101 results about "Runoff volume" patented technology

The invention discloses a river organic carbon flux inversion method based on multi-source data and related equipment. The method comprises the following steps: acquiring multi-source data of a river basin range in a target research area; obtaining a normalized difference index based on remote sensing image conversion; river parameters of the target research area are obtained based on the normalized difference water body indexes and multi-source data processing, and river runoff is obtained through conversion of a hydraulic geometrical relationship; on the basis of the remote sensing image, the normalized difference chlorophyll index and the normalized difference turbidity index, the particle organic carbon concentration is obtained through inversion by means of a first inversion model; based on the population number and the rainfall capacity, a second inversion model is used for inversion to obtain the concentration of dissolved organic carbon; and converting the river runoff, the particle organic carbon concentration and the dissolved organic carbon concentration to obtain the river organic carbon flux. The method can significantly improve the accuracy and reliability of monitoring on a high-dynamic river, accurately realizes the inversion of the organic carbon flux of the river, and can be widely applied to the technical field of data processing.

The invention discloses a drainage basin nitrogen and phosphorus pollution prediction method based on rainfall runoff migration, and relates to the technical field of water quality pollution prediction, and the method comprises the steps: accurately obtaining land utilization and water body distribution through a remote sensing technology, carrying out the space division of a pollution source through combining a nitrogen and phosphorus load coefficient, and constructing a rainfall runoff model. The runoff volume and time and space distribution under different rainfall events are reflected, and the spatial dynamic change of pollutants is captured through a hydrodynamic model, in combination with a convection diffusion equation and by introducing a conversion model of nitrogen and phosphorus in various forms. By integrating remote sensing data, rainfall runoff simulation, hydrodynamic force and nitrogen and phosphorus form transformation, the space-time migration and transformation process of nitrogen and phosphorus pollution in a drainage basin is described, so that the coupling relation between pollution source distribution and runoff power is revealed, the conveying and diffusion rule of pollutants in a river network is dynamically reflected, and the drainage effect is improved. And refining to transformation evolution of different nitrogen and phosphorus forms, and generating intuitive concentration distribution and thermodynamic diagrams through spatial interpolation.

The invention relates to the technical field of hydrological prediction, in particular to a drainage basin distributed runoff prediction method and system based on a graph neural network, and the method comprises the following steps: obtaining a drainage basin multi-source runoff data set to construct a multi-relation dynamic graph structure, and extracting node feature vectors; based on the node feature vector, obtaining a watershed evolution trend forward feature by establishing a watershed diffusion fitting architecture; constructing a distributed runoff probability prediction model, and taking the watershed trend forward features as model input to obtain runoff initial condition probability distribution of each sub-watershed in multiple periods in the future; establishing a mixed loss function, and performing physical constraint optimization on the runoff initial condition probability distribution to obtain distributed runoff optimization probability distribution; and performing uncertainty quantification on the distributed runoff optimization probability distribution to obtain a drainage basin distributed runoff prediction result. According to the method, the hydrological process simulation capability of the complex watershed is improved, and accurate prediction of the distributed runoff volume is realized.

The invention relates to the technical field of hydrological simulation and flood forecasting, and particularly discloses a watershed runoff simulation method and device with a graph neural network fused with hydrological priori knowledge, and the method comprises the steps: carrying out the grid division of a target watershed, and building a watershed graph structure in combination with the hydrological priori knowledge; introducing a trainable weight into the graph neural network model, and performing time delay weighting based on a spatial distance to obtain a fusion network model; the meteorological driving data, the rainfall data and the drainage basin attributes serve as input, the convergence process is simulated through the fusion network model, and the predicted downstream runoff volume is obtained; and comparing the predicted downstream runoff volume with the actually measured downstream runoff volume, and optimizing all learnable parameters of the fusion network model. According to the method, the defects that a traditional physical model is large in calculation amount and long in consumed time are overcome, rapid simulation of the runoff process is achieved, and the accuracy and reliability of a data driving model in a runoff simulation task are improved by fusing hydrological priori knowledge and introducing physical constraints.

The invention discloses a reservoir capacity prediction method considering underlying surface change conditions, and aims to solve the problems that the prediction value of an existing scheme is too large, deviation exists between the estimated reservoir capacity and the actual reservoir capacity, the functions of flood control, water supply, irrigation and the like of a reservoir are affected, and the economic benefits of the reservoir are directly affected. The method comprises the following steps: acquiring actually measured water and sediment series data of a reservoir in historical periods, analyzing the reservoir water and sediment change trend, determining reservoir runoff and reservoir sediment abrupt change points in combination with underlying surface factor changes, dividing different stages of reservoir sediment change, and calculating soil erosion amounts of drainage basins above the reservoir in different stages. And converting the reservoir sand amount in the early stage into the reservoir sand amount under the current underlying surface condition by using the conversion coefficient, estimating the reservoir sand amount in the planning level year or the design level year, and analyzing the reservoir capacity curve in the future. The reservoir capacity prediction method achieves the effects of more accurately predicting the reservoir capacity, exerting the reservoir function, improving the reservoir benefit and avoiding the waste of engineering investment.

The invention provides a flood model parameter calibration method and system based on isotopic tracing, and the method comprises the steps: determining a parameter set and prior distribution which need to be analyzed in a flood model, and the parameter set comprises parameters related to runoff volume and parameters related to isotope transmission in the flood model; randomly generating a group of parameter samples according to prior distribution of parameters; evaluating the sensitivity of each parameter in each group of parameter samples by analyzing the relationship between the parameter samples and the corresponding objective function values; screening out sensitive parameters according to a sensitivity evaluation result; according to the flood model parameter calibration method based on isotopic tracing provided by the invention, the problem of different parameters and the same effect of the flood model is solved, namely, the different parameter and the same effect of the model are greatly reduced (namely, the parameter range when the model evaluation index reaches a specific threshold value or above), and meanwhile, the simulation effect of the flood model in the verification period is greatly improved.

The present application belongs to the technical field of geological disaster prevention and mitigation, and relates to a debris flow power early warning method, specifically comprising: obtaining a startable source amount of a debris flow basin and a possible runoff under a forecast rainfall condition, calculating a water-rock mixture density ρ d and a water-rock mixture density ρ h under an extreme hourly rainfall condition of a target channel under the forecast rainfall condition according to the startable source amount and the possible runoff respectively; obtaining a final critical density ρ of a water-rock mixture when a debris flow is started in a main sliding direction, and determining a first early warning level T1 under the forecast rainfall condition in combination with a daily scale water-rock mixture density early warning level determination rule, and determining a second early warning level T2 under the extreme hourly rainfall condition in combination with an hourly scale water-rock mixture density early warning level determination rule; when the first early warning level T1 and the second early warning level T2 are inconsistent, selecting a higher early warning level as a published early warning level T of a target debris flow basin, which can effectively prevent group debris flow disasters.

The invention belongs to the technical field of resource optimization scheduling, and provides a drainage basin water volume accurate scheduling method and system based on precipitation real-time monitoring, and the method comprises the following steps: collecting hydrological data, early-stage influence rainfall and soil water content of each hydropower station in a target drainage basin; the earlier-stage influence rainfall is corrected through the collected hydrological data, and the corrected earlier-stage influence rainfall is obtained; in combination with hydrological data and the corrected early-stage influence rainfall, carrying out runoff forecasting in a short-term period of the drainage basin and middle-and-long-term end-of-period water level forecasting in the drainage basin, and obtaining end-of-period water level information; determining a flood prevention early warning level and a drought resistance early warning level according to the end-of-period water level information; and a scheduling decision is made according to the flood prevention early warning level and the drought resistance early warning level. According to the scheduling method provided by the embodiment of the invention, accurate scheduling of the water volume of the drainage basin can be realized.

The application provides a soil salinization response simulation method and device, the method comprises the following steps: obtaining the soil type, terrain grid data and meteorological data of a target area; inputting the terrain grid data into a first processing model to obtain a sub-basin division result; inputting the meteorological data and soil type into the first processing model according to the sub-basin division result to obtain sub-basin daily precipitation data and sub-basin evapotranspiration; inputting the sub-basin daily precipitation data and sub-basin evapotranspiration into a second processing model to obtain soil salt content and crop yield in combination with sub-basin irrigation; obtaining soil salinization prevention and control measures according to the soil salt content and crop yield; wherein the first processing model is obtained by calibrating a preset processing module according to the difference between the monthly runoff of a basin and the measured calibration data. The application can improve the accuracy, applicability and operability of soil salinization prevention and control.

The present disclosure relates to the technical field of watershed water and sediment regulation, and provides a method and device for predicting a watershed sediment transport ratio. The method comprises: obtaining a slope sediment yield parameter and a runoff parameter of a target watershed; calculating a target slope sediment yield according to the slope sediment yield parameter; calculating a target runoff according to the runoff parameter; and introducing the target slope sediment yield and the target runoff into a preset sediment transport ratio prediction model to predict a target sediment transport ratio. The present disclosure can greatly improve the application range and prediction accuracy of the prediction of the watershed sediment transport ratio.

The application discloses a snowmelt runoff prediction method and system based on direct measurement of snow water equivalent, and belongs to the technical field of hydrological monitoring. The system comprises a cooperative observation network composed of an acoustic source array emitting coded acoustic waves and a sensing optical fiber laid on the ground surface; a distributed optical fiber acoustic wave sensing demodulator for receiving a snow layer modulated vibration signal; a fusion inversion calculation unit adopting a physical information neural network algorithm, embedding an acoustic wave fluctuation equation as a constraint into network training, and directly inverting a continuous spatial distribution of the snow water equivalent from the signal; and a runoff prediction module for predicting runoff by using the distributed data. The application breaks through the limitations of traditional point measurement and satellite remote sensing, realizes direct, high-resolution and high-precision surface measurement of the snow water equivalent, fundamentally improves the input data quality of the snowmelt runoff prediction model, and has the characteristics of high precision, strong reliability and cost-effectiveness.

The application relates to the technical field of hydrological analysis and calculation, in particular to a water source area adjustable water volume prediction method and system. The method comprises the following steps: obtaining historical measured runoff data of a water source area downstream river section and calculating corresponding natural runoff; dividing a prediction period based on the natural runoff, determining a prediction object and a target prediction factor; determining a historical year prediction weight and screening a reference year through calculating a current year and a historical year difference measure; constructing a natural runoff prediction model based on the reference year to obtain a prediction result; obtaining a real-time reservoir water level and calculating a dynamic reservoir capacity; finally, combining the natural runoff prediction result and the dynamic reservoir capacity, calculating an adjustable water volume in each period and outputting a result. The application realizes high-precision comprehensive prediction of the water source area adjustable water volume, thereby significantly improving the scientificity and reliability of water resource regulation.

The application provides a well-canal combined scheduling method coupled with long short memory deep learning, first, a real-time hydrological forecast model is constructed, historical N-day meteorological, hydrological and different water source irrigation amounts are taken as inputs through a long short memory deep learning method to forecast runoff in a short future period, the different water source irrigation amounts include well water and canal water irrigation amounts; then, a real-time water amount scheduling model is constructed, a minimum runoff is taken as a constraint to maximize irrigation area water supply benefit to solve optimal irrigation amounts of different water sources in a short period; finally, the real-time hydrological forecast model and the real-time water amount scheduling model are iterated and simulated until balance is reached, so that optimal well water and canal water irrigation amounts are obtained. The application can realize precise irrigation of an irrigation area, fully utilize water supply capacity of a canal, exchange water supply time for water supply space, fully play water storage engineering capacity of the irrigation area and maximize reduction of drought.

A method for estimating glacial meltwater runoff is disclosed, comprising the following steps: S1, obtaining glacier geographical distribution information and calculating driving data; S2, simulating glacier ablation; S3, calculating glacial meltwater runoff. A relatively accurate glacial meltwater runoff estimation model is established, solving the problem of significant uncertainty in existing meltwater runoff estimation methods, and is able to simulate the changing characteristics of each glacier.

The invention relates to the technical field of water and soil conservation monitoring, and discloses an intelligent monitoring method and system for the water and soil loss amount of a runoff plot, and the method comprises the steps: synchronously collecting the original measurement data of a multi-source sensor, and carrying out the timestamp calibration and data frame packaging; performing preprocessing and quality evaluation on the multi-source original data sequence; carrying out real-time calculation on the instantaneous runoff volume and the sand content; calculating an accumulated runoff volume and an accumulated soil loss volume; rainfall events are divided, and statistics is carried out in combination with calculation results; performing intelligent analysis and prediction on the water and soil loss process; physical consistency verification and optimization calibration are carried out; intelligent early warning judgment is carried out, and data display and report automatic generation are carried out through a visual platform; according to the method, a complete technical chain of multi-source sensor synchronous acquisition, edge intelligent preprocessing and cloud deep learning analysis is constructed, so that the water and soil loss amount of the runoff plot is upgraded from discrete manual sampling to continuous automatic sensing.

The application discloses a kind of rice irrigation area field dynamic hydrological process simulation method and system, method includes: the spatial correlation information base of rice field and ditch is constructed;With day as step simulation cycle, according to the dynamic adjustment effective field height of rice growth stage, based on field type and water level potential energy difference calculation lateral recharge or discharge;According to the exponential decay correction of plow pan saturated hydraulic conductivity of consecutive flooding days, calculate vertical infiltration amount;Overflow drainage, lateral discharge, vertical infiltration are updated into water balance equation field water depth;According to the topological network of confluence of field spatial position and elevation, consider the peak-shaving superposition of confluence time delay, obtain sub-basin export total runoff.It realizes the whole process fine simulation of field ridge water storage, field-ditch lateral interaction, plow pan nonlinear infiltration and distributed confluence, significantly improves the precision of rice irrigation area hydrological process simulation, provides scientific basis for irrigation water resources management.

The invention provides a farmland nitrogen loss monitoring method and system based on multi-source data, and the method comprises the following steps: obtaining multi-source time sequence data of a target region; performing inversion to obtain the soil volumetric water content of each field parcel in the farmland parcel vector data; simulating to obtain the surface runoff and leakage of each field parcel in the farmland parcel vector data; constructing a nitrogen loss estimation model, and inputting the soil volumetric water content, the surface runoff and the leakage of each field in the farmland plot vector data into the nitrogen loss estimation model to obtain a nitrogen loss load corresponding to each field in the farmland plot vector data; and performing nitrogen loss risk monitoring on the field parcels in the target area according to the nitrogen loss load. Therefore, by applying the technical scheme of the application, the monitoring efficiency and the monitoring accuracy of the nitrogen loss of the farmland plot can be effectively improved.

The invention discloses a drainage basin scale snow melting runoff estimation method and system, electronic equipment and a storage medium, and the method comprises the steps: collecting reanalysis meteorological data, remote sensing accumulated snow monitoring data and digital elevation model data in a target drainage basin range, and carrying out the preprocessing to generate a drainage basin scale surface average data sequence; dividing the target drainage basin range into a plurality of calculation units; determining a critical temperature threshold value of the whole drainage basin based on a response relationship between the drainage basin average accumulated snow coverage rate and the drainage basin average temperature; on the basis of comparison of the basin average temperature and the critical temperature threshold value, the basin overall snow melting period is recognized by combining basin average accumulated snow coverage rate changes; an improved degree-day factor method is executed in parallel on each calculation unit, and the snow melting rate is calculated; and confluence calculation is carried out on runoff generated by the snow melting rate of each calculation unit, and the snow melting runoff of the drainage basin outlet section is output. According to the method, dynamic and accurate estimation of the drainage basin snow melting runoff is realized, and the simulation precision and practicability are remarkably improved.

The application introduces a kind of aquatic organism protection threshold identification method based on four water coupling models, comprising: S1, constructing a watershed hydrological model, then inputting rainfall data, obtaining runoff output data;S2, constructing a river water power and water quality model, then inputting water quantity data, water level data and water quality concentration data, obtaining water quantity output data, flow velocity output data, water level output data, flow direction output data and water quality concentration output data;S3, determine the aquatic organism sampling point, and obtain the hydrology, hydrodynamic, water quality data of the corresponding aquatic organism sampling point through the watershed hydrological model and the river water power and water quality model;S4, extract the hydrology, hydrodynamic, water quality data matched with the aquatic organism sampling point from the model, construct a multi-factor coupling model based on random forest, and identify the aquatic organism protection threshold. Through the above design, the technical effect of solving the data matching in the aquatic organism protection threshold identification process is realized.

The invention discloses a rainfall flood regulation and storage efficiency evaluation method and a related device, and relates to the technical field of data analysis, and the method comprises the steps: calculating the rainfall depth based on the rainfall scene parameters of a target region; determining a first maximum water storage depth of a blue space in the target area, and calculating a first net water storage volume of the blue space based on the first maximum water storage depth and the rainfall depth; determining a second maximum water storage depth of a green space in the target area based on a runoff curve method so as to analyze a second net water storage volume of the green space; determining the runoff volume of the impervious surface in the target area, and calculating a rainfall flood regulation and storage efficiency index based on the runoff volume, the first purified water storage volume and the second purified water storage volume; and comparing the rainfall flood regulation and storage efficiency index with a preset threshold value, and determining classification information of the rainfall flood regulation and storage efficiency based on a comparison result. According to the method, the calculation cost of a complex hydrological model and excessive simplification of an empirical parameter method are avoided, and efficient and accurate urban rainfall flood regulation and storage efficiency evaluation and classification are realized.

The invention provides a large river-lake interval confluence water quantity determination method, a medium, equipment and a product, and relates to the technical field of data fusion and hydrological information. The method comprises the following steps: acquiring a remote sensing image, terrain and hydrological observation data of a target area; establishing a water balance equation, and expressing the interval confluence water quantity as a function of the variable quantity of the water storage quantity of the river and the lake, the inlet flow, the outlet flow and the total flow of main branches (controlled by a hydrological station); performing spatial discretization on the calculation unit and generating an elevation grid and an initial water level grid; extracting an actual water area based on the remote sensing image; by comparing the calculated water area with the actual water area, the water level grid is iteratively adjusted until the convergence condition is met, and therefore the accurate water storage capacity is obtained; and finally, substituting into a water balance equation to calculate the interval confluence water quantity and analyze the change characteristics. According to the method, multi-source data and iterative optimization are fused, the precision and reliability of water volume calculation are improved, and the method is suitable for water resource management and hydrological analysis of large rivers and lakes.

The invention discloses a watershed runoff evolution trend prediction method based on multi-source data fusion, and relates to the technical field of hydrology and water resource prediction, and the method comprises the steps: obtaining historical hydrometeorological data, dividing rainfall runoff events, calculating the soil water absorption amount in unit time through water balance, and fitting wet and dry water absorption paths to construct a hysteretic loop; clustering the hysteretic loop set to extract a representative hysteretic loop family; and matching the most similar hysteretic loop during prediction, inputting the data into a reinforcement learning model to output a path advancing position, determining the corresponding soil water absorption amount, and calculating the predicted runoff volume in combination with meteorological data. By quantifying the nonlinear hysteretic characteristic of the soil water absorption behavior, the problem that the runoff prediction precision is low due to the fact that a traditional model ignores the water absorption capacity difference in the dry and wet process is solved.

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 relates to the technical field of flood management, in particular to a flood risk assessment method and system based on multi-dimensional data integration analysis. The method comprises the following steps: acquiring a multi-dimensional data set including geographic information data and hydro meteorological data; determining a water flow direction by using the geographic information data, and identifying a catchment area based on the water flow direction; carrying out drainage basin runoff production calculation according to the hydro-meteorological data, and recording runoff volume; executing runoff collection simulation in the catchment area by utilizing runoff volume to form a runoff process line; calculating flood peak flow based on the runoff hydrograph; evaluating the river water level according to the flood peak flow, and calculating the river overflow according to the river water level; generating a submerging depth distribution map according to the river overflow quantity; and performing flood risk prediction based on the submerging depth distribution map, and generating a flood management risk level. Fine evaluation and dynamic prediction of the flood risk are realized based on the flood management technology, and the flood prediction accuracy and the flood control decision efficiency are improved.

The invention discloses a river runoff prediction method and system based on a hybrid tree convolution residual network, and belongs to the technical field of artificial intelligence and hydrology and water resource prediction. The method aims at solving the problems that in the prior art, non-linear features of runoff sequences are not sufficiently captured, and time sequence dependence and spatial heterogeneity are difficult to consider at the same time. According to the technical scheme, original data are split into an input sequence and a target sequence and normalized, after splitting is conducted according to feature channels, a depth separable convolution module is adopted for independent modeling, local time sequence features are extracted, and preliminary predictor results of all the channels are generated and merged; splicing each channel prediction result and a corresponding sequence into a tree model input feature, calculating a residual error between a preliminary prediction result and a target sequence, and generating a correction amount through tree model fitting; and finally, adding the initial prediction and the correction, and outputting a prediction value through inverse standardization. The method is suitable for watershed water resource management, flood early warning, hydraulic engineering scheduling and other scenes.

The application discloses a modularized distributed water and heat coupling hydrological model in cold regions, and comprises the following steps: S1, a basin is divided into a plurality of grids, and files required by a confluence process are prepared; S2, driving data and grid characteristic description data of each grid are prepared; S3, an energy balance module is used to calculate an energy balance process of soil and a vegetation surface; S4, an evapotranspiration module is used to calculate an evapotranspiration process; S5, an ice and snow melt water module is used, and a degree day factor method is used to calculate an ice and snow melting process; S6, a soil water and heat transmission and infiltration module is used to calculate soil water infiltration and internal water and heat transmission processes; S7, a runoff calculation and confluence module is used to calculate runoff; and S8, a model evaluation index is used to evaluate the simulation performance of the model. The application has a flexible modular structure, can automatically adjust a time step, improves the simulation capability of a runoff process in cold regions, and can provide scientific guidance for water resource management in a river source region.

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.

This invention relates to the field of eco-hydrology technology and discloses a method for simulating and predicting runoff in bedrock mountainous areas. The method includes: acquiring an input dataset for an eco-hydrological model in bedrock mountainous areas; converting it to gridded data to unify spatial resolution; sequentially calculating the effective water content of the weathering crust, the vegetation-available water content, climate-underlying surface non-physical parameters, atmospheric potential evapotranspiration, cover layer potential evapotranspiration, evaporation coefficient, and simulated runoff; constructing a fitness function by combining deterministic coefficients and Nash efficiency coefficients; selecting parameters to be calibrated, and using precipitation and potential evapotranspiration for the simulation period as inputs to output simulated runoff. This invention extends the simulation depth to the weathering crust in bedrock mountainous areas, calculates water content layer by layer, and calculates vegetation-available water content segment by segment, reflecting the ability of deep-rooted vegetation to utilize deep water. The fitness function is optimized, and the simulated runoff is output, providing a basis for ecological and water resource management in bedrock mountainous areas.

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 invention discloses a drainage basin water resource evolution feature recognition method based on an SSA-HELM algorithm. The method comprises the steps of obtaining hydrometeorological time series data such as drainage basin runoff, precipitation and air temperature; dividing the data into a training set and a test set, performing normalization processing, and inputting the data into a hierarchical extreme learning machine (HELM) model optimized by adopting a sparrow search algorithm (SSA) for training; evaluating the precision of the model by using the decision coefficient R and the root mean square error RMSE, and judging whether a preset requirement is met or not; and if the precision reaches the standard, predicting the future water resource amount by using the trained SSA-HELM model, carrying out reverse normalization processing on the prediction result, and identifying the long-term evolution trend and periodic characteristics of the water resource pattern in combination with a Mann-Kendall trend test method and self-correlation analysis. According to the method, the input weight and the hidden layer bias of the HELM model are optimized through the SSA algorithm, the convergence speed, the generalization performance and the prediction stability of the model are improved, local optimum is effectively avoided, water resource changes under complex climate changes can be accurately predicted, and a scientific basis is provided for water resource management and planning.