Intelligent Estimation Method of Ecological Flow in Ungauged Mountainous River Basins Based on Multi-Scale Fusion

Abstract

The present invention provides an intelligent estimation method for ecological flow in data - scarce mountainous watersheds based on multi - scale fusion, including: collecting basic data of the research area and pre - processing it to obtain an effective data set; extracting a hydrological feature set and a process feature set according to the effective data set; generating dynamic response units based on the hydrological feature set and the process feature set, and performing multi - dimensional hydrological process decomposition to obtain a multi - scale feature matrix; performing ecological hydrological sequence matching on the multi - scale feature matrix to obtain a flow pattern set, and performing hierarchical uncertainty propagation analysis to generate an uncertainty index set; constructing an integrated model according to the flow pattern set and the uncertainty index set, and performing multi - constraint optimization to obtain the ecological flow estimation result. The present invention effectively solves the problem of parameter uncertainty caused by the spatio - temporal heterogeneity of rainfall in mountainous areas by introducing a terrain complexity index and a hydrological similarity index, and improves the accuracy of ecological flow estimation.

Description

Technical Field

[0001] The present invention relates to hydrological and hydrodynamic technologies, in particular to an intelligent estimation method for ecological flow in data - scarce mountainous watersheds based on multi - scale fusion. Background Art

[0002] Intelligent estimation of ecological flow in data - scarce mountainous watersheds based on multi - scale fusion is of great significance for maintaining the health of river ecosystems and ensuring regional sustainable development. Mountainous watersheds are characterized by complex terrain, variable climate, strong heterogeneity of the underlying surface, and few observation stations, which pose great challenges to the accurate estimation of ecological flow. Accurately estimating the ecological flow in mountainous watersheds can not only provide a scientific basis for ecological water allocation and river health assessment, but also provide decision - making support for watershed integrated management and ecological restoration, which has important practical significance for balancing ecological protection and economic development.

[0003] Currently, the estimation of ecological flow in mountainous watersheds mainly adopts hydrological methods, hydraulic methods, and habitat methods, etc. Hydrological methods determine ecological flow by setting fixed thresholds based on historical flow sequences, such as the Tennant method and the flow duration curve method, etc.; hydraulic methods establish estimation models based on the relationship between hydraulic parameters and ecological needs, such as the wet - perimeter method and the R2CROSS method, etc.; habitat methods evaluate ecological flow by constructing habitat suitability models, such as the PHABSIM model, etc. At the same time, researchers have also begun to try to introduce multi - source data such as remote sensing data and geographic information systems into ecological flow estimation and use methods such as machine learning to improve the estimation accuracy.

[0004] However, the existing research still has the following technical problems: the spatio - temporal heterogeneity of mountain rainfall is significant, and it is difficult for existing methods to accurately characterize this heterogeneity during the parameter calibration process, resulting in large parameter uncertainties; during the multi - source data fusion process, there are differences in the spatio - temporal resolutions of different data sources, and traditional scale - conversion methods fail to effectively handle this scale difference, causing significant scale - conversion errors; in terms of model integration, existing methods mostly adopt static weight - allocation strategies, which fail to fully adapt to the spatio - temporal heterogeneity characteristics of mountainous watersheds and affect the reliability of the estimation results. The existence of these technical problems severely restricts the accuracy and reliability of ecological flow estimation in data - scarce mountainous watersheds. Summary of the Invention

[0005] The object of the invention is to provide an intelligent estimation method for ecological flow in data - scarce mountainous watersheds based on multi - scale fusion, in order to solve the above - mentioned problems existing in the prior art.

[0006] Technical Solution: The intelligent estimation method for ecological flow in data - scarce mountainous watersheds based on multi - scale fusion includes the following steps:

[0007] Collect the basic data of the study area, and perform preprocessing to obtain an effective data set; the basic data includes DEM data, remote sensing image data, meteorological reanalysis data, and land use data;

[0008] Extract the hydrological feature set and process feature set according to the effective data set;

[0009] Generate dynamic response units based on the hydrological feature set and process feature set, and perform multi-dimensional hydrological process decomposition to obtain a multi-scale feature matrix;

[0010] Perform ecological hydrological sequence matching on the multi-scale feature matrix to obtain a flow pattern set, and perform hierarchical uncertainty propagation analysis to generate an uncertainty index set;

[0011] Construct an integrated model according to the flow pattern set and uncertainty index set, and perform multi-constraint optimization to obtain the ecological flow estimation result.

[0012] According to one aspect of the present application, the preprocessing includes:

[0013] Calculate the terrain complexity index value of the basic data to obtain a terrain decomposition factor;

[0014] Perform local decomposition and global decomposition on the basic data based on the terrain decomposition factor to generate a decomposition coefficient matrix;

[0015] Calculate the hydrological similarity value according to the decomposition coefficient matrix to obtain a similarity index;

[0016] Generate a data quality index based on the similarity index to obtain an effective data set.

[0017] According to one aspect of the present application, the steps of extracting the hydrological feature set and process feature set according to the effective data set include:

[0018] Extract the slope value, aspect value, and elevation value of the grid cell, and calculate the geomorphic position index to obtain a position index matrix;

[0019] Divide the geomorphic unit types based on the position index matrix, and extract the characteristic parameters of each type of unit to generate a geomorphic feature vector;

[0020] Construct a feature decomposition basis function based on the geomorphic feature vector to obtain a basis function set;

[0021] Calculate the time response relationship between rainfall data and runoff data to obtain a time lag coefficient matrix;

[0022] Perform feature extraction on the effective data set according to the basis function set and time lag coefficient matrix to obtain a feature component set;

[0023] Perform multi-scale combination on the feature component set to generate a hydrological feature set and a process feature set.

[0024] According to one aspect of the present application, the steps of generating a dynamic response unit include:

[0025] Calculating the topographic gradient values at each point in the watershed to obtain a gradient matrix;

[0026] Constructing a spatial neighborhood relationship based on the gradient matrix to generate a neighborhood matrix;

[0027] Calculating the confluence time of each region to obtain a time matrix;

[0028] Constructing a dynamic partitioning criterion based on the neighborhood matrix and the time matrix to obtain a partitioning criterion;

[0029] Performing adaptive partitioning according to the partitioning criterion to obtain an initial partitioning result;

[0030] Dynamically adjusting the initial partitioning result according to seasonal changes to obtain a dynamic response unit.

[0031] According to one aspect of the present application, the steps of performing multi-dimensional hydrological process decomposition include:

[0032] Extracting the topographic features of each unit to obtain a topographic feature set;

[0033] Calculating the hydrological response parameters of each unit to obtain a response parameter set;

[0034] Constructing a feature expression at the unit scale to obtain a feature function set;

[0035] Performing multi-dimensional decomposition based on the feature function set to obtain an initial decomposition set;

[0036] Optimizing the decomposition coefficients of the initial decomposition set to generate a decomposition coefficient set;

[0037] Calculating the energy values at each decomposition scale based on the decomposition coefficient set to obtain a multi-scale feature matrix.

[0038] According to one aspect of the present application, the steps of ecological hydrological sequence matching to obtain a flow pattern set include:

[0039] Extracting the water volume change sequence to obtain a water volume sequence;

[0040] Calculating the characteristics of the ecological sensitive period to obtain a sensitive period matrix;

[0041] Analyzing the ecological water demand to obtain a demand matrix;

[0042] Calculating the response intensity based on the water volume sequence and the sensitive period matrix to obtain an intensity matrix;

[0043] Combining the demand matrix and the intensity matrix to construct a response feature to generate a response feature set;

[0044] Perform pattern recognition based on the response feature set to generate a traffic pattern set.

[0045] According to one aspect of the present application, the steps of performing hierarchical uncertainty propagation analysis to generate an uncertainty index set include:

[0046] Construct a hydrological process transfer chain network to obtain a network structure;

[0047] Calculate the parameter sensitivity index of each node to generate a sensitivity matrix;

[0048] Analyze the uncertainty transfer path according to the network structure and the sensitivity matrix to obtain a transfer path set;

[0049] Evaluate the coupling effect of multi-source uncertainty to obtain a coupling coefficient set;

[0050] Calculate the comprehensive uncertainty based on the transfer path set and the coupling coefficient set to generate an uncertainty index set.

[0051] According to one aspect of the present application, the steps of constructing an integrated model include:

[0052] Extract terrain feature parameters to obtain a terrain feature matrix;

[0053] Calculate hydrological response indicators to obtain a response indicator matrix;

[0054] Analyze land use characteristics to obtain a land feature matrix;

[0055] Evaluate climate variability to obtain a climate indicator matrix;

[0056] Construct a model selection criterion based on the terrain feature matrix, the response indicator matrix, the land feature matrix, and the climate indicator matrix;

[0057] Calculate the spatio-temporal heterogeneity index of each region to obtain a heterogeneity matrix;

[0058] Determine the model weights according to the model selection criterion and the heterogeneity matrix to generate a weight vector;

[0059] Evaluate the reliability index of each model to obtain a reliability matrix;

[0060] Construct an integrated model using the weight vector and the reliability matrix.

[0061] According to one aspect of the present application, the steps of performing multi-constraint optimization to obtain an ecological flow estimation result include:

[0062] Construct a seasonal change function to obtain a change function;

[0063] Design an ecological demand constraint system to obtain a set of constraint conditions;

[0064] Construct an objective function considering seasonal changes;

[0065] Optimize and solve according to the constraint condition set and the objective function to obtain the initial result;

[0066] Analyze the response characteristics of the ecosystem to the initial result to obtain the response index;

[0067] Based on the response index, correct the initial result to generate the ecological flow estimation result.

[0068] According to one aspect of the present application, the ecological hydrological sequence matching further includes:

[0069] Extract the river network structure data to obtain the river network matrix;

[0070] Calculate the physical connectivity between river reaches to obtain the physical connectivity matrix;

[0071] Analyze the hydrological connectivity status to obtain the hydrological connectivity matrix;

[0072] Evaluate the ecological corridor function to obtain the corridor index matrix;

[0073] Integrate the physical connectivity matrix, the hydrological connectivity matrix and the corridor index matrix to generate the connectivity matrix;

[0074] The connectivity matrix and the response characteristic set are used for pattern recognition to generate the flow pattern set.

[0075] According to one aspect of the present application, the steps of analyzing the uncertainty transmission path include:

[0076] Design a perturbation experiment scheme to obtain the perturbation scheme;

[0077] Execute the parameter perturbation test to obtain the test result;

[0078] Calculate the response sensitivity to obtain the sensitivity matrix;

[0079] Analyze the parameter importance to obtain the importance vector;

[0080] Identify the uncertainty source to obtain the source matrix;

[0081] Analyze the coupling action mode to obtain the action mode matrix;

[0082] Generate the coupling coefficient set based on the action mode matrix.

[0083] Beneficial effects: By introducing the terrain complexity index and the hydrological similarity index, the problem of parameter uncertainty caused by the spatio-temporal heterogeneity of mountain rainfall is effectively solved, and the accuracy of ecological flow estimation is improved. Description of the Drawings

[0084] Figure 1 is the flowchart of the present invention.

[0085] Figure 2 is the flowchart of the preprocessing of the present invention.

[0086] Figure 3 is the flowchart of extracting the hydrological feature set and the process feature set according to the effective data set in the present invention.

[0087] Figure 4 is the flowchart of generating the dynamic response unit in the present invention.

[0088] Figure 5 is the flowchart of performing multi-dimensional hydrological process decomposition in the present invention. Detailed implementation manners

[0089] As Figure 1 shown, the intelligent estimation method for ecological flow in ungauged mountainous watersheds based on multi-scale fusion includes:

[0090] Step S1: Receive DEM data, remote sensing image data, meteorological reanalysis data, and land use data, process various input data by using the multi-scale adaptive decomposition method to generate a data quality index and an effective data set; according to the effective data set, perform adaptive extraction of hydrological process features to obtain a hydrological feature set and a process feature set.

[0091] Step S2: Receive the hydrological feature set and the process feature set, process them by using the terrain-guided adaptive zoning method to generate dynamic response units; perform multi-dimensional hydrological process decomposition based on the dynamic response units to obtain a multi-scale feature matrix.

[0092] Step S3: Receive the multi-scale feature matrix, process it by using the eco-hydrological sequence matching algorithm to obtain a flow pattern set; perform hierarchical uncertainty propagation analysis according to the flow pattern set to generate an uncertainty index set.

[0093] Step S4: Receive the flow pattern set and the uncertainty index set, construct an integrated model E by using the adaptive dynamic integration method; perform multi-constraint optimization according to the integrated model to obtain the ecological flow estimation result and the reliability evaluation index.

[0094] In addition to the problems in the background technology, it is found in the research that the ecosystem shows obvious non-linear response characteristics to flow changes, and existing methods mostly use linear or simple non-linear models to describe it, making it difficult to accurately depict this complex response relationship; in addition, under extreme climate conditions, the ecological flow shows unique dynamic change laws, but existing research mainly focuses on the flow characteristics under normal climate conditions and lacks in-depth analysis of the dynamic characteristics of ecological flow under extreme conditions; finally, current uncertainty analysis methods often consider each link separately and lack a systematic analysis of the uncertainty propagation mechanism on the hydrological process chain, making it difficult to quantify the coupling effect of multi-source uncertainties.

[0095] Through the technical route of multi-scale fusion, the intelligent estimation of ecological flow in mountainous basins without data is realized. Through innovative methods such as multi-scale adaptive decomposition, terrain-guided dynamic zoning, eco-hydrological sequence matching, and adaptive dynamic integration, this scheme effectively solves technical problems such as scarce data, complex terrain, and variable process responses in mountainous basins. Especially when dealing with multi-source heterogeneous data, an adaptive decomposition strategy based on terrain features is adopted to improve data quality and reliability; during the model construction process, through the division of dynamic response units and multi-dimensional feature extraction, a refined description of the hydrological process in mountainous basins is achieved; in the link of ecological flow estimation, combined with ecological demand constraints and multi-objective optimization, the scientificity and reliability of the estimation results are ensured. The overall scheme has strong universality and practicability, and can provide strong technical support for the ecological flow management of mountainous basins.

[0096] According to one aspect of the present application, step S1 is specifically as follows:

[0097] Step S11: Receive DEM data, remote sensing image data, meteorological reanalysis data, and land use data, calculate the terrain complexity index values of each data, and obtain terrain decomposition factors; based on the terrain decomposition factors, perform local decomposition and global decomposition on various input data to generate a decomposition coefficient matrix; according to the decomposition coefficient matrix, calculate the hydrological similarity values between each data point and adjacent data points to obtain similarity indicators; use the similarity indicators to identify outliers and generate data quality indicators; based on the data quality indicators, screen the original input data to obtain an effective data set.

[0098] Step S12: Receive the effective data set, extract the characteristic parameters of each watershed geomorphic unit, and construct a geomorphic feature vector; generate a feature decomposition basis function according to the geomorphic feature vector to obtain a set of basis functions; calculate the time response relationship between rainfall data and runoff data to obtain a time lag coefficient matrix; use the set of basis functions and the time lag coefficient matrix to extract features from the effective data set to obtain a set of feature components; perform multi-scale combination on the set of feature components to generate a hydrological feature set and a process feature set.

[0099] The input data is processed by a multi-scale adaptive decomposition method, achieving efficient fusion and quality control of multi-source heterogeneous data under complex terrain conditions in mountainous areas. Specifically, first, the spatial variability of terrain features is quantified by calculating the terrain complexity index. Based on this, the terrain decomposition factor constructed can adaptively adjust the scale and intensity of data decomposition, enabling a more detailed decomposition method to be adopted in complex terrain areas such as steep mountains and canyons, while a relatively rough decomposition is used in flat terrain areas, thus ensuring the matching degree of the decomposition result with the actual terrain features. At the same time, by calculating the hydrological similarity between data points, an outlier identification standard based on physical mechanisms is established, which can effectively screen out abnormal data affected by factors such as terrain and meteorology, improving the data quality for subsequent analysis. This data preprocessing method based on multi-scale decomposition not only improves the reliability of the data, but also the hydrological feature set and process feature set obtained through feature extraction can comprehensively characterize the hydrological process features of mountainous basins, providing a high-quality data basis for subsequent ecological flow estimation.

[0100] According to one aspect of the present application, step S2 is specifically as follows:

[0101] Step S21: Receive the hydrological feature set and process feature set, calculate the terrain gradient values of each point in the basin to obtain a gradient matrix; construct a spatial neighborhood relationship based on the gradient matrix to generate a neighborhood matrix; calculate the confluence time of each region to obtain a time matrix; construct a dynamic partitioning criterion based on the neighborhood matrix and time matrix to obtain a partitioning criterion; use the partitioning criterion to adaptively partition the study area to generate an initial partitioning result; dynamically adjust the initial partitioning result according to seasonal changes to obtain a dynamic response unit.

[0102] Step S22: Receive the dynamic response unit, extract the basin feature parameters of each unit to construct a feature function set; perform multi-dimensional decomposition on each response unit according to the feature function set to obtain a decomposition coefficient set; calculate the energy distribution of each component in the decomposition coefficient set to obtain an energy matrix; identify the dominant mode in the energy matrix to generate a mode matrix; calculate the information flux between each scale to obtain a flux index; perform feature recombination according to the mode matrix and flux index to generate a multi-scale feature matrix.

[0103] An adaptive zoning method guided by terrain is adopted to achieve a refined description and analysis of complex hydrological processes in mountainous watersheds. This method first constructs a spatial neighborhood relationship based on terrain gradients, and establishes a zoning criterion considering hydrological connectivity by calculating the flow concentration time, so that the zoning results can accurately reflect the hydrological response characteristics of mountainous watersheds. Especially in mountainous areas with significant seasonal variations, by dynamically adjusting the zoning boundaries, the changing characteristics of the watershed hydrological processes in different seasons can be effectively captured. This method of dividing dynamic response units breaks through the limitations of traditional static zoning methods and can better adapt to the spatio-temporal heterogeneity of hydrological processes in mountainous watersheds. At the same time, through the decomposition of multi-dimensional hydrological processes, the complex hydrological processes are decomposed into characteristic components at different scales, and the dominant patterns are identified through energy distribution analysis, realizing the extraction of multi-scale characteristics of hydrological processes in mountainous watersheds, providing a scientific basis for accurately estimating ecological flow.

[0104] According to one aspect of the present application, step S3 is specifically as follows:

[0105] Step S31: Receive the multi-scale feature matrix, extract the ecological response characteristics of each time series to obtain a response feature set; calculate the connectivity index of the water system in the watershed to generate a connectivity matrix; construct a similarity metric criterion based on the response feature set and the connectivity matrix to obtain a similarity index; perform multi-time scale decomposition on the data series to obtain a time series set; use the similarity index to perform pattern matching on the time series set to generate a flow pattern set.

[0106] Step S32: Receive the flow pattern set, construct a hydrological process transfer chain network to obtain a network structure; calculate the parameter sensitivity index of each node to generate a sensitivity matrix; analyze the uncertainty transfer path based on the network structure and the sensitivity matrix to obtain a transfer path set; evaluate the coupling effect of multi-source uncertainties to obtain a coupling coefficient set; calculate the comprehensive uncertainty based on the transfer path set and the coupling coefficient set to generate an uncertainty index set.

[0107] Through the ecological hydrological sequence matching algorithm and hierarchical uncertainty transfer analysis, the accurate identification and uncertainty quantification of ecological flow characteristics in mountainous watersheds are realized. This method first constructs a similarity metric criterion based on ecological response characteristics and water system connectivity. Through multi-time scale decomposition and pattern matching, the flow pattern characteristics under different hydrological scenarios can be accurately identified. Especially in mountainous watersheds with scarce data, this sequence matching-based method can make full use of limited observed data to extract representative flow patterns. At the same time, by constructing a hydrological process transfer chain network, analyzing the transfer law of uncertainty between different levels, and evaluating the coupling effect of multi-source uncertainties, the systematic quantification of the uncertainty of ecological flow estimation results is realized, providing a reliable uncertainty assessment basis for decision-making.

[0108] According to one aspect of the present application, step S4 is specifically as follows:

[0109] Step S41: Receive the flow pattern set and the uncertainty index set, extract the basin characteristic parameters, generate the characteristic index set; establish the model selection criteria according to the characteristic index set to obtain the selection criterion; calculate the spatio-temporal heterogeneity indexes of each region to obtain the heterogeneity matrix; determine the model weights based on the selection criterion and the heterogeneity matrix to generate the weight vector; evaluate the reliability indexes of each model to obtain the reliability matrix; construct an ensemble model using the weight vector and the reliability matrix to generate the ensemble model.

[0110] Step S42: Receive the ensemble model, establish the ecological demand constraint system to obtain the constraint condition set; construct the objective function considering seasonal variations to generate the objective function; perform optimization solution according to the constraint condition set and the objective function to obtain the initial result; analyze the response characteristics of the ecosystem to the initial result to obtain the response index; correct the initial result based on the response index to generate the ecological flow estimation result; evaluate the reliability of the ecological flow estimation result to obtain the reliability evaluation index.

[0111] Through the adaptive dynamic ensemble method and multi-constraint optimization, the accurate estimation and reliability evaluation of the ecological flow in mountainous basins are realized. This method first constructs the model selection criteria based on the basin characteristics, determines the model weights considering spatio-temporal heterogeneity, and establishes an ensemble model suitable for the characteristics of mountainous basins. Especially when considering seasonal variations, by establishing the ecological demand constraint system and multi-objective optimization, it can ensure the scientific rationality of the estimation result while meeting the needs of the ecosystem. At the same time, by analyzing the response characteristics of the ecosystem to correct the estimation result and establishing a systematic reliability evaluation system, the dynamic optimization and quality control of the estimation result are achieved, providing reliable technical support for the ecological flow management in mountainous basins.

[0112] According to one aspect of the present application, step S12 is specifically as follows:

[0113] Step S121: Receive the effective data set, extract the slope values of the grid cells to obtain the slope matrix; calculate the aspect values of the grid cells to obtain the aspect matrix; extract the elevation values of the grid cells to obtain the elevation matrix; calculate the geomorphic position index according to the slope matrix, aspect matrix and elevation matrix to obtain the position index matrix; divide the geomorphic unit types based on the position index matrix to obtain the unit type matrix; extract the parameters such as the area ratio, average elevation and slope of each type of unit according to the unit type matrix to generate the geomorphic feature vector.

[0114] Step S122: Receive the geomorphic feature vectors, construct the spatial distribution function of geomorphic units to obtain a set of spatial functions; calculate the confluence path lengths of each geomorphic unit to obtain a path length matrix; construct a basis function expression based on the set of spatial functions and the path length matrix to generate a set of basis functions; extract the rainfall data sequence from the valid dataset to obtain a rainfall sequence; extract the runoff data sequence from the valid dataset to obtain a runoff sequence; calculate the cross-correlation coefficient between the rainfall sequence and the runoff sequence to generate a correlation coefficient matrix; identify the peak response time based on the correlation coefficient matrix to obtain the time delay coefficient matrix R.

[0115] Step S123: Receive the set of basis functions, the time delay coefficient matrix, and the valid dataset, construct a projection matrix for feature extraction to obtain a projection matrix; use the projection matrix to transform the valid dataset to obtain a set of transformation coefficients; perform time scale adjustment on the set of transformation coefficients according to the time delay coefficient matrix to obtain a set of adjusted coefficients; calculate the principal components of the set of adjusted coefficients to generate a set of feature components.

[0116] Step S124: Receive the set of feature components, construct a multi-scale decomposition window sequence to obtain a window sequence; perform sliding decomposition on the set of feature components according to the window sequence to obtain a set of decomposition coefficients; calculate the energy distribution of the set of decomposition coefficients at different scales to obtain an energy distribution matrix; select the scale threshold for feature recombination based on the energy distribution matrix to obtain a threshold vector; perform recombination on the set of decomposition coefficients based on the threshold vector to generate a set of hydrological features; extract process feature parameters according to the set of hydrological features to obtain a set of process features.

[0117] Through the feature extraction and multi-scale combination method based on geomorphic units, the systematic extraction and characterization of hydrological features in mountainous basins are realized. This method first calculates the geomorphic position index based on topographic factors such as slope, aspect, and elevation, constructs feature basis functions through spatial distribution function and confluence path analysis, and establishes a feature extraction framework suitable for the topographic features of mountainous areas. Especially when dealing with the rainfall-runoff response relationship, the peak response time is identified by calculating the cross-correlation coefficient, accurately depicting the hydrological response characteristics of mountainous basins. At the same time, multi-scale sliding decomposition and energy distribution analysis are adopted to realize the multi-scale recombination of hydrological features, which not only retains the important feature information at different scales but also improves the accuracy and representativeness of feature expression. This feature extraction method fully considers the topographic features and hydrological process characteristics of mountainous basins, providing reliable feature support for subsequent ecological flow estimation.

[0118] According to one aspect of the present application, step S21 is specifically as follows:

[0119] Step S211: Receive the hydrological feature set and the process feature set, extract the terrain elevation data to obtain an elevation matrix; calculate the height differences between adjacent grids in the elevation matrix to obtain a height difference matrix; calculate the slope values according to the height difference matrix and the grid spatial resolution to obtain a slope matrix; combine the slope matrix and the flow direction information to calculate the terrain gradient and generate a gradient matrix.

[0120] Step S212: Receive the gradient matrix, define the spatial search radius to obtain the search parameter; identify the neighborhood units of each grid according to the search parameter to obtain a neighborhood set; calculate the spatial connection strength between the units in the neighborhood set to obtain a connection strength matrix; construct a neighborhood relationship network based on the connection strength matrix and generate a neighborhood matrix.

[0121] Step S213: Receive the neighborhood matrix, extract the river network structure information to obtain a river network matrix; calculate the flow velocity parameters of each river section in the river network matrix to obtain a flow velocity matrix; combine the gradient matrix and the flow velocity matrix to calculate the confluence path and obtain a path matrix; calculate the confluence time based on the path matrix and generate a time matrix.

[0122] Step S214: Receive the neighborhood matrix and the time matrix, construct a spatial constraint function to obtain the constraint function; calculate the similarity index between the units according to the constraint function to obtain a similarity matrix; set the partition threshold standard to obtain a threshold vector; comprehensively construct a partition criterion based on the similarity matrix and the threshold vector and generate a partition criterion.

[0123] Step S215: Receive the partition criterion, initialize the partition identifier to obtain an identifier matrix; iteratively optimize the identifier matrix according to the partition criterion C to obtain an optimized matrix; check the spatial continuity of the optimized matrix to obtain a continuity index; correct the partition result based on the continuity index and generate an initial partition result.

[0124] Step S216: Receive the initial partition result, extract the seasonal hydrological features to obtain a seasonal feature matrix; construct a seasonal adjustment function to obtain the adjustment function; calculate the dynamic change range of the partition boundary according to the adjustment function to obtain a change range matrix; adjust the initial partition result based on the change range matrix and generate a dynamic response unit.

[0125] The terrain-guided adaptive zoning method realizes the dynamic zoning and spatial unit division of mountainous watersheds. This method first calculates the terrain gradient based on elevation and slope, establishes a neighborhood relationship network through spatial search and connection strength analysis, and constructs a spatial zoning framework considering terrain features. Especially when calculating the confluence time, by combining the river network structure and flow velocity parameters, it accurately reflects the hydrological connectivity characteristics of mountainous watersheds. At the same time, by constructing spatial constraint functions and similarity indicators, the optimization and adjustment of the zoning results are realized, and the zoning boundaries are dynamically adjusted through seasonal feature analysis, so that the zoning results can better adapt to the spatio-temporal variation characteristics of the hydrological processes in mountainous watersheds. This dynamic zoning method breaks through the limitations of traditional static zoning and provides a more accurate basis for the spatial unit division in the analysis of hydrological processes in mountainous watersheds.

[0126] According to one aspect of the present application, step S22 is specifically as follows:

[0127] Step S221: Receive dynamic response units, extract the terrain features of each unit to obtain a terrain feature set; calculate the hydrological response parameters of each unit to obtain a response parameter set; construct a feature expression at the unit scale to obtain an expression matrix; generate a feature function according to the expression matrix to obtain a feature function set.

[0128] Step S222: Receive the feature function set, construct a multi-dimensional decomposition basis function to obtain a basis function matrix; design a decomposition scale sequence to obtain a scale sequence; perform multi-dimensional decomposition according to the basis function matrix and the scale sequence to obtain an initial decomposition set; optimize the decomposition coefficients of the initial decomposition set to generate a decomposition coefficient set.

[0129] Step S223: Receive the decomposition coefficient set, calculate the energy values of each decomposition scale to obtain an energy value matrix; construct an energy distribution function to obtain a distribution function; analyze the energy distribution characteristics using the distribution function to generate an energy matrix.

[0130] Step S224: Receive the energy matrix, set a pattern recognition criterion to obtain a recognition criterion; extract the dominant pattern features according to the recognition criterion to obtain a feature matrix; analyze the spatio-temporal distribution of the dominant pattern to obtain a distribution matrix; integrate the feature matrix and the distribution matrix to generate a pattern matrix.

[0131] Step S225: Receive the pattern matrix, construct a scale conversion function to obtain a conversion function; calculate the information transfer amount between adjacent scales to obtain a transfer matrix; evaluate the intensity of information flow to obtain an intensity index; synthesize the transfer matrix and the intensity index to generate a flow index.

[0132] Step S226: Receive the mode matrix and traffic metrics, determine the feature recombination weights to obtain a weight vector; perform feature recombination calculations based on the weight vector to obtain a recombination matrix; evaluate the representativeness of the recombination result to obtain a representativeness metric; optimize the recombination matrix based on the representativeness metric to generate a multi-scale feature matrix.

[0133] Through the multi-dimensional hydrological process decomposition and feature recombination method, the multi-scale feature extraction of the complex hydrological process in mountainous watersheds is realized. This method first constructs a feature function based on topographic features and hydrological response parameters, decomposes the process through multi-dimensional decomposition basis functions and decomposition scale sequences, and establishes a feature extraction framework suitable for the characteristics of mountainous watersheds. Especially in the process of energy distribution analysis, by identifying the dominant mode and analyzing the spatio-temporal distribution characteristics, the key features of the hydrological process are accurately captured. At the same time, by calculating the information flux between scales and constructing feature recombination weights, the optimized recombination of the decomposed features is realized, which not only ensures the representativeness of the recombination result but also improves the accuracy of feature expression. This multi-dimensional decomposition method can effectively process the multi-scale features of the hydrological process in mountainous watersheds and provides reliable feature support for ecological flow estimation.

[0134] According to one aspect of the present application, step S31 is specifically as follows:

[0135] Step S311: Receive the multi-scale feature matrix, extract the water volume change sequence to obtain a water volume sequence; calculate the ecological sensitive period characteristics to obtain a sensitive period matrix; analyze the ecological water demand to obtain a demand matrix; calculate the response intensity based on the water volume sequence and the sensitive period matrix to obtain an intensity matrix; combine the demand matrix and the intensity matrix to construct a response feature to generate a response feature set.

[0136] Step S312: Receive the multi-scale feature matrix, extract the river network structure data to obtain a river network matrix; calculate the physical connectivity between river reaches to obtain a physical connectivity matrix; analyze the hydrological connectivity status to obtain a hydrological connectivity matrix; evaluate the ecological corridor function to obtain a corridor index matrix; integrate the physical connectivity matrix, the hydrological connectivity matrix and the corridor index matrix to generate a connectivity matrix C 。

[0137] Step S313: Receive the response feature set and the connectivity matrix, construct a feature weight function to obtain a weight function; calculate the correlation between features to obtain a correlation matrix; construct a distance function based on the weight function and the correlation matrix to obtain a distance function; design a similarity calculation criterion to obtain a calculation criterion; generate a similarity index S according to the distance function and the calculation criterion.

[0138] Step S314: Receive the multi-scale feature matrix, determine the time window sequence to obtain the window sequence; construct the time-scale decomposition function to obtain the decomposition function; perform sequence decomposition based on the window sequence and the decomposition function to obtain the decomposition result; reorganize and optimize the decomposition result to generate the time series set.

[0139] Step S315: Receive the similarity metric and the time series set, construct the pattern matching criterion to obtain the matching criterion; calculate the dynamic similarity between sequences to obtain the dynamic similarity matrix; identify the key matching points to obtain the matching point set; perform pattern recognition based on the matching criterion and the matching point set to generate the traffic pattern set.

[0140] Through the ecological hydrological sequence matching and response feature analysis method, the accurate identification of the ecological flow pattern in mountainous watersheds is realized. This method first analyzes the ecological response characteristics based on the water volume change and the characteristics of the ecological sensitive period, establishes the hydrological connectivity relationship through the assessment of river network connectivity and ecological corridor functions, and constructs a pattern recognition framework considering ecological needs. Especially in the process of similarity calculation, by combining feature weights and correlation analysis, the dynamic similarity relationship between flow sequences is accurately characterized. At the same time, through time-scale decomposition and pattern matching, the accurate identification of flow patterns is achieved, which not only ensures the reliability of the identification results but also improves the accuracy of pattern expression. This pattern recognition method based on ecological response fully considers the ecological characteristics of mountainous watersheds and provides a scientific pattern support for ecological flow estimation.

[0141] According to one aspect of the present application, step S32 is specifically as follows:

[0142] Step S321: Receive the traffic pattern set, extract the key nodes in the process to obtain the node set; analyze the connection relationship between nodes to obtain the connection matrix; construct the process transfer function to obtain the transfer function; establish the network topology based on the node set and the connection matrix to obtain the topology matrix; construct the transfer chain network based on the transfer function and the topology matrix to generate the network structure.

[0143] Step S322: Receive the traffic pattern set and the network structure, design the perturbation experiment scheme to obtain the perturbation scheme; perform the parameter perturbation test to obtain the test result; calculate the response sensitivity to obtain the sensitivity matrix; analyze the parameter importance to obtain the importance vector; integrate the sensitivity matrix and the importance vector to generate the sensitivity matrix.

[0144] Step S323: Receive the network structure and the sensitivity matrix, construct the transfer path identification criterion to obtain the identification criterion; analyze the uncertainty propagation characteristics to obtain the propagation characteristic matrix; identify the key transfer paths to obtain the key path set; evaluate the path importance to obtain the importance matrix; integrate the key path set and the importance matrix to generate the transfer path set.

[0145] Step S324: Receive the transfer path set, identify the sources of uncertainty to obtain the source matrix; construct the coupling effect evaluation function to obtain the evaluation function; calculate the uncertainty interaction intensity to obtain the interaction matrix; analyze the coupling action mode to obtain the action mode matrix; generate the coupling coefficient set based on the interaction matrix and the action mode matrix C 。

[0146] Step S325: Receive the transfer path set and the coupling coefficient set, construct the comprehensive evaluation function to obtain the evaluation function; calculate the uncertainty contribution of each path to obtain the contribution matrix; evaluate the overall uncertainty level to obtain the level index; analyze the spatio-temporal distribution of uncertainty to obtain the distribution matrix; integrate the level index and the distribution matrix to generate the uncertainty index set.

[0147] Through the uncertainty transfer analysis and coupling effect evaluation method, the systematic quantification of the uncertainty of ecological flow estimation in mountainous watersheds is realized. This method first constructs a transfer chain network based on the key nodes of the process, identifies the key influencing factors through parameter perturbation tests and sensitivity analysis, and establishes an uncertainty analysis framework suitable for the characteristics of mountainous watersheds. Especially when analyzing the characteristics of uncertainty propagation, by identifying the key transfer paths and evaluating the importance of the paths, the transfer law of uncertainty is accurately characterized. At the same time, by analyzing the coupling effect of multi-source uncertainty and constructing a comprehensive evaluation function, the systematic quantification of uncertainty is realized, which not only ensures the comprehensiveness of the evaluation results but also improves the accuracy of uncertainty representation. This uncertainty analysis method can effectively evaluate the reliability of ecological flow estimation in mountainous watersheds and provide important support for decision-making.

[0148] According to one aspect of the present application, step S41 is specifically as follows:

[0149] Step S411: Receive the flow pattern set and the uncertainty index set, extract the terrain feature parameters to obtain the terrain feature matrix; calculate the hydrological response index to obtain the response index matrix; analyze the land use characteristics to obtain the land feature matrix; evaluate the climate variability to obtain the climate index matrix; integrate the terrain feature matrix, the response index matrix, the land feature matrix and the climate index matrix to generate the feature index set.

[0150] Step S412: Receive the feature index set, construct the model adaptability evaluation function to obtain the evaluation function; calculate the feature index weights to obtain the index weight matrix; design the model selection threshold to obtain the threshold vector; construct the selection rule according to the evaluation function and the index weight matrix to obtain the selection rule; generate the selection criterion based on the selection rule and the threshold vector.

[0151] Step S413: Receive the feature index set, divide the spatio-temporal calculation units to obtain a set of calculation units; analyze the spatial heterogeneity features to obtain a spatial heterogeneity matrix; calculate the time variability index to obtain a time variability matrix; evaluate the scale dependence to obtain a scale index matrix; integrate the spatial heterogeneity matrix, the time variability matrix and the scale index matrix to generate a heterogeneity matrix.

[0152] Step S414: Receive the selection criterion and the heterogeneity matrix, construct a weight allocation function to obtain an allocation function; calculate the model adaptability score to obtain a score matrix; optimize the weight allocation scheme to obtain an optimized scheme; calculate the weights according to the allocation function DF and the optimized scheme to generate a weight vector.

[0153] Step S415: Receive the flow pattern set, design a verification test scheme to obtain a test scheme; perform model verification calculations to obtain verification results; evaluate the prediction accuracy to obtain an accuracy matrix; analyze the stability index to obtain a stability index; integrate the accuracy matrix and the stability index to generate a reliability matrix.

[0154] Step S416: Receive the weight vector and the reliability matrix, construct a model integration framework to obtain an integration framework; design a model combination strategy to obtain a combination strategy; optimize the integration parameters to obtain a parameter matrix; construct an integrated model based on the integration framework and the combination strategy to generate an integrated model.

[0155] Through the adaptive dynamic integration and model selection method, the optimized integration of the ecological flow estimation model for mountainous watersheds is realized. This method first constructs a feature index set based on factors such as topographic features, hydrological responses, land use, and climate variability, and establishes an integration framework through model adaptability evaluation and weight optimization to construct a model integration system suitable for the characteristics of mountainous watersheds. Especially when considering spatio-temporal heterogeneity, by analyzing spatial heterogeneity, time variability, and scale dependence, the applicability characteristics of the model are accurately characterized. At the same time, through model verification and reliability evaluation, the optimized adjustment of the integrated model is realized, which not only ensures the prediction accuracy of the model but also improves the reliability of the estimation results. This adaptive integration method can make full use of the advantages of multiple models and provide a reliable model support for the ecological flow estimation of mountainous watersheds.

[0156] According to one aspect of the present application, Step S42 is specifically as follows:

[0157] Step S421: Receive the integrated model, extract the ecological demand characteristics to obtain a demand characteristic matrix; construct a water volume constraint condition to obtain a water volume constraint matrix; design an ecological protection target to obtain a protection target matrix; analyze the system carrying capacity to obtain a carrying capacity index; integrate the water volume constraint matrix, the protection target matrix and the carrying capacity index to generate a set of constraint conditions.

[0158] Step S422: Receive the set of constraint conditions, construct a seasonal variation function to obtain a variation function; design an objective optimization criterion to obtain an optimization criterion; establish a multi-objective trade-off function to obtain a trade-off function; construct an objective function based on the variation function and the trade-off function to generate an objective function.

[0159] Step S423: Receive the set of constraint conditions and the objective function, design the parameters of the optimization algorithm to obtain a parameter set; perform iterative optimization calculations to obtain an iterative result; analyze the convergence characteristics to obtain a convergence index; screen the optimal solution according to the convergence index to generate an initial result.

[0160] Step S424: Receive the initial result, construct an ecological response evaluation function to obtain an evaluation function; calculate the change in ecological indicators to obtain a change matrix; analyze the response sensitivity to obtain a sensitivity index; evaluate the response characteristics to obtain a characteristic matrix; integrate the sensitivity index and the characteristic matrix to generate a response index.

[0161] Step S425: Receive the initial result and the response index, construct a correction function model to obtain a correction function; calculate the adjustment parameters to obtain an adjustment matrix; perform result correction to obtain a corrected result; evaluate the correction effect to obtain an effect index; determine the final result according to the effect index to generate an ecological flow estimation result.

[0162] Step S426: Receive the ecological flow estimation result, construct an evaluation index system to obtain an index system; calculate the reliability index to obtain a reliability matrix; analyze the uncertainty level to obtain an uncertainty index; evaluate the adaptability characteristics to obtain an adaptability matrix; integrate the reliability matrix, the uncertainty index and the adaptability matrix to generate a reliability evaluation index.

[0163] Through the multi-constraint optimization and ecological response evaluation method, the dynamic optimization of the ecological flow estimation results in mountainous river basins is realized. This method first constructs constraint conditions based on the ecological demand characteristics, establishes an optimization framework through the seasonal variation function and multi-objective trade-off, and constructs an optimization system considering the needs of the ecosystem. Especially in the optimization process, through iterative calculations and convergence analysis, the optimal solution is accurately identified, and the results are corrected through ecological response evaluation. At the same time, through constructing an evaluation index system and analyzing the reliability characteristics, the systematic evaluation of the estimation results is realized, which not only ensures the scientific nature of the results, but also improves the comprehensiveness of the evaluation. This multi-constraint optimization method can effectively balance ecological demands and system constraints, providing a scientific basis for the ecological flow management in mountainous river basins.

[0164] Case 1:

[0165] Step S1: Preprocessing and feature extraction of multi-source heterogeneous data

[0166] The collected data includes: DEM data, 30m resolution raster data, containing elevation values H(i, j); remote sensing image data, 10m resolution Sentinel-2 images, containing multi-spectral band values B(i, j, k); meteorological reanalysis data, containing daily data such as rainfall P(t) and temperature T(t); land use data, 30m resolution raster data, containing land use type codes L(i, j).

[0167] Step S11: Specific process of data quality assessment

[0168] For each grid point (i, j), extract the elevation values within a 3×3 window; calculate the terrain complexity index T C (i, j): T C (i, j) = √[(∑(H(i, j) - H_mean) 2 ) / n]; where H_mean is the average elevation within the window and n is the number of grid cells within the window; generate the terrain decomposition factor T in the form of an n×m matrix

[0169] Divide the study area into k×k sub-regions; calculate the local statistical characteristics for each sub-region: mean μl = ∑X(i, j) / n; standard deviation σl = √[∑(X(i, j) - μl) 2 / n]; calculate the global statistical characteristics: mean μg = ∑X(i, j) / N; standard deviation σg = √[∑(X(i, j) - μg) 2 / N]; generate the decomposition coefficient matrix

[0170] Hydrological similarity calculation, including: defining the feature vector F(i, j) = [H(i, j), S(i, j), A(i, j)]; where S is the slope and A is the catchment area; calculating the similarity index SI(i, j) = exp(-||F(i, j) - F(i±1, j±1)|| / σ); generating the similarity index

[0171] Identifying outliers, including: setting the threshold θ = μ ± 2σ; marking the outlier positions: if |X(i, j) - μ| > θ, then M(i, j) = 1 otherwise M(i, j) = 0; generating the data quality index

[0172] Filter the data according to the data quality index: if Q(i, j) > Qthreshold, retain the data. Otherwise mark it as an invalid value; generate the valid data set

[0173] Step S12: Specific process of hydrological process feature extraction

[0174] Calculate the slope

[0175] S(i, j) = arctan[(H(i + 1, j) - H(i - 1, j))2 / 2dx + (H(i, j + 1) - H(i, j - 1)) 2 / 2dy];

[0176] Calculate the aspect A(i, j) = arctan[(H(i, j + 1) - H(i, j - 1)) / (H(i + 1, j) - H(i - 1, j))];

[0177] Obtain the geomorphic unit characteristics and generate the geomorphic feature vector;

[0178] Basis function construction, including: constructing the spatial distribution function: φ(x, y) = exp(-[(x - x o ) 2 + (y - y o ) 2 / 2σ 2 ); Calculate the length of the flow convergence path: L = ∑√[(xi + 1 - xi) 2 + (yi + 1 - yi) 2 ; Generate the basis function set B;

[0179] Time-delay response analysis, including: calculating the cross-correlation coefficient R(τ) = ∑[P(t) * Q(t + τ)] / [σp * σq];

[0180] Identify the maximum correlation time-delay τmax = argmax(R(τ)); Generate the time-delay coefficient matrix;

[0181] Feature projection: construct the projection matrix: W = B * B T ; Calculate the transformation coefficient: T C = W * V; Generate the feature component set;

[0182] Multi-scale combination: set the scale sequence s = [s1, s2,..., sn]; Calculate the energy at each scale E(s) = ∑|F(s)|2; Generate the hydrological feature set and the process feature set;

[0183] Among them, the data quality index is an n×m matrix with a value range of [0, 1]; The effective data set is a multi-dimensional data set containing data passing the quality inspection; The geomorphic feature vector is a triple of [slope, aspect, elevation] for each grid point; The basis function set B is a set of orthogonal basis functions {φ i (x, y)}; The time-delay coefficient matrix is a t×t cross-correlation coefficient matrix; The feature component set F is a set of k-dimensional feature vectors; The hydrological feature set and the process feature set: multi-dimensional feature matrix.

[0184] Step S2: Dynamic zoning and scale identification of mountainous watersheds

[0185] Hydrological feature set HF: a three-dimensional matrix of n×m×p, containing p hydrological feature layers; Process feature set PF: a t×q matrix, containing q process features at t time points.

[0186] Step S21: Specific process of the dynamic partitioning algorithm

[0187] Calculate the terrain gradient, including: calculating the x-direction gradient Gx(i, j) = [H(i + 1, j) - H(i - 1, j)] / 2Δx; calculating the y-direction gradient Gy(i, j) = [H(i, j + 1) - H(i, j - 1)] / 2Δy; synthesizing the total gradient G(i, j) = √[Gx(i, j) 2 + Gy(i, j) 2 ; Generate the gradient matrix G.

[0188] Construct the spatial neighborhood relationship: Define the search radius r; calculate the spatial distance matrix D(i, j, k, l) = √[(xi - xk) 2 +(yj - yl) 2 ; Construct the neighborhood connection strength C(i, j, k, l) = exp(-D(i, j, k, l) / r) • G(i, j) / G(k, l); Generate the neighborhood matrix N.

[0189] Calculation of the confluence time: Calculate the flow velocity V(i, j) = α • S(i, j) β • A(i, j) γ ; Calculate the time to the outlet T(i, j) = ∑[L(i, j) / V(i, j)]; Generate the time matrix T.

[0190] Step S22: Specific process of multi-scale decomposition

[0191] Construction of the characteristic function: Construct the terrain modulation function M(x, y) = exp(-κ • G(x, y)); Define the scale function: ψ(x, y, s) = M(x, y) • exp[-(x 2 + y 2 ) / 2s 2 ; Generate the characteristic function set F.

[0192] 2. Multi-dimensional decomposition: Set the scale sequence S = {s1, s2,..., s k}; Calculate the decomposition coefficient C (i, j, k) = JJf(x, y) • ψ(x, y, s k )dxdy; Generate the decomposition coefficient set S; JJ is the double integral.

[0193] Step S3: Identification of ecological flow patterns and uncertainty analysis

[0194] Read the multi-scale feature matrix M, an n×m×k matrix containing features of k scale levels;

[0195] Step S31: Specific process of ecological hydrological sequence matching

[0196] Ecological response feature extraction: Calculate the flow change rate ΔQ(t) = [Q(t) - Q(t - 1)] / Q(t - 1); Calculate the ecological sensitivity ES(t) = ΔQ(t) / ΔE(t), where ΔE(t) is the environmental variable change rate; Generate the response feature set R.

[0197] River system connectivity analysis, including: Construct the river reach connection matrix cN(i, j) = {1, if river reaches i and j are directly connected {0, otherwise; Calculate the connectivity index cI(i) = ∑[cN(i, j) • W(j)], where W(j) is the weight of river reach j; Generate the connectivity matrix;

[0198] Similarity calculation, including: Construct the feature distance function d(x, y) = √[∑(w i •(x i -y i ) 2 )], where w i is the feature weight; Calculate the temporal similarity S(i, j) = exp[-d(R i , R j ) / σ]; Generate the similarity index S;

[0199] Step S32: Specific process of uncertainty propagation analysis;

[0200] Propagation network construction: Define the node set V = {v i} for i = 1, 2,..., n; Construct the adjacency matrix A(i, j) = {w(i, j), if nodes i and j are connected {0, otherwise; Generate the network structure N;

[0201] Sensitivity analysis, including: Calculate the local sensitivity LSI(i) = dY / dX i ; Calculate the overall sensitivity TSI(i) = Var(E[Y|X i ) / Var(Y); Generate the sensitivity matrix S; d is the partial derivative symbol;

[0202] Step S4: Intelligent estimation of ecological flow

[0203] Step S41: Specific process of adaptive dynamic integration

[0204] Model weight calculation, including: Calculate the model error E(i, t) = [Q * i (t) - Q(t)] 2; Calculate the dynamic weight W(i, t) = exp[-E(i, t) / λ] / ∑exp[-E(j, t) / λ]; Generate the weight vector W;

[0205] Reliability assessment, including: calculating the prediction interval PI(t) = [Q*(t) ± z•σ(t)]; calculating the confidence level CR(t) = P[Q(t) ∈ PI(t)]; generating the reliability matrix;

[0206] Step S42: The specific process of multi-constraint optimization

[0207] Constraint condition construction, including: minimum ecological demand constraint Q(t) ≥ Q e (t); maximum available water volume constraint Q(t) ≤ Qm(t); generating the constraint condition set.

[0208] Objective function construction, including: ecological objective f1(Q) = ∑[w1*(Q(t) - Q e (t)) 2 ; water use objective f2(Q) = ∑[w2*(Q(t) - Q u (t)) 2 ; generating the objective function F;

[0209] Case 2: In some embodiments, the calculation process of the sub-steps can also be:

[0210] Step S121: Geomorphic unit feature extraction: Read the effective data set V: a multi-dimensional data set containing n×m grids;

[0211] Slope calculation, including: x-direction slope Sx(i, j) = [H(i + 1, j) - H(i - 1, j)] / 2Δx; y-direction slope Sy(i, j) = [H(i, j + 1) - H(i, j - 1)] / 2Δy; combined slope: S(i, j) = √[Sx(i, j)2 + Sy(i, j)2]; generating the slope matrix S1;

[0212] Aspect calculation, including: aspect angle A(i, j) = arctan[Sy(i, j) / Sx(i, j)]; direction adjustment A c (i, j) = {A(i, j), Sx(i, j) > 0{A(i, j) + π, Sx(i, j) < 0; generating the aspect matrix S2;

[0213] Elevation extraction, including: neighborhood averaging, Hm(i, j) = ∑H(i ± 1, j ± 1) / 9; anomaly detection, if |H(i, j) - Hm(i, j)| > θh: H(i, j) = Hm(i, j); generating the elevation matrix S3;

[0214] Δx and Δy are grid spacings, A(i, j) is the original slope aspect angle at position (i, j), and A c (i, j) is the corrected slope aspect angle at position (i, j) in radians. Hm(i, j) is the neighborhood average elevation at position (i, j), and θh is the elevation anomaly test threshold.

[0215] Step S122: Construction of the spatial feature function. Read the geomorphic feature vector G as an n×m×3 matrix, which contains three feature layers: slope, slope aspect, and elevation.

[0216] Process of constructing the spatial distribution function: Based on position (i, j), establish a local coordinate system and calculate the spatial distribution function SP(x, y) = exp[-(x - xi) 2 / 2σx 2 -(y - yi) 2 / 2σy 2 ·G(i, j), where σx and σy are feature diffusion parameters, generating a set of spatial functions. Use the D8 algorithm to trace the flow path from position (i, j) to the outlet, and calculate the path length as L(i, j) = ∑√[(xk + 1 - xk)2+(yk + 1 - yk)2], where (xk, yk) are the coordinates of the sequence points on the path, generating a path length matrix D. Combine the spatial distribution function and the path length to construct the basis function B(x, y) = SP(x, y)·exp[-L(x, y) / L o , where L o is the feature length parameter, generating a set of basis functions. Extract the rainfall sequence PR(t) and runoff sequence QR(t) from the valid dataset, where t = 1, 2,..., T, respectively generating the rainfall sequence and the runoff sequence. Calculate the rainfall-runoff cross-correlation coefficient C(τ) = ∑[PR(t)·QR(t + τ)] / [σp·σq], where τ is the time lag, generating a correlation coefficient matrix. Identify the time lag τmax = argmax(C(τ)) corresponding to the maximum correlation coefficient, and calculate the response time of each sub-basin, generating a time lag coefficient matrix.

[0217] SP(x, y) is the value of the spatial distribution function at position (x, y), σx and σy are the feature diffusion parameters of the spatial distribution function, L(i, j) is the length of the confluence path from position (i, j) to the outlet, and L o is the basis function feature length parameter, PR(t) is the rainfall intensity at time t, QR(t) is the runoff at time t, and C(τ) is the cross-correlation coefficient at time lag τ.

[0218] Step S123: Feature transformation and adjustment

[0219] Basis function set B: a p×q matrix containing q basis functions; time-delay coefficient matrix R: an n×m time-delay distribution matrix; effective data set V: an n×m×k multi-dimensional data matrix; according to the orthonormalization of the basis functions, the projection operator W(i, j)=∑[B(i, k)·B(j, k)] / λk, where λk is the eigenvalue of the k-th basis function, to generate the projection matrix W. Perform the transformation operation T on the effective data C (i, j)=∑[W(i, k)·V(k, j)], where i = 1, 2,..., n, j = 1, 2,..., m, to generate the transformation coefficient set T C . Adjust A according to the time-delay coefficient C (i, j)=T C (i, j)*exp[-R(i, j) / τ o , where τ o is the characteristic time parameter, to generate the adjustment coefficient set. Perform eigen-decomposition on the adjustment coefficient F(i)=∑[A C (i, j)*ek(j)], where ek is the eigenvector, and take the first k principal components to generate the characteristic component set.

[0220] Step S124: Multi-scale decomposition and recombination

[0221] Characteristic component set, an n×k matrix containing k principal components; set the multi-scale window length sequence K(i)={2 i |i = 1, 2,..., m}, where m is the maximum decomposition level, to generate the window sequence K. Perform wavelet decomposition D on each scale C (i, j, k)=∑[F(t)·ψ((t - iτ) / s j )] / √s j , where ψ is the wavelet basis function, s j is the scale parameter, to generate the decomposition coefficient set. Calculate the energy distribution E(j)=∑|D C (i, j, k)| 2 , and after normalization, generate the energy distribution matrix. Determine the important scale TV(j)=min{k|∑E(i) / ∑E(t)>η, i = 1,..., k}, where η is the energy threshold, to generate the threshold vector. Select the important scale for reconstruction HF(t)=∑[D C (i, j, k)·ψ((t - iτ) / s j )], to generate the hydrological feature set. Calculate the characteristic parameter PF(i)=Φ[HF(t)] based on the reconstruction result, where Φ is the feature extraction operator, to generate the process feature set.

[0222] In another embodiment of the present application, Step S211: Topographic gradient calculation

[0223] The hydrological feature set HF is an n×m×p matrix, containing p hydrological feature layers; the process feature set PF is a t×q matrix, containing q process features.

[0224] The second-order central difference scheme is used to calculate the gradient in the x direction Gx(i, j) = [H(i + 1, j) - H(i - 1, j)] / 2Δx and the gradient in the y direction Gy(i, j) = [H(i, j + 1) - H(i, j - 1)] / 2Δy, where H is the elevation value, and Δx, Δy are the grid spacings, to generate the elevation difference matrix H. Based on the directional gradients, the composite slope S(i, j) = arctan(√[Gx(i, j)^2 + Gy(i, j)^2]) is calculated to generate the slope matrix. Combining the elevation gradient and the slope, the comprehensive gradient G(i, j) = √[Gx(i, j) 2 + Gy(i, j) 2 · [1 + w · S(i, j)] is calculated, where w is the slope weight coefficient, to generate the gradient matrix G.

[0225] Step S212: Spatial neighborhood construction

[0226] Read the gradient matrix, an n×m gradient distribution matrix

[0227] Set the search radius r, and for each grid (i, j), identify the set of neighborhood cells that satisfy the distance condition d = √[(x - xi) 2 +(y - yj) 2 ≤ r to generate the neighborhood set N C . Calculate the connection strength K(i, j, k, l) = exp(-d / r o ) · exp(-|G(i, j) - G(k, l)| / g o ), where r o is the distance feature parameter, and g o is the gradient feature parameter, to generate the connection strength matrix K. Set the threshold θ c according to the connection strength. When K(i, j, k, l) > θ c , establish the neighborhood connection relationship to generate the neighborhood matrix.

[0228] Step S213: Confluence time calculation

[0229] Read the neighborhood matrix: an n×m×k×l neighborhood relationship matrix; read the gradient matrix: an n×m gradient distribution matrix.

[0230] Identify the river network cells based on the cumulative flow threshold θf, extract the river network orientation and connection relationship, construct the river network topological structure RN(i, j) = {1: river network, 0: non-river network} to generate the river network matrix RN. Use the Manning formula to calculate the channel velocity V(i, j) = (1 / n) · R(i, j)2 / 3 ·S(i, j) 1 / 2 , where n is the roughness coefficient, R is the hydraulic radius, and S is the hydraulic gradient, to generate the velocity matrix V. Based on the D8 algorithm, trace the flow path from each cell to the outlet, and calculate the path length and the cumulative time P(i, j) = {(x1, y1, t1),..., (x n , y n , t n )}, to generate the path matrix P. Cumulate the transmission time of each path segment T(i, j) = ∑[L(k) / V(k)], where L(k) is the length of the k-th path segment, to generate the time matrix.

[0231] Step S214: Construction of dynamic partitioning rules

[0232] Read the neighborhood matrix: the neighborhood relationship matrix; read the time matrix T: the flow concentration time distribution matrix.

[0233] Design the spatial constraint function F(i, j) = α·N(i, j) + β·exp[-|T(i) - T(j)| / τ c , where α and β are weight coefficients, and τ c is the time characteristic parameter, to generate the constraint function.

[0234] Calculate the similarity between cells SM(i, j) = F(i, j)·exp[-∑wk(xki - xkj)2] based on topographic features and hydrological responses, where wk is the feature weight and xk is the feature value, to generate the similarity matrix SM.

[0235] Use the hierarchical clustering method to determine the number of partitions and the similarity thresholds TH = {θ1, θ2,..., θn}, to generate the threshold vector TH.

[0236] Construct the partitioning criterion c(i, j) = {1: SM(i, j) > θk, 0: otherwise} by integrating spatial constraints and similarity thresholds, to generate the partitioning criterion. Where α and β are the weight coefficients of the constraint function, τ c is the time characteristic parameter, unit: hour, wk is the weight of the k-th feature, and θk is the similarity threshold of the k-th partition.

[0237] Step S215: Partition optimization and correction

[0238] Partition identification initialization process: Randomly assign an initial identification ID(i, j)=k to each cell, where k∈[1, K], K is the preset number of partitions, and generate an identification matrix. Update the identification based on the objective function J = ∑[w1·D(i)+w2·c(i, j)], where D(i) is the intra-group distance, and w1, w2 are weight coefficients, to obtain oM(i, j)=argmin(J), and generate an optimization matrix oM. Calculate the spatial continuity index of each partition C I(k)=N c (k) / N(k), where N c (k) is the number of connected cells, N(k) is the total number of cells, and generate a continuity index cI. Merge or split the regions that do not meet the continuity requirements, I(i, j)=f[oM (i, j), cI(k)], and generate an initial partition result.

[0239] Step S216: Seasonal dynamic adjustment

[0240] Read the initial partition result, and calculate the hydrological characteristics of each partition in different seasons SF(k, s)=∑[w(i)·X(i, s)], where w(i) is the feature weight, s is the season index, and generate a season feature matrix SF.

[0241] Design an adjustment function based on seasonal changes AF(t)=a o +∑[a i ·cos(2πit / T)+b i ·sin(2πit / T)], where T is the period, and generate an adjustment function AF. Calculate the dynamic change range of the partition boundary according to the adjustment function CR(k, t)=r o ·[1+λ·AF(t)], where r o is the reference range, λ is the adjustment coefficient, and generate a change range matrix CR. Combine the initial partition and the boundary change for dynamic adjustment D(i, j, t)=f[I(i, j), cR(k, t)], and generate a dynamic response unit.

[0242] Step S221: Feature function construction. Read the dynamic response unit.

[0243] Terrain feature extraction process: Calculate the morphological features of each cell TF(i)={H(i), S(i), A(i), TWI(i)}, where H is the elevation, S is the slope, A is the area, and TWI is the topographic wetness index, and generate a terrain feature set TF.

[0244] Calculate the response features RP(i)={t c (i), K c (i), η(i)} based on rainfall-runoff data, where t c is the confluence time, Kc is the runoff coefficient, η is the runoff efficiency, and the response parameter set RP is generated.

[0245] The feature expression EM(i, j) = φ(TF(i))·ψ(RP(i)) is constructed through orthogonalization processing, where φ and ψ are feature mapping functions, and the expression matrix is generated.

[0246] Based on the expression, the feature function F(x, y) = ∑[EM(i, j)·B(x, y)] is constructed, where B is the basis function, and the feature function set F is generated.

[0247] Step S222: Multidimensional decomposition implementation

[0248] Read the feature function set. Design the multidimensional orthogonal basis function BM(x, y, z) = ∑[αk·Pk(x)·Qk(y)·Rk(z)], where P, Q, and R are one-dimensional orthogonal polynomials, and the basis function matrix BM is generated.

[0249] Construct the geometric scale sequence SS(k) = s o ·q k , k = 0, 1,..., N, where s o is the initial scale, q is the scale ratio, and the scale sequence is generated. Perform multi-scale projection on the input function ID(i, j, k) = ∬F(x, y)·BM(x, y, k)dxdy to generate the initial decomposition set. Use the L1 norm constraint for sparse optimization S(i, j, k) = argmin{||F - BM·S|| 2 2 + μ||S||1}, and generate the decomposition coefficient set S.

[0250] Step S223: Energy distribution calculation

[0251] Read the decomposition coefficient set S: Calculate the energy EV(k) = ∑|S(i, j, k)| 2 / N for the decomposition coefficients at each scale, where N is the total number of samples, and generate the energy value matrix. Use the kernel density estimation method to construct the energy distribution function DF(k) = (1 / nh)∑K[(x - xi) / h], where h is the bandwidth parameter and K is the kernel function, and generate the distribution function. Calculate the relative energy contribution E(k) = EV(k) / ∑EV(i) at each scale and perform cumulative energy analysis to generate the energy matrix.

[0252] Step S224: Dominant mode recognition

[0253] Read the energy matrix E, set the energy threshold and the spatial correlation threshold IR = {θe, θr}, where θe is the energy contribution rate threshold and θr is the correlation coefficient threshold, and generate the recognition criterion IR. Based on the recognition criterion, filter the dominant modes FM(k) = {M(i)|E(i)>θe, R(i)>θr}, where R(i) is the spatial correlation coefficient, and generate the feature matrix FM. Calculate the spatio-temporal distribution characteristics of the dominant modes DM(i, j, k) = φ[FM(k)]·ψ(i, j), where φ and ψ are mapping functions, and generate the distribution matrix DM. Integrate the feature and distribution information P(i, j, k) = γ·FM(k)+(1 - γ)·DM(i, j, k), where γ is the weight coefficient, and generate the mode matrix P.

[0254] Step S225: Scale information flow analysis

[0255] Read the mode matrix P, design the scale conversion function TF(s1, s2) = ∫ψ(s1t)·ψ*(s2t)dt based on wavelet transform, where ψ is the wavelet function, and generate the conversion function TF. Calculate the information transfer amount between adjacent scales TM(i, j) = ∑P(i, k)·TF(sk, sj)·P(j, k), where sk, sj are scale parameters, and generate the transfer matrix TM. Evaluate the information flow intensity SI(k) = -∑TM(i, k)·l o g[TM(i, k)], and generate the intensity index SI. Synthesize the transfer matrix and the intensity index L(i, j) = TM(i, j)·[SI(i)+SI(j)] / 2, and generate the flow index L.

[0256] Step S226: Feature recombination implementation

[0257] Read the mode matrix P and the flow index L. Calculate the recombination weight W(k) = L(k)·E(k) / ∑[L(i)·E(i)] based on the information flow and energy distribution, and generate the weight vector W. Perform the feature recombination operation RM(i, j) = ∑[W(k)·P(i, j, k)], where k is the mode number, and generate the recombination matrix RM. Calculate the information retention rate RI(k) = ||RM(k)-P(k)|| / ||P(k)|| of the recombination result, and generate the representative index RI. Adjust the weight according to the representative index and recalculate M(i, j) = f[RM(i, j), RI(k)], and generate the multi-scale feature matrix M.

[0258] Step S311: Ecological response feature extraction

[0259] Read the multi-scale feature matrix M, calculate the flow change rate WS(t) = [Q(t) - Q(t - 1)] / Q(t - 1) for adjacent time periods, where Q(t) is the flow at time t, and generate the water volume sequence WS. Construct the sensitivity function SP(t) = α·B(t) + β·E(t) based on the ecological water demand characteristics, where B(t) is the basic water demand and E(t) is the ecological water demand, and generate the sensitive period matrix SP. Calculate the ecological water use demand DM(t) = max[Qmin(t), Qe co (t)], where Qmin is the minimum water demand and Qe co is the ecological water demand, and generate the demand matrix DM. Synthesize the water volume change and ecological demand IM(t) = WS(t)·SP(t) / DM(t), and generate the intensity matrix IM. Integrate each index R(t) = {WS(t), SP(t), DM(t), IM(t)}, and generate the response feature set R.

[0260] Step S312: River network connectivity analysis

[0261] Read the multi-scale feature matrix M. Use the threshold method to extract the river network structure RN(i, j) = H[F(i, j) - θf], where H is the step function and θf is the extraction threshold, and generate the river network matrix RN. Calculate the connectivity P C (i, j) = d(i, j)·exp[-L(i, j) / L o , where d is the connectivity indicator function and L is the river reach length, and generate the physical connection matrix P C . Analyze the flow transmission characteristics H C (i, j) = Q(i, j) / [Q(i) + Q(j)], where Q is the cross-section flow, and generate the hydrological connection matrix H C . Evaluate the ecological corridor function of the river reach C I(i) = w1·P C (i) + w2·H C (i), and generate the corridor index matrix C I. Integrate each index C (i, j) = f[P C (i, j), H C (i, j), CI(i)], and generate the connectivity matrix C .

[0262] Step S313: Similarity metric construction

[0263] Read the response feature set R, the connectivity matrix C , design the feature weight function WF(i) = exp[-λ·V(i)], where V(i) is the feature coefficient of variation and λ is the attenuation parameter, and generate the weight function WF. Calculate the correlation coefficient between featuresC M(i, j) = cov(Xi, Xj) / [σ(Xi)·σ(Xj)], where cov is the covariance and σ is the standard deviation, to generate the correlation matrix CM. Design the distance function DF(x, y) = √[(x - y) T Σ -1 (x - y)], where Σ is the covariance matrix, to generate the distance function. Calculate the similarity S(i, j) = WF(i)·exp[-DF(xi, xj) / h], where h is the bandwidth parameter, to generate the similarity index by combining the weight and the distance function.

[0264] Step S314: Time series decomposition

[0265] Read the multi-scale feature matrix and set the time window sequence WS(k) = {w1, w2,..., w k}, where w i is the window length, to generate the window sequence. Construct the decomposition function DF(t, s) = ∫x(τ)ψ[(t - τ) / s]dτ based on wavelet transform, where ψ is the wavelet basis function, to generate the decomposition function.

[0266] Perform multi-scale decomposition DR(t, s) = DF(t, s)·M(t), where s is the scale parameter, to generate the decomposition result.

[0267] Optimize the recombination parameter according to the energy distribution to obtain T(t) = ∑[α(s)·DR(t, s)], to generate the time series set.

[0268] Step S315: Pattern matching implementation

[0269] Read the similarity index S and the time series set T. Design the dynamic matching criterion M C (i, j) = S(i, j)·exp[-|t(i) - t(j)| / τ], where τ is the time feature parameter, to generate the matching criterion.

[0270] Use the dynamic programming algorithm to calculate the sequence similarity DS(i, j) = min{d(i, j) + min[DS(i - 1, j), DS(i, j - 1), DS(i - 1, j - 1)]}, where d is the local distance, to generate the dynamic similarity matrix.

[0271] Extract the feature matching points MP(k) = {(i, j)|DS(i, j) < θd, path(i, j) = 1} based on the optimal path, where θd is the distance threshold, to generate the matching point set. Integrate the matching results to generate the traffic pattern P(t) = f[MP(k), DS(i, j)], to generate the traffic pattern set.

[0272] Step S321: Transfer network construction

[0273] Read the traffic pattern set P, identify the key process nodes NP(i) = {n | P(n) > θp}, where θp is the node threshold, and generate a node set. Construct the connection matrix CM(i, j) = G(Xi → Xj) based on causal test, where G is the Granger causality test statistic, and generate the connection matrix. Design the traffic transfer function TF(i, j) = α·exp[-β·t(i, j)], where t is the transfer time, and generate the transfer function. Integrate the node and connection information N(i, j) = CM(i, j)·TF(i, j) to generate the network structure.

[0274] Step S322: Implementation of sensitivity analysis

[0275] Read the traffic pattern set P and the network structure N. Construct the parameter perturbation sequence DP(i) = P o (i)[1 + δ·ξ(i)], where δ is the perturbation intensity and ξ is the random perturbation, and generate the perturbation scheme DP. Calculate the response result TR(i, j) = M[DP(i)] for each group of perturbations, where M is the model response function, and generate the test result TR. Use the Sobol method to calculate the sensitivity index SM(i) = Var[E(Y|Xi)] / Var(Y), where Y is the response variable, and generate the sensitivity matrix SM. Calculate the importance score IV(i) = w1·SM(i) + w2·R(i) based on the sensitivity ranking, where R is the correlation index, and generate the importance vector IV. Synthesize each index to generate the sensitivity feature S(i, j) = f[SM(i), IV(j)], and generate the sensitivity matrix S.

[0276] Step S323: Transfer path analysis

[0277] Read the network structure N and the sensitivity matrix S. Design the path identification criterion IR(p) = ∏S(i)·∏N(i, j), where p is the path sequence, and generate the identification criterion IR. Calculate the path propagation feature PM(i, j) = ∑[w(k)·φ(x(k))], where φ is the feature function, and generate the propagation feature matrix. Screen the key paths KP(i) = {p | IR(p) > θr, PM(p) > θp} based on the feature importance, and generate the key path set. Calculate the path importance index IM(i) = KP(i)·S(i) / ∑S(j), and generate the importance matrix. Integrate the path information H(i) = f[KP(i), IM(i)] to generate the transfer path set.

[0278] Step S324: Coupling effect evaluation

[0279] Read the transfer path set H, determine the uncertainty source based on path tracing SM(i, j) = ∑H(k)·U(k, i, j), where U is the uncertainty distribution function, and generate the source matrix. Design the coupling effect evaluation function EF(i, j) = α·SM(i, j) + β·I(i, j), where I is the interaction intensity, and generate the evaluation function EF. Analyze the uncertainty interaction IM(i, j) = EF(i, j)·[1 + γ· C (i, j)], where C is the coupling coefficient, and generate the interaction matrix IM.

[0280] Action mode recognition process: Construct the mode PM(i, j) = ψ[IM(i, j)] based on the interaction features, where ψ is the mode recognition function, and generate the action mode matrix. Integrate the interaction features C (i) = f[IM(i), PM(i)], and generate the coupling coefficient set.

[0281] Step S325: Comprehensive uncertainty evaluation

[0282] Read the transfer path set and the coupling coefficient set, design the comprehensive evaluation function EF(x) = ∑[w(i)·x(i)]·c(i), where w is the weight coefficient, and generate the evaluation function. Analyze the contribution of each path C M(i, j) = H(i)· C (j) / ∑[H(k)· c(k)], and generate the contribution matrix. Calculate the overall uncertainty level LI(t) = ∑[cM(i, t)·EF(i)], and generate the level index LI. Construct the spatio-temporal distribution feature DM(i, j) = φ[LI(i)]·ψ(j), where φ and ψ are mapping functions, and generate the distribution matrix. The comprehensive evaluation result U(i) = f[LI(i), DM(i)], and generate the uncertainty index set.

[0283] Step S411: Watershed characteristic parameter extraction

[0284] Read the flow pattern set and the uncertainty index set. Extract the DEM-derived parameters TF(i, j) = {H(i, j), S(i, j), A(i, j), TWI(i, j)}, where TWI is the topographic wetness index, and generate the topographic feature matrix.

[0285] Calculate the hydrological response feature RI(i, t) = Q(i, t) / [P(t)·A(i)], where Q is the flow, P is the rainfall, and A is the area, and generate the response index matrix RI. Statistically analyze the land use composition LF(i) = {f1(i), f2(i),..., f n (i)}, where f kGenerate a land feature matrix for the proportion of the k-th type of land. Calculate the climate variation index cI(t) = [P(t) - P*] / σp, where P* is the average rainfall and σp is the standard deviation, and generate a climate index matrix. Integrate each characteristic index F(i) = g[TF(i), RI(i), LF(i), cI(i)] to generate a characteristic index set.

[0286] Step S412: Construction of model selection criteria

[0287] Read the characteristic index set F, design a model evaluation function EF(m) = ∑[w(i)·e(i, m)], where e is the evaluation index and w is the weight, and generate an evaluation function. Calculate the index weight IW(i) = [1 - H(i)] / ∑[1 - H(j)] based on the entropy weight method, where H is the information entropy, and generate an index weight matrix. Use the clustering method to determine the selection threshold TV(k) = μ k +σ k , where μ k , σ k are inter-class statistical parameters, and generate a threshold vector. Synthesize the evaluation function and the threshold SR(m) = {1: EF(m) > TV(k), 0: otherwise} to generate a selection rule. Integrate the rule and the weight K(m) = f[SR(m), IW(m)] to generate a selection criterion.

[0288] Step S413: Spatiotemporal heterogeneity analysis

[0289] Read the characteristic index set, divide the units based on feature similarity cU(i, j) = h[F(i, j)], where h is the classification function, and generate a set of calculation units cU. Calculate the spatial autocorrelation using Moran's I SH(i, j) = [n·∑∑wij(xi - x*)(xj - x*)] / [S o ·∑(xi - x*) 2 , and generate a spatial heterogeneity matrix. Calculate the time series variation index TM(i, t) = C V[x(i, t)], where cV is the coefficient of variation, and generate a time variation matrix. Analyze the multi-scale feature SI(i, s) = x(i, s) / x(i, s o ), where s is the scale parameter, and generate a scale index matrix. Integrate the spatiotemporal features H(i, j) = w1*SH(i, j) + w2*TM(i, j) + w3*SI(i, j) to generate a heterogeneity matrix.

[0290] Step S414: Optimize the weight assignment reading selection criteria and heterogeneity matrix. Design a weight assignment function DF(x) = exp[-λ·x] / ∑exp[-λ·x], where λ is the temperature parameter, to generate the assignment function DF. Calculate the model adaptability score SM(i, j) = K(i)·[1 - H(j)], where K is the selection score, to generate the score matrix. Construct the optimization objective function O P(w) = max{∑[SM(i)·w(i)]|∑w(i) = 1}, to generate the optimization plan. Solve the optimization plan to obtain the weight vector W(i) = DF[OP(i)], to generate the weight vector.

[0291] Step S415: Implement reliability assessment to read the traffic pattern set, construct a cross-validation scheme TP(i) = {T(i), V(i)}, where T is the training set and V is the validation set, to generate the test scheme. Execute model validation VR(i, j) = M[T(i), V(j)], where M is the model operation, to generate the validation result VR. Calculate the multiple precision metrics AM(i) = {NSE(i), RMSE(i), R2(i)}, to generate the precision matrix. Evaluate the model stability index SI(i) = σ[VR(i)] / μ[VR(i)], to generate the stability index. The comprehensive evaluation index R(i, j) = f[AM(i), SI(j)], to generate the reliability matrix. Among them, NSE is the Nash efficiency coefficient, RMSE is the root mean square error; R2 is the coefficient of determination, σ is the standard deviation operator, and μ is the mean operator.

[0292] Step S416: Construct an integrated model to read the weight vector and reliability matrix. Construct a hierarchical integration structure SF(i) = {L1(i), L2(i),..., L n (i)}, where L k is the k-th layer model group, to generate the integration framework. Design a model combination function C S(m) = α·W(m)·R(m) + β·D(m), where D is the diversity index, to generate the combination strategy. Use cross-validation to optimize the integration parameters PM(i, j) = argmin{∑[yt - f(xt, θ)]2}, to generate the parameter matrix. Based on the optimal parameters, construct an integrated model E(x) = ∑[CS(i)·Mi(x)], to generate the integrated model E.

[0293] Step S421: Construct constraint conditions to read the integrated model E, analyze the ecosystem requirements NF(t) = {Q1(t), Q2(t),..., Q n (t)}, where Q i is the i-th type of demand flow, to generate the demand feature matrix. Set the upper and lower limits of water volume constraints W C(t) = [Qmin(t), Qmax(t)], generating the water volume constraint matrix. Process of setting protection objectives: Construct the ecological protection index PM(i) = g[B(i), E(i)], where B is the reference value and E is the target value, generating the protection objective matrix. Calculate the system carrying capacity CI(t) = h[S(t), R(t)], where S is the supply and R is the demand, generating the carrying capacity index. Integrate various constraints C(i, t) = f[W C (t), PM(i), CI(t)], generating the set of constraint conditions C .

[0294] Step S422: Construction of the objective function Read the set of constraint conditions C. Construct the periodic function SF(t) = ao + ∑[a i cos(ω i t) + b i sin(ω i t)], where ω i is the frequency parameter, generating the variation function SF. Design the multi-objective optimization criterion OC (x) = {f1(x), f2(x),..., f n (x)}, where f i is the i-th objective function, generating the optimization criterion OC . Design the objective trade-off function TF(x) = ∑[w i ·f i (x)], where w i is the weight coefficient, generating the trade-off function TF. Integrate the seasonal variation and the trade-off function F(x, t) = TF(x)·SF(t), generating the objective function F. Among them, a o , a i , b i are the coefficients of the periodic function, ω i is the frequency parameter, unit: radian / day; f i is the i-th sub-objective function.

[0295] Step S423: Optimization solution implementation Read the set of constraint conditions and the objective function. Construct the optimization algorithm parameter set PS(i) = {p1(i), p2(i),..., p n (i)}, where p i is the algorithm parameter, generating the parameter set. Execute the optimization iteration IR(k) = argmin{F(x)| C (x) ≤ 0}, where k is the iteration number, generating the iteration result. Evaluate the optimization convergence characteristic C M(k) = |IR(k) - IR(k - 1)| / |IR(k - 1)|, generating the convergence index. Based on the convergence criterion, select the optimal solution I(x) = IR[k*], where k* is the optimal iteration number, generating the initial result.

[0296] Step S424: Ecological response assessment

[0297] Read the initial result and design an ecological response assessment function EF(x) = ∑[α i ·r i (x)], where r i is the response index and α i is the weight coefficient to generate the assessment function. Calculate the change of ecological index VM(i, t) = [y(i, t) - y o (i, t)] / y o (i, t), where y o is the benchmark value to generate the change matrix. Analyze the response sensitivity SI(i) = dy(i) / dx(i) · [x(i) / y(i)] to generate the sensitivity index SI. Extract the response characteristic parameter FM(i, j) = g[VM(i), SI(j)] to generate the characteristic matrix FM. Integrate the assessment result B(i) = h[SI(i), FM(i)] to generate the response index B. g and h are characteristic extraction functions. It should be noted that d is the partial derivative symbol.

[0298] Step S425: Result correction implementation

[0299] Read the initial result and the response index. Design a correction function MF(x) = x + λ·δ(x), where λ is the correction coefficient and δ is the deviation function to generate the correction function. Determine the adjustment parameter AM(i, j) = f[B(i), I(j)] based on the response index to generate the adjustment matrix AM. Perform the correction calculation MR(i) = MF[I(i)] · AM(i) to generate the correction result. Evaluate the correction effect EI(i) = |MR(i) - I(i)| / |I(i)| to generate the effect index. Determine the final result R(i) = k[MR(i), EI(i)] based on the effect index to generate the ecological flow estimation result.

[0300] Step S426: Reliability evaluation implementation

[0301] Read the ecological flow estimation result and construct an evaluation index system IS(i) = {q1(i), q2(i),..., q n (i)}, where q i is the evaluation index to generate the index system. Calculate the reliability index RM(i, j) = p[R(i, j) ≥ R o (i, j)], where R oGenerate a reliability matrix for the target value. The uncertainty of the evaluation result UI(i) = σ[R(i)] / μ[R(i)], and generate an uncertainty index. Analyze the adaptability feature AM(i, j) = f[RM(i), UI(j)], and generate an adaptability matrix. The comprehensive evaluation result A(i) = g[RM(i), UI(i), AM(i)], and generate a reliability evaluation index.

[0302] The preferred embodiments of the present invention have been described in detail above. However, the present invention is not limited to the specific details in the above embodiments. Within the scope of the technical concept of the present invention, various equivalent transformations can be made to the technical solutions of the present invention, and these equivalent transformations all belong to the protection scope of the present invention.

Claims

1. An intelligent estimation method for ecological flow in ungauged mountainous watersheds based on multi-scale fusion, characterized in that, It includes the following steps: Collect the basic data of the research area and preprocess it to obtain an effective data set; the basic data includes DEM data, remote sensing image data, meteorological reanalysis data, and land use data; Extract the hydrological feature set and process feature set according to the effective data set; Generate dynamic response units based on the hydrological feature set and process feature set, and perform multi-dimensional hydrological process decomposition to obtain a multi-scale feature matrix; Perform eco-hydrological sequence matching on the multi-scale feature matrix to obtain a flow pattern set, and perform hierarchical uncertainty propagation analysis to generate an uncertainty index set; Construct an integrated model according to the flow pattern set and uncertainty index set, and perform multi-constraint optimization to obtain the ecological flow estimation result.

2. The method according to claim 1, characterized in that, The preprocessing includes: Obtain and calculate the terrain complexity index value of the basic data to obtain the terrain decomposition factor; Perform local decomposition and global decomposition on the basic data based on the terrain decomposition factor to generate a decomposition coefficient matrix; Calculate the hydrological similarity value according to the decomposition coefficient matrix to obtain the similarity index; Generate a data quality index based on the similarity index to obtain an effective data set.

3. The method according to claim 1, characterized in that, The steps of extracting the hydrological feature set and process feature set according to the effective data set include: Extract the slope value, aspect value, and elevation value of the grid cell, and calculate the geomorphic position index to obtain the position index matrix; Divide the geomorphic unit types based on the position index matrix, and extract the characteristic parameters of each type of unit to generate a geomorphic feature vector; Construct a feature decomposition basis function based on the geomorphic feature vector to obtain a basis function set; Calculate the time response relationship between rainfall data and runoff data to obtain the time lag coefficient matrix; Perform feature extraction on the effective data set according to the basis function set and time lag coefficient matrix to obtain a feature component set; Perform multi-scale combination on the feature component set to generate a hydrological feature set and process feature set.

4. The method according to claim 1, characterized in that, The steps of generating dynamic response units include: Calculate the terrain gradient value of each point in the basin to obtain the gradient matrix; Construct a spatial neighborhood relationship according to the gradient matrix to generate a neighborhood matrix; Calculate the confluence time of each region to obtain the time matrix; Construct a dynamic partitioning criterion based on the neighborhood matrix and time matrix to obtain a partitioning criterion; Perform adaptive partitioning according to the partitioning criterion to obtain an initial partitioning result; Dynamically adjust the initial partitioning result according to seasonal changes to obtain dynamic response units.

5. The method according to claim 4, characterized in that, The steps of performing multi-dimensional hydrological process decomposition include: Extract the terrain features of each unit to obtain a terrain feature set; Calculate the hydrological response parameters of each unit to obtain a response parameter set; Construct a feature expression at the unit scale to obtain a feature function set; Perform multi-dimensional decomposition based on the feature function set to obtain an initial decomposition set; Optimize the decomposition coefficients of the initial decomposition set to generate a decomposition coefficient set; Calculate the energy value of each decomposition scale based on the decomposition coefficient set to obtain a multi-scale feature matrix.

6. The method according to claim 1, characterized in that, The steps of performing eco-hydrological sequence matching to obtain a flow pattern set include: Extract the water volume change sequence to obtain a water volume sequence; Calculate the characteristics of the ecological sensitive period to obtain a sensitive period matrix; Analyze the ecological water demand to obtain a demand matrix; Calculate the response intensity based on the water volume sequence and sensitive period matrix to obtain an intensity matrix; Combine the demand matrix and intensity matrix to construct a response feature, and generate a response feature set; Perform pattern recognition according to the response feature set to generate a flow pattern set.

7. The method according to claim 5, wherein The steps for hierarchical uncertainty transfer analysis to generate an uncertainty index set include: Construct a hydrological process transfer chain network to obtain the network structure; Calculate the parameter sensitivity index of each node to generate a sensitivity matrix; Analyze the uncertainty transfer path based on the network structure and the sensitivity matrix to obtain a transfer path set; Evaluate the coupling effect of multi-source uncertainties to obtain a coupling coefficient set; Calculate the comprehensive uncertainty based on the transfer path set and the coupling coefficient set to generate an uncertainty index set.

8. The method according to claim 1, characterized in that, The steps for constructing an integrated model include: Extract topographic feature parameters to obtain a topographic feature matrix; Calculate hydrological response indicators to obtain a response indicator matrix; Analyze land use characteristics to obtain a land feature matrix; Evaluate climate variability to obtain a climate indicator matrix; Construct a model selection criterion based on the topographic feature matrix, the response indicator matrix, the land feature matrix, and the climate indicator matrix; Calculate the spatio-temporal heterogeneity index of each region to obtain a heterogeneity matrix; Determine the model weights based on the model selection criterion and the heterogeneity matrix to generate a weight vector; Evaluate the reliability index of each model to obtain a reliability matrix; Construct an integrated model using the weight vector and the reliability matrix.

9. The method according to claim 1, wherein The steps for multi-constraint optimization to obtain the ecological flow estimation result include: Construct a seasonal change function to obtain a change function; Design an ecological demand constraint system to obtain a set of constraint conditions; Construct an objective function considering seasonal changes; Perform optimization solution based on the set of constraint conditions and the objective function to obtain an initial result; Analyze the response characteristics of the ecosystem to the initial result to obtain response indicators; Correct the initial result based on the response indicators to generate an ecological flow estimation result.

10. The method according to claim 5, characterized in that, Ecological hydrological sequence matching also includes: Extract river network structure data to obtain a river network matrix; Calculate the physical connectivity between river reaches to obtain a physical connectivity matrix; Analyze the hydrological connectivity status to obtain a hydrological connectivity matrix; Evaluate the ecological corridor function to obtain a corridor index matrix; Integrate the physical connectivity matrix, the hydrological connectivity matrix, and the corridor index matrix to generate a connectivity matrix; The connectivity matrix and the response feature set are used for pattern recognition to generate a flow pattern set.

11. The method according to claim 6, characterized in that, The steps for analyzing the uncertainty transfer path include: Design a perturbation experiment scheme to obtain a perturbation scheme; Execute a parameter perturbation test to obtain test results; Calculate the response sensitivity to obtain a sensitivity matrix; Analyze the parameter importance to obtain an importance vector; Identify the uncertainty sources to obtain a source matrix; Analyze the coupling action mode to obtain an action mode matrix; Generate a coupling coefficient set based on the action mode matrix.

Images