JP2004239901A - Earthquake predicting method, earthquake prediction system, earthquake prediction program and recording medium - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To find out the phenomenon of earthquake occurrence which depends with measure from the earthquake phenomena appearing in an earthquake occurrence area, observe earthquake phenomena by the measure and make a short-term earthquake prediction. <P>SOLUTION: The earthquake prediction system is provided with an earthquake element signal preparation means 2 for extracting prescribed earthquakes from earthquakes that have occurred in an area covered and numeralizing earthquake information as Dcj from the seismic center information and information of an earthquake mechanism solution of the earthquakes; a earthquake element time series preparing means 3 for obtaining Dc=(Dc1, Dc2, Dc3..., Dcm...), on the basis of the occurrence sequence of the earthquakes; a primary difference output means 4 obtaining as primary differences ΔDcm and n; a secondary difference output means 5 for obtaining as secondary differences Δ2Dcm, n and k, an amplitude and phase comparison and detection means 8 for comparing and detecting the amplitude and the phase of the secondary differences Δ2Dcm, n and k between different elements and outputting the time series of amplitude relationships and phase relation among elements and earthquake elements. <P>COPYRIGHT: (C)2004,JPO&NCIPI

Description

  The present invention relates to a method for predicting an earthquake, a system for predicting an earthquake, and a program for predicting an earthquake, and in particular, extracts a predetermined earthquake from an earthquake that has occurred in a target area, and uses the information on the source of the earthquake and the information on the focal mechanism solution. The present invention relates to an earthquake prediction method, an earthquake prediction system, an earthquake prediction program, and a recording medium for predicting a future earthquake occurrence time, an epicenter, and an earthquake magnitude.

  It is considered that the release of stress accumulated in the crust and faults near the boundary of the continental plate for a long time is a large earthquake, and there is a cycle where the accumulation of stress and the repetition time of the release are over 100 years, Such long periods are accompanied by fluctuations, which can be as little as 20-30 years. Therefore, short-term prediction of the source, time, and magnitude of a major earthquake of several years to several months, which is indispensable for daily life in a seismically active area, is impossible at the current technological level for the following reasons. It is said that there is no reliable earthquake prediction method or device.

  The reason why the short-term earthquake prediction is impossible is as follows. That is, in an earthquake occurrence other than aftershocks and swarm earthquakes, the occurrence location, occurrence time, and magnitude change randomly, and it is widely known that the statistical properties follow the Poisson process. For this reason, it is said that it is impossible to predict the time when an earthquake occurs beyond stochastic prediction.

  In addition, the universal law indicated by the random nature includes Gutenberg's linear relationship in semi-logarithmic relation in the relational expression between the cumulative number of earthquakes with magnitude greater than a certain magnitude and the magnitude. There is a universal statistical law. According to this universal law, the fact that magnitude is represented by the exponential function of the energy, moment, and rupture area of the generated earthquake indicates that the cumulative frequency of earthquakes with a magnitude value of M or more is a power of those parameters. Will be displayed. The statistical nature of such power variability phenomena is a theoretical guarantee that there is no inherent measure (for example, a measure that measures the magnitude of seismic energy to take a particular magnitude). He concludes that earthquake prediction itself is scientifically impossible.

  However, the following two prediction methods have been proposed for such seemingly random earthquake phenomena.

  The first method of prediction is a recent attempt. This applies a critical phenomenon to a seismic phenomenon because there is a critical phenomenon derived from the statistical physics law among phenomena in which a random phenomenon follows the power of the statistical law. In other words, this method predicts a large earthquake by assuming that the time when the phase of a substance changes in the critical phenomenon is the time when a large earthquake occurs. This method predicts the time at which the critical point appears in a time-varying critical phenomenon known as one of the theoretical solutions in the field of statistical physics, and the periodic oscillation proportional to the logarithm of the time interval of the current time. It is intended to be used.

  For example, if the time is plotted on the horizontal axis and the relative amount of subsidence is displayed on the vertical axis for the vibration phenomenon of subsidence in Omaezaki, the vibration phenomenon vibrates with a period that is approximately proportional to the logarithm of the previous time interval. However, it seems that it is approaching the critical state. Therefore, it is assumed that the theoretical solution in the field of statistical physics describes the oscillation phenomenon of relative land subsidence in Omaezaki. According to the predictions made in 1999, the time to reach the critical state is 2004.7 ± 1.7 years, and there are paper reports that the critical date is the time when the Tokai earthquake occurs. However, it is hard to say that the physical laws leading to such a critical vibration phenomenon describe the physical phenomena of subsidence in Omaezaki. In fact, according to the predictions in that paper, the subsidence of Omaezaki should have been in a state of rising from stoppage in 2002, but as of December 2002, it was still continuing to subside. Even if it occurs in .7 years ± 1.7 years, it must be concluded that the prediction method is incorrect.

  Regarding the prediction method, the theoretical critical vibration phenomenon itself, which was derived earlier, is different from the universally observed earthquake phenomenon that leads to the occurrence of an earthquake, and the critical vibration phenomenon cannot be linked to the earthquake occurrence phenomenon. It has been argued that there is no such thing, and it cannot be said to be a reliable prediction method for earthquakes.

  The second method of prediction is about 15 years ago. This is because some relatively large earthquakes have been observed from several days to several hours before the occurrence of the earthquake, such that the randomness of the time interval between the earthquakes is lost. And tried to use it for prediction.

  As an example, there is an example in which the Disaster Prevention Research Institute of Science and Technology uses a relatively large earthquake in the past in the Kanto Koshinetsu region to detect a point in time at which the statistical property changes from random and uses it for prediction. However, according to this report, there were cases where prediction was not possible, and it seems that it has not been widely used.

  On the other hand, the interpretation of the Universal Statistical Law itself, which has no inherent scale for random earthquake phenomena, has been questioned. For example, Gutenberg's universal law gives the relationship between M and the total number of earthquakes of a magnitude M or more, so the fluctuations of M are smoothed, mostly affected by aftershocks, Some claim that it does not represent a statistical rule that completely denies the possibility of prediction. For example, if only the main shock is expressed by Gutenberg's relational equation, there is multifractality that suggests the existence of a scale that predicts earthquakes, so it is said that there is a fact that there is some unique scale in earthquake phenomena It is clear from the report.

Some of these unique measures include the continental plate movement near the boundary (DB layer) between the hard granite layer (Brittle layer) and the soft basalt layer (Ductile layer) of the earth's lithosphere. Paper reports that when the stress from the basalt layer is transferred from the basalt layer to the transition layer of the basalt layer and granite layer, an earthquake with a unique magnitude of 3 to 4 occurs near the hard layer side of the transition layer. Has been made. Recently, it has been reported that earthquakes associated with asperities existing at the plate boundary have an inherent periodicity. However, many scientists are seeking to apply these scales to earthquake prediction, but have not yet found a prediction method using these scales. Therefore, short-term prediction of the epicenter and the time of occurrence of an earthquake such as the Great Hanshin-Awaji Earthquake (South Kobe Earthquake) is impossible at the current technological level.
JP-A-2000-131444 Patent No. 2858426

  However, if we observe seismic phenomena from another viewpoint, there is a possibility of short-term earthquake prediction. It assumes that the seismic phenomena that occur over a wide area are random, but that if an area where an earthquake is likely to occur is limited, anomalies observed in local crustal deformation may lead to a large earthquake. It is a possibility of short-term forecasting, in which electromagnetic phenomena, groundwater in the crust, changes in chemical substances, etc. caused by earthquakes are used as predictive phenomena of earthquakes, and earthquakes occurring locally are predicted immediately before the occurrence. A study investigating the possibility of such a short-term prediction included a clear predictive change in the chloride content that appeared in groundwater near the epicenter of the Great Hanshin-Awaji Earthquake. It is reported as a verification of predictive phenomena based on statistical data. Therefore, it is proposed to use such predictive phenomena, but for large earthquakes that will occur somewhere in the wide range of earthquakes in the future, the reproducibility, universality, operability of the predictive sign detection method, and There are many difficult problems to solve for reliability. Therefore, universal short-term earthquake prediction is considered to be scientifically difficult, as it was not possible to predict the Great Hanshin-Awaji Earthquake immediately before.

  At present, the Geospatial Information Authority of Japan (GEONET), a network of GPS (GEONET) networked throughout Japan, provides daily crustal displacement information for each GPS site. Therefore, abnormal crustal movements appearing at a GPS site located near the epicenter of a future earthquake can be detected using the motion change detection device of Japanese Patent No. 2787143. In this case, if it can be assumed that the abnormal crustal deformation observed at the GPS site is a sign of an earthquake that may occur in the vicinity, it has been demonstrated that the sign detection can be detected several weeks ago. Short-term or short-term prediction is possible. However, it is often difficult to predict not only a certain magnitude of an earthquake but also a distinction between a relatively large earthquake and a large earthquake only from the detection of abnormal crustal deformation.

  The present invention finds a phenomenon of earthquake occurrence depending on a scale from seemingly random earthquake phenomena appearing in an earthquake occurrence zone, regardless of the size of the area of the earthquake occurrence zone, and observes the earthquake phenomenon on the scale. It is an object of the present invention to provide a method and apparatus for short-term earthquake prediction (forecasting the time of occurrence of an earthquake, the epicenter, and the magnitude of magnitude). According to the present invention, the detection of an abnormal change in crustal deformation using the motion change detecting device of Japanese Patent No. 2787143 can be regarded as the detection of an earthquake sign. The purpose of the present invention is to provide a short-term prediction device for an earthquake.

  The invention according to claim 1 extracts a predetermined earthquake from earthquakes that occurred in a target area, and, based on source information of the earthquakes and information on a focal mechanism solution, sets a parameter representing each element of the earthquake information as c. A seismic element signaling step of quantifying the seismic information regarding each element c of the extracted j-th earthquake as Dcj in order to observe the seismic phenomenon of the selected area in the form of time-series data of each element c; The Dcj extracted in the earthquake element signalizing step is directly or predetermined selected or smoothed on the basis of the earthquake occurrence order j (j = 1, 2, 3,..., M,...). , And a time sequence of seismic elements to obtain Dc = (Dc1, Dc2, Dc3,..., Dcm,...); And an arbitrary earthquake occurrence order m of Dc obtained in the time series step of the seismic elements. At an arbitrary interval n (n = 1 ,...) And a difference proportional to the difference between Dcm and Dcm-n as a first-order differential value ΔDcm, n, and a time-series value ΔDcm output from the first-order differential output step. , n at an arbitrary interval k (k = 1, 2,...), that is, a quantity proportional to the difference between ΔDcm, n and ΔDcm−k, n is a secondary difference value in an arbitrary earthquake occurrence order m. The secondary differential output step obtained as Δ2Dcm, n, k and the amplitude and phase of the secondary differential value Δ2Dcm, n, k obtained in the secondary differential output step are compared and detected between different elements. And an amplitude / phase comparison / detection step of outputting a time series of an earthquake element and an amplitude relation and a phase relation.

  According to a second aspect of the present invention, according to the first aspect, an earthquake is generated from the amplitude relation and the phase relation between the elements and the time series of the earthquake elements, which are detected by the amplitude phase comparison detection step between the elements of the earthquake. An earthquake prediction method comprising the steps of predicting time, epicenter, and magnitude of an earthquake.

  According to a third aspect of the present invention, in the first or second aspect, the number of occurrences of an arbitrary earthquake is set to a (a = 1, 2,...) And extracted by the earthquake element signalizing step and the time series forming step. This is a method for predicting an earthquake, comprising a time interval accumulation detecting step of accumulating a number of time intervals apart from each other by one to an arbitrary order m of occurrence of an earthquake and setting it as CIm (a).

  According to a fourth aspect of the present invention, in any one of the first to third aspects, the number of occurrences of an arbitrary earthquake is a (a = 1, 2,...), And the m-th earthquake and the m-a-th earthquake This is a method for predicting an earthquake, comprising a time interval accumulation detection step of setting a difference between occurrence times to CIm (a).

  According to a fifth aspect of the present invention, in any one of the first to fourth aspects, the time series of the seismic elements obtained in the seismic element time series-converting step is filtered by a band-pass filter for each element c of the earthquake. An earthquake prediction apparatus comprising: a filter output unit that obtains Fcm in an earthquake occurrence order m and outputs the Fcm to the amplitude-phase comparison detection step.

  According to the invention of claim 6, a predetermined earthquake is extracted from an earthquake that occurred in a target area, and from the source information of the earthquakes and the information of the focal mechanism solution, a parameter representing each element of the earthquake information is set as c. Seismic element signalizing means for quantifying earthquake information on each element c of the extracted j-th earthquake as Dcj in order to observe the seismic phenomenon in the selected area as time-series data of each element c; The Dcj extracted by the earthquake element signal generator is directly or predetermined selected or smoothed on the basis of the earthquake occurrence order j (j = 1, 2, 3,..., M,...). , And Dc = (Dc1, Dc2, Dc3,..., Dcm,...), And an arbitrary earthquake occurrence order m of Dc obtained by the time-series means of the seismic element. At an arbitrary interval n (n = 1, 2, 3,...) Primary differential output means for obtaining an amount proportional to the difference, the difference between Dcm and Dcm-n as a primary differential value ΔDcm, n, and an arbitrary interval k (k) of a time series value ΔDcm, n output from the primary differential output means. = 1, 2,...), That is, an amount proportional to the difference between ΔDcm, n and ΔDcm−k, n is obtained as a secondary difference value Δ2Dcm, n, k in an arbitrary earthquake occurrence order m. The second derivative output means compares the amplitude and phase of the second difference value Δ2Dcm, n, k obtained in the second derivative output step between different elements, and detects the amplitude relation and phase relation between each element and the earthquake. An earthquake prediction device comprising: an amplitude / phase comparison / detection unit that outputs a time series of elements.

  According to a seventh aspect of the present invention, according to the sixth aspect, an earthquake is generated from the amplitude relation and the phase relation between the elements and the time series of the earthquake elements detected by the amplitude and phase comparison detection means between the elements of the earthquake. An earthquake prediction device including means for predicting time, epicenter, and magnitude of an earthquake.

  According to the invention of claim 8, in claim 6 or claim 7, the number of occurrences of an arbitrary earthquake is set to a (a = 1, 2,...) And extracted by the earthquake element signal generation means and the time series generation means. This is an earthquake prediction device that accumulates a number of time intervals of one earthquake and one earthquake apart from each other up to an arbitrary order m of occurrence of the earthquake to obtain CIm (a).

  According to a ninth aspect of the present invention, in any one of the sixth to eighth aspects, the arbitrary number of earthquakes is a (a = 1, 2,...), And the m-th and m-a-th earthquakes are generated. This is an earthquake prediction device including time interval accumulation detection means for setting a time difference to CIm (a).

  According to a tenth aspect of the present invention, in any one of the sixth to ninth aspects, the time series of the seismic elements obtained by the seismic element time-series means is filtered by a band-pass filter for each element c of the earthquake. An earthquake prediction device comprising: a filter output unit that obtains Fcm in an arbitrary earthquake occurrence order m and outputs the Fcm to the amplitude / phase comparison detection unit.

  The invention according to claim 11 extracts a predetermined earthquake from an earthquake that occurred in a target area, and, based on source information of the earthquake and information on a focal mechanism solution, sets a parameter representing each element of the earthquake information as c, A seismic element signaling step of quantifying the seismic information regarding each element c of the extracted j-th earthquake as Dcj in order to observe the seismic phenomenon of the selected area in the form of time-series data of each element c; The Dcj extracted in the seismic element signalizing step is subjected to direct or predetermined selection processing or smoothing processing based on the earthquake occurrence order j (j = 1, 2, 3,..., M,...). , Dc = (Dc1, Dc2, Dc3,..., Dcm,...), And an arbitrary seismic sequence m of Dc obtained in the seismic element time series step. Interval n (n = 1, , 3,...) And a difference proportional to the difference between Dcm and Dcm-n as a first-order differential value ΔDcm, n, and a time-series value ΔDcm, output from the first-order differential output step. n at an arbitrary interval k (k = 1, 2,...), that is, an amount proportional to the difference between ΔDcm, n and ΔDcm−k, n is a secondary difference value Δ2Dcm in an arbitrary earthquake occurrence order m. , n, k, the secondary differential output step, and the secondary differential value Δ2Dcm obtained in the secondary differential output step, the amplitude and the phase of n, k are compared and detected between different elements, An earthquake prediction program including an amplitude relationship, a phase relationship, and an amplitude phase comparison detection step of outputting a time series of earthquake elements.

  According to a twelfth aspect of the present invention, in accordance with the eleventh aspect, an earthquake is generated from the amplitude relation and the phase relation between the respective elements and the time series of the seismic elements, which are detected by the amplitude and phase comparison detection step between the respective elements of the earthquake. This is an earthquake prediction program that includes a step of predicting time, epicenter, and magnitude of an earthquake.

  According to a thirteenth aspect of the present invention, in the eleventh or twelfth aspect, the number of occurrences of an arbitrary earthquake is set to a (a = 1, 2,...) And extracted by the seismic element signalizing step and the time series forming step. This is a program for predicting an earthquake, comprising: a time interval accumulation detecting step of accumulating a number of time intervals of one earthquake and one time interval apart from each other up to an arbitrary order m of occurrence of an earthquake to obtain CIm (a).

  According to a fourteenth aspect of the present invention, in any one of the eleventh to thirteenth aspects, the number of occurrences of an arbitrary earthquake is a (a = 1, 2,...), And the m-th and m-a-th earthquakes are set. A time interval cumulative detection step of setting a difference between occurrence times to CIm (a).

  The invention according to claim 15 is characterized in that the time series of the seismic elements obtained in the seismic element time series step is filtered by a band-pass filter for each element c of the earthquake to obtain Fcm in an arbitrary order m of occurrence of the earthquake and obtain the amplitude. It is a program for predicting an earthquake, comprising a filter output means for outputting to a phase comparison detection step.

  According to a sixteenth aspect of the present invention, there is provided a recording medium storing the prediction program according to any one of the tenth to fifteenth aspects.

  According to the first, sixth, and tenth aspects of the present invention, the seismic element signaling step or the seismic element signaling means extracts predetermined earthquakes from the earthquakes that occurred in the target area and records the source information of those earthquakes. And the focal mechanism solution information, the parameter representing each information element of the earthquake is set as c, and the seismic phenomenon in the selected area is represented by time-series data of each element c, and the extracted j-th The earthquake information relating to each element c of the earthquake is quantified as Dcj. Specifically, the elements related to the geographical space, time interval, and magnitude of past earthquakes that have occurred are described as latitude (LAT), longitude (LON), depth (DEP), and time interval between earthquakes (INT ), The magnitude (MAG) is quantified, and the elements related to the rupture mechanism of the earthquake obtained from the information on the focal mechanism solution are expressed as the travel (R) and inclination (DEC) of the nodal surface 1 (N1) and 2 (N2). ), The azimuth (DIR) and vertical angle (VER) of the P axis (PA) and T (TA) axes are digitized.

  In other words, from the source information of the past earthquake, the element c is the following five parameters, the numerical data of the earthquake element represented by LAT, LON, DEP, INT and MAG, and from the mechanism solution, the element c is The following eight numerical data of seismic elements represented by N1-R, N1-DEC, N2-R2, N2-DEC, PA-DEC, PA-VER, TA-DIR and TA-VER .

  The time series forming step or the time series forming means converts the Dcj extracted by the seismic element signal forming step or the seismic element signal forming means into an earthquake occurrence order j (j = 1, 2, 3,..., M,. ) Based on the seismic element that obtains Dc = (Dc1, Dc2, Dc3,..., Dcm,...), Or performs a predetermined selection or smoothing process described later, and Dc = (Dc1, Dc2, Dc3). , ..., Dcm, ...) The time series forming step or time series forming means can regard the change of the seismic phenomenon from the past to the present in the selected area as a displacement time series of each element c of each earthquake generated by the movement of the seismic system. .

  According to the second, third, seventh and eighth aspects or the twelfth and thirteenth aspects, when a predetermined selection or smoothing step is performed, for example, when a moving average is obtained for w earthquakes, In addition, the randomness of occurrence of earthquakes can be averaged by w pieces, thereby facilitating observation of even a little random earthquake phenomenon.

  In particular, when the seismic element c is the epicenter element LAT, LON, DEP, these time series Dc describe the temporal transition of the epicenter to the present, and using the scale described later, the future epicenter location Can be estimated.

  The primary differential output step or the primary differential output means calculates a difference of arbitrary intervals n (n = 1, 2, 3,...) In an arbitrary earthquake occurrence order m of Dc obtained in the time-sequencing step of the seismic element; The amount proportional to the difference between Dcm and Dcm-n is determined as a primary difference value ΔDcm, n. A variable component in which Dcm and Dcm-n separated by an interval n are always substantially opposite signs is obtained in time series. It is extracted as a large value from random fluctuation of data.

  Therefore, the primary differential output step or the primary differential output means includes, in seemingly randomly varying time-series data, a variation having a different sign every n, that is, a periodic variation having a period of approximately 2n. Then, only the periodic fluctuation component is largely extracted from the random fluctuation, and at the same time, an amount proportional to the speed, which is the primary rate of change of Dc, is obtained.

The secondary differential output step or the secondary differential output means calculates a difference of the time series ΔDcm, n at an arbitrary interval k (k = 1, 2,...), That is, a difference between ΔDcj, n and ΔDcj−k, n. , Is calculated as a secondary difference value Δ ² Dcm, n, k in the earthquake occurrence order j. The secondary differential output step or the secondary differential output means performs the primary differential output from the time series data of the seismic element if there is a fluctuation component such that ΔDcm, n and ΔDcm−k, n are always substantially opposite signs. In the step, only the fluctuation component whose period becomes approximately 2k is extracted more greatly from the period fluctuation components of which the output period is approximately 2n.

  The secondary differential output step or secondary differential output means simultaneously obtains an amount proportional to the acceleration, which is the secondary rate of change of Dcm. Therefore, if the distances n and k are set to the same value (k = n), only the fluctuation component of the seismic element time-series data Dc having a periodic fluctuation component of approximately 2k can be largely extracted. .

  For example, when n = k, a periodic variation having a substantially periodicity of 2k is extracted as a scale from an earthquake phenomenon that seemingly appears to have no scale at all. Therefore, if a seemingly random seismic phenomenon is observed on this scale, not only can the periodic amplitude change of the seismic phenomenon expressed for each element c be measured in a predictive manner, but the period can also be used as a reference clock for time measurement. It is possible to predict future earthquake phenomena from past earthquake phenomena.

An amplitude-phase comparison detection step of comparing and detecting the amplitude and phase of the secondary difference value Δ ² Dcm, n, k of the order m of earthquake occurrence obtained by the secondary differentiation output step or the secondary differentiation output means between different elements. Can extract, from the element time series, information on an element time series having a special fluctuation period component and information on mutual comparison of amplitude and phase between elements of acceleration information.

  Therefore, the periodic oscillation selected at the interval k (n = k) described above appears as a periodic phenomenon (scale) in an earthquake phenomenon that seems to be seemingly random without a scale. Observing seismic phenomena on a periodicity scale, if a large earthquake in the past approaches, such as the Great Hanshin-Awaji Earthquake on January 17, 1995, the earthquakes that have occurred so far Anomalies appear in the cross-correlation information of the periodic oscillations. When an abnormality appearing in this periodic phenomenon is detected, it is possible to predict not only when a large earthquake will occur using the periodicity of the seismic phenomenon, but also the magnitude of the stress accumulated in the crust and the critical state of the stress.

  In this way, past earthquakes that occurred in a wide range of earthquake occurrence zones are detected by the amplitude-phase comparison detection step or the amplitude-phase comparison detection means, using the amplitude and phase relationships between the elements to predict the future earthquake. It is possible to configure an earthquake prediction method comprising detecting a phenomenon, predicting the occurrence time and magnitude from the prediction, and predicting the epicenter from the time series Dc of the earthquake elements LAT, LON, and DEP.

  Further, according to the second, third, sixth, seventh and eleventh aspects, the time interval cumulative detection step or the time interval cumulative detection means determines the number of occurrences of an arbitrary earthquake as a (a = 1, 2 , ..), and a time interval INT extracted by the seismic element signalizing step and the time-sequencing step is accumulated up to an arbitrary earthquake occurrence order m, and the result is CIm (a), or the m-th earthquake and m−a The difference between the times of occurrence of the th earthquake is CIm (a). This is because a seemingly random change of the time interval INT between earthquakes is continuously added by a, so that some randomness existing in the time series Dc of the element INT is removed, and CIm (a) becomes If it gets bigger, it means that a lot of time has been spent on the occurrence of a earthquakes, which means that the number of earthquakes per unit time has decreased, so it can be expressed that the calm of seismic activity has increased it can.

  At this time, if the number a to be accumulated (added) is set to k, which corresponds to the period of the scale for observing the seismic phenomenon described above, CIm (k) indicates the calm of the seismic activity that evolves into a large earthquake as follows. It can be classified into three stages and describe the progress of calm and the magnitude of the stage.

  The first stage is the beginning of the calm that has been observed several years before the earthquake. The second stage is a critical calm state where the stress state accumulated in the crust near the epicenter where a large earthquake is likely to occur about 1 to 6 months ago seems to have reached a critical state. The third stage is a stage in which the calm ends a few days and several hours before the occurrence of the large earthquake, and ends immediately before the occurrence of the large earthquake. Therefore, the transition of each stage of the CIm (k) and the magnitude and width of the CIm (k) profile have an effect of linking the occurrence time and magnitude of the earthquake that will occur to the parameters for predicting the magnitude.

  According to the fifth, tenth, and fifteenth aspects, the filter output step or the filter output means converts the time series of the seismic elements obtained in the time series of the seismic elements into a band for each element c of the earthquake. The Fcm, which is a measure of seismic phenomenon observation in an arbitrary earthquake occurrence order m, is filtered by a pass filter, and the amplitude and phase of the Fcm in claims 1, 2, 3, 3, 5, 6, 7, 9, 10, and 11 are calculated. Output to the comparison detection step.

  Earthquakes that occurred in an arbitrarily selected wide-area earthquake occurrence area are selected, and the earthquake elements are digitized in the earthquake element signalizing step. Then, in the time series forming step, the numerically converted seismic elements are made into a time series, and if only the periodic fluctuations are extracted by the secondary differential output step or the band pass filter of claim 3, the periodic fluctuations of the extracted elements are seemingly random. Can be used as a scale to observe seismic phenomena in selected areas. Since the scale to observe fluctuates periodically, if the amplitude and phase change of the vibration are monitored in the amplitude-phase comparison step, the phase inversion that occurs between the vibrations of the seismic elements INT and DEP and the vibration of the seismic element MAG A sign of a relatively large earthquake can be extracted from the relative amplitude information. Furthermore, after extracting the sign of calm, the periodicity of the vibration acts as a time base function of the clock, so the time at which the calm ends, that is, the time at which the earthquake occurs, can be accurately determined by the number of events to occur from the sign calm detection. I can foresee. In the time interval accumulation detecting step of claim 2, the CIm (k) obtained by accumulating the number of events corresponding to the fluctuation period observes the quietness of the seismic activity smoothed (accumulated) by the number of events corresponding to the substantially period. In other words, it is possible to clearly extract the process of the sign of the earthquake, the magnitude of the magnitude, and the process of the stress-strain reaching the critical state.

  At the same time as predicting the time of the earthquake, it is possible to estimate the average epicenter at the time of occurrence from the selected large earthquake zone from the time-series information of the LAT, LON, and DEP seismic elements. The source of a relatively large earthquake can be predicted. Whether the earthquake that occurs is as large as the Great Hanshin-Awaji Earthquake can be estimated from the size of the profile that characterizes the three-stage predictive tranquility that appears in CIm (k). In addition, the periodic fluctuation of the antiphase of DEP and MAG in the elemental time series that appeared for several years before the Great Hanshin-Awaji Earthquake, when a stress strain is generated by an earthquake and the strain is accumulated, eventually a critical state , The number of vibrations having the opposite phase can be used as a reference for determining whether or not the future earthquake is a large earthquake. Also, when the swarm earthquake ends, for example, in the Chugoku region, when the swarm earthquakes 10a and 10g end, the characteristic profile on CIm (k) appears approximately 70 days after the swarm earthquake ends. This is a sign of the 1984 Hyuga Nada Earthquake (10b) and the 2001 Akinada Earthquake (10h) that occurred on the sinking Philippine Plate. Predictions can also be made from earthquake observations.

  Furthermore, after predicting the epicenter of the earthquake that will occur, a GPS site near the epicenter predicted from the GPS network (GEONET), which was set up by the GSI like a network in Japan, was obtained. (If there is no site, install GPS) If the daily crustal movement is time-series using the motion change detection device of Japanese Patent No. 2787143, abnormal crustal deformation that may occur due to crustal movement change , 2-3 weeks to several weeks before, it has been proved that it can be reliably used in combination with the earthquake prediction method of the present invention.

  According to the sixteenth aspect, the program according to any one of the eleventh to fifteenth aspects can be executed by a predetermined processing device.

  Therefore, according to the present invention, in any region of the earthquake occurrence zone, each element information of the earthquake that is occurring every day, the sign of the earthquake is detected based on the crustal deformation information, the time of the occurrence of the earthquake, the epicenter, the magnitude, Predictions can be reliably made in real time.

  Hereinafter, the best mode for carrying out the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing an earthquake prediction device. This apparatus includes an earthquake source and focal mechanism information source 1, an earthquake element signal generator 2, a seismic element time series generator 3, a primary differential output unit 4, a secondary differential output unit 5, a filter output unit 6; a time interval accumulation detecting means 7; an amplitude / phase comparison detecting means 8; In this example, the earthquake prediction device is realized by executing the earthquake prediction program by a computer using information obtained from the earthquake source / focal mechanism information source 1.

  Next, an example in which the above-described earthquake prediction device is applied will be described in detail with reference to FIGS. 1 to 11, mainly taking into account the Great Hanshin-Awaji Earthquake of January 17, 1995 from the past earthquakes.

  The earthquake source and focal mechanism information source 1 extracts, for example, hypocenter information and focal mechanism information from a network of seismometers spread throughout Japan online or offline, and outputs seismic element signal generation means 2 to prediction means 9 for earthquakes. Input to the earthquake prediction device. Then, each means of the apparatus processes the earthquake information and makes predictions of the time of occurrence, the epicenter, and the magnitude of the earthquake.

  Figures 2, 3, and 4 show the results of a test conducted in the Chugoku region (32 ° -36 ° N, 131.5 ° -136 ° N) to test whether the prediction of the 1995 Great Hanshin-Awaji Earthquake that occurred in the Chugoku region was successful. This is the epicenter distribution of the Chugoku region earthquakes of magnitude ≥3.0 and hypocenter depth of less than 300km from the Japan Earthquake's National Catalog of Outbreaks from January 1983 to September 2001. Fig. 3 is a longitudinal cross-sectional view of the hypocenter distribution of each epicenter shown in Fig. 2 with the vertical axis representing latitude and the horizontal axis representing the epicenter depth in semi-logarithmic units in km. It is. FIG. 4 is a cross-sectional view of the epicenter distribution in the depth direction of each epicenter shown in FIG. 2, in which the horizontal axis represents the longitude of the epicenter and the vertical axis represents the depth in semilogarithmic units in km, in a latitudinal direction.

  FIG. 5 shows an embodiment in which the extracted source element information of the earthquake is time-series by a seismic element signalizing step, a seismic element time-series step, and a secondary differential output step.

  FIG. 6 is an embodiment showing that a periodic variation having a period of 60 events substantially common to the time series of five earthquake elements obtained from the source element information of the extracted earthquake exists in the time series spectrum. is there.

  FIG. 7 is an embodiment illustrating a filter configuration having a function equivalent to the second-order differential output step as an example of configuring the filter output step.

  FIG. 8 shows an embodiment in which the bandpass filter function of the filter output step shown in FIG. 7 is displayed in spectrum.

  FIG. 9 shows that the sign of quietness has been detected several years before the occurrence of a large earthquake such as the Great Hanshin-Awaji Earthquake from the Chugoku region of a wide selection area using CI (60) and CI (20) detected using the time interval cumulative detection step. This is an example showing what is observed.

  FIG. 10 is an embodiment in which a large earthquake in the Chugoku region and a sign of a relatively large earthquake are detected by the amplitude phase detection step.

  FIG. 11 shows a case where the number of events in FIGS. 5 and 10 is extracted from the range of 480 to 800, and after detecting a sign of the 1995 Great Hanshin-Awaji Earthquake, the prediction of the occurrence time (the prediction time indicated by the number of events) and This is an example showing how to obtain the prediction of the epicenter.

  In FIG. 1, an earthquake epicenter and focal mechanism information source 1 is a hypocenter and a focal mechanism information source obtained from a network of seismographs spread throughout Japan described above. The earthquake element signal generator 2 receives the earthquake information online or offline, and quantifies information on each element c of the j-th earthquake as Dcj, for example. Next, the time-series unit 3 performs a direct or predetermined selection or smoothing step on Dcj based on the earthquake occurrence order j (j = 1, 2, 3,..., M,...) The time series of each element c of the earthquake is Dc = (Dc1, Dc2, Dc3,..., Dcm,...).

Then, the first derivative output unit 4, the earthquake order m from the Dc, primary relates interval n and k (ΔDcm, n) the amount of difference, the second derivative output unit 5 in the secondary (delta ² Dcm, n, k) is obtained. The filter output means 6 filters the time series Dc of the earthquake elements obtained by the time series generation means 3 with a band-pass filter for each element c of the earthquake to obtain Fcm in an arbitrary earthquake occurrence order m.

The time interval accumulation detecting means 7 accumulates k INTs of the element c of the time series Dc obtained by the time series generating means 3 and obtains CIm (k) representing the three-stage silence of seismic activity. The amplitude / phase comparison and detection means 8 calculates the amplitude and phase of the Fcm obtained by band-pass filtering the secondary difference value Δ ² Dcm, n, k obtained by the secondary differentiation output means 5 or Dc obtained by the filter output means 6. Compare and detect between different elements. This is because of the relationship between the amplitude and the phase of the periodic fluctuation component selected and extracted by the secondary differential output means 5 or the filter output means 6 in each element c, a sign of an earthquake, that is, a crustal terrestrial likely to develop into a large earthquake The purpose of this is to detect the stress-strain situation and the sign of reaching the critical stress state.

  The prediction means 9 uses the information supplied from the time-series means 3 and the time interval accumulation detection means 7 on the basis of the earthquake sign detected by the amplitude / phase comparison detection means 8 to determine the occurrence time of the earthquake, the epicenter and the earthquake. Forecasting the magnitude (for example, the geographical location display of the prediction epicenter in FIGS. 1, 2, and 3 and the progress of the prediction using FIGS. 9 and 11) is analyzed in real time every day and the information is displayed.

  Next, we will verify whether it is possible to predict the past large earthquakes that occurred in the Chugoku region from 1984 to September 2001, using only the epicenter information, using the illustrated means. Above all, an example of the prediction verification of the 1995 Hanshin-Awaji Earthquake will be described in detail with reference to the drawings. In this example, in order to briefly explain the earthquake prediction, the focal mechanism solution that gives the spatiotemporal change information of the heterogeneous fault structure and the spatiotemporal distribution information of the asperity at the plate boundary is not used as the seismic information source.

  In Fig. 2, the earthquakes displayed on the map were in the Chugoku region (32 degrees-36 degrees N, 131.5 degrees-136 degrees N), and the magnitude of the earthquakes from January 1983 to September 2001 Earthquakes with an epicenter of more than 3.0 and a depth of less than 300 km were collected from the Japan Meteorological Agency's national epicenter catalog. The horizontal axis in the figure represents longitude, and the vertical axis represents latitude. The epicenter of each earthquake indicated by a dot has a magnitude (MAG) of 3.0 or more and less than 3.5, and the epicenter indicated by a circle is an earthquake with a MAG of 3.5 or more. Among the large earthquakes since 1984, six earthquakes with MAG of 6 or more and two swarm earthquakes are labeled 10a to 10h in the order of occurrence. 10a is a swarm earthquake of 1984/05/30 (year / month / day), 10b is a 1984/08/07 MAG earthquake of 7.1, 10c is a 1985/05/13 MAG of 6.0 earthquake, 10d Is the Great Hanshin-Awaji Earthquake with a MAG of 7.2 on January 17, 1995, 10e is a 6.3 earthquake with a MAG of June 25, 1997, 10f is a Tottori earthquake with a MAG of 2000/10/06, 7.3. 10g is the earthquake swarm of 2001/12/12 and 10h is the Akinada earthquake of 6.7 MAG on 03/24/2001.

  In FIG. 3, the quake mark of a circle with a MAG of 3.5 or more shown in the latitudinal sectional view indicates a situation in which the northern end of the Philippine plate is subsided below the southern end of the Eurasian plate. The linear trend labeled 11 is a trace of a periodic fluctuation (vibration) having a cycle of approximately 60 events, which will be described later, among earthquakes having a MAG of 3.5 or more. The coordinate origin of the vibration is approximately equal to the average value (34 degrees N, 133.75 degrees E, 30 km) of approximately 700 hypocenters immediately before the Great Hanshin-Awaji Earthquake. Therefore, it is assumed that the average value of the epicenter is the geographic center of the Chugoku region shown in FIG. 2 and the coordinate origin of the vibration is fixed.

  In Fig. 4, the seismic trace indicated by a circle with a MAG of 3.5 or more, which is shown in a longitudinal sectional view, shows that the Philippine plate is located in a substantially arc shape on a semilogarithmic axis below the Eurasian plate. Is represented.

  Therefore, in FIG. 2, FIG. 3 and FIG. 4, the Chugoku region of the earthquake prediction target area used in the example is the sedimentation zone of the Philippine plate (in FIG. 3, the hypocenter distribution where the lower right 10b, 10c, and 10h exist). Fig. 4 shows the trace distribution of the earthquake where the epicenter is deeper in the arc shape and the southern end of the Eurasian plate (10a, 10d, 10e, 10f, 10g at the upper left of Fig. 3). (Shallow horizontal band-like distribution).

  In FIG. 5, the time series Dc and the secondary difference information of each seismic element obtained in each step of 2, 3 and 5 in FIG. 1 will be described. The horizontal axis indicates the order of earthquake occurrence (the number of events), and all the sources of the collected earthquakes, the time of occurrence, and the magnitude information are shown in five rows (LAT, LON, DEP) for each of the two or three steps. , INT, and MAG), and the time series Dc directly obtained are shown as 12a in the LAT row, 15a in the LON row, 18a in the DEP row, 21a in the INT row, and 24a in the MAG row from the top. As an example of the smoothing step, similarly, a time series obtained by moving and averaging the directly obtained time series Dc by 20 events is similarly calculated from the top as 13a in the LAT row, 16a in the LON row, 19a in the DEP row, and 19a in the INT row. 22a and 25a in the MAG line.

  The reference values for the graph display of these time series Dc are as follows: LAT is 34 degrees latitude indicated by 12b, LON is 133 degrees longitude at 15b, DEP is 30km at 18b, INT is 200 hours at 21b, MAG is 5.0 hours at 24b. It is. Each scale indicates an offset value from the reference value of the graph. From the top, LAT and LON are 200 times, DEP is 10 times, INT is 2 times, and MAG is 400 times. A negative offset value is taken upward and a positive offset value is taken downward. Therefore, the positive and negative 600 of LAT and LON correspond to an increase and decrease of 3 degrees from 34 degrees and 133 degrees, respectively. That is, -600 is 31 degrees in the case of latitude (LAT), and LON is 130 degrees. In addition, DEP has a scale of -300 at 0 km, INT has -400 at 0 hour, MAG has -600 at -600 and +600 at 6.5.

  Furthermore, the latitude of 35.48 degrees indicated by 12c at the lower left of the LAT line indicates the latest epicenter latitude during monitoring, and thus indicates the latitude of the last event on the right end. Similarly, in the case of longitude (LON), 15c is the latest longitude being monitored, the depth of the epicenter (DEP), 18c is the latest DEP value being monitored, and in the case of INT, 21c is the latest value being monitored, MAG In the case of, 24c displays the latest MAG numerical value.

  Furthermore, when 24a representing the Dc of the MAG is 6 or more, it is marked with a bold line, and next to it, the labels corresponding to the Dc corresponding to 6 or more earthquakes excluding the swarm earthquakes are labeled 10b to 10h. Is written. For example, the Great Hanshin-Awaji Earthquake, which is described in detail as an example of prediction, corresponds to 10d.

  Time-series fluctuations of the seismic elements collected in the Chugoku region, 12a, 15a, 18a, 21a, and 24a, show that all events seem to appear at random. Very small fluctuations in the time series (except in the case of MAG) indicate that these events are swarms or aftershocks of large earthquakes. The localization of LAT, LON, and DEP, with very short INTs, reveals the location of these successive events. The time series 21a in which the earthquake element c is INT indicates that calm (long INT) appears randomly.

  Therefore, when looking back on the seismic activity in the earthquake occurrence area after the occurrence of the large earthquake, the sign of the presence of the silence is not recognized. However, as shown in the graphs of 13a, 16a, 19a, 22a, and 25a, the randomness of the element time series can be removed to some extent by taking a moving average of 20 events. At this time, the maximum peak of the time series 22a of the element INT is about 280 hours immediately before the Great Hanshin-Awaji Earthquake, and the longest maximum peak in the time series is only after reviewing the entire time series, It can be recognized as a sign of abnormal calm. This quietness is further determined by using the CIm (60), which accumulates 60 INTs and classifies the quietness of seismic activity into three levels in the description of FIG. 9 described later in the time interval cumulative detection step 7 using the Great Hanshin-Awaji Earthquake. Before the earthquake occurs, you can recognize that the earthquake is a clear sign of calm.

Further, in FIG. 5, using the Dc obtained by moving-averaging the time series of each element in 20 events, n = k = 30 is set in steps 4 and 5, and obtained in the second derivative output step 5 in FIG. The secondary difference value Δ ² Dcm, n, k (n = k = 30 events) is set to 14a, 17a, 20a, 23a, and 26a based on 14b, 17b, 20b, 23b, and 26b, respectively, as an earthquake element c. Are displayed relative to LAT, LON, DEP, INT, and MAG in order from the top. Since the time series of these secondary difference values is n = k = 30 in 5 in FIG. 1, the period of the fluctuation component of the seismic element time series Dc is approximately equal to twice the number of events of k. Only the periodical variation component of 60 events is extracted. The periodicity of the fluctuation component is a scale (observation window) for observing the seismic phenomenon. Then, the phase-amplitude relative change detection step 8 in FIG. 1 provides a specific observation step. As will be described later in detail with reference to FIG. 10, a sign of an earthquake is detected using the observation step, and an occurrence time indicated by the epicenter and the number of events of the earthquake is predicted.

  In FIG. 6, the fact that such a periodic fluctuation component exists in the time series Dc obtained directly or after smoothing in step 3 is shown by a time series spectrum of the seismic element. As an example, a time series spectrum obtained by moving average of the time series Dc directly obtained in step 3 by four events is displayed. The earthquake used to obtain the spectrum was 700 events right before the Great Hanshin-Awaji Earthquake. The vertical axis indicates the relative amplitude value, and the horizontal axis indicates the frequency (1 / number of events). 27 indicates LAT, 28 indicates LON, 29 indicates DEP, 30 indicates INT, and 31 indicates MAG spectrum. These spectra show that the time series Dc of each seismic element fluctuates almost randomly. In other words, the seismic phenomenon gives the impression that it appears to be appearing randomly, but there is also a periodicity of about 60 events of a periodic variation that is substantially common to each earthquake element. The cycle of approximately 60 events common to these elements is indicated by a vertical arrow 32. The frequency at this time is approximately 0.017 (1 / event), which is the reciprocal of the period 60 events. When the frequency is 0.1 (1 / event) or less, Dc has the same spectral structure whether or not Dc is moving averaged in four events. Also, the periodicity of approximately 60 events is directly observed from the moving average of the 20 elementary time series, that is, 13a, 16a, 19a, 22a, and 25a. This periodicity seems to be slightly different from the periodicity resulting from repeated mainshock and aftershock. Therefore, this periodic fluctuation (vibration) is used as an observation scale of the seismic phenomenon. Note that the horizontal arrow 33 indicates the band-pass range of the secondary differential output step 5 because it implies the role of a band-pass filter, as described in FIG. . The area corresponds to a range of 120 to 40 events in a cycle.

  In FIG. 7, rectangles 35, 36, and 37 are used to explain a function of extracting a specific periodic variation included in the second derivative output step 5 and to configure a band-pass filter equivalent to the function. It is an example of a block diagram of a wave and its combination. These figures are constructed based on Japanese Patent Nos. 2787143 and 3044228 and U.S. Pat. No. 5,562,141. In the figure, the vertical axis represents the magnitude of the earthquake element time series, and the horizontal axis represents the number of events. Reference numeral 34 denotes a directly obtained time series in which an element time series with an earthquake is not smoothed. For simplicity, the number of events is from 1 to 15. Accordingly, 34a, 34b, and 34c indicate the values of the element time series of the earthquake corresponding to the event numbers of 5, 10, and 15, respectively. Next, a detailed function of extraction (filter) included in step 5 in FIG. 1 will be described.

  An example of the selection or smoothing step performed immediately before the seismic element time series of No. 3 is a moving average of the time series 34. This moving average is obtained by moving the illustrated rectangular wave 35 to the right side by one event each time an arbitrary observation event m increases by one event, and sets a time series 34 indicated by three arrows in 35. By reading the value of 3 events (numbers), adding all the read values, dividing by 3 of the number of reads (averaging operation), and taking the average value as the value of the event indicated by the rightmost arrow in 35 is there. This averaging operation may reduce the length of each arrow to 1/3 in advance, and multiply each read value by the length of the arrow. At this time, when the arrow is upward, the magnification is positive, and when the arrow is downward, the magnification is negative. Also, three events corresponding to the number of readings are displayed with a width 35a of 35 (= 3 events). Therefore, the element time series 13a, 16a, 19a, 22a, and 25a in FIG. 5 are obtained by moving average of the unsmoothed element time series, 12a, 15a, 18a, 21a, and 24a, using the width 35a of the rectangular wave 35 as 20 events. It is an element time series smoothed.

  If these moving rectangular waves are arranged vertically in two as in the second stage 36, two upwards and one in the lower side (37a) as in the third stage 37, these rectangular waves 36 And 37, a continuous reading of the time series 34 is obtained, and the primary and secondary difference values of the time series 34 can be found. At this time, the arbitrary interval n = k shown by the interval 36a is an example of five events. The primary difference and the secondary difference are given by the following equations of Equations 1 and 2, respectively.

(Equation 1)

[(Dc15 + Dc14 + Dc13)-(Dc10 + Dc9 + Dc8)] / 3
(Equation 2)

[(Dc15 + Dc14 + Dc13) -2. (Dc10 + Dc9 + Dc8) + (Dc5 + Dc4 + Dc3)] / 3

  Therefore, the length of the downward arrow of the rectangular wave 37a is twice that of the other arrows and is 2/3. Such a reading operation of extracting only the fluctuation component of the time series 34 that matches the rectangular waveform is an operation for obtaining the cross-correlation between the element time series Dc and the rectangular waveform. At this time, the addition and the difference operation of the read values of Dc performed to obtain the cross-correlation are to apply low-pass and high-pass filter operations to Dc, respectively. Therefore, since the shape of the square wave determines these filter functions, the function is given by a Fourier integral transform equation (spectral) regarding the number m of events of all the square waveforms. The conversion formula is omitted, and only the calculation result is displayed in FIG.

In FIG. 8, the vertical axis indicates the relative amplitude value of the spectrum of the rectangular waveform, and the horizontal axis indicates the frequency (the unit is m ^−1, which is the reciprocal of the number m of events). Are displayed up to the area where is 0.05. The bandpass filter function of the rectangular waveform 37 when the number of moving averages is 20 events and n = k = 30 is graphically displayed on 38. In the case of 38, the functions of the band-pass filter in which the cutoff frequencies of the high-pass and low-pass filters are given by (4 × 30) ⁻¹ and (2 × 20) ⁻¹ are illustrated. These cutoff frequencies are defined where the highest intensity of 38 is halved. Also, its maximum strength is around (2 × 30) ⁻¹ . Further, in FIG. 6 described above, the range of the variable frequency component of the time series Dc passing through this band-pass filter is illustrated by a horizontal arrow 33 on the spectrum of each element time series Dc. For comparison, a band-pass filter function in a case where the moving average number 20 and the arbitrary interval n = k = 20 for taking the difference are all equal in the rectangular wave configuration example 37 is shown in FIG. From this, 38 indicates that the bandpass filter function of 38a is much better. The low pass filter function when the moving average number is 20 with 35 rectangular waveforms is also shown in FIG. Next, in observing the seismic phenomenon on a scale in which the periodic fluctuation component is a periodic vibration of approximately 60 events, the importance of selecting 60 events among the fluctuation periods substantially common among the seismic elements c is shown in FIG. It will be described based on.

  In FIG. 9, first, regarding the scale of the figure, the horizontal axis represents the number of events m. The vertical axis on the left is the number of days, which scales the graphs 40 and 41, with a maximum value of 800 days. 40 and 41 indicate the CIm (a) of the cumulative INT obtained by the time interval cumulative detection step of 7, and 40 (60) and 41 (20) INTs are continuously accumulated (added) by a number. In the time series, 40 is CIm (60) and 41 is CIm (20). If they are divided by the accumulated number 60 or 20, the moving average of the time series Dc of INT is obtained. Therefore, the increasing direction of 41 is the same as that of 22a in FIG. Forty accumulated events of 60 corresponding to the period of the periodic fluctuation represent the number of days spent generating 60 consecutive earthquakes, and 41 represents the number of days spent generating 20 consecutive earthquakes. Represents. Therefore, an increase in the number of days means that seismic activity with a MAG of 3.5 or more has calmed down. At this time, the calm of the seismic activity was observed over a wide area such as the Chugoku region selected in the example.

  Therefore, since it is not known from the detection of the sign of quietness in FIG. 9 where a large earthquake will occur, the prediction of the epicenter will be described with reference to FIG. FIG. 9 shows the same elemental time series of MAG and LON used for the display of 40 and 41 on the horizontal axis in order to clarify the situation in which the calm described below is a sign of an earthquake. Displayed using the number of events. The vertical axis on the right is a scale for the time series 24a in which the earthquake element is MAG, and only MAGs with 5 or more are displayed. The time series of LON is indicated by 15a, and its scale is a relative display using the vertical axis on the right. The element time series of LON and MAG is exactly the same as 15a and 24a in FIG. 5, but the increasing direction of their values is opposite to that in FIG. 5, and the upward direction is the increasing direction. The two swarm earthquakes 10a and 10g are shown on the LON time series, and for earthquakes with a MAG of 6 or more, 24a is labeled from 10b to 10h. In FIG. 40, eight labels 40a to 40h are attached to locations where signs of an earthquake that occurred in the Chugoku region in the selected area described with reference to FIG. 10 are detected.

  At CIm (60) 40, a clear predictive tranquility appears before the Great Hanshin-Awaji Earthquake 10d when the number of events m is 783, and there are three stages of tranquility. Is shown. The three stages appear between events m = 625 to 780. In the first phase, the CIm (60) starts with a steady rise from 300 days at m = 625 to 600 days at m = 700. In this way, the tranquility gradually becomes noticeable several years before the upcoming earthquake.

  Such a first-stage tranquility is universally observed as the tranquility of the seismicity before the earthquake. In the second stage, a sudden increase of 50 to 100 days is added to the first stage. In the third stage, just before the Great Hanshin-Awaji Earthquake, it will drop sharply for more than 100 days immediately after the second stage. In the third stage, the calm ends several days to several hours before the earthquake. As for the situation where this third stage of calm is over, there are reports that past MAGs in the Kanto region have been observed by more than 5 earthquakes. The peak 40d of the second CIm (60) in the second stage is a final sign of the Great Hanshin-Awaji Earthquake, and seems to indicate a critical state of stress accumulation. The peak also coincides with the CIm (20) peak just below 40d.

  Such a sudden tranquility in the second stage, consisting of 40c, 40e and 40f, occurred prior to the main earthquake or approximately 2-3 months before the earthquake. This situation is reflected in the sign of the earthquake in other regions. Also, when implementing actual earthquake prediction, the magnitude of the MAG of the earthquake that is predicted to occur in the selected area (for example, an earthquake with a MAG of 7 or more as in the case of the Great Hanshin-Awaji Earthquake) Parameters representing the profile and magnitude of the CIm (60) waveform based on past large earthquakes (eg, the maximum value at which the calm peaked, the magnitude of the waveform width, or the area surrounded by the waveform) Can be used to estimate. Therefore, when the quietness of the seismic activity smoothed (accumulated) at the cycle of about 60 events is observed by CIm (60), the predictive quietness of the earthquake, the magnitude of the magnitude, and the process of the stress strain reaching the critical state are clearly extracted. it can.

  Immediately after the end of the two earthquake swarms 10a and 10g that occurred in the shallow Eurasian plate (Hyogo prefecture in the northeast direction of Fig. 2), 70 days and 69 days later, a relatively large number of the submerged Philippine plates were settled. Regarding the Hyuga-Nada Earthquake 10b and the Aki-Nada Earthquake 10h, which occurred in deep places, from the profile of CIm (k), these swarms can be regarded as signs of an earthquake that occurred about 70 days later. The signs are indicated by 40a and 40h. When the shape of these small mountains is collected and the CIm (10) is shown by gathering earthquakes of magnitude 4 or more, the same sign of calm that rises sharply for more than 50 days in a step like a square wave immediately after the swarm earthquake is observed, These are signs of calm. Furthermore, regarding the fact that a swarm earthquake can be said to be a sign, the change in the crustal movement obtained from the GEONET of the Geographical Survey Institute at the GPS site near the epicenter of the earthquake 10g and 10h can be used as the motion change detection device of Japanese Patent No. 2787143. Time series analysis using, and the result is drawn as a trajectory of crustal deformation on the phase plane, revealing that the two earthquakes are related to each other and that the swarm earthquake triggered the Aki Nada earthquake 69 days later are doing.

  In FIG. 10, the amplitude and phase of 14a and 17a, 17a and 20a, 20a and 14a, 23a and 20a, 23a and 20a, and 26a and 20a from the top are compared with the elementary time-series periodic variation of the secondary difference extracted by 5 using 8 Is displayed. The arrangement of each row is the same as the LAT, LON, DEP, INT, and MAG rows in FIG. However, the display of 14a, 17a, 20a, 23a and 26a is standardized so that the maximum absolute value of each amplitude is 1. On the vertical axis of each row, the maximum is +1 and the minimum is -1 on the scale of the relative amplitude value. The plus and minus symbols in each row indicate the positive and negative amplitude regions, respectively. The number of events is displayed on each horizontal axis. The long arrow marked Great Hanshin-Awaji Earthquake indicates the event location (m = 783) where the earthquake occurred. Further, the quiet signs of the earthquake detected in the amplitude phase comparison detection step 8 are described as 40b, 40c, 40d, 40e, 40f, and 40g in the fourth INT line from the top. These signs of calm indicate that the seismic element reverses the phase between DEP 20a and INT 23a, and at the same time, MAG 26a exhibits a negative amplitude state indicating that seismic activity with a magnitude of MAG = 3.5 increases. I have. After such a sign of calm, a large earthquake always occurs as shown in FIG. 9A. However, the earthquake after the sign 40f is an earthquake with a MAG of 4.9. Next, for the different signs including the signs and quakes of the seismic phenomena found from the amplitude and phase relationships of 14a and 17a, 17a and 20a, and 20a and 14a, the factors of the occurrence of the signs will be outlined.

The Chugoku region selected in this example is located on the border between the Eurasian plate and the Philippine plate. According to the Geospatial Information Authority of Japan's GPS network (GEONET), the sedimentation plate of the Philippine sedimentation plate at this boundary is fixed at the epicenter of the earthquake 10e in Fig. 2 at a speed of about 3 cm / year from southeast to northwest. Is relatively moving toward. The relative motion of the plates creates an earthquake-prone zone near the transition zone (DB layer) from the soft to hard layer of the crust at the plate boundary, as shown in FIGS. As observed in this seismic zone, something that produces a long period of vibration (approximately 60 events) 14a, 17a, 20a, 23a and 26a that is substantially common to all seismic elements exists in the seismic zone Must be doing. If the transition zone from the hard layer to the soft layer in the crust can be considered as an elastic body, it is considered that the earthquakes occurring in those cycles repeat the increase and decrease in the stress accumulation caused by the increase and decrease in the strain of the crust .
Therefore, if a large earthquake is likely to occur, it is considered that something abnormal (predictive sign) appears in those repeated vibrations before the earthquake.

  First, the vibrations 14a, 17a, and 20a of the LAT, LON, and DEP of the seismic elements are associated with the traces of the earthquake in the geographical space shown in FIGS. 2, 3, and 4, and the appearance of a sign is shown. Assume that the zero reference of the vibration of the seismic elements LAT, LON and DEP is fixed at (34 degrees N, 133,75 degrees E, 30 km). Most of 14a and 17a are approximately in phase, and they are both approximately 20a out of phase. Therefore, the traces of many earthquakes that occurred in the Chugoku region oscillate along the diagonal line from southwest to northeast in FIG. 2 and along the linear trend line 11 in FIGS. Therefore, if the amplitude values of 14a and 17a are both positive and 20a is negative, the earthquake trace is in the shallow northeast region. However, two exceptions to these oscillations are observed in connection with earthquakes 10e and 10f. In both cases, the trace of the earthquake changed from southeast to northwest. The first is the opposite phase of 14a and 17a, the same phase of 17a and 20a, and the change of 20a and 14a to the same phase. These are predictive changes because they occurred almost one year before the earthquake 10e. The second is that the change of 14a and 17a to the opposite phase occurred after the occurrence of the earthquake 10f, but the changes of 17a and 20a and 20a and 14a to the same phase, which are the same as those of the earthquake 10f. Since it occurred months ago, it can be regarded as a predictive change.

  Next, it is shown that predictive calm is an abnormality that appears in such repeated vibration. The 23a and 20a vibrations have substantially the same phase except for the sign of quiet. This in-phase vibration means that as the amplitude value of 20a increases (DEP increases), 23a also increases (INT increases). Therefore, these in-phase oscillations indicate normal seismic activity in which the INT of deep-source earthquakes tends to be longer and the INT of shallow-source earthquakes tends to be shorter. However, when an action that causes a large earthquake is applied to the zone where the earthquake occurs, the phases are reversed and predictive tranquilities, 40b, 40c, 40d, 40e, 40f, and 40g, are observed.

  Furthermore, the vibrations 26a and 20a are also in substantially the same phase except for the section where the number of events several years before the aftershock and the Great Hanshin-Awaji Earthquake was 500 to 700. This in-phase oscillation reflects the normal seismic activity in which MAG increases with increasing DEP. However, as shown in the enlarged view of FIG. 11, the opposite phase in which the number of events started around 512 (December 1989) and ended around 719 (August 1993) indicates that when DEP becomes shallower than 30 km, MAG becomes Becomes larger than 3.9. From the vibration information at 14a and 17a, the vibrations of the opposite phases show that a large stress gradually accumulates in the transition region (DB layer) from the soft to hard layer of the crust in the northeast region, and in the process of reaching the critical state Suggests something. Therefore, those vibrations are a sign of the Great Hanshin-Awaji Earthquake. Since the frequency of the continuous antiphase is considered to be proportional to the amount of stress accumulation, the frequency can be used as an index for estimating MAG.

  In Fig. 11, we pick up the Great Hanshin-Awaji Earthquake that occurred in a wide area of the selected Chugoku region, observe the sign of calm, and predict how the time, epicenter, and size of MAG of the Great Hanshin-Awaji Earthquake would be predicted. explain. FIG. 11 is a diagram in which the number of events, m, in the range of 480 to 800 (1989/06/26 to 1995/01/17) is extracted and superimposed in FIGS. 5 and 10. However, on the lowermost MAG line, vibration 23a is superimposed in addition to vibrations 26a and 20a in order to clearly indicate the sign of calm. First, a description will be given of a method of detecting a sign of calm and then predicting the occurrence time of an earthquake by the number of events.

  From the earthquake phenomena in the Chugoku region during this period, the amplitude and phase comparison and detection means 8 detects the sign of calm as follows. 23a displayed in the INT line or MAG line is near the negative peak (m = 700, March 29, 1993), and an effect that may be the cause of calm works in the earthquake-prone area and starts delaying INT. I have. 23a slowly returns from its negative peak to 23b (zero) at the reference point, and the phase of the periodic vibration of 23a and 20a, which had been in the same phase until then, changes. This is the calm sign that appeared in seismic activity in the Chugoku region about two years after the Great Hanshin-Awaji Earthquake that occurred when the number of events (m) was 783, and the place of appearance was CIm (60), the second of 40 Corresponds to the beginning of the first steep sharp peak (m = 710) during the calm of the stage.

  The phase delay becomes apparent around October 1993 (m = 725 to 730). Further, the phase inversion of 23a and 20a (the valley of 23a and the peak of 20a are at m = 745 to 749) are observed from June 06, 1994 to September 3, 1994 (m = 743 to 757). is there. Furthermore, the immediately following m = 759 (October 14, 1994) corresponds to the second sharp increase 40d of the second stage shown in the profile of 40. At the time of this phase inversion, the vibration 26a of the element MAG slightly changes but has a substantially negative amplitude value, and has substantially the same phase as the DEP 20a. In step 8, these amplitude and phase relationships were detected, and the sign of the sign of the Great Hanshin-Awaji Earthquake that was observed was 23a, a positive amplitude value indicating that the INT of the earthquake was long, and the DEP of the earthquake was shallow. And a negative amplitude value 26a indicating that the seismic activity of the earthquake having a magnitude of approximately MAG = 3.5 becomes active while the seismic activity is in a calm state. In addition, the predictive calmness reliably extracts the predictive calmness of the second stage of the 40 profiles in FIG. The changes in the predictive calmness of the Great Hanshin-Awaji Earthquake extracted in this way agree very well with the changes in the predictive changes in the chemical content (chloride content) of groundwater that appeared before the Great Hanshin-Awaji Earthquake. I do. Immediately after the Great Hanshin-Awaji Earthquake, according to a paper report analyzing the chloride content of mineral water in stock PET bottles in order of production date, the chloride content in PET bottles remained unchanged until September 1993. It was stable. However, oscillations with a period of about three months began in November 1993. (In step 8, the first obvious sign of a predictive phase change was observed in October 1993.) At the beginning of August 1994, the oscillation changed its period and reached a high five months earlier. Reached the level. Then, the level reached a new peak and still oscillated. A steady increase in chloride content was finally established in October or November 1994.

  After the predictive calm is extracted, the predicted occurrence time of the Great Hanshin-Awaji Earthquake is the time when the predictive calm ends at the third stage of the CIm (60) profile 40, which is a negative time at which the vibration 23a reverses after approximately 30 events. This is near the peak of the vibration 23a. The peak of the negative vibration 23a indicates that INT has become shorter. Accordingly, since the peak of the positive vibration 23a at which the predictive calm was established is at m = 745, 30 events can be added, and the occurrence time (m = 788) of the Great Hanshin-Awaji Earthquake can be predicted to be near the number of events m = 775. . Actually, the oscillation cycle of 23a is a quiet event, and the half cycle read from the peak of the positive oscillation 23a observed at m = 745 and the immediately preceding negative peak is 43 events long, so at the latest, , M = 788, it can be predicted that the Great Hanshin-Awaji Earthquake will occur. The time when such a prediction can be made is at the time when this predictive calm can be confirmed, that is, at m = 756 (Aug. 22, 1994), which is slightly after the detection of the calm peak, approximately five months before the occurrence of the earthquake. . The prediction of the earthquake occurrence time using the number of events is accurate because the vibration cycle of 23a performs the time-based function of the clock.

  Next, a method of predicting the epicenter will be described. In FIG. 11, 14a and 17a are almost in phase in the range of the number of events m = 480-800, and 17a and 20a, and both 20a and 14a are in opposite phase. , Approximately in FIG. 2, in the direction from southwest to northeast, and between the two continental plates, along the linear line 11 in FIGS. 3 and 4 (34 ° N, 133.75 ° E). , 30 km). However, during the predictive calm period (m = 725-760), the signs of 14a and 17a remain positive, and 20a remains substantially negative. This means that the epicenter of most earthquakes is from the geographic origin to a shallow northeastern region (Kobe region).

  In addition, the origin of the vibration is slowly moving in the Kobe direction along with the linear components of the seismic element time-series variations 13a, 16a, and 19a obtained by moving average of the 20 pieces in Step 3. From m = 775 to around 788 of the number of events for which the occurrence of the Great Hanshin-Awaji Earthquake is predicted, the periodic signs of the secondary differences 14a, 17a, and 20a corresponding to the acceleration do not change in the vibration sign. Assuming that the movement tendency of 13a, 16a, 19a does not change and keeps the linearity, and extending each arrow 42a, 42b, 42c in FIG. 11 to the intermediate value m = 782 of the predicted earthquake occurrence time, The LAT, LON, and DEP readings indicated by the arrows are intermediate values of the predicted hypocenter. The readings are 34.5 degrees north latitude, 135 degrees east longitude, and the depth of the epicenter is 16 km, which is very close to the actual epicenter of the Great Hanshin-Awaji Earthquake, 34.70 degrees north, 135.08 degrees east longitude, and 16.08 km deep. In FIG. 11, the error between the actual epicenter and the predicted value is indicated by the fact that 12a, 15a, and 18a are changed from the earthquake just before the Great Hanshin-Awaji Earthquake to a thick line, so that the end of the thick line And the tip of each of the arrows 42a, 42b, 42c. Therefore, it can be seen that there is almost no error between the predicted epicenter and the actual epicenter.

  This earthquake prediction method was extracted regardless of the magnitude of the earthquake, regardless of the size of the CIm (k) profile and the number of phase-reversal oscillations observed between 20a and 26a several years before the Great Hanshin-Awaji Earthquake. The following summarizes the results of displaying the hypocenters and times of the prediction example applied to the two signs of calm (from 40b to 40g) in the format of (LAT, LON, DEP, m). Although the prediction of the hypocenter of the earthquake after 40f and 40g is slightly bad, it is generally good. As for the occurrence time, everything is good. The reason why the LAT prediction is 1.6 degrees south and the DEP is 30 km in the source prediction of the earthquake 7.3 magnitude 10f after 40g after the sign of calm is related to the occurrence of a water-related low frequency earthquake. It has been reported in a paper. The epicenter distribution of such a low-frequency earthquake is at a depth of about 30 km south of the latitude (LAT), so the epicenter prediction is not necessarily bad.

  For a magnitude 6.0 earthquake of 10/13/1985/13/13 (33.00 degrees, 132.59 degrees, 38.08 km, 230), the predictable epicenter and time were (33.6 degrees, 133.6 degrees, 34 km) at a sign of quiet 40b (extracted to m = 196). 226-229).

  For a magnitude 5.1 earthquake of 1988/07/29 earthquake (33.68 degrees, 132.51 degrees, 53.01 km, 428), the predictable hypocenter and time were (33.7 degrees, 132.7 degrees, 46.4 km) at a sign of quiet 40c (extracted to m = 396). , 423-426).

  The magnitude and magnitude of 10d (34.70 degrees, 135.08 degrees, 16.06 km, 783) of the 1917 Great Hanshin-Awaji Earthquake on January 17, 1995, and the predicted hypocenter and time were (34.5 degrees, 135 Degrees, 16km, 775-788).

  For a magnitude 6.3 earthquake of 10/06/1997/06/25 (34.48 degrees, 131.69 degrees, 8.29 km, 1089), a predictive quake source and time are (34 degrees, 132.5 degrees, 30 km) at a sign of quiet 40c (extracted to m = 1066). , 1088-1096).

  For an earthquake of magnitude 4.9 on March 16, 1999 (35.31 degrees, 135.96 degrees, 12.08 km, 1227), the predicted epicenter and time were (34.3 degrees, 133 degrees, 35.5 km) at a sign of quiet 40c (extracted to m = 1204). , 1226-1234).

  For 10f (35.35 degrees, 133.50 degrees, 11.26 km, 1350) of magnitude 10 earthquakes (35.35 degrees, 133.50 degrees, 11.26 km, 1350) with a magnitude of 7.3, the predicted hypocenter and time are (33.7 degrees, 133.5 degrees) , 30km, 1350-1354).

1 is a block diagram illustrating a method and apparatus for predicting an earthquake according to the present invention. It is a diagram showing the epicenter distribution where the magnitude generated in the Chugoku region is 3.0 or more and the epicenter depth is 300 km or less (shallow). It is a figure which shows the latitude-depth distribution of the epicenter of the Chugoku region. It is a figure which shows the longitude-depth distribution of the epicenter of the Chugoku region. It is a figure which shows the time series of each earthquake element obtained by each time series conversion step from the source element information of an earthquake. It is a figure which shows the spectrum of the time series of each earthquake element. It is a figure showing the example of composition of the band pass filter equivalent to the function of the second order differential output step. FIG. 3 is a diagram illustrating a function of a bandpass filter. It is a figure which shows a three-stage sign calm and a sign. It is a figure showing sign calm and a sign expressed by relative amplitude and phase inversion. It is a figure which shows the sign calm of the Great Hanshin-Awaji Earthquake, the occurrence time, the prediction of the epicenter, and the estimation of the magnitude.

Explanation of reference numerals

DESCRIPTION OF SYMBOLS 1 Earthquake source and focal mechanism information source 2 Seismic element signal generation means 3 Seismic element time series means 4 Primary differential output means 5 Secondary differential output means 6 Filter output means 7 Time interval accumulation detection means 8 Amplitude phase comparison detection means 9 Foresight means

Claims (16)

A predetermined earthquake is extracted from the earthquakes that occurred in the target area, and from the source information of the earthquakes and the information of the focal mechanism solution, the parameter representing each element of the earthquake information is set as c, and the seismic phenomenon in the selected area is determined. A seismic element signalizing step of quantifying earthquake information on each element c of the extracted j-th earthquake as Dcj in order to represent and observe the time series data of each element c;
The Dcj extracted in the seismic element signalizing step is directly or predetermined selected or smoothed based on the earthquake occurrence order j (j = 1, 2, 3,..., M,...). Performing a process to obtain a sequence of seismic elements to obtain Dc = (Dc1, Dc2, Dc3,..., Dcm,...);
An amount proportional to the difference between the arbitrary intervals n (n = 1, 2, 3,...) And the difference between Dcm and Dcm-n in the arbitrary earthquake occurrence order m of Dc obtained in the time series step of the earthquake element. As a primary difference value ΔDcm, n,
The amount of the time series value ΔDcm, n output from the primary differential output step at an arbitrary interval k (k = 1, 2,...), That is, an amount proportional to the difference between ΔDcm, n and ΔDcm−k, n As a secondary difference value Δ ² Dcm, n, k in an arbitrary earthquake occurrence order m;
The amplitude and phase of the secondary difference value Δ ² Dcm, n, k obtained in the secondary differential output step are compared and detected between different elements, and the amplitude relation and phase relation between each element and the time series of earthquake elements Amplitude phase comparison detection step of outputting
Earthquake prediction method equipped with. The method for predicting an earthquake according to claim 1,
The step of predicting the time of occurrence of the earthquake, the epicenter and the magnitude of the earthquake, from the amplitude relationship and the phase relationship between the elements and the time series of the earthquake element, detected by the amplitude phase comparison detection step between the elements of the earthquake. Prepared earthquake prediction method. In the method for predicting an earthquake according to claim 1 or 2, the number of occurrences of an arbitrary earthquake is set to a (a = 1, 2,...) And extracted by the earthquake element signalizing step and the time series forming step. A method for predicting an earthquake, comprising a time interval accumulation detecting step of accumulating a number of time intervals apart from each other by one to an arbitrary order m of occurrence of an earthquake to obtain CIm (a). 4. The method for predicting an earthquake according to claim 1, wherein an arbitrary number of earthquakes is a (a = 1, 2,...), And an m-th and m-a-th earthquakes are set. A method for predicting an earthquake, comprising a time interval cumulative detection step in which the difference between occurrence times is CIm (a). The method for predicting an earthquake according to any one of claims 1 to 4,
The time series of the seismic elements obtained in the seismic element time series forming step is filtered by a band-pass filter for each element c of the earthquake, Fcm is obtained in an arbitrary earthquake occurrence order m, and output to the amplitude phase comparison detecting step. An earthquake prediction device comprising a filter output means. A predetermined earthquake is extracted from the earthquakes that occurred in the target area, and from the source information of the earthquakes and the information of the focal mechanism solution, the parameter representing each element of the earthquake information is set as c, and the seismic phenomenon in the selected area is determined. A seismic element signal converting means for quantifying the seismic information on each element c of the extracted j-th earthquake as Dcj in order to represent and observe the time series data of each element c;
The Dcj extracted by the seismic element signal generator is directly or predetermined selected or smoothed based on the order of occurrence j of the earthquakes (j = 1, 2, 3,..., M,...). Means for performing a process and obtaining a time series of seismic elements to obtain Dc = (Dc1, Dc2, Dc3,..., Dcm,.
An amount proportional to the difference between arbitrary intervals n (n = 1, 2, 3,...) In the arbitrary earthquake occurrence order m of Dc obtained by the time series means of the earthquake element, and the difference between Dcm and Dcm-n As a primary difference value ΔDcm, n,
The amount of the time series value ΔDcm, n output from the primary differential output means at an arbitrary interval k (k = 1, 2,...), That is, an amount proportional to the difference between ΔDcm, n and ΔDcm−k, n As a secondary difference value Δ ² Dcm, n, k in an arbitrary earthquake occurrence order m;
The amplitude and phase of the secondary difference value Δ ² Dcm, n, k obtained in the secondary differential output step are compared and detected between different elements, and the amplitude relation and phase relation between each element and the time series of earthquake elements An earthquake prediction device comprising: an amplitude / phase comparison / detection unit that outputs a signal. The method of predicting an earthquake according to claim 6, wherein the time of occurrence of the earthquake and the time series of the earthquake elements are detected from the amplitude relation and the phase relation between the elements and the time series of the earthquake elements detected by the amplitude and phase comparison detection means between the elements of the earthquake. An earthquake prediction device comprising means for predicting the epicenter and magnitude of an earthquake.   8. The earthquake prediction device according to claim 6, wherein an arbitrary number of occurrences of the earthquake is a (a = 1, 2,...), And the earthquake extracted by the earthquake element signal generation means and the time series generation means An earthquake prediction device that accumulates a number of time intervals that are separated by one from each other up to an arbitrary earthquake occurrence order m and sets it as CIm (a). The earthquake prediction device according to any one of claims 6 to 8, wherein the number of occurrences of an arbitrary earthquake is a (a = 1, 2, ...), and the m-th and m-a-th earthquake occurrence times are set. An earthquake prediction device comprising time interval accumulation detecting means for setting the difference between the two as CIm (a). The earthquake prediction device according to any one of claims 6 to 9,
2. The amplitude-phase comparison according to claim 1, wherein the time series of the seismic elements obtained by the means for serializing the seismic elements is filtered by a band-pass filter for each element c of the earthquake to obtain Fcm in an arbitrary order m of occurrence of the earthquake. An earthquake prediction device comprising a filter output means for outputting to a detection means. A predetermined earthquake is extracted from the earthquakes that occurred in the target area, and from the source information of the earthquakes and the information of the focal mechanism solution, the parameter representing each element of the earthquake information is set as c, and the seismic phenomenon in the selected area is determined. A seismic element signalizing step of quantifying earthquake information on each element c of the extracted j-th earthquake as Dcj in order to represent and observe the time series data of each element c;
The Dcj extracted in the seismic element signalizing step is subjected to direct or predetermined selection processing or smoothing processing based on an earthquake occurrence order j (j = 1, 2, 3,..., M,...). Applying a seismic element time series to obtain Dc = (Dc1, Dc2, Dc3,..., Dcm,.
An amount proportional to the difference between the arbitrary intervals n (n = 1, 2, 3,...) And the difference between Dcm and Dcm-n in the arbitrary earthquake occurrence order m of Dc obtained in the time series step of the earthquake element. As a primary difference value ΔDcm, n,
The amount of the time series value ΔDcm, n output from the primary differential output step at an arbitrary interval k (k = 1, 2,...), That is, an amount proportional to the difference between ΔDcm, n and ΔDcm−k, n As a secondary difference value Δ ² Dcm, n, k in an arbitrary earthquake occurrence order m;
The amplitude and phase of the secondary difference value Δ ² Dcm, n, k obtained in the secondary differential output step are compared and detected between different elements, and the amplitude relation and phase relation between each element and the time series of earthquake elements And an amplitude / phase comparison / detection step of outputting a signal. In the earthquake prediction program according to claim 11,
The step of predicting the time of occurrence of the earthquake, the epicenter and the magnitude of the earthquake, from the amplitude relationship and the phase relationship between the elements and the time series of the earthquake element, detected by the amplitude phase comparison detection step between the elements of the earthquake. Prepared earthquake prediction program. In the earthquake prediction program according to claim 11 or 12, the number of occurrences of an arbitrary earthquake is set to a (a = 1, 2,...) And extracted by the earthquake element signalizing step and the time series forming step. An earthquake prediction program comprising: a time interval accumulation detection step of accumulating a number of time intervals separated by one from each other to an arbitrary order m of occurrence of an earthquake to obtain CIm (a). The earthquake prediction program according to any one of claims 11 to 13, wherein an arbitrary number of earthquakes is a (a = 1, 2, ...), and an m-th earthquake and an (ma) -th earthquake. An earthquake prediction program comprising: a time interval accumulation detection step in which a difference between occurrence times is set to CIm (a). In the earthquake prediction program according to claim 11 or claim 14,
The time series of the seismic elements obtained in the seismic element time series forming step is filtered by a band-pass filter for each element c of the earthquake, Fcm is obtained in an arbitrary earthquake occurrence order m, and output to the amplitude phase comparison detecting step. Earthquake prediction program with filter output means. A recording medium on which the earthquake prediction program according to claim 11 is recorded.

Images