Ecological safety assessment method for arid region drainage basin based on WSRS framework

Abstract

The invention discloses a WSRS framework-based ecological safety assessment method for a drainage basin in an arid region, relates to the technical field of ecological assessment, and solves the technical problems that the prior art is not adaptive to a coupling process of water, salt, evapotranspiration and underground water and lacks key factors, and an assessment method lacks weight uncertainty description and robustness analysis. The method comprises the steps of S1, acquiring multi-source data of a target drainage basin, and performing data preprocessing; s2, constructing a WSRS evaluation index system, and carrying out standardization processing on indexes; s3, based on the standardized indexes, determining the comprehensive weight of each index by adopting a subjective and objective combination weighting method; s4, performing comprehensive evaluation based on the comprehensive weight, and quantifying uncertainty of an evaluation result by adopting a cloud model to obtain an ecological safety index; s5, constructing a composite stress model and correcting an ecological safety index; and S6, carrying out risk grade division and result visualization. The method has the advantages that the arid region process adaptability is high, the evaluation result is stable and explainable, and the like.

Description

Technical Field

[0001] This invention belongs to the field of ecological assessment technology, specifically relating to a method for assessing the ecological security of watersheds in arid regions based on the WSRS framework. Background Technology

[0002] Arid watersheds are characterized by low precipitation, high evapotranspiration, scarce water resources, and extremely fragile ecosystems. The regional ecological security is not only affected by climate change, but also by a combination of factors such as surface water inflow and quality, inter-basin or main stream ecological water transport, groundwater depth and mineralization, soil salinity accumulation, and evapotranspiration intensity.

[0003] Currently, watershed ecological security assessments mainly adopt the following technical approaches: First, indicator frameworks such as PSR / DPSIR, typically constructing a Pressure-State-Response (PSR) or Driver-Pressure-State-Impact-Response (DPSIR) indicator system, primarily based on indicators such as land use, population density, economic development, water quality, water quantity, and ecological protection, and obtaining the ecological security index through a weighted comprehensive method; Second, conventional comprehensive evaluation methods, such as fuzzy evaluation and principal component analysis, determining comprehensive indicators through fuzzy comprehensive evaluation, principal component analysis (PCA), and analytic hierarchy process (AHP). However, in existing technologies, the indicator system is not customized for water, salinity, evapotranspiration, and groundwater processes in arid areas, and key factors such as policy water transfer, groundwater depth and mineralization, soil salinity, and evapotranspiration are missing; the assessment methods lack characterization of weight uncertainty, robustness analysis of results, and interpretability analysis; and there is a lack of an assessment framework closely coupled with scenario simulation and threshold back-calculation. Summary of the Invention

[0004] In view of this, the present invention discloses an ecological security assessment method for arid watersheds based on the WSRS framework, which aims to solve the technical problems of existing technologies not being adapted to the coupling processes of water, salinity, evapotranspiration and groundwater, lacking key factors, and lacking weight uncertainty characterization and robustness analysis.

[0005] To solve the aforementioned technical problems, the present invention adopts the following technical solution:

[0006] A method for assessing the ecological security of watersheds in arid areas based on the WSRS framework includes the following steps:

[0007] Step S1: Obtain multi-source data from the target watershed and perform data preprocessing to obtain a multi-source dataset;

[0008] Step S2: Based on the multi-source dataset obtained in Step S1, construct a WSRS assessment index system including four dimensions: hydrology and water resources, water and salt and evapotranspiration stress, ecological resilience and ecological support, and standardize the indicators of the index system.

[0009] Step S3: Based on the standardized indicators, determine the comprehensive weight of each indicator using a combination of subjective and objective weighting methods;

[0010] Step S4: Based on the comprehensive weights, a comprehensive evaluation is performed, and the uncertainty of the evaluation results is quantified using a cloud model to obtain the ecological security index;

[0011] Step S5: Construct a composite stress model based on water, salt, and evapotranspiration data and correct the ecological security index;

[0012] Step S6: Classify risk levels and visualize the results based on the revised ecological security index.

[0013] In this invention, the letters in the WSRS indicator system represent the following: W refers to an example of a hydrological and water resources dimension indicator, whose data comes from river basin management agencies and ecological water transfer scheduling platforms, including policy-managed water transfers, which include ecological water transfers and inter-basin water transfers, with units of 10. 8 m³ / a; surface water inflow and total watershed runoff; ecological flow satisfaction, i.e., actual flow / ecological water demand; surface water quality, including conductivity EC, total dissolved solids (TDS), chemical oxygen demand (COD), total nitrogen (TN), and total phosphorus (TP), which constitute a comprehensive water quality index; groundwater depth, in meters; groundwater salinity and total dissolved solids (TDS) of groundwater. gw Groundwater conductivity EC gw The data includes indicators such as salinity (Cl⁻, SO₄²⁻, etc.) and groundwater extraction rate (extraction / recharge). The first S represents an example of indicators related to water, salinity, and evapotranspiration stress, with data sourced from soil sampling monitoring and remote sensing inversion. These indicators include soil salinity (g / kg); soil electrical conductivity (Ece) (dS / m); salinization index (SI); surface water salinity index; actual evapotranspiration (ET) (mm / a); evapotranspiration ratio (ET / P); and surface temperature (LST), reflecting dry and hot stress. R represents an example of indicators related to ecological resilience, with data sourced from MODIS or calculated independently. These include the normalized difference vegetation index (NDVI), annual mean and interannual variability of net primary productivity (NPP); vegetation resilience; landscape connectivity, patch aggregation, fragmentation, etc.; baseflow index; and other hydrological resilience indicators. The second S represents an example of indicators for the ecological support / service dimension. The data comes from the Ministry of Natural Resources and the Department of Agriculture and Rural Affairs of the Ministry of Ecology and Environment, including the proportion of ecological water use in total water use; the frequency and total amount of ecological water replenishment; the proportion of water-saving irrigation area; the coverage rate of nature reserves and key ecological function zones; the proportion of farmland converted to grassland; and the intensity of soil and water conservation projects.

[0014] Furthermore, in step S1, data preprocessing includes: interpolating missing values, removing outliers, unifying the spatial resolution of multi-source data, and aggregating seasonal indicators annually.

[0015] After adopting this technical solution, the spatial resolution of multi-source data is unified to a unified grid, such as 250m / 1km, and the time scale is unified to an annual or multi-year average. For indicators with strong seasonality, such as NDVI, ET, LST, etc., annual normalization is performed first and then annual aggregation is performed to avoid the bias introduced by seasonal differences.

[0016] As a preferred method, interpolation for missing values ​​includes using interpolation of nearby years, seasonal quantile interpolation, or time series interpolation. For a certain evaluation unit i and index j, when the data is missing in year t, time series linear interpolation is used, as shown in the following formula:

[0017] ;

[0018] In the formula, Let be the value of the j-th index of evaluation unit i in year t.

[0019] For highly seasonal indicators such as NDVI and ET, missing values ​​are estimated using quantiles on historical samples from the same season, as shown in the following formula:

[0020]

[0021] In the formula, As an evaluation unit In the year ,season Indicators The estimated value, i.e. the result of filling in the missing values; As an indicator In the season Historical sample sets, such as the set of observations for this season from 2000 to 2022. ; This represents the quantile function, which is the function of taking the quantile from the historical sample set. The numerical value corresponding to the quantile; This is an empirical quantile parameter, typically taken as 0.5 or, based on climatic skewness, as a value between 0.25 and 0.75.

[0022] As a preferred method, outliers are removed using either the interquartile range or the 3σ method, with the following formulas:

[0023] ;

[0024] .

[0025] In the formula, Interquartile range; This represents the original observed value of the j-th index in evaluation unit i; The average value of index j across all samples; The standard deviation of the indicator reflects the degree of dispersion of the indicator value relative to the mean.

[0026] Further, in step S2, the standardization process includes: linear forward standardization of positive indices, linear reverse standardization of negative indices, and standardization of near-optimal indices using a piecewise linear or Gaussian membership function. The near-optimal indices have an optimal range of values. The linear forward standardization formula is:

[0027] ;

[0028] The linear inverse standardization formula is:

[0029] ;

[0030] The formula for the near-optimal index is: .

[0031] In the formula, The j-th index of evaluation unit i is a dimensionless value after near-optimal standardization, with a value range of [0,1]. The measured or calculated value of the i-th sample on index j; This is the lower boundary of the near-optimal interval; that is, once the index value reaches this level, it is considered to be close to the ideal state. This is the minimum threshold for the lower limit of the indicator; This is the upper boundary of the near-optimal interval; values ​​exceeding this are still within the ideal interval. This is the maximum threshold for the indicator.

[0032] Further, step S3 includes:

[0033] Step S3.1: Calculate the objective weights using information entropy weights and CRITIC weights respectively, and determine the subjective weights using the analytic hierarchy process (AHP).

[0034] In step S3.1, the information entropy weight refers to the entropy value and information utility value calculated by the dispersion of the indicator on the sample, and the entropy weight is obtained. Since the entropy weight focuses on the dispersion of the indicator's own value, while CRITIC further introduces the correlation with other indicators to avoid highly redundant information being repeatedly amplified. The two constitute a complementary source of objective weight.

[0035] The specific steps for information entropy weighting are as follows:

[0036] Standardized The probability matrix is ​​constructed by normalization, as shown in the following formula:

[0037] ;

[0038] Calculate information entropy The formula is as follows:

[0039] ;

[0040] Where n is the number of evaluation units; This is the entropy normalization constant, used to normalize information entropy. The value range of is normalized to [0, 1] to ensure that the entropy values ​​are comparable across different sample sizes. If the difference between index j and each unit is small, then Nearly uniform, information entropy The information entropy is relatively large; if the index j varies significantly across different units, then the information entropy is relatively large. Smaller, thus used to measure uniformity;

[0041] Calculate information utility value The formula is as follows:

[0042] ;

[0043] Information utility value The larger the value, the greater the amount of effective information in indicator j, and the greater its contribution to the overall evaluation.

[0044] The entropy value is obtained by normalization, as shown in the following formula:

[0045] ;

[0046] In the formula, m is the number of indicators;

[0047] The specific steps for CRITIC weighting are as follows:

[0048] The standard deviation is calculated using the following formula:

[0049] ;

[0050] In the formula, Let be the sample mean of indicator j;

[0051] The Pearson correlation coefficient between indicator j and indicator k is calculated using the following formula:

[0052] ;

[0053] In the formula, Let J be the Pearson correlation coefficient between index j and index k. Let j be the standardized value of the index. Let k be the standardized value of the index. Let be the mean of index j; The mean of index k; This represents the number of evaluation units.

[0054] The formula for calculating contrast intensity is as follows:

[0055] ;

[0056] The CRITIC weights are obtained by normalization, as shown in the following formula:

[0057] .

[0058] The subjective weights determined using the analytic hierarchy process include:

[0059] Constructing the judgment matrix: Based on the WSRS indicator system, experts in water resources, ecology, and management were invited to conduct pairwise comparisons of indicators at the same level, constructing a judgment matrix A=[ ].in, This indicates the importance of indicator j relative to indicator k, satisfying the following condition: =1, =1 / ;

[0060] The eigenvalue method is used to perform eigenvalue decomposition on the judgment matrix A to obtain the largest eigenvalue. and its corresponding eigenvector v = (v1, v2, ..., vm)ᵀ, i.e. ;

[0061] The AHP weights of each index are obtained by normalizing the feature vector v, as shown in the following formula:

[0062] ;

[0063] Based on the largest eigenvalue The consistency index is calculated using the following formula:

[0064] ;

[0065] In the formula, is a consistency index, a quantitative indicator that measures the degree of deviation of the consistency of the judgment matrix; m is the order of the judgment matrix; It is the largest eigenvalue; CR is the consistency ratio, which measures the deviation of the actual consistency of the judgment matrix from the random consistency; RI is the random consistency index of the same order. When CR < 0.1, the consistency of the judgment matrix is ​​considered to meet the requirements, and the AHP weights are calculated accordingly. efficient.

[0066] Step S3.2: Use game theory to combine and assign weights to the objective and subjective weights to obtain the final weight vector.

[0067] In step S3.2, the entropy weight, CRITIC weight, and AHP weight are considered as three independent weighting schemes, and the combined weight is set as follows:

[0068] ;

[0069] The three factors are coordinated and integrated using game theory, where α, β, and γ ≥ 0, and α + β + γ = 1;

[0070] The least squares optimization minimizes the sum of squared deviations between the combined weights and the individual weights, as shown in the following formula:

[0071] ;

[0072] The optimal values ​​of α, β, and γ are obtained by using the above formulas, resulting in the final weight vector w = (w1, ..., wm). Furthermore, the final weight vector can be used as a parameter of the Dirichlet distribution to simulate the impact of weight uncertainty on the ecological security index results through Monte Carlo sampling.

[0073] Further, step S4 includes:

[0074] Step S4.1: Use a weighted normalization matrix to determine the positive and negative ideal solutions;

[0075] In step S4.1, the weighted normalization matrix is: Positive ideal solution vector Negative ideal solution vector ;

[0076] Step S4.2: Calculate the distance from each evaluation unit to the positive and negative ideal solutions using Mahalanobis distance to obtain the proximity score;

[0077] In step S4.2, let the covariance matrix be... Its inverse matrix is Therefore, the Mahalanobis distance formula from evaluation unit i to the positive ideal solution is:

[0078] ;

[0079] The Mahalanobis distance formula to the negative ideal solution is:

[0080] ;

[0081] In the formula, Let be the standardized index vector for the i-th evaluation unit.

[0082] Relative closeness is calculated using the following formula:

[0083] ;

[0084] Step S4.3: Map the proximity to the cloud model to three feature quantities: expectation, entropy, and hyperentropy, to obtain the Ecological Security Index (ESI);

[0085] In step S4.3, the proximity of each evaluation unit is treated as a random variable sample, and the expected value Ex, entropy En, and hyperentropy He of the cloud model are introduced to achieve a robust estimate of the ecological security index ESI and its uncertainty. For example, the mean of the proximity sequence over many years can be used. As expected , with standard deviation As a fundamental quantity of entropy, and with hyperentropy defined as the proportionality coefficient of entropy:

[0086] ;

[0087] in, For functions that match the specific cloud model form, The empirical coefficients are used to generate cloud droplets based on the above three features, thus obtaining the ecological security index and its confidence interval for each evaluation unit.

[0088] Further, step S5 includes:

[0089] Step S5.1: Based on the multi-source dataset obtained in step S1, extract key stress factor data;

[0090] In step S5.1, key stress factors include groundwater depth, groundwater salinity, soil salinity, and evapotranspiration ratio.

[0091] Step S5.2: Construct the stress function of a single stress factor and superimpose them to obtain the comprehensive stress index. Then, correct the ecological security assessment method using the comprehensive stress index.

[0092] In step S5.2, the stress function of a single stress factor includes:

[0093] Groundwater depth stress function:

[0094] ;

[0095] Groundwater mineralization stress function:

[0096] ;

[0097] Soil salinity stress function:

[0098] ;

[0099] Evapotranspiration stress function:

[0100] ;

[0101] The comprehensive stress function is then obtained, as shown in the following formula:

[0102] .

[0103] Further, step S6 includes:

[0104] Step S6.1: Using the natural breakpoint method or the principle of cloud model entropy minimization, the ecological security index is divided into several levels;

[0105] Step S6.2: Draw ecological security index level distribution maps, uncertainty distribution maps, and water and salt stress distribution maps based on the ecological security index levels, and identify high-risk areas, sensitive areas, and key management areas.

[0106] As a preferred option, it also includes:

[0107] Step S7: Based on the multi-source data collected in Step S1, construct a dynamic model of the watershed water resources system, simulate the changing trends of ecological security under different climate, water allocation, ecological water replenishment, and land use scenarios, and use optimization algorithms to back-calculate the minimum combination of management measures required to achieve the target ecological security level.

[0108] In step S7, the watershed water resources system dynamic model includes four modules: water balance, groundwater recovery, salinity migration, and vegetation growth.

[0109] After adopting this technical solution, the model formulas for the four major modules—water balance, groundwater restoration, salinity migration, and vegetation growth—are as follows:

[0110] ,

[0111] In the formula, This refers to the change in the total water storage of a watershed per unit time. Natural precipitation over a certain period of time. This refers to the inflow of surface water from upstream or external sources. This refers to the artificial regulation of water input, such as ecological water replenishment and inter-basin water transfer. This includes total water loss from soil evaporation and vegetation transpiration. This refers to the amount of surface water infiltrating into the groundwater system. The amount of water discharged downstream or into other basins.

[0112] ,

[0113] In the formula, This refers to the renewed reserves after replenishment and extraction. This represents the total groundwater storage at time t. This refers to groundwater recharge. This refers to the amount of groundwater extracted. This is due to groundwater evaporation loss.

[0114] ,

[0115] .

[0116] In the formula, For the next moment, soil salinity; Let be the soil salinity at time t; k1 is the capillary rise coefficient; k2 is the rinsing coefficient. It is a capillary rising function; is the leaching function, and m is the natural degradation coefficient of vegetation.

[0117] The NDVI value at the next moment; Let be the normalized vegetation index at time t; Let be the vegetation response function, which represents the comprehensive response of vegetation growth to water W and salinity S conditions. It is usually an empirical nonlinear function.

[0118] Preferably, the optimization algorithm includes: constructing an objective function with the goal of minimizing management measures. The constraints of the objective function are:

[0119]

[0120] .

[0121] In the formula, For ecological security index; To achieve the target level of ecological security; The variance of the ecological security index; This represents the upper limit of the target variance.

[0122] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0123] 1. The present invention provides a watershed ecological security assessment method for arid areas based on the WSRS framework. By constructing the WSRS framework, key variables such as policy water transfer, surface water quantity and quality, groundwater depth and mineralization, soil salinity, and evapotranspiration are incorporated into a unified framework, which can systematically characterize the ecological risk mechanism of arid areas characterized by "low water - high salinity - strong evapotranspiration - groundwater sensitivity".

[0124] 2. The present invention provides a method for assessing the ecological security of arid watersheds based on the WSRS framework. By constructing a water-salt-evapotranspiration composite stress index, the latent ecological risks under conditions of groundwater level change, increased mineralization, soil salinity accumulation and high evapotranspiration can be made explicit and quantified, thereby improving the ability to identify important areas such as salinization hotspots and oasis-desert transition zones.

[0125] 3. The present invention provides a watershed ecological security assessment method based on the WSRS framework in arid areas. It adopts a combination of entropy weight, CRITIC, and AHP game weighting, and Mahalanobis distance and cloud model for comprehensive evaluation. It also introduces Monte Carlo uncertainty analysis of weight and index perturbation, which can show the reliability of regional results, which indicators have the greatest impact on the results, and key thresholds.

[0126] 4. The present invention provides a method for assessing the ecological security of arid watersheds based on the WSRS framework. By combining system dynamics models and scenario simulations, it answers questions such as "how will the ESI level change in the future under a certain water transfer scheme and groundwater control intensity" and "what is the minimum amount of policy water transfer, water replenishment frequency and salinization intensity required to upgrade a certain area from a lower level to a medium or higher level", and has direct decision support value. Attached Figure Description

[0127] The present invention will be described by way of example and with reference to the accompanying drawings, wherein:

[0128] Figure 1 This is a flowchart of the present invention. Detailed Implementation

[0129] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the embodiments and accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. The components of the embodiments of this application described and marked in the accompanying drawings can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely represents selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0130] In the description of the embodiments of this application, it should be noted that the terms "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, or the orientation or positional relationship conventionally placed when the invention is used. They are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application. In addition, the terms "first," "second," and "third," etc., are only used to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0131] Example 1

[0132] A method for assessing the ecological security of watersheds in arid regions based on the WSRS framework, such as Figure 1 As shown, it includes the following steps:

[0133] Step S1: Obtain multi-source data from the target watershed and perform data preprocessing to obtain a multi-source dataset;

[0134] Step S2: Based on the multi-source dataset obtained in Step S1, construct a WSRS assessment index system including four dimensions: hydrology and water resources, water and salt and evapotranspiration stress, ecological resilience and ecological support, and standardize the indicators of the index system.

[0135] Step S3: Based on the standardized indicators, determine the comprehensive weight of each indicator using a combination of subjective and objective weighting methods;

[0136] Step S4: Based on the comprehensive weights, a comprehensive evaluation is performed, and the uncertainty of the evaluation results is quantified using a cloud model to obtain the ecological security index;

[0137] Step S5: Construct a composite stress model based on water, salt, and evapotranspiration data and correct the ecological security index;

[0138] Step S6: Classify risk levels and visualize the results based on the revised ecological security index.

[0139] In step S1, data preprocessing includes: interpolating missing values, removing outliers, unifying the spatial resolution of multi-source data, and aggregating seasonal indicators annually.

[0140] In this embodiment, the spatial resolution of multi-source data is unified to a uniform grid, such as 250m / 1km, and the time scale is unified to an annual or multi-year average. For highly seasonal indicators, such as NDVI, ET, and LST, annual normalization is performed first, followed by annual aggregation to avoid bias introduced by seasonal differences. Interpolation for missing values ​​includes using interpolation of neighboring years, seasonal quantile interpolation, or time series interpolation methods. For a certain evaluation unit i and indicator j, when data is missing in year t, time series linear interpolation is used, as shown in the following formula:

[0141] ;

[0142] In the formula, Let be the value of the j-th index of evaluation unit i in year t.

[0143] For highly seasonal indicators such as NDVI and ET, missing values ​​are estimated using quantiles on historical samples from the same season, as shown in the following formula:

[0144] ;

[0145] In the formula, As an evaluation unit In the year ,season Indicators The estimated value, i.e. the result of filling in the missing values; As an indicator In the season Historical sample set, and ; This represents the quantile function, which is the function of taking the quantile from the historical sample set. The numerical value corresponding to the quantile; This is an empirical quantile parameter, typically taken as 0.5 or, based on climatic skewness, as a value between 0.25 and 0.75.

[0146] Outliers are removed using either the interquartile range or the 3σ method, with the formulas as follows:

[0147] ;

[0148] .

[0149] In the formula, Interquartile range; This represents the original observed value of the j-th index in evaluation unit i; The average value of index j across all samples; The standard deviation of the indicator reflects the degree of dispersion of the indicator value relative to the mean.

[0150] In step S2, the standardization process includes: linear forward standardization of positive indices, linear reverse standardization of negative indices, and standardization of near-optimal indices using piecewise linear or Gaussian membership functions. The formula for linear forward standardization is:

[0151] ;

[0152] The linear inverse standardization formula is:

[0153] ;

[0154] The formula for the near-optimal index is: .

[0155] In the formula, The j-th index of evaluation unit i is a dimensionless value after near-optimal standardization, with a value range of [0,1]. The measured or calculated value of the i-th sample on index j; This is the lower boundary of the near-optimal interval; that is, once the index value reaches this level, it is considered to be close to the ideal state. This is the minimum threshold for the lower limit of the indicator; This is the upper boundary of the near-optimal interval; values ​​exceeding this are still within the ideal interval. This is the maximum threshold for the indicator.

[0156] Step S3 includes:

[0157] Step S3.1: Calculate the objective weights using information entropy weights and CRITIC weights respectively, and determine the subjective weights using the analytic hierarchy process (AHP).

[0158] In step S3.1, the information entropy weight refers to the entropy value and information utility value calculated by the dispersion of the indicator on the sample, and the entropy weight is obtained. Since the entropy weight focuses on the dispersion of the indicator's own value, while CRITIC further introduces the correlation with other indicators to avoid highly redundant information being repeatedly amplified. The two constitute a complementary source of objective weight.

[0159] The specific steps for information entropy weighting are as follows:

[0160] Standardized The probability matrix is ​​constructed by normalization, as shown in the following formula:

[0161] ;

[0162] Calculate information entropy The formula is as follows:

[0163] ;

[0164] Where n is the number of evaluation units; This is the entropy normalization constant, used to normalize information entropy. The value range of is normalized to [0, 1] to ensure that the entropy values ​​are comparable across different sample sizes. If the difference between index j and each unit is small, then Nearly uniform, information entropy The information entropy is relatively large; if the index j varies significantly across different units, then the information entropy is relatively large. Smaller, thus used to measure uniformity;

[0165] Calculate information utility value The formula is as follows:

[0166] ;

[0167] Information utility value The larger the value, the greater the amount of effective information in indicator j, and the greater its contribution to the overall evaluation.

[0168] The entropy value is obtained by normalization, as shown in the following formula:

[0169] ;

[0170] In the formula, m is the number of indicators;

[0171] The specific steps for CRITIC weighting are as follows:

[0172] The standard deviation is calculated using the following formula:

[0173] ;

[0174] In the formula, Let be the sample mean of indicator j;

[0175] The Pearson correlation coefficient between indicator j and indicator k is calculated using the following formula:

[0176] ;

[0177] In the formula, Let J be the Pearson correlation coefficient between index j and index k. Let j be the standardized value of the index. Let k be the standardized value of the index. Let be the mean of index j; The mean of index k; This represents the number of evaluation units.

[0178] The formula for calculating contrast intensity is as follows:

[0179] ;

[0180] The CRITIC weights are obtained by normalization, as shown in the following formula:

[0181] .

[0182] The subjective weights determined using the analytic hierarchy process include:

[0183] Constructing the judgment matrix: Based on the WSRS indicator system, experts in water resources, ecology, and management were invited to conduct pairwise comparisons of indicators at the same level, constructing a judgment matrix A=[ ].in, This indicates the importance of indicator j relative to indicator k, satisfying the following condition: =1, =1 / ;

[0184] The eigenvalue method is used to perform eigenvalue decomposition on the judgment matrix A to obtain the largest eigenvalue. and its corresponding eigenvector v = (v1, v2, ..., vm)ᵀ, i.e. ;

[0185] The AHP weights of each index are obtained by normalizing the feature vector v, as shown in the following formula:

[0186] ;

[0187] Based on the largest eigenvalue The consistency index is calculated using the following formula:

[0188] ;

[0189] In the formula, in the formula, is a consistency index, a quantitative indicator that measures the degree of deviation of the consistency of the judgment matrix; m is the order of the judgment matrix; It is the largest eigenvalue; The consistency ratio (CR) measures the deviation of the actual consistency of the judgment matrix from the random consistency. RI is the random consistency index of the same order. In this embodiment, when CR < 0.1, the consistency of the judgment matrix is ​​considered to meet the requirements. The calculated AHP weights are then used. efficient.

[0190] Step S3.2: Use game theory to combine and assign weights to the objective and subjective weights to obtain the final weight vector.

[0191] In step S3.2, the entropy weight, CRITIC weight, and AHP weight are considered as three independent weighting schemes, and the combined weight is set as follows:

[0192] ;

[0193] The three factors are coordinated and integrated using game theory methods, where α, β, and γ ≥ 0, and α + β + γ = 1.

[0194] The least squares optimization minimizes the sum of squared deviations between the combined weights and the individual weights, as shown in the following formula:

[0195] ;

[0196] The optimal values ​​of α, β, and γ are obtained by using the above formulas, resulting in the final weight vector w = (w1, ..., wm). Furthermore, the final weight vector can be used as a parameter of the Dirichlet distribution to simulate the impact of weight uncertainty on the ecological security index results through Monte Carlo sampling.

[0197] Step S4 includes:

[0198] Step S4.1: Use a weighted normalization matrix to determine the positive and negative ideal solutions;

[0199] In step S4.1, the weighted normalization matrix is: Positive ideal solution vector Negative ideal solution vector ;

[0200] Step S4.2: Calculate the distance from each evaluation unit to the positive and negative ideal solutions using Mahalanobis distance to obtain the proximity score;

[0201] In step S4.2, let the covariance matrix be... Its inverse matrix is Therefore, the Mahalanobis distance formula from evaluation unit i to the positive ideal solution is:

[0202] ;

[0203] The Mahalanobis distance formula to the negative ideal solution is:

[0204] ;

[0205] In the formula, Let be the standardized index vector for the i-th evaluation unit;

[0206] Relative closeness is calculated using the following formula:

[0207] ;

[0208] Step S4.3: Map the proximity to the cloud model to three feature quantities: expectation, entropy, and hyperentropy, to obtain the Ecological Security Index (ESI);

[0209] In step S4.3, the proximity of each evaluation unit is treated as a random variable sample, and the expected value Ex, entropy En, and hyperentropy He of the cloud model are introduced to achieve a robust estimate of the ecological security index ESI and its uncertainty. For example, the mean of the proximity sequence over many years can be used. As expected , with standard deviation As a fundamental quantity of entropy, and with hyperentropy defined as the proportionality coefficient of entropy:

[0210] ;

[0211] in, For functions that match the specific cloud model form, The empirical coefficients are used to generate cloud droplets based on the above three features, thus obtaining the ecological security index and its confidence interval for each evaluation unit.

[0212] Step S5 includes:

[0213] Step S5.1: Based on the multi-source dataset obtained in step S1, extract key stress factor data;

[0214] In step S5.1, key stress factors include groundwater depth, groundwater salinity, soil salinity, and evapotranspiration ratio.

[0215] Step S5.2: Construct the stress function of a single stress factor and superimpose them to obtain the comprehensive stress index. Then, correct the ecological security assessment method using the comprehensive stress index.

[0216] In step S5.2, the stress function of a single stress factor includes:

[0217] Groundwater depth stress function:

[0218] ;

[0219] Groundwater mineralization stress function:

[0220] ;

[0221] Soil salinity stress function:

[0222] ;

[0223] Evapotranspiration stress function:

[0224] ;

[0225] The comprehensive stress function is then obtained, as shown in the following formula:

[0226] .

[0227] In this embodiment, the comprehensive stress index S reflects the combined stress intensity of water, salt, and evapotranspiration. This index can be used as a correction factor for the ecological security index or as the S-dimensional core sub-indicator in the WSRS system.

[0228] Step S6 includes:

[0229] Step S6.1: Using the natural breakpoint method or the principle of cloud model entropy minimization, the ecological security index is divided into several levels;

[0230] Step S6.2: Draw ecological security index level distribution maps, uncertainty distribution maps, and water and salt stress distribution maps based on the ecological security index levels, and identify high-risk areas, sensitive areas, and key management areas.

[0231] In this embodiment, it also includes:

[0232] Step S7: Based on the multi-source data collected in Step S1, construct a dynamic model of the watershed water resources system, simulate the changing trends of ecological security under different climate, water allocation, ecological water replenishment, and land use scenarios, and use optimization algorithms to back-calculate the minimum combination of management measures required to achieve the target ecological security level.

[0233] In step S7, the watershed water resources system dynamic model includes four modules: water balance, groundwater recovery, salinity migration, and vegetation growth.

[0234] After adopting this technical solution, the model formulas for the four major modules—water balance, groundwater restoration, salinity migration, and vegetation growth—are as follows:

[0235] ,

[0236] ,

[0237] ,

[0238] ;

[0239] In the above formula, This refers to the change in the total water storage of a watershed per unit time. Natural precipitation over a certain period of time. This refers to the inflow of surface water from upstream or external sources. This refers to the artificial regulation of water input, such as ecological water replenishment and inter-basin water transfer. This includes total water loss from soil evaporation and vegetation transpiration. This refers to the amount of surface water infiltrating into the groundwater system. The amount of water discharged downstream or into other basins. This refers to the renewed reserves after replenishment and extraction. This represents the total groundwater storage at time t. This refers to groundwater recharge. This refers to the amount of groundwater extracted. This is due to groundwater evaporation loss. For the next moment, soil salinity; Let be the soil salinity at time t; The capillary rise coefficient; The rinsing coefficient is... It is a capillary rising function; This is the rinsing function. The NDVI value at the next moment; Let be the normalized vegetation index at time t; Let be the vegetation response function, representing the comprehensive response of vegetation growth to water (W) and salinity (S) conditions, and m be the natural degradation coefficient of vegetation.

[0240] The optimization algorithm includes: constructing an objective function with the goal of minimizing management measures. The constraints of the objective function are:

[0241]

[0242] .

[0243] In the formula, For ecological security index; To achieve the target level of ecological security; The variance of the ecological security index; This represents the upper limit of the target variance.

[0244] Example 2

[0245] In this embodiment, an inland river basin in a typical arid region is taken as the object. Its annual precipitation is 50-150 mm, while the evapotranspiration (ET) is as high as 1500-2800 mm. The surface water inflow fluctuates greatly, and the groundwater depth is between 1.5-12 m. In this area, there are problems of groundwater over-extraction and long-term low river inflow. For example, the soil salinization SS is 8-25 g / kg, and the groundwater mineralization is increasing. In some areas, TDS_gw>4 g / L; the area with NDVI<0.15 exceeds 30%, and the vegetation degradation is obvious. Therefore, the ecological condition of this area is extremely sensitive to water regulation and is suitable as a validation object for the WSRS model.

[0246] In this region, following steps S1-S7, data indicators were collected to construct the required data. Three weighting systems—entropy weight, CRITIC, and AHP—were used, and the final weights were determined through game theory combinations. The results are shown in Table 1 below.

[0247] Table 1. Combined weights of an inland river basin in a typical arid region.

[0248]

[0249] In this region, taking a typical cross-section T1 as an example: by retrieving monitoring data from relevant departments and calculating through remote sensing monitoring, the groundwater depth at this cross-section is found to be GW=7.2m, and the groundwater TDS is [missing value]. gw =5.1 g / L, soil salinity SS=18 g / kg, evapotranspiration ratio ET / P=22. According to step 3, the values ​​of α, β, and γ are 0.3, 0.4, and 0.3, respectively. The groundwater depth stress function, groundwater mineralization stress function, soil salinity stress function, and evapotranspiration stress function can be obtained respectively.

[0250]

[0251]

[0252]

[0253]

[0254] Therefore, the combined coercion is:

[0255]

[0256] This indicates that the area is under extremely strong combined stress from water and salt evaporation.

[0257] In this embodiment, the indicator system has been constructed through S1–S3, which includes four dimensions: hydrology and water resources, water and salt evapotranspiration, ecological resilience, and ecological support. The relative fit is obtained through step S4. Cloud model expectations , , This indicates that the uncertainty of the results is low and the model is stable; ultimately, an ecological security index is derived. ;

[0258] In this embodiment, ESI is divided into five levels according to the natural breakpoint method, as shown in Table 2 below:

[0259] Table 2 Classification Table

[0260]

[0261] Therefore, the results indicate that, based on the above calculation of the ecological security index... The results showed that the area was in the Level I range of the five levels, indicating that although some indicators were acceptable, the overall ecological security was at extremely high risk due to groundwater and salinity stress, and water and salinity regulation was urgently needed.

[0262] In this embodiment, following the above steps, in addition to the typical cross-section T1, the typical cross-section T2 is further calculated, and the environmental parameters of point T2 (GW=3.5m, TDS) are input. gw After taking the concentrations (3.2 g / L, SS = 11.0 g / kg, ET / P = 16), the resulting ESI value is 0.37, which falls under the medium ecological security level.

[0263] Comparative Example 1

[0264] In this comparative example, a comprehensive evaluation was conducted on the use of a composite stress model based on water, salt and evapotranspiration data (i.e., Scheme B in Table 3 below) and the use of a composite stress model without water, salt and evapotranspiration data (i.e., Scheme A in Table 3 below). Typical arid watersheds with relatively complete salinization survey and monitoring data were selected, and the measured salinization degree and the distribution of oasis-desert ecotone were used as control groups.

[0265] In this comparative example, a typical highly salinized sample point P1 located on the edge of an oasis within a certain watershed was selected. Its environmental parameters were: NDVI=0.3, soil salinity SS=18g / kg, and groundwater mineralization TDS=4g / L. The evaluation table shown in Table 3 below was obtained:

[0266] Table 3 Evaluation Table for Sample Point P1

[0267]

[0268] As can be seen from Table 3 above, Method A, due to its use of linear weighting, is prone to the phenomenon of good indicators masking bad indicators, leading to misjudgments of areas with high salinity but where vegetation has not completely died. In contrast, Method B, by introducing a stress model, has achieved the identification of key disaster-causing factors and improved the sensitivity of risk identification.

[0269] Secondly, in this comparative study, the identification rate of high salinity areas, the proportion of low / medium salinity areas misclassified as high-risk, and the proportion of oasis-desert transition zones missed were used as indicators to evaluate identification ability. The comparison results are shown in Table 4.

[0270] Table 4 Comparison of models with and without stress

[0271]

[0272] The comparative results show that after introducing the composite stress model, the assessment system overcomes the smoothing effect of the traditional linear weighting method, and significantly improves the ability to identify high-risk salinization areas and oasis-desert ecotones. The identification rate of high-risk areas is improved, the proportion of misjudgment and omission is reduced, and the assessment results are more in line with the objective laws of water and salt constraints in arid areas and the results of actual surveys.

[0273] In this embodiment, method B refers to the composite stress model constructed based on water, salt, and evapotranspiration data according to the present invention.

[0274] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for assessing the ecological security of watersheds in arid regions based on the WSRS framework, characterized in that, Includes the following steps: Step S1: Obtain multi-source data from the target watershed and perform data preprocessing to obtain a multi-source dataset; Step S2: Based on the multi-source dataset obtained in Step S1, construct a WSRS assessment index system including four dimensions: hydrology and water resources, water and salt and evapotranspiration stress, ecological resilience and ecological support, and standardize the indicators of the index system. Step S3: Based on the standardized indicators, determine the comprehensive weight of each indicator using a combination of subjective and objective weighting methods; Step S4: Based on the comprehensive weights, a comprehensive evaluation is performed, and the uncertainty of the evaluation results is quantified using a cloud model to obtain the ecological security index; Step S5: Construct a composite stress model based on water, salt, and evapotranspiration data and correct the ecological security index; Step S6: Classify risk levels and visualize the results based on the revised ecological security index.

2. The method for assessing the ecological security of arid watersheds based on the WSRS framework according to claim 1, characterized in that, In step S1, data preprocessing includes: interpolating missing values, removing outliers, unifying the spatial resolution of multi-source data, and aggregating seasonal indicators annually.

3. The method for assessing the ecological security of arid watersheds based on the WSRS framework according to claim 1, characterized in that, In step S2, the standardization process includes: performing linear forward standardization on positive indices, performing linear reverse standardization on negative indices, and standardizing near-optimal indices using piecewise linear or Gaussian membership functions.

4. The method for assessing the ecological security of arid watersheds based on the WSRS framework according to claim 1, characterized in that, Step S3 includes: Step S3.1: Calculate the objective weights using information entropy weights and CRITIC weights respectively, and determine the subjective weights using the analytic hierarchy process (AHP). Step S3.2: Use game theory to combine and assign weights to the objective and subjective weights to obtain the final weight vector.

5. The method for assessing the ecological security of arid watersheds based on the WSRS framework according to claim 1, characterized in that, Step S4 includes: Step S4.1: Use a weighted normalization matrix to determine the positive and negative ideal solutions; Step S4.2: Calculate the distance from each evaluation unit to the positive and negative ideal solutions using Mahalanobis distance to obtain the proximity score; Step S4.3: Map the proximity to the cloud model to three feature quantities: expectation, entropy, and hyperentropy, to obtain the Ecological Security Index (ESI).

6. The method for assessing the ecological security of arid watersheds based on the WSRS framework according to claim 1, characterized in that, Step S5 includes: Step S5.1: Based on the multi-source dataset obtained in step S1, extract key stress factor data; Step S5.2: Construct the stress function of a single stress factor and perform comprehensive superposition to obtain the comprehensive stress index. Correct the ecological security assessment method through the comprehensive stress index.

7. The method for assessing the ecological security of arid watersheds based on the WSRS framework according to claim 1, characterized in that, Step S6 includes: Step S6.1: Using the natural breakpoint method or the principle of cloud model entropy minimization, the ecological security index is divided into several levels; Step S6.2: Draw ecological security index level distribution maps, uncertainty distribution maps, and water and salt stress distribution maps based on the ecological security index levels, and identify high-risk areas, sensitive areas, and key management areas.

8. A method for assessing the ecological security of arid watersheds based on the WSRS framework according to any one of claims 1-7, characterized in that, Also includes: Step S7: Based on the multi-source data collected in Step S1, construct a dynamic model of the watershed water resources system, simulate the changing trends of ecological security under different climate, water allocation, ecological water replenishment, and land use scenarios, and use optimization algorithms to back-calculate the minimum combination of management measures required to achieve the target ecological security level.

9. A method for assessing the ecological security of arid watersheds based on the WSRS framework according to claim 8, characterized in that, In step S7, the watershed water resources system dynamic model includes four modules: water balance, groundwater recovery, salinity migration, and vegetation growth.

10. A method for assessing the ecological security of arid watersheds based on the WSRS framework according to claim 8, characterized in that, The optimization algorithm includes: constructing an objective function with the goal of minimizing management measures. The constraints of the objective function are: 。

Images