Comprehensive quantitative traceability method for multi-source pollution of karst underground river system

Abstract

The invention discloses a karst underground river system multi-source pollution comprehensive quantitative traceability method, which comprises the following steps: S1, multi-dimensional sample collection and new pollutant screening: carrying out multi-period sampling on a water body, measuring conventional water quality indexes and nitrogen and oxygen isotopes, screening new pollutants, and carrying out targeted quantification on characteristic new pollutants; s2, principal component analysis dimension reduction and pollution factor identification: carrying out principal component analysis on conventional indexes and new pollutant data, extracting principal components of which the cumulative contribution rate exceeds 85%, identifying composite pollution factors, and eliminating secondary interference factors; s3, nitrogen and oxygen isotope auxiliary verification and denitrification elimination: eliminating denitrification interference by utilizing the ratio of delta 15N-NO3 <-> to delta 18O-NO3 <-> in combination with underground water burial depth and runoff velocity, and distinguishing the nitrate source as a chemical fertilizer, soil nitrogen or sewage and wastewater; and S4, PMF model optimization and pollution source quantification: inputting the conventional indexes and the new pollutant concentration matrix into a PMF model, and analyzing the contribution rate of each pollution source.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of groundwater pollution, and particularly relates to a comprehensive quantitative tracing method for multi-source pollution of karst underground river system. BACKGROUND

[0002] Karst landform accounts for 1 / 3 of the land area of China, among which karst development is most extensive in Guizhou Province. The underground river system in Guizhou has the typical characteristics of "frequent conversion of three waters, high ecological vulnerability, and strong pollution concealment". The research area (a certain underground river system in northern Guizhou) is a microcosm of karst areas, and is facing the problem of superimposed multi-source pollution such as agricultural non-point source pollution, alcohol enterprise wastewater, aquaculture wastewater, and domestic sewage. The indicators such as NH4 + , NO3 - -N, and TP in the groundwater quality exceed the standard by 3-954 times. Such pollution has the characteristics of "complex source, hidden migration path, and coexistence of new pollutants (such as antibiotics and organic acids)", and the traditional single tracing method cannot meet the precise analysis requirements.

[0003] The current tracing methods include the following types: (1) stable isotope technology (such as δ 15 N-NO3 - , δ 18 O-NO3 - ) quantifies the contribution of pollution sources by constructing an isotope model, but it faces two major bottlenecks in karst areas: the nitrogen and oxygen isotope value ranges of agricultural non-point source pollution (fertilizer), aquaculture wastewater (livestock and poultry manure), and alcohol enterprise wastewater (distiller's grains fermentation) overlap, for example, the δ 15 N-NO3 - of the alcohol enterprise wastewater in the research area ranges from -6.9‰ to 51.7‰, which overlaps with the ammonium nitrogen fertilizer and soil nitrogen value range, making it impossible to uniquely point to the pollution source by single isotope ratio. Secondly, the groundwater runoff speed in karst areas is fast (maximum 600 m / d), and the dissolved oxygen is sufficient, which inhibits denitrification, but microbial metabolism and mineral adsorption processes may still cause isotope fractionation, causing source analysis deviation. (2) The groundwater tracing method traces the pollution path by injecting fluorescent dyes, salts, and other tracers, but its application is limited in karst areas. The underground river pipeline length in the research area is 1-3 km, and is controlled by fault zones and conjugate joints. The connectivity of some shafts and sinkholes is unclear, and the tracer injection point is difficult to accurately locate. When multiple sources of pollution such as agricultural non-point source pollution, alcohol enterprise wastewater, and aquaculture wastewater exist simultaneously, the tracer signal may be masked by background pollutants, making it impossible to trace a single source. (3) Multivariate statistical analysis selects pollution main control factors by dimension reduction through factor analysis (FA) and principal component analysis (PCA), but the characteristic pollutants (such as NH4 +, COD) concentration are in the same order of magnitude, PCA can only identify the "composite pollution contribution", and cannot separate the two independent effects; Multivariate statistics rely on data correlation, without considering physical factors such as pollution source spatial distribution, hydrogeological conditions, etc., resulting in uncertainty in the results.

[0004] In summary, the essence of pollution in karst areas is a complex system of "multi-source input-multi-process coupling-multi-medium migration", and the existing methods only target a single dimension (such as isotopes, chemical indicators or spatial distribution), lacking systematic integration. Karst groundwater data have the characteristics of "low detection rate (new pollutants) and high spatial heterogeneity (water slope 7.25‰-1.28%)". Traditional PMF models and isotope mixing models have not been optimized for such data, resulting in parameter estimation bias. Therefore, a new quantitative source identification method is urgently needed to accurately identify the sources of groundwater pollution and improve the accuracy and efficiency of groundwater pollution source identification. SUMMARY

[0005] The present application aims to provide a comprehensive quantitative source identification method for multi-source pollution in karst underground river systems, which solves the defects in multi-source composite pollution and emerging pollutant source identification through the specificity of new pollutant tracing, the systematicness of multi-method coordination and the scene adaptability of model parameters.

[0006] To achieve the above-mentioned purpose, the present application provides the following technical solutions:

[0007] A comprehensive quantitative source identification method for multi-source pollution in karst underground river systems, comprising the following steps:

[0008] S1, multi-dimensional sample collection and new pollutant screening: multi-time period sampling of water bodies, determination of conventional water quality indicators, nitrogen and oxygen isotopes and screening of new pollutants, and targeted quantification of characteristic new pollutants;

[0009] S2, principal component analysis dimension reduction and pollution factor identification: principal component analysis of conventional indicator and new pollutant data, extraction of principal components with cumulative contribution rate exceeding 85%, identification of composite pollution factors of water bodies, and exclusion of secondary interference factors;

[0010] S3, nitrogen and oxygen isotope assisted verification and denitrification exclusion: using the ratio of δ 15 N-NO3 - and δ 18 O-NO3 - , combined with groundwater depth and runoff velocity, to exclude the interference of denitrification and distinguish the source of nitrate as fertilizer, soil nitrogen or waste water;

[0011] S4, PMF model optimization and pollution source quantification: inputting the conventional indicators, characteristic new pollutant targeted quantification results and new pollutant concentration matrix into the PMF model to analyze the contribution rate of each pollution source.

[0012] Furthermore, the water bodies include groundwater, surface water, winery wastewater, and aquaculture wastewater.

[0013] Furthermore, conventional water quality indicators include NO3 - -N, pH, Cl - , nitrogen and oxygen isotopes including δ 15 N-NO3 - , δ 18 O-NO3 - .

[0014] Furthermore, the characteristic new pollutants include ferulic acid and sulfonamide antibiotics.

[0015] Furthermore, the concentration of new pollutants not detected by S4 was set to 5 / 6 of the detection limit, and the model stability was improved by minimizing the robustness objective function.

[0016] Furthermore, the new pollutants include ferulic acid in winery wastewater with a concentration threshold of >1000ng / L and sulfamethoxazole in aquaculture wastewater with a concentration threshold of >500ng / L, as characteristic markers of pollution sources.

[0017] Furthermore, the PMF model input data includes targeted quantitative results of new pollutants.

[0018] Working principle and beneficial effects of the present invention:

[0019] Through high-resolution mass spectrometry non-targeted scanning, it was found that the concentration of ferulic acid, a fermentation product of wine companies, in sewage reached 25932.4ng / L, while that of agricultural non-point sources was only 20.3ng / L; the concentration of sulfamethoxazole in aquaculture wastewater was 6587.2ng / L, which was significantly higher than that of other pollution sources (<200ng / L). These characteristic pollutants constitute a "chemical fingerprint" independent of conventional indicators, which can specifically identify a single source in complex pollution. PCA dimensionality reduction: 3 principal components were extracted (PC1 wine companies + aquaculture, PC2 agricultural non-point sources, PC3 domestic sewage), with a cumulative contribution rate of 87.6%. Among them, the variables with the highest PC1 load were ferulic acid (0.91) and NH4 + (0.87), indicating that the two are a combination of winery wastewater characteristics.

[0020] Isotope verification: S01 spring point δ 15 N-NO3 - =4.2‰, δ 18 O-NO3 - =8.5‰, which is consistent with the fertilizer range (δ 15 N: -4‰~+8‰, δ 18O: +7‰~+25‰) completely matches, and the ferulic acid concentration at this point is <50ng / L, further eliminating the possibility of contamination from the winery. Input a three-dimensional matrix (sample × indicator × time) containing new pollutants, and the model automatically analyzes the spatiotemporal contribution of each pollution source. The PMF model formula is: X(n×m)=G(n×p)F(p×m)+E(n×m),

[0021] For example: upstream point K001: the contribution rate of wastewater from wineries is 96.07% (ferulic acid load 0.93), and that of aquaculture wastewater is only 0.68%; downstream point S004: the contribution rate of aquaculture wastewater is 84.22% (sulfamethoxazole load 0.89), and the contribution of wineries is reduced to 0.68%, which is consistent with the actual sewage outlet distribution.

[0022] This application introduces high-resolution mass spectrometry non-targeted screening technology into karst groundwater source tracing, establishes a feature library containing 5015 pollutants, and breaks through the limitations of traditional methods that rely only on conventional indicators. A closed-loop verification of "PCA dimensionality reduction → isotope verification → PMF quantification" is formed. Conventional technology is only a one-way process of "indicator screening + model calculation" and lacks data cross-validation. For example, the "precipitation dilution factor" excluded by PCA in this application was misjudged as "agricultural non-point source contribution" in the existing technology, resulting in biased results. In view of the "low detection rate and high heterogeneity" characteristics of karst area data, the undetected value is innovatively set to 5 / 6 of the detection limit. This parameter optimization method has not been reported in existing PMF model applications. The existing technology directly sets the undetected value to 0, resulting in a serious underestimation of the model's contribution rate to new pollutants. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 This is a sketch of the hydrogeology of the study area;

[0024] Figure 2 is the box plot of TP, TFe, and Mn concentrations in groundwater and sewage in the study area;

[0025] Figure 3 Piper three-line diagram of water bodies in the study area;

[0026] Figure 4 Groundwater Cl in the study area - With NO3 - / Cl - Molar concentration ratio change graph;

[0027] Figure 5 The results of principal component analysis of pollution causes are shown in Figure 2.

[0028] Figure 6 Groundwater and sewage δ 15 N-NO3 - , δ 18 O-NO3 - concentration box plot;

[0029] Figure 7 The water body δ 15 N-NO3 - and δ 18 O-NO3 - distribution map;

[0030] Figure 8 Contribution ratio of each pollution source to groundwater in the study area (PMF model).

[0031] Figure 1 The reference numerals in the figure include: 1-pure carbonate aquifer, 2-carbonate and clastic interbedded aquifer, 3-aquiclude, 4-underground river outlet and number, 5-karst descending spring, 6-karst vertical shaft, 7-underground river skylight, 8-monitoring well, 9-sampling point, 10-breeding farm, 11-isohydraulic line and groundwater flow direction, 12-groundwater system boundary, 13-inferred fault, 14-syncline axis, 15-geological boundary and stratigraphic code, 16-stratum occurrence, 17-rivers and streams, 18-potential pollution source boundary. DETAILED DESCRIPTION

[0032] The following is further described in detail through specific implementation methods:

[0033] Example:

[0034] 1. Overview of the Study Area

[0035] The study area is located in Zunyi City, Guizhou Province, in the karst mountainous area of ​​southwest China, with an area of ​​approximately 15 km 2 Located on the secondary terrace of the northern Guizhou Plateau, the study area has a highest point at the top of a mountain on the northeast side of the park, at an elevation of 1,081.1 meters, and a lowest point at the valley between the east and west rivers, at an elevation of 875 meters. The relative height difference is 206.1 meters. The landforms are mainly ridge valleys and karst-eroded low- and medium-mountain gullies. The study area has a subtropical monsoon climate, with an average annual precipitation of 1,039.0 to 1,203.9 mm, 60% of which falls between May and October. The main groundwater type in the area is carbonate karst water.

[0036] The study area is divided into the No. 1 and No. 2 underground river systems based on the surface watershed. The water-bearing rock group of the No. 1 underground river system is the second member of the Guanling Formation of the Middle Triassic (T2g 2 ) limestone, with water-bearing media consisting of karst caves and conduits. Groundwater within the system is primarily recharged by surface streams and ditches, fed by vertical shafts and sinkholes, supplemented by surface infiltration of atmospheric precipitation. Controlled by northeast-trending longitudinal fractures, groundwater generally flows from northeast to southwest. Blocked by northwest-trending faults, groundwater is discharged to the surface via underground river outlets (S02), with a discharge of 23.5 L·s in the high-season. -1 , dry season flow rate 2.34L·s-1 , hydraulic gradient 7.25‰, average groundwater flow rate 100-200m·d -1 , maximum flow rate 600m·d -1 , groundwater depth <20m( Figure 1 The water-bearing rock group of the No.2 underground river system is the Jialingjiang Formation of the Middle and Lower Triassic (T 1-2 j) Limestone and karst breccia. The water-bearing media are mainly caves and pipes. The groundwater in the system is mainly replenished by surface water injected into the sinkholes, supplemented by surface infiltration of atmospheric precipitation. Controlled by conjugate joints with an angle of 45° and 150°, the groundwater generally flows from west to east and is concentratedly discharged to the surface due to the cutting of gullies (S003 and S004). The flow rate in the high season is 120.8 L·s -1 , dry season flow rate 10.56L·s -1 , hydraulic gradient 1.28%, average groundwater velocity 139.63 m·d -1 , the maximum flow rate is 408.28m·d -1 , groundwater depth <15m( Figure 1 ).

[0037] The potential pollution sources of the No. 1 underground river system in the study area include a winery sewage treatment plant, wine lees storage yard, No. 1 fertilizer production workshop, No. 2 fertilizer production workshop, agricultural non-point sources, and domestic sewage; the potential pollution sources of the No. 2 underground river system include a winery brewing workshop, a farm, agricultural non-point sources and domestic sewage ( Figure 1 ).

[0038] 2. Research Methods

[0039] To comprehensively analyze the water quality characteristics within the system and trace them back to their source, sampling was conducted in May, September, and December 2023 at representative natural water points, boreholes, potential pollution sources, surface water, and atmospheric precipitation in the No. 1 and No. 2 underground river systems in the study area. The samples collected included 31 full-analysis samples (dry and wet seasons), 120 samples for characteristic index analysis, 24 samples for nitrogen and oxygen isotopes, and 12 samples for analysis of emerging pollutants. All samples were rinsed two to three times with purified water before collection and filled in 2.5L polyethylene bottles.

[0040] With the exception of new pollutants, which were filtered through a 0.45μm filter membrane, the filtrate was collected and stored in a -20°C refrigerator, and the sampling bottles were wrapped in tinfoil. All other samples were stored at room temperature and subsequently sent to the laboratory for testing. The storage, transportation, and testing of the samples strictly adhered to the current "Technical Specifications for Groundwater Environmental Monitoring" (HJ164-2020) to ensure the accuracy and reliability of the results.

[0041] Before indoor testing, pure nitric acid was added to the sample to acidify the sample to pH < 2, and the Ca 2+ Mg 2+ , K + 、Na + Isoconcentration; HCO3 - 、CO3 2- The concentration was determined by titration; the Cl - 、SO4 2- Isoconcentration [2] The nitrogen and oxygen isotope detection method uses the bacterial denitrification method and uses the MAT 253 stable isotope ratio mass spectrometer (Thermo Fisher) to measure the nitrogen and oxygen isotopes of N2O to determine the NO3 in the sample. - The new pollutant samples were enriched and purified by solid-phase extraction (SPE), and 5,015 new pollutants were screened non-targeted using a SCIEX X500R Q-TOF quadrupole time-of-flight high-resolution mass spectrometer. Subsequently, a SCIEX QTRAP 5500+ triple quadrupole mass spectrometer was used to further identify the 111 relatively abundant pollutant types identified in the original non-targeted screening and to determine the concentrations of 20 target new pollutants (including antibiotics, pesticides, and organic trace pollutants).

[0042] PCA is a fundamental method in multivariate statistical analysis, capable of reducing the dimensionality of multidimensional data and identifying the primary factors influencing the water quality of a study area. It objectively groups characteristic pollutants across multiple sample points and a series of other variables with complex relationships. The initial factor loading matrix is ​​then subjected to orthogonal rotation to reduce the dimensionality. Principal factors are then identified based on the factor loadings of each characteristic pollutant, yielding scores for each principal component.

[0043] The PMF model, first proposed by Paatero et al. in 1994, uses weighted least squares fitting to determine the pollution characteristic factor. The core of the PMF algorithm is to minimize the weighted residual through the equation. In recent years, studies have used the PMF model to identify groundwater pollution sources, achieving good results. In comparison with other multivariate statistical models, it is found that the PMF model results often have a better fit and less uncertainty. The mathematical principle of PMF is:

[0044] X(n×m)=G(n×p)F(p×m)+E(n×m)

[0045] Where X is the sample concentration matrix, G is the factor contribution matrix, F is the factor concentration matrix, E is the residual matrix, n is the number of samples, m is the number of chemical substances, and p is the number of factors.

[0046] The PMF model was used to conduct a pollution source analysis study on the groundwater in the study area. Due to the limited detection rate of the target new pollutants, there are a certain amount of new pollutants in the undetected state. To ensure the normal operation of the model, the concentration of the new pollutants in the undetected state is set as the detection limit concentration (50ng·L -1 ) to ensure the accuracy of the model calculation results.

[0047] The principal component analysis of the water quality parameters of the characteristic indicators of the study area was completed using Excel and IBM SPSS22.0, the EPAPMF 5.0 software was used to complete the source analysis of new groundwater pollutants, and data charts were constructed using Origin 2022.

[0048] 3. Research Results

[0049] The pH value of groundwater outcrops at the springs and wells in the study area is greater than 7, which is neutral to weakly alkaline water; the pH value of winery wastewater is between 2.87 and 4.89, which is strongly acidic water; the pH value of domestic sewage is between 6.3 and 7.2, which is weakly acidic to neutral water; the pH value of aquaculture wastewater is between 7.33 and 7.71, which is neutral water.

[0050] As shown in Table 1 and Figure 2 As shown in the figure, the groundwater K in the study area + Range: 0.7-142.8 mg·L -1 The average value is 12.52 mg·L -1 , Na + Range: 3.2-66.3 mg·L -1 The average value is 10.91 mg·L -1 , SO4 2- The range is 41.33 to 389.9 mg·L -1 The average value is 137.2 mg·L -1 , NH4 + Range 0.02~247mg·L -1 The average value is 13.3 mg·L -1 , COD range 4~1340mg·L -1 The average value is 77.9 mg·L -1 , NO3 - -N range: 0.04~148.3mg·L -1 The average value is 27.73 mg·L -1 TP range: 0.01~0.91mg·L -1 , median 0.17 mg·L-1 , TFe range 0.05~0.67mg·L -1 , median 0.17 mg·L -1 , Mn range 0.009~1.56mg·L -1 , median 0.26 mg·L -1 Among them, NH4 + 、NO3 - -N, TP, TFe, Mn + , COD, K + 、Na + 、Cl - 、SO4 2- The concentration (average or maximum) exceeds the background value of groundwater by 12 to 9540 times, and NH4 + 、NO3 - -N, TP, TFe, Mn + The concentrations of COD (average or maximum) exceeded the Class III standard of the "Groundwater Quality Standard" (GB / T 14848-2017) by 3 to 954 times, indicating that the groundwater environmental quality in the study area was significantly affected by external pollution sources.

[0051] Table 1 Statistics of physical and chemical parameters of different water bodies in the study area

[0052]

[0053]

[0054] like Figure 3 As shown in the anion triangle diagram in the lower right corner, most groundwater outcrop sampling points are in the lower left corner, and HCO3 - The percentage of milligram equivalent is greater than 50%, which is consistent with the natural background of groundwater in the karst area. However, K59, K29, and K61 are located in the upper triangle area, and SO4 2- The milliequivalent percentage is greater than 50%, which is due to the SO4 2- The content is high (as shown in Table 1), with the milliequivalent percentage greater than 70%. K59, K29, and K61 are hydraulically closely connected to surface water and receive recharge from upstream surface water, which causes the water chemistry of K59, K29, and K61 to change from HCO3-Ca·Mg to SO4·HCO3-Ca·Mg. In the cation triangle diagram in the lower left corner, most groundwater outcrops are in the lower left corner area, and Ca 2+ Between 55% and 90%, Mg 2+The Na+K mg equivalent percentage of the CK18 monitoring well is 58% in the lower right corner, while the Na+K mg equivalent percentage of the YY06 lees yard wastewater is 70%, indicating that the Na in the CK18 monitoring well is 10% to 40%, which is consistent with the natural background of groundwater in the karst area. + , K + The main source is leakage from the adjacent distiller's grains yard, which causes the water chemistry type of the CK18 monitoring well to change from HCO3-Ca·Mg to SO4·HCO3-K·Na·Ca.

[0055] Cl - It is very stable in natural water bodies and can be used to identify sewage leaks, fertilizers and feces, and natural pollution sources to a certain extent. - Sources with different Cl - With NO3 - / Cl - When there is high Cl in water - and low NO3 - / Cl - When the nitrate is affected by domestic sewage / livestock manure, the low Cl - and high NO3 - / Cl - This indicates that nitrate is mainly affected by agricultural activities.

[0056] Figure 4 NO3 in the groundwater of the study area - / Cl - With Cl - Molar ratio relationship, NO3 - / Cl - The values ​​range from 0.004 to 15.31, with an average value of 3.23. - Concentration ranges from 0.21 to 3.40 mmol·L -1 , with an average value of 0.65 mmol·L -1 , NO3 in monitoring well CK18 in area 1 - / Cl - The value is 0.11, Cl - The concentration is 3.40mmol·L -1 , Cl in the wastewater from the upstream winery in the winery yard - Concentration ranges from 0.8 to 9.3 mmol·L -1 , NO3 - Concentration ranges from 0.01 to 4.23 mmol·L -1 The CK18 monitoring well is adjacent to the wine lees dump, indicating that the nitrate and chloride ions in the groundwater of the CK18 monitoring well in Area 1 are mainly derived from the upstream wine lees dump. The natural outcrop of groundwater in Area 2 is located around and downstream of the 1# and 2# agricultural planting areas.- / Cl - The values ​​are in the range of 5.23 to 15.31, Cl - The concentration is between 0.58 and 0.81 mmol·L -1 , indicating that the nitrate in the groundwater of Area 2 mainly comes from the agricultural planting area. The groundwater outcrops in Area 3 are located in the fertilizer production workshop, wine production workshop, agricultural planting area, densely populated area and the downstream area of ​​H507 farm. - / Cl - The value is in the range of 0.004~0.95, Cl - The concentration is between 0.21 and 0.78 mmol·L -1 , indicating that nitrate in the groundwater of the three areas is the result of the combined effects of agricultural pollution, winery wastewater, aquaculture wastewater and domestic sewage.

[0057] Based on the consistency of data and the key points of groundwater recharge and drainage in the study area, six points (K001, K002, K501, K505, S003, and S004) from the upstream to the downstream of the No. 2 underground river system were selected, and the characteristic indicators COD, NH4 + 、NO3 - -N, TP, TFe, Mn + , pH water quality data to conduct principal component analysis (PCA) of regional pollution causes. Figure 5 As shown, the loadings for principal component 1 ranged from 48.14% to 77.66%, principal component 2 from 12.51% to 34.40%, and principal component 3 from 0% to 17.41%. This indicates that from November 17 to 21, 2023, groundwater quality was jointly influenced by principal components 1 and 2. The loading for principal component 1 gradually decreased over time. On November 22, the loading for principal component 3 gradually emerged, fluctuating between 15.81% and 18.44%, approaching a stable state. This suggests that this is a persistent and stable source of pollution. The upstream locations of the No. 2 underground river system include the winery production workshop, livestock farm, and agricultural planting area 2. Analysis suggests that principal component 3 represents an agricultural non-point source with long-term impact from agricultural planting area 2. Because the livestock farm and winery production workshop share the same pollution characteristic indicators and eigenvalues ​​of the same order of magnitude, principal component analysis cannot distinguish between the two, but only indicates the possibility that both contribute to groundwater pollution.

[0058] Table 2 Statistics of principal component load analysis results of pollution causes

[0059]

[0060] When there may be multiple sources of pollution in groundwater, simply using δ 15 N-NO3 -It cannot be used as an accurate basis for the source of nitrate. Nitrate oxygen isotopes (δ 18 O-NO3 - ) can provide more evidence of the source of pollution. Figure 6 As shown, (a) groundwater δ 15 N-NO3 - The range value is 3.52‰~12.2‰, the median is 8.57‰, and δ 18 O-NO3 - The range of values ​​is 0.8‰ to 25.7‰, with a median of 9.34‰. (b) Sewage δ 15 N-NO3 - The range value is -6.9‰~51.7‰, the median is 8.58‰, δ 18 O-NO3 - The range value is -5‰~30.1‰, with a median of 21.3‰. The nitrogen and oxygen isotope contents of nitrate in different water bodies are different. - No obvious denitrification occurred, and NO3 - An important prerequisite for pollution source analysis. Based on hydrogeological analysis, the groundwater depth in the area is mostly less than 20 meters. Karst is well developed, with sinkholes, vertical shafts, and karst skylights distributed in a "beaded" pattern, providing a good geological foundation for dissolved oxygen in groundwater. Furthermore, the karst underground river pipeline is 1 to 3 kilometers long, and the average groundwater runoff velocity is 100 to 300 meters per day. -1 , the groundwater flow rate is fast, and no favorable conditions for denitrification are created. In addition, Figure 8 It can be seen that the nitrogen and oxygen isotope values ​​of most samples in the study area are not between 1:1.3 and 1:2.1 for denitrification. Therefore, the nitrogen and oxygen isotope method can be used to analyze the NO3 - Identify the source of pollution.

[0061] like Figure 7 As shown, domestic sewage δ 15 N-NO3-range value -6.5‰~5.4‰, δ 18 The O-NO3- value range is -5‰~16.6‰. SH01 falls in the area affected by ammonium nitrogen fertilizer, soil nitrogen, sewage and wastewater, and feces. This is related to SH01's location in a densely populated village affected by indiscriminate discharge of domestic sewage, septic tank leakage, and soil cultivation. SH02 falls in the ammonium nitrogen fertilizer range, which is related to SH02's location near the 2# fertilizer production workshop. 15 N-NO3-range value -6.9‰~51.7‰, δ 18 The O-NO3- value range is 3.7‰~30.1‰, which basically falls within the range of ammonium nitrogen fertilizer and nitrate nitrogen fertilizer, indicating that the winery wastewater is related to the use of wine lees to make organic fertilizer in the study area.

[0062] Groundwater outcrop δ 15 N-NO3- range value is 3.5‰~12.2‰, δ 18 O-NO₃ values ​​range from 0.8‰ to 25.7‰, with the S01 spring falling within the fertilizer-NO₃ range, indicating that the primary source of nitrate at the S01 spring is related to leakage from an upstream fertilizer production plant. Other groundwater outcrops are located in areas where soil nitrogen, sewage, and feces combine, indicating that nitrate in these groundwaters is primarily influenced by a combination of soil nitrogen, sewage, and feces. This conclusion is reinforced by the presence of pollution sources upstream of these springs, such as fertilizer production plants, agricultural planting areas, and sewage treatment plants.

[0063] The aforementioned methods provide a certain degree of insight into the composition of pollution sources in the study area, but they cannot quantitatively analyze the contribution of these sources. Therefore, through non-targeted screening and targeted quantitative testing of new pollutants in the study area's groundwater, combined with the PMF model, we determined the contribution of pollution sources at each sampling point within the region, enabling the analysis of pollution sources in the regional groundwater system.

[0064] Given that the No. 2 underground river system is home to farms, winery production workshops, and agricultural planting areas, and that groundwater is contaminated, to identify the source and contribution of the pollution, samples from upstream to downstream were collected from K001-1, K001-2, K001-3, K501, K505, S003, S004-1, S004-2, S004-3, ZK4, aquaculture wastewater (H507), and winery cellar bottom water (JD10) for non-targeted and targeted quantitative screening of new pollutants. The results showed (Table 3) that the content of sulfamethoxazole in the H507 aquaculture wastewater was 6587.2 ng·L -1 , Chlortetracycline hydrochloride content 9984.2ng·L -1 , Flunixin content 158.9 ng·L -1 , sulfachloropyridazine content 5355.6ng·L -1 The content of ferulic acid in JD10 winery's cellar bottom water is 25932.4 ng·L -1 , salicylic acid content 1143.8ng·L -1 4-Hydroxy-3-methylbenzoic acid content: 519.8 ng·L -1 , p-Hydroxybenzaldehyde content 1970.7ng·L -1 The composition and concentration of characteristic new pollutants from the two potential pollution sources are quite different.

[0065] Table 3 Quantitative targeted screening analysis results of new pollutants in groundwater samples in the study area (ng·L -1 )

[0066]

[0067] Based on the above targeted screening results, the PMF model was used to identify the pollution sources. Figure 8 As shown in the figure, the pollution contribution of JD10 raw water is almost entirely from the winery wastewater (contribution rate 99.90%), and a very small amount of N, N-diethyl-3-methylbenzamide is detected in H507 raw water (0.6 ng·L -1 ), indicating that farm wastewater pollution accounts for an absolute majority (98.05%) of the raw water from farms, with a very small portion coming from surrounding agricultural non-point source pollution (1.95%). This is related to the sampling point being located in farmland outside the farm. K001 groundwater pollution sources are caused by liquor company wastewater, with a contribution rate of 85.06% to 96.07%, and agricultural non-point source pollution, with a contribution rate of 3.04% to 13.74%. The main source of groundwater pollution in the ZK4 monitoring well is agricultural non-point source pollution, with a contribution rate of 97.92%. Combined with hydrogeological analysis, although the groundwater in the ZK4 monitoring well originates from 30 meters below the ground, it has no obvious hydraulic connection with other sampling points in the pollution corridor, indicating that the groundwater quality in the area is contaminated by agricultural non-point source pollution, resulting in the detection of new pollutants mainly consisting of pesticides. K501 groundwater pollution sources are caused by farm wastewater, with a contribution rate of 61.80%, and agricultural non-point source pollution, with a contribution rate of 37.46%. Groundwater pollution in K505 is caused by livestock farm wastewater, which contributes 59.52%, agricultural non-point source pollution, which contributes 39.80%, and wastewater from distilleries, which contributes 0.68%. Groundwater pollution in S003 is caused by livestock farm wastewater, which contributes 51.24%, agricultural non-point source pollution, which contributes 48.13%, and wastewater from distilleries, which contributes 0.63%. Groundwater pollution in S004 is caused by livestock farm wastewater, which contributes 71.64% to 84.22%, agricultural non-point source pollution, which contributes 15.56% to 27.81%, and wastewater from distilleries, which contributes 0.22% to 0.55%.

[0068] In summary, (1) through the detection of the main ions in the study area, it was found that NH4 + 、NO3 - -N, TP, TFe, Mn + , COD, K + 、Na + 、Cl - 、SO4 2- The concentration (average or maximum) exceeded the background value by 12 to 9540 times, and NH4 + 、NO3 - -N, TP, TFe, Mn +The concentration of COD (average or maximum) exceeds the Class III standard of the "Groundwater Quality Standard" (GB / T 14848-2017) by 3 to 954 times. The corresponding characteristic indicators in winery wastewater, domestic sewage, and aquaculture wastewater are higher than those in groundwater. In addition, the chemical type of some groundwater has changed from HCO3-Ca·Mg type to SO4·HCO3-K·Na·Ca type, indicating that the groundwater in the area has been polluted.

[0069] (2) The main ion traceability, principal component analysis of characteristic pollutants, and nitrogen and oxygen isotope analysis revealed that groundwater pollution in the study area comes from the superposition of multiple pollution sources, including agricultural non-point source pollution, domestic sewage, aquaculture wastewater, and winery wastewater. Specifically, the CK18 monitoring well is polluted by the wine lees yard, while the pollution sources in the other groundwater monitoring wells, some springs, and underground river outlets are the result of the combined effects of winery wastewater, agricultural non-point source pollution, and domestic sewage. The groundwater pollution in K001, K002, K501, K505, S003, and S004 comes from the superposition of aquaculture wastewater, winery wastewater, and agricultural non-point source pollution.

[0070] (3) The PMF model based on the screening of new pollutants effectively reveals the contribution of pollution in the study area. Among them, the contribution rate of winery wastewater pollution sources in K001 is 85.06% to 96.07%, and the contribution rate of agricultural non-point sources is 3.04% to 13.74%. As the groundwater gradually migrates, the contribution rate of winery wastewater pollution sources gradually decreases. Downstream from K501 to S004, the main contributions to groundwater pollution come from farm wastewater (51.24% to 84.22%), agricultural non-point sources (15.56% to 48.13%), and winery wastewater (0.22% to 0.68%). By analyzing the correlation between the types and concentrations of new pollutants and the surrounding potential pollution sources, and using appropriate model tools, it is possible to effectively calculate the contribution rate of regional groundwater pollution sources, providing new ideas for regional groundwater pollution control.

[0071] The above is only an embodiment of the present invention, and the common knowledge such as the specific structure and characteristics of the scheme is not described in detail here. It should be pointed out that for those skilled in the art, without departing from the structure of the present invention, several variations and improvements can be made, which should also be regarded as the scope of protection of the present invention, and these will not affect the effect of the implementation of the present invention and the practicality of the patent. The scope of protection required by this application shall be based on the content of its claims, and the specific implementation methods and other records in the specification can be used to interpret the content of the claims.

Claims

1. A comprehensive quantitative tracing method for multi-source pollution in karst underground river systems, characterized by: The steps include: S1. Multi-dimensional sample collection and new pollutant screening: Sampling water bodies at multiple time periods to measure conventional water quality indicators, nitrogen and oxygen isotopes, screen for new pollutants, and conduct targeted quantification of characteristic new pollutants; S2. Principal component analysis dimensionality reduction and pollution factor identification: Conduct principal component analysis on conventional indicators and new pollutant data, extract principal components with cumulative contribution rates exceeding 85%, identify complex pollution factors in water bodies, and eliminate minor interfering factors; S3, Nitrogen and oxygen isotope-assisted verification and denitrification elimination: using δ 15 N-NO3 - and δ 18 O-NO3 - The ratio, combined with the groundwater depth and runoff velocity, eliminates the interference of denitrification and distinguishes the source of nitrate from fertilizer, soil nitrogen or sewage; S4. PMF model optimization and pollution source quantification: Input conventional indicators, targeted quantitative results of characteristic new pollutants, and new pollutant concentration matrix into the PMF model to analyze the contribution rate of each pollution source.

2. The comprehensive quantitative tracing method for multi-source pollution in karst underground river systems according to claim 1 is characterized in that: The water bodies include groundwater, surface water, winery wastewater, and aquaculture wastewater.

3. The comprehensive quantitative tracing method for multi-source pollution in karst underground river systems according to claim 2 is characterized in that: Conventional water quality indicators include NO3 - -N, pH, Cl - , nitrogen and oxygen isotopes including δ 15 N-NO3 - , δ 18 O-NO3 - .

4. The comprehensive quantitative tracing method for multi-source pollution in karst underground river systems according to claim 3 is characterized in that: The characteristic new pollutants include ferulic acid and sulfonamide antibiotics.

5. The comprehensive quantitative tracing method for multi-source pollution in karst underground river systems according to claim 4 is characterized in that: The concentration of undetected new pollutants in S4 was set to 5 / 6 of the detection limit, and the model stability was improved by minimizing the robustness objective function.

6. The comprehensive quantitative tracing method for multi-source pollution in karst underground river systems according to any one of claims 3 to 5, characterized in that: The new pollutants include ferulic acid in winery wastewater, with a concentration threshold of >1000ng / L, and sulfamethoxazole in aquaculture wastewater, with a concentration threshold of >500ng / L, which serve as characteristic markers of pollution sources.

Images