River flood flow monitoring method based on fusion of remote sensing imagery and hydrodynamic model

Abstract

The present invention relates to a river flood flow monitoring method based on the fusion of remote sensing images and hydrodynamic models, and belongs to the field of flood monitoring technology. First, the flood levels in the river basin are divided in combination with historical rainfall and flood conditions, flood process lines of different flood levels are drawn, and the 24-hour design rainstorms for each level of flood are determined. Then, a one-dimensional hydrodynamic model of the river is established using cross-section survey data, and a water level-water surface width relationship is constructed for the section to be measured. The model is solved to obtain the water level-flow relationship curves of the river under each level of flood, the recurrence level of this flood is determined, the current water surface width of the river section to be measured is calculated, and the corresponding water level is obtained. According to the flood level of the river and the stage of the flood flow process, the corresponding water level-flow relationship curve is selected and substituted into the curve to obtain the flood flow of the section to be measured in the current period. The present invention improves the accuracy of flood peak flow inversion using satellite remote sensing images by inverting water levels at different levels and stages.

Description

Technical Field

[0001] The invention relates to a river flood flow monitoring method based on the fusion of remote sensing images and hydrodynamic models, and belongs to the technical field of flood monitoring. Background Art

[0002] Flood flow monitoring is a crucial component of flood disaster forecasting. Rapidly and accurately capturing flood peaks during floods is crucial for minimizing losses from flood disasters. Currently, peak flow monitoring relies on two methods: manual flow measurement and automated monitoring stations. The former requires wading through water using equipment such as flowmeters, which poses safety risks and is inefficient during flood season. While the latter utilizes advanced instruments such as ultrasonic flowmeters, it requires supporting power and communication infrastructure, making implementation difficult in remote areas and subject to challenges such as flood damage and high sensor maintenance costs. Furthermore, fixed monitoring stations require regular equipment calibration and silt removal, imposing a heavy operational and maintenance burden.

[0003] For blind spots in small and medium-sized river monitoring, satellite remote sensing technology can be used, combined with cross-sectional mapping data to construct hydraulic models. Non-contact flow estimation can be achieved by inverting parameters such as water surface width and flow velocity distribution through remote sensing imagery. This collaborative air-ground-space monitoring model offers the advantages of rapid response and full coverage.

[0004] Patent CN118794495A invented a method for calculating peak flow by integrating hydrological calculations and remote sensing images. The method mainly uses the Manning formula to derive the corresponding water level-flow relationship, and then interpolates the calculated peak flow based on the obtained water level. However, since the Manning formula is an empirical formula under constant uniform flow conditions, this method is only applicable to artificial channels with constant width and cross-sectional shape along the way. It is not applicable to most natural river channels with irregular shapes and changing cross-sectional shapes along the way. Summary of the Invention

[0005] The purpose of the present invention is to overcome the above-mentioned shortcomings and provide a river flood flow monitoring method based on the fusion of remote sensing images and hydrodynamic models. By inverting water levels at different levels and stages, the accuracy of flood peak flow inversion from satellite remote sensing images is improved.

[0006] The technical solution adopted by the present invention is:

[0007] The river flood flow monitoring method based on the fusion of remote sensing images and hydrodynamic models includes the following steps:

[0008] S1. Flood levels within the river basin are classified based on historical rainfall and flood conditions. Flood hydrographs for different flood levels are drawn for the river section under test based on historical data, serving as representative flood flow processes for each flood level. The 24-hour design rainfall for each flood level within the river basin is then determined based on local hydrological manuals. This is used to determine the recurrence level of the current flood by comparing the maximum 24-hour rainfall in the basin prior to the current flood with the 24-hour design rainfall.

[0009] S2. Build a one-dimensional hydrodynamic model of the river channel based on the one-dimensional Saint-Venant equations. Construct a water level-water width relationship based on river channel topography data. Calculate the hydraulic radius, flow area, and Scheherazade coefficient, the parameters required for the hydrodynamic model, based on the actual river channel topography and ground cover data. Calculate the water level-discharge relationship for the downstream channel using the Scheherazade formula. Discrete the continuity and momentum equations in the Saint-Venant equations using the finite difference method. Discrete the river channel into water level and flow points at appropriate time and space steps. Solve the equations using the chase method, ultimately deriving the water level and flow rate at various times and locations in the river channel during a flood.

[0010] S3. Extract the water level and flow rate at each time before the peak flow rate of the measured section to obtain the water level-flow relationship during the flooding process. Then, extract the water level and flow rate at each time from the peak flow rate to the complete receding of the flood in the measured section to obtain the water level-flow curve during the receding process. This will result in the water level-flow relationship curves for the flooding and receding of the river under various flood magnitudes.

[0011] S4. Determine the recurrence level of this flood based on a comparison of the maximum 24-hour rainfall before the flood with historical data. Calculate the current water surface width of the river section under test using satellite remote sensing imagery data. Determine the corresponding water level based on the relationship between the water level at the river section and the water surface width.

[0012] S5. Select a corresponding water level-discharge relationship curve based on the determined flood level of the river channel and the stage of the flood flow process, substitute the corresponding water level for linear interpolation, and calculate the flood flow of the section to be measured in the current period.

[0013] In the above method, step S1 classifies floods into 5-year, 20-year, 50-year, and 100-year levels based on the rainfall within 24 hours before the flood. Rainfall less than 5 years is classified as a small flood, 5-20 years is classified as a medium flood, 20-50 years is classified as a major flood, and more than 50 years is classified as a major flood.

[0014] The one-dimensional Saint-Venant equation described in step S2 is:

[0015] ,

[0016] ,

[0017] Where: Q is the flow rate, m 3 / s;

[0018] q is the lateral inflow, m 3 / s;

[0019] A is the water flow area, m 2 / s;

[0020] h is the water level, m;

[0021] R is the hydraulic radius, m;

[0022] C is Xiecai coefficient;

[0023] α is the momentum correction coefficient;

[0024] g represents the acceleration due to gravity;

[0025] t is time;

[0026] x is the coordinate along the river channel, with positive values ​​toward the direction of water flow;

[0027] First, based on the actual topography and surface cover data of the river channel, representative sections were selected at intervals of 100-200m along the river channel, and the hydraulic radius, wetted perimeter, flow area and Xie Cai coefficient of each section at each water level were calculated to determine the input parameters required for the model calculation; then, the Saint-Venant continuity equation and momentum equation were discretized using the four-point implicit equidistant format, the river section points were set as water level variable points, and flow variable points were set between adjacent water level points; for the upstream calculation point, the flood process line was introduced as the input boundary condition; for the downstream boundary, the Xie Cai formula was used to calculate the water level-flow relationship curve of the section, and the water level-flow relationship was converted into a linear constraint equation and embedded in the system as the downstream boundary condition. Finally, the equations of all water level points and flow points were integrated to construct a linear equation system, which was solved using the pursuit method to directly obtain the water level-flow relationship curve under various working conditions for each section of the entire river channel.

[0028] The beneficial effects of the present invention are:

[0029] (1) Combining the water surface width inverted by remote sensing images with measured cross-sectional topographic data, it eliminates the need for manual wading or reliance on fixed monitoring stations, thus solving the problem of a small number of monitoring sites and narrow coverage of traditional hydrological stations;

[0030] (2) The water level inversion formulas of different magnitudes and stages are used for floods, which avoids the lag of traditional single curve interpolation in the dynamic process of floods and improves the accuracy of flood peak flow inversion from satellite remote sensing images. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 is a flow chart of the method of the present invention;

[0032] Figure 2 1. A typical flood flow process diagram of each level of flood according to an embodiment of the present invention;

[0033] Figure 3 1 is a water level-flow relationship diagram of the most downstream section of the river calculated according to Xie Cai's formula in an embodiment of the present invention;

[0034] Figure 4 Graphs showing the relationship between water level and flow during the rising and receding stages of floods of various levels according to an embodiment of the present invention; (a) represents a particularly large flood, (b) represents a large flood, (c) represents a moderate flood, and (d) represents a small flood.

[0035] Figure 5 1 is a graph showing the relationship between the water level and the water surface width of the section to be measured in an embodiment of the present invention. DETAILED DESCRIPTION

[0036] The present invention is further described below with reference to specific embodiments.

[0037] Example 1 A method for monitoring river flood flow based on the fusion of remote sensing images and hydrodynamic models, comprising the following steps (e.g. Figure 1 )as follows:

[0038] S1. Based on historical rainfall and flood conditions, the flood levels within the river basin were classified. Based on the rainfall-runoff correlation diagram in the local hydrological manual and the instantaneous unit line method, the flow process of typical floods of each level was calculated. The 5-, 20-, 50-, and 100-year design flood process lines of the test section were used as the typical flood flow processes for small floods, medium floods, large floods, and extremely large floods, respectively. The results are as follows: Figure 2 .

[0039] The local hydrological manual then determines the 24-hour design rainfall for the river basin with a return period of 5 years, 20 years, 50 years, and 100 years. This is used to determine the recurrence level of the flood based on the maximum 24-hour rainfall in the basin prior to the flood and the design rainfall. If the rainfall is less than the 5-year design rainfall, the flood is classified as a minor flood. A return period of 5-20 years is considered a moderate flood, a return period of 20-50 years is considered a major flood, and a return period of more than 50 years is considered a catastrophic flood. For example, the 24-hour design rainfall for a 5-year flood in a river is 134.3 mm, 191.9 mm for a 20-year flood, 228.3 mm for a 50-year flood, and 255.5 mm for a 100-year flood.

[0040] S2. Establish a one-dimensional hydrodynamic model of the river channel based on the one-dimensional Saint-Venant equations.

[0041] ,

[0042] ,

[0043] Where: Q is the flow rate, m 3 / s;

[0044] q is the lateral inflow, m 3 / s, lateral inflow is not considered in this method, q=0;

[0045] A is the water flow area, m 2 / s;

[0046] h is the water level, m;

[0047] R is the hydraulic radius, m;

[0048] C is Xiecai coefficient;

[0049] α is the momentum correction coefficient;

[0050] g represents the acceleration due to gravity;

[0051] t is time;

[0052] x is the coordinate along the river channel, with positive values ​​toward the direction of water flow;

[0053] The establishment of this model aims to obtain the accurate water level-flow relationship curve of the river section to be tested under various working conditions, and provide a basis for subsequent flow inversion. Its implementation method is: first, based on the actual topography and surface cover data of the river, representative sections are selected along the river at intervals of 100-200m, and the hydraulic radius, wetted perimeter, flow area and Xie Cai coefficient of each section at each water level are calculated to determine the input parameters required for model calculation; then, the four-point implicit equidistant format is used to discretize the Saint-Venant continuity equation and momentum equation, and the river section points are set as water level variable points (Z), and flow variable points (Q) are set between adjacent water level points; for the upstream calculation point, the flood process line is introduced as the input boundary condition; for the downstream boundary, the Xie Cai formula is used to calculate the water level-flow relationship curve of the section (such as Figure 3 ), the water level-flow relationship is converted into a linear constraint equation and embedded into the system as the downstream boundary condition. Finally, the equations of all water point (Z) and flow point (Q) are integrated to construct a linear equation system, which is solved by the pursuit method to directly obtain the water level-flow relationship curve under various working conditions in all sections of the entire river. The water level-water surface width relationship is constructed based on the river topography measurement data (such as Figure 5 ).

[0054] S3. Extract the water level and flow rate at each time before the flood peak at the section to be measured, and obtain the water level-flow relationship during the flooding process. Then extract the water level and flow rate at each time from the flood peak to the flood completely receding, and obtain the water level-flow curve during the receding process. Repeat this process for floods of different recurrence periods, and obtain the water level-flow curves for the flooding and receding processes of various magnitudes (such as Figure 4 ).

[0055] S4. The recurrence level of this flood was determined by comparing the maximum 24-hour rainfall before the flood with historical data. The current width of the river section under investigation was calculated using satellite remote sensing imagery. The corresponding water level was then determined based on the relationship between the water level and width of the river section. The rainfall data collected 24 hours before the flood was 171 mm, making this a moderate flood. Remote sensing image interpretation revealed that the river width at this section during a flood was 80 meters. The corresponding water level-water width curve was used to determine the current water level, which was 199.95 meters.

[0056] S5. Select the corresponding water level-discharge relationship curve based on the determined flood level and flood flow process stage of the river channel, substitute the corresponding water level and perform linear interpolation to calculate the flood flow of the measured section in the current period:

[0057] The water level-discharge relationship curve during the rising process of a medium-sized flood was selected, and the flood discharge at this location was found to be 203 m³ / s based on the water level.

[0058] The above is a detailed description of the present invention in conjunction with specific embodiments, and the protection scope of the present invention is not limited thereto.

Claims

1. A river flood flow monitoring method based on the fusion of remote sensing images and hydrodynamic models is characterized by: The steps are as follows: S1. Classify the flood levels within the river basin based on historical rainfall and flood conditions. Flood hydrographs of different flood levels for the river section under test are drawn based on historical data, serving as representative flood flow processes for each flood level. The 24-hour design rainfall for each flood level within the river basin is then determined based on the local hydrological manual. This is used to determine the recurrence level of the flood by comparing the maximum 24-hour rainfall in the basin prior to the flood with the 24-hour design rainfall. Floods are classified based on the rainfall within the 24 hours prior to the flood as occurring once every five years, once every 20 years, once every 50 years, and once every 100 years. A minor flood is defined as a rainfall event less than five years long, a moderate flood as a return period of 5-20 years, a major flood as a return period of 20-50 years, and a major flood as a return period exceeding 50 years. S2. Establish a one-dimensional hydrodynamic model of the river channel based on the one-dimensional Saint-Venant equations. Construct a cross-sectional water level-water width relationship based on river channel topography data. Calculate the hydraulic radius, flow area, and Scheherazade coefficient, parameters required for the hydrodynamic model, based on the actual river channel topography and ground cover data. Calculate the water level-discharge relationship of the downstream channel using the Scheherazade formula. Discrete the continuity and momentum equations in the Saint-Venant equations using the finite difference method. Discrete the river channel into water point and flow points at appropriate time and space steps. Solve the equations using the pursuit method, ultimately deriving the water level and flow rate of the river channel at various times and locations during a flood. S3. Extract the water level and flow rate at each time before the peak flow rate of the measured section to obtain the water level-flow relationship during the flooding process. Then, extract the water level and flow rate at each time from the peak flow rate to the complete receding of the flood in the measured section to obtain the water level-flow curve during the receding process. This will result in the water level-flow relationship curves for the flooding and receding of the river under various flood magnitudes. S4. Determine the recurrence level of this flood based on a comparison of the maximum 24-hour rainfall before the flood with historical data. Calculate the current water surface width of the river section under test using satellite remote sensing imagery data. Determine the corresponding water level based on the relationship between the water level at the river section and the water surface width. S5. Select a corresponding water level-discharge relationship curve based on the determined flood level of the river channel and the stage of the flood flow process, substitute the corresponding water level for linear interpolation, and calculate the flood flow of the section to be measured in the current period.

2. The river flood flow monitoring method based on the fusion of remote sensing images and hydrodynamic models according to claim 1 is characterized in that: The one-dimensional Saint-Venant equations described in step S2 are: , , Where: Q is the flow rate, m 3 / s; q is the lateral inflow, m 3 / s; A is the water flow area, m 2 / s; h is the water level, m; R is the hydraulic radius, m; C is Xiecai coefficient; α is the momentum correction coefficient; g represents the acceleration due to gravity; t is time; x is the coordinate along the river channel, with positive values ​​being in the direction of water flow.

3. The method for monitoring river flood flow based on the fusion of remote sensing images and hydrodynamic models according to claim 1 is characterized in that: In step S2, first, based on the actual topography and surface cover data of the river channel, representative sections are selected at intervals of 100-200m along the river channel, and the hydraulic radius, wetted perimeter, flow area and Xie Cai coefficient of each section at each water level are calculated to determine the input parameters required for the model calculation; then, the Saint-Venant continuity equation and momentum equation are discretized using a four-point implicit equidistant format, the river section points are set as water level variable points, and flow variable points are set between adjacent water level points; for the upstream calculation point, the flood process line is introduced as the input boundary condition; for the downstream boundary, the Xie Cai formula is used to calculate the water level-flow relationship curve of the section, and the water level-flow relationship is converted into a linear constraint equation embedded in the system as the downstream boundary condition. Finally, the equations of all water level points and flow points are integrated to construct a linear equation group, which is solved using the pursuit method to directly obtain the water level-flow relationship curve under each working condition for each section of the entire river channel.

Images