Drainage basin pollution tracing method, device and equipment and storage medium

Abstract

The invention relates to the technical field of pollution tracing, and discloses a drainage basin pollution tracing method, device and equipment and a storage medium. The method comprises the following steps: inputting obtained target pollution source emission data and basic data into a target SWAT model, then operating the target SWAT model, and simulating and outputting sub-basin pollution load of each sub-basin in a target basin; determining contribution proportions of different pollution source categories to the total pollution load of the target drainage basin based on an isotopic tracing method; determining a load vector of each pollution source category of the target drainage basin based on the contribution proportion and the drainage basin pollution simulation total load; constructing a sub-basin load distribution matrix; determining a unit load matrix based on the sub-basin load distribution matrix; and analyzing the unit load matrix, and tracing to obtain the dominant pollution category of each sub-basin. According to the scheme, by constructing the unit load matrix, the main source sub-drainage basins of various pollution sources are defined, the space traceability of the pollution sources is realized, and the traceability accuracy is improved.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of pollution tracing, and particularly relates to a method, device and equipment for tracing pollution in a river basin and a storage medium. BACKGROUND

[0002] The application belongs to the technical field of water environment management, river basin pollutant tracing and environmental informatization, and particularly relates to a method and system for fine tracing of nitrogen and phosphorus pollution loads in a river basin based on multi-dimensional coupling of a "source inventory-source model-source category", which realizes quantitative calculation of nitrogen and phosphorus nutrient pollution loads, sub-basin load distribution and accurate analysis of pollution source category contribution rate at a river basin scale.

[0003] With the rapid development of industrialization, urbanization and agricultural production activities, the concentration of nitrogen and phosphorus nutrients in the water body of the river basin increases year by year, leading to water eutrophication and ecosystem degradation, which seriously affects water environment safety and ecological function. Therefore, accurately identifying pollution sources and quantifying their contribution have become a key prerequisite for precise management of the river basin.

[0004] In the prior art, non-point source pollution has complex causes and obvious spatial and temporal heterogeneity. Agricultural non-point source pollution, livestock and poultry breeding and urban domestic sewage may all lead to an increase in nitrogen and phosphorus loads in the water body. Although existing hydrological and water quality simulation tools, such as the SWAT (Soil and Water Assessment Tool) model, can simulate total nitrogen (TN) and total phosphorus (TP) loads at a sub-basin scale, they are difficult to effectively support identification of dominant pollution types in a river basin with complex pollution source distribution and various types of pollution, which limits the development and implementation of differentiated and targeted pollution prevention and control measures. SUMMARY

[0005] In view of the deficiencies in the prior art, the purpose of the present application is to provide a method, device, equipment and medium for tracing pollution in a river basin to solve the problem of difficulty in identifying dominant pollution types in the prior art.

[0006] According to one aspect of the present application, a method for tracing pollution in a river basin is disclosed, the method comprising: obtaining target pollution source discharge data and basic data of a target river basin; inputting the target pollution source discharge data and basic data into a target SWAT model and running the target SWAT model to simulate and output sub-basin pollution loads of each sub-basin in the target river basin; determining a total pollution simulation load of the target river basin and a pollution source category system of the target river basin, the total pollution simulation load being the sum of the sub-basin pollution loads, and the pollution source category system comprising a plurality of pollution source categories; determining a contribution proportion of different pollution source categories to the total pollution load of the target river basin based on an isotope tracing method. determining a load vector of each pollution source category of the target watershed based on the contribution proportion and the total load of the watershed pollution simulation; constructing a sub-watershed load allocation matrix based on the load vector of each pollution source category of the target watershed and the proportion of the pollution load of each sub-watershed in the total load of the watershed pollution simulation; dividing each element in the sub-watershed load allocation matrix by the actual area corresponding to the pollution category associated with the pollution source category in the corresponding sub-watershed to obtain a unit load matrix; analyzing the unit load matrix to trace back to the dominant pollution category of each sub-watershed.

[0007] In some embodiments, after obtaining the target pollution source emission data and the basic data of the target watershed, the method further comprises constructing a target SWAT model, comprising: spatially allocating the target pollution source emission data to obtain model input parameters; dividing the watershed based on a digital elevation model to construct an initial SWAT model framework, wherein the initial SWAT model framework is constructed based on a plurality of sub-watersheds and a plurality of hydrological response units after watershed division; inputting the model input parameters and the basic data into the initial SWAT model framework to adjust the original model parameters in the initial SWAT model to obtain an optimized SWAT model with optimized model parameters; determining model calibration parameters, calibrating the optimized SWAT model based on the model calibration parameters to obtain the target SWAT model.

[0008] In some embodiments, the obtaining of the pollution source emission data of the target watershed comprises: obtaining pollution point sources and pollution area sources of the target watershed; determining point source pollution emission data of the pollution point sources based on an online monitoring method and determining area source pollution emission data of the pollution area sources based on a pollution discharge coefficient method, wherein the emission coefficient applied in the pollution discharge coefficient method is a target coefficient matched with the target watershed after modification of original coefficients of the pollution discharge coefficient method; determining the pollution source emission data of the target watershed based on the point source pollution emission data and the area source pollution emission data, wherein the point source pollution emission data and the area source pollution emission data each correspond to geographical location information and attribute information.

[0009] In some embodiments, the spatial allocation of the target pollution source emission data to obtain model input parameters comprises: spatialize the point source pollution emission data to the corresponding target sub-basin based on geographical position information of the point source pollution emission data; spatialize the non-point source pollution emission data to the corresponding target sub-basin by using a spatial allocation coefficient method, including: label the non-point source pollution emission data based on a non-point source identifier; determine associated parameter grid data related to the non-point source pollution emission data; for each target sub-basin, determine an allocation weight of each grid cell according to associated parameter grid data corresponding to each grid cell falling within the range of the target sub-basin and an area proportion of the grid cell in the target sub-basin; allocate the non-point source pollution emission data to each target sub-basin based on the allocation weight.

[0010] In some embodiments, the determining a model calibration parameter, calibrating the optimized SWAT model based on the model calibration parameter to obtain a target SWAT model includes: automatically calibrating the optimized SWAT model by using a SWAT-CUP tool; the automatic calibration process includes: selecting, from parameters of the optimized SWAT model, model calibration parameters sensitive to hydrological and water quality processes, the model calibration parameters including parameters CN2, ALPHA_BF, SOL_AWC affecting hydrological processes and parameters ERORGN, NPERCO affecting water quality processes; setting an objective function, the objective function including at least a Nash efficiency coefficient NSE and a determination coefficient R²; performing parameter optimization iteration by using a SUFI-2 algorithm, adjusting the model calibration parameters, and making the objective function between simulation output values of the optimized SWAT model and measured values of the target basin reach a target optimization range; verifying the model completed automatic calibration, and obtaining the target SWAT model when a verification simulation result meets a preset performance standard, wherein the performance standard includes that the Nash efficiency coefficient NSE is greater than a target efficiency coefficient threshold and the determination coefficient is greater than a target determination coefficient threshold.

[0011] In some embodiments, the determining an actual area corresponding to the pollution category includes: obtaining an original area proportion of each sub-basin target pollution category in each target sub-basin of the target basin; performing normalization processing on the original area proportion to determine an actual area corresponding to each sub-basin target pollution category, wherein a normalization processing formula is: ; In the formula, Sub-basin s No. i Land use normalization ratio; Sub-basin s No. i Original proportions of land use types; Actual area calculation formula: ; In the formula: Sub-basin s No. i Actual area of ​​land use category (ha); Sub-basin s Total area (ha).

[0012] The total simulated pollution load of the watershed is determined based on the following formula: ; In the formula: Sub-basin j Total pollutant load (kg / month), j=1,2,3......n; Sub-basin j No. i Actual area of ​​land use category (ha); : No. i Land use unit load (kg / ha / month), i=1,2,3......m; Total number of sub-basins; Total number of land use types; Matrix representation: ; ; The load vector is expressed based on the following formula: ; In the formula: Total pollutant load vectors for various land uses in the watershed (kg / month); Total pollutant load in the watershed (kg / month); : No. i Land use category proportion, when i=1, When i=m, ; The sub-basin load allocation matrix is ​​expressed based on the following formula: ; In the formula: : m × n Sub-basin load allocation matrix; Sub-basin j The Middle i Land use pollutant load (kg / month); Sub-basin j Total pollutant load (kg / month), j=1, When j=n, ; The unit load matrix is ​​expressed based on the following formula: ; In the formula: Sub-basin j The Middle i Land use unit load (kg / ha / month); Matrix form: ; In the formula: matrix U Each row corresponds to a sub-basin, and each column corresponds to a land use type.

[0013] According to another aspect of this application, a watershed pollution tracing device is also disclosed, the device comprising: The data acquisition module is used to acquire emission data of target pollution sources in the target watershed, as well as basic data. The simulation output module is used to input the emission data of the target pollution source and the basic data into the target SWAT model, run the target SWAT model, simulate and output the pollution load of each sub-basin in the target watershed; The category system determination module is used to determine the total simulated pollution load of the target watershed and the pollution source category system of the target watershed. The total simulated pollution load of the watershed is the sum of the pollution loads of multiple sub-watersheds, and the pollution source category system includes multiple pollution source categories. The contribution ratio determination module is used to determine the contribution ratio of different pollution source categories to the total pollution load of the target watershed based on the isotope tracing method. The load vector determination module is used to determine the load vector of each pollution source category in the target watershed based on the contribution ratio and the total simulated pollution load of the watershed. The sub-basin load allocation matrix construction module is used to construct the sub-basin load allocation matrix based on the load vectors of each pollution source category in the target watershed and the proportion of pollution load in each sub-basin to the total simulated pollution load of the watershed. The unit load matrix determination module is used to divide each element in the sub-basin load allocation matrix by the actual area corresponding to the pollution category associated with the pollution source category in the corresponding sub-basin to obtain the unit load matrix; The dominant pollution category tracing module is used to parse the unit load matrix and trace the dominant pollution category of each sub-basin.

[0014] In some embodiments, an electronic device is also disclosed, the electronic device including a memory and at least one processor, the memory storing instructions; the at least one processor invokes the instructions in the memory to cause the electronic device to perform various steps of the watershed pollution tracing method as described in any of the preceding embodiments.

[0015] In some embodiments, a computer-readable storage medium is also disclosed, wherein instructions are stored on the computer-readable storage medium, characterized in that, when executed by a processor, the instructions implement the various steps of the watershed pollution tracing method as described in any of the preceding embodiments.

[0016] The present invention includes, but is not limited to, the following beneficial effects: (1) This scheme transforms pollution source inventory data into model input parameters, realizing the transformation from simple emission statistics to model-driven, establishing a direct correlation between pollution source input and water load output, and solving the problem of source area identification being disconnected from actual ecological hazards; (2) This scheme improves the localization degree and simulation accuracy of the model by coupling the SWAT model with the source inventory data and combining it with SWAT-CUP calibration, and realizes the analysis of pollution contribution within sub-basins by classifying land use types and calculating unit load; (3) This scheme identifies the contribution rate of different pollution source categories by constructing a unit load matrix, reveals their spatial distribution characteristics, clarifies the main source sub-basins of various pollution sources, realizes spatial source tracing of pollution source contribution, and provides spatial targeting for precise policy implementation; (4) This scheme establishes a system integration framework of source inventory optimization model input, model support category analysis, and category analysis feedback optimization of the inventory, and provides an independent quantitative basis for inventory statistics and model calculation by using isotope tracing technology to determine the pollution source contribution rate, and the mutual verification of multiple methods improves the reliability of the overall source tracing results, and the unified matrix processing ensures the consistency of multi-scale calculations. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below.

[0018] Figure 1 This is a flowchart of the watershed pollution source tracing method according to an embodiment of this application; Figure 2 This is a flowchart illustrating the construction process of the target SWAT model in an embodiment of this application; Figure 3 This is a flowchart illustrating the data acquisition process of an embodiment of this application; Figure 4 This is a flowchart illustrating the model input parameter determination process according to an embodiment of this application. Figure 5 This is a flowchart illustrating the calibration of the target SWAT model according to an embodiment of this application; Figure 6 This is a structural block diagram of the watershed pollution tracing device according to an embodiment of this application; Figure 7 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application; Figure 8 This is a spatial distribution diagram of the unit load for the four pollution categories in the examples of embodiments of this application; Figure 9 yes Figure 8 The example shows the joint distribution of unit loads for the four pollution categories. Detailed Implementation

[0019] The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0020] For ease of understanding, the specific process of the embodiments of the present invention will be described below. Figure 1 For the watershed pollution source tracing method of this application embodiment, please refer to... Figure 1 It includes the following steps: S100: Obtain emission data and basic data of target pollution sources in the target watershed.

[0021] The basic data may include, but is not limited to, static background data, such as topography and soil, as well as dynamic driving data, such as meteorological and pollution emission data.

[0022] Furthermore, the emission data of the target pollution source can be referenced. Figure 3 The given flowchart is used to obtain the process steps, which will be described later.

[0023] S102. After inputting the target pollution source emission data and basic data into the target SWAT model, run the target SWAT model to simulate and output the sub-basin pollution load of each sub-basin in the target watershed.

[0024] S104. Determine the total simulated pollution load of the target watershed and the pollution source category system of the target watershed.

[0025] The total pollution load in the watershed simulation is the sum of the pollution loads of multiple sub-watersheds, and the pollution source category system includes multiple pollution source categories.

[0026] Specifically, a pollution source classification system can be determined through source inventory surveys and isotope tracing analysis. Isotope analysis is conducted at the watershed outlet, and its results reflect the overall contribution of various pollution sources flowing into the river network, thereby defining the pollution source categories that need attention at the target watershed level. This classification standard is applicable to the entire target watershed.

[0027] S106. Determine the contribution ratio of different pollution source categories to the total pollution load of the target watershed based on isotope tracing methods.

[0028] Specifically, isotope tracing technology measures the stable isotopic composition of nitrogen and phosphorus in water bodies, combines it with the characteristic isotopic values ​​of each pollution source end-member, and uses a Bayesian mixture model to quantitatively calculate the contribution rate of different pollution sources.

[0029] For example, when the pollution sources include both nitrogen and phosphorus, the nitrogen isotope analysis uses a nitrate nitrogen-oxygen dual isotope system. After filtration through a 0.45 μm filter membrane, the water sample is enriched with nitrate using an anion exchange resin, and NO3 is removed via bacterial denitrification. - It is converted to N2O, and then the δ-values ​​are determined using gas stable isotope ratio mass spectrometry (IRMS). 15 N-NO3 and δ 18 O-NO3 values. By comparing the characteristic value ranges of different pollution sources in Table 1, δ 15 N-NO3 can effectively distinguish between organic and inorganic nitrogen sources, δ 18 O-NO3 is mainly used to identify sources of atmospheric deposition.

[0030] Isotope analysis of phosphorus pollution sources uses phosphate oxygen isotopes (δ¹⁸O₂). 18The core indicator is Ag3PO4. Sample pretreatment requires collecting a large volume of water (typically 35-80 L). Phosphate is enriched by co-precipitation with MgCl2 and NaOH, followed by multiple steps including ammonium phosphomolybdate precipitation, magnesium ammonium phosphate precipitation, and cation exchange resin purification to finally obtain high-purity Ag3PO4 solid. δ¹⁻ is determined using a high-temperature pyrolysis-stable isotope mass spectrometry (TC / EA-IRMS) system. 18 The O-PO4 value analysis accuracy can reach ±0.5‰. Based on the characteristic value range shown in Table 1, combined with water chemistry parameters and watershed characteristics, different phosphorus pollution sources can be effectively distinguished.

[0031] Based on measured isotope data, a Bayesian stable isotope mixing model (such as SIAR or MixSIAR) is used for quantitative source apportionment. The model calculates the posterior probability distribution and uncertainty range of the contribution rate of each pollution source using the Markov Chain Monte Carlo (MCMC) algorithm. For nitrogen pollutants, isotope fractionation effects need to be considered for correction, while phosphates, due to their high stability in the environment, can be directly represented by observed values. Finally, the contribution rate and confidence interval of each pollution source are output, achieving quantitative source apportionment.

[0032] The contribution rates of various pollution sources obtained through isotope analysis are cross-validated with the statistical survey results to determine the contribution ratio of each land use type (such as aquaculture / pasture, urban areas, agriculture, etc.) to the total pollution load, providing a scientific basis for subsequent load allocation calculations.

[0033] Table 1. Isotopic ranges of common nitrogen and phosphorus pollution sources

[0034] S108. Based on the contribution ratio and the total simulated pollution load of the watershed, determine the load vector of each pollution source category in the target watershed.

[0035] Specifically, in this example, based on the output of the target SWAT model, at the sub-basin scale, the pollutant load of the sub-basin is expressed as the sum of the products of the actual area of ​​each pollution category and the corresponding unit load. Batch calculation and subsequent analysis are achieved through matrix representation, as shown in the following formula: ; In the formula: Sub-basin j Total pollutant load (kg / month), j=1,2,3......n; Sub-basin j No. i Actual area of ​​land use category (ha); : No. iLand use unit load (kg / ha / month), i=1,2,3......m; Total number of sub-basins; Total number of land use types; Matrix representation: ; ; Furthermore, the pollution load of a sub-basin is first allocated to different categories according to the proportion of pollution categories in the sub-basin, and then further allocated according to the proportion of pollution load in the sub-basin, so as to realize the matrix expression of the pollution load of the sub-basin.

[0036] The load vectors for pollution categories in each sub-basin are as follows: ; In the formula: Total pollutant load vectors for various land uses in the watershed (kg / month); Total pollutant load in the watershed (kg / month); : No. i Land use category proportion, when i=1, When i=m, .

[0037] S110. Based on the load vectors of each pollution source category in the target watershed and the proportion of pollution load in each sub-watershed to the total simulated pollution load of the watershed, construct the sub-watershed load allocation matrix.

[0038] Specifically, the sub-basin load allocation matrix is ​​as follows: ; In the formula: : m × n Sub-basin load allocation matrix; Sub-basin j The Middle i Land use pollutant load (kg / month); Sub-basin j Total pollutant load (kg / month), j=1, When j=n, .

[0039] S112. Divide each element in the sub-basin load allocation matrix by the actual area corresponding to the pollution category associated with that pollution source category in the corresponding sub-basin to obtain the unit load matrix.

[0040] Specifically, the calculation of the unit load matrix depends on the proportion of each land use type's contribution to the total load, which is determined using the aforementioned isotope tracing technique. Based on δ... 15 N-NO3, δ 18 O-NO3 (nitrogen pollution) or δ 18 The characteristic value range of O-PO4 (phosphorus pollution) was quantitatively analyzed using a Bayesian mixture model to obtain the contribution rate of different land use types such as aquaculture / pasture, urban areas, and agriculture, providing a scientific quantitative basis for load allocation and unit load calculation.

[0041] The unit loads of various types in a sub-basin are obtained by dividing the corresponding pollutant load of the sub-basin by the actual area, and output as a matrix, as shown in the following formula: ; In the formula: Sub-basin j The Middle i Land use unit load (kg / ha / month); Matrix form: ; In the formula: matrix U Each row corresponds to a sub-basin, and each column corresponds to a land use type.

[0042] S114. Analyze the unit load matrix to obtain the dominant pollution category for each sub-basin.

[0043] Specifically, through unit load matrix analysis, the spatial distribution characteristics of various pollution source categories can be systematically identified. First, high-load sub-basins are identified, and those with unit loads exceeding a threshold are selected as key control areas. Second, the dominant pollution sources are located to determine the dominant pollution source categories for each sub-basin. Simultaneously, spatial clustering analysis is performed to identify spatial clustering areas of similar pollution sources. Finally, by analyzing the unit load matrix, the dominant pollution category for each sub-basin can be traced back to its source.

[0044] The specific analysis is as follows: Based on the above analysis, the rows of the unit load matrix represent each specific sub-basin, and the columns represent different pollution source categories (e.g., agricultural non-point sources, livestock and poultry farming, urban domestic pollution, rural domestic pollution, industrial point sources, etc.). Each element in the matrix represents the unit area load (e.g., kg / ha / year) of pollutants generated by a certain type of pollution source within a specific sub-basin. This value standardizes the area impact and reflects the pollution emission intensity.

[0045] Further analysis and source tracing are conducted using both horizontal and vertical comparisons. Vertical comparisons are used to identify key areas, specifically high-load sub-basins. Specifically, the total unit load of all pollution categories in each sub-basin (row) can be calculated by summing the rows. A threshold (such as the basin-wide average or a back-calculated value based on water quality management targets) is then set to filter out the sub-basins with the highest total unit load. These sub-basins are pollution load hotspots within the basin, representing key spatial control areas, resulting in a pollution hotspot distribution map to illustrate the pollution situation in each sub-basin. Horizontal comparisons are used to identify key pollution types and locate dominant pollution sources. Specifically, for each sub-basin (each row), the unit load values ​​of each pollution source category (each column) within it are compared horizontally to identify the pollution source category with the highest unit load value or the top two (e.g., the first two) within that sub-basin. This category is the dominant pollution source for that sub-basin.

[0046] Furthermore, by overlaying the above longitudinal and transverse comparison results, the dominant pollution sources in the selected high-load hotspot sub-basins can be identified. For example, if the dominant source of a high-load sub-basin is found to be "livestock and poultry farming," then the main reason for its high load points to farming activities. Further, all sub-basins with the same dominant pollution source category can be rendered with the same color or pattern on the basin map to conduct spatial pattern analysis and identify clusters of similar pollution sources. After identifying these clusters, the precision of control can be improved. For example, for high-load sub-basins dominated by agriculture, the focus of control resources can be on controlling non-point source pollution in farmland (such as ecological interception ditches and fertilizer reduction and efficiency enhancement); for areas dominated by livestock farming, the focus can be on controlling livestock manure. Furthermore, the entire basin can be differentiated based on the spatial distribution of dominant pollution sources, such as dividing the spatial distribution into key agricultural control areas, key livestock farming control areas, and urban and rural domestic pollution control areas, to achieve zoned control.

[0047] In another example, a spatial distribution map of contribution rates can be drawn to visually demonstrate the spatial distribution pattern of contribution rates from various pollution sources.

[0048] Furthermore, Figure 2 This is a flowchart illustrating the construction process of the target SWAT model in an embodiment of this application. (See attached document.) Figure 2 It includes the following steps: S200. Spatially allocate the emission data of the target pollution source to obtain the model input parameters.

[0049] Specifically, this example provides an exemplary illustration of a flowchart for determining model input parameters; please refer to the following text for details. Figure 4 Introduction.

[0050] S202. Based on the digital elevation model, watershed division is carried out, and an initial SWAT model framework is constructed.

[0051] Specifically, the SWAT model is a physically based, semi-distributed watershed hydrological and water quality model capable of simulating the hydrological cycle and pollutant migration processes in complex watersheds. The model construction process begins with watershed delineation based on the DEM (Digital Elevation Model), generating sub-watersheds and hydrological response units (HRUs). Subsequently, basic data such as land use, soil type, and meteorological data are input, and point source and non-point source pollution inputs are set based on source inventory data. Finally, the model is run to obtain the flow and pollutant load outputs for each sub-watershed. The initial SWAT model framework is then constructed based on multiple sub-watersheds and multiple hydrological response units after watershed delineation.

[0052] S204. Input the model input parameters and basic data into the initial SWAT model framework to adjust the original parameters of the initial SWAT model and obtain an optimized SWAT model with optimized parameters.

[0053] S206. Determine the model calibration parameters, and calibrate the optimized SWAT model based on the model calibration parameters to obtain the target SWAT model.

[0054] Specifically, spatial allocation refers to converting target pollution source emission data into spatial raster data to meet the input requirements of the SWAT model. This process employs different processing methods for point sources and area sources, as detailed below: a) Point sources have clear location identifiers, and their corresponding grids can be directly located based on their latitude and longitude coordinates in their geographic location information. This geographic location information is collected during the survey phase.

[0055] b) Area source emissions are first identified by administrative region, and then allocated to each sub-basin using a spatial allocation coefficient method. Specifically, the emission data is spatialized to the sub-basin scale using raster data of proxy parameters directly related to the emission amount (such as total population, urban population, rural population, rainfall intensity under a specific underlying surface, etc.), and the allocation weight is determined according to the area proportion of each raster in the sub-basin, thereby completing the spatial allocation of area source emissions.

[0056] Furthermore, source inventory data, i.e., pollution source emission data of the target watershed, is used as input parameters for the SWAT model to optimize the model's regional settings. First, the emissions of various pollution sources are converted into point source and area source input data required by the model. Then, the land use parameters and management measure parameters of the model are adjusted according to the spatial distribution of pollution sources. Finally, the nutrient input intensity of the model is corrected using inventory data to improve the model's localization. Through the optimized model input from the source inventory, the SWAT model can more accurately reflect the actual pollution characteristics of the watershed, improving simulation accuracy.

[0057] Specifically, Figure 3 This is a flowchart illustrating the data acquisition process of an embodiment of this application. The flowchart details the implementation of step S100. (See attached document.) Figure 3 It includes the following steps: S300: Obtain the point sources and area sources of pollution in the target watershed.

[0058] Among them, point sources of pollution refer to pollution sources with fixed emission points that are easy to monitor and regulate, such as industrial enterprises and urban sewage treatment plants. Non-point sources of pollution refer to pollution sources that are dispersed and have no fixed emission outlets, such as agricultural non-point sources, livestock and poultry farming, rural domestic pollution, and aquaculture.

[0059] S302. Point source pollution emission data determined based on online monitoring methods and area source pollution emission data determined based on the emission coefficient method.

[0060] The emission coefficients used in the emission coefficient method are target coefficients that are corrected from the original coefficients of the emission coefficient method and matched with the target watershed.

[0061] Specifically, emission coefficients should be obtained primarily through local measurements. The original coefficients should be corrected based on the emission coefficients obtained from local measurements. When no measurement conditions are available, coefficient manuals or similar materials can be consulted for correction.

[0062] S304. Based on point source pollution emission data and area source pollution emission data, determine the pollution source emission data of the target watershed. The point source pollution emission data and area source pollution emission data each have their own geographical location information and attribute information.

[0063] Specifically, Figure 4 For a description of the process for determining model input parameters in embodiments of this application, please refer to... Figure 4 It includes the following steps: S400: Based on the geographic location information of point source pollution emission data, spatialize the point source pollution emission data to its corresponding target sub-basin.

[0064] Specifically, the latitude and longitude coordinates of point sources and their pollutant emissions, such as annual TN / TP emissions in tons, can be obtained. In a Geographic Information System (GIS), each point source is located to a specific spatial point based on its coordinates, and the emissions from that point source are entirely assigned to the target sub-basin where it is located. For example, if a wastewater treatment plant is located within sub-basin A, its annual COD emissions of 100 tons will be entirely included in the point source load of sub-basin A. This step achieves point-to-regional assignment of point source data.

[0065] S402, Marking surface source pollution emission data based on surface source identification.

[0066] Specifically, each type of non-point source pollution data, such as nitrogen loss from agricultural fertilizers and phosphorus discharge from rural domestic sewage, is labeled as a non-point source. This label is used to distinguish it from point sources in order to associate it with the correct spatial allocation weights. For example, data labeled with agricultural fertilizers will subsequently be allocated using arable land distribution maps, etc.

[0067] S404. Determine the raster data of the correlation parameters related to the area source pollution emission data.

[0068] Specifically, the purpose of this step is to establish a mapping relationship between non-point source types and associated parameters. Associated parameters include, but are not limited to, land use type raster maps for agricultural non-point sources, livestock and poultry farming quantity distribution maps or land use type maps (such as pasture / farmland land) for livestock and poultry farming non-point sources, and rural population distribution raster maps for rural domestic sources. Each pixel value in the raster data represents the intensity or probability of that type of pollution activity occurring at that location, serving as the basis for calculating the assigned weights.

[0069] S406. For each target sub-basin, determine the allocation weight of each grid cell based on the associated parameter grid data of each grid cell falling within its range and the area ratio of that grid cell in the target sub-basin.

[0070] Specifically, for each target sub-basin, all grid cells that fall completely or partially within its boundary are identified. For each grid cell, its corresponding association parameter value is obtained; for example, if fertilizer is allocated, the value is 1 if the grid cell is farmland, and 0 otherwise. Simultaneously, the area proportion of this grid cell within the sub-basin is calculated. For instance, if a grid cell occupies 60% of its area within the sub-basin, the proportion is set to 0.6. The association parameter value of this grid cell is multiplied by the area proportion to obtain a preliminary contribution value. The contribution values ​​of all grid cells within the sub-basin are normalized so that the sum of the weights of all grid cells is 1, ultimately yielding the final weight of each grid cell for allocating this type of non-point source pollution within the sub-basin.

[0071] S408. Based on the allocation weight, the non-point source pollution emission data are allocated to each target sub-basin.

[0072] Specifically, to obtain the total emissions of a type of non-point source pollution that needs to be weighted in the target area, for each target sub-basin, the total emissions are multiplied by the sum of the weights of all grids in that sub-basin to obtain the emissions data assigned to that sub-basin.

[0073] Furthermore, Figure 5 This is a flowchart illustrating the calibration of the target SWAT model according to an embodiment of this application. The flowchart is an exemplary description of step S206. (See attached document.) Figure 5 It includes the following steps: S500. From the parameters of the optimized SWAT model, select model calibration parameters that are sensitive to hydrological and water quality processes. The model calibration parameters include parameters CN2, ALPHA_BF, and SOL_AWC that affect hydrological processes, and parameters ERORGN and NPERCO that affect water quality processes.

[0074] S502. Set the objective function, which should include at least the Nash efficiency coefficient (NSE) and the coefficient of determination (R²).

[0075] S504. The SUFI-2 algorithm is used for parameter optimization iteration. By adjusting the model calibration parameters, the objective function between the simulated output value of the optimized SWAT model and the measured value of the target watershed is made to reach the target optimization range.

[0076] S506. Verify the model that has been automatically calibrated. When the verification simulation results meet the preset performance standards, the target SWAT model is obtained.

[0077] The performance criteria include a Nash efficiency coefficient (NSE) greater than the target efficiency coefficient threshold and a determination coefficient greater than the target determination coefficient threshold.

[0078] That is, this step uses SWAT-CUP (SWAT Calibration and Uncertainty Programs) for automatic calibration of model parameters. This process first selects sensitivity parameters, including hydrological parameters such as CN2, ALPHA_BF, and SOL_AWC, and water quality parameters such as ERORGN and NPERCO. Then, it sets the NSE and the coefficient of determination R0. 2 The objective function is defined, and the parameters are optimized using the SUFI-2 algorithm. Through multiple iterations, the simulated values ​​are optimally matched with the measured values. Finally, the model reliability is verified during the validation period to ensure that NSE > 0.5 and R0 > 0.5. 2 >0.6. The calibrated and validated SWAT model can accurately simulate the nitrogen and phosphorus loads of each sub-basin, providing reliable load data for subsequent analysis.

[0079] Furthermore, to facilitate understanding of the proposed solution, the following example is provided: This embodiment takes watershed A as the research object and uses the watershed pollution source tracing method of the present invention to calculate and analyze the total phosphorus (TP) pollution load. The watershed is divided into 88 sub-watersheds (numbered 1-88), and the method of the present invention is implemented systematically.

[0080] Furthermore, a dual-track accounting method was adopted to systematically investigate various pollution sources in Basin A. Emission data from key polluting units such as industrial enterprises and urban wastewater treatment plants were primarily obtained from online monitoring systems, while data from dispersed pollution sources such as agricultural non-point sources, livestock and poultry farming, and rural domestic pollution were calculated based on activity level data from on-site investigations, combined with locally measured emission coefficients or those from standard manuals. This resulted in a city-level pollution source emission inventory covering the entire basin, providing fundamental data for total phosphorus (TP) load calculation, as shown in Table 2.

[0081] Table 2. Example of a city-level pollution source list covering the entire river basin ; Furthermore, the source inventory data, after format conversion and spatial configuration, is input into the SWAT model. Point source pollution is directly loaded using coordinate positioning, while non-point source pollution is achieved by adjusting the nutrient input parameters of the corresponding HRUs. This parameter setting based on actual survey data significantly improves the regional applicability of the model.

[0082] Furthermore, a SWAT model for watershed A was constructed according to the method of this invention. Based on DEM data, 88 sub-watersheds were divided, and the model was completed by combining basic data such as land use, soil, and meteorology, as well as pollution input parameters provided by the source inventory. Calibration was performed using SWAT-CUP. The model's NSE coefficient was greater than 0.5 during both the calibration and validation periods, and the coefficient of determination R0 was [value missing]. 2 A value exceeding 0.6 allows for reliable simulation of TP load in each sub-basin. Land use data for each sub-basin is normalized to ensure the sum of the proportions is 1, for example: The original proportion of sub-basin 1 was 99.95%. After classification and normalization, the proportions of aquaculture / pasture, urban, agricultural, and other land use were 5.90%, 0.78%, 11.69%, and 81.63%, respectively. The original proportion of sub-basin 4 was 100.01%, and after classification and normalization, the proportions of urban, agricultural and other land use were 36.06%, 33.58% and 30.37%, respectively.

[0083] The actual area is then calculated to provide an accurate spatial data basis for load distribution.

[0084] Furthermore, to determine the contribution ratio of various pollution sources to the total load, 35L large-volume water samples were collected from representative sub-basins of watershed A, including high-load aquaculture areas, urban pollution areas, and agricultural pollution areas, for phosphate oxygen isotope (δ¹²) analysis.18 O-PO4) analysis. After enrichment and purification of water samples to obtain high-purity Ag3PO4 solid, the isotope values ​​were determined using a TC / EA-IRMS system with an analytical precision of ±0.5‰.

[0085] The measurement results showed that the δ of the aquaculture source 18 O-PO4 values ​​were concentrated in the range of 12.4-18.7‰, reflecting the mixed characteristics of animal feces and feed; agricultural sources were mainly distributed in the range of 8.5-14.8‰, reflecting a typical signal of fertilizer application; urban sources showed a comprehensive characteristic of 12.4-18.7‰. Quantitative analysis was performed using the MixSIAR Bayesian mixture model, and the posterior probability distribution of the contribution rate of each pollution source was generated using the MCMC algorithm. Finally, the contribution ratio of the four land use categories to TP load was determined, as shown in Table 3.

[0086] Table 3. Proportion of Nitrogen, Phosphorus, and Phosphorus Loading ; Furthermore, the actual area of ​​each land use type in each sub-basin is calculated based on the normalized land use ratio. The calculation method is the product of the normalized ratio and the total area of ​​the sub-basin.

[0087] Secondly, the matrix allocation method proposed in this invention is used to allocate the TP load of each sub-basin according to land use type. For each sub-basin, the TP load of each land use type is equal to the TP load of that sub-basin multiplied by the load allocation ratio of that land use type. According to the ratios shown in Table 3, aquaculture / ranchland land is allocated 40% of the load, urban land is allocated 16%, agricultural land is allocated 35%, and other land is allocated 9%.

[0088] Finally, based on this, the unit area load of each land use type in each sub-basin is further calculated. That is, the unit load is equal to the TP load allocated to that land use type divided by the actual area of ​​that land use type, and the unit is kg / ha / month. For some sub-basins where a certain land use type does not exist (actual area is zero), the corresponding unit load is marked as "not applicable" or "0". The specific results are shown in Table 4.

[0089] Table 4 Sub-basin unit load matrix (unit: kg / ha / month)

[0090] ; Through the above calculation process, an 88×4 unit load matrix is ​​generated. Each row in the matrix represents a sub-basin, and each column represents a land use type. The matrix elements represent the unit nitrogen and phosphorus (TP) loads for the corresponding land use type within the sub-basin. Based on this matrix system, pollution source contribution analysis and key source area identification are conducted. (1) Assessment of contribution rate at the whole basin scale Based on the load allocation results, the contribution characteristics of various land uses to TP pollution in watershed A are as follows: Aquaculture / pasture land: The total load contribution rate is 40%, approximately 1.9066 million kg / month. The unit load fluctuates significantly between different sub-basins, ranging from 0.68 to 128.45 kg / ha / month. The maximum and minimum values ​​differ by nearly 190 times, reflecting the spatial heterogeneity of aquaculture activity intensity. Agricultural land: The total load contribution rate is 35%, approximately 1.6683 million kg / month, with a unit load range of 0.70-10.41 kg / ha / month. The spatial distribution is relatively uniform, but there are high values ​​in some local areas, reflecting regional differences in planting structure and fertilization intensity. Urban land: The total load contribution rate is 16%, approximately 762,700 kg / month, with a unit load range of 0.11-48.86 kg / ha / month. The load is generally low, but there are extremely high values ​​in some sub-basins, reflecting the localized concentration of urban runoff pollution. Other land use: The total load contribution rate is 9%, approximately 429,000 kg / month. The unit load is relatively stable and generally low, and the pollution contribution is the smallest among the four types of land use.

[0091] (2) Identification of key polluted sub-basins Through extreme value analysis of the unit load matrix and Figure 8 The spatial comparison of the unit load spatial distribution of the four pollution categories shown identifies the following key pollution source areas: Sub-basin 10: The unit load of livestock / farm reached 128.45 kg / ha / month, the highest in the entire basin, indicating that there are extremely dense livestock activities or highly concentrated livestock emissions in this area, and source control measures should be prioritized. Sub-basin 62: The unit load of aquaculture / farmland is 86.04 kg / ha / month, ranking second in the entire basin, which also reflects the severity of aquaculture pollution. It should be treated in conjunction with sub-basin 10. Sub-basin 52: The urban land use unit load reached 48.86 kg / ha / month, which is the area with the most prominent urban pollution in the entire basin. It is necessary to focus on strengthening rainwater runoff control, improving the urban sewage collection and treatment system, and controlling non-point source pollution. Sub-basin 21: The urban land use unit load was 17.55 kg / ha / month, which was lower than that of sub-basin 52 but still significantly higher than the average level of the basin, indicating obvious urban non-point source pollution characteristics; Sub-basins 8 and 88: The unit load of agricultural land reached 10.41 and 10.28 kg / ha / month respectively, which are the areas with the most prominent agricultural pollution in the basin. This may be related to high-intensity planting, excessive fertilization or special farming methods, and agricultural production management measures need to be optimized.

[0092] (3) Spatial distribution characteristics analysis based on Figure 9 As shown, spatial analysis of the unit load matrix indicates that: High load concentration area: High unit load sub-basins are mainly distributed in specific areas of the basin. Among them, the pollution from aquaculture / farm in sub-basins 10 and 62, the urban pollution in sub-basin 52, and the agricultural pollution in sub-basins 8 and 88 all reach extremely high levels in the basin. Low-load concentrated area: In some sub-basins (such as sub-basins 1, 7, 16, etc.), the load of various land use units is low. The load of aquaculture / pastoral is mostly below 5 kg / ha / month, the load of urban and agricultural is mostly below 2 kg / ha / month, and the load of other land use is generally below 0.1 kg / ha / month. The ecological background is good and has a strong environmental carrying capacity. The spatial heterogeneity is extremely significant: the unit load of the same land use in different sub-basins can vary by more than 100 times (e.g., aquaculture / ranchage ranges from 0.68 to 128.45 kg / ha / month), reflecting the high spatial heterogeneity of watershed pollution and providing a scientific basis for precise zoning and differentiated management.

[0093] Through the above systematic analysis, a complete source tracing chain was realized from total load to sub-basin load and then to land use contribution, clarifying key pollution source areas and main pollution types, and providing reliable quantitative technical support for the scientific formulation of water environment zoning management, precise governance and pollution prevention and control measures in Basin A.

[0094] The unit load matrix output by this method provides a clear quantitative basis and scientific management suggestions for the precise control of TP pollution in watershed A: (1) Precision control strategy for aquaculture / farm source This study targets high-load areas of aquaculture / ranches, such as sub-basin 10 (unit load 128.45 kg / ha / month), sub-basin 62 (86.04 kg / ha / month), and sub-basin 77 (80.74 kg / ha / month).

[0095] Recommendations: Prioritize the implementation of projects for the resource utilization of livestock manure, promote the integrated farming and breeding ecological cycle model, strictly control the scale and density of livestock farming, and establish centralized treatment facilities for livestock waste.

[0096] (2) Urban source system governance plan For sub-basin 52 (unit load 48.86 kg / ha / month), sub-basin 71 (47.99 kg / ha / month), and sub-basin 21 (17.55 kg / ha / month), which are areas with high urban land use load.

[0097] Recommendations: Strengthen the management of rainwater runoff purification and discharge, improve the urban sewage collection and treatment system, promote the concept of sponge city construction, and implement initial rainwater interception and purification measures.

[0098] (3) Agricultural source classification and policy measures For sub-basin 8 (unit load 10.41 kg / ha / month), sub-basin 88 (10.28 kg / ha / month), and sub-basin 45 (8.21 kg / ha / month), which are areas with high agricultural land load, the study focuses on these areas.

[0099] Recommendations: Promote soil testing and formula fertilization technology, optimize the application amount and timing of nitrogen and phosphorus fertilizers, construct ecological interception ditch systems, and develop green ecological agricultural models.

[0100] (4) Multi-source collaborative land use optimization In high-load sub-basins, land use structure should be appropriately adjusted to increase the proportion of ecological land such as forest land, grassland, and wetlands, and a composite ecological buffer zone of farmland-forest land-water bodies should be constructed to enhance the natural purification capacity and environmental carrying capacity of the basin.

[0101] This scheme transforms pollution source inventory data into model input parameters, shifting from simple emission statistics to model-driven approaches. It establishes a direct correlation between pollution source input and water body load output, resolving the disconnect between source area identification and actual ecological hazards. By coupling the SWAT model with source inventory data and combining it with SWAT-CUP calibration, the scheme improves model localization and simulation accuracy. Through land use classification and unit load calculation, it analyzes pollution contributions within sub-basins. By constructing a unit load matrix, it identifies the contribution rates of different pollution source categories, revealing their spatial distribution characteristics and clarifying the main source sub-basins for various pollution sources, achieving spatial source tracing of pollution contributions and providing spatial targeting for precise policy implementation. By establishing a system integration framework that optimizes model input from the source inventory, supports category analysis through the model, and uses category analysis feedback to optimize the inventory, the pollution source contribution rate determined by isotope tracing technology provides independent quantitative evidence for inventory statistics and model calculations. Multiple methods mutually verify each other, improving the reliability of the overall source tracing results, and unified matrix processing ensures consistency across multi-scale calculations.

[0102] Furthermore, Figure 6 The above is a structural block diagram of the watershed pollution source tracing device according to the application embodiment, such as... Figure 6 As shown, the device includes: The data acquisition module is used to acquire emission data of target pollution sources in the target watershed, as well as basic data. The simulation output module is used to input the target pollution source emission data and basic data into the target SWAT model, run the target SWAT model, simulate and output the sub-basin pollution load of each sub-basin in the target watershed; The category system determination module is used to determine the total simulated pollution load of the target watershed and the pollution source category system of the target watershed. The total simulated pollution load of the watershed is the sum of the pollution loads of multiple sub-watersheds, and the pollution source category system includes multiple pollution source categories. The contribution ratio determination module is used to determine the contribution ratio of different pollution source categories to the total pollution load of the target watershed based on the isotope tracing method. The load vector determination module is used to determine the load vector of each pollution source category in the target watershed based on the contribution ratio and the total simulated pollution load of the watershed. The sub-basin load allocation matrix construction module is used to construct the sub-basin load allocation matrix based on the load vectors of each pollution source category in the target watershed and the proportion of pollution load in each sub-basin to the total simulated pollution load of the watershed. The unit load matrix determination module is used to divide each element in the sub-basin load allocation matrix by the actual area corresponding to the pollution category associated with that pollution source category in the corresponding sub-basin to obtain the unit load matrix; The dominant pollution category tracing module is used to parse the unit load matrix and trace the dominant pollution category of each sub-basin.

[0103] The application of the relevant modules of the device in this example can be referred to the relevant introduction of the method principle above, and will not be repeated here.

[0104] above Figure 6 The watershed pollution tracing device in this embodiment of the invention will be described in detail from the perspective of modular functional entities. The electronic equipment in this embodiment of the invention will be described in detail from the perspective of hardware processing.

[0105] Figure 7This is a schematic diagram of the structure of an electronic device 700 provided in an embodiment of the present invention. The electronic device 700 can vary significantly due to differences in configuration or performance, and may include one or more central processing units (CPUs) 710 (e.g., one or more processors) and a memory 720, and one or more storage media 730 (e.g., one or more mass storage devices) for storing application programs 733 or data 732. The memory 720 and storage media 730 can be temporary or persistent storage. The program stored in the storage media 730 may include one or more modules (not shown in the diagram), each module including a series of instruction operations on the electronic device 700. Furthermore, the processor 710 may be configured to communicate with the storage media 730 and execute the series of instruction operations in the storage media 730 on the electronic device 700.

[0106] Electronic device 700 may also include one or more power supplies 740, one or more wired or wireless network interfaces 750, one or more input / output interfaces 760, and / or one or more operating systems 731, such as Windows Server, MacOSX, Unix, Linux, FreeBSD, etc. Those skilled in the art will understand that... Figure 7 The illustrated electronic device structure does not constitute a limitation on electronic devices and may include more or fewer components than illustrated, or combine certain components, or have different component arrangements.

[0107] The present invention also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein the computer-readable storage medium stores instructions that, when executed on a computer, cause the computer to perform the steps of any of the above-described watershed pollution source tracing methods.

[0108] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the system, device, or unit described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0109] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0110] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for tracing the source of pollution in a watershed, characterized in that, The method includes: Acquire emission data and basic data of target pollution sources in the target watershed; After inputting the target pollution source emission data and basic data into the target SWAT model, the target SWAT model is run to simulate and output the sub-basin pollution load of each sub-basin in the target watershed. The total simulated pollution load of the target watershed and the pollution source category system of the target watershed are determined. The total simulated pollution load of the watershed is the sum of the pollution loads of multiple sub-watersheds, and the pollution source category system includes multiple pollution source categories. The contribution ratio of different pollution source categories to the total pollution load of the target watershed was determined based on the isotope tracing method. Based on the contribution ratio and the total simulated pollution load of the watershed, the load vector of each pollution source category in the target watershed is determined; Based on the load vectors of each pollution source category in the target watershed and the proportion of pollution load in each sub-watershed to the total simulated pollution load of the watershed, a sub-watershed load allocation matrix is ​​constructed. Divide each element in the sub-basin load allocation matrix by the actual area corresponding to the pollution category associated with that pollution source category in the corresponding sub-basin to obtain the unit load matrix; By analyzing the unit load matrix, the dominant pollution category of each sub-basin can be obtained.

2. The watershed pollution source tracing method according to claim 1, characterized in that, After acquiring the target pollution source emission data and basic data for the target watershed, the method further includes constructing a target SWAT model, including: Spatially allocate the emission data of the target pollution source to obtain the model input parameters; Watershed division is carried out based on digital elevation model, and an initial SWAT model framework is constructed. The initial SWAT model framework is constructed based on multiple sub-watersheds and multiple hydrological response units after watershed division. The model input parameters and the basic data are input into the initial SWAT model framework to adjust the original model parameters in the initial SWAT model, thereby obtaining an optimized SWAT model with model optimization parameters. Determine the model calibration parameters, and calibrate the optimized SWAT model based on the model calibration parameters to obtain the target SWAT model.

3. The watershed pollution source tracing method according to claim 1, characterized in that, The acquisition of pollution source emission data for the target watershed includes: Obtain information on point sources and non-point sources of pollution in the target watershed; The point source pollution emission data of the pollution point source is determined based on the online monitoring method, and the area source pollution emission data of the pollution area source is determined based on the emission coefficient method. The emission coefficient used in the emission coefficient method is a target coefficient that matches the target watershed after the original coefficient of the emission coefficient method is corrected. Based on the point source pollution emission data and the area source pollution emission data, the pollution source emission data of the target watershed is determined, wherein the point source pollution emission data and the area source pollution emission data each have their own geographical location information and attribute information.

4. The watershed pollution source tracing method according to claim 2, characterized in that, The spatial allocation of the target pollution source emission data to obtain model input parameters includes: Based on the geographic location information of point source pollution emission data, the point source pollution emission data is spatialized to its corresponding target sub-basin; The spatial allocation coefficient method is used to spatialize the area source pollution emission data to its corresponding target sub-basin, including: Data on area source pollution emissions are labeled based on area source identification. Determine the raster data of the correlation parameters related to area source pollution emission data; For each target sub-basin, the allocation weight of each grid cell is determined based on the associated parameter grid data of each grid cell falling within its range and the area ratio of that grid cell in the target sub-basin. Based on the allocation weights, the non-point source pollution emission data are allocated to each target sub-basin.

5. The watershed pollution source tracing method according to claim 2, characterized in that, The process of determining model calibration parameters and calibrating the optimized SWAT model based on these parameters to obtain the target SWAT model includes: The optimized SWAT model was automatically calibrated using the SWAT-CUP tool. The automatic calibration process includes: From the parameters of the optimized SWAT model, select model calibration parameters that are sensitive to hydrological and water quality processes. The model calibration parameters include parameters CN2, ALPHA_BF, and SOL_AWC that affect hydrological processes, and parameters ERORGN and NPERCO that affect water quality processes. Set an objective function, which includes at least the Nash efficiency coefficient (NSE) and the coefficient of determination (R²). 2 ; The SUFI-2 algorithm is used for parameter optimization iteration. By adjusting the model calibration parameters, the objective function between the simulated output value of the optimized SWAT model and the measured value of the target watershed is made to reach the target optimization range. The model that has been automatically calibrated is validated. When the validation simulation results meet the preset performance standards, the target SWAT model is obtained. The performance standards include a Nash efficiency coefficient (NSE) greater than the target efficiency coefficient threshold and a determination coefficient greater than the target determination coefficient threshold.

6. The watershed pollution source tracing method according to claim 1, characterized in that, Determining the actual area corresponding to the pollution category includes: Obtain the original area proportion of the target pollution category in each sub-basin within the target watershed; The original area ratios are normalized to determine the actual area corresponding to the target pollution category in each sub-basin. The normalization formula is as follows: ； In the formula: Sub-basin s No. i Land use normalization ratio; Sub-basin s No. i Original proportions of land use types; Actual area calculation formula: ； In the formula: Sub-basin s No. i Actual area of ​​land use category (ha); Sub-basin s Total area (ha).

7. The watershed pollution source tracing method according to claim 1, characterized in that, The total simulated pollution load of the watershed is determined based on the following formula: ； In the formula: Sub-basin j Total pollutant load (kg / month), j=1,2,3......n; Sub-basin j No. i Actual area of ​​land use category (ha); : No. i Land use unit load (kg / ha / month), i=1,2,3......m; Total number of sub-basins; Total number of land use types; Matrix representation: ； ； The load vector is expressed based on the following formula: ； In the formula: Total pollutant load vectors for various land uses in the watershed (kg / month); Total pollutant load in the watershed (kg / month); : No. i Land use category proportion, when i=1, When i=m, ; The sub-basin load allocation matrix is ​​expressed based on the following formula: ； In the formula: : m × n Sub-basin load allocation matrix; Sub-basin j The Middle i Land use pollutant load (kg / month); Sub-basin j Total pollutant load (kg / month), j=1, When j=n, ; The unit load matrix is ​​expressed based on the following formula: ； In the formula: Sub-basin j The Middle i Land use unit load (kg / ha / month); Matrix form: ； In the formula: matrix U Each row corresponds to a sub-basin, and each column corresponds to a land use type.

8. A watershed pollution source tracing device, characterized in that, The device includes: The data acquisition module is used to acquire emission data of target pollution sources in the target watershed, as well as basic data. The simulation output module is used to input the emission data of the target pollution source and the basic data into the target SWAT model, run the target SWAT model, simulate and output the pollution load of each sub-basin in the target watershed; The category system determination module is used to determine the total simulated pollution load of the target watershed and the pollution source category system of the target watershed. The total simulated pollution load of the watershed is the sum of the pollution loads of multiple sub-watersheds, and the pollution source category system includes multiple pollution source categories. The contribution ratio determination module is used to determine the contribution ratio of different pollution source categories to the total pollution load of the target watershed based on the isotope tracing method. The load vector determination module is used to determine the load vector of each pollution source category in the target watershed based on the contribution ratio and the total simulated pollution load of the watershed. The sub-basin load allocation matrix construction module is used to construct the sub-basin load allocation matrix based on the load vectors of each pollution source category in the target watershed and the proportion of pollution load in each sub-basin to the total simulated pollution load of the watershed. The unit load matrix determination module is used to divide each element in the sub-basin load allocation matrix by the actual area corresponding to the pollution category associated with the pollution source category in the corresponding sub-basin to obtain the unit load matrix; The dominant pollution category tracing module is used to parse the unit load matrix and trace the dominant pollution category of each sub-basin.

9. An electronic device, characterized in that, The electronic device includes a memory and at least one processor, the memory storing instructions; the at least one processor invokes the instructions in the memory to cause the electronic device to perform the various steps of the watershed pollution tracing method as described in any one of claims 1-7.

10. A computer-readable storage medium storing instructions thereon, characterized in that, When the instructions are executed by the processor, they implement the various steps of the watershed pollution tracing method as described in any one of claims 1-7.

Images