Method for simulating and forecasting flood in cold and cold mountainous area based on hydrological and hydrodynamic coupling

Abstract

The invention discloses a method for simulating and forecasting flood in a cold highland area based on hydrological and hydrodynamic coupling, and belongs to the technical field of disaster forecasting. The method specifically comprises the following steps: S1, multi-source basic data collection and preprocessing: collecting multi-type and multi-scale basic data for a target cold and cold mountainous area drainage basin; and S2, deep learning correction and fusion of the satellite rainfall data: aiming at the local overestimation and underestimation problems of the satellite rainfall data, a deep learning algorithm is adopted to carry out hour scale correction and fusion. Four types of core data of satellite remote sensing, reanalysis, ground observation and geographic space are collected, total factors of'rainfall-runoff-terrain-underlying surface 'required by flood simulation in the cold and cold mountainous area are covered, simulation one-sidedness caused by lack of data types in traditional modeling is avoided, rainfall and runoff abnormal values are eliminated by adopting a 3-sigma criterion, data formats and spatial-temporal scales are unified, and the modeling efficiency is improved. A standardized data set is formed, and interference of abnormal values, format incompatibility and space-time mismatching on subsequent model input is avoided.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of disaster prediction, and in particular to a high-cold mountainous area flood simulation and prediction method based on hydrological and hydrodynamic coupling. BACKGROUND

[0002] The high-cold mountainous area in Tibet is located in the core area of the Qinghai-Tibet Plateau, and the underlying surface conditions are extremely complex, with characteristics such as wide distribution of glaciers, development of permafrost, and large topographic drop. Moreover, the storm flood runoff process is influenced by superimposed factors such as glacier meltwater and permafrost freeze-thaw, and is significantly different from the flood mechanism in ordinary mountainous areas. At the same time, the meteorological and hydrological observation stations in this region are sparse, and the measured data of precipitation, runoff and the like are scarce, resulting in three major problems of the existing hydrological model (such as the traditional TOPMODEL and SWAT model): the precision of multi-source data (satellite precipitation, reanalysis runoff) is insufficient, which cannot directly support the model input; the influence of glacier meltwater and permafrost freeze-thaw on runoff generation is not considered, and the simulation error of runoff generation is large; the hydrological model is disconnected from the hydrodynamic model, and the flood concentration process and inundation range cannot be accurately simulated. SUMMARY

[0003] The purpose of the present application is to provide a high-cold mountainous area flood simulation and prediction method based on hydrological and hydrodynamic coupling, which can solve the problems of insufficient precision of multi-source data, not considering the influence of glacier meltwater and permafrost freeze-thaw on runoff generation, and disconnection between the hydrological model and the hydrodynamic model in the prior art.

[0004] Technical scheme: In order to solve the above technical problems, according to one aspect of the present application, more specifically, a high-cold mountainous area flood simulation and prediction method based on hydrological and hydrodynamic coupling, the method specifically comprises the following steps: S1, multi-source basic data collection and preprocessing: collecting multi-type and multi-scale basic data for the target high-cold mountainous area basin; S2, deep learning correction and fusion of satellite precipitation data: in view of the local overestimation and underestimation problems of satellite precipitation data, a deep learning algorithm is used for hour-scale correction and fusion; S3, GloFAS reanalysis runoff data correction and BTOP model parameter initial calibration: in view of the problem of scarcity of measured runoff data in the high-cold mountainous area, a random forest algorithm is used to correct the GloFAS reanalysis runoff data, and the corrected data is used for initial calibration of the BTOP hydrological model parameters; S4, optimization of initial conditions of soil water content: the correlation between the soil water content of the BTOP model and the ERA5-Land soil water content is established; S5, construction of improved BTOP hydrological runoff generation model: in view of the influence of glacier meltwater and permafrost freeze-thaw on runoff generation in the high-cold mountainous area, the improved BTOP runoff generation model is formed by coupling the snowmelt and ice melt runoff generation sub-model and the soil freeze-thaw runoff generation sub-model on the basis of the optimized BTOP model; S6, Hydrology-hydrodynamic coupling model construction and confluence simulation: a distributed hydrological model is coupled with a two-dimensional hydrodynamic model to realize accurate simulation of flood confluence process and inundation range; S7, Model verification and flood forecasting application: ensure model reliability through multi-dimensional verification, and build multi-time flood forecasting system based on the model.

[0005] Further, in step S1, the multi-source basic data includes: Satellite remote sensing data: near real-time satellite precipitation data, MODIS land cover data, and TPPG2017 glacier survey catalog data; Reanalysis data: GloFAS daily runoff data, ERA5-Land layered soil moisture data; Ground observation data: hourly precipitation data of weather stations in the basin, measured runoff data of hydrological stations, and soil freezing threshold temperature observation data; Geospatial data: digital elevation model of the basin, land use type map, and soil type map; The above data is formatted, spatially matched, and abnormal values of precipitation and runoff are removed using the 3σ criterion to form a standardized data set.

[0006] Further, in step S2, the specific method for deep learning correction and fusion of satellite precipitation data is: using hourly precipitation data of ground weather stations as true values, constructing an LSTM correction model for satellite precipitation data and measured precipitation data, using Adam optimizer in the training process, and using mean square error as the loss function. After correcting the three types of satellite precipitation data, weighted fusion method is used to generate hourly scale and 1km resolution fused precipitation data, and classification evaluation index and quantitative evaluation index are used to verify the applicability of the fused precipitation data.

[0007] Further, in step S3, the specific steps for GloFAS reanalysis runoff data correction and BTOP model parameter initial calibration are: S31, Trend analysis and error evaluation are performed on the original GloFAS daily runoff data to identify data bias characteristics; S32, using the measured daily runoff data of limited hydrological stations in the basin as the dependent variable, and GloFAS original runoff data, DEM slope, and land use type as independent variables, a random forest correction model is constructed, and correction is performed on time scale, spatial scale, and time-space coupled scale; S33, the corrected GloFAS runoff data is used as the observation value, which is input into the BTOP model, and the SCE-UA optimization algorithm is used for parameter initial calibration, and the objective function is NSE maximization. After initial calibration, the model NSE should not be less than the target value.

[0008] Further, in step S4, the specific method for optimizing the initial condition of soil moisture content is as follows: the key soil moisture content variables of the BTOP model and the ERA5-Land 4-layer soil moisture content data are extracted, multi-dimensional correlation analysis is performed, variable combinations with a correlation coefficient not lower than a threshold value are screened, a traditional curve fitting function and an LSTM network are respectively used to construct the mapping relationship of the screened variable combinations, the ERA5-Land soil moisture content is taken as the input and the BTOP soil moisture content is taken as the output, a mapping model with a higher NSE is selected through cross-validation, the soil moisture content calculated by the optimal mapping model is used to replace the initial soil moisture content of the BTOP model, and the optimization of the initial condition of the model is completed.

[0009] Further, in step S5, when the BTOP hydrological runoff model is improved, the TPPG2017 glacier data and the MODIS snow cover data are taken as initial conditions, the degree-day factor method is used to calculate the glacier / snow melt water, the temperature threshold method is used to divide the precipitation form, the melt water and the rainfall water are combined as total runoff input, the snow and ice runoff sub-model is constructed, the LSTM freeze-thaw runoff model is developed by using Python language, the input layer is the daily average air temperature and the precipitation, and the output layer is the runoff in the frozen soil region; when the daily average air temperature is greater than the soil freezing threshold, the original storage-full runoff mechanism of the BTOP is used for calculation; when the daily average air temperature is less than or equal to the soil freezing threshold, the excess infiltration runoff mechanism is used for calculation, the surface runoff is directly generated, the soil freeze-thaw runoff sub-model is constructed, the Python and FORTRAN hybrid programming technology is used to embed the two sub-models into the BTOP runoff module, and the model is verified through multiple historical flood events.

[0010] Further, step S6 specifically includes the following steps: S61, using the improved BTOP hydrological runoff model, calculating the total runoff of the basin at a 1km grid scale, using the concentration module of the BTOP to calculate the concentration flow of each grid to the river, and outputting the flow process of the target section of the river; S62, developing a two-dimensional hydrodynamic model by using the FORTRAN language, discretizing the rivers in the basin into single-line one-dimensional rivers, and retaining two-dimensional grids for the flooded risk area; S63, taking the river section flow output by the improved BTOP model as the upstream boundary condition of the two-dimensional hydrodynamic model and loading it to the corresponding river node; solving the two-dimensional diffusion wave equation by using the finite volume method to simulate the evolution process of the flood in the river and the water level and flow velocity distribution of the flooded area, and verifying the coupled model by using the historical flood inundation range.

[0011] Further, step S7 specifically includes the following steps: S71, verifying the accuracy of the sub-model by using field investigation data, and verifying the accuracy of the flow simulation and the inundation simulation of the coupled model by using historical flood events; S72, based on the improved BTOP-hydrodynamic coupling model, short-term, medium-term and long-term flood forecasting modules are constructed with the fusion precipitation data as input; S73, in combination with the "mountain flood gene library", the high-risk areas in the basin are identified through the flood depth and flow rate data output by the model, the prediction results and risk zoning map are generated, and the flood warning is supported.

[0012] Beneficial effects: By collecting satellite remote sensing, reanalysis, ground observation and geographic space 4 types of core data, covering the "precipitation-runoff-terrain-underlying surface" full element required for flood simulation in alpine mountainous areas, avoiding the one-sidedness of traditional modeling due to the lack of data types, using 3σ criterion to eliminate precipitation and runoff outliers, unifying data format and spatiotemporal scale, forming a standardized data set, avoiding the interference of outliers, incompatible formats and spatiotemporal mismatches on subsequent model inputs, and laying a reliable data foundation for the whole process modeling.

[0013] In view of the defects of satellite precipitation such as GSMaP, IMERG, PERSIANN and others, high-precision precipitation data is generated by correcting the hourly scale through the LSTM model and then by weighted fusion, generating 1km resolution hourly scale precipitation data, adapting to the simulation needs of "small range heavy precipitation" in complex terrain in alpine mountainous areas, solving the problem that traditional low-resolution precipitation data cannot depict local heavy rain, and providing accurate precipitation input for subsequent runoff simulation.

[0014] In view of the pain points of few hydrological stations and insufficient measured runoff data in alpine mountainous areas, GloFAS reanalysis runoff data is corrected in time / space / space scale through random forest algorithm, so that the corrected data can replace the measured data for model calibration, SCE-UA optimization algorithm is used, and BTOP model initial calibration is carried out with corrected GloFAS data as observation value, avoiding the model deviation caused by traditional "empirical setting of parameters when there is no measured data", and providing reliable initial parameters for subsequent model optimization.

[0015] Through multi-dimensional correlation analysis, the optimal variable combination of BTOP model soil moisture and ERA5-Land layered soil moisture is selected, and then the mapping model with higher NSE is selected through "curve fitting + LSTM network" comparison, and the calculation result is used to replace the initial soil moisture of BTOP to solve the problem of traditional model "empirical assignment of initial soil moisture", so that the initial parameters of the model are more consistent with the actual hydrological conditions of the basin, avoiding the transmission of initial condition error to subsequent runoff and concentration simulation, and improving the reliability of the overall simulation results.

[0016] Coupling "snow and ice melt sub-model", accurately calculating the amount of glacial / snow melt water, solving the problem of traditional model ignoring the contribution of melt water, coupling "soil freeze-thaw sub-model", dynamically simulating the conversion of "storage and runoff" and "excess seepage runoff" in frozen soil area, fitting the actual situation of high-cold frozen soil, embedding the two sub-models into BTOP through Python and FORTRAN hybrid programming, and verifying through multiple historical flood scenes, effectively reducing the error rate.

[0017] Improving the calculation of 1km grid scale confluence flow of BTOP model, outputting the process of river target section flow, ensuring the accuracy of confluence data, discretizing the river into one-dimensional through two-dimensional hydrodynamic model (FORTRAN development), keeping the two-dimensional grid in submerged area, solving the defect of traditional hydrological model "only outputting flow, unable to depict the submerged range", providing comprehensive data of "flow + spatial distribution" for flood risk zoning and risk avoidance route planning.

[0018] Based on the coupled model, three-level forecasting modules of short-term, medium-term and long-term are constructed, the sub-models are verified through field investigation data, and the coupled model is verified through historical flood, ensuring the reliability of the forecasting results and improving the practicality of disaster prevention: combined with "mountain flood gene library" to identify high-risk areas, assist local departments to quickly start emergency response and deploy flood control materials, directly support the whole process of flood "forecast, early warning, pre-play, pre-plan", effectively reduce personnel casualties and property losses. BRIEF DESCRIPTION OF DRAWINGS

[0019] Figure 1 is a method flow diagram. DETAILED DESCRIPTION

[0020] In order to make the technical scheme of the present application clearer, the present application will be further described in detail below in combination with the drawings and specific embodiments.

[0021] Embodiment 1 Step 1, multi-source basic data collection and pretreatment Data type: Satellite remote sensing data: ① Near real-time satellite precipitation: GSMaP (JAXA website, 0.1° resolution), IMERG (NASA website, 0.1° resolution), PERSIANN (UCI website, 0.25° resolution); ② MODIS land cover data (MOD13Q1, 1km resolution, NASA); ③ TPPG2017 glacier survey catalog (International Mountain Comprehensive Development Center).

[0022] Reanalysis data: ① GloFAS daily runoff data (ECMWF website, 0.1° resolution); ② ERA5-Land soil moisture data (Copernicus Climate Data Store, 0.1° resolution, 4 layers: 0-7cm, 7-28cm, 28-100cm, 100-289cm).

[0023] Ground monitoring data: ① Weather stations: Lhasa station (2022 July hourly precipitation, 0.1mm precision), Damxung station (same period hourly precipitation); ② Hydrological station: Lhasa hydrological station (2022 July daily runoff, 1m³ / s precision); ③ Soil freezing threshold: -2.0℃ measured at Damxung permafrost observation station.

[0024] Geospatial data: ① Digital Elevation Model (DEM, 30m resolution, Geospatial Data Cloud); ② 2020 land use type map (1km resolution, Chinese Academy of Sciences Resource Environment Data Center); ③ Soil type map (1:1 million, Ministry of Natural Resources).

[0025] Preprocessing operations: Format unification: Convert all data to NetCDF format (convenient for model calling); Temporal and spatial matching: Resample spatial resolution to 1km (using bilinear interpolation), and unify time scale to hourly (precipitation) / daily (runoff, soil moisture); Outlier removal: Use 3σ criterion to remove data with Lhasa station hourly precipitation > 50mm (extreme anomaly) and Lhasa hydrological station daily runoff > 2000m³ / s (exceeding historical extreme value), finally forming a standardized data set.

[0026] Step 2, Deep learning correction and fusion of satellite precipitation data LSTM correction model construction True value data: Lhasa and Damxung station hourly precipitation data from January to June 2022 (4320, 70% for training, 30% for verification); Model structure: Input layer (3 neurons corresponding to GSMaP, IMERG, PERSIANN precipitation) → Hidden layer (2 layers, 64 neurons each, ReLU activation function) → Output layer (1 neuron, corrected hourly precipitation); Training parameters: Adam optimizer (learning rate 0.001), loss function is mean square error (MSE), iteration times 100, batch size 32.

[0027] Correction effect verification The Nash-Sutcliffe efficiency (NSE) of the three types of satellite precipitation data after correction is: IMERG 0.81, GSMaP 0.77, and PERSIANN 0.73, which is more than 30% higher than the original data (NSE 0.52-0.65).

[0028] Weighted fusion Based on the NSE of the verification period, the weights are determined as follows: IMERG (0.4), GSMaP (0.35), and PERSIANN (0.25). The fused precipitation data with 1 km resolution and hourly scale is generated, and the NSE of the verification period reaches 0.82, with the root mean square error (RMSE) reduced to 4.2 mm / h.

[0029] Step 3: GloFAS reanalysis runoff data correction and BTOP model parameter initial calibration 1. Error analysis of original GloFAS Comparing the measured runoff at Lhasa Hydrological Station in 2020-2021, the original GloFAS daily runoff data is systematically 15.3% lower, and the peak time lags 2-3 days, so it needs to be corrected at multiple scales.

[0030] 2. Random forest correction model Independent variables: original GloFAS daily runoff, DEM slope (1 km grid average), and land use type (grassland / bare land / cropland code); Dependent variable: measured daily runoff at Lhasa Hydrological Station in 2020-2021 (total of 730, 80% training, 20% verification); Model parameters: 100 decision trees, maximum tree depth 10, 5-fold cross-validation, corrected time scale NSE 0.78, spatial scale (sub-basin) NSE 0.75, and spatiotemporal coupling scale NSE 0.80.

[0031] 3. BTOP model initial calibration Model parameter range: soil saturated hydraulic conductivity (1-10 mm / h), soil porosity (0.3-0.5), and surface roughness (0.02-0.05); Optimization algorithm: SCE-UA (mixed complex evolutionary algorithm), objective function is to maximize NSE; Initial calibration results: model NSE reaches 0.68, peak error is 12.5%, which meets the basic requirements for subsequent optimization.

[0032] Step 4: Optimization of initial soil moisture conditions Variable correlation analysis Extract the key variables of the BTOP model (root zone soil moisture Srz, unsaturated zone soil moisture Suz, and saturated soil water deficit SD) and ERA5-Land 4-layer soil moisture, calculate the Pearson correlation coefficient: Srz and ERA5-Land layer 2 (7-28 cm) correlation coefficient 0.72; Suz and ERA5-Land layer 3 (28-100 cm) correlation coefficient 0.68; Screening of the above two groups of variable combinations for mapping relationship construction.

[0033] Mapping model comparison Traditional curve fitting: power function (y=ax^b), Srz mapping R²0.65, Suz mapping R²0.62; LSTM network: input layer (4 neurons, ERA5-Land 4 layers of water content) → hidden layer (1 layer of 32 neurons) → output layer (2 neurons, Srz, Suz), cross-validation (8:2) NSE 0.76 (Srz), 0.73 (Suz); Select LSTM model as the optimal mapping scheme.

[0034] Initial condition replacement Replace the default initial value of the BTOP model with Srz and Suz calculated by LSTM, and the model NSE improves to 0.71 after re-calibration, and the initial condition error decreases by 28%.

[0035] Step 5, improve the construction of BTOP hydrological runoff model Snow and ice melt runoff sub-model Initial conditions: TPPG 2017 glacial data (determine glacial spatial distribution), MODIS snow cover data (MOD10A1, 8-day synthesis, 1km resolution); Degree-day factor: based on the observation of glacial melt water in 2019-2021 in Dangxiong Station, determine the ice melt degree-day factor 6.5mm / (℃・d), snow melt degree-day factor 3.2mm / (℃・d); Precipitation form division: 0℃ temperature threshold (measured at Dangxiong Station), air temperature ≥0℃ as rainfall, <0℃ as snowfall, and melt water and rainfall combined as total runoff input.

[0036] Soil freeze-thaw runoff sub-model Model development: Python language constructs LSTM model, input layer (daily average temperature, daily precipitation) → hidden layer (2 layers of 48 neurons) → output layer (frozen soil area runoff); Runoff mechanism switching: when the daily average temperature is >-2.0℃ (soil freezing threshold), calculate according to the original BTOP storage and runoff; ≤-2.0℃, calculate according to the excess infiltration runoff (precipitation directly forms surface runoff); Module coupling: Python (sub-model) and FORTRAN (BTOP core) are used in mixed programming, and sub-model embedding is achieved through dynamic link libraries (DLLs).

[0037] Model Validation The improved BTOP model was validated using the Lhasa River flood in September 2018 (measured peak flow of 980 m³ / s). The NSE of the improved model was 0.75, which is 21% lower than the original model (NSE 0.61).

[0038] Step 6: Construction of hydrological-hydrodynamic coupled model and confluence simulation 1. Hydrological model runoff calculation The improved BTOP model calculates the total runoff at a 1km grid scale. It uses the topographic index confluence module built into BTOP to calculate the confluence flow from each grid to the Lhasa hydrological station section and outputs the flood flow process in July 2022 (simulated peak flow 1250 m³ / s, measured 1280 m³ / s, error 2.3%).

[0039] 2. Construction of a two-dimensional hydrodynamic model Model development: Written in FORTRAN, based on two-dimensional diffuse wave equations, using unstructured mesh; Grid discretization: The main stream of the Lhasa River (from its source to the Lhasa Hydrological Station) is discretized into a single-line one-dimensional river channel (grid length 500m), while the flood risk area (surrounding Lhasa city and Duilongdeqing District) retains a two-dimensional grid (500m×500m). Parameter settings: Manning roughness coefficient (0.03 for river channels, 0.05 for grasslands, and 0.04 for cultivated land), time step 60s (to meet CFL conditions).

[0040] 3. Model Coupling and Simulation Boundary conditions: The Lhasa hydrological station cross-sectional flow rate output by the improved BTOP model is used as the upstream boundary of the two-dimensional hydrodynamic model, and the downstream boundary is the Lhasa River estuary tidal level (measured, average tidal level in July 2022 is 3590.2m). Numerical solution: The finite volume method is used to solve the two-dimensional diffusion wave equation to simulate the evolution of flood in the river channel and the distribution of water level and velocity in the inundation area; Verification: The flooding range interpreted by Sentinel-2 satellite (July 17, 2022, 10m resolution) showed a simulation overlap rate of 86% and a maximum flooding level error of 0.25m.

[0041] Step 7: Model Validation and Flood Forecasting Application 1. Multi-dimensional verification Sub-model validation: The meltwater volume error of the snow accumulation sub-model was 10%, and the flow rate error of the freeze-thaw sub-model was 8%. Coupling model validation: NSE 0.83, overlap rate of inundation area 86%, both meet the accuracy requirements of flood simulation in alpine mountainous areas.

[0042] 2. Multi-time forecast module construction Short-term forecast (0-12h): input fusion precipitation data (12h hourly forecast in the future), model calculation time 45 minutes, forecast peak error 4.8%; Medium-term forecast (12-72h): input ECMWF numerical prediction precipitation (updated every 12h), forecast peak error 7.5%; Long-term forecast (72-168h): input China Meteorological Administration T639 model precipitation (updated every 24h), forecast peak error 11.2%.

[0043] 3. High-risk area identification and early warning Mountain flood gene library: integrate the topographic (slope > 25°) and hydrological (peak flood > 800m³ / s) characteristics of 12 mountain flood events in the Lhasa River Basin from 2015 to 2021; Risk identification: identify high-risk areas (southeastern Lhasa urban area and northern Duilongdêqên District) through flood depth (> 1.5m) and flow velocity (> 2.0m / s) data output by the coupling model; Early warning release: generate forecast results and risk zoning map, and release through the Tibet Autonomous Region Flood Control Command Platform, with a data input to early warning output time ≤1.5 hours. Example 2

[0044] First step, multi-source basic data collection and preprocessing Satellite remote sensing data: ① Satellite precipitation: GSMaP (JAXA, 0.1°), IMERG (NASA, 0.1°), PERSIANN (UCI, 0.25°); ② MODIS land cover (MOD13Q1, 1km, NASA); ③ TPPG2017 glacier inventory (International Center for Integrated Mountain Development).

[0045] Reanalysis data: ① GloFAS daily runoff (ECMWF, 0.1°); ② ERA5-Land soil moisture (Copernicus, 0.1°, 4 layers).

[0046] Ground observation data: ① Weather stations: Bayi Station (2020 August hourly precipitation), Gongbogê Station (same period hourly precipitation); ② Hydrological stations: Bayi Hydrological Station (2020 August daily runoff); ③ Soil freezing threshold: Gongbogê frozen soil station measured -2.5℃.

[0047] Geospatial data: ① DEM (30 m, Geospatial Data Cloud); ② Land use map in 2020 (1 km, Resource Science Center of Chinese Academy of Sciences); ③ Soil type map (1:1 million, Ministry of Natural Resources).

[0048] Preprocessing operation Format unification: converted to NetCDF format for easy model calling; Temporal and spatial matching: spatial resampling to 1 km (bilinear interpolation), and temporal unification to hourly (precipitation) / daily (runoff); Outlier removal: 3σ criterion to remove data with hourly precipitation > 60 mm (extreme rainstorm) and daily runoff > 1500 m³ / s (historical extreme value) at Bajie and Gongboged stations, forming a standardized data set.

[0049] Step 2, deep learning correction and fusion of satellite precipitation data LSTM correction model True value data: hourly precipitation at Bajie and Gongboged stations from January to July 2019 (5040 pieces, 70% training, 30% verification); Model structure: input layer (3 neurons, three types of satellite precipitation) → hidden layer (2 layers of 48 neurons, ReLU activation) → output layer (1 neuron, corrected precipitation); Training parameters: Adam optimizer (learning rate 0.0015), MSE loss function, 120 iterations, batch size 64.

[0050] Correction effect NSE of corrected satellite precipitation: IMERG 0.83, GSMaP 0.79, PERSIANN 0.75, improved by 32% compared with original data (NSE 0.55-0.67).

[0051] Weighted fusion Weights: IMERG (0.45), GSMaP (0.3), PERSIANN (0.25), NSE of fused precipitation data 0.85, RMSE 3.8 mm / h (1 km, hourly scale).

[0052] Step 3, correction of GloFAS reanalysis runoff data and initial calibration of BTOP model parameters Compared with the measured runoff at Bajie hydrological station from 2018 to 2019, the original GloFAS systemically overestimated by 12.1% and the flood peak lagged by 1-2 days.

[0053] Random forest correction Variables: independent variables (GloFAS runoff, DEM slope, land use), dependent variable (measured runoff at Bajie station, 730 pieces, 80% training); Model parameters: decision tree 120, maximum depth 12, 5-fold cross-validation Correction effect: time scale NSE 0.80, spatial scale NSE 0.76, spatiotemporal scale NSE 0.82.

[0054] BTOP initial rate Parameter range: soil saturated hydraulic conductivity (1.5-12 mm / h), porosity (0.32-0.52), surface roughness (0.025-0.055); After SCE-UA optimization, the model NSE is 0.71, and the flood peak error is 10.8%.

[0055] Step 4: Optimization of initial soil moisture conditions Correlation analysis BTOP variable correlation with ERA5-Land: Srz correlation coefficient with layer 2 (7-28 cm) 0.75, Suz correlation coefficient with layer 3 (28-100 cm) 0.70.

[0056] Comparison of mapping models Curve fitting (power function): Srz R² 0.68, Suz R² 0.65; LSTM model: input (4 layers of ERA5-Land moisture) → output (Srz, Suz), cross-validation NSE 0.79 (Srz), 0.76 (Suz); Select the LSTM mapping model.

[0057] Replacement of initial conditions After replacement, the NSE of the BTOP model is improved to 0.74, and the initial error is reduced by 31%.

[0058] Step 5: Improvement of BTOP hydrological runoff model construction Snow and ice melting sub-model Initial conditions: TPPG 2017 glacial data, MOD10A1 snow data; Degree-day factor: based on Gongbujiangda station observation, ice melting 7.0 mm / (℃・d), snow melting 3.5 mm / (℃・d); Precipitation form: 0℃ threshold, meltwater + rainfall as runoff input.

[0059] Soil freezing and thawing sub-model Python-LSTM model: input (daily temperature, daily precipitation) → output (frozen soil runoff), hidden layer 2 layers of 56 neurons; Runoff switching: daily temperature >-2.5℃ (freezing threshold) storage runoff, ≤-2.5℃ excess runoff; Coupling method: Python-FORTRAN hybrid programming, embedded BTOP through API interface.

[0060] Verification Verified by the flood in June 2019 (measured peak 820 m³ / s), improved BTOP runoff NSE 0.78, original model NSE 0.63, error reduced by 22%.

[0061] Step 6: Hydrology-hydrodynamic coupling model construction and confluence simulation Hydrological confluence calculation Improved BTOP in 1 km grid calculation runoff, confluence to Bayi hydrological station, output August 2020 flood discharge process (simulated peak 1030 m³ / s, measured 1050 m³ / s, error 1.9%).

[0062] Two-dimensional hydrodynamic model FORTRAN development, two-dimensional diffusion wave equation, unstructured grid; Discrete method: one-dimensional river channel (above Bayi hydrological station) for Niyang River (500 m grid), two-dimensional grid (500 m x 500 m) for submerged area (Bayi Town, Nyingchi Town); Parameters: Manning roughness (river 0.028, forest 0.055, grassland 0.045), time step 60 s.

[0063] Coupling simulation Boundary: upstream (Bayi hydrological station flow), downstream (Niyang River estuary water level 3012.5 m); Solution: finite volume method, simulate maximum water level 3.8 m (Bayi Town), flow velocity 2.1 m / s; Verification: Sentinel-2 remote sensing submerged range (August 7, 2020), coincidence rate 88%.

[0064] Step 7: Model verification and flood forecasting application Verification results Sub-model error: 8% of melt water, 7% of freeze-thaw runoff; Coupling model: flow NSE 0.85, submerged coincidence rate 88%.

[0065] Forecasting module Short-term (0-12h): error 4.5%, time-consuming 50 minutes; Medium-term (12-72h): error 7.2%; Long-term (72-168h): error 10.8%.

[0066] Risk warning Mountain flood gene library: characteristics of 8 mountain flood events from 2016 to 2021; High-risk area: the northeast of Bayi Town and the west of Linzhi Town (flood depth > 1.5 m, flow velocity > 2.0 m / s); Early warning output: pushed to the flood control office of Linzhi City within 2 hours, to assist in flood avoidance in August 2020, and to transfer more than 1,200 people.

[0067] The above-described embodiments only express several embodiments of the present application, and the description is more specific and detailed, but it should not be understood as a limitation on the scope of the patent of the present application. It should be noted that for ordinary skilled persons in the art, without departing from the concept of the present application, a number of modifications and improvements can be made, which are all within the scope of protection of the present application. Therefore, the protection scope of the patent of the present application should be subject to the appended claims.

Claims

1. A method for flood simulation and forecasting in high-altitude cold mountainous areas based on hydrological-hydrodynamic coupling, characterized in that, The method specifically includes the following steps: S1. Collection and preprocessing of multi-source basic data: Collect multi-type and multi-scale basic data for the target high-altitude and cold mountainous watershed; S2. Deep learning correction and fusion of satellite precipitation data: To address the problem of local overestimation and underestimation in satellite precipitation data, deep learning algorithms are used for hourly-scale correction and fusion. S3. GloFAS Reanalysis Runoff Data Correction and BTOP Model Parameter Initial Calibration: To address the scarcity of measured runoff data in high-altitude and cold mountainous areas, the random forest algorithm is used to correct GloFAS reanalysis runoff data and apply it to the initial calibration of BTOP hydrological model parameters. S4. Optimization of initial conditions for soil moisture content: Constructing the correlation between soil moisture content in the BTOP model and soil moisture content in ERA5-Land; S5. Improved BTOP hydrological runoff model construction: In view of the impact of glacial meltwater and permafrost freeze-thaw on runoff in high-altitude and cold mountainous areas, an improved BTOP runoff model is formed by coupling the snowmelt and ice melt runoff sub-model and the soil freeze-thaw runoff sub-model on the basis of the optimized BTOP model. S6. Construction of Hydrological-Hydrodynamic Coupled Model and Confluence Simulation: A distributed hydrological model and a two-dimensional hydrodynamic model are coupled to achieve accurate simulation of the flood confluence process and inundation range; S7. Model Validation and Flood Forecasting Application: Ensure model reliability through multi-dimensional validation, and build a multi-time-effect flood forecasting system based on the model.

2. The flood simulation and forecasting method for high-altitude cold mountainous areas based on hydrodynamic coupling as described in claim 1, characterized in that: In step S1, the multi-source basic data includes: Satellite remote sensing data: near real-time satellite precipitation data, MODIS land cover data, TPPG2017 glacier survey catalog data; Further analysis of data: GloFAS daily runoff data, ERA5-Land stratified soil moisture data; Ground observation data: hourly precipitation data from meteorological stations within the basin, measured runoff data from hydrological stations, and soil freezing threshold temperature observation data; Geospatial data: watershed digital elevation model, land use type map, soil type map; The above data were formatted, spatiotemporally matched, and precipitation and runoff anomalies were removed using the 3σ criterion to form a standardized dataset.

3. The flood simulation and forecasting method for high-altitude cold mountainous areas based on hydrodynamic coupling as described in claim 1, characterized in that: In step S2, the specific method for deep learning correction and fusion of satellite precipitation data is as follows: using hourly precipitation data from ground meteorological stations as the true value, an LSTM correction model of satellite precipitation data and measured precipitation data is constructed. The training process uses the Adam optimizer, with the mean square error as the loss function. After correcting the three types of satellite precipitation data respectively, a weighted fusion method is used to generate fused precipitation data with an hourly scale and a 1km resolution. The applicability of the fused precipitation data is verified using classification evaluation indicators and quantitative evaluation indicators.

4. The flood simulation and forecasting method for high-altitude cold mountainous areas based on hydrodynamic coupling as described in claim 1, characterized in that: In step S3, the specific steps for GloFAS reanalysis runoff data correction and BTOP model parameter initial calibration are as follows: S31. Perform trend analysis and error assessment on the raw GloFAS daily runoff data to identify data deviation characteristics; S32. Using the measured daily runoff data from a limited number of hydrological stations within the basin as the dependent variable, and the GloFAS raw runoff data, DEM slope, and land use type as independent variables, a random forest correction model was constructed and corrected at the time scale, spatial scale, and spatiotemporal coupling scale, respectively. S33. Use the corrected GloFAS runoff data as observations and input them into the BTOP model. Use the SCE-UA optimization algorithm for initial parameter calibration. The objective function is to maximize NSE. After initial calibration, the model NSE should not be lower than the target value.

5. The flood simulation and forecasting method for high-altitude cold mountainous areas based on hydrodynamic coupling as described in claim 1, characterized in that: In step S4, the specific method for optimizing the initial soil moisture content conditions is as follows: extract the key soil moisture content variables of the BTOP model and the soil moisture content data of the ERA5-Land4 layer, perform multi-dimensional correlation analysis, screen variable combinations with correlation coefficients not lower than the threshold, construct the mapping relationship of the screened variable combinations using traditional curve fitting functions and LSTM networks respectively, take the ERA5-Land soil moisture content as input and the BTOP soil moisture content as output, select the mapping model with higher NSE through cross-validation, and replace the initial soil moisture content of the BTOP model with the soil moisture content calculated by the optimal mapping model to complete the optimization of the initial conditions of the model.

6. The flood simulation and forecasting method for high-altitude cold mountainous areas based on hydrodynamic coupling as described in claim 1, characterized in that: In S5, when constructing the improved BTOP hydrological runoff model, TPPG2017 glacier data and MODIS snow cover data are used as initial conditions, and the degree-day factor method is used to calculate the glacier / snow cover meltwater. The temperature threshold method was used to classify precipitation patterns. Meltwater and precipitation were combined into the total runoff input to construct a runoff sub-model for snowmelt and icemelt. An LSTM freeze-thaw runoff model was developed using Python. The input layer is the daily average temperature and precipitation, and the output layer is the runoff in the permafrost region. When the daily average temperature is greater than the soil freezing threshold, the runoff generation mechanism is calculated according to the original BTOP full storage mechanism; when the daily average temperature is less than or equal to the soil freezing threshold, the runoff generation mechanism is calculated according to the infiltration excess mechanism, and surface runoff is directly generated. A soil freeze-thaw runoff generation sub-model is constructed. The above two sub-models are embedded into the BTOP runoff generation module using Python and FORTRAN hybrid programming technology. The model is verified through multiple historical flood events.

7. The flood simulation and forecasting method for high-altitude cold mountainous areas based on hydrodynamic coupling as described in claim 1, characterized in that: Step S6 specifically includes the following steps: S61. Using the improved BTOP hydrological runoff model, the total runoff of the watershed is calculated at a 1km grid scale. The runoff from each grid to the river channel is calculated using the BTOP built-in runoff module, and the flow process of the target section of the river channel is output. S62. A two-dimensional hydrodynamic model was developed using FORTRAN language. The river channels in the basin were discretized into single-line one-dimensional channels, and a two-dimensional grid was retained for the inundation risk zone. S63. The cross-sectional flow rate of the river channel output by the improved BTOP model is used as the upstream boundary condition of the two-dimensional hydrodynamic model and loaded onto the corresponding river node. The two-dimensional diffusion wave equation is solved by the finite volume method to simulate the evolution process of the flood in the river channel and the water level and velocity distribution in the inundation area. The coupled model is verified by the inundation range of historical floods.

8. The flood simulation and forecasting method for high-altitude cold mountainous areas based on hydrodynamic coupling as described in claim 1, characterized in that: Step S7 specifically includes the following steps: S71. Verify the accuracy of the sub-model using field survey data, and verify the accuracy of the coupled model's flow simulation and inundation simulation using historical flood events; S72. Using fused precipitation data as input, construct short-term, medium-term, and long-term flood forecasting modules based on the improved BTOP-hydrodynamic coupling model; S73. By combining the "mountain torrent gene bank" and using the flood depth and flow velocity data output by the model, high-risk areas within the basin are identified, and forecast results and risk zoning maps are generated to support the issuance of flood warnings.

Images