JP6450129B2 - Slope failure prediction method and slope failure prediction device - Google Patents

Description

  The present invention relates to a slope failure prediction method and a slope failure prediction device for predicting the occurrence of slope failure such as landslides and landslides caused by rainfall.

  The occurrence of landslides such as landslides and landslides causes serious damage to human lives and property. For this reason, the establishment of an accurate prediction method for slope failure is important for taking measures to avoid slope failure and judging the timing of warning and evacuation.

  Slope failure tends to occur when the rainfall time is long or when the rainfall intensity is high. However, even if the rainfall time is short, if the rainfall intensity is very high, or if the rainfall intensity is low, the rainfall time is long, and if the total amount of rainfall up to that time is large, slope failure may occur. Thus, the occurrence of slope failure is affected by both the total rainfall and rainfall intensity. Therefore, it predicts the occurrence of slope failure by using the rainfall index (long-term rainfall index) for evaluating the influence of total rainfall and the rainfall index (short-term rainfall index) for evaluating the influence of rainfall intensity. It is effective.

  Therefore, a prediction method using a graph called a snake curve that represents a rainfall situation using a long-term rainfall index and a short-term rainfall index is generally known. The snake curve plots the points representing the long-term rainfall index and the short-term rainfall index at each time on a graph with the long-term rainfall index and the short-term rainfall index respectively on the horizontal axis and the vertical axis, and connects them with a line. Therefore, the occurrence of slope failure is predicted by tracking the locus of points according to the time series of rainfall.

Figure 1 shows a snake curve using an effective rainfall with a half-life of 72 hours (72h effective rainfall) as a long-term rainfall index and an effective rainfall with a half-life of 1.5 hours (1.5h effective rainfall) as a short-term rainfall index. An example is shown. The effective rainfall is a rainfall index that expresses the sustainability of the influence of preceding rainfall at a certain point on the accumulated rainfall at that point thereafter, with the half-life as a parameter. If the half-life is M [h], the effective rainfall amount X (M, t) [mm] at time t [h] is defined by the following equation (1).

Here, X (M, t-1) [mm] is an effective rainfall amount one hour ago, and R (t) [mm] is a rainfall amount (time rainfall amount) between times t-1 and t. The reduction coefficient α [h ⁻¹ ] is obtained by the following equation (2) using the half-life M.

In equation the effective rainfall (1), and so on rainfall R that falls in a certain time half-life M when [h] elapses R / 2, and becomes the R / 4 further half-life M [h] has passed, The impact of rainfall decreases exponentially. Therefore, when the half-life M is short, the influence of rainfall decreases rapidly with time, and the effective rainfall is an index that substantially represents the rainfall intensity. On the other hand, if the half-life is long, the influence of rainfall will last for a longer period. If the half-life is set to infinity, the effective rainfall will be equal to the integrated rainfall. FIG. 2 shows a graph (a) showing the hourly rainfall in each month of a certain year, and when the half-life M [h] is set to 1.5, 72, 216, 720, and 3000 for this hourly rainfall. Graphs (b) to (f) representing effective rainfall are shown. In FIG. 2, for example, the effective rainfall with a half-life of 1.5 [h] represents an index of the amount of water stored in the surface layer of the ground, and the effective rainfall with a half-life of 72 [h] is the water stored in the deep layer from the ground. Represents a quantity indicator.

  A boundary line CL on the snake curve indicates a rainfall line at the collapse occurrence limit. In the snake curve, it is determined that slope failure occurs when a point is plotted on the upper right side of the boundary line CL. Therefore, in the prediction of slope failure using a snake curve, it is important to select an appropriate rainfall index (half-life) for the vertical and horizontal axes and to set an appropriate boundary line CL.

Examples of the long-term rainfall index and the short-term rainfall index used for the horizontal and vertical axes of the snake curve are as follows.
The first example is an example that has been used for many years for prediction of slope failure, using accumulated rainfall as a long-term rainfall index and hourly rainfall as a short-term rainfall index (Non-Patent Document 1). .
The second example is an example in which the soil rainfall index obtained by evaluating the rainwater storage amount inside the slope with a tank model as shown in FIG. 3 is used as a long-term rainfall index (Non-patent Documents 2 and 3). Proposed by Makihara and Hirasawa in 1993, it has been widely used in recent years.
The third example is an example using the effective rainfall by adjusting the influence of the total amount and intensity of rainfall by the half-life as an index, and many studies have been made on various effective rainfalls with different half-lives. In particular, the effective rainfall with a half-life of 72 hours proposed by the General Sediment Disaster Countermeasures Study Group (Ministry of Construction) held in 1993 was used as the long-term rainfall index, and the effective rainfall with a half-life of 1.5 hours was used as the short-term rainfall. A method of using an index is widespread (Non-Patent Document 4).

  On the other hand, with respect to the boundary line, as shown in FIG. 4, for example, a straight line that can reduce both “missing” and “missing” as much as possible based on the past slope failure occurrence record and slope failure non-occurrence record in the target area. Has been used for a long time (Non-patent Document 1). In recent years, a method of drawing a non-linear boundary line CL that minimizes “missing” and “missing” by a method using a neural network has been proposed and widely used (Non-Patent Document 4). Yes.

Here, “missing” means that a case where slope failure has actually occurred is judged as non-occurrence, and “swinging” means that a case where slope failure has not actually occurred has occurred. It means to judge. When “missing” occurs, the slope collapse occurs in a situation where no warning or evacuation recommendations or instructions are given, and this increases human and property damage. On the other hand, if “swinging” is repeated, the residents in the target area will gradually become less aware of the slope failure, so even if a slope failure actually occurs, be wary of receiving recommendations and instructions, There is a risk that it will not take.
Therefore, in predicting slope failure, it is important how much “missing” and “missing” can be reduced.

Hideki Terada and Hiroaki Nakatani (2001): Sediment Disaster Warning and Evacuation Standard Rainfall Setting Method, National Institute for Land and Infrastructure Management, Material No.5, Ministry of Land, Infrastructure, Transport and Tourism, 58pp. Makihara Yasutaka and Hirasawa Masanobu (1993): Accuracy of tank model in slope failure risk prediction, JMA study time report, Vol.45, No.2, p.35-70 Ministry of Land, Infrastructure, Transport and Tourism Meteorological Agency homepage, "Soil Rainfall Index", [online], searched May 2, 2014], Internet <URL: http://www.jma.go.jp/jma/kishou/know/bosai/ dojoshisu.html> Kazumasa Kuramoto, Hiromi Tetsuga, Hirokazu Tetsuga, Masao Arakawa, Hirotaka Nakayama, Kohei Furukawa (2001): Research on setting of non-linear landslide occurrence limit rainfall line using RBF network, Proceedings of Japan Society of Civil Engineers, No.672 / VI-50 , P.117-132 Kenichiro Kosugi, Masamitsu Fujimoto, Yosuke Yamakawa, Naoya Masaoka, Satoshi Itobe, Takahisa Mizuyama, Atsuhiko Kinoshita (2013): Functional model based on effective rainfall to analyze groundwater level fluctuations in mountain base rocks, Journal of the Sabo Society of Japan, Vol.66, No.4, p.21-32

  From a mechanical point of view, slope failure can be considered to occur when the groundwater level rises due to rainfall and exceeds a certain critical level. Therefore, if the state of rising groundwater level and the critical level can be expressed appropriately, it is possible to reduce “missing” and “missing” in predicting the occurrence of slope failure.

How the groundwater level rises (the amount and timing of the rise) is determined by how it rains, that is, the amount of rainfall, the amount of rainfall per unit time, etc., but even if it rains, the groundwater level rises depending on the location. The way to do is different. This is because the nature of vegetation, topography, soil, and rock (water retention and water permeability) varies from point to point, so the speed at which rain permeates, the ease of collecting water from the upstream area, and the water to the downstream area This is because the ease of passage varies.
Similarly, the critical level of the groundwater level at which slope failure occurs depends on the vegetation at the site, topography, and the nature of the soil and rocks (the amount of voids and strength against fracture), and therefore varies from site to site.

  Therefore, the rainfall index that can appropriately express the rise of the groundwater level, and the boundary line CL that can explain the occurrence and non-occurrence boundaries of slope failures without any contradiction should be different for each point. However, in the past, regardless of the location, a uniform rainfall index was used to express the rise of the groundwater level, and the boundary line CL was set based on past occurrences of slope failures. There was a limit to the accuracy of slope failure prediction.

  The problem to be solved by the present invention is to provide a slope failure prediction method and a slope failure prediction device capable of accurately predicting slope failure occurring at each target point.

The slope failure prediction method according to the present invention made to solve the above problems is as follows.
A method for predicting the occurrence of slope failure due to a rain event at a target location,
a) Time series of effective rainfall up to the present time, which is the rainfall index indicating the sustainability of the influence of past rainfall events on the accumulated rainfall, with the half-life, which is the time until the accumulated rainfall is halved at a certain time, as a parameter The value is calculated for a predetermined long half-life and a short half-life shorter than the long half-life using the rainfall data of the past rainfall event at the target point,
b) A time series value of the first effective rainfall is plotted on an XY coordinate plane with the first effective rainfall with the long half-life as a parameter on the X-axis and the second effective rainfall with the short half-life on the parameter as the Y-axis. A determination XY coordinate plane is created by setting a first boundary line passing through the maximum value and parallel to the Y axis and a second boundary line passing through the maximum value of the time series value of the second effective rainfall and parallel to the X axis. And
c) From the rainfall data of the rainfall event that is the target of slope failure occurrence, the time series value of the third effective rainfall with the long half-life as a parameter and the time series of the fourth effective rainfall with the short half-life as a parameter. The value is sequentially calculated along with the progress of the forecasted rainfall event,
d) When the measurement point is the point where the time series value of the third effective rainfall and the time series value of the fourth effective rainfall at the same time are the X value and the Y value, respectively, the measurement point is the determination XY It is determined that there is a possibility of slope failure when located in a region exceeding the first boundary line or in a region exceeding the second boundary line in the coordinate plane.

  Here, for example, data provided by a public organization such as the Japan Meteorological Agency or the Ministry of Land, Infrastructure, Transport and Tourism, or data provided by a private weather information providing service company can be used as the rainfall data. Moreover, the time series value of effective rainfall means that the values of effective rainfall at a plurality of times are arranged according to the passage of time. The time intervals for obtaining the time series values may be equal intervals or unequal intervals. In general, considering the provision of rainfall data by the above-mentioned public institutions and private companies at regular intervals (for example, every hour), time series values of effective rainfall can be calculated at regular intervals. Although it is preferable, during and before and after a rain event, the interval for calculating the time series value of the effective rainfall is shortened so that the rainfall between one rain event and the next rain event is zero, or the rain event Even in the middle of the period, the interval for obtaining the time series value may be lengthened during the period when the rainfall is very small.

In the above slope failure prediction method, the time series value of the first effective rainfall and the time series value of the second effective rainfall at the same time obtained from the rainfall data of the past rainfall event at the target point are respectively calculated. Plotting a plurality of points as X value and Y value on the XY coordinate plane for determination,
When a point at a certain time among the plurality of points is a point of interest, a point having a larger Y value than the point of interest does not exist among points having the same X value as the point of interest, and than the point of interest If neither the point having the same X value as the target point or the point having a larger Y value than the target point exists among the points having a large X value, an operation of extracting the target point as a third boundary point is performed. repetition,
The XY coordinates as a third boundary line representing the maximum value of the effective rainfall due to the synergistic effect of the first effective rainfall and the second effective rainfall, with a line connecting all the extracted third boundary points in ascending order of the X value Set to a plane,
It is preferable to determine that a slope failure may occur when the measurement point is located in a region exceeding the third boundary line.

  In the above method, a plurality of points having the time series value of the first effective rainfall and the time series value of the second effective rainfall at the same time as the X value and the Y value, respectively, plotted on the determination XY coordinate plane Corresponds to a “snake curve”. The snake curve is a graph generally used for predicting the occurrence of slope failure. In the above method, the third boundary line is set using a plurality of points constituting the snake curve. For this reason, possibility of occurrence of slope failure can be predicted based on a more appropriate boundary line.

  In order to accurately predict the occurrence of slope failure, it is important to select a long half-life and a short half-life. If these two types of half-life are selected incorrectly, an appropriate boundary line cannot be set on the XY coordinate plane for judgment, and the measurement point reflects the nature of the rain event and the water retention / permeability of the target point. It will not be. As a result, even if the possibility of occurrence of a slope failure is predicted based on the position of the measurement point on the determination XY coordinate plane, the prediction may end in “missing” or “miss”.

Therefore, in the slope failure prediction method of the present invention, for a combination of a plurality of long half-lives and short half-lives, a determination XY coordinate plane in which the first to third boundary lines are set is obtained.
For all of the plurality of determination XY coordinate planes, the measurement points including the corresponding time series value of the third effective rainfall and the time series value of the fourth effective rainfall are the X axis, the Y axis, and the first boundary line. And a first region surrounded by the second boundary line and the third boundary line, a second region surrounded by the first boundary line, the second boundary line and the third boundary line, the first A third region that exceeds a boundary line and is closer to the X axis than the second boundary line; a fourth region that exceeds the second boundary line and is closer to the Y axis than the first boundary line; It is preferable to determine in which region of the fifth region that exceeds both the boundary line and the second boundary line, and predict the occurrence of slope failure from the total of the determination results at each time.
According to this method, it is possible to reduce “swinging” that is predicted that a slope failure is highly likely to occur despite the fact that slope failure does not occur.

Further, in the slope failure prediction method of the present invention, for a combination of a plurality of long half-lives and short half-lives, a determination XY coordinate plane in which the first to third boundary lines are set is obtained.
For all of the plurality of determination XY coordinate planes, the measurement points including the corresponding time series value of the third effective rainfall and the time series value of the fourth effective rainfall are the X axis, the Y axis, and the first boundary line. And a first region surrounded by the second boundary line and the third boundary line, a second region surrounded by the first boundary line, the second boundary line and the third boundary line, the first A third region that exceeds a boundary line and is closer to the X axis than the second boundary line; a fourth region that exceeds the second boundary line and is closer to the Y axis than the first boundary line; It is determined in which region of the fifth region that exceeds both the boundary line and the second boundary line, and the measurement point is one of the second to fifth regions in at least one determination XY coordinate plane. It is preferable to predict that the occurrence of slope failure is high.
According to this method, it is possible to avoid “missing” that overlooks the possibility of slope failure.

Furthermore, in the slope failure prediction method of the present invention, for a combination of a plurality of long half-lives and short half-lives, a determination XY coordinate plane in which the first to third boundary lines are set is obtained.
For all of the plurality of determination XY coordinate planes, the closest measurement point among the measurement points consisting of the corresponding time series value of the third effective rainfall and the time series value of the fourth effective rainfall is the X axis and the Y axis. And a first region surrounded by the first boundary line, the second boundary line, and the third boundary line, and a first region surrounded by the first boundary line, the second boundary line, and the third boundary line. 2 region, the first boundary line is exceeded, and the third region on the X axis side than the second boundary line, the second boundary line is exceeded, and the second boundary line is exceeded on the Y axis side from the first boundary line. 4 region, determine which region is located in the fifth region that exceeds both the first boundary line and the second boundary line,
You may make it produce the two-dimensional determination figure which plotted the symbol showing the result on the two-dimensional coordinate plane which has a long half life on a horizontal axis and a short half life on a vertical axis | shaft.
In the present invention, it is important to set a long half-life and a short half-life. By creating the two-dimensional determination diagram, the change in the determination result due to the difference in the combination of the long half-life and the short half-life can be visually confirmed. Can be recognized. Therefore, appropriate long half-life and short half-life can be set with reference to the two-dimensional determination diagram.

In the slope failure prediction method of the present invention, in the above-described determination XY coordinate plane, the most recent measurement point among the measurement points is the X axis, the Y axis, the first boundary line, the second boundary line, and the It is characterized by calculating the amount of rainfall per unit time when the next measurement point exceeds the first to third boundary lines when located in a region surrounded by the third boundary line.
According to this method, the possibility of slope failure can be predicted quantitatively.

In this case, for a combination of a plurality of long half-lives and short half-lives, a determination XY coordinate plane in which the first to third boundary lines are set is obtained,
For all of the plurality of determination XY coordinate planes, the nearest measurement point among the measurement points consisting of the corresponding time series value of the third effective rainfall and the time series value of the fourth effective rainfall is the X axis and the Y axis. When located in a region surrounded by the first boundary line, the second boundary line, and the third boundary line, the next measurement point per unit time that exceeds the first to third boundary lines Calculate rainfall,
A two-dimensional determination diagram in which the rainfall per unit time is plotted on a two-dimensional coordinate plane with the horizontal axis as the long half-life and the vertical axis as the short half-life may be created.
According to this method, it is possible to visually recognize a change in rainfall that causes slope failure due to a difference in combination of the long half-life and the short half-life. Therefore, appropriate long half-life and short half-life can be set with reference to the two-dimensional determination diagram.

The slope failure occurrence prediction device according to the present invention is a device that predicts the occurrence of slope failure due to a rain event at a target point,
a) Time series of effective rainfall up to the present time, which is the rainfall index indicating the sustainability of the influence of past rainfall events on the accumulated rainfall, with the half-life, which is the time until the accumulated rainfall is halved at a certain time, as a parameter A past effective rainfall calculation means for calculating a value for each of a predetermined long half-life and a short half-life shorter than the long half-life, using rainfall data of past rainfall events at the target point;
b) A time series value of the first effective rainfall is plotted on an XY coordinate plane with the first effective rainfall with the long half-life as a parameter on the X-axis and the second effective rainfall with the short half-life on the parameter as the Y-axis. A determination XY coordinate plane is created by setting a first boundary line passing through the maximum value and parallel to the Y axis and a second boundary line passing through the maximum value of the time series value of the second effective rainfall and parallel to the X axis. A determination XY coordinate plane creating means;
c) From the rainfall data of the rainfall event that is the target of slope failure occurrence, the time series value of the third effective rainfall with the long half-life as a parameter and the time series of the fourth effective rainfall with the short half-life as a parameter. A prediction effective rainfall amount calculating means for sequentially calculating a value along with the progress of the prediction target rain event;
d) When the measurement point is the point where the time series value of the third effective rainfall and the time series value of the fourth effective rainfall at the same time are the X value and the Y value, respectively, the measurement point is the determination XY Determining means for determining that a slope failure may occur when located in a region exceeding the first boundary line in a coordinate plane, or when located in a region exceeding the second boundary line. And

In the above slope failure prediction device,
The determination XY coordinate plane creating means uses the determination XY coordinates as a plurality of points having the time series value of the first effective rainfall and the time series value of the second effective rainfall at the same time as the X value and the Y value, respectively. When plotting on a plane and setting a point at a certain time among the plurality of points as a point of interest, a point having a larger Y value than the point of interest does not exist among points having the same X value as the point of interest; and If neither the point having the same Y value as the point of interest nor the point having the Y value larger than the point of interest exists among the points having an X value larger than the point of interest, the point of interest is designated as the third boundary point. The line that connects all the extracted third boundary points in ascending order of the X value represents the maximum value of the effective rainfall due to the synergistic effect of the first effective rainfall and the second effective rainfall. Set as 3 boundary lines on the XY coordinate plane for determination,
It is preferable that the determination unit determines that slope failure may occur when the measurement point is located in a region exceeding the third boundary line on the determination XY coordinate plane.

Further, the past effective rainfall calculation means calculates time series values of the effective rainfall for three or more half-lifes having different lengths,
The determination XY coordinate plane creating means forms a plurality of combinations of long half-life and short half-life, respectively, by selecting two types of half-life from the three or more types of half-life. ~ Find the XY coordinate plane for determination with the third boundary set,
The determination means has a measurement point consisting of a time series value of the third effective rainfall and a time series value of the fourth effective rainfall for all of the plurality of determination XY coordinate planes, the X axis and the Y axis. A first region surrounded by the first boundary line, the second boundary line, and the third boundary line; a second region surrounded by the first boundary line, the second boundary line, and the third boundary line; A region that exceeds the first boundary line and is a third region on the X-axis side from the second boundary line; a fourth region that exceeds the second boundary line and that is on the Y-axis side from the first boundary line; It is determined whether the region is located in the fifth region that exceeds both the region and the first boundary line and the second boundary line, and the occurrence of the slope failure is predicted from the total of the determination results at each time Anyway.

Further, the past effective rainfall calculation means calculates time series values of the effective rainfall for three or more half-lifes having different lengths,
The determination XY coordinate plane creating means forms a plurality of combinations of long half-life and short half-life, respectively, by selecting two types of half-life from the three or more types of half-life. ~ Find the XY coordinate plane for determination with the third boundary set,
The determination means has a measurement point consisting of a time series value of the third effective rainfall and a time series value of the fourth effective rainfall for all of the plurality of determination XY coordinate planes, the X axis and the Y axis. A first region surrounded by the first boundary line, the second boundary line, and the third boundary line; a second region surrounded by the first boundary line, the second boundary line, and the third boundary line; A region that exceeds the first boundary line and is a third region on the X-axis side from the second boundary line; a fourth region that exceeds the second boundary line and that is on the Y-axis side from the first boundary line; It is determined whether the region is located in the fifth region that exceeds both the region, the first boundary line, and the second boundary line, and the measurement points are second to second in at least one determination XY coordinate plane. It is also a good configuration to predict that there is a high probability of slope failure when located in any of the five regions

Furthermore, in the above slope failure prediction device,
The determination XY coordinate plane creating means obtains a determination XY coordinate plane in which first to third boundary lines are set for a plurality of combinations of long half-life and short half-life,
For all of the plurality of determination XY coordinate planes, the closest measurement point among the measurement points consisting of the corresponding time series value of the third effective rainfall and the time series value of the fourth effective rainfall is the X axis and the Y axis. And a first region surrounded by the first boundary line, the second boundary line, and the third boundary line, and a first region surrounded by the first boundary line, the second boundary line, and the third boundary line. 2 region, the first boundary line is exceeded, and the third region on the X axis side than the second boundary line, the second boundary line is exceeded, and the second boundary line is exceeded on the Y axis side from the first boundary line. It is determined which region is located in the fourth region, the fifth region exceeding both of the first boundary line and the second boundary line, and a symbol representing the result is represented by the long half-life on the horizontal axis and the vertical axis 2D decision diagram creation means for creating a two-dimensional decision diagram plotted on a two-dimensional coordinate plane with a short half-life It may be.

In addition, the determination XY coordinate plane creating means obtains the time series value of the first effective rainfall and the time series of the second effective rainfall at the same time, which are obtained from the rainfall data of the past rainfall event at the target point. When a plurality of points whose values are X and Y values are plotted on the XY coordinate plane for determination and a point at a certain time is selected as a point of interest, the point having the same X value as the point of interest If there is no point having a larger Y value than the point of interest, and a point having a larger X value than the point of interest, a point having the same Y value as the point of interest and a Y value greater than the point of interest When none of the large points exist, the operation of extracting the attention point as the third boundary point is repeated, and a line connecting all the extracted third boundary points in ascending order of the X value is used as the first effective point. Effective rainfall due to the synergistic effect of rainfall and the second effective rainfall The set determination XY coordinate plane as a third boundary line representing a Daine,
In the determination XY coordinate plane, the nearest measurement point among the measurement points is in a region surrounded by the X axis, the Y axis, the first boundary line, the second boundary line, and the third boundary line. When located, it may be provided with an excessive rainfall amount calculating means for calculating a rainfall amount per unit time at which the next measurement point exceeds the first to third boundary lines.
The possibility of slope failure can be quantitatively predicted from the magnitude of rainfall per unit time that exceeds the first to third boundaries.

In this case, the determination XY coordinate plane creating means obtains a determination XY coordinate plane in which the first to third boundary lines are set for a plurality of combinations of long half-life and short half-life,
The excess rainfall calculation means, for all of the plurality of determination XY coordinate planes, of the measurement points including the corresponding time series value of the third effective rainfall and the time series value of the fourth effective rainfall, When located in a region surrounded by the X axis, the Y axis, the first boundary line, the second boundary line, and the third boundary line, the next measurement point exceeds the first to third boundary lines. Calculate the amount of rainfall per unit time,
By providing a two-dimensional judgment diagram creation means for creating a two-dimensional judgment diagram in which the rainfall per unit time is plotted on a two-dimensional coordinate plane with the horizontal axis representing the long half-life and the vertical axis representing the short half-life, It is possible to visually recognize changes in rainfall that cause slope failures due to differences in the combination of period and short half-life, and to set appropriate long and short half-lives with reference to the above two-dimensional judgment chart Can do.

  In the present invention, the measurement point used for determining whether or not slope failure may occur is usually the latest measurement point, but is not limited to this. For example, when the acquisition interval (measurement interval) of the time series values of the third effective rainfall and the fourth effective rainfall is short, measurement points that are traced back one or more times from the most recent measurement point may be used. . Further, the number of measurement points used for determination is not limited to one, and may be plural.

  In the slope failure prediction method and apparatus according to the present invention, a determination XY coordinate plane in which a boundary line is set based on rainfall data of past rainfall events in a target area is created, and an object for predicting the occurrence of slope failure The possibility of occurrence of slope failure is determined based on the positional relationship between the measurement point obtained from the rainfall amount data obtained along with the progress of the rain event and the boundary set on the determination XY coordinate plane. Therefore, it is possible to reflect the water retention and water permeability of the target point, the nature of the rain event to be predicted, etc. in the prediction of the occurrence of slope failure, and accurate slope failure that avoids “missing” and “skipping” Predictions can be made.

The figure which shows the example of the snake curve figure used for generation | occurrence | production prediction of slope failure. The graph which shows the hourly rainfall (a) and the effective rainfall (b)-(f) of various half-lives. Explanatory drawing of the soil rainfall index using a tank model. The figure explaining optimization of the boundary line CL in a snake curve figure. The figure which shows the example which painted the point of each time according to the magnitude | size of the groundwater level observed at the time in a snake curve diagram. The figure for demonstrating the definition of the boundary line CL in a snake curve. The figure which shows the determination result of the snake curve of the rain event used as the prediction object using the graph for determination. The figure which shows an example of the two-dimensional determination figure which shows the symbol of the determination result calculated | required from the several snake curve from which the combination of long half life and short half life differs. The schematic block diagram of the slope failure prediction apparatus which concerns on this invention. The rainfall data at the time of the Shobara disaster used in Example 1 of the slope failure prediction method according to the present invention. Judgment result of exceeding past maximum value by 2D judgment chart at Shobara disaster. The map which shows Example 2 of the slope failure prediction method which concerns on this invention, and shows the generation | occurrence | production location of Nashizawa disaster and the observation station of the vicinity (added to the Geographical Survey Institute map). A graph showing observation data at three stations. The graph which shows the amount of rainfall and the amount of rainfall of this rainfall event which caused the past rainfall event and the Nashizawa disaster. For past rainfall events and current rainfall events, the snake curve with the horizontal axis as the soil rainfall index and the vertical axis as the hourly rainfall, the soil rainfall index historical maximum value, the historical rainfall maximum value, and the historical maximum value obtained from the AN boundary. FIG. Snake curve and half-life 51h effective rainfall past maximum value and half-life 0.9h effective rainfall, with the horizontal axis representing the effective rainfall with a half-life of 51h and the vertical axis representing the effective rainfall with a half-life of 0.9h. The figure which shows the past maximum value obtained by the past maximum value and AN boundary line. The two-dimensional determination figure which shows the rainfall amount of 1 hour required until a measurement point exceeds AN boundary line calculated | required from the several snake curve from which the combination of long half life and short half life differs. The snake curve (a) obtained by adding the predicted rainfall for the next hour to the rainfall data up to now and the minimum hourly rainfall exceeding the AN boundary to the current rainfall data Snake curve (b). The conceptual diagram of a slope failure prediction system.

  It is difficult to accurately predict the critical level of the groundwater level at which slope failure occurs on natural slopes. Therefore, when the groundwater level rises above the previous maximum value, it is not surprising when slope failure occurs. It should be considered that there is no state. In the present invention, based on such an idea, the critical level is set with reference to the past maximum value of the groundwater level. As a result, it is considered that “missing” in the prediction of slope failure occurrence can be eliminated as much as possible.

The prediction of the groundwater level during rainfall utilized the concept of a functional model proposed by the present inventor in 2013 to analyze the fluctuation of the groundwater level inside the base rock (Non-Patent Document 5).
In the snake curve in which the horizontal axis and the vertical axis represent the effective rainfall with a long half-life and the effective rainfall with a short half-life, respectively, the present inventor determines each time point according to the magnitude of the groundwater level observed at that time. It is a non-patent document that by optimizing the half-life of the effective rainfall on the horizontal axis and the vertical axis, the color of the points can be applied beautifully according to the magnitude of the effective rainfall on the horizontal axis and the vertical axis. Shown in Reference 5. In other words, when a snake curve is created from two types of effective rainfall with different half-lives, the groundwater level rise can be reproduced if the combination of the two types of half-lives is optimal.

  FIG. 5 shows a snake curve diagram (a) having an effective rainfall of half-life 72 hours on the horizontal axis, an effective rainfall of half-life 1.5 hours on the vertical axis, and an effective rainfall of half-life 1061 hours on the horizontal axis. The example of the snake curve figure (b) which made the vertical axis | shaft the effective rainfall of the half life 77 hours is shown. In the figure, the colors of the points at each time are painted in order from the lowest groundwater level observed at that time. Here, FIG. 5 is represented by white and black shading, but in actuality, it is painted in brown, red, yellow, yellow-green, and blue. In the same figure (a), while the point of a different color is mixed over a wide range, in the same figure (b), a plurality of points exist together for each color, and the effective rainfall and It is estimated that there is a correlation between groundwater levels.

  From the above, on the snake curve diagram of Fig. 5 (b), when the effective rainfall exceeds the past maximum value, it is considered that the groundwater level also exceeds the past maximum value, and this snake curve diagram is used. Therefore, when predicting the occurrence of slope failure, it is most reasonable to set the past maximum value of effective rainfall to the boundary line CL. The historical maximum value of the effective rainfall is the maximum value of the time series value of the first effective rainfall (first boundary line), the maximum value of the time series value of the second effective rainfall (second boundary line), and This corresponds to the maximum value (third boundary line) of the effective rainfall due to the synergistic effect of the first effective rainfall and the second effective rainfall. The first to third boundary lines will be described later.

  However, when the effective rain half-life taken on the vertical and horizontal axes of the snake curve diagram is set to what value, a snake curve diagram as shown in FIG. What is the half-life that best represents the rise in groundwater level varies from point to point. For this reason, it is difficult to accurately predict an appropriate combination of half-life at all target points.

  In contrast, the effective rainfall half-life (hereinafter also referred to as “long half-life”) set on the horizontal axis and the effective rainfall half-life (hereinafter also referred to as “short half-life”) set on the vertical axis are various. By changing the length and determining whether the maximum groundwater level has been exceeded for any combination of half-lives, the maximum groundwater level has been exceeded at any point where any groundwater level rise pattern has occurred. Can be detected. At the past maximum value of the groundwater level, at least slope failure has not occurred, and it is unclear whether slope failure will occur when exceeding it, so as described above, for all combinations of half-lives, Determining whether or not the past maximum value of the groundwater level has been exceeded leads to avoiding “skipping” and “missing” in predicting the occurrence of slope failure.

However, as described above, if the optimum combination of the long half-life and the short half-life can be selected, it is only necessary to determine whether or not the maximum groundwater level has been exceeded for one combination. It is possible to predict slope failure that avoids “skipping” and “missing”.
In addition, it is practically impossible to determine whether or not the maximum groundwater level has been exceeded for all combinations of the long half-life and the short half-life. In order to reduce “missing” and “missing”, it is effective to use a larger number of combinations of long half-life and short half-life for determination. However, the above-mentioned determination should be performed for at least two combinations. Thus, the prediction accuracy of slope failure can be improved as compared with the case where determination is made for one combination.

  The historical maximum value of effective rainfall, that is, the historical maximum value of the groundwater level is merely the maximum value of the groundwater level during the past rainfall, and does not represent the groundwater level when slope failure occurs. For this reason, while the rainfall observation period is short, even if the occurrence of slope failure is predicted based on the fact that the effective rainfall has exceeded the past maximum value, there is a high possibility that a situation of “swinging” frequently occurs. However, every time the “swinging” occurs, the past maximum value of the effective rainfall is updated and the boundary line CL is raised, so that the “swinging” gradually decreases by continuing to observe the rainfall.

  In addition, unlike the prior art, the rainfall index (two types of effective rainfall) used in the present invention has a clear physical meaning of “a rainfall index that can best reproduce the state of rising groundwater level” (FIG. 5). reference). For this reason, if a change in the groundwater level during actual rainfall can be measured, an appropriate half-life can be specified. That is, the present invention has an advantage that the observation result of the groundwater level at the target point can be directly reflected in the prediction of slope failure occurrence. For this reason, the accumulation of groundwater level observation data can effectively reduce “swinging”.

  Furthermore, in the present invention, it is determined whether or not the past maximum value of the effective rainfall has been exceeded, and the occurrence of slope failure is predicted based on the determination result. Such a judgment is made by many residents on a regular basis, such as “This kind of rain is okay because it has happened before”, “This kind of rain is dangerous because it happens for the first time” , Because it leads to the judgment of the risk of slope failure based on its own past experience, even if “empty swing” occurs, it is easier for residents to accept than conventional prediction technology, and awareness of warning and evacuation It is difficult to cause a decline.

Whether or not the effective rainfall exceeds the past maximum value is determined using, for example, the snake curve diagram shown in FIG. The snake curve diagram of FIG. 6 corresponds to the determination XY coordinate plane of the present invention.
First, using the long-term hourly rainfall data immediately before the occurrence of a rain event at the target point, two types of effective rainfall X (M ₁ ) with half-lives M ₁ and M ₂ (where M ₁ ≧ M ₂ ), respectively. , T), X (M ₂ , t), and draws a snake curve diagram with the effective rainfall X (M ₁ , t) on the horizontal axis and the effective rainfall X (M ₂ , t) on the vertical axis. In this snake curve diagram, “exceeding past maximum value” is classified into the following four cases.

First, on the snake curve diagram, the effective rainfall X (M ₁ , t) = X _max (M ₁ ) (the past maximum value of X (M ₁ , t)) and the effective rainfall X (M ₂ , t) = X _{max Two} straight lines of (M ₂ ) (the past maximum value of X (M ₂ , t)) are drawn. These two straight lines correspond to the first boundary line and the second boundary line of the present invention. Of the four regions formed by these two straight lines, region B corresponds to a case where the effective rainfall amount X (M ₁ , t) alone exceeds the past maximum value. A region C represents a case where the effective rainfall amount X (M ₂ , t) alone exceeds the past maximum value. Region D is a case where both the effective rainfall amount X (M ₁ , t) and the effective rainfall amount X (M ₂ , t) exceed the past maximum values.

On the other hand, the area (N + A) is a case in which neither the effective rainfall X (M ₁ , t) nor the effective rainfall X (M ₂ , t) exceeds the past maximum value, of which the area N is effective rainfall A case where the combination of X (M ₁ , t) and effective rainfall X (M ₂ , t) is within the past range is shown in the region A, where the effective rainfall X (M ₁ , t) and effective rainfall X (M ₂ , t In the case of the combination of t), it represents a case where the past maximum value is exceeded.

In order to define a boundary line (hereinafter referred to as “AN boundary line”) that separates the area A and the area N, points on the AN boundary line (hereinafter referred to as AN boundary points) are extracted by the following method. That is, there is no point where point p and X (M ₁ , t) are equal to point p on the snake curve at a certain time, and X (M ₂ , t) is larger than point p, and When there is no point where X (M ₁ , t) is greater than the point p, a point where the point p is equal to X (M ₂ , t), and a point where X (M ₂ , t) is greater than the point p Then, the point p is determined as the AN boundary point . Follow me, corresponds to the target point of the first to point chose p is the present invention, AN boundary point is equivalent to the third boundary point.
Then, all the extracted AN boundary points are rearranged in order from the smallest effective rainfall amount X (M ₁ , t), and are connected by a straight line as shown in FIG. The stepped line obtained in this way is used as the AN boundary line. The AN boundary line corresponds to the third boundary line of the present invention. Note that the effective rainfall amount X (M ₂ , t) at the first AN boundary point after the rearrangement coincides with the past maximum value X _max (M ₂ ), and the effective rainfall amount X (M ₁ , t at the last AN boundary point). ) Matches the past maximum value X _max (M ₁ ).

  After the region N and the regions A to D are determined by the above procedure, the region of the point of each time belongs is examined by superimposing the locus of the snake curve diagram during the event (FIG. 7). From the result, for each time, it is determined whether or not the past maximum value has been exceeded, and if it has occurred, it is determined which of the areas A to D corresponds to the case.

The analysis described above is performed with the half lives M ₁ and M ₂ set to various values. The results obtained at each time are displayed in the two-dimensional determination diagram of FIG. In the two-dimensional determination diagram, the horizontal axis is the half-life M ₁ , the vertical axis is the half-life M _2, and in the snake curve diagram by a combination of these half-life M ₁ and M ₂ , It is the figure which represented to which of N by the symbol corresponding to each area | region. Therefore, this two-dimensional determination diagram corresponds to the two-dimensional determination diagram according to claims 5 and 12. In addition, the point on the diagonal line of the two-dimensional judgment diagram is the result of the snake curve diagram of half-life M ₁ = half-life M ₂ , that is, whether or not the past maximum value is exceeded when one kind of effective rainfall is evaluated independently. The determination result is shown. In this case, it is either the region N or D and does not become the regions A to C.

  There are various forms of slope failure. “Surface layer collapse” in which the soil layer with a thickness of about 0.5-2m near the ground surface collapses, and large-scale collapse of several tens of meters in thickness, including the rock layer (base rock layer) constituting the mountain body Although “collapse” is extreme, there are many slope failures that show their intermediate form. In the method of the present invention, “exceeding the past maximum value is determined for two types of effective rainfall combinations having any half-life”, it is quantitatively determined that there is a difference in the rainfall index to be noted according to the slope failure type. Expected to be able to show. Therefore, it is possible to accurately predict the type of slope failure that can occur based on the rainfall data based on the results, and to establish a more detailed alert / evacuation system.

  In addition, the method of the present invention can evaluate not only the prediction of slope failure, but also the influence of slope changes caused by earthquakes, tree wind damage, artificial land modification, etc. on slope failure. For example, in an area that has experienced a large earthquake, there is a concern that collapse may occur even in rainfall of a scale that did not cause collapse in the past due to a decline in slope strength, and reduction of the boundary line CL is often considered. . Based on the method of the present invention, by continuously changing the half-life of the snake curve diagram and evaluating the long-term and short-term effects of rainfall, slope failure in situations where actual past maximum values do not actually occur It is possible to make an objective and quantitative examination as to whether or not this occurs.

  FIG. 9 shows a schematic diagram of a slope failure prediction apparatus for executing the slope failure prediction method according to the present invention. The slope failure prediction device 10 includes a data reception unit 11, a past effective rainfall calculation unit 12, a determination XY coordinate plane generation unit 13, a prediction effective rainfall calculation unit 14, a determination unit 15, a two-dimensional determination diagram generation unit 17, and the like. The control part 16 which controls is provided. An output device 30 such as a printer or a display is connected to the control unit 16.

When the data receiving unit 11 receives the rainfall data from the rainfall data providing device 20 in response to the command from the control unit 16, the control unit 16 sends the rainfall data to the past effective rainfall calculation unit 12.
The rainfall data providing device 20 refers to an organization having a database storing past rainfall data, such as a disaster prevention information providing center of the Ministry of Land, Infrastructure, Transport and Tourism, an observation station, and an AMeDAS observation station of the Japan Meteorological Agency. The data receiving unit 11 may directly receive the rainfall data from the rainfall data providing device 20, or may acquire the rainfall data by accessing a database via the Internet. The past effective rainfall calculation unit 12 that has received the rainfall data from the data receiving unit 11 via the control unit 16 uses the past rainfall data and newly acquired rainfall data for a plurality of preset half-lives. The time series value of effective rainfall is calculated.

The determination XY coordinate plane creation unit 13 extracts a plurality of arbitrary combinations of two types of half-life from these time series values of the effective rainfall, and the longer half of the two half-lives of each set is reduced by half. The snake curve is obtained by setting the shorter period and time as the short half-life. Then, a determination XY coordinate plan view in which the first to third boundary lines are set based on the snake curve is created and stored.
The determination XY coordinate plan view is created and stored every time new rainfall data is received.

  On the other hand, the prediction effective rainfall calculation unit 14 calculates effective rainfalls having a plurality of preset half-lives using newly acquired rainfall data. And the point (measurement point) which shows the time series value of the effective rainfall calculated about two types of half-life combinations is plotted on the XY coordinate plane view for judgment of the corresponding half-life combination. The determination unit 15 predicts the possibility of slope failure from the positional relationship between the position of the measurement point and the first to third boundary lines. The determination unit 15 predicts the possibility of slope failure with respect to the plurality of determination XY coordinate plan views created by the determination XY coordinate plane creation unit 13 and stores the result. In addition, if the measurement point does not exceed the first to third boundary lines, the amount of rainfall for the next fixed time (for example, 1 hour) (or “per unit time”) that will exceed the boundary line is predicted. To do. Therefore, here, the determination unit 15 also functions as an excess rainfall calculation unit.

The two-dimensional determination diagram creating unit 17 is a two-dimensional determination in which symbols representing prediction results stored in the XY coordinate plane creating unit 13 are plotted on a two-dimensional coordinate plane with the horizontal axis representing the long half-life and the vertical axis representing the short half-life. Create a diagram. As a symbol representing the prediction result, for example, a graphic having a different size or a graphic having a different color according to the possibility of slope failure can be used.
Next, an example of analysis using specific rainfall data will be described.

(1) Target disasters In Kawakita-cho, Shobara-shi, Hiroshima Prefecture, slope failure occurred intensively due to the heavy rain that occurred on July 16, 2010. One deceased, one seriously injured, and 38 houses were destroyed. Caused damage. The collapsed slope has a complex geology composed of volcanic rocks, sedimentary rocks, and kuroboku soil layers derived from volcanic ash. It is planted in cedar and cypress plantations and broad-leaved secondary forests, but includes logging sites. Most of the slope failures were classified as surface layer failures on a scale of 10-15m in width, 20-60m in length, and 0.5-1.5m in depth. Below, this disaster is called the Shobara disaster.

  FIG. 10 shows rainfall data observed at the nearest Hiroshima Prefectural Oto Observatory in Shobara City, heavy rain warning announcement time, slope failure start time, and earth and sand disaster warning information announcement time. From Fig. 10, the rainfall event that caused the Shobara disaster is considered to be a heavy rain with a high rainfall intensity, but the heavy rainfall of 260 to 270 mm of accumulated rainfall during the five days preceding this also caused a major disaster. It seems that it had an influence. As shown in Fig. 10, the Hiroshima Regional Meteorological Observatory announced a heavy rain warning to Shobara City at 16:39 on July 16, and the earth and sand disaster warning information was announced at 18:10. It was. However, sightings of slope failures and debris flows occurred at 16:55, and it is probable that many slope failures began to occur before 17:00.

(2) Rainfall data used for analysis The data of Hiroshima Oto Station shown in Fig. 10 cannot be acquired retroactively over a long period of time. Therefore, the maximum effective rainfall was calculated using the rainfall data of the Japan Meteorological Agency Amedas Shobara Observatory (34 ° 51.6min north, 133 ° 1.4min east longitude, 300m above sea level) located about 7km away from the affected area. The rainfall data at Shobara Observatory can be traced back to 1 o'clock on January 1, 1976, approximately 34 years and 7 months before the Shobara disaster. The effective rainfall was calculated continuously from the start time of the data with the initial value set to zero. However, the historical maximum was evaluated for a period 30 years after midnight on January 1 of the year when the disaster occurred. This is a measure aimed at removing the influence of the initial value on the calculation result of the effective rainfall.

The half-life M ₁ , M ₂ (where M ₁ ≧ M ₂ ) was in the range of 0.1 h to 3000 h, and the change amount of the logarithmic value was constant, and the total was changed in 21 ways. Here, the setting range of the half-life (0.1h to 3000h) is set with the intention that the effective rainfall sufficiently covers both extreme indicators relating to the sustainability of the influence of preceding rainfall such as rainfall intensity and accumulated rainfall. In particular, regarding the maximum value (3000h), in Non-Patent Document 5, the maximum value (1400h) used by the present inventors for reproduction of groundwater level data is taken into consideration, and the maximum value is set to about twice that value. The setting interval of the half-life is determined after confirming that a continuous tendency can be obtained.

(3) Results FIG. 11 shows a two-dimensional determination diagram regarding the Shobara disaster. At 16:00 on July 16, 1 hour after the start of rainfall (accumulated rainfall of 38 mm; see Fig. 10), region A is already in the range of 48.6 h ≤ M ₁ ≤ 1070 h, 0.100 h ≤ M ₂ ≤ 0.469 h It appears that the past maximum value has occurred (FIG. 11a). When focusing on the single effective rainfall, none of the effective rainfall exceeded the past maximum at this time.
At 17:00, just after the 72 mm hourly rainfall, many past combinations of half-lives M ₁ and M ₂ exceeded the maximum value (FIG. 11b). However, at this time, the past maximum value did not occur in the combination in the range of 136h ≦ M ₂ ≦ M ₁ . Furthermore, at 18:00 when the 63mm hourly rainfall has fallen, the past maximum excess has expanded to a larger range of half-life M ₁ and M ₂ (Fig. 11c), and at 19:00 at the end of the rain half-life _{M 1} is reduced in a small range (Figure 11d).

(4) Consideration The rainfall for 1 hour (15-16 o'clock) at the beginning of the Shobara disaster is 38 mm (Fig. 10), about 60% compared to the previous maximum of 64 mm (recorded on August 22, 1998). It was not so big value. However, there was a cumulative rainfall of 271 mm over the preceding 5 days. In the two-dimensional judgment chart (Fig. 11), the effective horizontal rainfall of 48.6h ≤ M ₁ ≤ 1070h has increased greatly due to the influence of the preceding rainfall, and the time rainfall of 38mm from 15 to 16 o'clock is 0.100h ≤ M ₂ ≤ The combination of the increase in the effective vertical rainfall of 0.469h indicates that the previous maximum value exceeded occurred at 16:00.

On the other hand, in the conventional method using an effective rainfall with a half-life of 72 hours as a long-term rainfall index and an effective rainfall with a half-life of 1.5 hours as a short-term rainfall index, it was 17:00 that the past maximum was first determined. (Not shown) and later than the onset of slope failure. In addition, among the conventional techniques, the results obtained by the method using the soil rainfall index as the long-term rainfall index and the hourly rainfall as the short-term rainfall index are the sub-windows in the two-dimensional determination diagram of FIG. 11 (solid squares in each panel). Shown in. Also in this case, it is 17:00 (see symbol “●” in the sub-window of FIG. 11b) that the past maximum value is determined for the first time, which is later than the start of slope failure.
Therefore, unlike the conventional determination method in which the determination of exceeding the past maximum value is delayed from the actual start of the occurrence of slope failure, according to the prediction method according to the present invention, the degree of risk is one hour before the occurrence of the Shobara disaster. Can be predicted, and “missing” can be reduced as compared with the prior art.

(1) Target disaster In Minamikiso, Nagano Prefecture, a heavy rain affected by typhoon No. 8 caused a debris flow in the tributary of the Kiso River called Nashizawa at around 17:40 on July 9, 2014. . Hereinafter, this disaster is referred to as “Nashizawa disaster”. On the day of the Rashizawa disaster, it was sunny until noon, and it started to rain about 2 hours before the debris flow. In this way, debris flow occurred unexpectedly, and the evacuation recommendation was issued to Nagiso Town at 17:50, 10 minutes after the occurrence of debris flow. The earth and sand disaster warning information was announced at 18:15, another 25 minutes later.

FIG. 12 shows the positions of the Nashizawa basin where debris flow occurred and the rain observation stations in the vicinity (Namiso Station of the Meteorological Agency, Mitsuno Station and Orchid Station of the Ministry of Land, Infrastructure, Transport and Tourism). As can be seen from FIG. 12, Nagiso Station is the nearest station to the debris flow generation basin.
Figure 10 shows the 10-minute rainfall on the day of the Nashizawa disaster at these three stations. At all stations, the rainfall was temporarily stopped after 3:00 on July 9, but it resumed after 15:30, and the rainfall intensity increased within a short time. The maximum rainfall for 10 minutes was 23 mm at the Orchid Observatory, followed by 17 mm at the Nagiso Observatory.

  The accumulated rainfall from 15:30 to 17:40, when debris flow occurs, reached 84 mm at the Nagiso Observatory nearest to the debris flow generation basin, but less than those at the Mitsuno and Orchid stations. In addition, the rainfall during the 1 hour (16: 40-17: 40) until the debris flow occurred reached 70 mm at the Nagiso Observatory, which was also less than the 97 mm at the Orchid Observatory. In this way, the rainfall at the Nagiso Observatory nearest to the debris flow basin tended to be slightly less than that of the other two stations. According to statistics from the Japan Meteorological Agency, the maximum rainfall for 1 hour at the Nagiso Observatory was 89 mm and the maximum rainfall for 10 minutes was 18 mm, which was not recorded in the current rainfall event.

(2) Comparison with past rainfall events (2-1) Comparison of hourly rainfall, accumulated rainfall, and preceding rainfall Going back to 1984, 30 years ago from the Nashizawa disaster, extracted major rainfall events that occurred in that period For these past rainfall events and the rain events that brought about the Rikozawa disaster, we compared the hourly rainfall (at the time of the hour) observed at the Nagiso Station and the accumulated rainfall. Hereinafter, the extracted past rainfall events are referred to as “S60 event”, “H16 event”, “H12 event”, and “H18 event” using the year of occurrence. Similarly, the rainfall event that brought about the Rikozawa disaster is called the “H26 event”.
FIGS. 14a to 14d are graphs showing the amount of rainfall and accumulated rainfall for three past rainfall events (H16 event, H12 event, H18 event) and H26 event. Table 1 below summarizes the specifications of the four past events (S60 event, H16 event, H12 event, H18 event) and H26 event.

  In addition, here, the rain event is separated by a no-rain period of 24 hours or more. According to this, the H26 event started at 18:00 on July 6, 2014. As shown in FIG. 14d and Table 1, the accumulated rainfall from the start of rainfall to the occurrence of debris flow in the H26 event was 194 mm.

The H16 event shown in FIG. 14a is an event in which the above-mentioned maximum rainfall of 1 hour (89 mm) is observed, and it can be said that it was a considerable torrential rain. The accumulated rainfall was relatively low, but it rose to a value comparable to the H26 event at the end.
In the H12 event shown in FIG. 14b, although the peak (maximum value) of the hourly rainfall was small, a large amount of rain fell in a relatively short time, and the accumulated rainfall was about 85 mm greater than the H26 event.
In the H18 event shown in FIG. 14c, the hourly rainfall peak is smaller than that of the H16 event, but the duration of the rainfall is as long as 150 hours, and the accumulated rainfall is very large.
The S60 event shown in Table 1 had an extremely long rainfall duration of 262 hours, and the accumulated rainfall reached 2.7 times that of the H26 event, but the peak of hourly rainfall was small.
As can be seen from Table 1, the rainfall during the two days preceding each rainfall event (the rainfall during the preceding two days) was the highest in the H26 event, but the other rainfall events other than the H26 event were the largest for the other preceding rainfall. The value is shown.

  Based on the above, when using hourly rainfall, accumulated rainfall, and preceding rainfall as indices, the specificity of the H26 event is not detected, and it cannot be concluded that the H26 event was an abnormal rainfall event that exceeded previous rainfall events. It was.

(2-2) Comparison Using Snake Curves In many cases, a snake curve with the horizontal axis as the soil rainfall index and the vertical axis as the hourly rainfall is used as the standard for landslide disaster alert information announcement. Therefore, the snake curve for each rainfall event was obtained.
First, the soil rainfall index was calculated using the hourly rainfall data (Nanoki Pass Observatory) for about 30 years from 1:00 on January 1, 1984 to just before the start of the H26 event, and a past snake curve was drawn (Figure 15a). Gray circles in ˜15d). The thick solid line in each figure connects the maximum values of the past snake curve, and is hereinafter referred to as “the past maximum value line”. When the snake curve enters the upper right area of the past maximum value line, the past maximum value is exceeded.

  Next, snake curves of past rainfall events (H16, H12, H18 events) and H26 events were written in FIGS. 15a to 15d with black circles and white circles. The snake curves for past rainfall events have all risen to the AN boundary, indicating that it was a rain event that formed part of the past maximum value just before the start of the H26 event. Further, from FIG. 15d, it can be seen that the snake curve of the H26 event has entered a region where the past maximum value is slightly exceeded and slightly exceeds the past maximum value line when the debris flow occurs.

  As described above, in addition to the snake curve shown in FIG. 15, a snake curve having a long half-life effective rainfall and a vertical axis having a short half-life effective rain is used for warning and evacuation of sediment disasters. There are many things. Therefore, the same analysis as in FIG. 15 was performed using a snake curve with the horizontal axis representing the effective rainfall with a half-life of 51 h and the vertical axis representing the effective rainfall with a half-life of 0.9 h. The result is shown in FIG. Referring to FIG. 16d, the snake curve of the H26 event markedly exceeded the past maximum value line when debris flow occurred, and it is clear that the past maximum value occurred. When such an analysis was performed with various combinations in which the half-life of the horizontal-axis effective rainfall and the half-life of the vertical-axis effective rainfall differed, the horizontal-axis half-life was 13 h to 173 h, and the vertical-axis half-life was 0.1 h to 2.0 h. It was found that when the debris flow of the H26 event occurred, the occurrence of exceeding the existing maximum value was detected.

  In the H26 event, there was a lot of rainfall in the preceding two days (Table 1), and the accumulated rainfall after the start of the event exceeded 100 mm, but it did not reach the previous maximum value, but it was raining with considerable intensity (Figure) 72 hours have passed since the start of the 14d event). The above-mentioned analysis results can be detected as “exceeding past maximum value” by using a snake curve with appropriately set rainfall indices on the vertical and horizontal axes, and as a result, exceed the past rainfall events. It was suggested that an abnormal rainfall event is possible.

(3) Prediction of occurrence of abnormal rainfall event In Figs. 15d and 16d, the snake curve that was within the past range at 17:00 on July 9 is the past maximum with 49 mm of rain that has fallen until the occurrence of debris flow. Exceeded. In this way, instead of calculating whether or not the past maximum value has been exceeded using the given rainfall, at 17:00 on July 9, how much rain will occur in the next hour will exceed the past maximum value. Calculated what to do. As a result, a value of 47.1 mm was obtained in the case of FIG. 15d and 35.6 mm in the case of FIG. 16d.

Snake curve created by changing the half-life (long half-life M ₁ ) of the effective rainfall on the horizontal axis and the half-life (short half-life M ₂ ) of the vertical rainfall on the vertical axis in the range of 0.1h to 3000h. It was performed using. The result is shown in FIG. In FIG. 17, the amount of rainfall exceeding the past maximum value obtained from the snake curve is represented by “horizontal axis effective rainfall half-life (h)” on the horizontal axis and “vertical effective rainfall half-life (h)” on the vertical axis. It is the graph shown on the two-dimensional coordinate and color-coded according to the magnitude of rainfall. This graph corresponds to the two-dimensional determination diagram according to claims 7 and 14. That is, in the region where the half-life of the effective rainfall on the horizontal axis is 40 to 100 h and the half-life of the effective rainfall on the vertical axis is 1.3 h, the rainfall exceeding the existing maximum value is 34.3 mm, which is the smallest. In fact, debris flow occurred as a result of 49 mm of rain exceeding this.

  It is difficult to predict the occurrence of debris flow this time, as it is difficult to predict the torrential rain caused by the rapidly developing clouds, as in the rain that caused the debris flow to occur directly at the H26 event (Fig. 13a). It is. However, as discussed in section (2-2), it is considered that the H26 event was also significantly affected by the preceding rainfall and the cumulative rainfall after the start of the rainfall event. For this reason, by performing the analysis of FIG. 17 in consideration of these effects, it was possible to predict that an abnormal rainfall event that would cause debris flow would occur at 17:00, due to a special rainfall of 34 mm. There is sex.

Therefore, the horizontal axis and the vertical axis effective rainfall of the predicted rainfall for the next hour at a certain point in the rain event are calculated, and based on whether the resulting snake curve exceeds the past maximum value line, The possibility of debris flow generation can be predicted. The predicted rainfall can be obtained from weather data (rainfall, wind direction, wind speed, temperature, etc.), a weather map, and the like.
Specifically, as shown in FIG. 18 (a), when the predicted rainfall for the next hour at 17:00 is 49 mm, the horizontal axis and the vertical axis effective rainfall points (“18:00 The point indicated by “” is located in an area beyond the past maximum value line. From this, it can be predicted that there is a risk of debris flow occurring in the next hour.

  Also, create a snake curve for various changes in rainfall over the next hour, and find the minimum value from the rainfall over the next hour when the past maximum value line is exceeded. It is also possible to quantitatively evaluate the possibility of debris flow generation (risk of debris flow generation) from comparison with the predicted rainfall for one hour. For example, as shown in FIG. 18B, when the rainfall for the next hour at 17:00 is 38 mm, the horizontal axis and the vertical axis indicate the effective rainfall points on the past maximum value line. In this case, this value (38 mm) and the predicted rainfall (49 mm) can be compared to determine the degree of possibility of debris flow occurring in the next hour. For example, if the rainfall amount on the past maximum value line is 0 to 50% of the predicted rainfall amount, the possibility is “very large”, and if it is 50 to 80%, the possibility is “large”, 80% When it is ˜100%, the possibility is “medium”, and when it is 100% ˜, the possibility is “small”.

Furthermore, a system for predicting the occurrence of debris flow can be constructed using the above-described two-dimensional determination diagram (FIG. 8) and the two-dimensional determination diagram used in this embodiment (FIG. 17). For example, FIGS. 19A and 19B show two types of two-dimensional determination diagrams at the present time obtained from past rainfall data at a certain point. In this system, these two-dimensional determination diagrams are displayed on a liquid crystal display or the like. Then, select the point the user is optional in any of the decision diagram displayed on the liquid crystal display (clicked), the long half-life M ₁ of points the selection and short half-life M _2, respectively horizontal and vertical axes A snake curve is automatically created (FIG. 19 (c)), and an estimated point obtained from the difference in rainfall over the next hour is written. FIG. 19C shows an example in which estimated points are displayed when the rainfall amount for the next hour is 0 mm, 20 mm, 38 mm, 40 mm, 49 mm, and 60 mm.
Therefore, the user can estimate the amount of rainfall for the next hour that causes the occurrence of debris flow from the position of the estimated point.

  As described above, the first embodiment and the second embodiment described above are useful as slope failure prediction methods, respectively. Furthermore, if the methods shown in these embodiments are used in combination, the slope failure prediction accuracy is further improved. be able to.

  The present invention can be put into practical use immediately and has a great effect of improving the accuracy of slope failure prediction. However, the following problems can be raised for further accuracy improvement.

  It is necessary to examine how to set the period for calculating the maximum effective rainfall. In the above example, about 30 years were targeted for analysis by using data that was released from the Japan Meteorological Agency and easy to obtain, but if more long-term rainfall data is used, a more accurate past maximum value is calculated. It is possible to reduce the “missing” in the collapse prediction. On the other hand, considering that there is a secular change in the characteristics of the slope, it is not always good to go back to the past, and there may be an appropriate period. It is necessary to accumulate knowledge by analyzing more disaster cases.

DESCRIPTION OF SYMBOLS 10 ... Slope failure prediction apparatus 11 ... Data receiving part 12 ... Past effective rainfall calculation part 13 ... XY coordinate plane preparation part 14 for judgment 14 ... Effective prediction rainfall calculation part 15 ... Determination part 16 ... Control part 17 ... Two-dimensional determination figure preparation Unit 20: Rainfall data providing device 30 ... Output device

Claims (14)

A slope failure prediction method for predicting the occurrence of slope failure due to a rain event at a target location,
a) Time series of effective rainfall up to the present time, which is the rainfall index indicating the sustainability of the influence of past rainfall events on the accumulated rainfall, with the half-life, which is the time until the accumulated rainfall is halved at a certain time, as a parameter The value is calculated for a predetermined long half-life and a short half-life shorter than the long half-life using the rainfall data of the past rainfall event at the target point,
b) A time series value of the first effective rainfall is plotted on an XY coordinate plane with the first effective rainfall with the long half-life as a parameter on the X-axis and the second effective rainfall with the short half-life on the parameter as the Y-axis. A determination XY coordinate plane is created by setting a first boundary line passing through the maximum value and parallel to the Y axis and a second boundary line passing through the maximum value of the time series value of the second effective rainfall and parallel to the X axis. And
c) From the rainfall data of the rainfall event that is the target of slope failure occurrence, the time series value of the third effective rainfall with the long half-life as a parameter and the time series of the fourth effective rainfall with the short half-life as a parameter. The value is sequentially calculated along with the progress of the forecasted rainfall event,
d) When the measurement point is the point where the time series value of the third effective rainfall and the time series value of the fourth effective rainfall at the same time are the X value and the Y value, respectively, the measurement point is the determination XY Slope failure prediction characterized in that it is determined that slope failure may occur when located in a region exceeding the first boundary line or located in a region exceeding the second boundary line in a coordinate plane Method. In the slope failure prediction method according to claim 1,
A plurality of time series values of the first effective rainfall and time series values of the second effective rainfall obtained at the same time obtained from rainfall data of past rainfall events at the target point are set as an X value and a Y value, respectively. Plot points on the XY coordinate plane for determination,
When a point at a certain time among the plurality of points is a point of interest, a point having a larger Y value than the point of interest does not exist among points having the same X value as the point of interest, and than the point of interest If neither the point having the same X value as the target point or the point having a larger Y value than the target point exists among the points having a large X value, an operation of extracting the target point as a third boundary point is performed. repetition,
For the determination, a line connecting all the extracted third boundary points in ascending order of the X value is a third boundary line representing the maximum value of the effective rainfall due to the synergistic effect of the first effective rainfall and the second effective rainfall. Set to XY coordinate plane,
A slope failure prediction method, wherein it is determined that slope failure may occur when the measurement point is located in a region exceeding the third boundary line on the determination XY coordinate plane. In the slope failure prediction method according to claim 2,
For a plurality of combinations of a long half-life and a short half-life, a determination XY coordinate plane in which the first to third boundary lines are set is obtained,
For all of the plurality of determination XY coordinate planes, the measurement points including the corresponding time series value of the third effective rainfall and the time series value of the fourth effective rainfall are the X axis, the Y axis, and the first boundary line. And a first region surrounded by the second boundary line and the third boundary line, a second region surrounded by the first boundary line, the second boundary line and the third boundary line, the first A third region that exceeds a boundary line and is closer to the X axis than the second boundary line; a fourth region that exceeds the second boundary line and is closer to the Y axis than the first boundary line; A slope characterized by determining which area of the fifth area that exceeds both the boundary line and the second boundary line is present, and predicting the occurrence of slope failure from the total of the discrimination results at each time Collapse prediction method. In the slope failure prediction method according to claim 2,
For a plurality of combinations of a long half-life and a short half-life, a determination XY coordinate plane in which the first to third boundary lines are set is obtained,
For all of the plurality of determination XY coordinate planes, the measurement points including the corresponding time series value of the third effective rainfall and the time series value of the fourth effective rainfall are the X axis, the Y axis, and the first boundary line. And a first region surrounded by the second boundary line and the third boundary line, a second region surrounded by the first boundary line, the second boundary line and the third boundary line, the first A third region that exceeds a boundary line and is closer to the X axis than the second boundary line; a fourth region that exceeds the second boundary line and is closer to the Y axis than the first boundary line; It is determined in which region of the fifth region that exceeds both the boundary line and the second boundary line, and the measurement point is one of the second to fifth regions in at least one determination XY coordinate plane. A slope failure prediction method characterized by predicting that there is a high probability of slope failure when located in the area. In the slope failure prediction method according to claim 2,
For a plurality of combinations of a long half-life and a short half-life, a determination XY coordinate plane in which the first to third boundary lines are set is obtained,
For all of the plurality of determination XY coordinate planes, the closest measurement point among the measurement points consisting of the corresponding time series value of the third effective rainfall and the time series value of the fourth effective rainfall is the X axis and the Y axis. And a first region surrounded by the first boundary line, the second boundary line, and the third boundary line, and a first region surrounded by the first boundary line, the second boundary line, and the third boundary line. 2 region, the first boundary line is exceeded, and the third region on the X axis side than the second boundary line, the second boundary line is exceeded, and the second boundary line is exceeded on the Y axis side from the first boundary line. 4 region, determine which region is located in the fifth region that exceeds both the first boundary line and the second boundary line,
A slope failure prediction method characterized by creating a two-dimensional determination diagram in which a symbol representing the result is plotted on a two-dimensional coordinate plane having a long half-life on the horizontal axis and a short half-life on the vertical axis. In the slope failure prediction method according to claim 1,
A plurality of time series values of the first effective rainfall and time series values of the second effective rainfall obtained at the same time obtained from rainfall data of past rainfall events at the target point are set as an X value and a Y value, respectively. Plot the points on the XY coordinate plane for determination,
When a point at a certain time among the plurality of points is a point of interest, a point having a larger Y value than the point of interest does not exist among points having the same X value as the point of interest, and than the point of interest If neither the point having the same X value as the target point or the point having a larger Y value than the target point exists among the points having a large X value, an operation of extracting the target point as a third boundary point is performed. repetition,
For the determination, a line connecting all the extracted third boundary points in ascending order of the X value is a third boundary line representing the maximum value of the effective rainfall due to the synergistic effect of the first effective rainfall and the second effective rainfall. Set to XY coordinate plane,
In the determination XY coordinate plane, the nearest measurement point among the measurement points is in a region surrounded by the X axis, the Y axis, the first boundary line, the second boundary line, and the third boundary line. A slope failure prediction method, comprising: calculating a rainfall amount per unit time when the next measurement point exceeds the first to third boundary lines when positioned. In the slope failure prediction method according to claim 6,
For a plurality of combinations of a long half-life and a short half-life, a determination XY coordinate plane in which the first to third boundary lines are set is obtained,
For all of the plurality of determination XY coordinate planes, the nearest measurement point among the measurement points consisting of the corresponding time series value of the third effective rainfall and the time series value of the fourth effective rainfall is the X axis and the Y axis. When located in a region surrounded by the first boundary line, the second boundary line, and the third boundary line, the next measurement point per unit time that exceeds the first to third boundary lines Calculate rainfall,
A slope failure prediction method, comprising: creating a two-dimensional judgment chart in which the rainfall per unit time is plotted on a two-dimensional coordinate plane with the horizontal axis representing the long half-life and the vertical axis representing the short half-life. A slope failure prediction device that predicts the occurrence of slope failure due to a rain event at a target location,
a) Time series of effective rainfall up to the present time, which is the rainfall index indicating the sustainability of the influence of past rainfall events on the accumulated rainfall, with the half-life, which is the time until the accumulated rainfall is halved at a certain time, as a parameter A past effective rainfall calculation means for calculating a value for each of a predetermined long half-life and a short half-life shorter than the long half-life, using rainfall data of past rainfall events at the target point;
b) A time series value of the first effective rainfall is plotted on an XY coordinate plane with the first effective rainfall with the long half-life as a parameter on the X-axis and the second effective rainfall with the short half-life on the parameter as the Y-axis. A determination XY coordinate plane is created by setting a first boundary line passing through the maximum value and parallel to the Y axis and a second boundary line passing through the maximum value of the time series value of the second effective rainfall and parallel to the X axis. A determination XY coordinate plane creating means;
c) From the rainfall data of the rainfall event that is the target of slope failure occurrence, the time series value of the third effective rainfall with the long half-life as a parameter and the time series of the fourth effective rainfall with the short half-life as a parameter. A prediction effective rainfall amount calculating means for sequentially calculating a value along with the progress of the prediction target rain event;
d) When the measurement point is the point where the time series value of the third effective rainfall and the time series value of the fourth effective rainfall at the same time are the X value and the Y value, respectively, the measurement point is the determination XY Determining means for determining that a slope failure may occur when located in a region exceeding the first boundary line in a coordinate plane, or when located in a region exceeding the second boundary line. Slope failure prediction device. In the slope failure prediction apparatus according to claim 8,
The determination XY coordinate plane creating means uses the determination XY coordinates as a plurality of points having the time series value of the first effective rainfall and the time series value of the second effective rainfall at the same time as the X value and the Y value, respectively. When plotting on a plane and setting a point at a certain time among the plurality of points as a point of interest, a point having a larger Y value than the point of interest does not exist among points having the same X value as the point of interest; and If neither the point having the same Y value as the point of interest nor the point having the Y value larger than the point of interest exists among the points having an X value larger than the point of interest, the point of interest is designated as the third boundary point. The line that connects all the extracted third boundary points in ascending order of the X value represents the maximum value of the effective rainfall due to the synergistic effect of the first effective rainfall and the second effective rainfall. Set as 3 boundary lines on the XY coordinate plane for determination,
The slope failure prediction apparatus, wherein the judgment means judges that slope failure may occur when the measurement point is located in a region exceeding the third boundary line on the judgment XY coordinate plane. . In the slope failure prediction device according to claim 9,
The historical effective rainfall calculation means calculates time series values of the effective rainfall for three or more half-lives having different lengths,
The determination XY coordinate plane creating means forms a plurality of combinations of long half-life and short half-life, respectively, by selecting two types of half-life from the three or more types of half-life. ~ Find the XY coordinate plane for determination with the third boundary set,
The determination means has a measurement point consisting of a time series value of the third effective rainfall and a time series value of the fourth effective rainfall for all of the plurality of determination XY coordinate planes, the X axis and the Y axis. A first region surrounded by the first boundary line, the second boundary line, and the third boundary line; a second region surrounded by the first boundary line, the second boundary line, and the third boundary line; A region that exceeds the first boundary line and is a third region on the X-axis side from the second boundary line; a fourth region that exceeds the second boundary line and that is on the Y-axis side from the first boundary line; It is determined which region is located in the fifth region that exceeds both the region, the first boundary line, and the second boundary line, and the occurrence of slope failure is predicted from the total of the determination results at each time Slope failure prediction device characterized by In the slope failure prediction device according to claim 9,
The historical effective rainfall calculation means calculates time series values of the effective rainfall for three or more half-lives having different lengths,
The determination XY coordinate plane creating means forms a plurality of combinations of long half-life and short half-life, respectively, by selecting two types of half-life from the three or more types of half-life. ~ Find the XY coordinate plane for determination with the third boundary set,
The determination means has a measurement point consisting of a time series value of the third effective rainfall and a time series value of the fourth effective rainfall for all of the plurality of determination XY coordinate planes, the X axis and the Y axis. A first region surrounded by the first boundary line, the second boundary line, and the third boundary line; a second region surrounded by the first boundary line, the second boundary line, and the third boundary line; A region that exceeds the first boundary line and is a third region on the X-axis side from the second boundary line; a fourth region that exceeds the second boundary line and that is on the Y-axis side from the first boundary line; It is determined whether the region is located in the fifth region that exceeds both the region, the first boundary line, and the second boundary line, and the measurement points are second to second in at least one determination XY coordinate plane. A slope characterized by predicting that there is a high probability of slope failure when located in any of the five regions Corrupted prediction device. In the slope failure prediction device according to claim 9,
The determination XY coordinate plane creating means obtains a determination XY coordinate plane in which first to third boundary lines are set for a plurality of combinations of long half-life and short half-life,
For all of the plurality of determination XY coordinate planes, the closest measurement point among the measurement points consisting of the corresponding time series value of the third effective rainfall and the time series value of the fourth effective rainfall is the X axis and the Y axis. And a first region surrounded by the first boundary line, the second boundary line, and the third boundary line, and a first region surrounded by the first boundary line, the second boundary line, and the third boundary line. 2 region, the first boundary line is exceeded, and the third region on the X axis side than the second boundary line, the second boundary line is exceeded, and the second boundary line is exceeded on the Y axis side from the first boundary line. 4 region, determine which region is located in the fifth region that exceeds both the first boundary line and the second boundary line,
A slope having a two-dimensional determination diagram creating means for creating a two-dimensional determination diagram in which a symbol representing the result is plotted on a two-dimensional coordinate plane with a long half-life on the horizontal axis and a short half-life on the vertical axis Collapse prediction device. In the slope failure prediction apparatus according to claim 8,
The determination XY coordinate plane creating means obtains the time series value of the first effective rainfall and the time series value of the second effective rainfall at the same time obtained from the rainfall data of past rainfall events at the target point. A plurality of points each having an X value and a Y value are plotted on the XY coordinate plane for determination, and when a point at a certain time among the plurality of points is a point of interest, the point of interest and the X value are the same. If there is no point having a larger Y value than the point of interest, and a point having a larger X value than the point of interest, a point having the same Y value as the point of interest and a point having a larger Y value than the point of interest If none of these exist, the operation of extracting the target point as a third boundary point is repeated, and a line connecting all the extracted third boundary points in ascending order of the X value is represented by the first effective rainfall and Maximum effective rainfall due to the synergistic effect of the second effective rainfall It represents the set determination XY coordinate plane as a third boundary line,
In the determination XY coordinate plane, the nearest measurement point among the measurement points is in a region surrounded by the X axis, the Y axis, the first boundary line, the second boundary line, and the third boundary line. A slope failure prediction device comprising: an excess rainfall calculating means for calculating a rainfall per unit time when the next measurement point exceeds the first to third boundary lines when positioned. In the slope failure prediction device according to claim 13,
The determination XY coordinate plane creating means obtains a determination XY coordinate plane in which first to third boundary lines are set for a plurality of combinations of long half-life and short half-life,
The excess rainfall calculation means, for all of the plurality of determination XY coordinate planes, of the measurement points including the corresponding time series value of the third effective rainfall and the time series value of the fourth effective rainfall, When located in a region surrounded by the X axis, the Y axis, the first boundary line, the second boundary line, and the third boundary line, the next measurement point exceeds the first to third boundary lines. Calculate the amount of rainfall per unit time,
It has a two-dimensional determination diagram creating means for creating a two-dimensional determination diagram in which the rainfall per unit time is plotted on a two-dimensional coordinate plane in which the horizontal axis is the long half-life and the vertical axis is the short half-life. To predict slope failure.