epoch analysis (SEA) for positive GIM-TEC anomalies. The upper graph is the SEA result for the 52M P 6.0 earthquakes in region (A) (0 6 R < 1000 km) and the lower one is that for the earthquakes in region (B) (1000 6 R < 2224 km).
before M P 6.0 earthquakes within 1000 km from the epicenter in Japan. 4.2.2. The dependence of the magnitude In the Section 4.2.1, positive anomalies in region (A) are found to be important. Therefore, magnitude dependence of SEA in region (A) is examined. The step of magnitude investigation is set 0.1 from 5.5. Fig. 9 shows the result. The same SEA procedure is performed as shown in Figs. 7 and 8. With increasing the magnitude, there is a tendency of the increase of the counts 1–5 days before the earthquake. This result indicates the clear magnitude dependence of the GIM-TEC anomalies; that is, there is a significant enhancement 1–5 days prior to the 79M P 5.8 (especially 52M P 6.0) out of the 140M P 5.5 earthquakes. 5. Discussion and conclusion In this paper, ionospheric anomalies associated with earthquakes in the Japan area have been investigated with using normalized GIM-TEC (GIM-TEC) obtained from GIM. The analyzed period is from May 13, 1998 to May 13, 2010 and the study area is centered at (37.5°N, 140°E). A superposed epoch analysis (SEA) has been applied to statistically investigate the existence of
ionospheric anomalies associated with the M P 6.0 earthquakes with focal depth 640 km around Japan. The result of statistical study indicates that positive GIM-TEC anomalies appear 1–5 days before (especially 1 day before with significance) the selected earthquakes within 1000 km of the epicenter, as seen in Fig. 8. In this paper, the statistical significance of the GIM-TEC anomalies related to the earthquakes around Japan has been presented for the first time. Temporal variations of GIM-TEC for the four specific earthquakes have been examined. Although there are some positive and negative anomalies before and after the earthquakes, one can find out that positive GIM-TEC anomalies appear 1–5 days before all the earthquakes even during relatively quiet geomagnetic condition. These results of the case studies are consistent with that of the statistical study. The obtained statistical result in this paper looks like different from the previous statistical studies, which show significance of negative TEC anomalies before large earthquakes in Taiwan and China (Liu et al., 2004a,b, 2009). The difference may be caused by the geomagnetic latitude dependence. At the lower latitude, equatorial ionospheric anomaly (EIA) controls the ionospheric phenomena strongly. Although the generating mechanism is not understood, the perpendicular component of the upward electric field and the Earth’s magnetic field would be able to produce a
418
S. Kon et al. / Journal of Asian Earth Sciences 41 (2011) 410–420
Fig. 9. The result of superposed epoch analysis (SEA) for positive GIM-TEC anomalies in region (A) (0 6 R < 1000 km). (a) for 140M P 5.5 earthquakes, (b) for 119M P 5.6 earthquakes, (c) for 98M P 5.7 earthquakes, (d) for 79M P 5.8 earthquakes, (e) for 66M P 5.9 earthquakes, (f) for 52M P 6.0 earthquakes.
westward plasma E B drift (Kelley, 1989) which in turn results in the reductions of the GPS-TEC prior to the large earthquakes (Liu et al., 2009). However, the mid-latitudes such as the Japan area the ionospheric variation is different. As for earthquake-related ionospheric phenomena in midlatitude, Zakharenkova et al. (2007) reports the extreme TEC
enhancement 1–5 days before the 2003 Hokkaido earthquake (September 25, 2003, M8.3, 41.81°N, 143.91°E). Moreover, Zhao et al. (2008) and Liu et al. (2009) observe the extreme and localized TEC enhancement in afternoon period 3 days before the 2008 Wenchuan earthquake (May 12, 2008, M7.9, 31.0°N, 103.4°E) in China region. For the Wenchuan earthquake, anomalous enhancement of
S. Kon et al. / Journal of Asian Earth Sciences 41 (2011) 410–420
the electron density is also observed by ionosondes (Zhao et al., 2008). These results of case studies in mid-latitude seem to be consistent with our result. The magnitude dependence of the positive GIM-TEC anomalies has been detected. This result is quite similar in the magnitude dependence of TEC reduction in the China region (Liu et al., 2009). They show the significant increase of TEC reduction 3–5 days before the earthquake. Furthermore, Figs. 7 and 8 show the epicentral distance dependence. The existence of these dependences is highly suggestive of the earthquake-related ionospheric anomalies. This paper has demonstrated the existence of the positive TEC anomaly 1–5 days prior to the large earthquakes in mid-latitude, especially around Japan. This tendency is different from that in equatorial region. Since it is reasonable to consider a physical mechanism of lithosphere–atmosphere–ionosphere coupling is unique, further statistical tests in other mid-latitude regions are essential. If similar results are obtained, positive TEC anomaly could be dominant in mid-latitude. Acknowledgements The authors wish to thank Prof. J.Y. Liu in Institute of Space Science, National Central University for useful comments and discussions. The GIM data were isolated from the Center for Orbit Determination in Europe (CODE). The GPS data observed by ground-based GPS receivers of GEONET were provided by the Geographical Survey Institute of Japan. The GAMIT software developed at the Massachusetts Institute of Technology (MIT) and the Scripps Institution of Oceanography (SIO) was used to extract the GPS-TEC from GPS data. The Dst index and the Kp index were supplied by World Data Center (WDC) for Geomagnetism, Kyoto. The F10.7 solar flux data were provided by the Solar Radio Monitoring Program operated jointly by the National Research Council and the Canadian Space Agency. Also we thank the United States Geological Survey for providing the earthquake data. This research was partly supported by a Grand-in-Aid for Scientific Research of Japan Society for Promotion of Science (19403002) and National Institute of Information and Communication Technology (R&D promotion funding International Joint Research). References ~ o-like Adams, J.B., Mann, M.E., Ammann, C.M., 2003. Proxy evidence for an el Nin response to volcanic forcing. Nature 426 (6964), 274–278. Afraimovich, E.L., Astafyeva, E.I., 2008. TEC anomalies – local TEC changes prior to earthquakes or TEC response to solar and geomagnetic activity changes? Earth, Planets and Space 60, 961–966. Afraimovich, E.L., Voeykov, S.V., Perevalova, N.P., Ratovsky, K.G., 2008. Large-scale traveling ionospheric disturbances of auroral origin according to the data of the GPS network and ionosondes. Advances in Space Research 42, 1213–1217. Chen, Y.I., Liu, J.Y., Tsai, Y.B., Chen, C.S., 2004. Statistical tests for pre-earthquake ionospheric anomaly. Terrestrial, Atmospheric and Oceanic Sciences 15, 385–396. Chree, C., 1912. Some phenomena of sunspots and of terrestrial magnetism, at Kew observatory. Philosophical Transactions Royal Society London Series A 212, 75. Chuo, Y.J., Chen, Y.I., Liu, J.Y., Pulinets, S.A., 2001. Ionospheric foF2 variations prior to strong earthquakes in Taiwan area. Advances in Space Research 27, 1305–1310. doi:10.1016/S0273-1177(01)00209-5. Chuo, Y.J., Liu, J.Y., Pulinets, S.A., Chen, Y.I., 2002. The ionospheric perturbations prior to the Chi-Chi and Chia-Yi earthquakes. Journal of Geodynamics 33, 509– 517. doi:10.1016/S0264-3707(02)00011-X. Dautermann, T., Calais, E., Haase, J., Garrison, J., 2007. Investigation of ionospheric electron content variations before earthquakes in southern California, 2003– 2004. Journal of Geophysical Research 112, B02106. doi:10.1029/2006JB004447. Davies, K., 1990. Ionospheric Radio. Peter Peregrinus Ltd., London, p. 580. Haung, Y., Lee, Y., 1975. On the variation of solar wind velocity following solar flares. Journal of Geophysical Research 80, 2863–2865. Hayakawa, M. (Ed.), 1999. Atmospheric and Ionospheric Electromagnetic Phenomena with Earthquakes. Terra Sci. Pub. Co., Tokyo, p. 996. Hayakawa, M., Fujinawa, Y. (Eds.), 1994. Electromagnetic Phenomena Related to Earthquake Predication. Terra Sci. Pub. Co., Tokyo, p. 667. Hayakawa, M., Molchanov, O.A. (Eds.), 2002. Seismo Electromagnetics: Lithosphere–Atmosphere–Ionosphere Coupling. TERRAPUB, Tokyo, p. 477.
419
Hobara, Y., Parrot, M., 2005. Ionospheric perturbations linked to a very powerful seismic event. Journal of Atmospheric and Solar–Terrestrial Physics 67, 677– 685. doi:10.1016/j.jastp. 2005.02.006. Hocke, K., 2008. Oscillations of global mean TEC. Journal of Geophysical Research A: Space Physics 113, A04302. Kelley, M.C., 1989. The Earth’s Ionosphere. Academic Press, San Diego, p. 487. Liperovskaya, E.V., Parrot, M., Bogdanov, V.V., Meister, C.-V., Rodkin, M.V., Liperovsky, V.A., 2006. On variations of foF2 and F-spread before strong earthquakes in Japan. Natural Hazards and Earth System Sciences 6, 735–739.
420
S. Kon et al. / Journal of Asian Earth Sciences 41 (2011) 410–420
Zakharenkova, I.E., Shagimuratov, I.I., Tepenitzina, N.Y., Krankowski, A., 2008. Anomalous modification of the ionospheric total electron content prior to the 26 September 2005 Peru earthquake. Journal of Atmospheric and Solar– Terrestrial Physics 70, 1919–1928.
Zhao, B., Wang, M., Yu, T., Wan, W., Lei, J., Liu, L., Ning, B., 2008. Is an unusual large enhancement of ionospheric electron density linked with the 2008 great Wenchuan earthquake? Journal of Geophysical Research A: Space Physics 113, A11304. doi:10.1029/2008JA013613.