Method for constructing a water level prediction model, program for constructing a water level prediction model, method for outputting water level prediction data, program for outputting water level prediction data, method for interpolating test data, and method for calculating the accuracy of water level predictions.

Abstract

This disclosure aims to improve the accuracy of water level prediction in tidal zones to the same level as in non-tidal zones, using a water level prediction model. [Solution] In this disclosure, the training data extraction unit 4 extracts, as training data necessary for learning the water level at a water level prediction point in the tidal zone, tidal zone water level data EW, at least one of the upstream water level data UW and rainfall data RD, and pseudo-tidal level data that shows a pseudo-tidal level at the water level prediction point at that time, in the past before that time, and in the future before that time. The water level prediction model construction unit 5 constructs a water level prediction model M for predicting the water level at a water level prediction point in the tidal zone using the tidal zone water level data EW, at least one of the upstream water level data UW and rainfall data RD, and the pseudo-tidal level data.

Description

[Technical Field]

[0001] This disclosure relates to a technology for learning and predicting water levels at water level prediction points in a tidal zone. [Background technology]

[0002] A technique for learning and predicting water levels at water level prediction points in a river area using machine learning models such as neural networks is disclosed in Patent Document 1, etc.

[0003] First, a water level prediction model is constructed using training data. Specifically, as training data necessary to learn the water level at a given time at a water level prediction point, at least one of the following is extracted: (1) water level data showing the water level at the time and at least one of the past time at at least one of the water level prediction point and upstream of the water level prediction point; and (2) rainfall data showing the rainfall at the time, at least one of the past time and at least one of the future time at at least one of the water level prediction point and upstream of the water level prediction point. Then, a water level prediction model is constructed to predict the water level at the water level prediction point using at least one of the water level data and rainfall data.

[0004] Next, the water level at the prediction point is predicted using the water level prediction model constructed with training data. Specifically, as test data necessary to predict the water level at a given time at the prediction point, at least one of the following is input into the water level prediction model: (1) water level data showing the water level at the time and at least one of the past times at at least one of the prediction point and upstream of the prediction point, and (2) rainfall data showing the rainfall at the time, at least one of the past times and at least one of the future times at at least one of the prediction point and upstream of the prediction point. The water level prediction model then outputs the predicted water level data for that time at the prediction point. [Prior art documents] [Patent Documents]

[0005] [Patent Document 1] Japanese Patent Publication No. 2019-095240 [Overview of the project] [Problems that the invention aims to solve]

[0006] Figure 1 illustrates the challenges of conventional water level prediction processing. The upper part of Figure 1 shows the temporal changes in water level and rainfall in non-tidal zones, which are not affected by low and high tides. It can be seen that the water level in non-tidal zones rises in conjunction with an increase in rainfall in those zones. The lower part of Figure 1 shows the temporal changes in water level and rainfall in tidal zones, which are affected by low and high tides. It can be seen that even if the rainfall in tidal zones does not increase, the water level in those zones repeatedly rises and falls.

[0007] Therefore, unlike in non-tidal areas, in tidal areas, even if the water level prediction model for non-tidal areas described in Patent Document 1 is used, it is not possible to improve the accuracy of water level prediction.

[0008] Therefore, in order to solve the aforementioned problems, this disclosure aims to improve the accuracy of water level prediction in tidal zones to the same level as in non-tidal zones by using a water level prediction model. [Means for solving the problem]

[0009] To solve the aforementioned problem, the following are extracted as training data necessary for learning the water level at a water level prediction point in the tidal zone: tidal zone water level data, at least one of upstream water level data and rainfall data, and pseudo-tidal level data showing a pseudo-tidal level at the water level prediction point at that time, in the past, and in the future. Here, the tidal level data shows the observed or predicted tidal level at the tidal level observation point or tidal level prediction point, while the pseudo-tidal level data shows a pseudo-tidal level (based on the river mouth shape of the water level prediction point) at a water level prediction point upstream of the tidal level observation point or tidal level prediction point. Then, a water level prediction model for predicting the water level at a water level prediction point in the tidal zone is constructed using the tidal zone water level data, at least one of upstream water level data and rainfall data, and the pseudo-tidal level data.

[0010] Specifically, this disclosure includes a tidal zone data extraction step which extracts, as training data necessary for learning the water level at a water level prediction point in the tidal zone, (1) tidal zone water level data showing at least one of the water levels at the water level prediction point at that time and in the past before that time, and (2) pseudo-tidal level data showing at least one of the pseudo-tidal levels at the water level prediction point located upstream of the tidal level observation point or the water level prediction point, at least one of the pseudo-tidal levels at that time, in the past before that time and in the future before that time, and as training data, (3) data showing at least one of the water levels at a point upstream of the water level prediction point at that time and in the past before that time The method for constructing a water level prediction model is characterized by comprising: an upstream water level data extraction step of extracting at least one of the following: (4) upstream water level data and rainfall data indicating rainfall at at least one of the following: rainfall at the water level prediction point and the water basin upstream of the water level prediction point; and a water level prediction model construction step of constructing a water level prediction model for predicting the water level at the water level prediction point located in the tidal zone using the tidal zone water level data, the pseudo-tidal level data, and at least one of the upstream water level data and the rainfall data.

[0011] With this configuration, it is possible to construct a water level prediction model that improves the accuracy of water level predictions to the same level as in non-tidal areas, by using pseudo-tidal level data in tidal zones.

[0012] Furthermore, this disclosure provides a method for constructing a water level prediction model, characterized in that it further includes a pseudo-tide level data conversion step before the tidal zone data extraction step, in which tidal level data indicating observed or predicted tidal levels at the tidal level observation point or the tidal level prediction point is used as input data, the tidal zone water level data (however, limited to data within a period not affected by rainfall) is used as training data, and the pseudo-tide level data is used as output data.

[0013] This configuration allows for the conversion of tidal level data into pseudo-tidal level data (based on the river mouth shape at the water level prediction point) using machine learning models such as neural networks.

[0014] Furthermore, this disclosure provides a method for constructing a water level prediction model, characterized in that the tidal zone data extraction step and the upstream zone data extraction step subtract the pseudo-tidal zone data from the tidal zone water level data, convert the tidal zone water level data into deemed water level data from which the influence of tides has been removed, extract the training data necessary for learning the water level of the water level prediction point during the water level peak period of the deemed water level data, and exclude the water level of the water level prediction point outside of the water level peak period of the deemed water level data from the learning target.

[0015] With this configuration, when extracting training data necessary for learning water levels in tidal zones, water level peaks caused by tidal influences are excluded from the learning target, and water level peaks caused by rainfall are extracted as the learning target, thereby improving the learning accuracy of water level peaks caused by rainfall.

[0016] Furthermore, this disclosure is a method for constructing a water level prediction model, characterized in that the upstream data extraction step extracts at least one of the following as rainfall data: (1) instantaneous rainfall data showing the instantaneous rainfall at at least one of the time, past time, and future time at at least one of the water level prediction point and the water basin upstream of the water level prediction point; (2) total rainfall data showing the total rainfall from the start of rainfall to at least one of the time, past time, and future time at at least one of the time when rainfall does not start (however, it is reset to 0 when a predetermined time has elapsed from the time when no rainfall starts); and (3) average rainfall data showing the average rainfall within a predetermined average period at at least one of the time, past time, and future time at at least one of the time when rainfall does not start.

[0017] With this configuration, it is possible to construct a water level prediction model that improves the accuracy of water level prediction by using total rainfall data as the rainfall data and considering the time required from rainfall in the tidal zone or upstream area to inflow into the tidal zone or upstream area. Furthermore, it is possible to construct a water level prediction model that improves the accuracy of water level prediction by using average rainfall data as the rainfall data, unlike with total rainfall data, without abruptly resetting the data to zero.

[0018] Furthermore, this disclosure provides a method for constructing a water level prediction model, characterized in that it further includes a water level prediction model selection step after the water level prediction model construction step, in which verification data for verifying the water level prediction model is input into multiple types of water level prediction models constructed using multiple types of training data, and a water level prediction model is selected that performs water level predictions that better meet the issuance criteria.

[0019] According to this configuration, an optimal water level prediction model can be selected from a plurality of types of water level prediction models constructed using a plurality of types of training data, so that the generalization performance near the water level peak (particularly, near the water level peak that serves as a criterion for issuing evacuation advisories, etc.) is improved.

[0020] Moreover, the present disclosure is a water level prediction model construction program for causing a computer to execute each processing step included in the water level prediction model construction method described above.

[0021] According to this configuration, a program having the effects described above can be provided.

[0022] Moreover, in the present disclosure, when predicting the water level at the water level prediction point in the tidal reach using the water level prediction model constructed by the water level prediction model construction method described above, as test data necessary for predicting the water level at a certain time at the water level prediction point, the tidal reach water level data, the pseudo tide level data, and at least any one of the upstream water level data and the rainfall data are input to the water level prediction model, and water level prediction data indicating the predicted water level at that time at the water level prediction point is output from the water level prediction model, and a water level prediction data output method is provided, characterized by comprising a water level prediction data output step.

[0023] According to this configuration, in the tidal reach, the prediction accuracy of the water level can be improved equivalently to that in the non-tidal reach by using a water level prediction model constructed using pseudo tide level data as well.

[0024] Moreover, the present disclosure is a water level prediction data output program for causing a computer to execute the water level prediction data output step included in the water level prediction data output method described above.

[0025] According to this configuration, a program having the effects described above can be provided.

[0026] To address further challenges, as test data, when tidal zone water level data is missing in the tidal zone, the data for the missing time in the tidal zone is interpolated using predicted water level data for the missing time, which is output after an arbitrary time after the predicted time in the tidal zone, reflecting tidal, rainfall, and actual water level conditions at a predicted time arbitrarily prior to the missing time.

[0027] Specifically, the present disclosure is a test data interpolation method characterized in that, when the test data extracted by the water level prediction data output method described above is missing tidal zone water level data at the water level prediction point, the test data interpolation step comprises the water level prediction data output at the water level prediction point, which is output at a prediction time arbitrarily before the time of the missing data, and which interpolates the tidal zone water level data at the time of the missing data at the water level prediction point at an arbitrary time after the prediction time at the water level prediction point.

[0028] With this configuration, even when test data for the water level in the tidal zone is missing, the accuracy of water level prediction in the tidal zone can be improved by interpolating the test data for the water level in the tidal zone while reflecting the tidal, rainfall, and actual water level conditions.

[0029] To address further challenges, we extract tidal zone water level data, pseudo-tidal level data, upstream water level data, past observed rainfall data, and future forecast rainfall data as verification data. We then calculate the water level prediction accuracy rate by determining the proportion of water level prediction data that falls within the error range surrounding the observed water level data.

[0030] Specifically, this disclosure provides a verification data extraction step for extracting verification data for verifying the water level prediction model constructed by the water level prediction model construction method described above, which includes the tidal zone water level data, the pseudo-tidal zone data, the upstream water level data, and past observed rainfall data and future forecast rainfall data as rainfall data, and the verification data for verifying the water level at multiple times at the water level prediction point, which includes the tidal zone water level data, the pseudo-tidal zone data, the upstream water level data, the past observed rainfall data and The method for calculating water level prediction accuracy is characterized by comprising: a water level prediction data output step of inputting the aforementioned future forecast rainfall data into the water level prediction model and outputting water level prediction data from the water level prediction model that shows the predicted water levels at the water level prediction point for the multiple time periods; and a water level prediction accuracy calculation step of calculating the proportion of the water level prediction data for the multiple time periods at the water level prediction point that falls within the error range around the water level observation data at the water level prediction point for the multiple time periods as the prediction accuracy of the water level at the water level prediction point.

[0031] With this configuration, since future observed rainfall data cannot be extracted as test data during the prediction phase, even when future forecast rainfall data is extracted, the reliability of the water level prediction data can be visualized by calculating the water level prediction accuracy rate during the verification phase.

[0032] Furthermore, the inventions disclosed above can be combined as much as possible. [Effects of the Invention]

[0033] Thus, this disclosure demonstrates that, in tidal zones, the water level prediction accuracy can be improved to the same level as in non-tidal zones by using a water level prediction model. [Brief explanation of the drawing]

[0034] [Figure 1] This figure shows the challenges of conventional water level prediction processing techniques. [Figure 2] This figure shows the configuration of the water level prediction device disclosed herein. [Figure 3]This figure shows the procedure for water level prediction processing in this disclosure. [Figure 4] This figure shows the concept of pseudo-tide level data as disclosed in this disclosure. [Figure 5] This figure shows the conversion process to pseudo-tide level data as disclosed herein. [Figure 6] This figure shows the learning target for the tidal zone water level data in this disclosure. [Figure 7] This figure shows the total rainfall data and average rainfall data disclosed herein. [Figure 8] This figure shows examples of prior art and the water level prediction process described herein. [Figure 9] This figure shows examples of prior art and the water level prediction process described herein. [Figure 10] This figure shows the configuration of the water level prediction device disclosed herein. [Figure 11] This diagram shows the procedure for selecting a water level prediction model in this disclosure. [Figure 12] This figure shows a specific example of the synthesis process of training data as described in this disclosure. [Figure 13] This figure shows a specific example of the selection process for the water level prediction model disclosed herein. [Figure 14] This figure shows the configuration of the water level prediction device disclosed herein. [Figure 15] This figure shows the procedure for the test data interpolation process described herein. [Figure 16] This figure shows a specific example of the water level test data interpolation process described herein. [Figure 17] This figure shows a specific example of the rainfall test data interpolation process described herein. [Figure 18] This figure shows a specific example of the tidal level test data interpolation process described herein. [Figure 19] This figure shows the configuration of the water level prediction device disclosed herein. [Figure 20] This diagram shows the procedure for calculating the accuracy rate of water level predictions as disclosed herein. [Figure 21] This figure shows a specific example of the water level prediction error range setting process disclosed herein. [Figure 22]This figure shows a specific example of the water level prediction accuracy calculation process disclosed herein. [Modes for carrying out the invention]

[0035] Embodiments of the present disclosure will be described with reference to the attached drawings. The embodiments described below are examples of the implementation of the present disclosure, and the present disclosure is not limited to these embodiments.

[0036] (Outline of the water level prediction process in this disclosure) Figure 2 shows the configuration of the water level prediction device disclosed herein. Figure 3 shows the procedure for the water level prediction process disclosed herein. A technology for learning and predicting water levels at water level prediction points in a tidal zone using machine learning models such as neural networks is disclosed in the following description of this disclosure.

[0037] The water level prediction device W is equipped with a pseudo-tide level data conversion unit 1, a deemed water level data conversion unit 2, a rainfall data conversion unit 3, a training data extraction unit 4, and a water level prediction model construction unit 5 to execute the water level prediction model construction method (steps S1 to S6). The water level prediction model M is constructed, and the water level prediction model construction program (steps S1 to S6) shown in Figure 3 is installed on a computer to realize this.

[0038] The water level prediction device W is equipped with a pseudo-tide level data conversion unit 1, a rainfall data conversion unit 3, a test data extraction unit 6, and a water level prediction data output unit 7 to execute the water level prediction data output method (steps S7 to S10). It can be realized by applying the water level prediction model M and installing the water level prediction data output program (steps S7 to S10) shown in Figure 3 onto a computer.

[0039] In the following, we will first describe the training phase processing (steps S1 to S6) of the water level prediction model M disclosed herein, then describe the prediction phase processing (steps S7 to S10) of the water level prediction model M disclosed herein, and finally describe examples of the prior art and the water level prediction processing disclosed herein.

[0040] The training data extraction unit 4 extracts the following as training data necessary for learning the water level at a water level prediction point in the tidal zone at a given time: (1) Tidal zone water level data EW showing at least one of the water levels at the water level prediction point at that time and in the past before that time; and (2) Pseudo-tidal level data showing at least one of the pseudo-tidal levels at the water level observation point or a water level prediction point upstream of the tidal level prediction point at that time, in the past before that time, and in the future before that time (step S5).

[0041] The training data extraction unit 4 then extracts at least one of the following: (3) upstream water level data UW, which shows the water level at a point upstream of the water level prediction point at the time in question and at least one of the past times prior to that time; and (4) rainfall data RD, which shows the rainfall at a point upstream of the water level prediction point and at least one of the past times prior to that time and at least one of the future times prior to that time. (Step S5). Here, for the rainfall data RD, "upstream watershed" may also include "upstream points only".

[0042] Here, it is desirable for the training data extraction unit 4 to extract rainfall and water level data from periods with a high correlation to the water level during floods at the water level prediction points in the tidal zone, and to discard rainfall and water level data from periods with a low correlation.

[0043] For example, the training data extraction unit 4 may, as training data, (1) set a water level peak discrimination threshold that rejects periods of no water level fluctuation and then select a period for extracting water level data; (2) set a water level peak extraction width that rejects periods of no water level fluctuation and then select a period for extracting water level data; (3) set a predetermined water level period for extracting periods of rising and falling water levels and then select a period for extracting water level data; (4) set a predetermined rainfall period that takes into account the time it takes for rainfall to reach the water level prediction point and then select a period for extracting rainfall data; or (5) set a predetermined cumulative period that takes into account the time it takes for rainfall to flow from the soil or reservoir into the river and then select a period for extracting rainfall data (see Patent Document 1).

[0044] The water level prediction model construction unit 5 constructs a water level prediction model M for predicting water levels at water level prediction points in the tidal zone using tidal zone water level data EW, pseudo-tidal zone data, and at least one of the upstream zone water level data UW and rainfall data RD (step S6).

[0045] In this way, by using pseudo-tidal level data in the tidal zone, it is possible to construct a water level prediction model M that improves the accuracy of water level prediction to the same level as in the non-tidal zone.

[0046] The test data extraction unit 6 extracts the following test data necessary for predicting the water level at a water level prediction point in the tidal zone at a given time: (1) Tidal zone water level data EW showing at least one of the water levels at the water level prediction point at that time and in the past; and (2) Pseudo-tidal level data showing at least one of the pseudo-tidal levels at the water level observation point or a water level prediction point upstream of the tidal level prediction point at that time, in the past and in the future (step S9).

[0047] The test data extraction unit 6 then extracts at least one of the following: (3) upstream water level data UW, which shows the water level at a point upstream of the water level prediction point at the time in question and at least one of the water levels in the past before that time; and (4) rainfall data RD, which shows the rainfall at a point upstream of the water level prediction point and at least one of the rainfall at the time in question, at least one of the rain levels in the past before that time and at least one of the rain levels in the future before that time (step S9). Here, for the rainfall data RD, "upstream watershed" may also include "upstream point only".

[0048] The water level prediction data output unit 7 inputs the tidal zone water level data EW, pseudo-tidal zone data, and at least one of the upstream water level data UW and rainfall data RD as test data to the water level prediction model M (step S9), and outputs water level prediction data P from the water level prediction model M that shows the predicted water level at the water level prediction point in the tidal zone at that time (step S10).

[0049] In this way, in tidal zones, the water level prediction accuracy can be improved to the same level as in non-tidal zones by using a water level prediction model M constructed with pseudo-tidal level data.

[0050] (Details of the water level prediction process in this disclosure) The concept of the pseudo-tide level data described in this disclosure is shown in Figure 4. As the river flows from the upstream point US through the water level prediction point WP to the tide level observation / prediction point TO, the river mouth width changes from the width w' at the water level prediction point WP to the width w at the tide level observation / prediction point TO, the river mouth depth changes from the depth d' at the water level prediction point WP to the depth d at the tide level observation / prediction point TO, the distance from the water level prediction point WP to the tide level observation / prediction point TO is l, and the elevation at the water level prediction point WP is h.

[0051] Here, the tidal level data TD represents the observed / predicted tidal level at the tidal level observation / prediction point TO, while the pseudo-tidal level data represents the pseudo-tidal level at the water level prediction point WP, which is upstream of the tidal level observation / prediction point TO (based on the river mouth shape of the water level prediction point WP). In other words, the period p' of the pseudo-tidal level is approximately equal to the period p of the observed / predicted tidal level, the amplitude a' of the pseudo-tidal level is determined by the amplitude a of the observed / predicted tidal level and the river mouth shape of the water level prediction point WP, and the delay t of the pseudo-tidal level relative to the observed / predicted tidal level is determined by the distance l from the tidal level observation / prediction point TO to the water level prediction point WP and the tidal level propagation speed. However, the period p' of the pseudo-tide level is not necessarily equal to the period p of the observed / predicted tide level, the amplitude a' of the pseudo-tide level is not simply determined by the amplitude a of the observed / predicted tide level and the shape of the river mouth at the water level prediction point WP (the lower limit of the amplitude a' may be determined depending on the elevation h of the riverbed at the water level prediction point WP), and the propagation speed of the tide level is not trivial depending on the shape of the river mouth at the water level prediction point WP.

[0052] Figure 5 shows the conversion process to pseudo-tide level data according to this disclosure. In the training phase of the water level prediction model M, the pseudo-tide level data conversion unit 1 takes tide level data TD, which shows the observed or predicted tide level at a tide level observation point or tide level prediction point, as input data, and tidal area water level data EW (however, limited to data within a period not affected by rainfall) as training data, and outputs pseudo-tide level data (step S1). In the prediction phase of the water level prediction model M, the pseudo-tide level data conversion unit 1 takes tide level data TD, which shows the observed or predicted tide level at a tide level observation point or tide level prediction point, as input data, and outputs pseudo-tide level data (step S7).

[0053] Here, it is desirable for the pseudo-tide level data conversion unit 1 to discard tidal area water level data EW within a predetermined period (for example, within the rainy season) or within a predetermined time before and after the time of rainfall (for example, within 12 hours) as pseudo-tide level data. Furthermore, the pseudo-tide level data conversion unit 1 may output tidal level data TD as pseudo-tide level data at tidal level observation points or tidal level prediction points.

[0054] In this way, machine learning models such as neural networks can be used to convert tidal level data (TD) into pseudo-tidal level data (based on the river mouth shape of the water level prediction point).

[0055] Figure 6 shows the learning targets for the tidal zone water level data of this disclosure. During the training phase of the water level prediction model M, the deemed water level data conversion unit 2 subtracts the pseudo-tidal level data (indicated by the left arrow) from the tidal zone water level data EW (indicated by the left arrow) to convert the tidal zone water level data EW into deemed water level data (indicated by the left arrow) from which the influence of tides has been removed (step S2). Then, the training data extraction unit 4 extracts the training data necessary for learning the water level of the water level prediction points during the water level peak period of the deemed water level data, and excludes the water level of the water level prediction points outside of the water level peak period of the deemed water level data from the learning targets (step S4).

[0056] Here, the tidal zone water level data EW includes water level peaks due to rainfall (indicated by the right arrow) and may also include water level peaks due to tidal influences. Therefore, if the water level peaks of the tidal zone water level data EW are used as the learning target, then water level peaks due to rainfall and water level peaks due to tidal influences may also be used as the learning target. On the other hand, the estimated water level data includes water level peaks due to rainfall but does not include water level peaks due to tidal influences. Therefore, if the water level peaks of the estimated water level data are used as the learning target, then water level peaks due to rainfall will be used as the learning target, but water level peaks due to tidal influences will not be used as the learning target.

[0057] In this way, when extracting training data necessary for learning water levels in tidal zones, it is possible to exclude water level peaks caused by tidal influences from the learning target and extract water level peaks caused by rainfall as the learning target, thereby improving the learning accuracy of water level peaks caused by rainfall.

[0058] The total rainfall data and average rainfall data of this disclosure are shown in Figure 7. During the training and prediction phases of the water level prediction model M, the rainfall data conversion unit 3 generates at least one of the following as rainfall data RD: (1) instantaneous rainfall data showing the instantaneous rainfall at at least one of the time, the time in the past, and the time in the future at at least one of the water level prediction point and the water basin upstream of the water level prediction point; (2) total rainfall data showing the total rainfall from the start of rainfall to at least one of the time, the time in the past, and the time in the future at at least one of the time in the past, and the time in the future at at least one of the time in the past, and the time in the future (however, it is reset to 0 when a predetermined time has elapsed from the time when no rain falls); and (3) average rainfall data showing the average rainfall within a predetermined averaging period at at least one of the time, the time in the past, and the time in the future at at least one of the time in the past, and the time in the future at at least one of the time in the past, and the time in the future (steps S3, S8). Here, the instantaneous rainfall data, total rainfall data, and average rainfall data may be either the "spatial average value for the entire upstream basin" or the "spatial sum for the entire upstream basin."

[0059] In the upper part of Figure 7, total rainfall data is extracted, showing the total rainfall from the start of rainfall to the time of concern (instantaneous rainfall data is indicated by a left arrow, and total rainfall data is indicated by a right arrow). However, the total rainfall data is reset to 0 after a predetermined time has elapsed from the start of no rainfall. Here, the "predetermined elapsed time" should be set to construct a water level prediction model M with improved water level prediction accuracy. In the lower part of Figure 7, average rainfall data is extracted, showing the average rainfall within a predetermined averaging period up to the time of concern (instantaneous rainfall data is indicated by a left arrow, and average rainfall data is indicated by a right arrow). Here, the "predetermined averaging period" should be set to construct a water level prediction model M with improved water level prediction accuracy. Note that since the total rainfall data is greater than the average rainfall data, it is desirable to standardize either the total rainfall data or the average rainfall data so that the water level prediction data P based on the total rainfall data is approximately equal to the water level prediction data P based on the average rainfall data.

[0060] Thus, by using total rainfall data as the rainfall data RD, a water level prediction model M can be constructed that improves the accuracy of water level prediction by considering the time required from rainfall in the tidal zone or upstream area to inflow into the tidal zone or upstream area. Furthermore, by using average rainfall data as the rainfall data RD, a water level prediction model M can be constructed that improves the accuracy of water level prediction without abruptly resetting the data to zero, unlike with total rainfall data.

[0061] Examples of conventional technology and the water level prediction processing described herein are shown in Figures 8 and 9. In conventional technology, a water level prediction model M is constructed using only tidal zone water level data EW, upstream water level data UW, and rainfall data RD. In this disclosure, a water level prediction model M is constructed using not only tidal zone water level data EW, upstream water level data UW, and rainfall data RD, but also pseudo-tidal level data.

[0062] Figures 8 and 9 show the temporal changes in actual water level data (indicated by the left arrow), water level forecast data P (water level forecast results for 1 and 6 hours later, indicated by the left arrow), and rainfall data RD (indicated by the right arrow) during rainfall. It can be seen that, in the conventional technology, the water level forecast data P differs significantly from the actual water level data (because the tidal level data TD and, consequently, the pseudo-tidal level data have not been learned, and the water level peaks due to rainfall have not been properly learned). In particular, the water level forecast result for 1 hour later differs from the actual water level data in that it has not reached the planned high water level. On the other hand, in this disclosure, the water level forecast data P is in close agreement with the actual water level data (because the tidal level data TD and, consequently, the pseudo-tidal level data have been learned, and the water level peaks due to rainfall have been properly learned). In particular, the water level forecast result for 1 hour later reaches the planned high water level compared to the actual water level data, making it possible to issue evacuation guidance orders.

[0063] (Procedure for selecting a water level prediction model in this disclosure) The configuration of the water level prediction device disclosed herein (particularly the water level prediction model selection process) is also shown in Figure 10. The procedure for the water level prediction model selection process disclosed herein is shown in Figure 11. The water level prediction device W also includes a verification data extraction unit 8 and a water level prediction model selection unit 9, and the water level prediction model selection program shown in Figure 11 can also be realized by installing it on a computer.

[0064] The training data extraction unit 4 extracts multiple types of training data (step S11, see Figure 12). The water level prediction model construction unit 5 constructs multiple types of water level prediction models M1, ..., Mn using the multiple types of training data (step S11).

[0065] The verification data extraction unit 8 extracts the following as verification data necessary for verifying the water level at a water level prediction point in the tidal zone at a given time: (1) Tidal zone water level data EW showing at least one of the water levels at the water level prediction point at that time and in the past before that time; and (2) Pseudo-tidal level data showing at least one of the pseudo-tidal levels at the water level observation point or a water level prediction point upstream of the tidal level prediction point at that time, in the past before that time, and in the future before that time (step S12).

[0066] The verification data extraction unit 8 then extracts at least one of the following: (3) upstream water level data UW, which shows the water level at a point upstream of the water level prediction point at the time in question and at least one of the past times prior to that time; and (4) rainfall data RD, which shows the rainfall at a point upstream of the water level prediction point and at least one of the past times prior to that time and at least one of the future times prior to that time. (Step S12). Here, for the rainfall data RD, "upstream watershed" may also include "upstream points only". The extraction period for the verification data is the same as that for the training data.

[0067] The water level prediction model selection unit 9 inputs verification data for verifying the water level prediction model M into multiple types of water level prediction models M1, ..., Mn, which are constructed using multiple types of training data (step S12), and selects a water level prediction model (for example, M1) that better satisfies the issuance criteria (i.e., has a small prediction error near the water level peak) (step S13).

[0068] A specific example of the training data synthesis process described in this disclosure is shown in Figure 12. The training data extraction unit 4 (and the verification data extraction unit 8) extracts the tidal zone water level data EW, pseudo-tidal level data, and upstream water level data UW as mandatory data, and arbitrarily extracts the converted data (at least one of instantaneous rainfall data, total rainfall data, and average rainfall data) as rainfall data RD.

[0069] In the upper left column of Figure 12, only the tidal zone water level data EW, pseudo-tidal zone data, and upstream water level data UW from time t1 to time t2 are extracted as the first training data, and the converted rainfall data RD is not extracted. In the lower left column of Figure 12, the tidal zone water level data EW, pseudo-tidal zone data, upstream water level data UW, and instantaneous rainfall data are extracted as the second training data from time t1 to time t2. In the upper middle column of Figure 12, the tidal zone water level data EW, pseudo-tidal zone data, upstream water level data UW, and total rainfall data are extracted as the third training data from time t1 to time t2. In the lower middle column of Figure 12, the tidal zone water level data EW, pseudo-tidal zone data, upstream water level data UW, instantaneous rainfall data, and total rainfall data are extracted as the fourth training data from time t1 to time t2. In the upper right section of Figure 12, the fifth training data is extracted, consisting of tidal zone water level data EW, pseudo-tidal zone water level data, upstream water level data UW, and average rainfall data from time t1 to time t2. In the lower right section of Figure 12, the sixth training data is extracted, consisting of tidal zone water level data EW, pseudo-tidal zone water level data, upstream water level data UW, instantaneous rainfall data, and average rainfall data from time t1 to time t2.

[0070] A specific example of the water level prediction model selection process in this disclosure is shown in the left column of Figure 13. The water level prediction model selection unit 9 should select a water level prediction model (e.g., M1) when the predicted time of the water level rise period near the water level peak is the same as the actual time of the water level rise period. Even if the predicted time of the water level rise period near the water level peak is not the same as the actual time of the water level rise period, a water level prediction model (e.g., M1) should be selected according to the following policy.

[0071] In other words, as in Model B, the water level prediction model is considered superior when the predicted time of the water level rise period is earlier than the actual time of the water level rise period. On the other hand, as in Model A, the water level prediction model is considered inferior when the predicted time of the water level rise period is later than the actual time of the water level rise period. In this way, the possibility of delays in issuing evacuation advisories and other warnings is reduced.

[0072] A specific example of the water level prediction model selection process in this disclosure is also shown in the right column of Figure 13. The water level prediction model selection unit 9 should select a water level prediction model (e.g., M1) when the predicted water level near the water level peak (peak water level, etc.) is the same as the actual water level near the water level peak. Even if the predicted water level near the water level peak (peak water level, etc.) is not the same as the actual water level near the water level peak, a water level prediction model (e.g., M1) should be selected according to the following policy.

[0073] In other words, as in Model B, the water level prediction model is considered superior when the predicted water level near the peak is higher than the actual water level near the peak. On the other hand, as in Model A, the water level prediction model is considered inferior when the predicted water level near the peak is lower than the actual water level near the peak. In this way, the possibility of the criteria for issuing evacuation advisories not being met is reduced.

[0074] In this way, from multiple types of water level prediction models M1, ..., Mn constructed using multiple types of training data, the optimal water level prediction model M can be selected so as to enhance generalization performance near the water level peak (especially near the water level peak that serves as the basis for issuing evacuation advisories, etc.).

[0075] (Procedure for test data interpolation processing in this disclosure) The configuration of the water level prediction device of this disclosure (particularly the test data interpolation process) is also shown in Figure 14. The procedure for the test data interpolation process of this disclosure is shown in Figure 15. A specific example of the water level test data interpolation process of this disclosure is shown in Figure 16. A specific example of the rainfall test data interpolation process of this disclosure is shown in Figure 17. A specific example of the tidal level test data interpolation process of this disclosure is shown in Figure 18. The water level prediction device W also includes a test data interpolation unit 10, and the test data interpolation program shown in Figure 15 can also be realized by installing it on a computer.

[0076] The test data interpolation unit 10 interpolates the tidal zone water level data EW at the water level prediction point when it is missing data (step S21) (step S22). Here, the test data interpolation unit 10 interpolates using the water level prediction data P for the missing time at the water level prediction point, which was output by the water level prediction data output unit 7 at a prediction time arbitrarily before the missing time (step S22). The test data extraction unit 6 inputs the test data (including the tidal zone water level data EW for the missing time interpolated by the test data interpolation unit 10, the rainfall data RD for the missing time will be described later, and the tidal level data TD for the missing time will also be described later) into the water level prediction model M (step S22).

[0077] Furthermore, if the upstream water level data UW at the upstream location is missing, even if the tidal zone water level data EW at the water level prediction location is not missing, the water level prediction data P at the water level prediction location cannot be output. Therefore, the upstream water level data UW for the time of the missing data at the upstream location is interpolated with the water level prediction data P at the upstream location.

[0078] Furthermore, the arbitrary time set by the test data interpolation unit 10 is equal to the time interval at which the water level prediction data output unit 7 outputs the water level prediction data P. In other words, the rainfall forecast for a short period of time is highly accurate, and the arbitrary time set by the test data interpolation unit 10 is short.

[0079] In the first section of Figure 16, in order to predict the water level at 12:30 (30 minutes from the current time), the following test data are extracted: 3.0m (1 hour prior to the current time, at 11:00), 4.0m (30 minutes prior to the current time, at 11:30), and 5.0m (the current time, at 12:00). In sections 3-5 of Figure 16, some water level data is missing as test data.

[0080] In the second section of Figure 16, the predicted water level 30 minutes from the current time is output as follows: predicted water level data of 2.9m for 11:00 (one hour prior to the current time), predicted water level data of 3.9m for 11:30 (30 minutes prior to the current time), and predicted water level data of 4.8m for 12:00 (the current time). In the second section of Figure 16, the arbitrary time set by the test data interpolation unit 10 is 30 minutes.

[0081] In the third panel of Figure 16, the water level data for 11:00 (3.0m) and 11:30 (4.0m) are missing. Therefore, these are interpolated using the predicted water level data for 11:00 (2.9m) and 11:30 (3.9m). In the fourth panel of Figure 16, the water level data for 11:30 (4.0m) and 12:00 (5.0m) are missing. Therefore, these are interpolated using the predicted water level data for 11:30 (3.9m) and 12:00 (4.8m). In the fifth panel of Figure 16, the water level data for 11:30 (4.0m) is missing. Therefore, this is interpolated using the predicted water level data for 11:30 (3.9m).

[0082] In this way, even when test data for tidal water levels in the tidal zone are missing, the accuracy of water level prediction in the tidal zone can be improved by interpolating the test data for tidal water levels in the tidal zone while reflecting tidal, rainfall, and actual water level conditions.

[0083] Then, the missing data can be interpolated using water level prediction data P, which is output at a prediction time shortly before the missing data time (the time interval for water level prediction), reflecting tidal, rainfall, and actual water level conditions, and then output at a prediction time shortly after the prediction time (the time interval for water level prediction) in the tidal zone.

[0084] When rainfall data RD is missing at a water level prediction point (step S23), the test data interpolation unit 10 interpolates the rainfall data RD for the time of the missing data at the water level prediction point (step S24). Here, the test data interpolation unit 10 interpolates using the forecast rainfall data RF for the time of the missing data at the water level prediction point (step S24). The test data extraction unit 6 inputs the test data (including the rainfall data RD for the time of the missing data interpolated by the test data interpolation unit 10, the tidal zone water level data EW for the time of the missing data as described above, and the tidal level data TD for the time of the missing data as described below) into the water level prediction model M (step S24).

[0085] Furthermore, if rainfall data RD is missing in the upstream basin, even if rainfall data RD is available at the water level prediction point, the water level prediction data P at the water level prediction point cannot be output. Therefore, the rainfall data RD for the missing time in the upstream basin is interpolated using the forecast rainfall data RF from the upstream basin.

[0086] In the first section of Figure 17, in order to predict the water level at 12:30 (30 minutes from the current time), rainfall data of 30 mm at 11:00 (1 hour prior to the current time), rainfall data of 40 mm at 11:30 (30 minutes prior to the current time), and rainfall data of 50 mm at 12:00 (the current time) are extracted as test data. In sections 3 to 5 of Figure 17, some rainfall data is missing as test data.

[0087] In the second row of Figure 17, the forecast rainfall data RF (short-term precipitation forecast) is entered as follows: forecast rainfall data of 29 mm for 11:00 (one hour before the current time), forecast rainfall data of 39 mm for 11:30 (30 minutes before the current time), and forecast rainfall data of 48 mm for 12:00 (the current time).

[0088] In the third panel of Figure 17, the rainfall data for 11:00 (30 mm) and 11:30 (40 mm) are missing. Therefore, these are interpolated using the forecast rainfall data for 11:00 (29 mm) and 11:30 (39 mm). In the fourth panel of Figure 17, the rainfall data for 11:30 (40 mm) and 12:00 (50 mm) are missing. Therefore, these are interpolated using the forecast rainfall data for 11:30 (39 mm) and 12:00 (48 mm). In the fifth panel of Figure 17, the rainfall data for 11:30 (40 mm) is missing. Therefore, this is interpolated using the forecast rainfall data for 11:30 (39 mm).

[0089] In this way, even when rainfall test data is missing in the tidal zone and upstream areas, the accuracy of water level prediction in the tidal zone can be improved by incorporating forecast rainfall data (RF) in the tidal zone and upstream areas and then interpolating the rainfall test data in the tidal zone and upstream areas.

[0090] When the tidal level data TD at a tidal level observation point is missing (step S25), the test data interpolation unit 10 interpolates the tidal level data TD for the time the data is missing at the tidal level observation point (step S26). Here, the test data interpolation unit 10 interpolates using the astronomical tidal data AD for the time the data is missing at the tidal level observation point (step S26). The test data extraction unit 6 inputs the test data (including the tidal level data TD for the time the data is missing, which has been interpolated by the test data interpolation unit 10; the tidal area water level data EW for the time the data is missing is as described above; and the rainfall data RD for the time the data is missing is also as described above) into the water level prediction model M (step S26).

[0091] In the first section of Figure 18, in order to predict the water level at 12:30 (30 minutes from the current time), the following tidal level data are extracted as test data: 3.0m at 11:00 (1 hour prior to the current time), 4.0m at 11:30 (30 minutes prior to the current time), and 5.0m at 12:00 (the current time). In sections 3-5 of Figure 18, some tidal level data is missing as test data.

[0092] In the second row of Figure 18, the following are entered as astronomical tidal data AD (tidal forecast data): 2.9m for 11:00 (one hour before the current time), 3.9m for 11:30 (30 minutes before the current time), and 4.8m for 12:00 (the current time).

[0093] In the third panel of Figure 18, the tide level data for 11:00 (3.0m) and 11:30 (4.0m) are missing. Therefore, they are interpolated using the astronomical tidal data for 11:00 (2.9m) and 11:30 (3.9m), respectively. In the fourth panel of Figure 18, the tide level data for 11:30 (4.0m) and 12:00 (5.0m) are missing. Therefore, they are interpolated using the astronomical tidal data for 11:30 (3.9m) and 12:00 (4.8m), respectively. In the fifth panel of Figure 18, the tide level data for 11:30 (4.0m) is missing. Therefore, it is interpolated using the astronomical tidal data for 11:30 (3.9m).

[0094] In this way, even when tidal level test data is missing at tidal level observation points, the accuracy of water level prediction in tidal zones can be improved by incorporating astronomical tidal data AD at tidal level observation points and then interpolating the tidal level test data from those points.

[0095] (Procedure for calculating the accuracy rate of water level predictions in this disclosure) The configuration of the water level prediction device disclosed herein (in particular, the water level prediction accuracy calculation process) is also shown in Figure 19. The procedure for the water level prediction accuracy calculation process disclosed herein is shown in Figure 20. The water level prediction device W also includes a water level prediction error range setting unit 11 and a water level prediction accuracy calculation unit 12, and the water level prediction accuracy calculation program shown in Figure 20 can also be realized by installing it on a computer.

[0096] The training data extraction unit 4 extracts, as training data necessary for learning the water level at a given time at a water level prediction point, past observed rainfall data RO for at least one of the water level prediction point and at least one of the upstream watersheds from the water level prediction point, and rainfall data RO for "future observations" from that time onward.

[0097] The test data extraction unit 6 extracts, as test data necessary for predicting the water level at a given time at a water level prediction point, the following: RO, past observed rainfall data RO from at least one of the water level prediction point and at least one of the water basins upstream of the water level prediction point, at the time in question and at least one of the past times prior to that time; and RF, "future forecast" rainfall data from that time onward.

[0098] However, because the "future forecast" rainfall data RF is less accurate than the "future observation" rainfall data RO, the water level "prediction" data P can deviate significantly from the water level "observation" data. Therefore, this disclosure visualizes the reliability of the water level "prediction" data P.

[0099] Here, the training data is known data for water level learning and cannot be applied to the calculation of the water level prediction accuracy rate H, as it is set to approximately 100%. The test data is unknown data for water level prediction, but it is data specific to a particular rainfall situation and therefore cannot be applied to the calculation of the water level prediction accuracy rate H. On the other hand, the validation data is unknown data for validating the water level prediction model M and is not data specific to a particular rainfall situation, so it can be applied to the calculation of the water level prediction accuracy rate H.

[0100] A specific example of the water level prediction error range setting process in this disclosure is shown in Figure 21. The verification data extraction unit 8 extracts tidal zone water level data EW, pseudo-tidal level data, upstream water level data UW, past observed rainfall data RO, and future observed rainfall data RO as verification data for verifying the water level prediction model M (step S31), but does not extract future forecast rainfall data RF.

[0101] The verification data extraction unit 8 inputs, as verification data, the tidal level data EW, the pseudo-tidal level data, the upstream water level data UW, the past observed rainfall data RO, and the future observed rainfall data RO for verifying the water levels at a plurality of times at the water level prediction location to the water level prediction model M (step S32). The water level prediction error range setting unit 11 acquires water level prediction data P indicating the predicted water levels at the plurality of times at the water level prediction location from the water level prediction model M (step S32).

[0102] The water level prediction error range setting unit 11 calculates the predicted error probability distribution E of the water level at the water level prediction location and sets the error range of the water level prediction data P with respect to the water level observation data in the water level prediction median rate calculation step S35 (step S33). Here, the water level prediction error range setting unit 11 calculates the predicted error probability distribution E for each command criterion for water level warning at the water level prediction location and / or for each period length from the prediction output time in the water level prediction data output step S32 to each prediction target time, and sets the error range of the water level prediction data P (step S33).

[0103] In the column of "prediction error after each N hours" in FIG. 21, the predicted error probability distribution E of the water level at the water level prediction location is calculated for the observed water level less than the command criterion water level A of 0 m or more, the observed water level less than the command criterion water level B of the command criterion water level A or more, and the observed water level of the command criterion water level B or more. In the column of "error range after each N hours" in FIG. 21, for the observed water level less than the command criterion water level A of 0 m or more, the observed water level less than the command criterion water level B of the command criterion water level A or more, and the observed water level of the command criterion water level B or more, among the predicted error probability distribution E, e lоw-min above e lоw-max below and e mid-min above e mid-max below and e high-min above e high-max below, the error range of the water level prediction data P with respect to the water level observation data is set so as to be within the error range. Here, taking the minimum value and the maximum value of the predicted error probability distribution E as 0% and 100%, 0.1% ≤ e high-min ≤ e mid-min ≤ e lоw-min ≤ 20% and 80% ≤ e lоw-max ≤ e mid-max ≤ e high-maxThe condition ≤99.9% is satisfied. In other words, the error range of the water level prediction data P is set wider at higher water levels, narrower at lower water levels, and in the middle at medium water levels. The upper and lower fluctuation ranges of the error range of the water level prediction data P may be equal or different, depending on the distribution shape of the prediction error probability distribution E.

[0104] Thus, when using machine learning to predict water levels in tidal zones, during the validation phase, it is possible to calculate the probability distribution E of the prediction error for water levels, set the error range P for the water level prediction data relative to the water level observation data, and prepare to calculate the water level prediction accuracy H.

[0105] Furthermore, since future observed rainfall data RO cannot be extracted as test data during the prediction phase, even when extracting future forecast rainfall data RF, the water level prediction accuracy rate H can be prepared during the verification phase to support the determination of whether each issuance criterion has been reached.

[0106] A specific example of the water level prediction accuracy calculation process in this disclosure is shown in Figure 22. The verification data extraction unit 8 extracts tidal zone water level data EW, pseudo-tidal level data, upstream water level data UW, past observed rainfall data RO, and future forecast rainfall data RF as verification data for verifying the water level prediction model M (step S31), but does not extract future observed rainfall data RO.

[0107] The verification data extraction unit 8 inputs the following data into the water level prediction model M as verification data: tidal zone water level data EW, pseudo-tidal level data, upstream water level data UW, past observed rainfall data RO, and future forecast rainfall data RF, for verifying the water levels at multiple times at the water level prediction point (step S34). The water level prediction data output unit 7 outputs water level prediction data P, which shows the predicted water levels at the water level prediction point for those multiple times, from the water level prediction model M (step S34).

[0108] The water level prediction accuracy calculation unit 12 calculates the water level prediction accuracy rate H at the water level prediction point as the proportion of the water level prediction data P at the water level prediction point at the water level prediction point at the water level prediction point at the water level prediction point at the water level prediction point at the water level prediction point at the water level prediction point at the water level prediction point at the water level prediction point at the water level prediction point at the water level prediction point at the water level prediction point at the water level prediction point at the water level prediction point (step S35). Here, the water level prediction accuracy rate H is calculated for each issuance criterion for water level warning at the water level prediction point, and / or for each period length from the prediction output time in the water level prediction data output step S34 to each prediction target time (step S35).

[0109] In the "Predicted Data for Each N Hours Later" column of Figure 22, water level prediction data P for multiple time points at the water level prediction point is output for observed water levels of 0m or more but less than the warning threshold water level A, observed water levels of warning threshold water level A or more but less than the warning threshold water level B, and observed water levels of warning threshold water level B or higher. For observed water levels of 0m or more but less than the warning threshold water level A, there are 80 data points that fall within the error range, 10 data points that are below the error range, and 10 data points that are above the error range. For observed water levels of warning threshold water level A or more but less than the warning threshold water level B, there are 30 data points that fall within the error range, 15 data points that are below the error range, and 15 data points that are above the error range. For observed water levels of warning threshold water level B or higher, there are 10 data points that fall within the error range, 10 data points that are below the error range, and 10 data points that are above the error range.

[0110] In the "Accuracy of Prediction After Each N Hours" column of Figure 22, the accuracy of water level prediction H at the water level prediction point is calculated for observed water levels of 0m or more but less than the warning threshold water level A, observed water levels of warning threshold water level A or more but less than the warning threshold water level B, and observed water levels of warning threshold water level B or higher. For observed water levels of 0m or more but less than the warning threshold water level A, the accuracy of water level prediction is 80 / (80+10+10)×100=80%. For observed water levels of warning threshold water level A or more but less than the warning threshold water level B, the accuracy of water level prediction is 30 / (30+15+15)×100=50%. For observed water levels of warning threshold water level B or higher, the accuracy of water level prediction is 10 / (10+10+10)×100=33%.

[0111] Thus, since future observed rainfall data RO cannot be extracted as test data during the prediction stage, even when extracting future forecast rainfall data RF, the reliability of the water level prediction data P can be visualized by calculating the water level prediction accuracy rate H during the verification stage.

[0112] Furthermore, since future observed rainfall data RO cannot be extracted as test data during the prediction phase, even when extracting future forecasted rainfall data RF, the water level prediction accuracy rate H can be calculated during the verification phase to support the determination of whether each issuance criterion has been reached. [Industrial applicability]

[0113] The water level prediction model construction method, water level prediction model construction program, water level prediction data output method, and water level prediction data output program disclosed herein can improve the accuracy of water level prediction in tidal zones to the same level as in non-tidal zones by using the water level prediction model.

[0114] The test data interpolation method disclosed herein can improve the accuracy of water level predictions in tidal zones even when test data for water level in tidal zones are missing. The water level prediction accuracy calculation method disclosed herein can visualize the reliability of water level prediction data by calculating the water level prediction accuracy. [Explanation of symbols]

[0115] W: Water level prediction device M, M1, Mn: Water level prediction model 1: Simulated tide level data conversion unit 2: Deemed water level data conversion unit 3: Rainfall data conversion unit 4: Training data extraction unit 5: Water Level Prediction Model Construction Department 6: Test data extraction unit 7: Water level prediction data output unit 8: Data extraction unit for verification 9: Water Level Prediction Model Selection Section 10: Test data interpolation section 11: Water level prediction error range setting section 12: Water level prediction accuracy calculation unit EW: Tidal zone water level data TD: Tidal level data UW: Upper basin water level data RD: Rainfall data P: Water level prediction data RO: Observed rainfall data RF: Predicted rainfall data AD: Astronomical tidal data E: Prediction error probability distribution H: Water level prediction accuracy rate V: Verification data US:Upstream point WP: Water level prediction point TO: Tidal level observation / prediction site

Claims

1. A tidal zone data extraction step extracts the following as training data necessary for learning the water level at a water level prediction point in the tidal zone at a given time: (1) Tidal zone water level data showing at least one of the water levels at the water level prediction point at that time and in the past before that time; and (2) Pseudo-tidal level data showing at least one of the pseudo-tidal levels at the water level prediction point located upstream of the tidal level observation point or tidal level prediction point at that time, in the past before that time and in the future before that time. As training data, the upstream data extraction step extracts at least one of the following: (3) upstream water level data showing the water level at a point upstream of the water level prediction point at at least one of the water levels at that time and in the past at that time; and (4) rainfall data showing the rainfall at at least one of the rainfall at that time, in the past at that time and in the future at at least one of the water level prediction point and the upstream water basin above the water level prediction point. A water level prediction model construction step involves constructing a water level prediction model for predicting the water level at the water level prediction point in the tidal zone, using the tidal zone water level data, the pseudo-tidal zone data, and at least one of the upstream water level data and the rainfall data. A method for constructing a water level prediction model, characterized by comprising the following features.

2. The following pseudo-tide level data conversion step is further provided before the tidal zone data extraction step: the pseudo-tide level data conversion step uses tidal level data indicating observed or predicted tidal levels at the tidal level observation point or the tidal level prediction point as input data, the tidal zone water level data (limited to data from periods unaffected by rainfall) as training data, and the pseudo-tide level data as output data. A method for constructing a water level prediction model according to claim 1, characterized in that...

3. The steps for extracting tidal zone data and the upstream zone data are as follows: subtract the pseudo-tidal zone data from the tidal zone water level data; convert the tidal zone water level data into deemed water level data from which the influence of tides has been removed; extract the training data necessary for learning the water level of the water level prediction point during the peak water level period of the deemed water level data; and exclude the water level of the water level prediction point outside of the peak water level period of the deemed water level data from the learning target. A method for constructing a water level prediction model according to claim 1 or 2, characterized in that

4. The upstream data extraction step extracts at least one of the following as rainfall data: (1) instantaneous rainfall data showing the instantaneous rainfall at at least one of the time, past time, and future time at at least one of the water level prediction point and the water basin upstream of the water level prediction point; (2) total rainfall data showing the total rainfall from the start of rainfall to at least one of the time, past time, and future time at at least one of the time, past time, and future time at at least one of the time, past time, and future time at at least one of the time, past time, and future time at the water level prediction point and the water basin upstream of the water level prediction point (however, it is reset to 0 when a predetermined time has elapsed from the time when no rain falls); and (3) average rainfall data showing the average rainfall within a predetermined average period at at least one of the time, past time, and future time at at least one of the time, past time, and future time at at least one of the time, past time, and future time. A method for constructing a water level prediction model according to claim 1 or 2, characterized in that

5. A water level prediction model selection step is further provided after the water level prediction model construction step, in which verification data for verifying the water level prediction model is input into multiple types of water level prediction models constructed using multiple types of training data, and a water level prediction model that more closely satisfies the issuance criteria is selected. A method for constructing a water level prediction model according to claim 1 or 2, characterized in that

6. A water level prediction model construction program for causing a computer to perform each processing step of the water level prediction model construction method described in claim 1 or 2.

7. In order to predict the water level at the water level prediction point in the tidal zone using the water level prediction model constructed by the water level prediction model construction method according to claim 1 or 2, the water level prediction data output step involves inputting the tidal zone water level data, the pseudo-tidal level data, and at least one of the upstream water level data and the rainfall data as test data necessary for predicting the water level at a certain time at the water level prediction point, and outputting water level prediction data from the water level prediction model that shows the predicted water level at that time at the water level prediction point. A method for outputting water level prediction data, characterized by comprising the following features.

8. A water level prediction data output program for causing a computer to perform the water level prediction data output step included in the water level prediction data output method according to claim 7.

9. In the test data extracted by the water level prediction data output method according to claim 7, when the tidal zone water level data at the water level prediction point is missing, the water level prediction data output step interpolates the tidal zone water level data for the time of the missing data at the water level prediction point using the water level prediction data output at a prediction time arbitrarily before the time of the missing data, and the water level prediction data for the time of the missing data at a time arbitrarily after the prediction time at the water level prediction point. A test data interpolation method characterized by comprising the following:

10. The water level prediction model construction method according to claim 1 or 2 includes a verification data extraction step of extracting the following as verification data for verifying the constructed water level prediction model: the tidal zone water level data, the pseudo-tidal zone data, the upstream water level data, and the past observed rainfall data and future forecast rainfall data as rainfall data; As verification data, the water level data for verifying water levels at multiple times at the water level prediction point is input into the water level prediction model, the tidal zone water level data, the pseudo-tidal zone water level data, the upstream water level data, the past observed rainfall data, and the future forecast rainfall data is output from the water level prediction model, and water level prediction data showing the predicted water levels at the water level prediction point at those multiple times is output. A water level prediction accuracy calculation step, in which the proportion of the water level prediction data for multiple time points at the water level prediction point that falls within the error range around the water level observation data for multiple time points at the water level prediction point is calculated as the water level prediction accuracy at the water level prediction point, A method for calculating the accuracy rate of water level predictions, characterized by comprising the following:

Images