River health evaluation method suitable for northern water-deficient cities

Abstract

The invention discloses a river health evaluation method suitable for northern water-deficient cities, and relates to the technical field of environmental science and engineering.The system comprises the following steps that the pH value, the dissolved oxygen concentration and the total electrolyte amount are monitored through a pH sensor, a dissolved oxygen sensor and a conductivity sensor, the river water surface, the river channel and the surrounding environment are monitored through a GIS, and the river health evaluation result is obtained. Comprising a land utilization type, a vegetation coverage rate and a river bank erosion condition, collecting environmental problems uploaded by the public through a mobile phone APP, and investigating the satisfaction degree of the public to the environment; a river health evaluation system is constructed, the trend of river health is predicted, and potential factors affecting river health are identified; and according to a prediction result, configuring a decision support system to assist in making a management strategy, and regularly reviewing and evaluating system effectiveness. The river health evaluation method and management system are continuously optimized by dynamically adjusting river management and protection work according to a dynamic adjustment result, and the river ecological system of northern water-deficient cities is effectively protected for a long time.

Description

Technical Field

[0001] The present invention relates to the field of environmental science and engineering technology, and particularly to a method for evaluating the health of rivers suitable for water-scarce cities in the north. Background Art

[0002] With the increasing global awareness of ecological environment protection, rivers, as important natural resources and ecosystems, have received growing attention for their health status. Traditional water quality evaluation methods can no longer meet the current needs, and river health evaluation has gradually shifted from a single water quality evaluation to a comprehensive evaluation combining water quality, habitat, and aquatic biological communities. In terms of evaluation content, river health evaluation not only focuses on water quality cleanliness but also attaches importance to multiple dimensions such as environmental safety and ecological beauty, aiming to comprehensively reflect the health status of rivers. At the same time, evaluation indicators have gradually become richer, expanding from physical and chemical indicators to biological indicators, especially benthic animal communities, which have become important references for evaluating river health. In terms of evaluation methods, river health evaluation pays more attention to scientificity and systematicness. By establishing the correlation between habitat factors and biological factors, elements such as hydrological conditions, water quality conditions, and habitat quality are clarified, and a health assessment system is established for each river. In addition, with the development of intelligent and big data technologies, the data collection, processing, and analysis capabilities of river health evaluation have also been significantly improved. In the future, river health evaluation methods will continue to develop in a more refined and intelligent direction.

[0003] Therefore, in response to the two core problems existing in the current river health evaluation method in urban water ecosystems - dysfunction, reduced service function, and weak regeneration and restoration ability, as well as the limitations of the evaluation system, a method for evaluating the health of rivers suitable for water-scarce cities in the north has emerged. This method aims to go beyond the traditional evaluation framework with environmental quality compliance as the single standard and evaluate the health status of rivers more comprehensively and deeply, especially in the context of water-scarce cities in the north. This method integrates ecological principles with the actual needs of water resource management. It not only focuses on basic water quality indicators but also deeply examines the ecological structure and functional integrity of rivers to more accurately reflect the true health status of rivers. By introducing multi-dimensional indicators such as comprehensive water quality index, biodiversity index, and ecological integrity index, this method effectively makes up for the deficiencies in the evaluation of regeneration and restoration ability in the existing evaluation system, providing a scientific basis for urban water ecosystem management. In addition, this method also emphasizes dynamic monitoring and adaptive management. Combining big data and intelligent analysis technologies, it realizes real-time tracking and early warning of river health status, provides timely and accurate information support for decision-makers, promotes the sustainable management and development of urban water ecosystems, better serves the new requirements of ecological environment management, and helps water-scarce cities in the north build a green and harmonious water ecological environment. Summary of the Invention

[0004] In view of the deficiencies of the prior art, the present invention provides a method for evaluating the health of rivers applicable to water-scarce cities in the north, solving the problems in the above-mentioned background art.

[0005] To achieve the above objectives, the present invention is realized through the following technical solutions: A method for evaluating the health of rivers applicable to water-scarce cities in the north, which consists of the following steps:

[0006] S1. At representative water areas and sections, evaluate the current status of water environment quality. Through pH sensors, dissolved oxygen sensors, and conductivity sensors, real-time monitor the pH value, dissolved oxygen concentration, and total electrolyte content. Use GIS to monitor the land use types and vegetation coverage rates of the river water surface, river channels, and their surrounding environments. Through a mobile application, collect environmental problems uploaded by the public;

[0007] S2. Introduce a comprehensive water quality index, biodiversity index, and ecological integrity index to construct a river health evaluation system. Use the long short-term memory network algorithm to predict the trend of river health. Through ecological modeling techniques, identify potential factors affecting river health;

[0008] S3. According to the prediction results, configure a decision support system to assist in formulating management strategies. Through an adaptive control system, regularly review the effectiveness of the river health evaluation system to achieve dynamic adjustment of river management and protection work.

[0009] Furthermore, in the above-mentioned S1, the process of real-time monitoring of the pH value, dissolved oxygen parameters, and conductivity parameters through pH sensors, dissolved oxygen sensors, and conductivity sensors includes:

[0010] Deploy pH sensors, dissolved oxygen sensors, and conductivity sensors on fixed monitoring stations;

[0011] The pH sensor is based on glass membrane electrode technology, measures the hydrogen ion activity in the river, and converts it into an electrical signal. The reference electrode provides a voltage reference to measure the pH value. The dissolved oxygen sensor uses the fluorescence method and measures the dissolved oxygen concentration based on the quenching effect caused by the presence of oxygen after the fluorescent material is illuminated. The conductivity sensor measures the resistance change by applying an alternating voltage to two parallel electrodes and calculates the conductivity according to Ohm's law to monitor the total electrolyte content in the river;

[0012] Preprocess the collected data of pH value, dissolved oxygen concentration, and total electrolyte content. The preprocessing operations include preliminary filtering and denoising to eliminate the influence of short-term fluctuations and random noise, perform linearization correction and temperature compensation to eliminate the influence of air pressure and salinity on dissolved oxygen, and use an outlier detection algorithm to identify and exclude incorrect readings caused by sensor failures and external interferences;

[0013] Extract statistical features from the preprocessed data, where the statistical features include the mean, standard deviation, and maximum and minimum values, to characterize the stability and change trend of water quality parameters.

[0014] Further, in the step S1, the process of using GIS to monitor the land use type and vegetation coverage rate of the river and its surrounding environment includes:

[0015] Utilize the collected data to construct a GIS platform based on cloud services. Use multi-spectral and hyperspectral remote sensing technologies to collect remote sensing images. Distinguish land use types according to the different electromagnetic wave reflection characteristics of ground objects. Calculate the vegetation coverage rate through the normalized difference vegetation index. Conduct time series analysis in combination with historical data sets to track land use and cover changes. Use classification algorithms to improve the land use classification accuracy. Establish a land use and cover change prediction model. Integrate natural factors and socio-economic driving factors to predict the change trend. Perform radiometric correction, atmospheric correction, and geometric correction on the original remote sensing images to eliminate imaging errors. Integrate data with different resolutions and sources to generate comprehensive land use maps and vegetation coverage maps. Use filters to remove outliers and noise. Apply unsupervised classification methods to generate land use classification maps. For vegetation coverage, extract vegetation information through threshold methods. Compare maps of different periods to identify areas of land use and vegetation cover changes, and quantify their area, location, and change rate. Define ecological health indicators, where the ecological health indicators include green space ratio, fragmentation degree, and connectivity, to evaluate the overall ecological environment status of the river basin, analyze long-term change trends, and reveal potential influencing factors.

[0016] Further, in the step S1, the process of collecting environmental problems uploaded by the public through a mobile application includes:

[0017] Develop a mobile application. The public uploads environmental problems in the form of taking pictures, recording videos, and text descriptions. This application integrates the GPS function to automatically record the geographical location where the event occurs, and provides problem type tags, where the problem type tags include water quality anomalies, garbage discharge, and illegal construction, to assist users in describing problems and simplify the background classification process. Establish a two-way communication mechanism to enable users to receive the processing progress and result feedback of the uploaded reports.

[0018] Conduct basic verification on the received reports to check whether they contain geographical location and timestamp information. For incomplete reports, the system prompts users to supplement information. Use geofencing technology and time window algorithms to identify and merge duplicate reports at the same location and time period. Conduct a preliminary quality assessment of the photos and videos submitted by users to make the images clear and the content relevant. Use image recognition technology to automatically screen media files. During the processing, retain anonymized data for analysis.

[0019] Parse the text description provided by the user using natural language processing techniques to extract key information, including the nature of the problem, severity, and possible causes. Analyze the uploaded pictures and videos using computer vision algorithms to automatically identify pollution sources and signs of ecological damage. Combine geographical location and time information to construct a spatio-temporal distribution map, identify hotspots and peak occurrence times, quickly locate key monitoring areas, and long-term track the occurrence frequency and changing trends of different types of problems. Analyze the public's attitude and emotional tendency towards environmental problems by analyzing the tone and vocabulary in the analysis report to assist in understanding social reactions and public concerns.

[0020] Further, in S2, the process of constructing a river health assessment system includes:

[0021] The process of constructing a river health assessment system is divided into four stages, and the four stages include: data integration and standardization stage, index system construction stage, comprehensive scoring model construction stage, and status assessment and classification stage.

[0022] Further, in S2, the process of introducing the comprehensive water quality index, biodiversity index, and ecological integrity index to improve the river health assessment system includes:

[0023] In the data integration and standardization stage, integrate the collected pH value, dissolved oxygen concentration, total electrolyte amount, land use type, vegetation coverage rate, and environmental problems uploaded by the public. For the data of pH value, dissolved oxygen concentration, and total electrolyte amount, perform Z-score standardization to convert them into values between 0 and 1. For land use type and vegetation coverage rate, use one-hot encoding to convert them into numerical forms. When it comes to area ratios, use percentages as the standardized values. For the text description provided by the public, perform word segmentation, remove stop words, and stem extraction, and use TF-IDF text embedding technology to convert the text into a numerical vector. For the uploaded pictures and videos, extract key frames through computer vision algorithms and apply deep learning models to generate fixed-length feature vectors. For data with timestamps, convert the time information into periodic features. For the fluctuating data in the time series, use the sliding window method to calculate the moving average to smooth short-term fluctuations and retain long-term trends. Before standardization, use the Z-score threshold method to identify and remove potential incorrect readings;

[0024] In the stage of constructing the index system, the comprehensive water quality index, biodiversity index and ecological integrity index are introduced as key indicators. The analytic hierarchy process is used for weight allocation. The river health assessment problem is decomposed into an objective layer, a criterion layer and a scheme layer. Among them, the objective layer represents the river health status, the criterion layer represents the key indicators, and the scheme layer represents the specific monitoring areas. For each pair of criteria, experts are invited to score according to their relative importance to construct a pairwise comparison matrix. The maximum eigenvalue and eigenvector of the judgment matrix are solved to obtain the weights of each criterion. The consistency ratio of the judgment matrix is checked to be less than 0.1 to verify the effectiveness of the results;

[0025] In the stage of constructing the comprehensive scoring model, membership functions are defined for each indicator to represent the probability distribution of the indicator values falling into different levels. The different levels include very healthy, healthy, sub-healthy, unhealthy and morbid. According to the membership functions, a fuzzy matrix is constructed to represent the relationship between the scores of each indicator and different levels. Through the method of weighted summation, the comprehensive score is calculated and then mapped to different health levels;

[0026] In the stage of state evaluation and classification, based on historical data, the distribution ranges of each key indicator under different health levels are statistically analyzed to determine the threshold intervals. Experts are invited to adjust and improve the threshold settings according to their experience and professional knowledge to generate a health index map to display the health status of different areas of the river.

[0027] Furthermore, in S2, the process of using the long short-term memory network algorithm to predict the future development trend includes:

[0028] Construct a long short-term memory network model, which includes an input layer, a hidden layer and an output layer;

[0029] The input layer receives the preprocessed pH value, dissolved oxygen concentration and total electrolyte data, and integrates the spatial geographical information of land use type and vegetation coverage rate from the GIS platform, including text descriptions, picture and video feature vectors after natural language processing and computer vision analysis, as well as geographical location and timestamp information. The above multi-source heterogeneous data are integrated into a multi-modal feature matrix as the input of the LSTM;

[0030] The hidden layer processes the multi-source heterogeneous time series data through its internal gating mechanism and memory cell structure. The gating mechanism includes a forget gate, an input gate and an output gate. Among them, the forget gate removes historical information from the memory cell, filters short-term fluctuations and outliers, the input gate adds new information to the memory cell, captures the characteristics of the current environmental conditions, calculates candidate values to generate new information for updating the memory cell state, and the output gate screens the information representing the current river health status, enabling the LSTM model to capture the historical change laws of water quality parameters and their relationship with the environment, and effectively modeling the long-term dependence relationship of river health;

[0031] Based on the memory cell state of the LSTM layer and the result of the output gate, the output layer generates the prediction results of the comprehensive water quality index, biodiversity index, and ecological integrity index through its internal fully connected layer. At the same time, the mean square error and mean absolute error are used to measure the difference between the predicted value and the actual value.

[0032] Further, in the step S2, the process of identifying the potential factors affecting river health through ecological modeling technology includes:

[0033] Extract natural factors and socioeconomic driving factors as auxiliary features, share the hidden layer of the LSTM model, enable mutual learning between different prediction tasks, evaluate the importance of input features to the prediction results through the Shapley value interpretation tool, and identify the factors affecting river health;

[0034] By changing the input feature values, analyze the changes in the prediction results, determine the sensitive factors, clarify the potential impact mechanisms, construct simulation experiments under different scenarios based on historical data and expert knowledge, evaluate the impact of different scenarios on river health. The different scenarios include climate change and policy implementation. Use causal inference methods to extract causal relationships from the LSTM model, understand the interactions between different factors and their impacts on river health, integrate LSTM with other machine learning models, and use stacking generalization to train the model with the outputs of multiple models as new features.

[0035] Further, in the step S3, the process of using a decision support system to assist in formulating management strategies based on the prediction results includes:

[0036] The decision support system receives and integrates the prediction results of the comprehensive water quality index, biodiversity index, and ecological integrity index from the LSTM, integrates the prediction results with the collected original data, constructs a data foundation and simulation experiments under different scenarios, evaluates the potential impacts of various scenarios on river health, conducts sensitivity analysis to determine the impacts of factor changes on river health, identifies high-risk areas and periods, and provides early warnings for emergency response;

[0037] Apply the analytic hierarchy process and fuzzy comprehensive evaluation method, combined with expert knowledge, to quantitatively evaluate different management plans, calculate the comprehensive scores of each plan, and map them to the preset health levels;

[0038] Based on the prediction results and risk assessment, optimize the allocation of water resources, the layout of pollution treatment facilities, and ecological protection, determine the key monitoring areas that need to be prioritized for treatment, establish a two-way communication mechanism, and track the occurrence frequencies and change trends of different types of problems;

[0039] Generate an analysis report covering prediction results, impact factor analysis, and management suggestions to assist decision-makers in understanding information and making decisions.

[0040] Furthermore, in the step S3, the process of periodically reviewing the effectiveness of the evaluation system through an adaptive control system to achieve dynamic adjustment of river management and protection work includes:

[0041] The adaptive control system periodically evaluates the effectiveness of the river health evaluation system through a closed-loop feedback mechanism. In the data collection link, the range of monitoring parameters is expanded, and water temperature sensors, biochemical oxygen demand sensors, chemical oxygen demand sensors, total phosphorus sensors, total nitrogen sensors, and ammonia nitrogen sensors are deployed at fixed monitoring stations, representative waters, and sections to collect water temperature, biochemical oxygen demand, chemical oxygen demand, total phosphorus, total nitrogen, and ammonia nitrogen data in real time. At the same time, referring to the Surface Water Environment Quality Standard (GB3838-2002), priority is given to using the water environment status information uniformly released by the competent department of ecological environment protection to construct an evaluation index system. Based on the collected data and the latest prediction results, the effects of existing management strategies are reviewed regularly to identify potential impact factors and long-term change trends;

[0042] Integrate real-time monitoring data and public reports, use natural language processing and computer vision technologies for effectiveness analysis, and dynamically adjust water resource allocation, pollution treatment facility layout, and ecological protection according to the evaluation results, optimize resource allocation, and continuously improve the adaptive control system;

[0043] Use GIS technology to generate a health index map to display the river health status, and establish a two-way communication mechanism to enable decision-makers to understand and adjust management measures in a timely manner.

[0044] The present invention provides a river health evaluation method applicable to water-scarce cities in the north. It has the following beneficial effects:

[0045] First, the river health assessment method is tailored to the special environment of water-scarce cities in the north. By real-time monitoring of key water quality indicators such as pH value, dissolved oxygen concentration, and total electrolyte content, and by using GIS technology to macroscopically monitor land use and vegetation coverage, it can comprehensively and accurately reflect the health status of the river and its surrounding environment. This not only helps to timely detect potential environmental problems but also provides detailed data support for subsequent treatment work. Second, by introducing the long short-term memory network algorithm for prediction, this method can accurately predict the development trend of river health, providing forward-looking reference information for decision-makers. At the same time, the application of ecological modeling technology further reveals the potential factors affecting river health, providing a scientific basis for formulating targeted management strategies. Moreover, this invention combines a decision support system and an adaptive control system, which can dynamically adjust river management and protection work according to the prediction results and the actual effects of management strategies. This flexible management method not only improves the treatment efficiency but also ensures the continuous effectiveness of the river health assessment system. In summary, the river health assessment method not only improves the accuracy and timeliness of river health monitoring in water-scarce cities in the north but also provides strong technical support for the sustainable management of rivers. Its application will strongly promote the protection and restoration of river ecosystems in water-scarce cities in the north, laying a solid foundation for the sustainable development of cities. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 It is a flowchart of a river health assessment method applicable to water-scarce cities in the north according to an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0047] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0048] As Figure 1 shown, the present invention provides a technical solution: a river health assessment method applicable to water-scarce cities in the north, which consists of the following steps:

[0049] S1. At representative water areas and sections, evaluate the current status of water environment quality. Through pH sensors, dissolved oxygen sensors, and conductivity sensors, real-time monitor the pH value, dissolved oxygen concentration, and total electrolyte content. Use GIS to monitor the land use types and vegetation coverage rates of the river water surface, river channels, and their surrounding environments. Through a mobile application, collect environmental problems uploaded by the public;

[0050] S2. Introduce the comprehensive water quality index, biodiversity index, and ecological integrity index to construct a river health assessment system. Use the long short-term memory network algorithm to predict the trend of river health, and through ecological modeling techniques, identify potential factors affecting river health;

[0051] S3. According to the prediction results, configure a decision support system to assist in formulating management strategies. Through an adaptive control system, regularly review the effectiveness of the river health assessment system to achieve dynamic adjustment of river management and protection work.

[0052] In the above S1, the process of real-time monitoring of pH value, dissolved oxygen parameter, and conductivity parameter through pH sensor, dissolved oxygen sensor, and conductivity sensor includes:

[0053] Deploy pH sensor, dissolved oxygen sensor, and conductivity sensor at fixed monitoring stations;

[0054] The pH sensor is based on glass membrane electrode technology, measures the hydrogen ion activity in the river, and converts it into an electrical signal. The reference electrode provides a voltage reference to measure the pH value. The dissolved oxygen sensor uses fluorescence method, and measures the dissolved oxygen concentration based on the quenching effect caused by the presence of oxygen after the fluorescent material is illuminated. The conductivity sensor measures the resistance change by applying an alternating voltage to two parallel electrodes, and calculates the conductivity according to Ohm's law to monitor the total amount of electrolytes in the river;

[0055] Perform preprocessing operations on the collected data of pH value, dissolved oxygen concentration, and total electrolyte amount. The preprocessing operations include preliminary filtering and denoising to eliminate the influence of short-term fluctuations and random noise, perform linearization correction and temperature compensation to eliminate the influence of air pressure and salinity on dissolved oxygen, and use outlier detection algorithms to identify and exclude incorrect readings caused by sensor failures and external interferences;

[0056] Extract statistical features from the preprocessed data. The statistical features include mean value, standard deviation, maximum and minimum values to characterize the stability and change trend of water quality parameters.

[0057] In the above S1, the process of using GIS to monitor the land use type and vegetation coverage rate of the river and its surrounding environment includes:

[0058] Using the collected data, construct a GIS platform based on cloud services. Utilize multi - spectral and hyperspectral remote sensing technologies to collect remote sensing images. Distinguish land use types according to the different electromagnetic wave reflection characteristics of ground objects. Calculate the vegetation coverage rate through the normalized difference vegetation index. Conduct time - series analysis by combining with historical datasets to track land use and land cover changes. Use classification algorithms to improve the accuracy of land use classification. Establish a prediction model for land use and land cover changes. Integrate natural factors and socio - economic driving factors to predict the change trend. Perform radiometric correction, atmospheric correction, and geometric correction on the original remote sensing images to eliminate imaging errors. Integrate data with different resolutions and sources to generate comprehensive land use maps and vegetation coverage maps. Use filters to remove outliers and noise. Apply unsupervised classification methods to generate land use classification maps. For vegetation coverage, extract vegetation information through threshold methods. Compare maps of different periods to identify the changed areas of land use and vegetation coverage, and quantify their areas, locations, and change rates. Define ecological health indicators, which include green space ratio, fragmentation degree, and connectivity, to evaluate the overall ecological environment status of river basins, analyze long - term change trends, and reveal potential influencing factors.

[0059] In S1, the process of collecting environmental problems uploaded by the public through a mobile application includes:

[0060] Develop a mobile application. The public uploads environmental problems by taking photos, videos, and text descriptions. The application integrates the GPS function to automatically record the geographical location where the event occurs, provides problem type tags, and the problem type tags include abnormal water quality, garbage discharge, and illegal construction, which assist users in describing problems and simplify the background classification process. Establish a two - way communication mechanism so that users can receive the processing progress and result feedback of the uploaded reports.

[0061] Conduct basic verification on the received reports to check whether they contain geographical location and timestamp information. For incomplete reports, the system prompts users to supplement information. Use geofencing technology and time - window algorithms to identify and merge duplicate reports at the same location and time period. Conduct a preliminary quality assessment on the photos and videos submitted by users to make the images clear and content - relevant. Use image recognition technology to automatically screen media files. During the processing, retain anonymized data for analysis.

[0062] Use natural language processing technology to parse the text description provided by the user, extract key information, including the nature of the problem, severity, and possible causes. Apply computer vision algorithms to analyze the uploaded pictures and videos, automatically identify pollution sources and signs of ecological damage. Combine geographical location and time information to construct a spatio-temporal distribution map, identify hotspots and peak periods, quickly locate key monitoring areas, and long-term track the occurrence frequency and change trends of different types of problems. Analyze the public's attitude and emotional tendency towards environmental problems by analyzing the tone and vocabulary in the analysis report to assist in understanding social responses and public concerns.

[0063] In S2, the process of constructing a river health assessment system includes:

[0064] The process of constructing a river health assessment system is divided into four stages, including: data integration and standardization stage, index system construction stage, comprehensive scoring model construction stage, and status assessment and classification stage.

[0065] In S2, the process of improving the river health assessment system by introducing the comprehensive water quality index, biodiversity index, and ecological integrity index includes:

[0066] In the data integration and standardization stage, integrate the collected pH value, dissolved oxygen concentration, total electrolyte amount, land use type, vegetation coverage rate, and environmental problems uploaded by the public. For the data of pH value, dissolved oxygen concentration, and total electrolyte amount, perform Z-score standardization to convert them into values between 0 and 1. For land use type and vegetation coverage rate, use one-hot encoding to convert them into numerical forms. When it comes to area ratios, use percentages as the standardized values. For the text description provided by the public, perform word segmentation, stop word removal, and stemming, and use TF-IDF text embedding technology to convert the text into numerical vectors. For the uploaded pictures and videos, extract key frames through computer vision algorithms and apply deep learning models to generate fixed-length feature vectors. For data with timestamps, convert the time information into periodic features. For the fluctuating data in the time series, use the sliding window method to calculate the moving average to smooth short-term fluctuations and retain long-term trends. Before standardization, use the Z-score threshold method to identify and delete potential incorrect readings;

[0067] In the stage of constructing the index system, the comprehensive water quality index, biodiversity index, and ecological integrity index are introduced as key indicators. The analytic hierarchy process is used for weight allocation. The river health assessment problem is decomposed into an objective layer, a criterion layer, and a scheme layer. Among them, the objective layer represents the river health status, the criterion layer represents the key indicators, and the scheme layer represents the specific monitoring areas. For each pair of criteria, experts are invited to score according to their relative importance to construct a pairwise comparison matrix. The maximum eigenvalue and eigenvector of the judgment matrix are solved to obtain the weights of each criterion. The consistency ratio of the judgment matrix is checked to be less than 0.1 to verify the effectiveness of the results;

[0068] In the stage of constructing the comprehensive scoring model, membership functions are defined for each indicator to represent the probability distribution of the indicator values falling into different levels. The different levels include very healthy, healthy, sub-healthy, unhealthy, and morbid. According to the membership functions, a fuzzy matrix is constructed to represent the relationship between the scores of each indicator and different levels. Through the method of weighted summation, the comprehensive score is calculated and then mapped to different health levels;

[0069] In the stage of status assessment and classification, based on historical data, the distribution ranges of each key indicator at different health levels are statistically analyzed to determine the threshold intervals. Experts are invited to adjust and improve the threshold settings according to their experience and professional knowledge to generate a health index map to display the health status of different regions of the river.

[0070] In S2, the process of using the long short-term memory network algorithm to predict the future development trend includes:

[0071] Construct a long short-term memory network model, which includes an input layer, a hidden layer, and an output layer;

[0072] The input layer receives the preprocessed pH value, dissolved oxygen concentration, and total electrolyte data, and integrates the spatial geographical information of land use type and vegetation coverage rate from the GIS platform, including text descriptions, pictures, and video feature vectors after natural language processing and computer vision analysis, as well as geographical location and timestamp information. The above multi-source heterogeneous data is integrated into a multi-modal feature matrix as the input of the LSTM;

[0073] The hidden layer processes the multi-source heterogeneous time series data through its internal gating mechanism and memory unit structure. The gating mechanism includes a forget gate, an input gate, and an output gate. Among them, the forget gate removes historical information from the memory unit, filters short-term fluctuations and outliers, the input gate adds new information to the memory unit, captures the characteristics of the current environmental conditions, calculates candidate values to generate new information for updating the memory unit state, and the output gate screens the information representing the current river health status, enabling the LSTM model to capture the historical change laws of water quality parameters and their relationship with the environment, and effectively modeling the long-term dependence relationship of river health;

[0074] The output layer generates the prediction results of the comprehensive water quality index, biodiversity index, and ecological integrity index through its internal fully connected layer based on the memory cell state of the LSTM layer and the result of the output gate. At the same time, the mean squared error and mean absolute error are used to measure the difference between the predicted value and the actual value.

[0075] In the above-mentioned S2, the process of identifying potential factors affecting river health through ecological modeling technology includes:

[0076] Extract natural factors and socioeconomic driving factors as auxiliary features, share the hidden layer of the LSTM model, enable mutual reference between different prediction tasks, use the Shapley value interpretation tool to evaluate the importance of input features to the prediction results, and identify factors affecting river health;

[0077] By changing the input feature values, analyze the changes in the prediction results, determine sensitive factors, clarify potential impact mechanisms, construct simulation experiments under different scenarios based on historical data and expert knowledge, evaluate the impact of different scenarios on river health. The above-mentioned different scenarios include climate change and policy implementation. Use causal inference methods to extract causal relationships from the LSTM model, understand the interactions between different factors and their impact on river health, integrate LSTM with other machine learning models, and use stacking generalization to train the model with the outputs of multiple models as new features.

[0078] In the above-mentioned S3, the process of using a decision support system to assist in formulating management strategies based on the prediction results includes:

[0079] The decision support system receives and integrates the prediction results of the comprehensive water quality index, biodiversity index, and ecological integrity index from the LSTM, integrates the prediction results with the collected original data, constructs a data foundation and simulation experiments under different scenarios, evaluates the potential impact of various scenarios on river health, conducts sensitivity analysis to determine the impact of factor changes on river health, identifies high-risk areas and periods, and provides early warnings for emergency response;

[0080] Apply the analytic hierarchy process and fuzzy comprehensive evaluation method, combined with expert knowledge, to quantitatively evaluate different management plans, calculate the comprehensive scores of each plan, and map them to the preset health levels;

[0081] Based on the prediction results and risk assessment, optimize the allocation of water resources, the layout of pollution treatment facilities, and ecological protection, determine the key monitoring areas to be prioritized, establish a two-way communication mechanism, and track the occurrence frequency and change trends of different types of problems;

[0082] Generate an analysis report covering prediction results, impact factor analysis, and management suggestions to assist decision-makers in understanding information and making decisions.

[0083] In S3, the process of regularly reviewing the effectiveness of the evaluation system through an adaptive control system to achieve dynamic adjustment of river management and protection work includes:

[0084] The adaptive control system regularly evaluates the effectiveness of the river health evaluation system through a closed-loop feedback mechanism. In the data collection link, it expands the range of monitoring parameters, and deploys water temperature sensors, biochemical oxygen demand sensors, chemical oxygen demand sensors, total phosphorus sensors, total nitrogen sensors, and ammonia nitrogen sensors at fixed monitoring stations, representative waters, and sections to collect water temperature, biochemical oxygen demand, chemical oxygen demand, total phosphorus, total nitrogen, and ammonia nitrogen data in real time. At the same time, referring to the Surface Water Environment Quality Standard (GB3838-2002), it preferentially adopts the water environment status information uniformly released by the competent department of ecological environment protection to construct an evaluation index system. Based on the collected data and the latest prediction results, it regularly reviews the effects of existing management strategies, identifies potential influencing factors and long-term change trends;

[0085] Integrate real-time monitoring data and public reports, use natural language processing and computer vision technologies for effectiveness analysis, and dynamically adjust water resource allocation, pollution treatment facility layout, and ecological protection according to the evaluation results, optimize resource allocation, and continuously improve the adaptive control system;

[0086] Use GIS technology to generate a health index map to display the river health status, and establish a two-way communication mechanism to enable decision-makers to timely understand and adjust management measures.

[0087] It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article or device. Without further limitation. An element defined by the statement "comprising an..." does not exclude the presence of additional identical elements in the process, method, article or device comprising the element.

[0088] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for evaluating the health of rivers applicable to water-scarce cities in the north, characterized in that Including the following steps: S1. Evaluate the current status of water environment quality at representative water areas and sections. Through pH sensors, dissolved oxygen sensors, and conductivity sensors, real-time monitor the pH value, dissolved oxygen concentration, and total electrolyte content. Use GIS to monitor the land use types and vegetation coverage rates of the river surface, river channels, and their surrounding environments. Through a mobile application, collect environmental problems uploaded by the public; S2. Introduce comprehensive water quality indices, biodiversity indices, and ecological integrity indices to construct a river health assessment system. Use the long short-term memory network algorithm to predict the trend of river health. Through ecological modeling techniques, identify potential factors affecting river health; S3. According to the prediction results, configure a decision support system to assist in formulating management strategies. Through an adaptive control system, regularly review the effectiveness of the river health assessment system to achieve dynamic adjustment of river management and protection work.

2. The river health assessment method applicable to water-scarce cities in the north according to claim 1, wherein: In the above S1, the process of real-time monitoring of the pH value, dissolved oxygen parameters, and conductivity parameters through pH sensors, dissolved oxygen sensors, and conductivity sensors includes: Deploy pH sensors, dissolved oxygen sensors, and conductivity sensors at fixed monitoring stations; The pH sensor is based on glass membrane electrode technology, measures the hydrogen ion activity in the river, and converts it into an electrical signal. The reference electrode provides a voltage reference to measure the pH value. The dissolved oxygen sensor uses the fluorescence method and measures the dissolved oxygen concentration based on the quenching effect caused by the presence of oxygen after the fluorescent material is illuminated. The conductivity sensor measures the resistance change by applying an alternating voltage to two parallel electrodes and calculates the conductivity according to Ohm's law to monitor the total electrolyte content in the river; Perform preprocessing operations on the collected data of pH value, dissolved oxygen concentration, and total electrolyte content. The preprocessing operations include preliminary filtering and denoising to eliminate the influence of short-term fluctuations and random noise, perform linearization correction and temperature compensation to eliminate the influence of air pressure and salinity on dissolved oxygen, and use an outlier detection algorithm to identify and exclude incorrect readings caused by sensor failures and external interferences; Extract statistical features from the preprocessed data. The statistical features include mean value, standard deviation, maximum and minimum values, to characterize the stability and change trend of water quality parameters.

3. A method for evaluating river health applicable to water-scarce cities in the north, as claimed in claim 2, wherein: In the above S1, the process of using GIS to monitor the land use types and vegetation coverage rates of the river and its surrounding environments includes: Using the collected data, a GIS platform based on cloud services is constructed. Remote sensing images are collected using multi-spectral and hyper-spectral remote sensing technologies. Land use types are distinguished according to the different characteristics of the reflected electromagnetic waves of ground objects. The vegetation coverage rate is calculated through the normalized difference vegetation index. Time series analysis is carried out in combination with historical data sets to track the changes in land use and land cover. Classification algorithms are used to improve the accuracy of land use classification. A prediction model for land use and land cover changes is established. By integrating natural factors and socio-economic driving factors, the change trend is predicted. Radiometric correction, atmospheric correction, and geometric correction are performed on the original remote sensing images to eliminate imaging errors. Data with different resolutions and sources are fused to generate comprehensive land use maps and vegetation coverage maps. Filters are used to remove outliers and noise. Unsupervised classification methods are used to generate land use classification maps. For vegetation coverage, vegetation information is extracted through threshold methods. By comparing maps of different periods, the areas of land use and vegetation cover changes are identified, and their areas, locations, and change rates are quantified. Ecological health indicators are defined, and the ecological health indicators include green space ratio, fragmentation degree, and connectivity, to evaluate the overall ecological environment status of the river basin, analyze long-term change trends, and reveal potential influencing factors.

4. The river health assessment method applicable to water-scarce northern cities according to claim 3, characterized in that: In S1, the process of collecting environmental problems uploaded by the public through a mobile application includes: Develop a mobile application. The public uploads environmental problems by taking pictures, recording videos, and text descriptions. The application integrates the GPS function to automatically record the geographical location where the event occurs, provides problem type tags, and the problem type tags include abnormal water quality, garbage discharge, and illegal construction, to assist users in describing problems and simplify the background classification process. A two-way communication mechanism is established to enable users to receive the processing progress and result feedback of the uploaded reports. Perform basic verification on the received reports to check whether they contain geographical location and timestamp information. For incomplete reports, the system prompts users to supplement information. Using geofencing technology and time window algorithms, identify and merge duplicate reports at the same location and time period. Conduct a preliminary quality assessment of the photos and videos submitted by users to make the images clear and the content relevant. Use image recognition technology to automatically screen media files. During the processing, anonymized data is retained for analysis. Use natural language processing technology to parse the text descriptions provided by users and extract key information. The key information includes the nature of the problem, severity, and possible causes. Use computer vision algorithms to analyze the uploaded pictures and videos to automatically identify pollution sources and signs of ecological damage. Combine geographical location and time information to construct a spatio-temporal distribution map, identify hotspots and peak occurrence periods, quickly locate key monitoring areas, and long-term track the occurrence frequency and change trends of different types of problems. By analyzing the tone and vocabulary in the reports, analyze the public's attitude and emotional tendency towards environmental problems to assist in understanding social responses and public concerns.

5. The river health assessment method applicable to water-scarce cities in the north according to claim 4, wherein: In S2, the process of constructing a river health assessment system includes: The process of constructing a river health assessment system is divided into four stages, and the four stages include: data integration and standardization stage, index system construction stage, comprehensive scoring model construction stage, and status evaluation and classification stage.

6. The river health assessment method applicable to water-scarce cities in the north according to claim 5, wherein: In S2, the process of introducing the comprehensive water quality index, biodiversity index, and ecological integrity index to improve the river health assessment system includes: Data integration and standardization stage: Integrate the collected pH value, dissolved oxygen concentration, total electrolyte amount, land use type, vegetation coverage rate, and environmental problems uploaded by the public. For the data of pH value, dissolved oxygen concentration, and total electrolyte amount, perform Z-score standardization to convert them into values between 0 and 1. For land use type and vegetation coverage rate, use one-hot encoding to convert them into numerical forms. When it comes to area ratios, use percentages as the standardized values. For the text descriptions provided by the public, perform word segmentation, stop word removal, and stemming, and use TF-IDF text embedding technology to convert the text into numerical vectors. For the uploaded pictures and videos, extract key frames through computer vision algorithms and apply deep learning models to generate fixed-length feature vectors. For data with timestamps, convert the time information into periodic features. For the fluctuating data in the time series, use the sliding window method to calculate the moving average to smooth short-term fluctuations and retain long-term trends. Before standardization, use the Z-score threshold method to identify and remove potential error readings; Index system construction stage: Introduce the comprehensive water quality index, biodiversity index, and ecological integrity index as key indicators, and use the analytic hierarchy process for weight assignment. Decompose the river health assessment problem into an objective layer, a criterion layer, and a scheme layer. Among them, the objective layer represents the river health status, the criterion layer represents the key indicators, and the scheme layer represents the specific monitoring areas. For each pair of criteria, invite experts to score according to their relative importance to construct a pairwise comparison matrix, solve the maximum eigenvalue and eigenvector of the judgment matrix to obtain the weights of each criterion, and check the consistency ratio of the judgment matrix to make it less than 0.1 to verify the effectiveness of the results; Comprehensive scoring model construction stage: Define membership functions for each indicator to represent the probability distribution of the indicator values falling into different levels. The different levels include very healthy, healthy, sub-healthy, unhealthy, and morbid. Construct a fuzzy matrix according to the membership functions to represent the relationship between the scores of each indicator and different levels. Calculate the comprehensive score through weighted summation and then map it to different health levels; Status evaluation and classification stage: Based on historical data, statistically analyze the distribution ranges of each key indicator at different health levels to determine the threshold intervals. Invite experts to adjust and improve the threshold settings according to their experience and professional knowledge, and generate a health index map to display the health status of different areas of the river.

7. A method for river health assessment applicable to water-scarce cities in the north, as claimed in claim 6, wherein: In S2, the process of using the long short-term memory network algorithm to predict future development trends includes: Construct a long short-term memory network model, and the long short-term memory network model includes an input layer, a hidden layer, and an output layer; The input layer receives the preprocessed pH value, dissolved oxygen concentration, and total electrolyte data, integrates the spatial geographical information of land use type and vegetation coverage rate from the GIS platform, including text descriptions, picture and video feature vectors after natural language processing and computer vision analysis, as well as geographical location and timestamp information, and integrates the above multi-source heterogeneous data into a multi-modal feature matrix as the input of the LSTM; The hidden layer processes the multi-source heterogeneous time series data through its internal gating mechanism and memory unit structure. The gating mechanism includes a forget gate, an input gate, and an output gate. Among them, the forget gate removes historical information from the memory unit, filters short-term fluctuations and outliers, the input gate adds new information to the memory unit, captures the characteristics of the current environmental conditions, calculates candidate values to generate new information for updating the memory unit state, and the output gate filters the information representing the current river health status, enabling the LSTM model to capture the historical change laws of water quality parameters and their relationships with the environment, and effectively modeling the long-term dependence relationship of river health; The output layer generates the prediction results of the comprehensive water quality index, biodiversity index, and ecological integrity index according to the memory unit state of the LSTM layer and the results of the output gate. At the same time, the mean square error and mean absolute error are used to measure the difference between the predicted value and the actual value.

8. A method for evaluating river health applicable to water-scarce cities in the north, as claimed in claim 7, wherein: In step S2, the process of identifying potential factors affecting river health through ecological modeling techniques includes: Extracting natural factors and socioeconomic driving factors as auxiliary features, sharing the hidden layer of the LSTM model, enabling mutual reference between different prediction tasks, and evaluating the importance of input features to the prediction results through the Shapley value interpretation tool to identify factors affecting river health; By changing the input feature values, analyzing the changes in the prediction results, determining sensitive factors, clarifying potential impact mechanisms, constructing simulation experiments under different scenarios based on historical data and expert knowledge, evaluating the impact of different scenarios on river health, where the different scenarios include climate change and policy implementation, using causal inference methods to extract causal relationships from the LSTM model, understanding the interactions between different factors and their impacts on river health, integrating LSTM with other machine learning models, and training the model with the outputs of multiple models as new features through stacking generalization.

9. The river health assessment method applicable to water-scarce northern cities according to claim 8, characterized in that: In step S3, the process of using a decision support system to assist in formulating management strategies according to the prediction results includes: The decision support system receives and integrates the prediction results of the comprehensive water quality index, biodiversity index, and ecological integrity index from the LSTM, integrates the prediction results with the collected original data, constructs a data basis and simulation experiments under different scenarios, evaluates the potential impacts of various scenarios on river health, conducts sensitivity analysis to determine the impacts of factor changes on river health, identifies high-risk areas and time periods, and provides early warnings for emergency responses; Applying the analytic hierarchy process and fuzzy comprehensive evaluation method, combined with expert knowledge, quantitatively evaluating different management plans, calculating the comprehensive scores of each plan, and mapping them to the preset health levels; Optimize water resource allocation, pollution treatment facility layout, and ecological protection based on prediction results and risk assessment, determine key monitoring areas that need to be prioritized for treatment, establish a two-way communication mechanism, and track the occurrence frequency and changing trends of different types of problems; Generate an analysis report covering prediction results, impact factor analysis, and management suggestions to assist decision-makers in understanding information and making decisions.

10. A method for evaluating river health applicable to water-scarce cities in the north, as claimed in claim 9, characterized in that: In S3, the process of realizing the dynamic adjustment of river management and protection work by regularly reviewing the effectiveness of the evaluation system through an adaptive control system includes: The adaptive control system regularly evaluates the effectiveness of the river health evaluation system through a closed-loop feedback mechanism. In the data collection link, expand the range of monitoring parameters, deploy water temperature sensors, biochemical oxygen demand sensors, chemical oxygen demand sensors, total phosphorus sensors, total nitrogen sensors, and ammonia nitrogen sensors at fixed monitoring stations and representative waters and sections to collect water temperature, biochemical oxygen demand, chemical oxygen demand, total phosphorus, total nitrogen, and ammonia nitrogen data in real time. Based on the collected data and the latest prediction results, regularly review the effects of existing management strategies, identify potential influencing factors and long-term changing trends; Integrate real-time monitoring data and public reports, conduct effectiveness analysis using natural language processing and computer vision technologies, dynamically adjust water resource allocation, pollution treatment facility layout, and ecological protection according to the evaluation results, optimize resource allocation, and continuously improve the adaptive control system; Use GIS technology to generate a health index map to display the river health status, establish a two-way communication mechanism, and enable decision-makers to timely understand and adjust management measures.

Images