76 results about "SWAT model" patented technology

The invention relates to the technical field of data processing, and discloses a water quality and water quantity accurate simulation method and system based on an SWAT model and a storage medium. The method comprises the following steps: analyzing rainfall and evapotranspiration data through a two-parameter trend test, and generating standard hydrological data by adopting a triple configuration algorithm; dividing sub-basins by combining topographic information, and calculating a hydrological response unit; performing water balance operation and water conservation evaluation; pollutant diffusion is analyzed, and a diffusion equation is solved; running state monitoring is implemented, and an early warning system is designed; and comparing and correcting measured data, and constructing a hierarchical monitoring mechanism. According to the invention, the whole-process precise management from hydrological data acquisition, water quality evaluation to early warning monitoring is realized, and the accuracy and reliability of water quality and water quantity monitoring are remarkably improved.

The invention discloses a cold region SWAT model construction method considering a glacier melting process. A calculation method comprises the following steps: step 1, obtaining corrected rainfall data; corrected air temperature data are obtained; evapotranspiration correction data is obtained; actually measured runoff is obtained; 2, obtaining a plurality of sub-basin boundaries and spatial distribution of each sub-basin glacier; 3, calculating the glacier melting amount of each sub-drainage basin in the research time period by adopting a day-to-day factor method to obtain the day-by-day glacier melting amount of each sub-drainage basin; step 4, performing double-element parameter calibration based on the actually measured runoff data and the evapotranspiration correction data to obtain parameters suitable for the watershed hydrological process; and 5, calculating to obtain the relative contribution rate of the glacier melting process to runoff and evapotranspiration. The simulation precision of hydrological elements such as runoff of the cold region hydrological model is improved, and a scientific basis can be provided for sustainable utilization and planning of water resources of a local drainage basin.

The invention discloses a water quality prediction method based on machine learning and an SWAT model, and belongs to the technical field of water environment simulation and water quality early warning. The method comprises the steps that S1, hydrology, water quality, meteorology, land utilization types, pollution sources and spatial elevation multi-temporal data are collected from multiple channels in a drainage basin, and noise reduction, normalization and missing value filling preprocessing methods are adopted for original data; s2, performing sensitivity analysis on the nonlinear relationship between the SWAT model parameters and the output by adopting SVR, and screening out key parameters with remarkable influence; s3, on the basis of the water quality monitoring data, constructing a water quality time sequence prediction model by adopting an LSTM model; and S4, constructing a self-learning mechanism based on the LSTM and the SVR model, and retraining the model by adjusting a learning time window and regularly utilizing latest data. The method is excellent in the aspects of model efficiency, prediction precision and application adaptability, and has remarkable engineering application value and popularization potential.

The invention discloses an SSP scene reservoir group flood control scheduling method considering hydrological forecast uncertainty, and the method comprises the steps: obtaining SSP climate scene meteorological data to drive an SWAT model to generate daily runoff, and deducing and designing a flood hydrograph; a flow error sequence is calculated based on historical simulation and measured data, optimal distribution is optimized through fitting of a multi-probability distribution model, and a 95% confidence interval is generated through Monte Carlo simulation to correct design flood; and constructing an optimization model with the goal of minimizing the highest water level of flood regulation, bringing in various constraints, solving an optimal scheduling process by adopting a discrete differential dynamic programming algorithm, and finally refining an adaptive scheduling rule. The method solves the problems that prediction uncertainty is not considered and solving efficiency is low in the prior art, flood control safety and dispatching adaptability of the reservoir group are improved, and the method is suitable for flood control dispatching of the reservoir group under climate change.

The invention discloses a drainage basin water quality pollution source flux identification method based on an SWAT model, computer equipment and a storage medium, and the method comprises the steps: obtaining topographic data, land utilization data, soil type data and meteorological data, and enabling the topographic data, the land utilization data, the soil type data and the meteorological data to form a drainage basin spatial database; based on the preprocessed digital elevation model, determining a water flow direction and setting a reasonable water collection area threshold value, and dividing sub-basins to construct a basin calculation unit network topology structure; obtaining pollution source data, and constructing an SWAT-CUP model according to the pollution source data; performing system calibration and verification on the watershed hydrological process by adopting an SUFI-2 algorithm in the SWAT-CUP model; carrying out system calibration and verification on the total nitrogen load of the drainage basin by adopting an SWAT-CUP model; and based on the calibrated SWAT hydrological model, the input condition of the pollution source is changed. The method can cover the whole watershed and quantify the contribution rate of different pollution sources; cost is high, operation is simple, and long-term change can be dynamically simulated.

The invention relates to the technical field of tailing pond model construction, in particular to an earth surface groundwater combined tailing pond model construction method and system. According to the method, the SWAT-MODFLOW coupling model is established, the model combines the advantages of the SWAT model in surface hydrological process simulation and the accuracy of the MODFLOW model in underground water flowing process simulation, the characteristics of the hydrodynamic field and the seepage field of the tailings pond can be more comprehensively described, and the seepage flow of the dam body can be accurately predicted. And meanwhile, a dam body slope stability CPSO-BP prediction model is established by utilizing a coordinated particle swarm optimization algorithm and a BP network, and the model can describe a complex nonlinear relationship between the slope stability and influence factors thereof, so that the prediction precision is improved. The influence of the surface water and the underground water on the tailings pond is comprehensively considered, and compared with a traditional model only considering the surface water, the running state and the instability mechanism of the tailings pond can be simulated more truly.

The invention discloses an ecological flow guarantee method based on runoff prediction and supply. Meteorological, hydrological and other data of the Tongkeng river basin are collected, an SWAT model is constructed and verified to simulate runoff, and extended rainfall is analyzed based on an R / S method and a wavelet method. And constructing a surface water and underground water coupling model, and predicting underground water volume, water level and water quality. And ecological flow safeguard measures are formulated after water resource supply and demand balance analysis, wherein the ecological flow safeguard measures comprise optimal allocation of water resources, enhanced management, ecological water supplementation, water environment treatment, ecological restoration project development and the like. The method has the advantages that measure implementation subjects and progresses are defined, monitoring system evaluation effects are established, adjustment and optimization are performed according to results, ecological flow is guaranteed, sustainable utilization of drainage basin water resources and ecological environment protection are realized, remarkable social, economic and environmental benefits are achieved, and a scientific and effective method is provided for water resource management.

The invention provides a method for predicting medium and long-term inflow runoff of a large lake in a river network area. The method comprises the following steps: collecting data and dividing branches into low, medium and high influence branches; a SWAT model and a physically constrained LSTM model are adopted to predict low and medium influence branch runoffs respectively; screening runoff prediction results of low and medium influence branches related to the high influence branches as prediction factors, and constructing a DBN-BP model to realize Euclidean prediction of the runoff of the high influence branches; and constructing runoff probability distribution by adopting a conditional normalization flow model, and realizing error correction through a nonparametric residual model. According to the invention, for river network regional differentiation modeling, physical constraint and data driving methods are combined, the problems of human activity interference and uncontrolled interval runoff prediction are solved, the prediction precision is effectively improved, and decision support is provided for water resource management.

The invention relates to the technical field of water and soil loss dynamic monitoring and early warning methods in water conservancy projects, and discloses a water and soil loss dynamic monitoring and early warning method, which comprises the following steps: synchronously acquiring multi-dimensional data of terrain, rainfall, soil humidity and vegetation coverage through unmanned aerial vehicle remote sensing and a ground sensor network; registering and correcting the multi-source heterogeneous data based on a space-time fusion algorithm; the improved SWAT model is combined with a machine learning algorithm to dynamically simulate a water and soil loss process; and triggering graded early warning according to a preset risk threshold and generating prevention and control suggestions. The system comprises a data acquisition module, a data fusion processing module, a water and soil loss simulation prediction module and an early warning response module. According to the scheme, the real-time performance, the spatial resolution and the early warning accuracy of water and soil loss monitoring are remarkably improved, and efficient decision support is provided for ecological protection of water conservancy projects.

The invention provides a large-scale agricultural irrigation area grid scale underground water level spatial and temporal distribution efficient prediction method, and belongs to the technical field of underground water prediction.The method comprises the steps that firstly, an SWAT model of a drainage basin where an irrigation area is located and an MODFLOW model covering the irrigation area are constructed respectively, the SWAT model and the MODFLOW model are coupled to obtain an SWAT-MODFLOW model, and parameter correction is conducted; secondly, an irrigation area underground water level data set is constructed through scene combination and model simulation and used for training and screening the LSTM model with the highest precision, the optimal control range of each LSTM model is determined through space compression and reconstruction, and a multi-LSTM model is obtained; and finally, carrying out parallel calculation on the multiple LSTM models to predict the spatial and temporal distribution of the grid scale underground water level of the irrigation area. According to the method, the optimal control range of each LSTM model is determined through space compression-reconstruction, the prediction of the underground water level spatial and temporal distribution of the grid scale of the large-scale agricultural irrigation area is finally realized through multi-LSTM model parallel calculation, and the prediction efficiency is improved on the premise of ensuring the prediction precision of the underground water level of the irrigation area.

The invention relates to the technical field of environmental science, and discloses an SWAT-based BMP optimal configuration method adaptive to climate change, and the method comprises the steps: collecting pollution data of a target agricultural region, and constructing an NPS simulation model of the target agricultural region based on an SWAT model; collecting basic data of a target agricultural area, determining implementation data of a plurality of preset optimal management measures, and constructing an optimal management measure database; based on the optimal management measure database, performing parameter optimization on the NPS simulation model to form an NPS simulation optimization model; simulating climate change in future time according to the climate observation data and the historical meteorological data, and constructing a meteorological database; and based on the meteorological database, operating the NPS simulation optimization model, and performing optimal management measure optimization configuration analysis through a genetic algorithm to obtain an optimal management measure optimization result. According to the method, resources can be reasonably configured, and the flexibility, adaptability and actual effect of optimal management measures are improved.

The invention discloses a basin hydrological simulation method and system embedded with a canopy interception mechanism, and the method comprises the steps: embedding a revised Gash canopy interception model module in an SWAT model, dividing a rainfall event into a wetting stage, a saturation stage and a drying stage according to the parameters of rainfall, a leaf area index and a tree height, dynamically calculating the canopy interception amount and the interception evaporation amount, and carrying out the calculation of the canopy interception amount and the interception evaporation amount. The effective rainfall capacity is used for replacing original rainfall input in the SWAT model, the adjusting effect of the forest canopy on the rainfall process is reflected more truly, and the basin hydrological simulation precision is improved. According to the method, the response and simulation capability of the SWAT model to the hydrological process of the underlying surface of the forest is effectively improved, the dynamic response to effective rainfall and evaporation terms is enhanced, the method is suitable for watershed hydrological process modeling of various forest types, and the reliability of flood forecasting and water conservation evaluation is improved.

The invention discloses a watershed ditch and pond-road low-cost ecological sand control method based on hydrological connectivity, which belongs to the field of ecological environment protection and mainly comprises the steps of data acquisition, DEM data correction, confluence network construction, hydrological structure connectivity identification, SWAT model construction, key source area identification, scene analysis and cost benefit analysis. Under the influence of a ditch pond-road system, the key source area identification technology based on hydrological structure connectivity and the drainage basin key source area ditch pond road low-cost water regulation and soil conservation method achieve the dual purposes of reducing the water and soil regulation and control cost and continuously improving the ecological benefit by identifying the key source area.

The invention discloses a hydrological analysis method and system based on SWAT model runoff reconstruction and a storage medium. The hydrological analysis method comprises the following steps: 1, obtaining research basin runoff data, meteorological data, digital elevation model data, land utilization data and soil data; 2, based on the data in the step 1, analyzing the evolution trend and potential mutation points in the runoff time sequence period by adopting a mutability test method; 3, dividing a natural period and a variation period based on the mutation points, and obtaining a natural runoff sequence by using an SWAT model; and 4, based on a SWAT model simulation result, extracting multiple hydrological indexes representing hydrological situation changes by adopting an IHA method, extracting ecologically-related hydrological indexes ERHIs from the hydrological indexes by adopting a PCA method, and calculating contribution rates of climate changes and human activities to the ecologically-related hydrological indexes by adopting a quantitative attribution method. The runoff driving mechanism of climate factors and human activities is disclosed, and the physical interpretability and decision support capability of the SWAT model are enhanced.

The invention discloses an artificial wetland space optimization arrangement method and system, a non-point source pollution model is constructed based on a hydrological response unit, target meteorological data, target pollution data and an SWAT model, and the simulation precision of the non-point source pollution process can be improved. Climate scene data is constructed through future climate data, a preset path rule and a simple quantile algorithm, and the influence of climate change on pollution load can be reflected. Furthermore, a wetland optimization model is constructed based on a non-point source pollution model and a target function, the scientificity of an optimization strategy is enhanced, candidate wetland areas are identified by a spatial analysis algorithm based on a complex terrain index CTI, and the maximum potential area of each candidate wetland area is counted, so that efficient identification of potential arrangement areas can be realized. The constructed wetland optimization regulation and control scheme is generated based on the preset genetic algorithm, and the constructed wetland space optimization arrangement is performed according to the scheme, so that the optimization and practicability of the constructed wetland layout can be realized.

The invention discloses a runoff prediction method based on a physical information gating network. The runoff prediction method comprises the following steps: S1, constructing an SWAT model; s2, constructing a data set; s3, global feature extraction; s4, extracting local features; s5, gate control weight calculation is carried out; and S6, performing adaptive fusion and prediction. According to the runoff prediction method, whether the model focuses on surface runoff or underground runoff can be dynamically determined according to the current moment state and the historical time sequence memory of the drainage basin, and the adaptability of the model to different hydrological scenes is remarkably improved. A double-branch architecture combining Autoformer and TCN is adopted, so that the problem that a single model is difficult to consider both periodic trend and mutation characteristics is effectively solved. And an introduced feature attention and time attention mechanism is combined with the change of a physical gating value, so that multi-level physical interpretability is provided for a prediction result of the deep learning model. The end-to-end joint optimization strategy avoids the problem of separation of traditional physical model parameter calibration and deep learning model training, the convergence speed is high, and the simulation precision is high.

The invention discloses an SWAT model correction and application method based on remote sensing data, and relates to the technical field of hydrology and water resource model optimization. Comprising the following steps: S1, acquiring various remote sensing data in a hydrological process; s2, preprocessing various remote sensing data; s3, a corresponding SWAT model is established, and parameter correction is realized through the established SWAT model; and S4, applying various remote sensing data to the established SWAT model, and analyzing and monitoring landform and river channel information in the hydrological process. According to the method, parameter correction is realized through the established SWAT model, various remote sensing data are applied to the established SWAT model, and landform and river channel information in the hydrological process is analyzed and monitored.

The invention discloses a water conservancy regulation and control information fused plain river network area pollutant flux simulation correction method, which realizes deep fusion of a physical mechanism and deep learning, utilizes a physical reference of an SWAT model as a basic constraint, utilizes deep learning to excavate a nonlinear residual error, and cooperates with ReLU physical constraint and a graph topology transmission mechanism to realize simulation correction of the flux of pollutants in a plain river network area. The method has the interpretability of a physical model and the high precision of a data-driven model, and is particularly suitable for monitoring a complex watershed with scarce data. High-precision inversion and correction are carried out on the day-by-day runoff, the total nitrogen flux and the total phosphorus flux of the whole watershed through a small amount of key node monitoring data, and powerful support is provided for space-time analysis of plain river network area non-point source pollution and water environment fine management. The problem of false runoff production in a plain river network area is solved, cascade correction logic with runoff as quality is established, and the problems of runoff simulation distortion and large pollution flux estimation error of a traditional hydrological model are avoided.

The invention discloses a pixel scale water environment safety assessment method based on a non-point source water body pollution risk, and belongs to the technical field of environmental protection. According to the method, through an agricultural non-point source pollution direct evaluation method, the pollution risk of terrestrial agricultural non-point source output to surface water is accurately quantified, a DPeRS model is introduced, quantitative evaluation of the remote sensing pixel scale agricultural non-point source river entering amount is achieved, and the pixel scale agricultural non-point source river entering concentration is quantified as a non-point source pollution risk evaluation index in combination with an SWAT model. With surface water environment quality improvement as a target, an agricultural non-point source pollution in-water risk grade threshold value is determined based on a nitrogen and phosphorus concentration limit value in a surface water environment quality standard and a national secondary pollution general survey agricultural non-point source proportion, and a pixel scale agricultural non-point source in-water pollution risk prominent area is rapidly screened. High-precision target positioning of remote sensing pixel scale agricultural non-point source pollution is realized, and source-targeted agricultural non-point source pollution precise prevention and control and agricultural non-point source pollution treatment effect evaluation are facilitated.

The application discloses a method for reducing nitrogen and phosphorus loss of slope farmland, which comprises the following steps: step 1, determining a slope farmland catchment area and estimating nitrogen and phosphorus loss load of a SWAT model; step 2, designing a drainage ditch, a rain and sewage diversion device and a diversion pipe according to the estimation result of the model; and step 3, designing a vegetation retention pool according to the runoff of the slope farmland; the vegetation retention pool effectively reduces the loss of nitrogen and phosphorus of the slope farmland through the absorption of plants, the filtration of sediments and the retention effect of water flow. The method can realize diversion treatment of small rain and high-concentration nitrogen and phosphorus runoff in the early stage of heavy rain, and the nitrogen reduction rate of the slope farmland drainage reaches 70-80%, and the phosphorus reduction rate reaches 80-90%.

The invention provides a comprehensive monitoring method for nutrient loss of a freeze-thaw basin, which comprises the following four steps of: 1, analyzing hydrometeorological characteristics of the basin based on remote sensing data, and determining a key period of nutrient loss; 2, dividing a sub-basin and a river network structure by using an SWAT model, and scientifically arranging representative sampling points; thirdly, surface water collection is carried out in combination with the determined sampling scheme, and laboratory analysis is carried out on key indexes such as total nitrogen and total phosphorus; and 4, identifying a key period and a source region of pollutant loss based on the river network convergence relationship and the monitoring data. The method integrates three technologies of remote sensing identification, model simulation and field observation, has the advantages of low cost, high efficiency, high representativeness and strong adaptability, and is especially suitable for data scarce areas such as medium and high latitude freeze thawing areas. The method has a good application prospect in the fields of water environment monitoring and pollution identification.

The present invention discloses a method and system for emergency treatment and analysis of snowmelt non-point source pollution, relating to the field of data processing technology. The method comprises: collecting basic regional information of a target area; dividing the area based on the basic regional information; parsing the basic regional information to generate land use data, soil type data, soil property data, and historical meteorological data; using a SWAT model to obtain non-point source pollution simulation results; collecting snow cover data of the target area; and predicting snowmelt floods based on the snow cover data and the basic regional information using a WRF-driven two-layer grid overlay model to obtain prediction results; and conducting emergency management based on the prediction results and the non-point source pollution simulation results. The present invention solves the technical problems of low intelligence and low management accuracy in the prior art for emergency decision-making regarding snowmelt non-point source pollution, thereby achieving the technical effect of improving the efficiency and management quality of emergency treatment of snowmelt non-point source pollution.

The application provides a large-scale agricultural irrigation area grid scale groundwater level space-time distribution efficient prediction method, belongs to the technical field of groundwater prediction, and steps include: firstly, a SWAT model of a basin where an irrigation area is located and a MODFLOW model covering the irrigation area are respectively constructed, the two are coupled to obtain a SWAT-MODFLOW model, and parameter correction is carried out; secondly, an irrigation area groundwater level data set is constructed through scenario combination and model simulation, is used for training and screening the highest-precision LSTM model, and the optimal control range of each LSTM model is determined through space compression and reconstruction, and multiple LSTM models are obtained; finally, the multiple LSTM models are parallelly calculated to predict the space-time distribution of the grid scale groundwater level of the irrigation area. The optimal control range of each LSTM model is determined through space compression-reconstruction, finally, the prediction of the space-time distribution of the grid scale groundwater level of the large-scale agricultural irrigation area is realized through the parallel calculation of the multiple LSTM models, and the prediction efficiency is improved on the premise of ensuring the prediction accuracy of the irrigation area groundwater level.

The invention relates to the technical field of pollution tracing, and discloses a drainage basin pollution tracing method, device and equipment and a storage medium. The method comprises the following steps: inputting obtained target pollution source emission data and basic data into a target SWAT model, then operating the target SWAT model, and simulating and outputting sub-basin pollution load of each sub-basin in a target basin; determining contribution proportions of different pollution source categories to the total pollution load of the target drainage basin based on an isotopic tracing method; determining a load vector of each pollution source category of the target drainage basin based on the contribution proportion and the drainage basin pollution simulation total load; constructing a sub-basin load distribution matrix; determining a unit load matrix based on the sub-basin load distribution matrix; and analyzing the unit load matrix, and tracing to obtain the dominant pollution category of each sub-basin. According to the scheme, by constructing the unit load matrix, the main source sub-drainage basins of various pollution sources are defined, the space traceability of the pollution sources is realized, and the traceability accuracy is improved.

This invention provides a method and system for identifying agricultural non-point source pollution risks in lake basins. The method includes: acquiring meteorological and crop data and inputting them into a constructed basin SWAT model; running the model monthly to obtain the runoff of each river and the concentration and output of non-point source pollutants in each river; constructing a river-lake estuary MIKE21 model to simulate the changes in lake non-point source pollutants for the current month and predict the next month, obtaining lake water quality indicators, and determining whether the lake currently meets the requirements of the lake's water function zone. If so, the risk area under the minimum reduction inflow load scenario is determined, and the basin SWAT model is called again for basin-wide agricultural non-point source pollution control. If not, the input load of each inflow river sub-basin is reduced one by one until the water quality meets the standards. Based on the basin's agricultural characteristics and meteorological and hydrological forecasts, this invention identifies and predicts key areas and time periods for agricultural non-point source pollution and proposes small-basin and segmented area management, demonstrating significant environmental and ecological benefits as well as socio-economic benefits.

The invention belongs to the technical field of groundwater pollution control and treatment, and discloses a method for effectively reducing migration of groundwater pollutants in a karst basin, and the method comprises the following specific use steps: S1, data preparation and preprocessing; s2, constructing and improving an SWAT model; s3, performing SWAT model simulation and result analysis; s4, an MODFLOW model is constructed; s5, carrying out MODFLOW model simulation and reaction migration process description; and S6, coupling the SWAT model and the MODFLOW model. According to the method, an SWAT model is improved, distribution of hydrological response units is optimized, a surface karst zone module is added, description of the model on the underground migration process of nitrogen and phosphorus along with karst water is completed, a verification code is obtained by constructing a needed space and attribute database and utilizing actually measured data verification and model adaptability, and meanwhile, the model can be used for analyzing the nitrogen and phosphorus content of the karst water. In the gurge drainage basin, the simulation precision of the model meets the requirement, and the certainty coefficient R2 and the Nash efficiency coefficient are both larger than 0.66.

The invention discloses a drainage basin base flow change attribution analysis method and system, and relates to the field of drainage basin water resource management, and the method comprises the steps: obtaining multi-source basic data of a to-be-researched region; based on daily flow data in the multi-source basic data, a heuristic segmentation algorithm is adopted to determine an annual runoff abrupt change point set; dividing the data according to the first mutation point to obtain reference period data and influence period data; based on the reference period multi-source data and an improved SWAT model of an embedded base flow extraction module, obtaining a base flow simulation sequence; and obtaining the contribution rate of the climate change and the human activity to the basic flow change of the drainage basin through a scene reconstruction and difference decomposition calculation method in combination with the data of the reference period and the influence period. According to the method, by accurately depicting the basin base flow process and the base flow-runoff synergistic effect, multi-time-scale refined quantitative analysis of the contribution rate of human activities and sub-items thereof and climate changes to the base flow change is realized, and the scientificity of basin water resource management decision is remarkably improved.

The invention relates to the technical field of water and soil loss dynamic monitoring and early warning methods in water conservancy projects, and discloses a water and soil loss dynamic monitoring and early warning method, which comprises the following steps: synchronously acquiring multi-dimensional data of terrain, rainfall, soil humidity and vegetation coverage through unmanned aerial vehicle remote sensing and a ground sensor network; registering and correcting the multi-source heterogeneous data based on a space-time fusion algorithm; the improved SWAT model is combined with a machine learning algorithm to dynamically simulate a water and soil loss process; and triggering graded early warning according to a preset risk threshold and generating prevention and control suggestions. The system comprises a data acquisition module, a data fusion processing module, a water and soil loss simulation prediction module and an early warning response module. According to the scheme, the real-time performance, the spatial resolution and the early warning accuracy of water and soil loss monitoring are remarkably improved, and efficient decision support is provided for ecological protection of water conservancy projects.

The invention belongs to the related technical field of hydrology and water resource engineering, and discloses a drainage basin runoff prediction method and system fusing a graph attention network and physical constraints. The method comprises the steps of inputting preprocessed hydrological and geographic parameters of a to-be-processed watershed into an SWAT model, and obtaining attributes of each sub-watershed in the to-be-processed watershed, a topological relation of the watershed and simulated runoff volume of each sub-watershed in a specified time sequence; encoding the output of the SWAT model by using an encoder to obtain static features and dynamic features of each sub-basin; taking the static characteristics and the dynamic characteristics as input of a multi-head attention network to obtain a runoff corrected value of each sub-basin; and correcting the simulated runoff volume by using the runoff volume correction value to obtain a final predicted value of the runoff volume in a specified time sequence, so as to predict the runoff volume of the drainage basin. According to the invention, the problems of high parameter sensitivity, lack of physical constraints and simple attention mechanism design in hydrological prediction are solved.

The application relates to the field of karst basin runoff simulation calculation, and particularly discloses a karst SWAT model runoff simulation method coupled with an early-stage precipitation index, which comprises the following steps: step one, establishing a SWAT model with the characteristics of a karst area, generating a land use map with the characteristics of the karst area and hydrological response units (HRUs) with the distribution characteristics of the karst and non-karst; step two, modifying the source code of the SWAT model to construct a K-SWAT model with the function of karst runoff generation and concentration; step three, embedding a normalized early-stage precipitation index (MIWAPI) in the K-SWAT model to construct a K-W-D-SWAT model with the function of identifying the effect of dry and wet events on the karst runoff generation and concentration process; and step four, optimizing the parameters of the SWAT model, the K-SWAT model and the K-W-D-SWAT model by using the SUFI-2 algorithm in the SWAT-CUP software, and evaluating the runoff simulation level of each model in the karst area. The karst SWAT model runoff simulation method coupled with the early-stage precipitation index is suitable for the simulation of the runoff process in the karst area with complex climate change, and can improve the simulation accuracy and rationality of the hydrological process in the karst area.