JP2019045290A - River water level prediction system - Google Patents

Abstract

To provide a river water level prediction system having a higher safety in consideration of a prediction error.SOLUTION: A river water level prediction system 1 predicts a change in a river water level by localized torrential downpour into a medium sized-river, and includes: a river basin information part 2 for dividing an object river basin into a mesh state and determining an outflow rate and a flood arrival time; a raining information processing part 3 for acquiring raining information of an object river basin and extracting the same as raining amount data; river information part 4 for acquiring a water level flow curve which is a relation between the water level at a prediction point for predicting a river water level and a flow rate; a flow rate analysis part 51 for inputting the flow rate and the flood arrival time as well as the raining amount data into a conceptual model and performing a flow rate prediction; a water level calculation part 52 for calculating the river water level from a result of the flow rate prediction and the water level flow rate curve; and an output part 6 for outputting the river water level calculated by the water level calculation part as the river water level curve.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a river water level prediction system that predicts changes in river water level due to heavy rainfall in small and medium-sized rivers.

  In recent years, local torrential rains have caused the water levels of small and medium rivers to rise rapidly, causing frequent flooding and flooding, and even for smooth flood control and evacuation activities, river water levels Development of a system capable of prediction is desired.

  Patent Documents 1-3 and Non-Patent Document 1 disclose a rainfall flood forecasting system and an information distribution system for so-called small and medium rivers. In particular, urban rivers, as disclosed in Non-Patent Document 1, have a short flood arrival time from the precipitation to the target river basin and a high rate of rising water level, so various analyzes can be made in a short time. Methods have been developed.

Japanese Patent Application Laid-Open No. 2002-107462 JP 2012-194738 A JP, 2008-50903, A

Toshiyoshi Shugo, "Special Feature: Cutting-edge technology for rainwater management (2: Water level) Flood forecast in Tokyo," VOL. 105 Water circulation Storage and penetration, Association for Rainwater Infiltration Technology, July 7, 2017 p .19-22

  However, with a simplified runoff analysis model for analysis in a short time, it is difficult to make high-accuracy predictions. In addition, in the case of short time prediction, the influence of the prediction error also increases. For example, considering the safety side and predicting the earliest time and water level to be reached by a flood, a larger water level rise may occur at a later time.

  Then, an object of the present invention is to provide a safer river water level prediction system in consideration of a prediction error.

  In order to achieve the above object, a river water level prediction system of the present invention is a river water level prediction system that predicts a change in river water level due to heavy rainfall in small and medium rivers. River basin information unit that determines the runoff rate and flood arrival time, a rainfall information processing unit that acquires rainfall amount information of the target river basin and extracts it as rainfall data, and a prediction point that predicts the river water level A river information unit for obtaining a water level flow curve which is a relationship between water level and flow rate; a flow rate analysis unit for performing flow rate prediction by inputting the outflow rate and flood arrival time and the rainfall amount data to an outflow analysis model; A water level calculation unit that calculates a river level from the result and the water level flow curve, and an output unit that outputs the river level calculated by the water level calculation unit as a river level curve The flood arrival time is set with a width for each mesh, and the flow rate analysis unit performs at least two types of flow rate prediction: a model for quickening the flood arrival time and a model for slowing the flood, and the output unit Is characterized in that at least two river level curves are output.

  Here, it is preferable that in the output unit, a threshold value of the river water level is set, and the output unit includes notification means for notifying when any of the predicted river water level curves exceeds the threshold value.

  The mesh preferably has a side of 250 m or less. Further, in the output unit, a rainfall display area for displaying the rainfall amount information and a predicted water level display area for displaying the predicted river level curve may be provided adjacent to each other.

  The river water level prediction system according to the present invention configured as described above has two models, ie, a model in which at least the flood arrival time is quicker and a model in which it is delayed, when the flow rate is predicted by the runoff analysis model from rainfall information of the target river basin. Make a flow rate forecast.

  Therefore, even if there is a prediction error, it is possible to consider a plurality of models for both the flood arrival time and the water level, and therefore it is possible to predict the river water level with higher safety.

  In addition, if the threshold value of the river water level is set in the output unit, and if any of the predicted river water level curves exceeds the threshold value, the water level increase is likely to occur. In the event of a situation, the manager can shift to alerting without missing. Furthermore, if one side of the mesh is 250 m or less, prediction with higher accuracy can be performed.

  Then, if the rainfall display area and the predicted water level display area are provided adjacent to each other in the output unit, the predicted river water level curve can be viewed while comparing with the real-time rainfall amount information. It will be possible to give the response a wide range, such as making a decision.

It is explanatory drawing which showed the flow of a structure and process of the river water level prediction system of embodiment of this invention. It is a figure showing an example which divided an object river basin into mesh shape. It is the figure which showed the relationship between the peak river water level and the flood arrival time. It is the figure which showed an example of the water level flow curve (HQ curve). It is the figure which showed an example of the prediction result of two models from which precipitation time data and a river water level curve are arranged side by side up and down, and the generation time of a peak river water level differs. It is a figure showing an example of an output part where a rainfall display area and a predicted water level display area were displayed adjacent to each other.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is an explanatory view showing the configuration and process flow of a river level prediction system 1 according to the present embodiment.

The river level prediction system 1 according to the present embodiment aims to predict changes in the river level due to heavy rainfall in small and medium-sized rivers. Here, "medium and small rivers" means rivers with a small basin area and which do not need to consider storage phenomena. For example, rivers with a basin area of approximately 800 km ² or less, urban rivers, etc. are included.

  In addition, “intensive torrential rainfall” targets rainfall with a rainfall intensity of 20 mm / h or more, and especially if rainfall with a rainfall intensity of 50 mm / h or more falls on the target river basin, prediction by the river water level prediction system 1 will always be performed Become.

  As shown in FIG. 1, the river level prediction system 1 includes a river basin information unit 2, a rainfall information processing unit 3, a river information unit 4 and an analysis unit 5 (flow rate analysis unit 51, water level calculation unit 52), It is mainly comprised by the output part 6 (The screen output part 61, the alert part 62 used as alerting | reporting means). Moreover, the observation part 7 which observes the actual river water level can also be provided.

  First, in the river basin information unit 2, the target river basin is divided into meshes. FIG. 2 shows an example in which the watershed 22 (target river basin) of the target river 21 which is a medium or small river is divided into meshes.

  The basin 22 is set as an area which is a rainfall that affects the water level at the prediction point 24 where the river water level is predicted. For example, within a radius of 10 km of the prediction point 24, a range in which precipitation affecting the river water level can be set can be set as the basin 22.

  The mesh 23 which divides the basin 22 can be, for example, a square mesh having a side of 250 m. That is, the whole area of the basin 22 is divided by meshes 250, 250 m square per side.

  An outflow rate (outflow coefficient) is set to each mesh 23. The runoff rate is set to a different value depending on the topography and land use conditions, and it is 0.8-0.9 (general urban area 0.8, dense urban area 0.9) in general urban areas where there are many paved ground surfaces and there are many releases from sewers, If it is a field or a field, 0.6 is set if it is 0.6, if it is a paddy field.

  In addition, flood arrival time is also set to each mesh 23 respectively. This “flood arrival time” is the time required for the rainfall falling on a certain mesh 23 to reach the predicted point 24.

  In the river level prediction system 1 of the present embodiment, the flood arrival time is set with a width in each mesh 23. For example, two times, the earliest time and the latest time of flood arrival time, are set for each mesh.

  Here, FIG. 3 shows the relationship between the peak river water level and the flood arrival time. The illustrated plot shows the observed values at three points (point A, point B and point C (see FIG. 2)) in the watershed 22 and the estimated values by the practical estimation formula.

  As can be seen from this figure, the flood arrival time is dispersed with a width between 20 and 60 minutes. Moreover, the straight line drawn by the least squares method based on the plot is almost horizontal, which indicates that the case of late flood arrival time may become the peak river level.

  Therefore, a model (the early delivery model 611) in which the flood arrival time is set to the earliest time and a model (the late delivery model 612) in which the flood arrival time is set to the latest time are constructed (see FIGS. ).

  On the other hand, the rainfall information processing unit 3 acquires rainfall amount information of the basin 22. As rainfall information, rainfall information by numerical forecast can be obtained from the Meteorological Agency. For example, in the case of precipitation nowcasting, it is possible to obtain a short-term forecast of precipitation at a resolution of 1 km square.

  Furthermore, in the case of high resolution precipitation nowcasting, it is possible to obtain a short-term forecast of precipitation at a resolution of 250 m square. That is, by setting the mesh 23 to a 250 m square, it can be made to coincide with the range in which the amount of rainfall can be obtained by high resolution precipitation nowcasting.

  Therefore, the rainfall information processing unit 3 extracts rainfall data corresponding to the mesh 23 of the basin 22 from the rainfall information of the high resolution precipitation nowcast. The extraction of rainfall data by the rainfall information processing unit 3 is performed, for example, one hour after the current time, three hours before the current time, and the flood arrival time.

  Further, in the river information unit 4, a water level flow curve (H-Q curve) which is a relationship between the water level H of the prediction point 24 where the river water level is predicted and the flow rate Q is obtained in advance. An example of a water level flow curve (HQ curve) is shown in FIG.

  This water level flow curve (HQ curve) is created by determining the limited range (limited range) affected by the river of the prediction point 24 and the flood. Here, the basin 22 shown in FIG. 2 is a limited range. The river area (the cross-sectional area of the river 41), the river gradient, the depth of the river, the roughness coefficient of Manning are determined in the river 41 of the limited range. Here, the riverbed and the depth change depending on the river water level 411.

  In FIG. 4, the relationship between the river water level increase amount H and the flow rate Q is shown. By preparing such a water level flow curve (H-Q curve) in advance, it becomes possible to immediately replace the result of flow rate prediction with the river water level.

  Then, when there is rainfall above the predetermined value in the basin 22, the flow rate analysis unit 51 of the analysis unit 5 predicts the flow rate of the prediction point 24 by the outflow analysis model. The runoff analysis model is preferably a conceptual model that can directly calculate the river flow output from the rainfall falling in the basin 22 because it can reduce input information and calculation load. In the river level prediction system 1 of the present embodiment, a combined rational equation is used as a runoff analysis model.

  The synthetic rational formula is a runoff analysis model that is often used for planning small and medium-sized rivers and sewers at the practical level. In the synthetic rational formula, the target river basin is divided into small basins, and the peak discharge of each of the small basins obtained by the rational formula is calculated from the average rainfall of the basin considering the runoff rate and the flood arrival time, This is a method of predicting the flow rate of the prediction point 24.

  The flow rate analysis unit 51 acquires the outflow rate and the flood arrival time of each mesh 23, ... from the river basin information unit 2, and acquires the rainfall amount data of each mesh 23, ... from the rainfall information processing unit 3 Analysis.

  In the flow rate analysis unit 51, two types of analysis of the early delivery model 611 in which the flood arrival time is early and the late delivery model 612 in which the flood arrival time is late are performed, so two flow rate predictions are calculated.

  And with respect to the calculated flow volume prediction, it is calculated in the water level calculation part 52 as prediction of a river water level. That is, the predicted river flow can be calculated by replacing the predicted flow with the river water level using the water flow curve (HQ curve) of the river information unit 4.

  Here, since the flow rate prediction is obtained as an elapsed time value, the river predicted water level can also be calculated as an elapsed time value and graphed as a river water level curve. An example of the predicted river level curve is shown in FIG.

  In FIG. 5, rainfall data are shown in the upper part, and temporal changes in river water level are shown in the lower part. Here, the rainfall amount data is shown as a weighted average with five areas of M1 area, M2 area, M3 area, M4 area and M5 area. These five areas correspond to small basins that divide the target river basin according to the synthetic rational formula, and in the present embodiment, exemplifying that the basin 22 is divided into five areas (small basins).

  And in the lower part of FIG. 5, the predicted results of the early delivery model 611, which is a model for which the peak river water generation time is early, and the predicted results of the late delivery model 612, which is a model for which the peak river water generation time is late, are It is shown as a water level curve.

  Moreover, the actual river water level measured by the observation part 7 in the prediction point 24 is shown as an observation value for comparison. According to the comparison with this observation value, it can be said that in this case of rainfall, the prediction result of the express delivery model 611 is closer.

  FIG. 6 shows an example of the output unit 6. The output unit 6 corresponds to a monitor screen or a dedicated terminal screen of a computer. Also, a printer can be used as an output unit.

  The output unit 6 is provided with a rainfall display area 6a, a predicted water level display area 6b adjacent thereto, a wind speed information display area 6c, a presentation information display area 6d, and a numerical information display area 6e.

  In the rainfall display area 6a, a display of high resolution precipitation nowcast obtained periodically from the Meteorological Agency is output. In the wind speed information display area 6c and the presentation information display area 6d, the wind speed information on the ground and the sky obtained similarly from the Japan Meteorological Agency and the information announced by the Meteorological Agency such as an alarm and a warning are displayed. The announcement information display area 6d can be displayed in yellow in the case of a warning, or in red in the case of an alarm.

  On the other hand, two river water level curves are output as the screen output unit 61 in the predicted water level display area 6b. Here, a river water level curve by the early delivery model 611 and a river water level curve by the late delivery model 612 are output. The forecast of the river level curve is updated at intervals of up to 5 minutes.

  With respect to the prediction of the two river level curves, the screen output unit 61 also displays two river level thresholds. Here, the lower threshold of the water level is a primary control value 621 (for example, 2.2 m), and the higher threshold of the water level is a secondary management value 622 (for example, 3.5 m).

  Then, if the peak river level of the predicted river level curve is equal to or less than the primary management value 621, the alert unit 62 serving as a notification means displays "No river level alert".

  On the other hand, when the peak river water level of one of the two predicted river water level curves exceeds the primary management value 621, the alert unit 62 displays "Warning" and is registered. The alert message is sent to the mobile terminal that has been

  Furthermore, when the peak river level of either of the two predicted river level curves exceeds the secondary control value 622, the alert unit 62 displays "danger" and is installed at the site. The rotating light flashes and an alert mail is sent to the mobile terminal.

  Here, the value of the water level of the river water level curve can also be read from the numerical information display area 6e. In the numerical information display area 6e, the observed value of the river water level at the current time measured by the observation unit 7 and a plurality of predicted values one hour ahead from the current time are displayed as numerical values.

  Next, the operation of the river level prediction system 1 of the present embodiment will be described.

  First, a prediction point 24 for monitoring the river water level is determined, and a target river basin serving as rainfall affecting the river water level of the prediction point 24 is set as the water basin 22 (see FIG. 2). The watershed 22 is divided by the mesh 23 with reference to the resolution of the rainfall amount data that can be acquired by the rainfall information processing unit 3 and the like.

  Subsequently, with respect to the divided meshes 23, ..., in the river basin information unit 2, the setting of the runoff rate and the flood arrival time to the predicted point 24 is performed respectively. Here, for the flood arrival time, an early time for the express delivery model 611 and a late time for the late delivery model 612 are set.

  On the other hand, in the river information unit 4, the water level flow curve (HQ curve) in the channel 41 of the limited range of the prediction point 24 is obtained. In addition, a water level gauge or the like is installed as the observation unit 7 at the prediction point 24 so that the river water level prediction system 1 can be captured.

  Then, when these initial conditions are met, the operation program of the river level prediction system 1 installed in the computer or stored on the cloud is activated. Each process described below is executed by an instruction that causes a computer to function.

  When the operation of the river level prediction system 1 is started, rainfall information is acquired from the Meteorological Agency to the rainfall information processing unit 3 at intervals of, for example, five minutes, and extracted as rainfall data. The acquired rainfall amount information is displayed on the rainfall display area 6 a of the monitor screen which is the output unit 6. Rainfall radar images from three hours ago to one hour ahead are displayed in the rainfall display area 6a.

  In the flow rate analysis unit 51, on the basis of the extracted rainfall amount data and the outflow rate and flood arrival time for each of the meshes 23,... Flow prediction is performed at the prediction point 24 by the model 612.

  The two predicted flow rates are converted into river predicted water levels by the water level flow curve (H-Q curve) fetched from the river information unit 4 and displayed in the predicted water level display area 6b of the monitor screen that becomes the output unit 6 Be done. In the predicted water level display area 6b, two river water level curves (611, 612) from two hours ago (or three hours ago) to one hour ahead are displayed.

  Here, in the predicted water level display area 6b, in addition to the two predicted river level curves (611, 612), a primary management value 621 and a secondary management value 622 are displayed. Moreover, the observation value of the river water level to the present time measured by the observation part 7 can also be displayed collectively.

  When none of the peak river levels of the two predicted river level curves exceeds the primary management value 621, the processing from acquisition of rainfall information to prediction of the river level is repeated at predetermined intervals.

  On the other hand, when any of the predicted peak river levels exceeds the primary management value 621, "Attention" is displayed on the alert unit 62 of the monitor screen serving as the output unit 6, and the manager's mobile phone Alert emails are sent to

  Furthermore, when any of the predicted peak river levels exceeds the secondary management value 622, "danger" is displayed on the alert unit 62 of the monitor screen that is the output unit 6, and the manager's mobile phone Alert emails are sent to In addition, turning lights can be activated to make many people give visual reminders.

  Next, the operation of the river level prediction system 1 of the present embodiment will be described.

  The river level prediction system 1 according to the present embodiment configured as described above is at least a flood when the flow rate is predicted by the runoff analysis model such as the synthetic rational formula from the rainfall amount information of the basin 22 which is the target river basin. Two flow rate predictions are performed: an early delivery model 611 that the arrival time is earlier and a late delivery model 612 that is the late arrival time.

  For this reason, even if there are actual river water elevations and different prediction errors, multiple models (611, 612) are considered for both flood arrival time and water level, and therefore the safer river water level prediction and Can manage.

  In addition, the threshold (621, 622) of the river water level is set in the output unit 6, and informing means (alert unit 62) performs notification when any of the predicted river level curves (611, 612) exceeds the threshold. If there is a notification transmission function (alert email notification, rotating light, etc.), the administrator can shift to a vigilant system without overlooking in situations where the water level is likely to rise. .

  For example, according to the prediction of the express delivery model 611, it is possible to start flood control activities and evacuation activities at an earlier time, but when the peak river level of the express delivery model 611 does not exceed the primary management value 621, notification is not given and the flood control It does not lead to the start of activities. On the other hand, even if the peak river level of the express delivery model 611 does not exceed the primary management value 621, the peak river level of the late delivery model 612 may exceed the threshold, so prediction by the express delivery model 611 is sufficient. Even if not, by adding the prediction of the late delivery model 612, it becomes possible to start the flood control activity etc. on the safe side, and it becomes possible to carry out more secure management.

  In addition, in the case where the flood control activity or the like is started after the water level rise is actually measured by the observation unit 7, the dangerous water level may be reached in about 10 minutes, so that only evacuation may be possible. On the other hand, if the river water level can be predicted with high accuracy, construction sites can secure time for removing materials and materials, and perform useful management with high safety and little damage. become able to.

  In addition, if there are two predictions, the early delivery model 611 and the late delivery model 612, it will be easier to predict the time when the river water level will return during peace after local heavy rainfall ceases, ensuring security. It will be possible to quickly shift to evacuation and recovery actions in a state of

  Furthermore, if the environment is such that one side of mesh 23, ... is 250m or less and detailed rainfall information in a narrow range can be acquired like high resolution precipitation nowcast, river water level prediction will be performed with higher accuracy. be able to.

  Then, if the rainfall display area 6 a and the predicted water level display area 6 b are provided adjacent to each other in the output unit 6, the predicted river water level curve can be viewed while comparing with the real-time rainfall amount information. In this case, regardless of the display of the alert unit 62, for example, the manager can complement the prediction while looking at the rainfall amount information, so that it is possible to have a wider response, such as making a judgment on the safe side. be able to.

  Furthermore, in addition to the rainfall display area 6a and the predicted water level display area 6b, if the wind speed information display area 6c and the presentation information display area 6d from the Japan Meteorological Agency are provided on the monitor screen of the output unit 6, the administrator It will be possible to make a comprehensive judgment from a lot of information.

  As mentioned above, although the embodiment of the present invention has been described in detail with reference to the drawings, the specific configuration is not limited to this embodiment, and the design change to the extent not departing from the gist of the present invention Included in the invention.

  For example, although the above embodiment has described the case of predicting the two river level curves of the early delivery model 611 and the late delivery model 612, the present invention is not limited to this, and three or more river level curves are described. Can also be displayed and displayed.

  Moreover, in the said embodiment, although the synthetic | combination rational formula was demonstrated to the example as a runoff analysis model, it is not limited to this, A tank model, a unit map method, etc. can also be used as a runoff analysis model.

  Furthermore, although the said embodiment demonstrated the case where the basin 22 was divided | segmented into the square mesh 23, ... of which one side is 250 m, it is not limited to this, for example, one side is larger than 250 m, for example It is also possible to divide one side into a 1-km square mesh.

1 River water level prediction system 2 River basin information part 21 target river (medium and small rivers)
22 basin (target river basin)
23 mesh 24 prediction point 3 rainfall information processing unit 4 river information unit 411 river water level 5 analysis unit 51 flow rate analysis unit 52 water level calculation unit 6 output unit 6 a rainfall display area 6 b predicted water level display area 61 screen output unit 611 early delivery model (flood Model with faster arrival time)
612 Delayed delivery model (flood arrival time delayed model)
62 Alert section (informing means)
621 Primary management value (threshold)
622 Secondary control value (threshold)

Claims (4)

A river water level prediction system that predicts changes in river water level due to heavy rainfall in small and medium-sized rivers.
A river basin information unit that divides the target river basin into meshes and determines the runoff rate and flood arrival time for each mesh;
A rainfall information processing unit that acquires rainfall information of the target river basin and extracts it as rainfall data;
A river information unit for obtaining a water level flow curve which is a relationship between the water level and the flow rate at a prediction point where the river water level is predicted;
A flow rate analysis unit that estimates the flow rate by inputting the runoff rate, the flood arrival time, and the rainfall amount data to a runoff analysis model;
A water level calculation unit that calculates a river water level from the flow rate prediction result and the water level flow curve;
An output unit that outputs the river water level calculated by the water level calculation unit as a river water level curve;
The flood arrival time is set with a width for each mesh, and the flow rate analysis unit performs two types of flow rate predictions, a model in which the flood arrival time is earlier and a model in which the delay is slower, and at least the output unit A river level prediction system characterized in that two river level curves are output.   The threshold value of the river water level is set to the said output part, It has an alerting | reporting means which alert | reports, when either of the estimated said river water level curves exceeds the said threshold value. River water level forecasting system.   The river level prediction system according to claim 1 or 2, wherein one side of the mesh is 250 m or less.   The rainfall display area for displaying the rainfall amount information and the predicted water level display area for displaying the predicted river level curve are provided adjacent to each other in the output unit. The river level forecasting system described in any one of the items.