Environmental risk step-by-step assessment method under multi-medium multi-scale complex scene

Abstract

The invention discloses an environmental risk step-by-step assessment method in a multi-medium and multi-scale complex scene, and belongs to the technical field of environmental risk assessment. The method comprises the following steps: obtaining the heavy metal concentration of the waste residue through a leaching experiment, and determining the toxicity grade by using a single pollution index method and a potential ecological risk index method; dividing volume grades based on quantiles, dividing local diffusion risk grades by using an inverse distance weighted interpolation method and the like, dividing remote receptor sensitivity grades according to a surrounding land and water source construction model, and generating priority control grades by integrating the grades; and finally, fusing multi-source data to construct a spatio-temporal conjoint analysis model verification result. Therefore, by integrating pollution source toxicity, volume, diffusion risk and receptor sensitivity to carry out risk step-by-step assessment, a multi-source data fusion and spatio-temporal conjoint analysis method is proposed for the first time, and the problems of single assessment dimension, insufficient data fusion, lack of dynamic verification and the like of an existing assessment method in a multi-medium and multi-scale scene are effectively solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of environmental risk assessment, and specifically relates to a method for gradually assessing environmental risks in a multi-medium and multi-scale complex scenario. Background Art

[0002] With the long-term development of mineral resources, the problem of heavy metal pollution caused by the stockpiling of historical legacy waste residues has become increasingly prominent, posing a potential threat to the surrounding soil, water bodies and ecological environment. At present, there has been some research on the environmental risk assessment method for waste residues in mining areas. For example, the invention patent with the application number CN202210922204.5 proposes a method for assessing the comprehensive environmental pollution risk of abandoned mining areas. This assessment method is based on the analytic hierarchy process to grade and score the pollution sources and sensitivity indicators of abandoned mining areas. However, many factors such as the toxicity, volume of pollution sources and multi-medium diffusion are not fully considered in its assessment system. Therefore, it is difficult to reflect the associated risks of pollution migration and receptor exposure in complex scenarios. In addition, the invention patent with the application number CN202110962898.0 discloses a method for analyzing the environmental risk of pollutants based on multi-dimensional evaluation factors. This method optimizes the calculation of risk factors by introducing the environmental persistence coefficient. However, this method mainly relies on chemical analysis data and lacks the integration of spatial information such as remote sensing images and multi-temporal monitoring. There is also a problem of insufficient spatio-temporal dynamics in risk assessment.

[0003] In the prior art, there are still certain limitations in the comprehensive assessment in a multi-medium and multi-scale scenario at the present stage. For example, the invention patent with the application number CN202310489715.7 uses a three-dimensional neural network interpolation model to predict the distribution of heavy metal concentrations. Although it can improve the accuracy of spatial analysis, it still does not consider the joint assessment of pollution source toxicity, local diffusion and remote receptor sensitivity. The verification of this model mainly relies on static sampling data. Therefore, it is difficult to support dynamic risk early warning. At the same time, traditional methods generally ignore the quantitative impact of pollution source volume on the risk level. For example, the relevance between the waste residue accumulation scale and the pollutant release potential has not been systematically incorporated into the assessment system, resulting in inaccurate division of control priorities and making it difficult to effectively guide pollution control and ecological restoration.

[0004] Based on the above analysis, the existing assessment methods have the following main problems: First, the assessment dimensions are limited to single indicators, and key levels such as the toxicity intensity of pollution sources, the spatial volume distribution, the multi-media migration and diffusion laws, and the differences in receptor sensitivity are not systematically integrated; Second, the effectiveness of multi-source heterogeneous data fusion is insufficient, and there are certain technical barriers in the analysis between remote sensing interpretation results, on-site sampling data, and historical monitoring information; Third, the risk assessment model lacks a dynamic feedback verification mechanism, resulting in a systematic deviation between theoretical predictions and the actual diffusion paths of pollutants. Therefore, how to design a new environmental risk assessment method through multi-dimensional feature coupling and dynamic simulation technology to accurately identify high-risk areas and formulate a scientific control priority division strategy has become one of the technical problems to be solved urgently at present. Summary of the Invention

[0005] Aiming at the deficiencies of the existing assessment methods in the background technology, the present invention integrates the toxicity, volume, diffusion risk of pollution sources, and receptor sensitivity, and through multi-source data fusion and model verification, realizes the step-by-step assessment and accurate classification of risks, providing a scientific basis for priority control in environmental management. At the same time, by constructing a multi-dimensional assessment index system, introducing a spatio-temporal joint analysis model, and establishing comprehensive control rules, the comprehensiveness and accuracy of risk assessment in complex scenarios are solved.

[0006] Based on this, the present invention provides an environmental risk step-by-step assessment method in a multi-media and multi-scale complex scenario.

[0007] To achieve the above object, the present invention adopts the following technical solutions: An environmental risk step-by-step assessment method in a multi-media and multi-scale complex scenario, which includes the following steps: Step 1 Pollution source toxicity assessment: Obtain the heavy metal concentration data in the waste residue through a leaching experiment, calculate the single pollution risk level by using the single pollution index method and evaluate the comprehensive ecological risk level by using the potential ecological risk index method respectively, and determine the pollution source toxicity level according to the higher value of the two risk levels; Step 2 Pollution source volume assessment: Divide the volume level based on the quantile value of the pollution source volume to identify the pollution sources that need to be prioritized for control; Step 3 Local diffusion risk assessment: Use the inverse distance weighted interpolation method to perform spatial interpolation on the pollutant concentrations in soil, water, and agricultural products, and combine the weighted calculation results of the direct diffusion risk index and the indirect diffusion risk index to divide the local diffusion risk level; Step 4 Remote diffusion receptor sensitivity assessment: Analyze the land use types and water source distribution within a 2-kilometer buffer zone around the pollution source to construct a sensitivity index model, calculate the sensitivity index of all pollution sources through the quantile value, and divide the remote receptor sensitivity level; Step 5 Comprehensive risk assessment: The toxicity level, volume level, local diffusion risk level, and remote receptor sensitivity level obtained in Steps 1 to 4 are weighted and summed according to preset weights to generate the priority control level of the pollution source. Step 6 Multi-source data fusion and model verification: Multitemporal high-resolution remote sensing images, on-site sampling data, and historical environmental monitoring data are fused to construct a spatio-temporal joint analysis model to verify the reliability of the risk assessment results.

[0008] As a further supplementary explanation of the above technical solution, in Step 1, according to PI i The single pollution risk level is divided and mapped into five levels. The calculation formula of the single pollution index method is: , where PI i is the single pollution index, C i is the measured concentration of the i th heavy metal, S i is the i th environmental quality standard limit value corresponding to the heavy metal; In Step 1, according to RI The comprehensive ecological risk level is divided and mapped into five levels. The calculation formula of the potential ecological risk index method is: , where , in the formula: is the single potential risk index of the i th heavy metal, is the toxicity coefficient of the i th heavy metal; Based on the larger numerical level in the single pollution risk level and the comprehensive ecological risk level as the pollution source toxicity level, that is , where L PI , L RI are the 1-5 numerical levels of the single pollution risk and the comprehensive ecological risk respectively. The larger the value, the higher the risk.

[0009] As a further supplementary explanation of the above technical solution, the specific method for dividing the volume level based on the quantile value of the pollution source volume in Step 2 is as follows: Arrange the waste residue volumes of all blocks in ascending order to form a pollution source volume dataset, calculate the 25%, 50%, 75%, and 90% quantile values of the pollution source volume dataset, and divide the pollution source volume into five levels according to the quantile values. The calculation formula of the quantile value by linear interpolation method is: , where n is the total number of data, k is the quantile point, X is the sorted dataset.

[0010] As a further supplementary description of the above technical solution, the calculation formula of the inverse distance weighted interpolation method in step 3 is: , weight , where Z ( x 0 ) is the predicted value of the pollutant concentration at the unknown point ，Z ( x j ) is the pollutant concentration at the j-th sampling point, d j is the Euclidean distance between the unknown point and the j -th sampling point ，p is the distance attenuation coefficient ， The value range is from 1 to 3.

[0011] As a further supplementary description of the above technical solution, in step 3, according to RDI 总 value, the local diffusion risk is divided into five levels, and the calculation formula of the comprehensive diffusion risk index is: RDI 总 = RDI 直接 ·0.5 + RDI 间接 ·0.5, where RDI 直接 is the diffusion risk index of soil and water media, RDI 间接 is the diffusion risk index of agricultural product media.

[0012] As a further supplementary description of the above technical solution, the calculation formula of the sensitivity index model in step 4 is: SI = A 1 ×5 + A 2 ×3 + A 3 ×1 + A 4 ×(7 + N ), where SI represents the sensitivity index; A 1 、A 2 、A 3 、A 4 are the area ratios of main agricultural land, other agricultural land, residential land and water area respectively, and their area ratios are calculated based on the total area of the 2-kilometer buffer zone around the pollution source; Nis the number of water sources in the buffer zone, 7+N is the water area ratio coefficient; Arrange the sensitivity indexes of all pollution sources in ascending order to form a sensitivity index data set. Calculate the 25%, 50%, 75%, and 90% quantiles of the sensitivity index data set according to the linear interpolation calculation formula in step 3, and divide the sensitivity index into five levels according to the quantile values.

[0013] As a further supplement to the above technical solution, the weight allocation rule in step 5 is that the weight of the pollution source toxicity level is 40%, the weight of the volume level is 20%, the weight of the local diffusion risk level is 25%, and the weight of the remote receptor sensitivity level is 15%. The comprehensive risk value is obtained by weighted summation, and the priority control level is divided into five levels according to the comprehensive risk value. The comprehensive risk value calculation formula is as follows: CR=W1·G1+W2·G2+W3·G3+W4·G4, wherein W1, W2, W3, and W4 represent the pollution source toxicity level weight, volume level weight, local diffusion risk level weight, and remote receptor sensitivity level weight, respectively; G1, G2, G3, and G4 represent the standardized scoring values of the pollution source toxicity level, volume level, local diffusion risk level, and remote receptor sensitivity level, respectively, and the value range is 0-10.

[0014] As a further explanation and limitation of the above technical solution, the multi-source data fusion and model verification in step 6 specifically includes the following steps: Step 6.1 Data acquisition and preprocessing: geometric correction, radiation correction, noise removal, band fusion and uniform color mosaic are performed on remote sensing images to ensure that the spatial resolution is not less than 1 meter; the Z-score standardization method is used for sampling data to eliminate dimensional differences; linear interpolation is used for historical monitoring data to fill missing values, and invalid data with a time span of more than 10 years is eliminated; Step 6.2 Feature extraction and fusion: extract the normalized vegetation index, normalized water index and texture entropy features from the remote sensing image; extract the pollutant concentration gradient features and spatial clustering features from the sampling data; extract the annual average growth rate and seasonal fluctuation characteristics of pollutant concentration from the historical data, and construct a multidimensional feature matrix; Step 6.3: Construction of spatiotemporal joint analysis model: Using spectral index, pollutant concentration gradient, and historical trend as input variables, the random forest algorithm is used, the number of decision trees is set to 100-500, the maximum depth is 5-15 layers, and Gini impurity is used as the node splitting criterion. The parameters of split feature selection strategy, minimum number of leaf node samples, and split threshold tolerance are optimized through cross-validation; Model output: Pollution diffusion probability P diffusion RSS S sensitiVity The comprehensive risk weight is calculated by the following formula: , where α is the diffusion probability weight coefficient, β is the sensitivity score weight coefficient, and the coefficients satisfy the constraint condition α + β = 1, α The value range is 0.6 - 0.8, β The value range is 0.2 - 0.4; Step 6.4 Model verification: Calculate the Pearson correlation coefficient, root mean square error, and Kappa coefficient through the cross-validation method to verify the model prediction accuracy and correct the risk level classification results.

[0015] As a further supplementary explanation of the above technical solution, the following indicators are used in the step 6.4 model verification: The calculation formula for the Pearson correlation coefficient is: , where x i is the measured concentration value of the pollutant at the i th sampling point, y i is the model predicted concentration value at the corresponding point, , are the average values of the measured and predicted values respectively, n is the total number of samples; The calculation formula for the root mean square error is: , where x i is the measured concentration value of the pollutant at the i th sampling point, y i is the model predicted concentration value at the corresponding point; Kappa The calculation formula for the coefficient is: where N is the total number of samples, m is the number of risk level categories, O kk is the number of correctly classified samples in the k th risk level category, R k and C k are the total number of samples in the k th risk level category in the manual interpretation result and the model output result respectively, Kappa The closer the

[0016] As a further supplementary description of the above technical solution, the output results of the spatio-temporal joint analysis model in step 6 are respectively used to dynamically update the threshold for classifying the volume level of pollution sources, optimize the distance attenuation coefficient in the inverse distance weighting interpolation method, generate a risk heat map, and identify the priority control level.

[0017] Compared with the prior art, the present invention has the following remarkable advantages: 1. The present invention comprehensively considers key factors such as the toxicity, volume, local diffusion risk of pollution sources, and the sensitivity of remote receptors. By combining the single pollution index method and the potential ecological risk index method to evaluate the toxicity of pollution sources, it can comprehensively reflect the hazards of single and multiple pollutants; using the quantile value division technique to evaluate the volume of pollution sources, and collaborating with other evaluation indicators, it realizes the full-chain risk classification from pollution sources to receptor exposure. At the same time, with scientific and reasonable weight allocation and quantitative evaluation, the comprehensiveness and accuracy of the evaluation results are further improved.

[0018] 2. The present invention first proposes a multi-source data fusion and spatio-temporal joint analysis method, deeply fuses remote sensing images, sampling data, and historical monitoring data, constructs a spatio-temporal joint analysis model, and verifies and corrects it by calculating indicators such as the Pearson correlation coefficient, root mean square error, and Kappa coefficient through the cross-validation method, further enhancing the accuracy and reliability of environmental risk assessment, and providing a solid technical support for dynamic risk early warning in complex scenarios.

[0019] 3. The present invention constructs a remote receptor sensitivity assessment model based on the land use type and water source distribution within a 2-kilometer buffer zone around the pollution source, accurately identifying potential risks in sensitive areas such as drinking water source protection areas and farmlands. At the same time, an objective weight allocation and comprehensive risk calculation mechanism is established to avoid the subjectivity of traditional weight setting, providing a quantifiable scientific basis for environmental management decisions and facilitating the formulation of targeted environmental management and protection measures.

[0020] 4. The environmental risk hierarchical assessment method of the present invention is not limited to the environmental risk assessment of mine waste residues. Its multi-dimensional assessment idea, data fusion method, and model construction method are universal. It can be widely applied to other fields such as environmental pollution risk assessment in chemical industrial parks and environmental risk assessment around urban landfills, opening up new ways for risk assessment in different scenarios, providing new technical means and research ideas. Description of the Drawings

[0021] Figure 1 It is a flowchart of the environmental risk hierarchical assessment method in the present invention.

[0022] Figure 2 It is a flowchart of the pollution local diffusion risk assessment method in the embodiment of the present invention.

[0023] Figure 3This is a schematic diagram of the risk level of local pollution diffusion in the embodiments of the present invention.

[0024] Figure 4 This is a schematic diagram of the environmental risk heat in the embodiments of the present invention. Detailed implementation manners

[0025] In order to further elaborate the technical solution of the present invention, the present invention will be further described below through embodiments.

[0026] An environmental risk step-by-step assessment method in a multi-medium and multi-scale complex scenario mainly realizes the full-chain risk classification from the pollution source to the receptor through the coupling assessment of local diffusion and remote receptors, the dynamic fusion of multi-source data, and the adaptive adjustment of weights.

[0027] As shown in the appendix Figure 1 The assessment method includes the following steps: Step 1: Pollution source toxicity assessment: Obtain the heavy metal concentration data in the waste residue through a leaching experiment, calculate the single pollution risk level by using the single pollution index method and evaluate the comprehensive ecological risk level by using the potential ecological risk index method, and determine the pollution source toxicity level according to the higher value of the two risk levels; Step 2: Pollution source volume assessment: Divide the volume level based on the quantile value of the pollution source volume to identify the pollution sources that need to be preferentially controlled; Step 3: Local diffusion risk assessment: Use the inverse distance weighted interpolation method to perform spatial interpolation on the pollutant concentrations in soil, water, and agricultural products, and combine the weighted calculation results of the direct diffusion risk index and the indirect diffusion risk index to divide the local diffusion risk level; Step 4: Remote diffusion receptor sensitivity assessment: Analyze the land use types and water source distribution within a 2-kilometer buffer zone around the pollution source to construct a sensitivity index model, calculate the sensitivity index of all pollution sources through the quantile value, and divide the remote receptor sensitivity level; Step 5: Comprehensive risk assessment: Weight and sum the toxicity level, volume level, local diffusion risk level, and remote receptor sensitivity level obtained in Steps 1 to 4 according to the preset weights to generate the priority control level of the pollution source; Step 6: Multi-source data fusion and model verification: Integrate multi-temporal high-resolution remote sensing images, on-site sampling data, and historical environmental monitoring data to construct a spatio-temporal joint analysis model. The output results of the spatio-temporal joint analysis model are respectively used to dynamically update the threshold for dividing the pollution source volume level, optimize the distance attenuation coefficient in the inverse distance weighted interpolation method, generate a risk heat map, and identify the priority control level, so as to verify the reliability of the risk assessment results.

[0028] In this embodiment, the calculation formula of the single pollution index method is: , where PI iis the single pollution index, C i is the i measured concentration of the S i th i heavy metal, and its value refers to the screening value of the corresponding construction land type in GB 36600-2018 "Soil Environmental Quality - Risk Control Standards for Soil Pollution of Construction Land".

[0029] According to PI i divide the single pollution risk level and map it into the following five levels: safe level, warning level, light pollution, moderate pollution and heavy pollution.

[0030] Table 1 Classification Criteria for Assessing Heavy Metal Pollution Risk Levels by the Single Pollution Index Method

[0031] In step 1, according to RI divide the comprehensive ecological risk level and map it into five levels. The calculation formula of the potential ecological risk index method is: , where In the formula: is the single potential risk index of the i th heavy metal, is the toxicity coefficient of the i th heavy metal. This toxicity coefficient is determined with reference to GB 36600-2018 "Soil Environmental Quality - Risk Control Standards for Soil Pollution of Construction Land" and toxicological research, and the values are: arsenic 10, cadmium 30, hexavalent chromium 2, copper 5, lead 5, mercury 40, nickel 5.

[0032] According to RI divide the comprehensive ecological risk level and map it into the following five levels: slight harm, medium harm, high harm, relatively high harm and extremely high harm.

[0033] Table 2 Classification Criteria for Assessing Heavy Metal Pollution Risk Levels by the Potential Ecological Risk Index Evaluation Method

[0034] Take the larger numerical level in the single pollution risk level and the comprehensive ecological risk level as the toxicity level of the pollution source, that is , where, L PI , L RI are the 1-5 level numerical levels of the single pollution risk and the comprehensive ecological risk respectively. The larger the value, the higher the risk.

[0035] The specific toxicity levels and numerical ranges of the pollution sources are as follows: Safety level: The numerical range is PI ≤0.7 and RI <150, corresponding to both the single pollution risk level and the comprehensive ecological risk level being level 1, with the lowest risk.

[0036] Warning level: The numerical range is 0.7 < PI ≤1.0 or 150 ≤ RI <300, and the larger numerical level of the two is level 2, indicating the existence of potential pollution risks and the need to strengthen monitoring.

[0037] Light pollution level: The numerical range is 1.0 < PI ≤2.0 or 300 ≤ RI <600, and the larger numerical level of the two is level 3, with a certain pollution impact, and preliminary control measures need to be taken.

[0038] Moderate pollution level: The numerical range is 2.0 < PI ≤3.0 or 600 ≤ RI <800, and the larger numerical level of the two is level 4, with a relatively high pollution risk, and targeted treatment is required.

[0039] Severe pollution level: The numerical range is PI >3.0 or RI ≥800, and the larger numerical level of the two is level 5, with an extremely high pollution risk, and priority emergency control is required.

[0040] In this embodiment, the accumulation of waste residues will occupy a large area of land, which not only affects the utilization of land resources but also becomes a source of environmental pollution. The larger the volume of waste residues, the greater the harm to the environment and human health. Therefore, it is necessary to comprehensively consider various factors such as pollutant toxicity, pollutant leaching amount, and waste residue volume to more accurately evaluate the potential ecological risks of waste residue pollution sources. The specific method for dividing the volume level based on the quantile value of the pollution source volume in step 2 is as follows: Arrange the waste residue volumes of all blocks in ascending order to form a pollution source volume data set, calculate the 25%, 50%, 75%, and 90% quantile values of the pollution source volume data set, and divide the pollution source volume into five levels according to the quantile values. The calculation formula for the quantile value by linear interpolation method is: , where n is the total number of data, k is the quantile point, X is the sorted data set; calculate the 25%, 50%, 75%, and 90% quantile values of the pollution source volume data set, and obtain P 25 , P 50 , P 75 , P90 As a threshold, according to the quantile value P k The source volume is divided into the following five levels: extremely small volume V < P 25 , relatively small volume P 25 ≤ V <P 50 , medium volume, P 50 ≤ V <P 75 , relatively large volume P 75 ≤ V <P 90 , extremely large volume V ≥P 90 . Thus, the blocks that need to be preferentially controlled due to the large volume of waste residue are identified.

[0041] As shown in the appendix Figure 2 , the migration and diffusion of heavy metal pollution in waste residue in the medium are divided into two cases according to the medium type. The first is the migration and diffusion in natural environmental media such as soil, surface water, groundwater, and sediment. The second is the migration and diffusion from natural environmental media to agricultural products related to human health. At the same time, five typical polluting heavy metals, namely mercury, cadmium, arsenic, lead, and chromium, are selected as the regional heavy metal pollution diffusion risk assessment indicators. For direct pollution, the detection data of four natural environmental media, namely soil, surface water, groundwater, and sediment, are selected for calculating the direct pollution level. Since the detection data of different samples have different units, the single pollution index method is used to calculate the pollution level of each type of sample respectively to eliminate the influence of dimension differences, so as to form unified dimensionless pollution level data, and the single pollution level of heavy metals in natural environmental media of mercury, cadmium, arsenic, lead, chromium, and characteristic pollutants is calculated respectively. For indirect pollution, the detection data of agricultural product samples are selected for calculating the indirect pollution level, and the single pollution index method is used to calculate the agricultural product pollution level. The direct pollution levels of five heavy metals in agricultural products, namely mercury, cadmium, arsenic, lead, and chromium, are calculated respectively. According to the characteristics of the attenuation of heavy metal pollution diffusion with distance, the calculation formula of the inverse distance weight interpolation method is selected as: , weight , Among them, Z ( x 0 )is the predicted value of the pollutant concentration at the unknown point ，Z ( x j )is the pollutant concentration at the jth sampling point, d j is the Euclidean distance between the unknown point and the j th sampling point ，p is the distance attenuation coefficient ，The value range is from 1 to 3. For the direct and indirect pollution environmental media in the study area, regional interpolation is carried out for mercury, cadmium, arsenic, lead, and chromium one by one to obtain the spatial distribution of the pollution indices of different types of pollutants. Since it is the spatial interpolation of the risk levels of single pollutants in different environmental media, rather than the interpolation of pollutant concentrations, the interpolation can reflect the pollution risks of single pollutants in local areas.

[0042] The construction of the local pollution diffusion risk index consists of two parts: direct diffusion risk and indirect diffusion risk. Among them, direct diffusion mainly includes the direct diffusion risk of heavy metals in environmental media under natural actions such as soil wind erosion, surface runoff, groundwater migration, and sediment deposition. Indirect diffusion is mainly caused by human activities within and around the polluted area, and is the diffusion risk of heavy metals in agricultural products. The calculation method of the local pollution diffusion risk index is as follows: RDI 总 = RDI 直接 ·0.5 + RDI 间接 ·0.5, where RDI 直接 is the diffusion risk index of soil and water media, RDI 间接 is the diffusion risk index of agricultural product media; RDI 总 The RDI value divides the local diffusion risk into the following five levels: no diffusion RDI ≤1.0, mild diffusion 1.0 < RDI ≤2.0, moderate diffusion 2.0 < RDI ≤3.0, severe diffusion 3.0 < RDI ≤4.0 and extremely severe diffusion

[0043] In this embodiment, the calculation formula of the sensitivity index model in step 4 is: SI = A 1 ×5 + A 2 ×3 + A 3 ×1 + A 4 ×(7 + N ), where SI represents the sensitivity index; A 1 、A 2 、A 3 、A 4The area ratios of prime agricultural land, other agricultural land, residential land, and water area respectively, and the area ratios are calculated based on the total area of the 2-kilometer buffer zone around the pollution sources; N It is the number of water source areas in the buffer zone. 7 + N is the water area ratio coefficient, which is set according to the sensitivity weights of water source area protection in HJ338-2018 "Technical Specification for the Division of Drinking Water Source Protection Areas". Here, N is the number of water source areas in the buffer zone. For each additional water source area, the sensitivity weight is increased by 7 units to strengthen the priority of drinking water safety protection.

[0044] Arrange the sensitivity indices of all pollution sources in ascending order to form a sensitivity index data set, and calculate the 25%, 50%, 75%, and 90% quantile values of the sensitivity index data set according to the above linear interpolation method calculation formula to obtain P 25 ´ 、 P 50 ´ 、 P 75 ´ 、 P 90 ´ 。

[0045] According to P k ´ The quantile values, divide the sensitivity indices into five levels: Level Ⅰ: extremely high sensitivity SI ≥P 90 ´ 、Level Ⅱ: high sensitivity P 75 ´ ≤ SI < P 90 ´ 、Level Ⅲ: medium sensitivity P 50 ´ ≤ SI < P 75 ´ 、Level Ⅳ: low sensitivity P 25 ´ ≤ SI < P 50 ´ 、Level Ⅴ: extremely low sensitivity SI < P 25 ´ 。

[0046] In this embodiment, step 5 is determined based on historical monitoring data regression analysis and expert Delphi method. We invited 15 environmental experts to conduct three rounds of anonymous questionnaire consultation, focusing on the impact of pollution source toxicity, volume, diffusion risk and receptor sensitivity on actual risk events, and obtaining the importance score of each indicator. At the same time, multiple linear regression was performed based on 100 sets of historical monitoring data, and the weight adjustment formula was: , verify the correlation between each indicator and the actual risk event, and combine the expert opinions to finally determine the weights as 40% for toxicity, 20% for volume, 25% for local diffusion, and 15% for remote sensitivity, and meet the following constraints: W1+W2+W3+W4=100%. The comprehensive risk value is calculated by the following formula: CR=W1·G1+W2·G2+W3·G3+W4·G4, where W1, W2, W3, and W4 represent the pollution source toxicity level weight, volume level weight, local diffusion risk level weight, and remote receptor sensitivity level weight, respectively; G1, G2, G3, and G4 represent the standardized scoring values of the pollution source toxicity level, volume level, local diffusion risk level, and remote receptor sensitivity level, respectively, with a value range of 0-10.

[0047] The priority control level is divided into five levels according to the comprehensive risk value: extremely high risk CR ≥ 8.0, high risk 6.0 ≤ CR < 8.0, medium risk 4.0 ≤ CR < 6.0, low risk 2.0 ≤ CR < 4.0, and extremely low risk CR < 2.0. The higher the priority control level, the greater the relative risk, and it should be included in the environmental management and risk control list first.

[0048] In this embodiment, the multi-source data fusion and model verification in step 6 specifically include the following steps: Step 6.1 Data acquisition and preprocessing: geometric correction, radiation correction, noise removal, band fusion and uniform color mosaic are performed on remote sensing images to ensure that the spatial resolution is not less than 1 meter; the Z-score standardization method is used for sampling data to eliminate dimensional differences; linear interpolation is used for historical monitoring data to fill missing values, and invalid data with a time span of more than 10 years is eliminated; Step 6.2 Feature extraction and fusion: extract the normalized vegetation index, normalized water index and texture entropy features from the remote sensing image; extract the pollutant concentration gradient features and spatial clustering features from the sampling data; extract the annual average growth rate and seasonal fluctuation characteristics of pollutant concentration from the historical data, and construct a multidimensional feature matrix; Step 6.3 Spatiotemporal Joint Analysis Model Construction: Using spectral index, pollutant concentration gradient, and historical trend as input variables, the random forest algorithm is adopted, with the number of decision trees set to 100 - 500 and the maximum depth to 5 - 15 layers. Using Gini impurity as the node splitting criterion, the parameters of the splitting feature selection strategy, minimum number of samples in leaf nodes, and splitting threshold tolerance are optimized through cross - validation; The model outputs the pollution diffusion probability P diffusion and the receptor sensitivity score S sensitiVity , and the comprehensive risk weight is calculated through the following formula:,[[]]END]] , where α is the diffusion probability weight coefficient, β is the sensitivity score weight coefficient, and the coefficients satisfy the constraint condition α + β = 1, α with a value range of 0.6 - 0.8, β with a value range of 0.2 - 0.4; Step 6.4 Model Validation: Calculate the Pearson correlation coefficient, root - mean - square error, and Kappa coefficient through the cross - validation method to verify the model prediction accuracy and correct the risk level classification results. The specific model validation uses the following indicators: The Pearson correlation coefficient calculation formula is: , where x i is the measured pollutant concentration value at the i th sampling point, y i is the model - predicted concentration value at the corresponding point, , are the average values of the measured and predicted values respectively, n is the total number of samples; The Pearson correlation coefficient is used to measure the linear correlation degree between the model - predicted value and the measured value. The closer its value is to 1, the higher the correlation between the model - predicted value and the measured value.

[0049] The root - mean - square error calculation formula is: , where x i is the measured pollutant concentration value at the i th sampling point, y i is the model - predicted concentration value at the corresponding point. The root - mean - square error is used to quantify the deviation degree between the model - predicted value and the measured value. The smaller its value, the higher the model prediction accuracy.

[0050] Kappa The coefficient calculation formula is as follows: , wherein, N is the total number of samples, m is the number of risk level categories, O kk is the number of correctly classified samples of the k th risk level, R k and C k are respectively the total number of samples of the k th risk level in the manual interpretation result and the model output result, Kappa The closer the coefficient value is to 1, the higher the consistency between the model classification result and the manual interpretation result.

[0051] Next, we will elaborate in detail on the specific implementation process of the environmental risk step-by-step assessment method in a multi-medium and multi-scale complex scenario to ensure that each technical feature in the claims can be fully supported by the specification.

[0052] This implementation project takes the legacy waste residue pile in a historical mining area in Hainan Province as the research object. There are heavy metal pollution risks such as cadmium, mercury, arsenic, lead, and chromium in this area. Farmland, residential areas, and surface water bodies are distributed within a radius of 2 kilometers around. It is necessary to conduct environmental risk assessments for various environmental media such as soil, surface water, groundwater, and agricultural products at different spatial scales of 500 meters locally and 2 kilometers remotely.

[0053] I. Basic data collection and preprocessing Remote sensing image data: Collect satellite images with a resolution of 0.3 - 2 meters from 2000 to 2022. After geometric correction, radiometric correction, and band fusion processing, extract the normalized difference vegetation index and normalized difference water index features for waste residue pile boundary identification and land use type classification.

[0054] On-site sampling data: For waste residue piles with an area > 100 m², distribute sampling points according to the 50 m × 50 m grid method. A total of 50 waste residue leachate samples, 80 soil samples, 60 agricultural product samples, 40 surface water samples, and 30 sediment samples are collected. The sampling points of the waste residue pile are evenly distributed on the surface of the waste residue pile, and the soil sampling points are set at different distances around the waste residue pile. Samples of waste residue leachate, soil, agricultural products, surface water, and sediment are collected to detect heavy metal concentrations. The detection results are as follows: The Cd concentration in the waste residue leachate is 150 μg / L, and the Hg concentration is 50 μg / L; the Cd concentration in the surrounding soil is 25 mg / kg, and the Cd concentration in rice grains is 0.3 mg / kg.

[0055] Historical monitoring data: Soil and surface water heavy metal concentration data from 2013 to 2022 were collated, missing values were filled using linear interpolation, invalid data with a time span of more than 10 years were eliminated, and the average annual growth rate and seasonal fluctuation characteristics were extracted.

[0056] 2. Toxicity Assessment of Pollution Sources Single pollution index method: refer to the screening value of construction land in GB36600-2018, Cd is taken as 50μg / L, Hg is taken as 0.1μg / L. According to the calculation formula: , calculate PI Cd =3.0, i.e. moderate Cd pollution, level 4; PI Hg =500; Hg severe pollution, level 5, therefore, the single pollution risk level of this area takes the highest value of level 5.

[0057] Potential ecological risk index method: According to the calculation formula: , the toxicity coefficient refers to GB36600-2018 and toxicological research, and Cd=30, Hg=40 are selected, that is, RI=20090 in this case, which is far beyond the "extremely high hazard" threshold, and the comprehensive ecological risk level is level 1. Finally, based on the toxicity level of the pollution source, the higher value of the single item and the comprehensive level is taken, and the toxicity of the pollution source in this area is determined to be level 5 severe pollution.

[0058] 3. Assessment of the volume of pollution sources We counted the volumes of 100 waste slag piles in the area, arranged them in ascending order, and calculated the 25%, 50%, 75%, and 90% quantiles, i.e., P 25 =200m 3 , P 50 =500m 3 , P 75 =1000m 3 , P 90 =2000m 3 The threshold is determined by the linear interpolation formula. The target waste slag pile volume in this area is 1500m 3 Located in P 75 -P 90 The interval is classified as “large volume”, indicating that it has a higher potential for pollutant release.

[0059] IV. Local Spread Risk Assessment We use the inverse distance weighted interpolation method to analyze the spatial distribution of pollutants in soil, water and agricultural products and calculate the diffusion risk index: , weight Get the distance attenuation coefficient p =2, generate a 50m resolution pollution distribution map, such as Figure 3As shown, Figure (a) is a schematic diagram of the direct diffusion PI value and pollution level of this block, and Figure (b) is a schematic diagram of the indirect diffusion PI value and pollution level of this block. From this, it can be identified that the area proportion of the high-risk area of direct diffusion is 20%. The direct diffusion risk and the indirect diffusion risk each account for 50% of the weight. According to the calculation method of the pollution local diffusion risk index: RDI 总 =RDI 直接 ·0.5 + RDI 间接 ·0.5, the direct diffusion RDI 直接 = 3.6, the indirect diffusion RDI 间接 = 3.2, calculating RDI 总 = 3.4. It can be seen that this area belongs to medium diffusion, that is, corresponding to level 3.

[0060] V. Remote receptor sensitivity assessment Analyze the land use types and water source distribution within 2 kilometers, construct a buffer sensitivity index model, and calculate the sensitivity index according to the following formula: SI = A 1 ×5 + A 2 ×3 + A 3 ×1 + A 4 ×(7 + N ), where A1 is 40%, A2 is 20%, and N is the number of water source areas, which is 3. The calculated SI = 4. Sort all the pollution source SI values, and determine the target SI is located in the P 50 ´ - P 75 ´ interval, and determine that this area belongs to medium sensitivity, classified as level 3, indicating that the exposure risk of the remote receptor is medium.

[0061] VI. Comprehensive risk assessment Based on the regression analysis of historical monitoring data and the Delphi method of experts, determine the weights of toxicity 40%, volume 20%, local diffusion 25%, and sensitivity 15%. Map each level to get toxicity level 5 → 10 points, volume level 4 → 8 points, local diffusion level 3 → 6 points, sensitivity level 3 → 6 points. According to the formula, calculate CR = 0.4×10 + 0.2×8 + 0.25×6 + 0.15×6 = 8.0, corresponding to "extremely high risk", classified as priority control level I. Therefore, it is currently recommended that risk control measures should be immediately implemented for the remaining waste residue piles in this historical mining area.

[0062] VII. Multi-source data fusion and model verification After geometric correction, the resolution of the remote sensing image is ≤1 meter. The sampling data is standardized by Z-score. Missing values in the historical data are filled and invalid data is removed. Three-dimensional features such as NDVI, NWI, and texture entropy are extracted from the remote sensing image. Two-dimensional features such as concentration gradient and spatial clustering are extracted from the sampling data, and two-dimensional features such as annual average growth rate and seasonal fluctuation are extracted from the historical data to construct a seven-dimensional feature matrix. The random forest algorithm is used for model construction, with 300 decision trees set, a maximum depth of 10 layers, grid search to optimize the node splitting criterion, the minimum number of samples in the leaf node is 5, the splitting threshold tolerance is 0.01, the output pollution diffusion probability is 0.85, the receptor sensitivity score is 4.0, and the comprehensive risk weight W = 0.7×0.85 + 0.3×4.0 = 1.795 is calculated. Through cross-validation calculation R = 0.91, RMSE = 0.9, and Kappa = 0.82. Thus, it can be confirmed that the model has high prediction accuracy and reliable risk assessment results.

[0063] During the verification process of this model, from the pollution diffusion probability of 0.85 output by the model, it can be judged that the existing volume of waste residue has a relatively high impact on the risk. There is a significant positive correlation between the pollution diffusion probability and the volume of waste residue. Since the waste residue pile with a volume of 1500m 3 is evaluated as an extremely high risk, its management priority needs to be improved. Therefore, we adjusted the original 90th percentile value from 2000m 3 to 1800m 3 . For the optimization of the distance attenuation coefficient, by comparing the interpolation accuracies of different p values, when p = 1, 2, 3, the root mean square errors are calculated to be 1.2, 0.9, and 1.1 respectively. Finally, it is determined that p = 2 is the optimal parameter. Thus, it can be seen that the model after optimizing the distance attenuation coefficient can more accurately predict the pollution diffusion range, and then guide the formulation of more effective environmental restoration plans. As Figure 4 shown, the environmental risk heat map in this paper has removed the satellite map information and only retains the longitude and latitude markings. The longitude and latitude are used to assist in positioning the geographical location of the risk area. Different color areas represent different environmental risk levels. For example, red represents extremely high risk (CR≥8.0), and yellow represents high risk (6.0≤CR<8.0). Through this map, the distribution differences of environmental risks in different longitude and latitude areas can be intuitively seen, helping relevant departments and personnel quickly locate high-risk areas, reasonably allocate environmental management resources, and formulate scientific and effective environmental risk management plans.

[0064] In summary, through this method, the waste residue pile is accurately identified as a priority-controlled Class I pollution source, and the assessment results are highly consistent with on-site rapid detection and historical image interpretation. The model verification indicators show that this method can be applied to the risk assessment and management of historical mining areas.

[0065] The main features and advantages of the present invention have been shown and described above. For those skilled in the art, it is obvious that the specific implementation manners of the present invention are not limited to the details of the above exemplary embodiments, and various changes and improvements can be made without departing from the spirit or basic features of the present invention. For example, new variables such as wind speed and precipitation pattern can be added to the feature matrix to further improve the adaptability of the model to climate factor changes; or the parameters of the random forest algorithm can be fine-tuned to adapt to the differences in the environmental characteristics of specific regions. In addition, geographic information system technology can be combined to achieve the visual display of risk assessment results, assisting decision-makers to allocate resources and intervene in risks more efficiently. Therefore, from any perspective, the embodiments should be regarded as exemplary and non-restrictive. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be included in the present invention.

[0066] In addition, it should be understood that although this specification is described according to embodiments, not every embodiment only contains an independent technical solution, but may contain a combination of multiple technical solutions, and these technical solutions can be independent of each other or related to each other. For example, when dealing with multiple pollution sources, the risks of each pollution source can be evaluated separately, and then the results can be integrated to achieve the comprehensive management of the pollution risk in the entire region. Such a processing method can not only control precisely, but also provide more comprehensive data support when formulating environmental policies. In this way, we can ensure the effectiveness and adaptability of environmental risk assessment and management, thereby protecting the ecological environment and promoting sustainable development. Furthermore, in order to improve the generalization ability of the model under different environmental conditions, the cross-validation technology in machine learning can be introduced. This narrative way of the specification is only for clarity. Those skilled in the art should regard the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

Claims

1. An environmental risk step-by-step assessment method in a multi-medium and multi-scale complex scenario, characterized in that, The following steps are involved: Step 1: Toxicity assessment of pollution sources: obtain the heavy metal concentration data in the waste residue through leaching experiments, calculate the single pollution risk level using the single pollution index method, and evaluate the comprehensive ecological risk level using the potential ecological risk index method, and determine the toxicity level of the pollution source based on the higher value of the two risk levels; Step 2: Pollution source volume assessment: The volume level is divided based on the quantile value of the pollution source volume to identify the pollution sources that need to be controlled first; Step 3: Local diffusion risk assessment: Use the inverse distance weighted interpolation method to spatially interpolate the pollutant concentrations in soil, water and agricultural products, and combine the weighted calculation results of the direct diffusion risk index and the indirect diffusion risk index to divide the local diffusion risk level; Step 4: Remote diffusion receptor sensitivity assessment: Analyze the land use types and water source distribution within the 2 km buffer zone around the pollution source to build a sensitivity index model, calculate the sensitivity index of all pollution sources through quantile values and divide the remote receptor sensitivity level; Step 5 Comprehensive risk assessment: The toxicity level, volume level, local diffusion risk level and remote receptor sensitivity level obtained in steps 1 to 4 are weighted and summed according to preset weights to generate a pollution source priority control level; Step 6: Multi-source data fusion and model verification: Integrate multi-temporal high-resolution remote sensing images, field sampling data, and historical environmental monitoring data to build a spatiotemporal joint analysis model to verify the reliability of risk assessment results.

2. The method according to claim 1, characterized in that In step 1, according to PI i the single - item pollution risk level is divided and mapped into five levels. The calculation formula of the single - item pollution index method is: , where PI i is the single - item pollution index, C i is the i measured concentration of the S i th heavy metal, i and i is the environmental quality standard limit value corresponding to the i th heavy metal; In the said Step 1, according to RI the comprehensive ecological risk level is divided and mapped into five levels, and the calculation formula of the potential ecological risk index method is: , where In the formula: is the single potential risk index of the i th heavy metal, is the toxicity coefficient of the i th heavy metal; The greater value level between the single pollution risk level and the comprehensive ecological risk level is used as the pollution source toxicity level, that is , where L PI and L RI are the numerical levels of 1-5 for the single pollution risk and the comprehensive ecological risk respectively. The larger the value, the higher the risk.

3. The method according to claim 2, wherein The specific method for dividing the volume level based on the quantile value of the pollution source volume in step 2 is as follows: Arrange the waste residue volumes of all blocks in ascending order to form a pollution source volume data set, calculate the 25%, 50%, 75%, and 90% quantile values of the pollution source volume data set, and divide the pollution source volume into five levels according to the quantile values. The calculation formula for the quantile value by linear interpolation method is: , where n is the total number of data, k is the quantile point, X is the sorted data set.

4. The method according to claim 1, wherein The calculation formula of the inverse distance weighting interpolation method in step 3 is as follows: , weight , Among them, Z ( x 0 ) is the predicted value of the unknown point pollutant concentration ，Z ( x j ) is the pollutant concentration at the j-th sampling point, d j is the Euclidean distance between the unknown point and the j th sampling point ，p is the distance attenuation coefficient ， The value range is from 1 to 3.

5. The method according to claim 4, characterized in that In step 3, according to RDI 总 values, the local diffusion risk is divided into five levels, and the calculation formula for the comprehensive diffusion risk index is: RDI 总 = RDI 直接 ·0.5 + RDI 间接 ·0.5, where RDI 直接 is the diffusion risk index of soil and water media, RDI 间接 is the diffusion risk index of agricultural product media.

6. The method according to claim 3, characterized in that, The calculation formula of the sensitivity index model in step 4 is: SI = A 1 ×5+ A 2 ×3+ A 3 ×1+ A 4 ×（7+ N ）， Among them SI represents the sensitivity index; A 1 、A 2 、A 3 、A 4 are respectively the area proportion of prime agricultural land, other agricultural land, residential land and water area, and the area proportion is calculated based on the total area of the 2-kilometer buffer zone around the pollution source; N is the number of water sources in the buffer zone, and 7 + N is the water area proportion coefficient; Arrange the sensitivity indexes of all pollution sources in ascending order to form a sensitivity index data set. Calculate the 25%, 50%, 75%, and 90% quantiles of the sensitivity index data set according to the linear interpolation calculation formula in step 3, and divide the sensitivity index into five levels according to the quantile values.

7. The method according to claim 6, characterized in that The weight allocation rule in step 5 is that the weight of the pollution source toxicity level is 40%, the weight of the volume level is 20%, the weight of the local diffusion risk level is 25%, and the weight of the remote receptor sensitivity level is 15%. The comprehensive risk value is obtained by weighted summation. The priority control level is divided into five levels according to the comprehensive risk value. The calculation formula of the comprehensive risk value is as follows: CR=W1·G1+W2·G2+W3·G3+W4·G4, Among them, W1, W2, W3, and W4 represent the weights of the toxicity level of the pollution source, the volume level, the local diffusion risk level, and the remote receptor sensitivity level, respectively; G1, G2, G3, and G4 represent the standardized scoring values of the toxicity level of the pollution source, the volume level, the local diffusion risk level, and the remote receptor sensitivity level, respectively, and the value range is 0-10.

8. The method according to claim 1, wherein The multi-source data fusion and model verification in step 6 specifically include the following steps: Step 6.1 Data acquisition and preprocessing: geometric correction, radiation correction, noise removal, band fusion and uniform color mosaic are performed on remote sensing images to ensure that the spatial resolution is not less than 1 meter; the Z-score standardization method is used for sampling data to eliminate dimensional differences; linear interpolation is used for historical monitoring data to fill missing values, and invalid data with a time span of more than 10 years is eliminated; Step 6.2 Feature Extraction and Fusion: Extract the normalized difference vegetation index, normalized difference water index, and texture entropy features from remote sensing images; extract the pollutant concentration gradient features and spatial clustering features from sampling data; extract the annual average growth rate and seasonal fluctuation features of pollutant concentration from historical data, and construct a multi-dimensional feature matrix; Step 6.3 Spatiotemporal Joint Analysis Model Construction: Using spectral indices, pollutant concentration gradients, and historical trends as input variables, adopt the random forest algorithm, set the number of decision trees to 100 - 500, the maximum depth to 5 - 15 layers, use the Gini impurity as the node splitting criterion, and optimize the parameters of the splitting feature selection strategy, the minimum number of samples in leaf nodes, and the splitting threshold tolerance through cross-validation; Model output of pollution diffusion probability P diffusion and the receptor sensitivity score S sensitiVity , and calculate the comprehensive risk weight through the following formula: ， Among them, α is the diffusion probability weight coefficient, β is the sensitivity score weight coefficient, and the coefficients satisfy the constraint condition α + β = 1, α has a value range of 0.6 - 0.8, β has a value range of 0.2 - 0.4; Step 6.4 Model Validation: Calculate the Pearson correlation coefficient, root mean square error, and Kappa coefficient through the cross-validation method to verify the model prediction accuracy and correct the risk level classification results.

9. The method according to claim 8, characterized in that, The following indicators are used for the model verification in Step 6.4: The calculation formula of the Pearson correlation coefficient is as follows: , Among them, x i is the measured concentration value of pollutants at the i th sampling point, y i is the model predicted concentration value at the corresponding point, , are respectively the average values of the measured value and the predicted value, n is the total number of samples; The root mean square error calculation formula is as follows: , Among them, x i is the measured concentration value of pollutants at the i th sampling point, y i and is the model predicted concentration value at the corresponding point; Kappa The coefficient calculation formula is as follows: , Among them, N is the total number of samples, m is the number of risk level categories, O kk is the number of correctly classified samples of the k th risk level, R k and C k are the total number of samples of the k th risk level in the manual interpretation result and the model output result respectively, Kappa The closer the coefficient value is to 1, the higher the consistency between the model classification result and the manual interpretation result.

10. The method according to claim 7, wherein The output results of the spatiotemporal joint analysis model in Step 6 are respectively used to dynamically update the threshold for classifying the source volume level, optimize the distance attenuation coefficient in the inverse distance weighting interpolation method, generate a risk heat map, and identify the priority control level.

Images