A Method for Evaluating the Effectiveness of River Restoration Based on Multi-Source Data

Abstract

This invention relates to the field of environmental water conservancy and ecological restoration technology, and discloses a method for evaluating the effectiveness of river restoration based on multi-source data. The method includes: collecting target and measure-type indicator data; preprocessing the data; dividing the indicators into time-series and non-time-series types; applying exponential time decay weighting to the time-series indicators to generate time-series characteristic weighted indicators; performing robust standardization of the median absolute deviation with the non-time-series indicators to generate a standardized indicator dataset; calculating the proportion and information entropy of each standardized indicator within each sample to construct the total information entropy; calculating the residual average sum using the equally weighted average as the measured effectiveness value and the initial weighted average as the simulated effectiveness value; constructing a joint loss function; solving for the function with the goal of minimizing the joint loss function to generate the optimal weights for each standardized indicator; and finally calculating a comprehensive restoration effectiveness index to quantitatively evaluate the river restoration effect. This invention enables quantifiable and comparable evaluation of restoration effects in different watersheds and at different times.

Description

Technical Field

[0001] This invention relates to the field of environmental water conservancy and ecological restoration technology, specifically to a method for evaluating the effectiveness of river restoration based on multi-source data. Background Technology

[0002] The assessment of the effectiveness of river ecosystem restoration (repair) is an important basis for watershed water environment management and ecological restoration decisions. With the increasing abundance of multi-source data such as remote sensing monitoring, automatic ground stations, and statistical yearbooks, how to comprehensively utilize these high-dimensional, heterogeneous, and time-series and non-time-series data to scientifically quantify the implementation effects and target achievement of river restoration measures has become a research hotspot in the field of environmental water conservancy.

[0003] Existing methods for assessing the health or recovery effectiveness of rivers typically employ the following two categories:

[0004] One type of evaluation is based on a single indicator or a few key indicators, such as judging the effectiveness by comparing changes in water quality or biodiversity index before and after treatment. However, this type of method is difficult to comprehensively reflect the overall achievement of multi-dimensional recovery goals (such as water quantity, water quality, and aquatic ecology), and is easily affected by the randomness of annual hydrological and meteorological conditions.

[0005] Another type is based on multi-indicator comprehensive evaluation methods, such as using the analytic hierarchy process (AHP) or entropy weight method to assign weights to each indicator and sum them to obtain a comprehensive index. However, these methods generally have the following limitations:

[0006] First, there is a lack of effective processing of time-series data. Existing methods often simplify long-term monitoring time-series data directly to annual averages or values ​​at specific moments, ignoring the differences in the contribution of monitoring data from different periods to the current state of the river. For example, recent water quality improvements should be more representative than earlier data.

[0007] Second, it cannot effectively separate and evaluate the synergistic relationship between "measures" and "goals." Traditional comprehensive evaluations typically weight all indicators (including engineering measures and effectiveness targets) together to obtain a total score. This approach cannot diagnose whether "high effectiveness" is due to effective measures or external factors such as favorable natural hydrological conditions; that is, it cannot reveal the degree of inherent coupling and coordination between governance measures and recovery effectiveness.

[0008] Third, data standardization methods lack robustness. Traditional min-maximum standardization or Z-score standardization is susceptible to outliers, which can significantly distort evaluation results when outliers are present in multi-source data. Summary of the Invention

[0009] To address the aforementioned shortcomings in existing technologies, this invention provides a method for evaluating the effectiveness of river restoration based on multi-source data. This method solves the problems of existing methods, such as difficulty in integrating multi-source time-series data, resistance to outlier interference, and inability to scientifically analyze the synergistic relationship between restoration goals and governance measures.

[0010] To achieve the above-mentioned objectives, the technical solution adopted by this invention is as follows:

[0011] A method for evaluating the effectiveness of river restoration based on multi-source data includes the following steps:

[0012] We collected target-type and measure-type indicator data related to river recovery, preprocessed missing and outlier values, and then divided the data into time-series and non-time-series indicators based on the data collection frequency and time dimension characteristics.

[0013] The time-series index data is weighted by exponential time decay and compressed into a time-series characteristic weighted index reflecting the recent state of the river;

[0014] Robust standardization of the median absolute deviation is performed on time-series weighted indicators and non-time-series indicator data to generate a standardized indicator dataset.

[0015] Based on the standardized indicator dataset, the weight and information entropy of each standardized indicator in each sample are calculated, and the information entropy of all standardized indicators are added together to generate the total information entropy.

[0016] The arithmetic mean of each standardized indicator within each sample is used as the measured performance value for each sample. At the same time, each standardized indicator is assigned an initial weight. The initial weight is then weighted and summed with each standardized indicator within each sample to generate the simulated performance value for each sample. Finally, the sum of squared residuals for all samples is calculated.

[0017] A joint loss function is constructed to maximize the total information entropy and minimize the sum of squared residuals. With minimizing the joint loss function as the objective and the initial weights as the optimization variables, the sequential quadratic programming method is used to solve the problem under the conditions of weight normalization and non-negativity constraints, thereby generating the optimal weights for each standardized index.

[0018] By selecting target-type and measure-type indicators from the standardized indicator dataset, and combining them with optimal weights, the degree of achievement of recovery goals and the degree of implementation of measures are calculated respectively. Then, the coupling degree and the coupling coordination degree are calculated. Finally, the coupling coordination degree is subjected to an S-shaped nonlinear mapping to generate a comprehensive recovery effectiveness index, which is used to quantitatively evaluate the river recovery effect.

[0019] The present invention has the following beneficial effects:

[0020] This invention proposes a method for evaluating the effectiveness of river restoration based on multi-source data. Through multi-source data fusion and time-series decay weighting, it accurately preserves recent river state information while employing robust standardization to enhance data resistance to interference. Based on entropy weighting and joint loss function optimization, it achieves objective dynamic weighting that balances indicator discrimination and fitting accuracy, avoiding subjective weighting bias. Through "target-measure" separation evaluation and coupled coordination analysis, it quantifies restoration effectiveness and diagnoses imbalances such as "high targets, low measures" or "high measures, low targets," overcoming the inability to attribute causes in traditional assessments. Finally, through S-shaped nonlinear mapping, it generates an intuitive percentage-based comprehensive index, enabling quantifiable and comparable evaluation of restoration effects in different watersheds and at different times. This provides scientific, stable, and interpretable decision support for optimizing river governance strategies. Attached Figure Description

[0021] Figure 1 This is a flowchart illustrating a method for evaluating the effectiveness of river restoration based on multi-source data proposed in this invention. Detailed Implementation

[0022] The specific embodiments of the present invention are described below to enable those skilled in the art to understand the present invention. However, it should be understood that the present invention is not limited to the scope of the specific embodiments. For those skilled in the art, various changes are obvious as long as they are within the spirit and scope of the present invention as defined and determined by the appended claims. All inventions utilizing the concept of the present invention are protected.

[0023] A method for evaluating the effectiveness of river restoration based on multi-source data includes the following steps:

[0024] Step 1: Collect target and measure indicator data related to river recovery. After preprocessing for missing and outlier values, classify the data into time-series and non-time-series indicators based on the data collection frequency and time dimension characteristics.

[0025] In this step, target-type indicator data is used to characterize the effectiveness of river restoration, including hydrological and water resource indicators, water environment indicators, and water ecology indicators. These indicators are time-series data, consistent with the operational characteristics of long-term dynamic monitoring and reflecting changes in river status. Specifically, hydrological and water resource indicators include overall river connectivity, river length with water, duration of water flow, ecological flow (i.e., water volume and level) compliance rate, ecological water replenishment, and groundwater level. Water environment indicators include water quality categories at major control sections, water quality compliance rate, and water flow velocity. Water ecology indicators include the proportion of clean and unobstructed river channels, the number of native fish species, fish population index, ecological water supply guarantee rate of nature reserves, and ecological water supply guarantee rate of important wetlands.

[0026] Measures-based indicators are used to characterize the intensity of human intervention, including engineering implementation indicators and management and scheduling indicators; these indicators are non-time-series data. Engineering implementation indicators include the completion rates of ecological flow release facility renovation or construction, river regulation, water system connectivity, groundwater extraction reduction, irrigation district water-saving renovation, and monitoring system construction (such as water conservancy project construction management systems). Management and scheduling indicators include the initial water rights allocation completion rate, ecological scheduling plan execution rate, water withdrawal permit control rate, and annual task ledger completion rate.

[0027] In addition, when preprocessing the collected data for missing and outlier values, linear interpolation is used to fill missing values ​​for time-series indicators, and median is used to fill missing values ​​for non-time-series indicators. Outlier detection employs a dual rule: for data that conforms to a normal distribution, extreme outliers exceeding the mean ± 3 times the standard deviation are removed based on the 3σ principle; for data that does not conform to a normal distribution, box plots are used to identify and remove outliers, eliminating the interference of outlier data on subsequent calculations.

[0028] Based on the preprocessed data, further classification was completed according to the data collection frequency and time dimension characteristics: target indicators reflecting the natural state of rivers, such as hydrology, water resources, water environment, and water ecology, were classified as time-series indicator data; indicators reflecting human governance intervention measures, such as engineering implementation and management scheduling, were classified as non-time-series indicator data, laying the data foundation for the subsequent introduction of a time decay weighting mechanism and the conduct of classification evaluation.

[0029] In summary, this step systematically categorizes multi-source data scattered across different departments such as hydrology, environmental protection, engineering, and management according to two major dimensions: "target-based" and "measure-based." This clarifies the correspondence between governance measures and recovery effectiveness, and solves the problem of mixed data types and difficulty in separating and evaluating them in existing technologies. Furthermore, by dividing data into time-series and non-time-series indicators, it avoids simply compressing long-term monitoring data into averages, preserving the original time dimension information. This provides support for subsequently introducing time-decay weighting to highlight recent river dynamic changes, ultimately achieving the systematic integration and standardized processing of multi-source heterogeneous data, providing comprehensive and reliable data support for subsequent accurate assessment of river recovery effectiveness.

[0030] Step 2: Apply exponential time decay weighting to the time-series index data to compress it into a time-series weighted index that reflects the recent state of the river.

[0031] Specifically, the formula for calculating the exponential time decay weighted average of time-series indicator data is as follows:

[0032]

[0033] In the formula, For the first Time weighting of moments; It is an exponential function; The time decay coefficient is determined using the median half-life method, and the corresponding formula is: , The number of time periods required to halve the set data weights should be consistent with the data collection period for time-series indicators. For example, if the time-series indicator data is monitored on a monthly basis, then... Values ​​can be 3 months, 6 months, etc.; This refers to the total number of time periods for time-series indicator data, i.e., the total number of time-series data cycles involved in the calculation; A weighted index for time-series characteristics; For time series indicators in the first The original value at time.

[0034] In this step, exponential time decay weighting is used to compress time-series indicators into single values ​​reflecting the river's recent state, which meets the need for assessing river recovery effectiveness by emphasizing the near term and downplaying the long term; among these, the time weighting... Using exponential decay And through normalization Ensure that the sum of all weights is 1; Calculation formula for the time-series feature weighted index. It is a standard weighted average form, which can effectively compress time series information.

[0035] In summary, this step employs exponential time decay weighting, compressing long-term monitored time-series indicators into a single feature-weighted indicator. This avoids the loss of time dimension information caused by directly using annual averages and solves the dimensionality curse problem caused by directly inputting data from multiple periods, achieving an effective balance between data compression and information preservation. Furthermore, the exponential decay mechanism ensures that monitoring data closer to the present contributes more to the final indicator, while data further back contributes less. This perfectly aligns with the practical need in river restoration effectiveness assessment that "recent conditions better reflect current governance effectiveness," resolving the assessment lag problem caused by treating historical and recent data equally in traditional methods. In addition, the introduction of half-life... As an adjustable parameter, the time decay coefficient The determination of has a clear physical meaning: This represents the number of time periods required to halve the weight of the representative data. Evaluators can flexibly set this value based on the response characteristics of different indicators. Value – The rapid response indicator is set relatively small. (e.g., 3 months), set the slow response index to a relatively large value. (e.g., 1 year) to make the method adaptable to different scenarios.

[0036] Step 3: Perform robust standardization of the median absolute deviation between the time-series weighted index and the non-time-series index data to generate a standardized index dataset.

[0037] Specifically, the formula for robustly standardizing the median absolute deviation of time-series characteristic-weighted indicators and non-time-series indicator data is as follows:

[0038]

[0039] In the formula, Standardized indicator data; This includes time-series weighted index data or non-time-series index data; For data the median; For data The absolute deviation of the median; To prevent extremely small values, the value is set to 10. -6 To avoid the denominator being 0.

[0040] In summary, this step employs robust standardization to map indicator data of different dimensions and orders of magnitude to the same scale, eliminating the dimensional influence between different unit indicators such as ecological flow (m³ / s), water quality compliance rate (%), and fish species count, thus making subsequent weighted summation and coupling analysis mathematically additive. Meanwhile, traditional Z-score standardization relies on the mean and standard deviation. When outliers exist in multi-source monitoring data (such as sudden pollution events causing abrupt changes in water quality indicators or sensor malfunctions leading to abnormal records), the mean and standard deviation are severely skewed, resulting in distorted standardization results. This step, however, is based on the median and median absolute deviation (MAD). Since the median is insensitive to outliers, even with extreme values ​​in the data, the standardization results can still accurately reflect the distribution characteristics of the main data, significantly improving the robustness of the method. Furthermore, in actual river assessments, there may be situations where all sample values ​​for a certain indicator are exactly the same (e.g., all monitoring sections in a certain area meet the standard for a certain indicator), in which case MAD=0. A zero-prevention minimum is added to the denominator. This avoids division by zero errors, ensures numerical stability in the calculation process, and because The impact on the standardization results of normal data is negligible; ultimately, this standardization method is applied to both the time-series feature weighted index after compression in step two and the original non-time-series index, so that the two types of data enter the same feature space after the same mathematical transformation, laying the foundation for subsequent unified weighting and coupling analysis.

[0041] Step 4: Based on the standardized indicator dataset, calculate the weight and information entropy of each standardized indicator in each sample, and sum the information entropies of all standardized indicators to generate the total information entropy.

[0042] Specifically, the formula for calculating the weight of each standardized indicator and the information entropy within each sample is as follows:

[0043]

[0044]

[0045] In the formula, For the first Within the sample, the first Standardized indicators The proportion of; For the first Within the sample, the first One standardized indicator; The number of samples; For the first The information entropy of a standardized indicator, and the information entropy The larger the value, the more evenly the data is distributed, meaning the less information is contained, and the lower the information entropy. The smaller the value, the greater the difference in the indicator data, which means the greater the amount of information. The correction coefficient is used to adjust the information entropy. Normalize to the [0, 1] interval; It is a logarithmic function.

[0046] In this step, before calculating the weight and information entropy of each standardized indicator within each sample, it is necessary to check whether any of the standardized indicator data contains negative values. If so, all standardized indicator data are non-negatively processed before calculation using the above formula; otherwise, the above formula can be used directly. Non-negativity processing involves uniformly shifting all standardized indicators to address the potential negative values ​​that may be included in the robust standardization of the median absolute deviation. This shift operation only changes the absolute position of the indicator data without altering the data's distribution structure and information content, satisfying the requirement of non-negativity for input data in the entropy weight method, while preserving all statistical information of the original data.

[0047] Therefore, the process of preprocessing all standardized index data by nonnegation is as follows:

[0048] First, obtain the global minimum value among all standardized indicator data, defined as... .

[0049] Secondly, determine the translation constant. The calculation formula is as follows:

[0050]

[0051] in, It is an extremely small positive number, such as 0.001, used to ensure that the minimum value after translation is strictly greater than 0.

[0052] Then, all indicators are uniformly shifted, that is:

[0053]

[0054] in, For the first After nonnegating the nth sample, the i-th One standardized indicator.

[0055] Finally, after nonnegating each standardized indicator, the nonnegated indicator is substituted into the above formula to calculate the weight and information entropy of each standardized indicator in each sample.

[0056] In summary, this step, based on the standardized indicator dataset, quantitatively measures the information content of each standardized indicator through weight calculation and the information entropy formula. This step first performs a coordinate shift on the robust standardization results of the median absolute deviation to ensure that the standardized indicator data are non-negative, and then calculates the proportion of each sample on each standardized indicator. and the information entropy of each standardized indicator The lower the information entropy, the greater the difference between different samples and the richer the identification information contained. This step provides the foundation for the subsequent construction of the total information entropy maximization objective, realizing the objective quantification of the information content of the indicator.

[0057] Step 5: Calculate the arithmetic mean of each standardized indicator within each sample to obtain the measured performance value for each sample. At the same time, assign initial weights to each standardized indicator, and then perform a weighted summation of these initial weights with each standardized indicator within each sample to generate the simulated performance value for each sample. Finally, calculate the sum of squared residuals for all samples.

[0058] Specifically, the formula for calculating the measured performance value of each sample is as follows:

[0059]

[0060] In the formula, For the first The measured performance value of each sample; The total number of standardized indicators for the sample; For the first Within the sample, the first One standardized indicator.

[0061] Specifically, the formula for calculating the simulated performance value for each sample is as follows:

[0062]

[0063] In the formula, For the first Simulated performance values ​​for each sample; The total number of standardized indicators for the sample; For the first The initial weights of each standardized indicator, which can be randomly initialized or set to equal values, are given below. .

[0064] Specifically, the sum of squared residuals for all samples The calculation formula is:

[0065] .

[0066] In summary, this step, based on standardized indicator data, calculates the measured performance value for each sample using an arithmetic mean. At the same time, initial weights are assigned to each standardized indicator. The simulated performance value for each sample is calculated by weighted summation. And further calculate the sum of squared residuals for all samples. An explicit mapping relationship between weights and evaluation results was established, providing a verifiable benchmark and a differentiable optimization objective for subsequent joint loss function optimization, while fully preserving the effective information of the standardized values.

[0067] Step 6: Construct a joint loss function that maximizes the total information entropy and minimizes the sum of squared residuals. With minimizing the joint loss function as the objective and the initial weights as the optimization variables, the sequential quadratic programming method is used to solve the problem under the conditions of weight normalization and non-negativity constraints, generating the optimal weights for each standardized index.

[0068] Specifically, the formula for calculating the joint loss function that maximizes total information entropy and minimizes the sum of squared residuals is as follows:

[0069]

[0070]

[0071] In the formula, For the joint loss function; This is the initial weight matrix for the standardized indicators; The total number of indicators in the sample; For the first Information entropy of a standardized indicator; The total information entropy; The sum of squared residuals for all samples; This is a balancing coefficient used to adjust the relative importance between the two objectives of maximizing information entropy and minimizing fitting error; , , They are the 1st, 2nd, and 3rd respectively. The initial weights of each standardized indicator.

[0072] In this step, the balance coefficient The value of is determined using the order of magnitude balance method, and its calculation formula is:

[0073]

[0074] In the formula, The simulated performance value is calculated using the initial weights. The magnitude balancing method achieves adaptive balance by making the two terms of similar magnitude in the initial state. That is, by making the information entropy term and the residual sum of squares term of similar magnitude in the initial state, it avoids one objective dominating the optimization process and is completely based on data adaptive determination without the need for manual parameter tuning.

[0075] Specifically, the weight normalization and nonnegativity constraints are expressed as follows:

[0076]

[0077] In the formula, For the first The initial weights of each standardized indicator.

[0078] In this step, the objective is to minimize the joint loss function, with the initial weights as the optimization variables. The calculation formula, under the conditions of weight normalization and non-negativity constraints, can be expressed as follows:

[0079]

[0080] in, To perform the minimum value operation, It means to make it satisfy a certain condition.

[0081] In this step, the item This is used to convert maximizing total information entropy into minimizing negative entropy, guiding weights towards indicators with high discriminative power (i.e., the lower the entropy, the greater the information content, and the greater the weight), ensuring that the indicators can effectively identify the recovery levels of different samples; Then, the sum of squared residuals is introduced as a fitting accuracy constraint to guide the weights to converge in the direction of higher fitting accuracy, ensuring that the deviation between the simulated performance value and the measured performance value is minimized.

[0082] In summary, this step constructs a joint loss function that maximizes the total information entropy and minimizes the sum of squared residuals. Under the conditions of weight normalization and non-negativity constraints, it uses a sequential quadratic programming method to solve the problem, generating the optimal weights for each standardized index. Simultaneously, it uses a balancing coefficient... This approach achieves synergistic optimization of information discrimination and fitting accuracy, ensuring that the final weights not only preserve the information differences between indicators but also accurately reflect the actual performance level of the samples, providing high-quality input for subsequent coupled and coordinated analysis.

[0083] Step 7: Select target-type and measure-type indicators from the standardized indicator dataset. Combine the optimal weights to calculate the achievement rate of recovery goals and the implementation rate of measures. Then calculate the coupling degree and coupling coordination degree. Finally, apply an S-shaped nonlinear mapping to the coupling coordination degree to generate a comprehensive recovery effectiveness index, which is used to quantitatively evaluate the river recovery effect. Specifically:

[0084] Filter the target-type and measure-type indicators within the standardized indicator dataset.

[0085] Based on the target indicators and their corresponding optimal weights, the degree of achievement of the recovery target is calculated, i.e.:

[0086]

[0087] In the formula, For the first The degree of achievement of recovery targets for each sample; For the first The optimal weights for each target category indicator; For the first Within the sample, the first Target-type indicators.

[0088] Based on the measures-related indicators and their corresponding optimal weights, the implementation rate of the measures is calculated, i.e.:

[0089]

[0090] In the formula, For the first The implementation rate of measures for each sample; For the first The optimal weights for each measure category indicator; For the first Within the sample, the first Indicators categorized by measures.

[0091] In this step, the degree of achievement of the recovery target is calculated separately. With the degree of implementation of measures This is so that the coupling degree between the two can be calculated further. Coupling Coordination This method enables the separate evaluation and synergistic analysis of "the natural state of the river" and "the degree of human intervention." It can effectively diagnose whether high effectiveness stems from effective measures or favorable natural conditions, solving the problem of traditional comprehensive evaluation conflating objectives and measures and failing to attribute causes.

[0092] Based on the degree of achievement of recovery goals and the degree of implementation of measures, the coupling degree is calculated, namely:

[0093]

[0094] In the formula, For the first The coupling degree of each sample; To prevent extremely small values, the value is set to 10. -6 This is used to avoid the denominator being 0.

[0095] In this step, the coupling degree The quantitative analysis depicted the matching relationship between the achievement of recovery goals and the implementation of measures: when and When the values ​​are similar, the coupling degree is close to 1, indicating good coordination between the measures and the target. When the difference between the two is large, the coupling degree decreases, suggesting an imbalance between "high target, low measures" or "high measures, low target." Meanwhile, the zero minima ensures computational stability: adding a zero-preservation minima to the denominator of the coupling degree... This effectively avoids when and Simultaneously, the computational collapse caused by a zero denominator when the denominator is zero is prevented, thus ensuring the numerical stability of the method.

[0096] Based on the coupling degree, the coupling coordination degree is calculated, i.e.:

[0097]

[0098] In the formula, For the first The degree of coupling coordination of each sample; , Both are contribution coefficients, representing the weighting of the achievement of recovery goals and the implementation of measures in the overall coordinated development of river recovery, and satisfying the following conditions: ,in , The value of is determined using the equal weight method, that is , .

[0099] In this step, the coupling coordination degree Coupling (Reflecting the quality of collaboration) and the comprehensive coordination index (Reflecting the level of development) organically combined, avoiding a situation where a high degree of coupling results in a low level of development (such as...). and The evaluation distortion caused by the fact that all of them are low but equal makes the evaluation results more comprehensive and scientific.

[0100] By applying an S-shaped nonlinear mapping to the coupling coordination degree, a comprehensive recovery effectiveness index with values ​​ranging from 0 to 100 is generated, namely:

[0101]

[0102] In the formula, For the first The comprehensive recovery effectiveness index of each sample; It is an exponential function; This is a nonlinear amplification factor used to adjust the steepness of the S-shaped function. Its value can be adjusted within the range of 5 to 10 according to actual evaluation needs. The larger the value, the better the function. The higher the differentiation in the vicinity.

[0103] In this step, an S-shaped function is used. This maps the coupling coordination degree in the [0, 1] interval to an intuitive score of 0-100. This mapping has the following characteristics:

[0104] (1) Threshold effect: with This is an inflection point; when the score is below 0.5, the score decreases rapidly, and when it is above 0.5, the score increases rapidly.

[0105] (2) Boundary compression: in As the score approaches 0 or 1, the change slows down, which is consistent with the law of diminishing marginal utility.

[0106] Based on the comprehensive recovery effectiveness index, a quantitative assessment of the river recovery effect can be achieved.

[0107] In this step, a higher comprehensive recovery effectiveness index indicates a better river recovery effect, meaning a higher degree of goal achievement, more thorough implementation of measures, and better coupling and coordination between the two. Conversely, a lower index indicates a poorer recovery effect, suggesting problems such as unmet goals, inadequate implementation of measures, or imbalance between the two. Therefore, by calculating this index, the recovery effectiveness of different rivers, at different times, or under different governance schemes can be directly compared, providing a scientific basis for river management.

[0108] In summary, this step calculates the degree of achievement of recovery goals based on the screened target and measure indicators and their optimal weights. With the degree of implementation of measures Then calculate the coupling degree. Coupling Coordination Finally, a comprehensive recovery effectiveness index of 0-100 points is generated through nonlinear mapping using an S-shaped function. This approach enables the separate evaluation and collaborative analysis of objectives and measures. By comprehensively reflecting the matching quality and development level of the two through the coupling coordination degree, it outputs intuitive scores through S-shaped mapping, thus solving the problems of traditional methods being unable to attribute causes and having unintuitive scores.

[0109] Specific embodiments have been used to illustrate the principles and implementation methods of this invention. The descriptions of the embodiments above are only for the purpose of helping to understand the method and core ideas of this invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this invention. Therefore, the content of this specification should not be construed as a limitation of this invention.

[0110] Those skilled in the art will recognize that the embodiments described herein are intended to help the reader understand the principles of the invention, and should be understood that the scope of protection of the invention is not limited to such specific statements and embodiments. Those skilled in the art can make various other specific modifications and combinations based on the technical teachings disclosed in this invention without departing from the spirit of the invention, and these modifications and combinations are still within the scope of protection of this invention.

Claims

1. A method for evaluating the effectiveness of river restoration based on multi-source data, characterized in that, Includes the following steps: We collected target-type and measure-type indicator data related to river recovery, preprocessed missing and outlier values, and then divided the data into time-series and non-time-series indicators based on the data collection frequency and time dimension characteristics. The time-series index data is weighted by exponential time decay and compressed into a time-series characteristic weighted index reflecting the recent state of the river; Robust standardization of the median absolute deviation is performed on time-series weighted indicators and non-time-series indicator data to generate a standardized indicator dataset. Based on the standardized indicator dataset, the weight and information entropy of each standardized indicator in each sample are calculated, and the information entropy of all standardized indicators are added together to generate the total information entropy. The arithmetic mean of each standardized indicator within each sample is used as the measured performance value for each sample. At the same time, each standardized indicator is assigned an initial weight. The initial weight is then weighted and summed with each standardized indicator within each sample to generate the simulated performance value for each sample. Finally, the sum of squared residuals for all samples is calculated. A joint loss function is constructed to maximize the total information entropy and minimize the sum of squared residuals. With minimizing the joint loss function as the objective and the initial weights as the optimization variables, the sequential quadratic programming method is used to solve the problem under the conditions of weight normalization and non-negativity constraints, thereby generating the optimal weights for each standardized index. By selecting target-type and measure-type indicators from the standardized indicator dataset, and combining them with optimal weights, the degree of achievement of recovery goals and the degree of implementation of measures are calculated respectively. Then, the coupling degree and the coupling coordination degree are calculated. Finally, the coupling coordination degree is subjected to an S-shaped nonlinear mapping to generate a comprehensive recovery effectiveness index, which is used to quantitatively evaluate the river recovery effect.

2. The method for evaluating the effectiveness of river restoration based on multi-source data according to claim 1, characterized in that, The target-type and measure-type indicator data associated with river restoration include hydrological and water resources indicator data, water environment indicator data, water ecology indicator data, engineering implementation indicator data, and management and scheduling indicator data; among them, hydrological and water resources indicator data, water environment indicator data, and water ecology indicator data are target-type indicator data, while engineering implementation indicator data and management and scheduling indicator data are measure-type indicator data. Hydrological and water resources indicators include the entire line being connected, the length of the river with water, the duration of water availability, the ecological flow compliance rate, the amount of ecological water replenishment, and the groundwater level; Water environment index data include water quality category, water quality compliance rate, and water flow velocity at major control sections; Water ecological indicators include the proportion of clean and unobstructed waterways, the number of native fish species, the fish population index, the ecological water supply guarantee rate of nature reserves, and the ecological water supply guarantee rate of important wetlands. The project implementation indicators include the completion rate of ecological flow release facility renovation or construction, river regulation completion rate, water system connectivity completion rate, groundwater reduction and extraction completion rate, irrigation area water-saving renovation completion rate, and monitoring system construction completion rate. The management and scheduling indicators include the initial water rights allocation completion rate, ecological scheduling plan execution rate, water abstraction permit control rate, and annual task ledger completion rate.

3. The method for evaluating the effectiveness of river restoration based on multi-source data according to claim 1, characterized in that, The formula for calculating exponential time decay weighted index data is as follows: in, For the first Time weighting of moments It is an exponential function. The time decay coefficient, This represents the total number of time periods for time-series indicator data. As a time-series characteristic weighted index, For time series indicators in the first The original value at time.

4. The method for evaluating the effectiveness of river restoration based on multi-source data according to claim 1, characterized in that, The formula for robustly standardizing the median absolute deviation of time-series weighted indicators and non-time-series indicator data is as follows: in, To standardize indicator data, For time-series weighted index data or non-time-series index data, For data the median; For data The absolute deviation of the median, To prevent extremely small quantities.

5. The method for evaluating the effectiveness of river restoration based on multi-source data according to claim 1, characterized in that, The formula for calculating the weight and information entropy of each standardized indicator within each sample is as follows: in, For the first Within the sample, the first Standardized indicators The proportion of For the first Within the sample, the first A standardized indicator, For the sample size, For the first Information entropy of a standardized indicator For correction factors, It is a logarithmic function.

6. The method for evaluating the effectiveness of river restoration based on multi-source data according to claim 1, characterized in that, The formula for calculating the measured performance value of each sample is as follows: in, For the first The measured performance value of each sample. The total number of standardized indicators for the sample. For the first Within the sample, the first One standardized indicator.

7. The method for evaluating the effectiveness of river restoration based on multi-source data according to claim 1, characterized in that, The formula for calculating the simulated performance value for each sample is as follows: in, For the first The simulated performance value for each sample The total number of standardized indicators for the sample. For the first The initial weights of each indicator, For the first Within the sample, the first One standardized indicator.

8. The method for evaluating the effectiveness of river restoration based on multi-source data according to claim 1, characterized in that, The formula for calculating the joint loss function that maximizes total information entropy and minimizes the sum of squared residuals is as follows: in, For the joint loss function, This is the initial weight matrix for the standardized indicators. The total number of standardized indicators for the sample. For the first Information entropy of a standardized indicator For total information entropy, For the sample size, For the first The measured performance value of each sample. For the first The simulated performance value for each sample The sum of squared residuals for all samples. For balance coefficient, , , They are the 1st, 2nd, and 3rd respectively. The initial weights of each standardized indicator.

9. The method for evaluating the effectiveness of river restoration based on multi-source data according to claim 8, characterized in that, The weight normalization and nonnegativity constraints are expressed as follows: in, For the first The initial weights of each standardized indicator.

10. The method for evaluating the effectiveness of river restoration based on multi-source data according to claim 9, characterized in that, The process involves selecting target-type and measure-type indicators from a standardized indicator dataset, calculating the achievement rate of recovery targets and the implementation rate of measures based on optimal weights, then calculating the coupling degree and coupling coordination degree. Finally, the coupling coordination degree is subjected to an S-shaped nonlinear mapping to generate a comprehensive recovery effectiveness index, which is used to quantitatively evaluate the river recovery effect. Filter target-type and measure-type indicators within the standardized indicator dataset; Based on the target indicators and their corresponding optimal weights, the degree of achievement of the recovery target is calculated, i.e.: in, For the first The degree of achievement of recovery targets for each sample For the first The optimal weights for each target category indicator. For the first Within the sample, the first One target category indicator; Based on the measures-related indicators and their corresponding optimal weights, the implementation rate of the measures is calculated, i.e.: in, For the first The implementation rate of measures for each sample For the first The optimal weights for each measure-type indicator. For the first Within the sample, the first Indicators categorized by measures; Based on the degree of achievement of recovery goals and the degree of implementation of measures, the coupling degree is calculated, namely: in, For the first The coupling degree of each sample, To prevent extremely small quantities; Based on the coupling degree, the coupling coordination degree is calculated, i.e.: in, For the first The degree of coupling coordination of each sample , Both are contribution coefficients, representing the weighting of the achievement of recovery goals and the implementation of measures in the overall coordinated development of river recovery, respectively. By applying an S-shaped nonlinear mapping to the coupling coordination degree, a comprehensive recovery effectiveness index with values ​​ranging from 0 to 100 is generated, namely: in, For the first The comprehensive recovery effectiveness index of the sample It is an exponential function. This is the nonlinear amplification factor; Based on the comprehensive recovery effectiveness index, a quantitative assessment of the river recovery effect can be achieved.

Images