66 results about "Catchment runoff" patented technology

The invention discloses a drainage basin runoff forecasting method fusing an attention mechanism and a physical recurrent neural network, and the method comprises the steps: collecting historical hydrometeorological data and geographic attribute data of a research drainage basin, selecting a daily scale hydrological model, and determining the intermediate flux and state variables of the model; a PRNN circulation network structure with physical mechanism constraints is constructed, a time sequence attention mechanism is introduced into the structure, and relative contribution weights of different water storage state variables to runoff simulation are dynamically estimated; simulating and generating internal physical parameters of the model by taking hydro-meteorological driving factors and drainage basin geographic attributes as input, and carrying out collaborative optimization on the parameters in combination with an attention weight and a back propagation mechanism of a target function; and carrying out drainage basin runoff forecasting based on the trained coupled PRNN model. According to the method, the physical consistency and hydrological process characterization capability of the model are enhanced, and the interpretability and stability of the machine learning model in drainage basin runoff forecasting are effectively improved.

The invention relates to a multi-source data and physics combined driven drainage basin runoff uncertainty forecasting method. The method comprises the following steps: acquiring runoff sequence data and external forecasting factor data of a target drainage basin; constructing a runoff uncertainty generation model, performing probability diffusion on the runoff sequence data through a diffusion generator with constraints, and generating approximate runoff data; designing a joint loss function to update the weight and offset terms of the runoff uncertainty generation model; training loss convergence under joint guidance, and outputting a runoff prediction result. The method has the beneficial effects that a loss function based on Fourier transform and a physical theory guide item are combined, the randomness in a runoff physical system can be effectively quantified through introduction of physical constraints, more selectable predicted values can be provided while the physical consistency of prediction results can be ensured, and the prediction accuracy of the runoff physical system is improved. The inherent random uncertainty in the physical drought and flood process is effectively quantified, so that the physical consistency and prediction precision of the model are enhanced.

The invention discloses a flood peak enhanced physical base flow and residual error correction collaborative runoff prediction method, which belongs to the field of hydrological prediction, and comprises the following steps of: dividing a training set and a verification set according to a proportion, performing oversampling processing on flood peak samples, and constructing a time sequence window; a Xinanjiang model is discretized and expressed by adopting an ordinary differential equation, rainfall and potential evaporation data are input, and intermediate variables are obtained. A physical base flow and residual error correction dual-channel module is constructed, and physical base flow and residual error correction is calculated through two full-connection networks. And calculating a final runoff predicted value by adopting a residual connection structure, taking basic NSE loss as a core, superposing a flood peak sample error weighted item, strengthening flood peak fitting precision, and updating physical parameters and neural network weight through a back propagation algorithm. And verifying the model, and respectively calculating prediction indexes of the training set and the verification set. According to the method, fusion of a traditional hydrological model and a deep learning method is realized, the physical interpretation of the model is enhanced, and the basin runoff prediction precision is improved.

The invention discloses a space-time deep learning-based data-free watershed runoff prediction method and system. The method comprises the following steps: collecting static geographic raster data and dynamic hydro meteorological time series data of multiple watersheds, and performing preprocessing; constructing a double-flow deep learning model fusing space and time features; training a double-flow deep learning model based on the multi-region large-sample watershed data to obtain a general hydrological model; evaluating the adaptability of each network layer in the general hydrological model to a target watershed through a layered unfreezing test, and screening out a key adaptive layer; and under a leave-one-out method cross validation framework, layered progressive unfreezing transfer learning is carried out on the general hydrological model based on the key adaptation layer, and data-free drainage basin runoff prediction is realized. According to the method, multi-basin data driving, spatio-temporal feature fusion and a transfer learning mechanism are organically combined, the conversion of a hydrological modeling norm from local adaptation to global generalization is promoted, and a new path is provided for intelligent prediction of data-free basin runoff.

The invention discloses a flood control reservoir flood season application method giving consideration to water resource utilization, and the method comprises the steps: calculating and designing a water inflow process, a water demand process and a minimum reserve water amount and water level process according to the obtained long series of data of a drainage basin; according to the activation mechanism and the quit mechanism, timely developing a flood-control reservoir flood-season water resource application mode or a flood-season reservoir floating water level operation control strategy, and finally, after flood-season operation is finished, evaluating and evaluating the effect of the flood-control reservoir flood-season water resource application mode. The method can more scientifically guide the reserve water amount of the flood control reservoir at the proper time in the flood season to cope with the influence of drought events on downstream water supply safety, and the drainage basin flood and drought disaster prevention and water resource optimal allocation capacity is improved.

The invention discloses a less-data watershed runoff prediction method fusing multi-scale features and prediction results. Determining a watershed with data most similar to the watershed with less data; dividing the watershed with data into different watershed clusters by using a clustering algorithm, and determining the watershed cluster to which the watershed with less data belongs; constructing a multi-scale watershed LSTM model, namely a regional watershed LSTM model, a cluster watershed LSTM model and an individual watershed LSTM model; extracting multi-scale features of the less-data watershed by using a multi-scale watershed LSTM model and obtaining a multi-scale runoff prediction result of the less-data watershed; and by taking the multi-scale features and the runoff prediction result of the less-data watershed as input features and taking the final runoff prediction result of the less-data watershed as an output label, constructing a less-data watershed runoff prediction model. The method can effectively improve the runoff prediction precision and generalization ability of the watershed with less data, and has a wide application prospect in the field of runoff prediction.

The invention discloses a runoff prediction method, a runoff prediction device and a storage medium. The runoff prediction method comprises the following steps: calculating dynamic response characteristics of each grid unit at each moment by using meteorological input; extracting static influence characteristics of the static attributes on the runoff error; extracting features reflecting time and space influences at the same time, and obtaining an error correction value generated based on meteorological input and underlying surface information; constructing a reservoir scheduling substitution network, and outputting a correction component of the reservoir scheduling to the runoff; and constructing a loss function, superposing the correction component and the output of the preliminary hydrological model, and updating parameters of each module through end-to-end back propagation to obtain runoff prediction. According to the runoff prediction method, the influence of dynamic changes such as rainfall and air temperature, underlying surface distribution and reservoir regulation on drainage basin runoff is fully captured, and the precision and stability of runoff simulation are remarkably improved.

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 relates to a multi-scale response analysis method of drainage basin runoff to a landscape pattern, ecological hydrological research is the basis of comprehensive management of a drainage basin, and the response relationship of the runoff to the landscape pattern is an important decision basis for integrated protection and systematic management of landscapes, forests, lakes, grasses and sand in the drainage basin. Taking the Poyang Lake basin as an example, the method comprises the following steps: establishing a basin SWAT model; determining parameter values of the Poyang Lake basin based on an SUFI-2 method and verifying the correctness of the parameter values; inputting parameters, and calculating the runoff volume of the Poyang Lake drainage basin through an SWAT model; constructing a landscape type division system of'landscape, water, forest field, lake, grass and sand 'based on land utilization / cover; selecting landscape indexes through collinearity diagnosis; the influence of landscape composition on the runoff is discussed by utilizing correlation analysis methods such as a scatter diagram, a Pearson's correlation coefficient and a geographically weighted regression model, and the response of the runoff to the landscape pattern is revealed in landscape and type scales. According to the method, the Poyang Lake basin is taken as a research object, runoff simulation and landscape pattern analysis are respectively carried out by utilizing an SWAT model and Fragstats software, and nonlinear characteristics of runoff response to landscape composition and landscape pattern are revealed.

The invention discloses a drainage basin runoff prediction method and system coupling a physical mechanism and deep learning, and relates to the technical field of water resource management, and the method comprises the steps: obtaining historical meteorological driving data and historical runoff data of a to-be-predicted drainage basin, and obtaining the surface vegetation information of the to-be-predicted drainage basin; the physical information recurrent neural network module, the dynamic plant long-short-term memory network module and the grid full-distributed module are fused, and a drainage basin runoff prediction model is constructed; training a drainage basin runoff prediction model based on the historical meteorological driving data, the historical runoff data and the surface vegetation information to obtain a trained drainage basin runoff prediction model; and predicting the runoff of the to-be-predicted drainage basin based on the trained drainage basin runoff prediction model. The technical problem that runoff prediction errors are large in the prior art is solved.

The invention belongs to the technical field of runoff forecasting, and particularly discloses a deep learning runoff forecasting method and system considering spatial-temporal characteristics of multi-source data, and the method comprises the steps: training a runoff forecasting model based on a deep neural network based on a training set, and obtaining a trained runoff forecasting model; the training set comprises watershed meteorological data, underlying surface data and site data, and the site data comprises runoff data and rainfall data; the runoff forecasting model extracts meteorological element spatial-temporal features and underlying surface spatial features based on the meteorological data and the underlying surface data; based on time correlation characteristics of the runoff data and the rainfall data, time characteristics of the site data are extracted; matching the meteorological element spatial-temporal features with the underlying surface spatial features to obtain runoff and confluence features; and obtaining predicted runoff data according to the runoff production and convergence characteristics and the time characteristics of the station data. According to the method, accurate and effective drainage basin runoff forecasting can be realized, and the uncertainty of future runoff is reduced.

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 present application belongs to the technical field of hydrological prediction and water resources management, and specifically relates to a prediction method and system based on multi-scale decomposition of runoff sequence and dynamic weight reconstruction. The method first constructs a physical hydrological model based on meteorological driving data and generates a runoff simulation sequence; the simulation sequence is subjected to multi-scale decomposition by using variational mode decomposition, the decomposition parameters are adaptively determined by particle swarm optimization, and a plurality of mode components are obtained; a long short-term memory network is constructed to establish a mapping relationship between the contribution weights of the mode components, and dynamic weights varying with time and normalized are output; the mode components are weighted and reconstructed according to the weights, so as to realize deviation correction of the simulated runoff of the physical hydrological model on different time scale structures. In the prediction stage, the same decomposition is performed on the future runoff simulation sequence, and the trained model is directly used to output the runoff prediction result. The present application converts the runoff prediction problem into a dynamic weight distribution problem of multi-scale structure components, improves the prediction precision and migration ability while maintaining physical interpretability, and is suitable for scenarios such as basin runoff prediction, flood simulation, water resources scheduling and the like.

The application discloses a large-scale basin deep learning flood forecasting method based on runoff lag information and relates to the technical field of hydrological simulation and flood forecasting of machine learning, and comprises the following steps: collecting basin hydrological observation site information data, basin characteristic data, basin historical hydrological data and meteorological forecasting characteristic data; dividing a large-scale basin into a plurality of sub-basins; preprocessing data, extracting runoff lag information obtained by using a convolutional neural network model, and constructing a runoff lag information database; adopting a deep learning long short-term memory model to couple the runoff lag information database, obtaining a plurality of sub-basin runoff lag information models; obtaining optimal runoff lag information parameters of the plurality of sub-basins according to a model precision evaluation index, obtaining an optimal hydrological forecasting model, and realizing accurate flood forecasting of the large-scale basin. The application makes up for the deficiency of flood forecasting under the condition of lack of hydrological data of a large-scale basin and has a strong application prospect in areas with little or no data.

The invention discloses a reservoir scheduling method and system based on hydrological simulation, and belongs to the field of reservoir optimal scheduling method.The method comprises the steps that a hydrological environment simulation model is built based on hydrological environment data and a preset water and soil conservation model; obtaining initial drainage basin runoff data based on the hydrological environment simulation model; based on the initial drainage basin runoff data and a preset drought evaluation model, obtaining initial drainage basin drought characteristics; constructing a reservoir balance model based on the reservoir feature information; constructing a multi-target reservoir scheduling model based on the initial drainage basin runoff data, the initial drainage basin drought characteristics and a reservoir balance model; and solving the multi-target reservoir scheduling model to obtain a reservoir scheduling optimization scheme so as to realize the scheduling of the multi-target reservoir. Therefore, the problem that in the prior art, a reservoir dispatching optimization scheme is limited to a certain extent in practical application is solved, flexible dispatching and real-time regulation and control of the multi-target reservoir are achieved by building the model, and the hydrological drought problem is avoided.

The invention discloses an urban rainfall runoff calculation method based on drainage node downstream tracking. The method comprises the following steps: step 1, obtaining drainage pipe network data of a research area; 2, analyzing the flow direction of the drainage pipe network; step 3, dividing sub catchment areas and setting runoff production parameters; 4, generalizing drainage pipe network data; step 5, node generalization marking; 6, new outlet nodes of the sub catchment areas are determined; step 7, calculating confluence delay; and step 8, calculating runoff production of the sub catchment areas. According to the method, on the basis of a drainage node downstream tracking mode, the accuracy of runoff production calculation of the sub catchment areas is improved, the defect that a traditional method can change a catchment path and a spatial pattern is overcome, the accuracy of urban rainfall runoff calculation is improved, and powerful support is provided for urban flood control and disaster reduction.

The invention provides a watershed runoff prediction method and system, and relates to the technical field of watershed runoff prediction.The watershed runoff prediction method comprises the steps that a watershed to be predicted is subjected to gridding processing, multiple watershed grids are obtained, the center of the watershed grids is obtained, and the following operations are conducted on each watershed grid: multiple meteorological state models are established for the center of the watershed grids, each meteorological state model is used for describing one kind of meteorological data of the center of the drainage basin grid; periodically predicting various meteorological prediction data of the center of the drainage basin grid according to various meteorological state models by taking a first period as an interval; and periodically predicting runoff data of the center of the drainage basin grid based on the various meteorological prediction data and the initial runoff data of the center of the drainage basin grid by taking a second period as an interval, the second period being greater than the first period and being an integral multiple of the first period. The method has the advantage of improving the runoff prediction precision and efficiency.

The invention discloses a method for determining the contribution rate of the influence of climate and land cover change on runoff and carbon sequestration, and relates to the technical field of carbon neutralizer.The method comprises the steps that benchmark period meteorological data, benchmark period land cover data, change period meteorological data and change period land cover data of a target drainage basin are utilized; calculating a first runoff variable quantity and a first total primary productivity variable quantity of the target drainage basin in a climate change scene, and a second runoff variable quantity and a second total primary productivity variable quantity of the target drainage basin in a land cover change scene; according to the first runoff variation and the second runoff variation, determining the contribution rate of climate and land cover change to the runoff variation; according to the first total primary productivity variable quantity and the second total primary productivity variable quantity, determining the contribution rate of climate and land cover change to the total primary productivity variable quantity, and determining the contribution rate of the climate and land cover change to the runoff and carbon sequestration influence of the target drainage basin; and the reliability of determining the contribution rate of climate and land cover change to runoff and carbon sequestration is improved.

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.

The invention provides a method and system for calculating snow melting runoff based on a snow water equivalent distribution curve, and relates to the technical field of hydrological forecasting, and the method comprises the steps: carrying out the cumulative calculation of collected environment data through an air temperature threshold segmentation method, and obtaining the total snow water equivalent; constructing a snow water equivalent distribution curve according to the total snow water equivalent and the area proportion of the accumulated snow area, and performing integral correlation on the snow water equivalent distribution curve and the total snow water equivalent to obtain a maximum snow water equivalent value; on the basis of the environment data, calculating a potential accumulated snow melting amount through a degree-day factor method; according to the snow water equivalent distribution curve, the maximum snow water equivalent value and the potential accumulated snow melting amount, the snow melting amount in the unit time period is calculated; and converting the snowmelt amount in the unit time period into snowmelt runoff in each time period through a linear reservoir method. The method is low in data demand, simple in parameter and high in adaptability, can effectively avoid dependence on remote sensing observation, and improves the reliability of runoff forecasting and water resource scheduling management of the cold region drainage basin.

The invention discloses a watershed runoff simulation method fusing a random Xinanjiang model and machine learning, and belongs to the technical field of watershed runoff simulation. The method comprises the following steps: collecting hydro-meteorological data in a research basin; constructing a stochastic three-water-source Xinanjiang model based on a stochastic differential equation; utilizing Monte Carlo simulation to generate a runoff probability trajectory and extracting statistical characteristics; extracting a multi-scale feature vector of the runoff sequence based on discrete wavelet transform; and constructing machine learning models with different architectures, inputting a mean trajectory or a full feature set subjected to wavelet decomposition, and simulating the drainage basin runoff by adopting different information combinations. According to the method, the three-water-source Xinanjiang model based on the stochastic differential equation is constructed, and noise reduction and multi-scale decomposition of the runoff random trajectory are realized in combination with Monte Carlo simulation and discrete wavelet transform, so that a machine learning model can more accurately capture multi-scale hydrological signals; and the runoff simulation effect of the model is improved by improving the low and high flow simulation precision of the drainage basin.

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.

The application discloses a water regime monitoring abnormal data management system based on multi-source heterogeneous data fusion, and belongs to the technical field of hydrological monitoring and data processing.The application reflects the river flow direction and the confluence relationship by constructing a directed and weighted topological atlas, extracts the spatial and temporal characteristics of the station, and fuses the catchment runoff model and the upstream water level evolution law to calculate the causal confidence index, so that the sensor failure and the real hydrological process are accurately distinguished from the physical cause level; sliding filtering, spatial weighted interpolation or variational autoencoder network are matched according to different abnormal levels for adaptive repair; the edge gateway and the cloud are cooperated to realize real-time response and global optimization. The application greatly reduces the false alarm rate of water regime monitoring, improves the accuracy and robustness of abnormal data identification and repair, and provides reliable data support for intelligent water conservancy.

The invention relates to the technical field of hydrological data analysis, and provides a runoff reanalysis data error analysis method and system based on panel regression analysis, and the method comprises the following steps: collecting a watershed hydro meteorological data set, watershed runoff reanalysis data and meteorological reanalysis data, extracting drainage basin runoff reanalysis time series data and drainage basin meteorological reanalysis time series data; constructing a fixed effect panel regression model of the meteorological input deviation and the runoff simulation deviation based on the deviation between the reanalysis data and the observation value; the method comprises the following steps of: clustering all drainage basins according to drainage basin site attribute data to obtain spatial clusters of the drainage basins and attribute characteristics of each cluster, resampling each cluster through a bootstrap method to form first panel data, performing fixed effect panel regression on the first panel data to obtain a corresponding regression coefficient, and performing fixed effect panel regression on the first panel data to obtain a regression result; a first coefficient distribution for measuring spatial heterogeneity effects of meteorological deviations is generated.

Embodiments of the present application disclose a small hydropower group output prediction method and device, computer equipment and a storage medium, wherein the method comprises: finding a runoff prediction model corresponding to a target single watershed from a target runoff prediction model set corresponding to a target watershed class to obtain a search result; if the search result is successful, inputting target data corresponding to the target single watershed into the runoff prediction model corresponding to the search result to obtain a single watershed runoff prediction result; if the search result is unsuccessful, inputting the target data corresponding to the target single watershed into each runoff prediction model in the target runoff prediction model set to obtain a to-be-analyzed runoff set, inputting the to-be-analyzed runoff set into a runoff fitting model corresponding to the target single watershed for fitting to obtain the single watershed runoff prediction result; and obtaining a small hydropower group output prediction result according to each single watershed runoff prediction result; thereby obtaining a prediction result with excellent prediction performance and an accuracy meeting use requirements.

The present invention discloses a method for diagnosing the problem of water volume imbalance in a basin hydropower station and correcting the outflow discharge, which includes the following steps: Step 1, organizing basic data; Step 2, calculating the probability distribution of the basin runoff coefficient; Step 3, for the data of unreasonable points and the data of points not falling within the confidence interval in Step 2, it is necessary to diagnose the problem according to the actual situation and judge whether there is a problem with the outflow discharge of the hydropower station corresponding to this point; Step 4, for Step 3, if there is a problem with the outflow discharge, establish the mathematical relationship between the outflow discharge of this hydropower station, the power generation discharge, and the flood discharge; calibrate the parameters α and β based on the data during the calibration period by calculating the runoff coefficient formula of this hydropower station and the downstream hydropower station; Step 5, calculate the outflow discharge of each period during the inspection period through the parameters α and β obtained during the calibration period. The present invention can handle the problem of large errors in the outflow discharge caused by water volume imbalance.

The invention discloses a multi-year regulation reservoir cascade scheduling method considering a high-water annual runoff utilization rate, and the method comprises the steps: dividing a high-water annual scene according to long-series natural runoff data; setting a discrete flood control water level set for regulating the reservoir for many years according to the situation of the wet water year and the actual scheduling experience; constructing a multi-year sea-entering water abandoning rate index to represent a drainage basin runoff utilization rate; taking the discrete flood control water level set and the actual operation water level as boundary conditions, and taking the maximum power generation amount, the maximum sand discharge amount and the minimum sea-entering water abandoning rate as targets to construct a cascade reservoir group combined dispatching model; performing multi-objective optimization calculation on the discrete flood control water levels to obtain Pareto improved solution sets under different flood control water levels; according to the Pareto improved solution set, the optimal combination of the flood control water level of the multi-year regulation reservoir and the last-year water elimination and falling level is determined; and according to the optimal combination, dispatching the multi-year regulation reservoir and the cascade reservoir. The water resource utilization efficiency of the Yellow River basin can be improved.

The application discloses a watershed runoff probability prediction method based on Bayesian deep learning and a differentiable physical model, and the method comprises the following steps: firstly, a digital elevation model is used to construct a watershed differentiable distributed computing architecture, a parameter inference network is connected with an HBV hydrological model and a Muskingum confluence equation in series to form an end-to-end differentiable computation graph; then, Bayesian inference is realized by introducing a random inactivation layer and an input disturbance mechanism, and a random parameter group and a random rainfall scenario conforming to a posterior distribution are generated; finally, parallel set physical evolution is performed, and a runoff probability prediction and a flood risk assessment interval are output; the method combines a physical mechanism and deep learning, realizes high-precision deterministic prediction and reasonable uncertainty quantification, and provides more reliable decision support for flood warning.

This invention discloses a method for predicting watershed runoff evolution trends based on multi-source data fusion, belonging to the field of hydrological and water resources prediction technology. The method includes acquiring historical hydrological and meteorological data and classifying rainfall-runoff events; calculating soil water absorption per unit time using water balance; constructing hysteresis loops by fitting wet and dry water absorption paths; clustering the hysteresis loop set to extract representative hysteresis loop families; matching the most similar hysteresis loop during prediction; inputting the data into a reinforcement learning model to determine the output path advancement position; determining the corresponding soil water absorption; and combining meteorological data to calculate the predicted runoff. This invention solves the problem of low runoff prediction accuracy caused by traditional models neglecting the differences in water absorption capacity during wet and dry processes by quantifying the nonlinear hysteresis characteristics of soil water absorption behavior.