JP7201143B2 - Earthquake prediction device and earthquake prediction method - Google Patents

Description

The present invention relates to an earthquake prediction device and an earthquake prediction method.

Japan has several places where plates are subducting in its coastal waters, and the land of Japan itself is located in the collision zone of four plates. Along with this, Japan has been frequently hit by large earthquakes since ancient times, causing enormous damage each time. Therefore, there is a strong demand for a technique for predicting earthquakes.

Various earthquake prediction techniques have been studied so far.
For example, Patent Document 1 discloses a technique that attempts to predict the occurrence of natural disasters for each region by receiving electromagnetic signals generated from the natural world, detecting abnormal signals, and collecting and analyzing the abnormal detection signals for each region. is disclosed.

In addition, Patent Document 2 continuously measures the ion concentration of the positive aerosol that stays around the measurement point, and collects or analyzes the measurement data and measurement condition data at the analysis center to generate prediction information data as an earthquake precursor. An earthquake prediction method is disclosed for predicting the scale of an earthquake that may occur in a range including all measurement points, the date and time of the occurrence of an earthquake, or the epicenter of an earthquake from the prediction information data.

JP 2005-134316 A JP 2006-194844 A

However, in order to acquire data for use in predicting earthquakes that occur in Japan, it is necessary to newly install facilities for acquiring such data throughout Japan, which incurs a large amount of cost.

For example, when the technology disclosed in Patent Document 1 is put into practical use, a plurality of mobile terminals having functions of detecting electromagnetic signals generated from the ground and detecting abnormalities in electromagnetic signals are newly manufactured, and the newly manufactured mobile terminals It is necessary to detect the electromagnetic signals generated from the ground all over Japan using , which incurs not only high economic costs but also significant human costs.

In addition, when the technology according to Patent Document 2 is put into practical use, it is necessary to newly install equipment for continuously measuring the ion concentration of the positive aerosol throughout Japan. not only the economic cost of

An object of the present invention is to provide an earthquake prediction method that can be implemented at low cost by using data that can be acquired by existing facilities.

The present invention is an earthquake prediction apparatus using observation data acquired by a seismic observation network, which is observation data at a plurality of points among the observation data at the occurrence of past earthquakes, and which is the observation data at a predetermined time from the occurrence of the earthquake at the points. A screening unit that screens the second observation data, which is the most recent observation data at the plurality of points, using the first observation data in a period going back in the past, and the time point of the earthquake using the screening result. , a prediction result calculation unit that calculates a prediction result of at least one of an earthquake occurrence location and an earthquake magnitude;
It relates to an earthquake prediction device comprising

Further, in the earthquake prediction device, the screening unit extracts, as the first observation data, observation data when an earthquake having a magnitude equal to or greater than a predetermined value occurs from the observation data acquired by the seismic observation network. is preferred.

Further, in the earthquake prediction device, both the first observation data and the second observation data are numerical data, and the screening unit numerically analyzes the first observation data and the second observation data. It is preferable to screen by

Further, both the first observation data and the second observation data are image data, and the screening unit performs screening by pattern matching the second observation data using the first observation data. preferably.

In addition, it is preferable that the prediction result calculation unit predicts the location where the earthquake will occur based on a combination of a plurality of observation points of the first observation data actually used in the screening.

Also, the seismic observation network is preferably a high-sensitivity seismic observation network.

Also, the seismic observation network is preferably a broadband seismic observation network.

The present invention also provides an earthquake prediction apparatus using observation data acquired by a seismic observation network, which is historical data at a plurality of locations installed in the seismic observation network within a predetermined period of time prior to the occurrence of an earthquake. a history data generation unit that generates history data including moving averages of seismometer output values, locations of occurrence of earthquakes, and magnitudes of earthquakes; An observation data acquisition unit that continuously acquires a moving average value of observation data; The present invention relates to an earthquake prediction apparatus including a prediction result calculation unit that calculates prediction results of at least one or more of magnitudes of earthquakes.

The present invention also provides an earthquake prediction apparatus using observation data acquired by a seismic observation network, wherein historical data at one or a plurality of points is obtained from the seismic observation network within a predetermined period of time prior to the occurrence of an earthquake. A history that generates history data including the location of the earthquake and the magnitude of the earthquake when an abnormality is found in any of the output value of the seismometer installed in the a data generation unit; an observation data acquisition unit that continuously acquires observation data of seismometer output values, which are observation data at the plurality of locations; The present invention relates to an earthquake prediction apparatus including a prediction result calculation unit that calculates prediction results of at least one of the time of occurrence of an earthquake, the location of an earthquake, and the magnitude of an earthquake.

Further, the present invention is an earthquake prediction method using observation data acquired by a seismic observation network, which is observation data at a plurality of points among the observation data when an earthquake occurred in the past, a first step of screening second observation data, which is the most recent observation data at the plurality of points, using first observation data that goes back a predetermined time; and a second step of calculating prediction results of at least one of an earthquake occurrence location and an earthquake magnitude.

According to the present invention, it is possible to provide an earthquake prediction method that can be implemented at low cost by using data that can be acquired by existing equipment.

FIG. 4 is a diagram showing the installation locations of high-sensitivity seismometers in the Hi-net used in the present invention; FIG. 4 is a diagram showing an example of seismic waveform data acquired by Hi-net used in the present invention; FIG. 4 is a diagram showing the installation locations of broadband seismometers in the F-net used in the present invention; FIG. 4 is a diagram showing an example of seismic waveform data acquired by F-net used in the present invention; FIG. 4 is a diagram showing characteristics of seismic waveform data acquired by F-net used in the present invention; FIG. 4 is a diagram showing characteristics of seismic waveform data acquired by F-net used in the present invention; FIG. 4 is a diagram showing characteristics of seismic waveform data acquired by F-net used in the present invention; FIG. 4 is a diagram showing characteristics of seismic waveform data acquired by F-net used in the present invention; FIG. 4 is a diagram showing characteristics of seismic waveform data acquired by F-net used in the present invention; FIG. 4 is a diagram showing characteristics of seismic waveform data acquired by F-net used in the present invention; 1 is a functional block diagram of an earthquake prediction device according to a first embodiment of the present invention; FIG. It is a flow chart which shows operation of an earthquake prediction device concerning a 1st embodiment of the present invention. It is a figure which shows the example of each phase contained in a seismic waveform. It is a figure which shows the data example as an earthquake precursor phenomenon in 1st Embodiment of this invention. It is a figure which shows the data example as an earthquake precursor phenomenon in 1st Embodiment of this invention. It is a figure which shows the data example as an earthquake precursor phenomenon in 1st Embodiment of this invention. It is a figure which shows the data example as an earthquake precursor phenomenon in 1st Embodiment of this invention. It is a figure which shows the data example as an earthquake precursor phenomenon in 1st Embodiment of this invention. It is a figure which shows the data example as an earthquake precursor phenomenon in 1st Embodiment of this invention. It is a figure which shows the data example as an earthquake precursor phenomenon in 1st Embodiment of this invention. It is a figure which shows the data example as an earthquake precursor phenomenon in 1st Embodiment of this invention. It is a figure which shows the data example as an earthquake precursor phenomenon in 1st Embodiment of this invention. It is a figure which shows the data example as an earthquake precursor phenomenon in 1st Embodiment of this invention. It is a figure which shows the data example as an earthquake precursor phenomenon in 1st Embodiment of this invention. It is a functional block diagram of an earthquake prediction device concerning a 2nd embodiment of the present invention. It is a flow chart which shows operation of an earthquake prediction device concerning a 2nd embodiment of the present invention. FIG. 10 is a diagram showing an example of data as an earthquake precursor phenomenon in the second embodiment of the present invention; It is a flow chart which shows operation of an earthquake prediction device concerning a 3rd embodiment of the present invention.

[1 First embodiment]
[1.1 Outline of the Invention]
The earthquake prediction method according to the first embodiment of the present invention roughly predicts the occurrence of an earthquake by the following method.

First, seismic waveform data acquired by the seismic observation network is stored in a database.

In this database, for example, the seismic waveform data at point A, which is characterized by Pa at time Ta, and the seismic waveform data at point B, which is characterized by Pb at time Tb It is assumed that waveform data and seismic waveform data at point C, which indicate that an earthquake of a predetermined magnitude or more occurred at time Tc within a predetermined period from Ta and Tb, have been accumulated.

At this time, new seismic waveform data with the characteristic Pa at point A and seismic waveform data with the characteristic Pb at point B are obtained. Predict the occurrence of earthquakes of a predetermined magnitude or greater.

That is, in this embodiment, using the seismic waveform data accumulated in the database, from among the newly obtained seismic waveform data, earthquakes having a characteristic of Pa at point A acquired within a predetermined period of time Waveform data and seismic waveform data having a feature of Pb at point B are screened, and based on the screening results, the occurrence of an earthquake of a predetermined magnitude or greater at point C is predicted.

[1.2 Seismic waveform data]
[1.2.1 Seismic Observation Network]
As seismic waveform data used in the earthquake prediction method according to the embodiment of the present invention, mainly data obtained by Hi-Net (high-sensitivity seismic observation network) and F-net (broadband seismic observation network) are used.

Hi-net is an observation network in which high-sensitivity seismometers capable of recording seismic motion with high sensitivity uniformly cover the whole of Japan at approximately 20km intervals. is. In Hi-net, high-sensitivity seismometers are installed at the bottom of wells with a depth of 100m or more to reduce the effects of artificial vibrations caused by automobiles, factories, and the like. Signals recorded by high-sensitivity seismometers are constantly transmitted to the National Research Institute for Earth Science and Disaster Resilience, and epicenter information and observed waveforms are widely available on the Internet.

FIG. 1 is a map showing observation points where seismic waveform data can be obtained from the Internet, which are part of the installation locations of broadband seismometers in Hi-net. In the map shown in FIG. 1, from roughly north to south, Shari North, Shibetsu South, Hamanaka, Saroma, Omu, Takinoue North, Oketo East, Wassamu, Niigata North, Shimizu, Taiki, Monbetsu East, Kimobetsu, Kuromatsunai, Assabu, Sai, Towada Kosai, Tago, Kunohe, Iwaizumi, Kawai Minami, Towa, Fujisawa, Kahoku, Naruko, Rifu, Ueyama, Miharu, Inawashiro, Nishiaizu, Iwaki Nishi, Naganuma, Bato, Tadami, Kuriyama Higashi, Iwase, Tone, Tatebayashi , Yuzawa, Nakano, Naguri, Itoigawa, Shimonita, Takeishi, Fujimi, Tatsuno, Hayakawa, Takagi, Izumi, Yaotsu, Hashima, Taga, Miyama, Shigaraki, Miyagawa, Sasayama, Higashiyoshino, Santo, Konohana, Yumesaki, Chizu, Totsukawahigashi, Tajiri, Hanazono, Yubara, Hinase, Oto, Hirokawa, Takebe, Kamigo, Bihoshi, Yoshida, Sadamitsu, Kisa, Iwami, Shingu, Kitagawa, Hiroshima, Kure, Saijo, Tobe, Mikawa, Hiyoshi, Mito, Tsushima, Kitakyushu , Yamaga, Nozuhara, Saigawa, Ume-Nishi, Tachiarai, Kikuchi, Fuji, Hyuga, Yabe, Nishimera, Kinkai, Ashikita, Miyakonojokita, and Miyanojo.

FIG. 2 shows an example of seismic waveform data acquired by Hi-net. The image shown in Fig. 2 is called "100-trace continuous waveform", and the seismic waveform for 60 minutes at the above 100 observation points is arranged so that the vertical axis is in order from north to south from top to bottom. The images are arranged horizontally. In the example of FIG. 2, seismic waveforms for 60 minutes from 8:00 am to 9:00 am on October 10, 2018 are shown.

As will be described later, in the present embodiment, when earthquake prediction is performed using Hi-net seismic waveform data, numerical analysis is performed using the numerical data that is the basis of this seismic waveform, so that the occurrence location of the earthquake can be determined. and predict magnitude.

On the other hand, F-net is an observation network in which approximately 100 broadband seismometers capable of recording seismic motion over a wide frequency range, from fast to slow, are deployed at approximately 100km intervals (planned) nationwide. In F-net, the broadband seismometers are installed deep in tunnels several tens of meters deep to reduce the influence of temperature and atmospheric pressure changes on seismic waveforms. Observation data from F-net is constantly transmitted to the National Research Institute for Earth Science and Disaster Resilience, and waveforms are automatically analyzed, stored, and published.

FIG. 3 is a map showing observation points where seismic waveform data can be obtained from the Internet, which are part of the installation locations of broadband seismometers in F-net. In the map shown in Fig. 3, Mt. Abu, Akadomari, Amami Oshima, Aogashima, Ashio, Choshi, Fujikawa, Fukue, Gojome, Hidaka, Hachijo, Honkawane, Hirono, Sapporo, Ishigaki, Imakane, Nakatsu, Tokushima, Iwatsuki , Yamagata, Izuhara, Nakaizu, Kofu, Kunigami, Kiwa, Kamitonda, Kamikyuusu, Kanayama, Kunikofu, Kanaya, Kawasaki, Kesennuma, Kushiro, Nagata, Kashiwazaki, Kozushima, Mimaya, Asahi, Nakagawa, Nemuro, Nogami, Nishiokibe, Nariwa, Nishiki, Osuka, Okawa, Onishi, Owani, Ogasawara, Saigo, Sefuri, Shibata, Tsuru Kanno, Shari, Mt. Shibi, Shimizu, Shiramine, Sotome, Tashiro, Taga, Tamagawa, Takakuma, Takeda, Takaoka, Tomari, Tomari , Nakagawa, Tenryu, Nishitosa, Tsukuba, Takato, Tateyama, Tonoyamazaki, Umaji, Urahoro, Wajima, Watarai, Yasaka, Yamizo, Yonaguni, Yoshida, Toyoda, Yamazaki, and Zamami.

FIG. 4 shows an example of seismic waveform data acquired by F-net. The example shown in Fig. 4 is the seismic waveform acquired at 9:00 am on May 13, 2018 in Nakaizu. It is an image in which 60 seismic waveforms for 60 seconds are arranged horizontally in parallel.

In this embodiment, earthquake prediction may be performed using only seismic waveform data acquired by Hi-net, or earthquake prediction may be performed using only seismic waveform data acquired by F-net. By combining these predictions, it is possible to make earthquake predictions with higher accuracy.

[1.2.2 Characteristics of seismic waveforms]
The features of the seismic waveform data acquired by F-net are shown below with reference to FIGS. 5A to 5F. The seismic waveforms acquired by F-net have the following six features from type A to type F.

[1.2.2.1 Type A]
FIG. 5A shows Type A features. A feature of type A is that the amplitude of each of 60 waveforms for one hour is greater than a predetermined value, as shown in FIG. 5A.

[1.2.2.2 Type B]
FIG. 5B shows Type B features. A feature of type B is that the interval between adjacent waveforms exceeds a predetermined value, as indicated by the arrow (b) in FIG. 5B.

[1.2.2.3 Type C]
FIG. 5C shows Type C features. As shown by the arrow (c) in FIG. 5C, the characteristics of type C are that, in the seismic waveform, the period is equal to or less than a predetermined value, and the amplitude is greater than or equal to a predetermined value that exceeds the adjacent waves for a first predetermined period or longer. A waveform having a plurality of peaks over a period of time is characterized by sporadically occurring at intervals of a second predetermined period or longer.

[1.2.2.4 Type D]
FIG. 5D shows Type D features. The feature of type D is that when the seismic waveform is divided into a trend component and a noise component, as shown by the arrow (d) in FIG. It is characteristic that the wave itself looks thick because of the

[1.2.2.5 Type E]
FIG. 5E shows Type E features. A feature of type E is that, as shown by the arrow (e) in FIG. This is a sporadically generated waveform.

[1.2.2.6 Type F]
FIG. 5F shows Type F features. A feature of type F is that, as shown by the arrow (f) in FIG. This is a waveform that occurs sporadically at

[1.3 Configuration of the Invention]
FIG. 6 is a functional block diagram of an earthquake prediction device according to the present invention. An earthquake prediction device 10 according to this embodiment includes a communication unit 11 , a storage unit 12 and a control unit 13 .

The communication unit 11 is a communication interface used by the data acquisition unit 14 to acquire seismic waveform data obtained by the seismic observation network via the Internet, as will be described later.

The storage unit 12 stores seismic waveform data acquired by the data acquisition unit 14, data obtained by processing the seismic waveform data by the control unit 13, and the like.

The control unit 13 has a CPU, a ROM, a RAM, a CMOS memory, etc., which are known to those skilled in the art and are configured to communicate with each other via a bus.
The CPU is a processor that controls the earthquake prediction device 10 as a whole. The CPU reads the system program and the application program stored in the ROM via the bus, and controls the entire earthquake prediction apparatus 10 according to the system program and the application program, thereby operating the control unit 13 as shown in FIG. It is configured to realize the functions of the data acquisition unit 14 , the screening unit 15 , the prediction result calculation unit 16 , and the display unit 17 . Various data such as temporary calculation data and display data are stored in the RAM. The CMOS memory is backed up by a battery (not shown) and configured as a non-volatile memory that retains the memory state even when the power of the earthquake prediction device 10 is turned off.

The data acquisition unit 14 acquires seismic waveform data obtained from the seismic observation network from the Internet via the communication unit 11, which is a communication interface. Furthermore, the data acquisition unit 14 stores the acquired seismic waveform data in the storage unit 12 .

The screening unit 15 selects observation data at a plurality of locations from among the observation data at the time of past earthquakes stored in the storage unit 12, and selects the first observation data within a period going back a predetermined time from the occurrence of the earthquake at the location. The observation data is used to screen the second observation data, which is the most recent observation data at the plurality of points.

More specifically, the screening unit 15 selects, from among the observation data of past earthquakes stored in the storage unit 12, observation data at a plurality of points when an earthquake having a magnitude equal to or greater than a predetermined value occurs, Numerical analysis is used to screen the most recent numerical data at the plurality of points by using the numerical data within a period that goes back a predetermined period of time from the occurrence of the earthquake at the point.

Alternatively, the screening unit 15 selects observation data at a plurality of points when an earthquake having a magnitude of a predetermined value or more occurs among the observation data at the time of past earthquakes stored in the storage unit 12, Using the image data of seismic waveforms within a period that goes back a predetermined time from the occurrence of the earthquake, the image data of the most recent seismic waveforms at the plurality of points are screened by pattern matching.

Details of the operation of the screening unit 15 will be described later in [1.4 Earthquake Prediction Method].

The prediction result calculation unit 16 uses the results of screening by the screening unit 15 to calculate a prediction result of at least one of the time of occurrence of the earthquake, the location of the occurrence of the earthquake, and the magnitude of the earthquake.

More specifically, the prediction result calculator 16 predicts the location and magnitude of an earthquake using the results of screening by the screening unit 15, and predicts the time of occurrence of the earthquake from the screened seismic waveform.

Details of the operation of the prediction result calculation unit 16 will be described later in [1.4 Earthquake Prediction Method].

The display unit 17 displays the prediction result calculated by the prediction result calculation unit 16 on a display device (not shown) connected to the earthquake prediction device 10 . For example, a monitor may be used as the display device, and in this case, the display unit 17 displays the prediction result on the screen of the monitor. Alternatively, the display device may be, for example, a printer. In this case, the display unit 17 uses the printer to print the prediction result.

[1.4 Earthquake prediction method]
[1.4.1 Overall flow]
FIG. 7 is a flowchart showing an outline of an earthquake prediction method, which is the operation of the earthquake prediction device 10 according to this embodiment.

In step S11, the data acquisition unit 14 acquires seismic waveform data obtained by the seismic observation network via the Internet.

In step S12, the screening unit 15 screens the seismic waveform data newly acquired in step S11 using past seismic waveform data.

In step S<b>13 , the prediction result calculation unit 16 calculates prediction results of locations and/or magnitudes of future earthquakes based on the results of screening by the screening unit 15 .

In step S14, the prediction result calculator 16 calculates the prediction result of the occurrence time of an earthquake based on the seismic waveform data screened in step S12.

In step S<b>15 , the display unit 17 displays at least one prediction result of the location, magnitude, and time of occurrence of an earthquake on the display device.

Details of steps S12 to S14 among the above steps will be described below.

[1.4.2 Calculation method of prediction result of earthquake occurrence location and magnitude]
[1.4.2.1 Prediction result calculation method using Hi-net seismic waveform data]
When calculating an earthquake prediction result using Hi-net seismic waveform data, the prediction result is calculated by performing numerical analysis using the numerical data of the seismic waveform data.

In the seismic waveform data accumulated in the storage unit 12, the _{amplitude values} at time t ₁ at _each observation point p ₁ , p ₂ . . . _1,100 . Also, the amplitude values at time t2 at _each observation point p ₁ , p ₂ . . . p ₁₀₀ are v _2,1 , _{v 2,2} . At _{each observation point p 1 , p 2 . , 100} .

Here, the screening unit 15, for each amplitude v _{1,1 ,} v _{1,2 .} is equal to or greater than a predetermined value V, and the value is 0 when less _{than the} predetermined _{value V.} Set up a matrix F ₀ of ₁₀₀ .

Furthermore, the screening unit 15, when an earthquake of magnitude M equal to or larger than a predetermined magnitude occurs at point _C within a predetermined period T2 from time _tn , the matrix _F0 and the location and magnitude of this earthquake (magnitude).

Here, each time the data acquisition unit 14 acquires new seismic waveform data, the screening unit ₁₅ successively continues to generate the matrix F, and the newly generated matrix F1 and the already accumulated matrix _The newly generated matrix F1 is screened when the degree of matching with _F0 is greater than or equal to a predetermined value, that is, when the number of elements with matching values between matrix _F0 and matrix F1 is greater than or equal to _a predetermined ratio.

The prediction result calculation unit 16 is based on the newly generated matrix F ₁ and the fact that the matrix F ₀ whose degree of matching with this matrix F ₁ is equal to or greater than a predetermined value is associated with an earthquake of magnitude M at point C. , a prediction result that an earthquake of magnitude M will occur at point C is calculated.

[1.4.2.2 Prediction result calculation method using F-net seismic waveform data]
When calculating the prediction result of an earthquake using F-net seismic waveform data, the prediction result is calculated by pattern matching using the image data of the seismic waveform data.

The screening unit 15, as described in [1.2.2 Characteristics of seismic waveforms] above, for the seismic waveform data obtained at time t1 from the F _- net stored in the storage unit 12, It is determined whether or not there are six types of features from type A to type F.

Furthermore, the screening unit 15, if an earthquake of magnitude M equal to or greater than a predetermined magnitude occurred at _{point C} within _a predetermined period of time T2 from time t1, the seismic waveform data obtained at time t1 The combination of characteristics and observation points is linked to the location and magnitude of the earthquake.

Here, each time seismic waveform data is newly acquired by the data acquisition unit 14, the screening unit 15 successively checks that the newly acquired seismic waveform data includes six types of features from type A to type F. Whether or not it is determined by pattern matching, and when it is determined that it matches, the seismic waveform data including this feature is screened.

More specifically, the screening unit 15 is the past seismic waveform data accumulated in the storage unit 12, which includes any of the above six types of features from type A to type F, and also identifies the location of the earthquake. And from the seismic waveform data linked to the magnitude, patterns corresponding to the features of types A to F are extracted, and using these patterns, pattern matching is performed on newly acquired seismic waveform data. Screen by

If the combination of features and observation points of the seismic waveform data, which is the result of screening by the screening section 15, coincides with the combination of features and observation points associated with an earthquake of magnitude M at point C, the prediction result calculation unit 16 , a prediction result that an earthquake of magnitude M will occur at point C is calculated.

For example, among the seismic waveform data accumulated in the storage unit 12, the seismic waveform data acquired at Tamagawa at time t1 includes the feature of _type D, and the seismic waveform data acquired at Kanaya at time t1 includes _type D features. Suppose that E features are included. Further, assume that an earthquake of magnitude 5.6 occurred in Nagano within _a predetermined period of _time T2 from time t1.

At this time, among the earthquake data newly acquired by the data acquisition unit 14, the seismic waveform data acquired at Tamagawa contains the type D feature, and the seismic waveform data acquired at Kanaya contains the type E feature. , the prediction result calculation unit 16 calculates a prediction result that an earthquake of magnitude 5.6 will occur in Nagano.

The prediction result calculation unit 16 uses the above [1.4.2.1 Prediction result calculation method using Hi-net seismic waveform data] and [1.4.2.2 F-net seismic waveform Prediction Result Calculation Method Using Data] can be combined to calculate a more accurate prediction result of the location and magnitude of an earthquake.

[1.4.3 Prediction method for when an earthquake will occur]
FIG. 8 shows, as an example, seismic waveforms over 36 hours when an earthquake occurred in Fukushima on November 22, 2016. From 36 hours before the earthquake, movement and amplitude above a predetermined value can be seen in the seismic waveform. (this period is referred to herein as the "Active Phase"). Furthermore, from 12 hours before the occurrence of the earthquake, the motion and amplitude of the seismic waveform become smaller (this period is called the "Quiet Phase" here), and soon the earthquake will occur.

That is, 36 hours before the occurrence of the earthquake, movement occurred in the seismic waveform. Movement becomes quiet from before the time until the earthquake occurs.

The prediction result calculator 16 predicts when an earthquake will occur based on this change in movement. Specifically, whether the screening unit 15 screens the Hi-net seismic waveform data or the F-net seismic waveform data, the prediction result calculation unit 16, in the screened seismic waveform data, , which phase exemplified in FIG. 8 is currently present, and based on this determination result, a prediction result for predicting how much time has passed since the present time when an earthquake will occur is calculated.

Alternatively, the prediction result calculation unit 16 determines that the latest seismic waveform data acquired at observation points around the location where an earthquake is predicted to occur in step S13 in FIG. It is also possible to determine whether or not the current phase is in phase, and based on this determination result, calculate a prediction result that predicts how much time will elapse from the present time when an earthquake will occur.

[1.5 Example]
As an example of the present embodiment, data as precursors of the Northern Osaka Earthquake that occurred on June 18, 2018 are shown.

9A and 9B are seismic waveform data obtained by Hi-net. FIG. 9A is the seismometer output readings generally from May 30 to June 19, 2018, at Miyagawa. Also, FIG. 9B shows seismometer outputs generally from May 30 to June 19, 2018, at Saijo.

As shown in FIG. 9A, in the seismic waveform acquired at Miyagawa, for example, on May 30, 2018, June 8, 2018, June 10, 2018, June 12, 2018, June 17, 2018 and, as shown in FIG. 9B, in the seismograms acquired at Saijo, e.g. On June 12, 2018, and June 17, 2018, high displacement values were shown. is possible.

10A to 10C are seismic waveform data at Mt. Abu obtained by F-net. More specifically, FIG. 10A is the data obtained at 23:00 on June 14, 2018, and FIG. 10B is the data obtained at 0:00 on June 15, 2018. 10C is data obtained on June 15, 2018 at 1:00.

The data in FIG. 10A includes the Type A features described above. The data in FIG. 10B includes the Type A features described above. The data in FIG. 10C includes the Type A features described above.

11A to 11F are seismic waveform data at Okawa obtained by F-net. More specifically, FIG. 11A is the data obtained at 20:00 on June 14, 2018, and FIG. 11B is the data obtained at 21:00 on June 14, 2018. 11C is data obtained between 22:00 on June 14, 2018, FIG. 11D is data obtained between 23:00 on June 14, 2018, and FIG. The data was obtained at 12:00 on June 15, and FIG. 11F is the data obtained at 1:00 on June 15, 2018.

The data in FIG. 11A contains the Type D features described above. The data in FIG. 11B includes the Type D features described above. The data in FIG. 11C includes the Type D features described above. The data in FIG. 11D includes the Type D features described above. The data in FIG. 11E includes the Type D features described above. The data in FIG. 11F contain the Type D features described above.

As shown in FIG. 10A, the seismic waveform acquired at Mt. The feature of type A is included in the 0:00 range, and as shown in FIG. 10C, the feature of type A is included in the 1:00 range on June 15, 2018. As shown in FIG. 11A, the feature of type D is included at 20:00 on June 14, 2018, and the feature of type D is included at 21:00 on June 14, 2018, as shown in FIG. 11B. 11C, the feature of type D is included between 22:00 on June 14, 2018, and the feature of type D is included between 23:00 on June 14, 2018, as shown in FIG. , and includes Type D features at midnight on June 15, 2018, as shown in FIG. In other words, it is possible to predict the occurrence of the Northern Osaka Prefecture Earthquake based on the acquisition location and each feature of these data.

[1.6 Effects of First Embodiment]
The earthquake prediction device 10 according to the first embodiment of the present invention is an earthquake prediction device that uses observation data acquired by a seismic observation network. a screening unit 15 that screens the second observation data, which is the most recent observation data at the plurality of points, using the first observation data that goes back a predetermined time from the occurrence of the earthquake at the point; A prediction result calculation unit 16 that calculates prediction results of at least one of the time of occurrence of an earthquake, the location of an earthquake, and the magnitude of an earthquake, using the results.

This makes it possible to provide an earthquake prediction method that can be implemented at low cost by using data that can be acquired with existing equipment.

The screening unit 15 also extracts, as first observation data, observation data obtained when an earthquake having a magnitude equal to or greater than a predetermined value occurs from the observation data acquired by the seismic observation network.

By narrowing down the first observation data, the prediction result calculation unit 16 can save the amount of calculation required to calculate the prediction result.

Moreover, when both the first observation data and the second observation data are numerical data, the screening unit 15 performs screening by numerically analyzing the first observation data and the second observation data.

As a result, the screening unit 15 can perform screening by omitting the procedure of converting the numerical data into an image.

Moreover, when both the first observation data and the second observation data are image data, the screening unit 15 performs screening by pattern matching the second observation data.

Accordingly, when a plurality of patterns are provided as patterns used in pattern matching, observation data can be screened corresponding to a plurality of features.

Also, the prediction result calculation unit 16 predicts the location of an earthquake based on a combination of a plurality of points, which are the observation points of the first observation data actually used in the screening.

For example, instead of using only the seismic waveform data at point A to predict an earthquake that will occur at point A, combine seismic waveform data from multiple points and predict the location of the earthquake based on the type of combination. This makes it possible to calculate earthquake prediction results based on complex plate tectonics.

Also, the seismic observation network is a high-sensitivity seismic observation network (Hi-net) and/or a broadband seismic observation network (F-net).

This makes it possible to calculate earthquake prediction results using readily available seismic waveform data.

[2 Second Embodiment]
[2.1 Outline of the Invention]
The earthquake prediction method according to the second embodiment of the present invention roughly predicts the occurrence of an earthquake by the following method.

For example, based on the earthquake catalog data published by the Japan Meteorological Agency, etc., the output values of the seismometers at each observation point in Hi-net and F-net for a predetermined period of time going back from the occurrence of earthquakes that have occurred so far get. Also, the moving average value a is calculated within the data of each point in the period. Furthermore, a value indicating how long before the earthquake occurred when each moving average value a exceeded a specified value (σ), the area where the earthquake occurred, and the magnitude of the earthquake are stored as historical data in a database.

Here, the "moving average value a" is obtained by taking the moving average of the time-series data in the predetermined period.

On top of that, in the present embodiment, from the newly acquired seismometer output values in Hi-net and F-net, at a certain measurement location, the calculation of the moving average value b in a continuous data group is continued. In the meantime, when data exceeding the specified value (σ) is observed in the moving average value b, the occurrence time, occurrence area, and magnitude of the earthquake are predicted by collating with the above-mentioned historical data database.

Here, the "moving average value b" is the moving average of continuous time-series data of continuous data groups at a certain measurement location of the seismometers in the above Hi-net and F-net. is.

For example, when the moving average value b calculated from the continuous output values of the seismometer in Hi-net or F-net exceeds a specified value at point A, the history data is referred to. In this historical data, if the moving average value a at point A exceeds the specified value (σ), B days later, and if an earthquake of magnitude D occurs at point C, then B days from now , C, an earthquake of magnitude D is predicted to occur.

[2.2 Configuration of the Invention]
FIG. 12 is a functional block diagram of an earthquake prediction system according to this embodiment. An earthquake prediction system 1A according to this embodiment includes an earthquake prediction device 10A, a first server 30A, and a second server 30B.

The earthquake prediction device 10A extracts the above history data from the catalog data, which is the data of the earthquake catalog stored in the first server 30A, and the output values of the seismometers before the earthquake occurred obtained from Hi-net and F-net. is generated and stored in the second server 30B.
In addition, the earthquake prediction device 10A continuously calculates the moving average value b of a continuous data group based on the observation data, which is the output value of the seismograph, acquired from Hi-net and F-net, Earthquake prediction is performed by comparing the moving average value b with the history data stored in the second server 30B.

As shown in FIG. 12, the earthquake prediction device 10A, the first server 30A and the second server 30B are communicably connected to each other. Although not shown, the earthquake prediction device 10A, the first server 30A, and the second server 30B may be connected to each other via a network.

10 A of earthquake prediction apparatuses are provided with 11 A of communication parts, and 13 A of control parts.
The communication unit 11A is a communication interface used by the catalog data acquisition unit 18, which will be described later, to acquire catalog data from the first server 30A. The communication unit 11A is also a communication interface used by the history data generation unit 19, which will be described later, to store history data in the second server 30B. The communication unit 11A is also a communication interface used by the observation data acquisition unit 20 described later to acquire earthquake numerical data obtained by seismic observation networks such as Hi-net and F-net via the Internet. be. The communication unit 11A is also a communication database used by a prediction result calculation unit 16A, which will be described later, to refer to history data from the second server 30B in order to calculate earthquake prediction results.

The control unit 13A has a CPU, a ROM, a RAM, a CMOS memory, etc., which are known to those skilled in the art and are configured to communicate with each other via a bus.
The CPU is a processor that controls the earthquake prediction device 10A as a whole. The CPU reads the system program and the application program stored in the ROM via the bus, and controls the entire earthquake prediction apparatus 10A according to the system program and the application program. It is configured to implement the functions of the catalog data acquisition unit 18 , the history data generation unit 19 , the observation data acquisition unit 20 , the prediction result calculation unit 16A, and the display unit 17 . Various data such as temporary calculation data and display data are stored in the RAM. The CMOS memory is backed up by a battery (not shown) and configured as a non-volatile memory that retains the stored state even when the power of the earthquake prediction device 10A is turned off.

The catalog data acquisition unit 18 acquires catalog data from the first server 30A via the communication unit 11A, which is a communication interface.
The catalog data includes at least the time, area and magnitude of each earthquake, but may also include the output values of the seismometer before the earthquake occurred.

Further, the catalog data acquisition unit 18 may acquire the catalog data from the first server 30A after narrowing down the earthquake occurrence period based on the value of the occurrence period input by the user of the earthquake prediction device 10A. .

Further, the catalog data acquisition unit 18 may acquire the catalog data from the first server 30A after narrowing down the earthquake occurrence area based on the occurrence area input by the user of the earthquake prediction device 10A. The user can set the earthquake occurrence area to, for example, the whole world, the country where the earthquake occurred, or a rectangular range specified by latitude and longitude. It is also possible to specify the occurrence area by other methods.

The catalog data acquisition unit 18 may acquire catalog data from the first server 30A after narrowing down the magnitude of the earthquake based on the magnitude range input by the user of the earthquake prediction device 10A. The user can narrow down the magnitude range using values between 4.0 and 9.5 as the magnitude of the earthquake.

The history data generation unit 19 uses the catalog data acquired by the catalog data acquisition unit 18 and the output values of the seismographs before the earthquake occurrence acquired from Hi-net and F-net via the communication unit 11A. , to generate historical data. Here, the historical data means the moving average value a calculated from the seismometer output values obtained from Hi-net and F-net during a predetermined period before the occurrence of the earthquake, and the moving average value a. This data includes a value indicating how long before the earthquake occurred when the value σ was exceeded, the area where the earthquake occurred, and the magnitude of the earthquake. Further, the predetermined period may be, for example, 14 days before the occurrence of the earthquake, 7 days before the occurrence of the earthquake, or 3 days before the occurrence of the earthquake.
When the output values of the seismometer before the earthquake are included in the catalog data as described above, the history data generation unit 19 obtains from Hi-net and F-net via the communication unit 11A Instead of using the output value of the seismometer before the occurrence of the earthquake, the output value of the seismometer originally included in the catalog data may be used.
In historical data, the magnitude of an earthquake is, for example, 4.0 or more and less than 4.5, 4.5 or more and less than 5.0, 5.0 or more and less than 5.5, etc., in increments of 0.5. categorized by.
Further, the history data generator 19 stores the generated history data in the second server 30B.

The observation data acquisition unit 20 acquires output values of the seismometer obtained from the seismic observation network from the Internet via the communication unit 11A, which is a communication interface.

The prediction result calculation unit 16A continues to calculate the moving average value b of the continuous data group from the newly acquired seismometer output values in Hi-net and F-net, and calculates the moving average value b as described above. Predict the timing, area, and magnitude of earthquakes by comparing them with the database of historical data.

Since the display section 17 is the same as the display section 17 in the first embodiment, description of its function is omitted.

The first server 30A includes a communication section 31A and a storage section 32A.

The communication unit 31A is a communication interface used for acquiring the catalog data 33, for example, through the Internet. The communication unit 31A is also a communication interface used to transmit catalog data to the earthquake prediction device 10A.

The storage unit 32A is a device that stores the catalog data 33. FIG.

The second server 30B includes a communication section 31B and a storage section 32B.

The communication unit 31B is a communication interface used for transmitting/receiving the history data 34 to/from the earthquake prediction device 10A.

The storage unit 32B is a device that stores the history data 34. FIG.

[2.3 Earthquake prediction method]
FIG. 13 is a flow chart showing an outline of an earthquake prediction method, which is the operation of the earthquake prediction device 10A according to this embodiment.

In step S21, the catalog data acquisition unit 18 acquires catalog data from the first server 30A.

In step S22, the history data generator 19 uses the catalog data 33 to generate history data.

In step S23, the history data generator 19 stores the history data 34 in the second server 30B.

In step S24, the observation data acquisition unit 20 acquires observation data, which is the output value of the seismograph, from the seismic observation network, for example, via the Internet.

In step S25, the prediction result calculation unit 16A calculates the moving average value b using the observed data.

In step S26, the prediction result calculator 16A compares the moving average value a with the specified value. If the moving average value b exceeds the specified value (S26: YES), the process proceeds to step S27. If the moving average value b is equal to or less than the specified value (S26: NO), the process is terminated.

In step S27, the prediction result calculation unit 16A refers to history data.

In step S28, the prediction result calculation unit 16A calculates at least one of the earthquake location, time of occurrence, and magnitude of the earthquake based on the reference result.

In step S<b>29 , the display unit 17 displays the prediction result of at least one of the location, magnitude, and time of occurrence of an earthquake on the display device.

[2.4 Examples]
As an example of this embodiment, data as precursors of the Kumamoto earthquake that occurred on April 16, 2016 are shown. Figure 14 shows the above moving average value b calculated from the output values of the seismometers acquired in Kure from March 16, 2016, before the earthquake to April 16, 2016, when the earthquake occurred. It is a graph which shows a transition.

In the graph of Fig. 14, most of the moving average values from March 16, 2016 to April 16, 2016 before the earthquake show a high value of -3.0 or more, which is a sign of the occurrence of the earthquake. phenomenon.

[2.5 Effects of Second Embodiment]
An earthquake prediction device 10A according to the second embodiment of the present invention is an earthquake prediction device that uses observation data acquired by a seismic observation network. a history data generation unit 19 for generating history data including a value corresponding to the deviation of the output value of the seismograph installed in the seismic observation network, the location of the occurrence of the earthquake, and the magnitude of the earthquake within the period specified; The observation data acquisition unit 20 continuously acquires the moving average of the observation data of the output value of the seismometer, which is the observation data at the point, and the observation data is collated with the historical data to obtain the time of occurrence of the earthquake, the earthquake and a prediction result calculation unit 16A that calculates a prediction result of at least one of the occurrence location of the earthquake and the magnitude of the earthquake.

As a result, as with the earthquake prediction apparatus 10 according to the first embodiment, it is possible to provide an earthquake prediction method that can be realized at low cost by using data that can be acquired by existing equipment.

[3 Third Embodiment]
[3.1 Outline of the Invention]
The earthquake prediction method according to the third embodiment of the present invention roughly predicts the occurrence of an earthquake by the following method.

In the second embodiment, as described above, based on the data in the earthquake catalog, seismometers at each observation point in Hi-net and F-net for a predetermined period of time going back from the occurrence of earthquakes that have occurred so far I was getting the output value of Moreover, the moving average value a was calculated within the data of each point in the said period. In addition, a database was created as historical data that indicates how long before the earthquake occurred when each moving average value a exceeded a specified value (σ), the area where the earthquake occurred, and the magnitude of the earthquake. .

Furthermore, in the second embodiment, the calculation of the moving average value b in a continuous data group at a certain measurement location is continued from the newly acquired seismometer output values in Hi-net and F-net. When data exceeding the specified value (σ) was observed in the moving average value b, the timing, area, and magnitude of the earthquake were predicted by comparing it with the above-mentioned historical data database.

On the other hand, in this embodiment, one of the data of the output value of the seismograph at each observation point, the moving average value a of the data, and the seismic waveform for a predetermined period before the occurrence of an earthquake of magnitude 4.0 or more If any abnormalities are observed, such as exceeding the specified values or the waveforms of types A to F above are observed, for each observation point, the period until the occurrence of the earthquake, The time, the place where the earthquake occurred, and the magnitude of the earthquake are used as a historical data database.

Here, in the newly acquired Hi-net and F-net seismometer observation data, if the above-mentioned anomaly is found at one or more observation points, from the history data database, the above-mentioned For each observation point, extract information on the period until the earthquake occurrence, the time of occurrence of the earthquake, the location of the occurrence of the earthquake, and the magnitude of the earthquake.

At one or more observation points, information on the period until the occurrence of an earthquake, the time of occurrence of the earthquake, the location of the occurrence of the earthquake, and the magnitude of the earthquake are extracted. If there is information, information with a high degree of redundancy is displayed as a predicted earthquake.

In addition, if an earthquake occurs even though the above-mentioned anomaly was not observed at the observation point among the newly acquired observation data of the seismographs in Hi-net and F-net, the earthquake The date and time of the occurrence of the earthquake, the location of the occurrence of the earthquake, and the magnitude of the earthquake are stored in the history data database.

[3.2 Configuration of the Invention]
A functional block diagram of the earthquake prediction system 1B according to the third embodiment is the same as that of the earthquake prediction system 1A according to the second embodiment, so illustration thereof is omitted.

Note that in the earthquake prediction system 1B according to the third embodiment, the history data generation unit 19B generates the data of the output values of the seismographs at each observation point in a predetermined period before the occurrence of an earthquake with a magnitude of 4.0 or more, If the moving average value a of the data or one or more of the seismic waveforms exceeds the specified value, or if an abnormality such as the above type A to type F waveforms is observed, the earthquake will be monitored. For each point, the period until the occurrence of the earthquake, the time of occurrence of the earthquake, the location of the occurrence of the earthquake, and the magnitude of the earthquake are generated as history data and stored in the second server 30B.

Further, the prediction result calculation unit 16B, in the observation data acquired by the observation data acquisition unit 20, when the above abnormality is found at one or a plurality of observation points, the prediction result calculation unit 16B, from the database of the history data 34, retrieves the above For each observation point, extract information on the period until the earthquake occurred, the time when the earthquake occurred, the location of the earthquake, and the magnitude of the earthquake. If there is information, information with a high degree of redundancy is calculated as a predicted earthquake and output to the display unit 17 .

[3.3 Earthquake prediction method]
FIG. 15 is a flowchart showing an outline of an earthquake prediction method, which is the operation of the earthquake prediction device 10B according to this embodiment.

In step S31, the catalog data acquisition unit 18 acquires catalog data from the first server 30A.

In step S<b>32 , the history data generator 19 uses the catalog data 33 to generate history data.

In step S33, the history data generator 19 stores the history data 34 in the second server 30B.

In step S34, the observation data acquisition unit 20 acquires observation data, which is the output value of the seismograph, from the seismic observation network, for example, via the Internet.

In step S35, if there is an abnormality in the observation data (S35: YES), the process proceeds to step S36. If there is no abnormality in the observed data (S35: NO), the process ends.

In step S36, the prediction result calculation unit 16B determines from the history data 34 if the above abnormality is found at one or a plurality of observation points in the observation data acquired by the observation data acquisition unit 20. For each observation point, extract information on the period until the earthquake occurrence, the time of occurrence of the earthquake, the location of the occurrence of the earthquake, and the magnitude of the earthquake.

In step S37, the prediction result calculation unit 16B, if there is overlapping information among the information related to a plurality of earthquakes related to one or more observation points, the information with a high degree of overlap is calculated as the predicted earthquake. Calculate as

In step S38, the display unit 17 displays the prediction result calculated by the prediction result calculation unit 16B on the display device.

[3.4 Effects of Third Embodiment]
An earthquake prediction device 10B according to the third embodiment of the present invention is an earthquake prediction device that uses observation data acquired by a seismic observation network. In the above, if an abnormality is found in any of the output values of the seismometers installed in the seismic observation network, the moving average value of the output values, or the seismic waveform, the history including the location of the earthquake and the magnitude of the earthquake A history data generation unit 19B that generates data, an observation data acquisition unit 20 that continuously acquires observation data of seismometer output values, which is observation data at multiple points, and compares the observation data with the history data. Accordingly, a prediction result calculation unit 16B is provided for calculating the prediction result of at least one of the time of occurrence of the earthquake, the location of the occurrence of the earthquake, and the magnitude of the earthquake.

As a result, similar to the earthquake prediction device 10 according to the first embodiment and the earthquake prediction device 10A according to the second embodiment, by using data that can be acquired with existing equipment, earthquakes that can be realized at low cost It becomes possible to provide a prediction method.

[4 Modifications]
[4.1 Modification 1]
In the first embodiment, the seismic waveform data to be screened or the seismic waveform data of observation points around locations where an earthquake is expected to occur are classified into "Active Phase" or "Quiet Phase". Based on which phase of "Quiet Phase", the time of occurrence of an earthquake is predicted, but the present invention is not limited to this.

For example, in the above [1.4.2.1 Prediction result calculation method using Hi _- net seismic waveform data], within a predetermined period T2 from the time _tn , at point C, a magnitude of a predetermined magnitude or more When an earthquake of M has occurred and the matrix F ₀ is associated with the occurrence location and magnitude (magnitude) of this earthquake, the prediction result calculation unit 16 calculates the acquisition time of the seismic waveform data that is the basis of the matrix F ₁ You may calculate the prediction result that an earthquake will occur within _T2 from .

Alternatively, in [1.4.2.2 Prediction result calculation method using F-net seismic waveform data], the screening unit 15 determines _each feature and observation point included in the seismic waveform data at time t1 and the location and magnitude of an earthquake of magnitude M that occurred at point _C within _a predetermined period of time T2 from time t1, the prediction result calculation unit 16 determines that the seismic waveform data, which is the screening result, is A prediction result may be calculated that predicts that an earthquake will occur within T2 from the acquisition _time of this seismic waveform data when the combination of features and observation points that it possesses matches the above combination.

[4.2 Modification 2]
In the second embodiment, the earthquake prediction device 10A, the first server 30A, and the second server 30B are separate entities, but the present invention is not limited to this.

For example, the earthquake prediction device 10A may include a storage unit, and the storage unit may store catalog data and/or history data. That is, in this aspect, the first server 30A and/or the second server 30B may not be essential components.

Although the embodiments of the present invention have been described above, the present invention is not limited to the above-described embodiments. Moreover, the effects described in the present embodiment are merely enumerations of the most suitable effects resulting from the present invention, and the effects of the present invention are not limited to those described in the present embodiment.

An earthquake prediction method by the earthquake prediction device 10 or 10A is realized by software. When realized by software, a program that constitutes this software is installed in the computer (earthquake prediction device 10 or 10A). These programs may be recorded on removable media and distributed to users, or may be distributed by being downloaded to users' computers via a network. Furthermore, these programs may be provided to the user's computer (earthquake prediction device 10 or 10A) as a web service via a network without being downloaded.

1A Earthquake prediction system 10 10A Earthquake prediction device 11 11A 31A 31B Communication unit 12 32A 32B Storage unit 13 13A Control unit 14 Data acquisition unit 15 Screening unit 16 16A Budget result calculation unit 17 Display unit 18 Catalog data acquisition unit 19 History data generation unit 20 observation data acquisition unit 30A 30B server

Claims (10)

An earthquake prediction device using observation data acquired by a seismic observation network,
Of the observation data at the time of past earthquakes, the observation data at multiple points, and using the first observation data within a period that goes back a predetermined time from the occurrence of the earthquake at the points, the most recent observation data at the multiple points a screening unit that screens second observation data, which is observation data;
a prediction result calculation unit that calculates a prediction result of at least one of the time of occurrence of an earthquake, the location of an earthquake, and the magnitude of an earthquake using the results of the screening;
An earthquake prediction device comprising 2. The earthquake according to claim 1, wherein said screening unit extracts, as said first observation data, observation data when an earthquake having a magnitude equal to or greater than a predetermined value occurs from observation data acquired by said seismic observation network. prediction device. Both the first observation data and the second observation data are numerical data,
The earthquake prediction device according to claim 1 or 2, wherein said screening unit performs screening by numerically analyzing said first observation data and said second observation data. Both the first observation data and the second observation data are image data,
The earthquake prediction apparatus according to any one of claims 1 to 3, wherein said screening unit performs screening by pattern matching with said second observation data using said first observation data. 5. The prediction result calculation unit predicts the occurrence location of the earthquake based on a combination of a plurality of observation points of the first observation data actually used in the screening. The earthquake prediction device according to item 1. The earthquake prediction device according to any one of claims 1 to 5, wherein said seismic observation network is a high-sensitivity seismic observation network. The earthquake prediction device according to any one of claims 1 to 5, wherein said seismic observation network is a broadband seismic observation network. An earthquake prediction device using observation data acquired by a seismic observation network,
History data at multiple locations, including moving average output values of seismometers installed in the seismic observation network, locations of occurrence of earthquakes, and magnitude of earthquakes within a predetermined period of time prior to the occurrence of an earthquake. a history data generation unit that generates history data; an observation data acquisition unit that continuously acquires a moving average value of the observation data of the output values of the seismometer, which is the observation data at the plurality of points;
Prediction results for calculating prediction results for at least one of the time of occurrence of an earthquake, the location of an earthquake, and the magnitude of an earthquake by comparing the specified value of the moving average value of the observation data with the history data. An earthquake prediction device comprising a calculator. An earthquake prediction device using observation data acquired by a seismic observation network,
Historical data at one or more locations, which are output values of seismometers installed in the seismic observation network, moving average values of the output values, or seismic waveforms within a predetermined period of time before the occurrence of an earthquake. a history data generation unit that generates history data including the location of the earthquake and the magnitude of the earthquake when an abnormality is found in the an observation data acquisition unit that continuously acquires
a prediction result calculation unit that calculates prediction results of at least one of the time of occurrence of an earthquake, the location of an earthquake, and the magnitude of an earthquake by comparing the observation data with the history data. Device. An earthquake prediction method using observation data acquired by a seismic observation network,
Of the observation data at the time of past earthquakes, the observation data at multiple points, and using the first observation data within a period that goes back a predetermined time from the occurrence of the earthquake at the points, the most recent observation data at the multiple points A first step of screening second observation data, which is observation data;
a second step of calculating prediction results of at least one of the time of occurrence of an earthquake, the location of an earthquake, and the magnitude of an earthquake using the results of the screening;
An earthquake prediction method comprising:

Images