A method and system for early warning of flash floods based on meteorological and hydrological collaborative analysis

Abstract

This invention belongs to the field of data processing technology, specifically relating to a method and system for flash flood disaster early warning based on meteorological and hydrological collaborative analysis. The method includes: acquiring and processing topographic datasets and rainfall data sequences as model inputs for the target early warning area; acquiring multiple topographic geometric parameters for calculating terrain energy; calculating the current terrain energy; accumulating the unit previous impact rainfall, the mean of previous target rainfall across all grid cells, and the standard deviation of previous target rainfall; calculating the current dynamic runoff coefficient; obtaining the current runoff intensity; calculating the current runoff velocity based on the regulatory effect of dynamic hydrological conditions on the static base flow velocity determined solely by topography; and inputting the current runoff velocities of all grid cells within the target early warning area into a distributed hydrological model for early warning. This invention solves the problem that existing distributed hydrological models cannot dynamically reflect differences in runoff generation and runoff changes.

Description

Technical Field

[0001] This invention relates to the field of data processing technology. More specifically, this invention relates to a method and system for early warning of flash floods based on meteorological and hydrological collaborative analysis. Background Technology

[0002] In mountainous areas, early warning of flash floods is crucial, as it directly determines the safety of people's lives and property downstream. Only by accurately predicting flood risks can we buy precious time for disaster prevention and mitigation. Therefore, it is necessary to conduct accurate monitoring of flood risks.

[0003] Traditional methods of monitoring flood risk rely on manually deployed rain gauges. Experienced flood control personnel judge the risk by observing whether the rainfall intensity of a single rain gauge reaches a preset threshold. However, this method depends entirely on sparse, point-like observations, is highly subjective, and the threshold standards set for different river basins vary. Furthermore, the monitoring range of rain gauges is limited. For localized sudden rainstorms not covered by rain gauges, this method will produce missed reports, resulting in a slow response or failure of the early warning system, making it difficult to achieve standardized monitoring of risks across the entire river basin.

[0004] Existing technologies use distributed hydrological models for flood risk prediction. This method is based on rainfall data sequences and topographic datasets covering the entire region, and enables the simulation of the flood evolution process.

[0005] However, this method has limitations for flash flood early warning in small watersheds in mountainous areas. The complex terrain and underlying surface conditions, such as the huge spatial differences in soil moisture content in the early stage, cause the runoff generation and confluence processes of the watershed to exhibit highly nonlinear characteristics. Distributed hydrological models can only reflect physical processes based on fixed parameters, such as fixed runoff coefficients and roughness, and cannot accurately calculate the differences in runoff generation caused by different underlying surface saturation, which are crucial for judging the true risk, and the dynamic changes in flow velocity caused by changes in confluence intensity. Therefore, its monitoring results still have biases and cannot fully meet the needs of accurate early warning. Summary of the Invention

[0006] To address the technical problem that the aforementioned distributed hydrological models, due to their use of fixed parameters, cannot dynamically reflect differences in runoff generation and changes in runoff concentration, this invention provides solutions in the following aspects.

[0007] In a first aspect, the present invention provides a flash flood disaster early warning method based on meteorological and hydrological collaborative analysis, comprising:

[0008] The system acquires and processes terrain datasets and rainfall data sequences for the target warning area, generating them as model inputs. Based on the terrain change characteristics of the target warning area, it acquires multiple terrain geometric parameters for calculating terrain energy, including terrain slope, upstream grid number, upstream elevation, and upstream cell distance. It calculates the current terrain energy based on the interaction between the current grid cell and the surrounding terrain conditions. Based on the rainfall data sequence, it accumulates the unit previous impact rainfall, the mean of previous target rainfall across all grid cells, and the standard deviation of previous target rainfall. Finally, it calculates the current terrain energy based on the previous rainfall amount of the current grid cell and its relationship with the previous target rainfall. Soil moisture content is assessed by the difference between mean rainfall values, and the current dynamic runoff coefficient is calculated. The current topographic energy is multiplied by the total upstream runoff determined by the current dynamic runoff coefficient and the previous influencing rainfall to obtain the current runoff intensity. The mean and standard deviation of the current runoff intensity of all grid cells within the target warning area are calculated and denoted as the target runoff intensity mean and target runoff intensity standard deviation. The current runoff velocity is calculated based on the regulating effect of dynamic hydrological conditions on the static base velocity determined only by geomorphology. The current runoff velocity of all grid cells within the target warning area is input into the distributed hydrological model for warning purposes.

[0009] This invention addresses the problems of traditional manual methods relying on sparse rain gauges, high subjectivity, and a high risk of missed reports. It obtains unified topographic and rainfall data across the entire region, covering areas without rain gauges, thus reducing subjective judgment errors. For existing distributed hydrological models that use fixed parameters and cannot reflect differences in runoff generation and confluence changes, this invention first uses a dynamic runoff coefficient, combined with previous rainfall, to determine soil moisture levels, accurately reflecting the differences in rainfall conversion into runoff in different areas. Then, it assesses the topographic water catchment capacity using topographic properties, combining confluence intensity and dynamic flow velocity to allow the confluence velocity to adjust with water conditions, truly reflecting confluence changes. Finally, it automatically issues warnings by comparing historical data, eliminating the need for manually setting thresholds and avoiding monitoring biases caused by fixed parameters and subjective thresholds. Overall, it achieves end-to-end optimization from data processing to warning determination, accurately reflecting the actual runoff generation and confluence situation, significantly improving the accuracy and reliability of flash flood warnings, and effectively reducing missed and false alarms.

[0010] Preferably, the process of acquiring and processing terrain datasets and rainfall data sequences for the target warning area to generate as model input includes:

[0011] The system performs coordinate projection transformation and spatial resampling on the digital elevation model data of the target warning area to generate a standard terrain raster with a uniform grid resolution, thus obtaining a terrain dataset. Simultaneously, within a preset time period, the meteorological radar data center acquires the rainfall estimation data of the target warning area, and parses the areal radar data and rain gauge station data. It calculates the difference between the rainfall estimation data and the rain gauge station data and performs spatial interpolation to generate an error distribution field. After correction, it performs the same coordinate projection transformation and spatial resampling as the terrain dataset to obtain the rainfall data sequence of the target warning area.

[0012] Preferably, multiple topographic geometric parameters for calculating terrain energy are obtained, including:

[0013] Read the elevation data corresponding to each grid cell in the target warning area, and calculate the elevation change rate in the horizontal and vertical directions with the current grid cell as the center. Then, take the square root of the two to obtain the terrain slope set. Analyze the maximum slope direction between each grid cell and its neighboring cells in the terrain dataset, construct a global flow direction network, and then identify all upstream neighbors that directly flow into the current grid cell based on the network, count the number, read the elevation value, and obtain the number of upstream grid cells and the upstream elevation value of the current grid cell. Calculate the distance between the center point of the current grid cell and the center point of a single upstream grid cell, and record it as the upstream cell distance.

[0014] Preferably, the current situation satisfies the following expression:

[0015] ;

[0016] In the formula, P represents the current terrain energy, which is dimensionless; N represents the number of upstream grid cells, which is dimensionless; U represents the upstream elevation value of the current grid cell, which is height and the unit is meters; C represents the current grid cell elevation value, which is height and the unit is meters; D represents the upstream cell distance, which is meters; and S represents the slope of the upstream grid cell of the current grid cell, which is dimensionless.

[0017] This invention combines factors such as the height, distance, and slope of upstream grid units to comprehensively assess the impact of upstream on the current grid water catchment. It also uses averaging to make the potential energy at different locations comparable, thus accurately distinguishing which areas are prone to water flow and which areas have weak water catchment capacity. This allows the system to identify high-risk points on the terrain in advance and provides accurate terrain references for subsequent early warning.

[0018] Preferably, the unit of anterior impact rainfall, the mean of the anterior target rainfall compared with all grid cells, and the standard deviation of the anterior target rainfall are accumulated, including:

[0019] Based on the rainfall data sequence, the rainfall of individual grid cells within the historical time period is filtered and accumulated to obtain the unit previous impact rainfall. The average and standard deviation of the previous impact rainfall of all grid cells in the target warning area are calculated and denoted as the average previous target rainfall and the standard deviation of the previous target rainfall.

[0020] Preferably, the current dynamic runoff coefficient satisfies the following expression:

[0021] ;

[0022] In the formula, G is the current dynamic runoff coefficient, which is dimensionless; A is the unit anterior impact rainfall of the current grid cell, in millimeters; This represents the average of the target rainfall in the preceding period, in millimeters. The standard deviation of the target rainfall in the previous period is expressed in millimeters; 0.5 and 1 played a role in... The output value of the function is scaled and translated so that G is within [0,1]. Represents a very small positive number, ensuring that the denominator is not zero; The function represents a non-linear mapping function.

[0023] This invention combines the differences between the previous rainfall of a single grid cell and the regional average, and takes into account the influence of soil moisture, allowing the coefficient to vary within a reasonable range. The coefficient is small during drought, indicating that more rainfall infiltrates; the coefficient is large during wet conditions, indicating that more rainfall is converted into water flow. This avoids the inaccurate judgment of runoff situation caused by using a fixed coefficient and improves the accuracy of runoff analysis.

[0024] Preferably, the current flow strength satisfies the following expression:

[0025] ;

[0026] In the formula, F represents the current runoff intensity in cubic meters; P represents the current geological energy, which is dimensionless; A is the current antecedent rainfall per unit of the current grid cell in meters, converted from millimeters; and G is the current dynamic runoff coefficient, which is dimensionless. This represents the area of ​​the current grid cell, in square meters.

[0027] This invention combines topographic energy and upstream effective runoff, integrating the topographic water collection capacity with the actual rainfall-generated water flow for calculation. Areas with strong topographic water collection capacity and abundant rainfall runoff have high flow intensity, while those with weak topographic capacity have low flow intensity. This can accurately reflect the flood convergence situation at different locations, help accurately identify current high-risk areas, and provide a reliable basis for subsequent flow velocity calculation and early warning.

[0028] Preferably, the current merging speed satisfies the following expression:

[0029] ;

[0030] In the formula, This indicates the current flow velocity, expressed in meters per second. It represents the acceleration due to gravity and is a universal physical constant; This represents the side length of a standard terrain grid with uniform grid resolution, in meters. This indicates the current situation; Indicates the terrain slope of the current grid cell; This indicates the current current flow intensity of the current grid cell, in cubic meters. , Indicates the mean and standard deviation of the target flow intensity; The function represents a nonlinear mapping function; It represents a very small positive number, and guarantees that the denominator is not 0.

[0031] This invention considers both the basic driving force provided by the terrain and the influence of real-time confluence intensity, allowing the flow velocity to adjust with changes in water conditions. The flow velocity is fast when the confluence intensity is high and slow when the confluence intensity is low. It eliminates the need for manually setting fixed parameters, avoiding deviations in flow velocity calculation caused by unreasonable parameters, making the flow velocity more closely match the actual flood movement, and improving the accuracy of flood evolution simulation.

[0032] Preferably, the current flow velocity of all grid cells within the target warning area is input into the distributed hydrological model for warning purposes, including:

[0033] The current flow velocity of all grid cells within the target warning area is calculated and input into the distributed hydrological model. The output peak flow value and its growth rate are compared with the historical data of each grid cell. When any of the indicators exceeds the historical record, an early warning is automatically triggered.

[0034] Secondly, the present invention provides a flash flood disaster early warning system based on meteorological and hydrological collaborative analysis, including a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned flash flood disaster early warning method based on meteorological and hydrological collaborative analysis is implemented.

[0035] By adopting the above technical solution, a computer program for a flash flood disaster early warning method based on meteorological and hydrological collaborative analysis is generated and stored in a memory so that it can be loaded and executed by a processor. A terminal device can then be made based on the memory and processor for convenient use.

[0036] The beneficial effects of this invention are as follows: In terms of safeguarding people's lives and property, more accurate and timely early warnings allow downstream areas to prepare in advance, reducing casualties and property losses, and providing strong protection for public safety. Regarding improving the efficiency of flood control work, the system automates the complex early warning analysis process, reducing manual operations and the workload of staff, making flood control work more efficient, especially in emergencies such as sudden rainstorms, enabling rapid response and buying valuable time for disaster prevention and mitigation. From the perspective of regional development, reliable early warnings can reduce the damage of flash floods to local agriculture, transportation, and infrastructure, minimizing the impact of disasters on economic and social development and contributing to stable regional development. Attached Figure Description

[0037] Figure 1 This is a flowchart illustrating a flash flood disaster early warning method based on meteorological and hydrological collaborative analysis according to the present invention. Detailed Implementation

[0038] This invention discloses a flash flood disaster early warning method based on meteorological and hydrological collaborative analysis, referring to... Figure 1 This includes steps S1-S4:

[0039] S1: Acquire and process the terrain dataset and rainfall data sequence for the target warning area to generate as input to the model.

[0040] It should be noted that the early warning method of this invention is based on two common types of gridded data covering the entire early warning basin. First, this raw information needs to be acquired from the data source and standardized to generate the basic data required for subsequent calculations. Taking a small mountainous watershed as an example, and designating this area as the target early warning area, this watershed has complex terrain and sparse traditional rain gauges. Historically, it has repeatedly experienced sudden flash floods caused by localized short-term heavy rainfall, resulting in severe losses. A common characteristic of these disasters is that the heavy rain belt does not completely cover the rain gauges, leading to slow response or failure to issue warnings by the station-based early warning system. Traditional methods of manually interpreting radar images are inefficient and highly subjective. This invention uses technical means to automatically and standardizedly acquire relevant data from these fully covered target early warning areas, establishing a clean and well-organized data foundation for subsequent intelligent analysis.

[0041] Specifically, the process involves acquiring and processing terrain datasets and rainfall data sequences for the target warning area to generate as model inputs, including:

[0042] The system performs coordinate projection transformation and spatial resampling on the digital elevation model data of the target warning area to generate a standard terrain raster with a uniform grid resolution, thus obtaining a terrain dataset. Simultaneously, within a preset time period, the meteorological radar data center acquires rainfall estimation data of the target warning area and parses out a two-dimensional array of rainfall intensity, denoted as area radar data. Real-time rainfall data from automatic rain gauges deployed within and around the watershed of the target warning area are introduced, denoted as rain gauge station geomorphic data. The difference between the rainfall estimation data and the rain gauge station geomorphic data is calculated and spatially interpolated to generate an error distribution field. Based on the error distribution field, the area radar data is corrected, and the same coordinate projection transformation and spatial resampling as the terrain dataset are performed to generate a standard rainfall raster spatially aligned with the standard terrain raster. The standard rainfall raster is then integrated in chronological order to obtain the rainfall data sequence of the target warning area.

[0043] Thus, the topographic dataset and rainfall data sequence of the target warning area were obtained.

[0044] S2: Based on the topographic change characteristics of the target warning area, obtain multiple topographic geometric parameters for calculating the topographic energy, including topographic slope, upstream grid number, upstream elevation value, and upstream cell distance; calculate the current topographic energy based on the interaction between the current grid cell and the surrounding topographic conditions; based on the rainfall data sequence, accumulate the unit previous impact rainfall, the average previous target rainfall of all grid cells, and the standard deviation of the previous target rainfall; assess the soil moisture content based on the amount of previous rainfall in the current grid cell and its difference from the average previous target rainfall, and calculate the current dynamic runoff coefficient; multiply the current topographic energy by the upstream total runoff determined by the current dynamic runoff coefficient and the previous impact rainfall to obtain the current runoff intensity.

[0045] It should be noted that this invention no longer relies solely on the single, fragile threshold logic of rainfall intensity for flash flood warnings. Instead, in flash flood warning practice, it focuses on distinguishing between "disaster-causing" and "non-disaster-causing" rainfall. This involves analyzing the natural runoff capacity of each geographical location within the target warning area. In real mountainous scenarios, a flat, open mountaintop has extremely low runoff capacity, while a narrow, steep gully accelerates flood accumulation. Therefore, by constructing a terrain energy index, the system can automatically identify and calculate these terrain differences, assigning a runoff potential weight to the method of this invention far higher than existing technologies, thereby achieving the identification of high-risk areas at the starting point of the analysis.

[0046] Specifically, based on the terrain change characteristics of the target warning area, multiple terrain geometric parameters are obtained for calculating terrain energy, including terrain slope, upstream grid number, upstream elevation value, and upstream cell distance, including:

[0047] Read the elevation data corresponding to each grid cell in the target warning area, and calculate the elevation change rate in the horizontal and vertical directions with the current grid cell as the center. Then, take the square root of the two to obtain the terrain slope set. Analyze the maximum slope direction between each grid cell and its neighboring cells in the terrain dataset, construct a global flow direction network, and then identify all upstream neighbors that directly flow into the current grid cell based on the network, count the number, read the elevation value, and obtain the number of upstream grid cells and the upstream elevation value of the current grid cell. Calculate the distance between the center point of the current grid cell and the center point of a single upstream grid cell, and record it as the upstream cell distance.

[0048] It should be noted that this invention requires the construction of a single index that can comprehensively reflect the physical meaning of these topographic geometric parameters, namely, topographic energy. The ability of a grid cell to collect floodwater is not determined by a single topographic factor, but by the combined effect of all upstream water inflow paths. The contribution of each upstream water inflow path is mainly composed of two factors: firstly, the hydraulic gradient of the incoming water, which is determined by the relative elevation difference and distance between the upstream grid cell and the current grid cell, representing the initial gravitational potential energy carried by the incoming water; secondly, the topographic slope of the upstream path itself, which represents the secondary acceleration capacity of the topography on the water flow of the path. Therefore, by combining these two factors, the confluence contribution intensity of a single upstream path can be calculated. The total confluence potential of a current grid cell is the sum of the contributions of all its upstream paths. However, to make this indicator comparable across different locations, it needs to be averaged, i.e., the average contribution intensity of each upstream grid cell is calculated. In this way, the geodynamic index can more accurately describe the confluence characteristics of a current grid cell, which is dominated by multiple slow streams and a few fast streams, thus more accurately assessing its inherent risk as a flood confluence point.

[0049] Preferably, the current terrain energy is calculated based on the interaction between the current grid cell and the surrounding terrain conditions, including:

[0050] The current situation satisfies the following expression:

[0051] ;

[0052] In the formula, P represents the current terrain energy, which is dimensionless; N represents the number of upstream grid cells, which is dimensionless; U represents the upstream elevation value of the current grid cell, which is height and the unit is meters; C represents the current grid cell elevation value, which is height and the unit is meters; D represents the upstream cell distance, which is meters; and S represents the slope of the upstream grid cell of the current grid cell, which is dimensionless.

[0053] In the formula, Indicates the relative height between the upstream grid cell and the current grid cell; This represents the slope, which is the spatial relationship from the upstream grid cell to the current grid cell. The calculation of the terrain steepness affects the potential energy. The larger the area and the steeper the terrain, the stronger the potential energy transfer force. This indicates that the contribution of terrain to potential energy is further adjusted by the slope of the upstream grid cells. The larger the area, the more concentrated the terrain, and the stronger the potential energy accumulation effect. The average potential energy of the target warning area is divided by the number of upstream grid cells to obtain the average ground potential energy contribution of each upstream grid cell to the current grid cell. To handle boundary cases, when the grid cell is located on a watershed or mountain top and its upstream grid cell number is 0, the ground potential energy of the grid cell is defined as 0.

[0054] For example, a grid cell located on a gentle hillside has only one upstream grid cell. The number of upstream grid cells is 1. The relative height of the first upstream grid cell is 2 meters, the distance between upstream cells is 50 meters, and the slope of the upstream grid cells is 0.04. The current terrain energy is: There is also a scenario where a grid cell located within a channel has two upstream grid cells. The upstream grid cell has a relative height of 15 meters to the first upstream grid cell, a distance of 30 meters, and a slope of 0.4. The upstream grid cell has a relative height of 10 meters to the second upstream grid cell, a distance of 30 meters, and a slope of 0.3. The current geodynamic energy is: It is evident that the terrain energy of the upstream grid cell slope (0.15) is significantly greater than that of the grid cell located on a gentle slope (0.0012). This invention effectively identifies high-risk points in the target warning area based on the terrain. The above calculation results... , All values ​​are rounded to four decimal places.

[0055] It should be noted that after obtaining the static terrain energy, it is necessary to further examine the dynamic interaction between rainfall and terrain energy. To enable the system to provide early warning of flooding in rainfall-affected areas, this invention combines the dynamic rainfall process with the static terrain energy, calculating runoff intensity in real time, thereby achieving accurate capture of real risks. Furthermore, to more accurately simulate the surface runoff process, this invention first introduces a current dynamic runoff coefficient before calculating the runoff intensity. The reason for this is that the proportion of rainfall converted into surface runoff is not constant but closely related to the previous moisture level of the surface. Dry soil absorbs a large amount of rainwater, resulting in a low runoff coefficient; while saturated soil after continuous rainfall has a runoff coefficient close to 1. This invention calculates this effect by constructing a current dynamic runoff coefficient. When constructing the current dynamic runoff coefficient, the antecedent rainfall in the target warning area is a key indicator for measuring the surface moisture level. It directly affects the proportion of rainfall converted into surface runoff. For example, in areas that have experienced long-term drought, the antecedent rainfall is small, the soil has a strong ability to absorb rainwater, and the runoff generation efficiency is low. Conversely, after continuous rainfall, the antecedent rainfall is large, the soil is saturated, and the runoff generation efficiency is high.

[0056] Preferably, based on the rainfall data series, the average of the previous impact rainfall per unit, the average of the previous target rainfall across all grid cells, and the standard deviation of the previous target rainfall are accumulated, including:

[0057] Based on the rainfall data sequence, the rainfall of individual grid cells within the historical time period is filtered and accumulated to obtain the unit previous impact rainfall. The average and standard deviation of the previous impact rainfall of all grid cells in the target warning area are calculated and denoted as the average previous target rainfall and the standard deviation of the previous target rainfall.

[0058] Specifically, soil moisture content is assessed based on the amount of previous rainfall in the current grid cell and its difference from the average previous target rainfall, and the current dynamic runoff coefficient is calculated, including:

[0059] The current dynamic runoff coefficient satisfies the following expression:

[0060] ;

[0061] In the formula, G is the current dynamic runoff coefficient, which is dimensionless; A is the unit anterior impact rainfall of the current grid cell, in millimeters; This represents the average of the target rainfall in the preceding period, in millimeters. The standard deviation of the target rainfall in the previous period is expressed in millimeters; 0.5 and 1 played a role in... The output value of the function is scaled and translated so that G is within [0,1]. Represents a very small positive number, ensuring that the denominator is not zero; The function represents a non-linear mapping function.

[0062] In the formula, This represents the standardization process for the unit of previous rainfall in the current raster cell, which converts the current raster's previous rainfall into the degree of deviation from the average previous target rainfall, denoted as the previous rainfall deviation. The larger the value, the more excessive the previous rainfall in the current grid cell; the smaller the value, the less excessive the previous rainfall. This indicates that the S-shaped curve describes the relationship between the degree of deviation of previous rainfall and the response of the current dynamic runoff coefficient. Specifically, when the unit of previous impact rainfall is much greater than the average previous target rainfall, tanh approaches 1, the current dynamic runoff coefficient approaches 1, and almost all rainfall is lost. When the unit of previous impact rainfall is much less than the average previous target rainfall, tanh approaches -1, the current dynamic runoff coefficient approaches 0, and almost all rainfall is infiltrated. When the unit of previous impact rainfall is close to the average previous target rainfall, the soil moisture state is at a moderate level, and the distribution of rainfall infiltration and runoff is also in a middle state, with tanh approaching 0 and the runoff coefficient approaching 0.5, meaning that runoff and infiltration each account for about half. This means that by scaling and translation, the current dynamic runoff coefficient of the nonlinear mapping is constrained within the interval [0,1], so that the current dynamic runoff coefficient can directly represent the proportion of rainfall converted into runoff and reflect the dynamic impact of previous rainfall on the runoff generation process.

[0063] For example, suppose the average rainfall of the target area in the previous period at a certain moment is... millimeters, standard deviation In millimeters, there exists a scenario where a certain grid cell was relatively dry in the early stages, and its unit of early-stage rainfall influence A = 20 millimeters. Then, the current dynamic runoff coefficient is... Scenario 2 exists: a certain grid cell is highly humid due to continuous rainfall, with an initial rainfall A = 100 mm and a current dynamic runoff coefficient of [missing value]. As can be seen, this invention dynamically differentiates the runoff generation capacity of surfaces with different levels of moisture through calculation. The above calculation results... , All values ​​are rounded to four decimal places.

[0064] It should be noted that, after obtaining the current geomorphic energy and the current dynamic runoff coefficient, in order to assess the probability risk of flooding in the current grid cell, this invention proposes a calculation mode that combines the source and sink. That is, the current geomorphic energy, which represents the sink, is used as a risk amplification factor and multiplied by the current dynamic runoff coefficient and the upstream total effective runoff, which represents the source, determined by the rainfall of the current grid cell, thereby dynamically calculating the actual flood energy collected at the current point.

[0065] Preferably, the current runoff intensity is obtained by multiplying the current geological energy with the total upstream runoff determined by the current dynamic runoff coefficient and the previous influencing rainfall, including:

[0066] The current flow strength satisfies the following expression:

[0067] ;

[0068] In the formula, F represents the current runoff intensity in cubic meters; P represents the current geological energy, which is dimensionless; A is the current antecedent rainfall per unit of the current grid cell in meters, converted from millimeters; and G is the current dynamic runoff coefficient, which is dimensionless. This represents the area of ​​the current grid cell, in square meters.

[0069] In the formula, The calculation of this item is the effective runoff depth of the current grid cell, which represents the actual depth of rainwater that is converted into surface runoff. Its product with the area of ​​the current grid cell is the runoff of the current grid cell. By summing up the runoff of all units in the upstream catchment area, the total effective runoff volume upstream of the current grid unit is obtained, which serves as the basis for the water volume driving the downstream flood. By multiplying the current geomorphic energy, which represents the confluence capacity, with the water volume base that drives downstream floods, a risk amplification model that better reflects the physical process is constructed, namely the current confluence intensity, which intuitively characterizes the flood energy collected in the current grid cell.

[0070] For example, suppose the flow rates of all cells within the entire catchment area upstream of the current grid cell are summed, i.e. ,in The precipitation is 4,500,000,000 cubic millimeters, which translates to 4,500 cubic meters. Consider a scenario where this 4,500 cubic meters of precipitation falls on the aforementioned gentle slope area, where the current geological energy P = 0.0012, and the area was previously relatively dry. The calculated current dynamic runoff coefficient is 0.2. Therefore, the current runoff intensity for this area is calculated as follows: Another scenario exists: the same 4500 cubic meters of precipitation falls in the upstream area of ​​the aforementioned gully, where the terrain energy P=0.15, and the area is already saturated due to previous rainfall. The calculated current dynamic runoff coefficient G=0.9. Therefore, the runoff intensity F at the gully mouth is calculated as follows: As can be seen, after considering the influence of surface saturation, the difference in runoff intensity between higher and lower current runoff intensities is significantly amplified, greatly enhancing the model's ability to identify real flood risks. The above calculation results... , All values ​​are rounded to two decimal places.

[0071] S3: Calculate the average and standard deviation of the current runoff intensity of all grid cells within the target warning area, denoted as the target runoff intensity mean and target runoff intensity standard deviation; calculate the current runoff velocity based on the regulating effect of dynamic hydrological conditions on the static base velocity determined solely by topography.

[0072] It should be noted that the confluence velocity of a real flood event is not constant, but is determined by both the basic driving force provided by the terrain and the real-time hydrological pressure. This invention no longer relies on parameters such as the basic flow velocity that need to be manually calibrated in traditional hydrological models, but proposes a dynamic flow velocity calculation method that is entirely driven by terrain and real-time hydrological conditions. For example, in a channel with the same slope, the greater the volume of water collected upstream, the stronger the resulting head pressure, which leads to a significant increase in confluence velocity.

[0073] Specifically, the average and standard deviation of the current runoff intensity of all grid cells within the target warning area are calculated, denoted as the target runoff intensity mean and target runoff intensity standard deviation; based on the regulating effect of dynamic hydrological conditions on the static base flow velocity determined solely by topography, the current runoff velocity is calculated, including:

[0074] The current flow rate satisfies the following expression:

[0075] ;

[0076] In the formula, This indicates the current flow velocity, expressed in meters per second. It represents the acceleration due to gravity and is a universal physical constant; This represents the side length of a standard terrain grid with uniform grid resolution, in meters. This indicates the current situation; Indicates the terrain slope of the current grid cell; This indicates the current current flow intensity of the current grid cell, in cubic meters. , Indicates the mean and standard deviation of the target flow intensity; The function represents a nonlinear mapping function; It represents a very small positive number, and guarantees that the denominator is not 0.

[0077] In the formula, This indicates that by introducing gravitational acceleration and resolution, an objective natural velocity scale with velocity dimensions has been constructed. Indicates passage and As a dimensionless topographic driving factor, this natural velocity scale is modulated to obtain a static base flow velocity determined by topography that is unaffected by instantaneous water volume. This represents a dimensionless dynamic hydrological regulation coefficient, where, The function will set the current raster cell's The relative prominence across the entire basin is nonlinearly mapped to an adjustment factor used to amplify the flow velocity in high-risk areas; This means that the current confluence velocity is obtained by multiplying the static base velocity by the dynamic hydrological adjustment coefficient, making the current confluence velocity a dynamic physical quantity with complete physical meaning that does not require manual parameter setting and can adaptively respond to real-time hydrological changes.

[0078] For example, suppose the raster resolution of the terrain dataset is... 30 meters, gravitational acceleration The value is 9.8 meters / For a terrain slope The current grid cell, its slope term The value is 0.3, and its geomorphic energy is calculated. The average target runoff intensity in the current watershed cubic meters, standard deviation cubic meters, then the calculated static foundation velocity The current flow rate is 1.03 m / s; however, there is a case where the current grid cell is located in an area with low runoff intensity, where F is 31.5 cubic meters. meters per second; another scenario exists where the unit is located in a region of extremely high current density. cubic meter, m / s. It is evident that this invention achieves dynamic adaptation of the confluence velocity to the real-time confluence intensity, eliminating to some extent the dependence on a manually set base velocity coefficient. The above calculation results... , To retain two decimal places.

[0079] S4: Input the current flow velocity of all grid cells within the target warning area into the distributed hydrological model to issue a warning.

[0080] It should be noted that the ultimate goal of this invention is to achieve objective and automated flash flood disaster early warning through a computational framework that is entirely data-driven and constrained by physical processes. To this end, this invention has completed the standardized processing of raw meteorological and hydrological data, followed by progressive calculations of current geological energy, current dynamic runoff coefficient, current runoff intensity, and current runoff velocity obtained through the collaborative analysis of both. Finally, this invention will utilize these intermediate calculation results to construct an early warning judgment system that requires no manual threshold setting and is capable of self-learning and evolution, thus forming a complete flood early warning data management and processing system serving the collaborative analysis of meteorology and hydrology.

[0081] Specifically, the current flow velocity of all grid cells within the target warning area is input into the distributed hydrological model for warning purposes, including:

[0082] The current flow velocity of all grid cells within the target warning area is calculated and input into the distributed hydrological model. The output peak flow value and its growth rate are compared with the historical data of each grid cell. When any of the indicators exceeds the historical record, an early warning is automatically triggered.

[0083] Thus, the early warning system for flash floods based on the coordinated efforts of meteorology and hydrology was completed.

[0084] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.

[0085] While this specification has shown and described numerous embodiments of the invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Many modifications, alterations, and alternatives will occur to those skilled in the art without departing from the spirit and essence of the invention. It should be understood that various alternatives to the embodiments of the invention described herein may be employed in the practice of this invention.

Claims

1. A flash flood disaster early warning method based on meteorological hydrological collaborative analysis, characterized in that, The method comprises the following steps: Obtaining and processing a terrain dataset and a rainfall data sequence as model inputs for a target warning area; Based on the terrain variation characteristics of the target warning area, a plurality of terrain geometric parameters for calculating terrain potential energy are obtained, including terrain slope, upstream grid number, upstream elevation value, and upstream cell distance; the current terrain potential energy is calculated according to the mutual influence of the current grid cell and the surrounding terrain conditions; based on the rainfall data sequence, the unit pre-impact rainfall, the pre-impact target rainfall mean value of all grid cells, and the pre-impact target rainfall standard deviation are obtained by accumulation; the soil moisture content is evaluated according to the amount of pre-impact rainfall of the current grid cell and the difference between it and the pre-impact target rainfall mean value, and the current dynamic runoff coefficient is calculated; and the current flow concentration intensity is obtained by multiplying the current terrain potential energy and the total upstream runoff determined by the current dynamic runoff coefficient and the pre-impact rainfall. The average value and the standard deviation of the current flow concentration intensity of all grid cells in the target warning area are calculated, which are denoted as the target flow concentration intensity mean value and the target flow concentration intensity standard deviation. The current flow concentration velocity is calculated according to the adjusting effect of the dynamic hydrological conditions on the static basic flow velocity determined only by the topography, which satisfies the expression: ； wherein, represents the current flow velocity, with the unit of meter per second; represents the gravity acceleration, which is a universal physical constant; represents the standard terrain grid length with uniform grid resolution, with the unit of meter; represents the current terrain potential; represents the current terrain slope of the grid cell; represents the current flow intensity of the current grid cell, with the unit of cubic meter; , represents the target flow intensity mean value and the target flow intensity standard deviation; the function represents a nonlinear mapping function; represents a minimum positive number, which ensures that the denominator is not 0; represents a natural velocity scale with the dimension of velocity, which is constructed by introducing the gravity acceleration and the resolution; represents the terrain slope of the current grid cell; and as a dimensionless terrain driving factor, modulate the natural velocity scale, and together obtain a static basic flow rate determined by the terrain and landform, which is not affected by the instantaneous water volume; represents a dimensionless dynamic hydrological adjustment coefficient;​ The current flow concentration velocity of all grid cells in the target warning area is input into the distributed hydrological model for warning. 2.The flash flood early warning method based on meteorological and hydrological collaborative analysis according to claim 1, characterized in that, The method comprises the following steps: The system performs coordinate projection transformation and spatial resampling on the digital elevation model data of the target warning area to generate a standard terrain grid with uniform grid resolution, and then obtains the terrain dataset; at the same time, the rainfall estimation data of the target warning area is obtained from the center of the weather radar data within a preset time, and the planar radar data and the rainfall station data are parsed; the difference between the rainfall estimation data and the rainfall station data is calculated and spatial interpolation is performed to generate an error distribution field; and correction is performed to obtain the rainfall data sequence of the target warning area by performing the same coordinate projection transformation and spatial resampling as the terrain dataset. 3.The flash flood early warning method based on meteorological and hydrological collaborative analysis of claim 1, wherein, The method comprises the following steps: The elevation data corresponding to each grid cell of the target warning area is read, and the elevation change rates in the horizontal and vertical directions are calculated with the current grid cell as the center; then the terrain slope set is obtained by taking the square root of the two values; the maximum slope direction between each grid cell and its adjacent cells in the terrain dataset is analyzed to construct a global flow direction network, and then all upstream neighbors directly flowing into the current grid cell are identified according to the network and the number is counted; the elevation values are read to obtain the upstream grid number and the upstream elevation value of the current grid; the center point of the current grid cell and the center point of a single upstream grid cell are calculated, which are denoted as the upstream cell distance.

4. The flash flood early warning method based on meteorological and hydrological collaborative analysis according to claim 1, characterized in that, The current terrain potential energy satisfies the following expression: ； In the formula, P represents the current terrain potential energy, which is dimensionless; N represents the upstream grid number, which is dimensionless; U represents the upstream elevation value of the current grid cell, which has a height dimension and a meter unit; C represents the elevation value of the current grid cell, which has a height dimension and a meter unit; D represents the upstream cell distance, which has a meter unit; and S represents the upstream grid cell slope in the terrain slope set of the current grid cell, which is dimensionless.

5. The flash flood early warning method based on meteorological and hydrological collaborative analysis according to claim 1, characterized in that, The accumulated unit previous period influence rainfall, the previous period target rainfall mean of all grid units, and the previous period target rainfall standard deviation are obtained, including: Based on the rainfall data sequence, the rainfall of a single grid unit in a historical period is screened and accumulated to obtain the unit previous period influence rainfall, and the previous period influence rainfall mean and standard deviation of all grid units in the target early warning area are calculated, denoted as the previous period target rainfall mean and the previous period target rainfall standard deviation.

6. The flash flood early warning method based on meteorological and hydrological collaborative analysis according to claim 1, characterized in that, The current dynamic runoff coefficient satisfies the following expression: ； In the formula, G is the current dynamic runoff coefficient, dimensionless; A is the unit antecedent impact rainfall of the current grid cell, unit: mm; is the mean of the antecedent target rainfall, unit: mm; is the standard deviation of the antecedent target rainfall, unit: mm; 0.5, 1 play the role of scaling and shifting the output value of the function, so that G is in [0, 1]; Function of scaling and shifting the output value of the function, so that G is in [0, 1]; Indicates a very small positive number, to ensure that the denominator is not 0; The function represents a nonlinear mapping function.

7. The flash flood early warning method based on meteorological and hydrological collaborative analysis according to claim 1, characterized in that, The current concentration intensity satisfies the following expression: ； In the formula, F represents the current confluence intensity, with the unit of cubic meter; P represents the current terrain potential, with no dimension; A is the unit pre-impact rainfall of the current grid unit, with the unit of meter, which is converted from millimeter; G is the current dynamic runoff coefficient, with no dimension; is the area of the current grid unit, with the unit of square meter. 8.The flash flood warning method based on meteorological and hydrological collaborative analysis of claim 1, wherein, The current concentration speed of all grid units in the target early warning area is input into the distributed hydrological model for early warning, including: The current concentration speed of all grid units in the target early warning area is calculated and input into the distributed hydrological model, and the output flood peak flow value and its growth rate are compared with the historical data of each grid unit, and when any index exceeds the historical record, the early warning is automatically triggered.

9. A flash flood disaster early warning system based on meteorological hydrological collaborative analysis, characterized in that, Including: A processor and a memory, the memory stores computer program instructions, when the computer program instructions are executed by the processor, a mountain torrent disaster early warning method based on meteorological and hydrological collaborative analysis according to any one of claims 1-8 is realized.

Images