A method and system for agricultural waste recycling environmental risk assessment and safety threshold determination

Abstract

This invention discloses a method and system for environmental risk assessment and safety threshold determination in the recycling of agricultural waste. The method includes: acquiring the spatial distribution of pollutants, pollutant types, and the physicochemical properties of farmland soil; identifying pollutant transfer pathways and related parameters based on the spatial distribution, pollutant types, and farmland soil physicochemical properties; calculating the soil ecological environment safety threshold based on the pollutant transfer pathways and related parameters; obtaining the annual carrying capacity safety threshold of soil pollutants based on the pollutant transfer pathways, related parameters, and the soil ecological environment safety threshold; and conducting a risk assessment of the environmental carrying capacity of farmland soil based on the annual carrying capacity safety threshold of soil pollutants. This invention achieves accurate assessment, graded early warning, and scientific determination of safe return-to-field amounts of soil ecological risk in the recycling of agricultural waste by constructing a multi-pollutant composite exposure-ecological response model and machine learning prediction methods.

Description

Technical Field

[0001] This invention belongs to the interdisciplinary field of agricultural environmental engineering and environmental risk assessment, specifically relating to a method and system for environmental risk assessment and safety threshold determination of agricultural waste recycling. Background Technology

[0002] For a long time, high intensity of rural production and living, separation of planting and breeding, disconnect between production and daily life metabolism, and large-scale agricultural and rural waste have led to increasingly tight resource and environmental constraints and prominent ecological and environmental problems in some areas. Agricultural waste recycling and synergistic emission reduction are important pathways to ecological security and improving the quality and efficiency of the agricultural system, and are the only way to achieve green and low-carbon agricultural development. However, existing scientific and technological achievements are insufficient to support the realization of these goals. Therefore, it is urgent to make breakthroughs in agricultural waste recycling and synergistic emission reduction technologies, and to establish a risk assessment framework and safety threshold determination method for farmland soil ecological health during the agricultural waste recycling process that is suitable for different regions and ecological types, coordinates production and ecology, and emphasizes both quality improvement and efficiency enhancement.

[0003] An agricultural research center developed a nitrogen and phosphorus loss assessment model, and a university developed a dynamic nutrient loss prediction model, revealing the main pathways of soil nitrogen loss under different farmland management practices. However, their research results have limited reference value for revealing the cyclical metabolic pathways of nitrogen and phosphorus elements under complex farming systems and crop-livestock farming patterns in a certain central and southern region. A university established a pesticide environmental risk prediction model, and an environmental research center analyzed the driving effects of various environmental factors on the migration and transformation of pesticides and heavy metals. However, the mechanisms of their cross-media migration and transformation are still unclear, and the correlation between their environmental behavior and soil ecological environment safety remains undetermined. In recent years, an ecological environment research center has made significant progress in the study of the multi-media distribution characteristics, spatiotemporal evolution patterns, and bioaccumulation effects of new pollutants such as antibiotics and microplastics. However, breakthroughs are still needed in the cross-media diffusion fluxes of antibiotics and microplastics in soil and groundwater, exposure pathways, and their quantitative relationship with soil ecological environment safety. In particular, during the recycling of agricultural waste such as livestock and poultry manure, the coexistence of multiple pollutants, their varying emission characteristics, the intertwined cross-media migration and transformation pathways of "water-soil-crops," and the cumulative environmental impacts exacerbate the risks to soil ecological safety. Summary of the Invention

[0004] To address the problems existing in the prior art, this invention provides a method and system for environmental risk assessment and safety threshold determination in the recycling of agricultural waste. By constructing a multi-pollutant composite exposure-ecological response model and machine learning prediction methods, it achieves accurate assessment, graded early warning, and scientific determination of the safe amount of agricultural waste returned to the field in the recycling of agricultural waste.

[0005] To achieve the above objectives, the present invention provides the following solution: A method for environmental risk assessment and safety threshold determination for agricultural waste recycling, the method comprising: To obtain information on the spatial distribution of pollutants, pollutant types, and the physical and chemical properties of farmland soil; Based on the spatial distribution of pollutants, pollutant types, and the physical and chemical properties of farmland soil, the pollutant transfer pathways and related parameters are identified. Based on pollutant transfer pathways and related parameters, calculate the soil ecological environment safety threshold; Based on pollutant transfer pathways and related parameters, as well as soil ecological environment safety thresholds, the annual carrying capacity safety threshold of soil pollutants is obtained. Risk assessment of farmland soil environmental carrying capacity is conducted based on the annual carrying capacity safety threshold of soil pollutants.

[0006] Preferred methods for calculating soil ecological environment safety thresholds based on pollutant transfer pathways and related parameters include: If the pollutant transfer pathways and related parameters meet the data requirements of species sensitivity distribution, the soil environmental ecological safety threshold can be determined based on the species sensitivity distribution curve method. If the pollutant transfer pathways and related parameters do not meet the data requirements for species sensitivity distribution, the soil environmental ecological safety threshold should be determined based on the assessment factor extrapolation method, pollution index method, or ecological function response method.

[0007] Preferably, methods for obtaining the annual carrying capacity safety threshold of soil pollutants based on pollutant transfer pathways and related parameters, and soil ecological environment safety thresholds include: Based on pollutant transfer pathways and related parameters, and soil ecological environment safety thresholds, calculate soil pollutant input and output fluxes; Calculate the annual cumulative amount of soil pollutants based on the input and output fluxes of soil pollutants; Calculate the annual carrying capacity of soil pollutants based on the annual cumulative amount of soil pollutants; Based on the annual carrying capacity of soil pollutants, the safe threshold of the annual carrying capacity of soil pollutants is calculated.

[0008] Preferably, methods for calculating the annual accumulation of soil pollutants based on soil pollutant input and output fluxes include: ; In the formula, I in K represents the average annual input of pollutants, and K represents the residual rate of soil pollutants.

[0009] Preferably, the method for calculating the annual carrying capacity of soil pollutants based on the annual accumulation of soil pollutants includes: ; In the formula, SECC represents the regional soil environmental carrying capacity; d represents the soil depth. is the soil bulk density; A is the area of agricultural land; is the current value of the pollutant; is the soil environmental quality standard or hazard concentration under the corresponding land use type; F1 is the soil buffer coefficient, 0 < F1 ≤ 1; F2 is the accumulation coefficient, 0 < F2 ≤ 1; F3 is the risk coefficient, 0 < F3 ≤ 1.

[0010] Preferably, the method for risk assessment of farmland soil environmental carrying capacity based on the annual carrying capacity safety threshold of soil pollutants includes: Determine the soil ecological risk level based on the machine learning prediction model of soil pollution carrying capacity; Calculate the maximum allowable annual amount of agricultural waste returned to the field based on the soil ecological risk level; Achieve the ecological safety risk assessment of soil pollution based on the maximum allowable annual amount of agricultural waste returned to the field.

[0011] Preferably, the method for determining the soil ecological risk level based on the machine learning prediction model of soil pollution carrying capacity includes: ; In the formula, is the average contribution of the feature on the entire data set, is the benchmark prediction value, is the number of features, is the input data; The calculation process of ; In the formula, is the variable importance, is the number of samples, represents the th feature's contribution value to the prediction result of the th sample.

[0012] Preferably, the method for calculating the maximum allowable annual amount of agricultural waste returned to the field based on the soil ecological risk level includes: ; In the formula, is the maximum application amount of feces, is the maximum input amount of heavy metals, is the maximum input flux of antibiotics, and are both pollutant concentrations in feces.

[0013] This invention also provides a system for environmental risk assessment and safety threshold determination of agricultural waste recycling. The system is used to implement the aforementioned method and includes: a data acquisition and preprocessing module, a pollutant input flux calculation module, a composite ecological risk calculation module, an ecological safety threshold and carrying capacity inversion module, and a return-to-field intensity decision and early warning output module. The data acquisition and preprocessing module is used to receive data on agricultural waste type, application intensity, pollutant content, as well as soil physicochemical properties, crop type, and regional environmental parameters, and to perform unit unification, missing value processing, and standardization transformation on the received data to form the input dataset required for calculation. The pollutant input flux calculation module is used to calculate the annual input flux of heavy metals, antibiotics and persistent organic pollutants based on the type and application intensity of agricultural waste and according to a preset flux calculation model. The composite ecological risk calculation module is used to construct a multi-pollutant composite exposure index based on the predicted environmental concentration, calculate the composite pollution index or potential ecological impact ratio, and determine the soil ecological risk level based on the composite pollution index or potential ecological impact ratio. The ecological security threshold and carrying capacity inversion module is used to compare the composite ecological risk index with the preset ecological security threshold, and invert the regional soil environmental carrying capacity or the maximum allowable annual input flux of pollutants under the premise of meeting the risk constraints. The field return intensity decision and early warning output module is used to calculate the maximum allowable amount of agricultural waste to be returned to the field based on the constraints corresponding to the most restrictive pollutant, and output the risk zoning results and over-threshold early warning information.

[0014] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention addresses the multi-source input, combined exposure, and amplified ecological effects of heavy metals, antibiotics, and persistent organic pollutants during the recycling of agricultural waste. Taking farmland soil ecosystems as the research object, it comprehensively considers pollutant input intensity, soil background characteristics, environmental fate and transformation processes, and differences in ecological sensitivity to construct a multi-pollutant combined exposure-ecological response model. This model reveals the influence mechanism of synergistic / antagonistic effects between different pollutants on the evolution of ecological risks. Based on this, a graded risk threshold system based on environmental carrying capacity is proposed, providing theoretical basis and technical support for the safe return of agricultural waste to the field and the precise management of farmland soil risks.

[0015] The method provided by this invention can be used to determine the soil ecological safety thresholds (warning values ​​and limit values) of various pollutants (heavy metals, antibiotics, microplastics and organic pesticides) in the process of recycling agricultural waste in different regions.

[0016] The method provided by this invention can conduct risk assessments based on soil pollution carrying capacity, resulting in more reliable assessment results. It can be used to evaluate the ecological health risks of various pollutants during the recycling of agricultural waste in different regions and to provide timely early warnings.

[0017] The method provided by this invention can be used to assess the maximum amount of different wastes (such as livestock and poultry manure) returned to the field during the recycling of agricultural waste in different regions, providing a reference for the formulation of relevant policies.

[0018] The method provided by this invention can construct a soil environmental risk prediction model based on advanced machine learning methods (such as XGBoost / LSTM / GCN, etc.). This model has high accuracy, can assess multiple types of pollution simultaneously, and can greatly reduce the complexity of risk assessment and reduce workload. Attached Figure Description

[0019] To more clearly illustrate the technical solution of the present invention, the drawings used in the embodiments are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a schematic diagram of a method for environmental risk assessment and safety threshold determination for the recycling of agricultural waste according to an embodiment of the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0023] Example 1 like Figure 1 As shown, this invention provides a method for environmental risk assessment and safety threshold determination of agricultural waste recycling, the method comprising: To obtain information on the spatial distribution of pollutants, pollutant types, and the physical and chemical properties of farmland soil; Based on the spatial distribution of pollutants, pollutant types, and the physical and chemical properties of farmland soil, the pollutant transfer pathways and related parameters are identified. Calculate the soil ecological environment safety threshold based on the pollutant transfer pathway and related parameters; Obtain the annual carrying capacity safety threshold of soil pollutants based on the pollutant transfer pathway, related parameters, and soil ecological environment safety threshold; Conduct a risk assessment of the farmland soil environment carrying capacity based on the annual carrying capacity safety threshold of soil pollutants.

[0024] In this embodiment, the method for calculating the soil ecological environment safety threshold based on the pollutant transfer pathway and related parameters includes: If the pollutant transfer pathway and related parameters meet the data requirements of the species sensitivity distribution, determine the soil environmental ecological safety threshold based on the species sensitivity distribution curve method; If the pollutant transfer pathway and related parameters do not meet the data requirements of the species sensitivity distribution, determine the soil environmental ecological safety threshold based on the assessment factor extrapolation method, pollution index method, or ecological function response method.

[0025] In this embodiment, the method for obtaining the annual carrying capacity safety threshold of soil pollutants based on the pollutant transfer pathway, related parameters, and soil ecological environment safety threshold includes: Calculate the input-output flux of soil pollutants based on the pollutant transfer pathway, related parameters, and soil ecological environment safety threshold; Calculate the annual accumulation amount of soil pollutants based on the input-output flux of soil pollutants; Calculate the annual carrying capacity of soil pollutants based on the annual accumulation amount of soil pollutants; Calculate the annual carrying capacity safety threshold of soil pollutants based on the annual carrying capacity of soil pollutants.

[0026] In this embodiment, the method for calculating the annual accumulation amount of soil pollutants based on the input-output flux of soil pollutants includes: ; In the formula, I in is the annual average input amount of pollutants, and K is the residual rate of soil pollutants.

[0027] In this embodiment, the method for calculating the annual carrying capacity of soil pollutants based on the annual accumulation amount of soil pollutants includes: ; In the formula, SECC is the regional soil environment carrying capacity; d is the soil depth; is the soil bulk density; A is the agricultural land area; is the current value of pollutants; is the soil environmental quality standard or hazard concentration under the corresponding land use type; F1 is the soil buffer coefficient, 0 < F1 ≤ 1; F2 is the accumulation coefficient, 0 < F2 ≤ 1; F3 is the risk coefficient, 0 < F3 ≤ 1.

[0028] In this embodiment, the method for calculating the safe threshold of annual carrying capacity of soil pollutants based on the annual carrying capacity of soil pollutants includes: ; In the formula, This represents the annual carrying capacity safety threshold for soil pollutants.

[0029] In this embodiment, the method for risk assessment of farmland soil environmental carrying capacity based on the annual carrying capacity safety threshold of soil pollutants includes: A machine learning prediction model based on soil pollution carrying capacity is used to determine the level of soil ecological risk. Calculate the maximum allowable annual amount of agricultural waste returned to the field based on the soil ecological risk level; Based on the maximum permissible annual amount of agricultural waste returned to the field, an ecological safety risk assessment of soil pollution is conducted.

[0030] In this embodiment, the method for calculating the maximum permissible annual amount of agricultural waste returned to the field based on the soil ecological risk level includes: ; In the formula, This is the maximum amount of feces that can be applied. This represents the maximum input amount of heavy metals. This represents the maximum input flux of antibiotics. and All values ​​represent the concentration of pollutants in feces.

[0031] Example 2 The following is a detailed description of preferred embodiments of the present invention.

[0032] Hazard identification: ① Method 1: Collect relevant data and information obtained during the farmland soil pollution investigation phase, including information on farmland soil types, pollution types and pollutants, and environmental quality. Define risk assessment and data collection methods. Sampling of farmland soil is conducted according to GB / T 36197-2018, GB / T36199-2018, and HJ / T166. The samples are then tested, recorded, and analyzed according to HJ / T166 and GB15618-2018 to determine pollution items and their spatial distribution. Physicochemical properties of polluted farmland soil (pH value, organic matter (SOM), clay content, and anion exchange capacity (CEC)) are determined according to standards such as NY / T1377-2007 and NY / T 85-1988.

[0033] Method 2: Key Exposure Pathway Identification Based on Machine Learning. This method integrates machine learning with pollutant migration and transformation models to analyze the interactions and key exposure pathways of pollutants in farmland soil. First, multi-source data (including soil moisture content, spatiotemporal concentration of pollutants, crop uptake, and groundwater level) of the soil-crop-groundwater system are collected. A solute transport model (such as HYDRUS-1D) is used to describe the pollutant migration and transformation process. This equation provides a dynamic migration basis for exposure pathway tracking. The proposed core governing equation is as follows: in the formula Soil moisture content (cm) 3 ·cm -3 C represents the pollutant solubility (mg·L). -1 z represents the vertical depth (m or cm), and S represents the adsorption and degradation processes (mg·L⁻¹). -1 ·d -1 ). t For time (d or s). x The coordinates are for the formula (m or cm). D Hydrodynamic dispersion coefficient (m 2 ·d -1 ), v Void water velocity (m·d) -1 ).

[0034] Subsequently, machine learning algorithms (such as random forests and graph convolutional neural networks) were introduced to construct an exposure pathway contribution assessment model. The cumulative amounts of pollutants in crop roots, groundwater aquifers, and surface runoff zones, as output by the model, were used as feature variables. The contribution weights of pathways such as crop uptake, groundwater leaching, and surface runoff were quantified using a feature importance ranking formula. The model parameters were optimized by cross-validation, and the key exposure pathways with a contribution share exceeding 30% were finally identified.

[0035] (1) Overall structure of the model: A contribution assessment model M for exposure pathways is constructed. Its input is the cumulative amount of pollutants in each exposure pathway calculated by the solute transport model, and its output is the contribution weight of different exposure pathways.

[0036] The model structure can be represented as: Where: X = (x1,x2,...xn) is the feature vector, xn is the nth feature, Y is the pollutant exposure risk or pollution re-load index, and M is the machine learning model.

[0037] The characteristic variables are obtained from the output of the solute transport model: In the formula, A crop A represents the cumulative amount absorbed by crops. gw A represents the cumulative amount of groundwater leaching. runoff C represents surface runoff migration. soil Let θ be the average concentration of soil pollutants, θ be the soil moisture content, and q be the soil water flux.

[0038] (2) Random Forest Contribution Evaluation Self-Model First, construct a random forest regression model: In the formula, T represents the number of decision trees. Let be the t-th regression tree. Each tree is partitioned by minimizing the mean squared error, and the output of the random forest is used to calculate the importance of each feature.

[0039] (3) Calculation of feature importance The importance of feature xj is defined as the amount of error reduction that feature brings across all decision trees: in the formula For the importance of the j-th feature, To use feature x j The set of nodes to be split. The information gain brought about by node splitting. , The number of features.

[0040] Normalization yields the contribution weights: in the formula The contribution weight for the j-th exposure pathway.

[0041] (5) Identification of key exposure pathways Key exposure pathways are calculated based on contribution weights: Set the discrimination threshold: If the value is greater than 0.30, then the j-th path is determined to be a critical exposure path.

[0042] Ecotoxicological data collection: Species toxicity data can be obtained from authoritative sources such as soil ecotoxicity tests, various toxicity databases, and publicly published literature. Ecotoxicity testing methods should follow currently accepted standard ecotoxicity testing methods; toxicity data obtained without using standard testing methods should be evaluated based on the specific circumstances.

[0043] Commonly used toxicity databases include the ECOTOX toxicity database established by the U.S. Environmental Protection Agency and the TOXLINE toxicology database established by the National Library of Medicine (NLM). The selection of toxicity data obtained from toxicity databases or literature should follow the principles of reliability, richness, and practicality.

[0044] Determination of soil ecological environment safety threshold: This invention is based on a composite exposure ecological risk model, which uses PI = 0.5–1.0 as the ecological risk early warning threshold and PI ≥ 1.0 as the limit threshold. These thresholds are then mapped to the risk screening value and risk control value in the "Standard for Risk Control of Agricultural Land Soil Pollution", respectively, to achieve environmental carrying capacity-oriented hierarchical management under multi-pollutant scenarios.

[0045] (1) Threshold determination based on species sensitivity distribution curve method The Species Sensitivity Distribution (SSD) method is used to characterize the sensitivity of different biological species to pollutant toxicity and to derive soil environmental ecological safety thresholds based on this. Its core idea is to fit a probability distribution to multi-species toxicity data and calculate the low quantile concentration (e.g., HC5) using the distribution function, which serves as the ecological protection threshold. The main steps in deriving soil environmental ecological safety thresholds using the SSD method include: collecting and screening species toxicity data; and plotting the quantiles of the biological toxicity data arranged by magnitude. First, collect toxicity data of the target pollutant on different species. Toxicity indicators may include EC50 (median effective concentration), LC50 (median lethal concentration), NOEC (no observable effect concentration), and LOEC (lowest observable effect concentration).

[0046] Suppose there are n species toxicity data: T={t1,t2,t3,...,t n} Where t i Let X be the toxicity concentration value of the i-th species. To reduce the influence of different dimensions and distribution skewness, the toxicity data is logarithmically transformed: X i =ln(t i ) Transformed toxicity dataset: X={x1,x2,x3, ..., x n} Toxicity data sorting and cumulative quantile calculation: Sort the toxicity data in ascending order: x (1) ≤x (2) ≤...≤x (n)Then, the cumulative probability (quantile) for each species is calculated, often using the Hazen or Blom quantile formula: in the formula Let be the cumulative probability corresponding to the i-th species, and n be the number of species in the sample; This step yields the raw scatter plot data of the SSD: A specific distribution is selected to fit the parameters of these data; the SSD curve is determined by evaluating the goodness of fit. Suitable fitting models include the logistic distribution, normal distribution, and Weibull distribution. The fit can be judged using indicators such as the coefficient of determination (R²) and p-value, and the optimal model is selected through comparison. For example, the closer R² is to 1, the better the fit.

[0047] In deriving the ecological safety threshold of soil pollutants using methods such as SSD, due to the high spatial heterogeneity of the study area and the differences in soil physicochemical properties and other environmental conditions, it is necessary to determine the standard soil scenario and normalize the toxicological data. For toxicity data obtained using the same species and the same test soil at the same endpoint, the geometric mean is taken; for toxicity data obtained using the same species and the same test soil at different endpoints, the minimum value is taken; for toxicity data obtained using the same species and different test soils at the same endpoint, the minimum value is taken. A specific regression model example is shown below: Formula EC x ECx represents the concentration of pollutants that cause x% biological effect (ECx: the concentration of pollutants that causes x% biological effect, an effect level parameter representing the percentage of biological response (0–100%)), a, b, c, and d are regression model parameters, pH is the pH value of the tested soil, OC is the organic matter content of the tested soil, and CEC is the cation exchange capacity of the tested soil.

[0048] (2) Threshold determination by evaluation factor (AF) extrapolation method The assessment factor method is a method for determining the ecological screening value of a pollutant by dividing the selected lowest reported toxicity value by an uncertainty factor or safety factor. The toxicity data used in the assessment factor method are typically the lowest known effective concentration obtained from laboratory tests, and the range of assessment factor values ​​is determined based on scientific experience and professional judgment. Taking the determination of the safety threshold for antibiotics in soil as an example, the no-effect concentration (PNEC) of the antibiotic is... waterCalculation: HC5 is calculated based on the optimal SSD model. An assessment factor (AF) is introduced, and the calculation formula is as follows. Referring to existing research, AF can be set to 1000 to obtain the ecological risk threshold for each antibiotic. Soil antibiotic safety threshold calculation: This is done using the soil moisture distribution coefficient. Calculate the antibiotic-free concentration in soil The proposed calculation formula is as follows: (3) Derivation of threshold based on pollution index Based on pollution risk and ecological risk assessment, a comprehensive risk level is derived using methods such as the pollution index, and the threshold for heavy metal pollution zoning is determined. The corresponding effective state threshold is then calculated using a total-effective state model. The method of inferring zoning thresholds from the comprehensive assessment results is simple and convenient.

[0049] ① Single-factor pollution index method Using a single-factor pollution index ( Quantitatively characterize the ecological security risk of individual pollutants and determine the risk level. These are actual measured values ​​of pollutants in farmland soil. The following calculation formula is proposed to be used as the baseline concentration (soil quality standard value of pollutants or soil background pollutant concentration): The Nemerow integrated pollution index is used to quantify the cumulative ecological security risk of multiple pollutants in a single type and to determine the risk level. The proposed calculation formula is as follows: Where P ave This represents the impact of the inherent toxicity characteristics of different heavy metal pollutants on the soil environment, and is used to describe the average value of a single-factor pollution index. For quantity, This is a weighted importance value of pollutants to the environment. This value is calculated using a weighted counting method or provided by experts, combining data on trace elements belonging to different environmental importance categories; P i，max P represents the maximum value of the pollution index for a single factor. t The Nemerow Comprehensive Pollution Index is for a single type of pollutant.

[0050] The proposed calculation formula for the comprehensive risk index (PI) of multiple pollutions in farmland soil is as follows: In the formula, P tX P tYP tZ and P tJ These represent the comprehensive pollution indices for heavy metals, pesticide residues, antibiotics, and microplastics, respectively. PI is the comprehensive pollution index for farmland soil. Referring to the HJ / T 166—2004 standard document, this index is divided into five risk levels.

[0051] ② Entropy weight method Entropy weighting is an objective weighting method that calculates a comprehensive index after considering the information provided by various factors. In this invention, entropy weighting is based on three single-factor pollution indices. Potential ecological risk coefficient With the Earth Accumulation Index I geo The entropy weight method determines the weights based on the information entropy of each indicator's variability, thus resolving the issues of inconsistent evaluation results and difficulty in quantification. The main steps in applying the entropy weight method include data standardization and calculating the information entropy E of the j-th indicator. j The proposed calculation formula is as follows: In the formula: j = 1, 2, 3; The weight of the indicator value of the i-th evaluation object under the j-th indicator is... Let be the index value of the i-th evaluation object under the j-th index.

[0052] Calculate the weight of each indicator The formula is as follows: Calculate the comprehensive pollution risk of each evaluation object S i The formula is as follows: Standardized values ​​for each evaluation index (Table 1) Potential ecological risk coefficient With the Earth Accumulation Index I geo The calculation method is as follows: Potential ecological risk coefficient The potential ecological risk coefficient method is a method that combines heavy metal content, ecological and environmental effects, and toxicological effects to evaluate pollutants based on their properties and their migration, transformation, and deposition behavior in the environment. Its reference calculation formula is: In the formula, The potential ecological risk coefficient of heavy metal i in soil or sediment of a certain area. Let i be the toxicity response coefficient.

[0053] Geographic Cumulative Index The geoaccumulation index method is a quantitative method for studying the degree of heavy metal pollution in soil and sediments, which comprehensively considers factors such as anthropogenic pollution, environmental geochemical background, and background value variations caused by natural diagenesis. The calculation formula is as follows: In the formula, The coefficient is chosen to account for the background value variations caused by natural diagenesis in different regions. It is generally set to 1.5 and can be adjusted according to the actual situation.

[0054] (4) Ecological function response method This method analyzes the impact of pollutant concentration changes on key functions of the soil ecosystem (such as microbial activity, enzyme activity, and nutrient cycling capacity), establishes a pollutant concentration-ecological function response model, and identifies the critical concentration at which ecological functions change significantly, thereby determining the soil safety threshold.

[0055] The core of the ecological function response method is to establish a quantitative relationship between pollutant concentration (CCC) and ecological function index (FFF): F = f(C). Where C in the soil represents the concentration of the target pollutant (mg·kg⁻¹) -1 F represents the soil ecological function index. When pollutant concentrations increase, ecological function indicators usually show inhibition or decline. The safety threshold can be obtained by identifying the critical point of functional change.

[0056] Constructing a concentration-ecological function response model (Logistic model): in the formula The ecological function level at concentration C. The maximum ecological function value, Let 'a' be the slope of the curve, representing the concentration that would cause a 50% reduction in function. Model parameters are estimated using either the least squares method or the maximum likelihood method.

[0057] Functional response threshold calculation: When the ecological function declines to a certain percentage (such as 10% or 20%), the corresponding concentration is used as a safety threshold: = or = Where E The concentration that causes a 10% change in function, E The concentration that causes a 20% functional change.

[0058] Risk assessment based on the environmental carrying capacity of farmland soil: Construct a multi-scenario input model for heavy metals, antibiotics, and organic pollutants under the scenario of returning agricultural waste to the field. Combine the soil background, degradation and migration, and crop absorption processes to predict the environmental exposure levels of pollutants; Based on Composite ecological risk indicators such as etc., divide the safe area, early warning area, and restricted area, and map the early warning threshold to the risk accumulation area below the screening value of agricultural land soil pollution risk, and map the restricted threshold to the risk control level to reverse the regional environmental carrying capacity and the maximum allowable amount of returning to the field.

[0059] According to the T / SSSC 009-2024 standard document, the work content of calculating the regional soil environmental carrying capacity includes calculating the regional soil environmental carrying capacity, clarifying the net input flux of soil pollutants in the region, and calculating the regional soil environmental carrying capacity. For example, the input pathways of heavy metals may include atmospheric dry and wet deposition, organic fertilizers and chemical fertilizers, irrigation, and runoff, and crop harvest and straw removal are used as output pathways. Set the allowable limit of the content according to the ecological safety threshold of farmland soil for pollutants, etc.; Measure the actual pollutant content and soil physical and chemical properties of farmland soil; Set different time lengths, and divide the environmental capacity by the time to obtain the maximum annual increment of pollutants; Comprehensively consider the input and output pathways of pollutants in farmland soil, and calculate the annual accumulation amount of pollutants. The calculation method can refer to the following formula and be adjusted according to the actual situation.

[0060] (1) The calculation formula for the soil environmental carrying capacity is as follows: SECC (soil environmental carrying capacity) is the regional soil environmental carrying capacity (kg); d is the soil depth (m), generally only considering the surface layer of 20 cm; ρ is the soil bulk density (kg·m -3 ); A is the area of agricultural land (km2); is the current value of the pollutant (mg·kg -1 ); is the soil environmental quality standard (mg·kg -1 ) under the corresponding land use type, or the hazard concentration; F1 is the soil buffer coefficient, 0 < F1 ≤ 1; F2 is the accumulation coefficient, 0 < F2 ≤ 1; F3 is the risk coefficient, 0 < F3 ≤ 1.

[0061] ① Characterize the soil buffer coefficient F1 with the natural attenuation capacity (NAC), and the calculation formula is as follows: In the formula, ref represents the reference value of the corresponding index, and ω i The weights of each indicator are i=1, 2, 3, 4. Taking soil organic matter (SOM), clay content (CLAY), soil pH and cation exchange capacity (CEC) as examples, the indicators characterize the ability of pollutants to be retained in the soil or the purification effect of the soil on pollutants.

[0062] ②The cumulative coefficient F2 is to be calculated using the following formula: In the formula, I in The average annual input of pollutants, kg·a -1 ; Annual average output of pollutants (kg·a) -1 ); other meanings are the same as before.

[0063] For agricultural land, atmospheric deposition (I) is the primary consideration. atm ), irrigation water (I irr ), manure fertilization (I fert The specific transfer pathways and key parameters of pollutants need to be adjusted based on specific circumstances or relevant experimental results, including input pathways such as agricultural films (microplastics) and output processes such as crop harvesting (U), straw removal (H), and drainage (D).

[0064] ③ Annual net accumulation: Due to the complexity of the migration and transformation process of pollutants in the soil, the model parameters are numerous and some data are not easy to obtain. Therefore, in practice, the annual net accumulation (annual accumulation of soil pollutants) is often estimated by using the soil pollutant residue rate (K). The K value can be obtained through experimental measurement or by referring to the literature. The calculation formula is as follows.

[0065] ④ Calculation of risk factor F3: For agricultural land, based on literature review and field surveys, the crop pollutant enrichment factor (PUF) and the proportion (P) of the regional population ingesting crops with excessive levels of a certain pollutant are obtained. Taking heavy metal pollution in rice as an example, the risk coefficient is proposed to be calculated using the following formula: (2) Calculation of pollutant input flux Taking livestock and poultry manure fertilization as an example, the pollutant input flux is calculated. First, the manure emissions of different livestock and poultry are calculated based on the excretion coefficient method. The calculation formula is as follows: M represents the amount of manure produced by a certain type of livestock (pigs, cattle, sheep) or poultry (unit: tons, fresh weight); N represents the number of such livestock or poultry raised; K represents the manure excretion coefficient per unit animal (unit: kg / day / head, fresh weight); D represents the animal feeding cycle (unit: days). f This indicates a unified unit conversion factor, with a value of 10. - ³; P represents the rate of returning animal manure to the field (unit: %).

[0066] The proposed calculation formulas for the pollutant emissions in the manure of different livestock and poultry are as follows: M represents the emission amount of a certain type of pollutant in the manure of a certain type of livestock (pig, cattle, sheep) or poultry (unit: tons, dry weight); M represents the emission amount of manure of this type of livestock or poultry in the province (unit: tons, fresh weight); W represents the average moisture content of the manure of this type of livestock or poultry (unit: %), and the data can be obtained from previous research reports; This indicates the average concentration of a certain type of pollutant in the manure of this type of livestock or poultry (unit: mg / kg, dry weight). f' This indicates a unit conversion factor, which can be 10. -6 Adjustments can be made based on the actual situation. This represents the total amount of pollutants introduced through the return of livestock and poultry manure to the fields.

[0067] (3) Calculation of pollutant output flux Taking the straw removal pathway as an example, the output pollutant flux (H) is calculated using the following formula: In the formula, H i For some kind of pollutant, f j The crop straw removal rate can be set as needed (e.g., it can be set to 100%). M represents the ratio of pollutants in crop straw to the content of a certain type of pollutant in grains; M represents the ratio of crop straw yield to crop yield, which can be adjusted based on actual experimental results.

[0068] (4) Risk assessment and early warning In farmland management, maintaining the integrity of soil ecological conditions is crucial for the sustainable use of land. If SECC (Sequential Soil Contamination Control Classification) is used as an acceptable threshold for soil pollution, basic calculations can estimate the time required to reach that threshold. The proposed calculation formula for issuing early warnings is as follows: In the formula, This is a risk warning time.

[0069] The calculation steps for the soil bearing capacity safety threshold are the same as those for the soil bearing capacity itself. However, when calculating the soil bearing capacity safety threshold, the safety or limit threshold for a specific pollutant needs to be substituted into the parameters. In the calculation formula Perform the calculation.

[0070] Screening values ​​and control values ​​were selected as thresholds for the soil environmental quality of farmland, and warning levels were assigned. The higher the warning level, the greater the ecological health risk. Low alert level: Pollutant concentration is below the screening value. Level 1: Predicted exceedance period > 30 years; Level 2: Predicted exceedance period < 30 years. Medium alert level: Pollutant concentration is between the screening value and the control value.

[0071] High alert level: Pollutant concentration exceeds control limits, which may have significant health impacts and requires mandatory soil remediation.

[0072] (5) Construction of risk assessment machine learning model To address the need for ecological risk management of farmland soil caused by the combined input of multiple pollutants during the recycling of agricultural waste, this paper constructs a machine learning-based risk assessment method under mechanistic constraints to achieve the engineering identification of soil pollution carrying capacity and risk thresholds. First, based on the application intensity of agricultural waste, pollutant occurrence characteristics, and soil physicochemical properties, a mechanistic model for predicting pollutant migration, transformation, and exposure is established. Predicted environmental concentrations and initial ecological risk indicators under different scenarios are calculated, providing physical constraints and prior information for machine learning modeling. Second, based on the output of the mechanistic model, machine learning methods are introduced to characterize the nonlinear response relationship of ecological risk to pollution input intensity under combined exposure conditions of multiple pollutants. A three-level risk assessment model of "safety-early warning-restriction" is constructed to achieve rapid screening and risk classification of agricultural waste return intensity. By systematically changing the pollutant input scenario, key intervals in which risk levels shift are identified, and ecological risk early warning and restriction thresholds are inverted. Soil pollution carrying capacity is expressed in interval form, improving the safety margin of engineering decisions. During model application, the risk assessment results are validated using monitoring data from typical demonstration areas to evaluate the applicability and stability of the model under different soil types and application conditions. Ultimately, this method was embedded into the agricultural waste return-to-field management decision-making system to achieve automatic calculation of the maximum allowable amount of waste returned to the field and dynamic output of risk warning information, providing technical support for the safe recycling of agricultural waste and the refined management of farmland soil risks.

[0073] Input The intensity of agricultural waste returning to the field and the characteristics of pollutant occurrence (heavy metals, antibiotics, persistent organic pollutants) were combined with crop type, soil physicochemical properties and regional environmental parameters to construct a multi-pollutant composite input and soil environmental background dataset.

[0074] Model Under the constraints of the pollutant migration, transformation and exposure prediction mechanism model, machine learning methods (such as XGBoost / LSTM / GCN) are introduced to characterize the nonlinear response relationship of ecological risk to pollution input intensity under multi-pollutant compound exposure conditions, and a three-level risk assessment model of "safety-early warning-limitation" is constructed to realize the engineering identification of soil pollution carrying capacity threshold.

[0075] The proposed machine learning model is: In the model For each pollutant The warning time, These are environmental and management characteristic variables (such as pH and SOM). For each pollutant For calculating the regional soil environmental carrying capacity, machine learning models such as XGBoost regression model, Random Forest, and LSTM (long-term monitoring data) can be used.

[0076] Model training objective function: To identify the key factors influencing the duration of pollutant exceedances, an interpretive analysis of the machine learning model is performed. The prediction model can be represented as: in the formula Features The average contribution across the entire dataset can be expressed as follows: The calculation formula is derived from this; The number of features; This is the baseline predicted value, which is usually the average of the predicted values ​​from the training data.

[0077] Key drivers are identified by calculating the mean absolute SHAP value: in the formula For variable importance, For the sample size, Indicates the first The feature is related to the first The contribution value of the prediction results for each sample.

[0078] Output Decision parameters such as the maximum allowable amount of agricultural waste to be returned to the field, ecological risk early warning thresholds, and restriction thresholds are established to provide intelligent support for the risk-based classification and control of farmland soil pollution and its safe recycling.

[0079] (6) Calculation of the maximum allowable annual input flux of pollutants Based on the multi-source input-output quality balance and ecological risk probability inversion method of farmland soil, the maximum allowable annual input flux of pollutants is deduced under the premise of setting an acceptable ecological risk level, and the maximum safe return intensity of agricultural waste to the field is determined by the most limiting pollutant, thereby realizing scenario-based risk management based on soil carrying capacity.

[0080] Set maximum warning time limit (T) limit The timeframe is 30 years, and the maximum annual input of pollutants is calculated. The proposed calculation formula is as follows: Based on the different sources and proportions of pollution, the maximum annual usage of corresponding agricultural waste is calculated. For example, the formula for calculating the maximum usage of livestock and poultry manure is as follows: in the formula This is the maximum amount of feces that can be applied. This represents the maximum input amount of heavy metals. This represents the maximum input flux of antibiotics. and All values ​​represent the concentration of pollutants in feces.

[0081] Example 3 The present invention also provides a system for environmental risk assessment and safety threshold determination of agricultural waste recycling. The system is used to implement the method described in Embodiment 1. The system includes: a data acquisition and preprocessing module, a pollutant input flux calculation module, a composite ecological risk calculation module, an ecological safety threshold and carrying capacity inversion module, and a return-to-field intensity decision and early warning output module. The modules are connected sequentially or in parallel via data interfaces to enable automatic assessment of soil ecological risks and output of decision parameters in the context of agricultural waste returning to the field.

[0082] The data acquisition and preprocessing module is used to receive or read data on agricultural waste type, application intensity, pollutant content, soil physicochemical properties, crop type, and regional environmental parameters, and to perform unit unification, missing value processing, and standardization transformation on the data to form the input dataset required for model calculation.

[0083] The pollutant input flux calculation module is used to calculate the annual input flux of heavy metals, antibiotics and persistent organic pollutants based on the type and intensity of agricultural waste and according to a preset flux calculation model, providing boundary conditions for subsequent risk calculations.

[0084] The composite ecological risk calculation module is used to construct a composite exposure index for multiple pollutants based on predicted environmental concentrations, calculate the composite pollution index or the proportion of potential ecological impacts, and determine the soil ecological risk level accordingly.

[0085] The ecological security threshold and carrying capacity inversion module is used to compare the composite ecological risk index with the preset ecological security threshold, and, under the premise of meeting the risk constraints, invert the regional soil environmental carrying capacity or the maximum allowable annual input flux of pollutants.

[0086] The agricultural waste return intensity decision and early warning output module is used to calculate the maximum allowable amount of agricultural waste to be returned to the field based on the constraints corresponding to the most restrictive pollutants, and output the risk zoning results and over-threshold early warning information.

[0087] Example 4 The present invention also provides an electronic device for implementing the above-mentioned method for soil environmental risk assessment and ecological safety threshold determination of agricultural waste recycling. The electronic device can be a server, workstation, cloud computing node or other computing device with data processing capabilities.

[0088] (1) Hardware composition of electronic devices The electronic device includes at least: Processor; Memory; Communication interface (optional); The processor is electrically connected to the memory and is used to execute computer program instructions stored in the memory.

[0089] (2) Memory and program structure The memory stores a computer program for performing soil environmental risk assessments for agricultural waste recycling. The computer program includes multiple instructions, which, when executed by a processor, cause the electronic device to perform at least the following steps: ① Steps to obtain input data To obtain data on agricultural waste types, application intensity, pollutant occurrence characteristics, as well as soil physicochemical properties, crop systems, and regional environmental parameters; ② Flux Accounting Steps The pollutant input flux calculation command is invoked to calculate the annual input flux of pollutants based on the application intensity of agricultural waste and the pollutant content. ③ Steps for calculating complex ecological risks The composite risk calculation command is invoked to convert the exposure levels of multiple pollutants into a unified composite risk index, which is used to characterize the soil ecological risk level under conditions of coexistence of multiple pollutants. ④ Steps for determining ecological security thresholds The composite risk index is compared with a preset ecological security threshold to determine the risk level under the current scenario; ⑤ Steps for inverting bearing capacity and return-to-field intensity Under the condition of meeting the ecological security threshold constraint, the inversion calculation instruction is invoked to deduce the maximum allowable input flux of pollutants or the maximum allowable amount of agricultural waste returned to the field. ⑥ Result Output Steps The output includes assessment results such as risk level, early warning information, maximum allowable amount of pollutants to be returned to the field, and the most restricted pollutants.

[0090] (3) Technical effects Through the execution of the aforementioned electronic equipment and computer programs, the automatic assessment of the complex ecological risks of multiple pollutants under the conditions of agricultural waste recycling and the calculation of the decision parameters for the intensity of returning waste to the field are realized. This enables the assessment method to be implemented stably and repeatedly in the form of a computer system, thereby improving the engineering feasibility and decision-making efficiency of soil ecological environment risk management.

[0091] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made to the technical solutions of the present invention by those skilled in the art without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. A method for agricultural waste recycling environmental risk assessment and safety threshold determination, characterized in that, The method includes: To obtain information on the spatial distribution of pollutants, pollutant types, and the physical and chemical properties of farmland soil; Based on the spatial distribution of pollutants, pollutant types, and the physical and chemical properties of farmland soil, the pollutant transfer pathways and related parameters are identified. Based on pollutant transfer pathways and related parameters, calculate the soil ecological environment safety threshold; Based on pollutant transfer pathways and related parameters, as well as soil ecological environment safety thresholds, the annual carrying capacity safety threshold of soil pollutants is obtained. Risk assessment of farmland soil environmental carrying capacity is conducted based on the annual carrying capacity safety threshold of soil pollutants.

2. The method of claim 1, wherein, Methods for calculating soil ecological and environmental safety thresholds based on pollutant transfer pathways and related parameters include: If the pollutant transfer pathways and related parameters meet the data requirements of species sensitivity distribution, the soil environmental ecological safety threshold can be determined based on the species sensitivity distribution curve method. If the pollutant transfer pathways and related parameters do not meet the data requirements for species sensitivity distribution, the soil environmental ecological safety threshold should be determined based on the assessment factor extrapolation method, pollution index method, or ecological function response method.

3. The method of claim 1, wherein, Methods for obtaining the annual carrying capacity safety threshold of soil pollutants based on pollutant transfer pathways and related parameters, as well as soil ecological environment safety thresholds, include: Based on pollutant transfer pathways and related parameters, and soil ecological environment safety thresholds, calculate soil pollutant input and output fluxes; Calculate the annual cumulative amount of soil pollutants based on the input and output fluxes of soil pollutants; Calculate the annual carrying capacity of soil pollutants based on the annual cumulative amount of soil pollutants; Based on the annual carrying capacity of soil pollutants, the safe threshold of the annual carrying capacity of soil pollutants is calculated.

4. The method according to claim 3, characterized in that, Methods for calculating the annual accumulation of soil pollutants based on soil pollutant input and output fluxes include: ； In the formula, I in is the annual average input of pollutants, and K is the soil pollutant residual rate.

5. The method according to claim 3, characterized in that, Methods for calculating the annual carrying capacity of soil pollutants based on the annual accumulation of soil pollutants include: ； In the formula, SECC represents the regional soil environmental carrying capacity; d represents the soil depth. A represents soil bulk density; A represents agricultural land area. This represents the current status of pollutants; F1 represents the soil environmental quality standard or hazard concentration under the corresponding land use type; F2 represents the soil buffer coefficient, 0 < F1≤1; F3 represents the cumulative coefficient, 0 < F2≤1; and F3 represents the risk coefficient, 0 < F3≤1.

6. The method according to claim 1, characterized in that, Methods for risk assessment of farmland soil environmental carrying capacity based on annual carrying capacity safety thresholds for soil pollutants include: A machine learning prediction model based on soil pollution carrying capacity is used to determine the level of soil ecological risk. Calculate the maximum allowable annual amount of agricultural waste returned to the field based on the soil ecological risk level; Based on the maximum permissible annual amount of agricultural waste returned to the field, an ecological safety risk assessment of soil pollution is conducted.

7. The method according to claim 6, characterized in that, Methods for determining soil ecological risk levels based on machine learning prediction models of soil pollution carrying capacity include: ； In the formula, Features Average contribution across the entire dataset As the baseline forecast value, For the number of features, Input data; The calculation process is as follows: ； In the formula, For variable importance, For the sample size, Indicates the first The feature is related to the first The contribution value of the prediction results for each sample.

8. The method according to claim 7, characterized in that, Methods for calculating the maximum permissible annual amount of agricultural waste returned to the field based on soil ecological risk levels include: ； In the formula, This is the maximum amount of feces that can be applied. This represents the maximum input amount of heavy metals. This represents the maximum input flux of antibiotics. and All values ​​represent the concentration of pollutants in feces.

9. A system for environmental risk assessment and safety threshold determination of agricultural waste recycling, said system being used to implement the method described in any one of claims 1-8, characterized in that, The system includes: a data acquisition and preprocessing module, a pollutant input flux calculation module, a composite ecological risk calculation module, an ecological security threshold and carrying capacity inversion module, and a land return intensity decision and early warning output module. The data acquisition and preprocessing module is used to receive data on agricultural waste type, application intensity, pollutant content, as well as soil physicochemical properties, crop type, and regional environmental parameters, and to perform unit unification, missing value processing, and standardization transformation on the received data to form the input dataset required for calculation. The pollutant input flux calculation module is used to calculate the annual input flux of heavy metals, antibiotics and persistent organic pollutants based on the type and application intensity of agricultural waste and according to a preset flux calculation model. The composite ecological risk calculation module is used to construct a multi-pollutant composite exposure index based on the predicted environmental concentration, calculate the composite pollution index or potential ecological impact ratio, and determine the soil ecological risk level based on the composite pollution index or potential ecological impact ratio. The ecological security threshold and carrying capacity inversion module is used to compare the composite ecological risk index with the preset ecological security threshold, and invert the regional soil environmental carrying capacity or the maximum allowable annual input flux of pollutants under the premise of meeting the risk constraints. The field return intensity decision and early warning output module is used to calculate the maximum allowable amount of agricultural waste to be returned to the field based on the constraints corresponding to the most restrictive pollutant, and output the risk zoning results and over-threshold early warning information.

Images