A method for identifying mixed endmembers in groundwater considering the influence of seasonal variations in river water.

By screening river and groundwater sample data and combining principal component analysis and least squares optimization methods, the problem of accurately assessing the impact of river water infiltration on groundwater was solved, and the accurate identification and proportion estimation of mixed end-members were achieved.

CN119901891BActive Publication Date: 2026-11-10INST OF GEOGRAPHICAL SCI & NATURAL RESOURCE RES CAS +2
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202510103469.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2025-01-22
Publication Date
2026-11-10
Estimated Expiration
2045-01-22

AI Technical Summary

Technical Problem

Existing technologies, when analyzing groundwater systems connected to river water or surface water, do not fully consider the spatiotemporal variability of the physicochemical parameters of each end-member, resulting in inaccurate calculation of mixing ratios and an inability to accurately assess the impact of infiltration of different end-member water bodies on groundwater.

Method used

A groundwater mixing end-member identification method considering the influence of seasonal variations in river water was adopted. By screening river and groundwater sample data, principal components were analyzed to screen mixing end-members. Principal component analysis (PCA) and least squares optimization methods were used to quantitatively estimate the mixing ratio and the physicochemical parameters of the identified end-members.

Benefits of technology

It improves the accuracy and effectiveness of groundwater mixing end-member identification, enabling intuitive and quantitative assessment of the impact of river water infiltration on groundwater and achieving accurate estimation of mixing ratio.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119901891B_ABST
    Figure CN119901891B_ABST
Patent Text Reader

Abstract

This application discloses a method for identifying groundwater mixing end-members considering the influence of seasonal variations in river water, belonging to the field of water environment engineering technology. The embodiments of this application adopt the following technical solution: a method for identifying groundwater mixing end-members considering the influence of seasonal variations in river water, including the following steps: screening data from river and groundwater samples; analyzing the principal components of river water samples; screening river water mixing end-members; analyzing the principal components of groundwater samples; screening groundwater mixing end-members. This application can effectively identify river water mixing end-members and further screen representative physicochemical parameters, achieving quantitative estimation of the mixing ratio and physicochemical parameter data of the mixing end-members. Simultaneously, it can effectively identify groundwater end-members in the mixing process of different water bodies, intuitively and quantitatively assessing the impact of river water infiltration on groundwater, and improving the accuracy and effectiveness of groundwater mixing end-member identification.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This application belongs to the field of water environment engineering technology, specifically, it relates to a method for identifying groundwater mixed end-members that takes into account the influence of seasonal changes in river water. Background Technology

[0002] In water-scarce regions, reclaimed water or inter-regional water transfer is regarded as a vital available resource. It is not only widely used as a source of water for agricultural irrigation, urban industry and greening, but also plays an important role in restoring depleted river channels and optimizing the river ecological environment. With the large-scale application of cross-regional transfer of surface water and reclaimed water in rivers, its impact on the groundwater environment has increasingly become a focus of public attention.

[0003] The use of surface water or reclaimed water can bring ecological benefits to urban rivers in arid environments, but the potential negative impacts on aquatic systems and human health cannot be ignored. Currently, when analyzing the mixing effect in groundwater systems connected to river water or surface water, the spatiotemporal variability of the physicochemical parameters of each end-member has not been fully considered, which brings uncertainty to the accuracy of calculating the mixing ratio. This may not be sufficient to accurately assess the impact of different end-member water bodies infiltration on groundwater. Summary of the Invention

[0004] To address the aforementioned problems and technical deficiencies, this application adopts the following technical solution: a method for identifying groundwater mixing end-members considering the influence of seasonal variations in river water, comprising the following steps:

[0005] Step S1: Screening data from river and groundwater samples;

[0006] Step S2: Analyze the principal components of the river water sample;

[0007] Step S3: Screening river water mixing end-members;

[0008] Step S4: Analyze the main components of the groundwater sample;

[0009] Step S5: Screening groundwater mixed end-members.

[0010] Furthermore, the screening data parameters for the water sample in step S1 include: field index, cations, anions, nitrogen, total organic carbon, permanganate index, and total dissolved solids;

[0011] The screening method uses a direct elimination method.

[0012] Preferably, the specific steps for analyzing the principal components of the river water sample in step S2 are as follows:

[0013] Step S201: Set up a dataset X containing m samples and n chemical and physical parameters;

[0014] Step S202: Calculate the mean vector of the samples, center the dataset to obtain the centered dataset, and then calculate the covariance matrix of the data.

[0015] Step S203: Perform eigenvalue decomposition on the covariance matrix to obtain eigenvalues ​​and corresponding eigenvectors;

[0016] Step S204: Sort the feature values ​​according to their size;

[0017] Step S205: Obtain the principal component score based on the eigenvectors corresponding to the sorted eigenvalues.

[0018] Furthermore, the formula for calculating the covariance matrix in step S202 is as follows:

[0019]

[0020] X centered It is data that has been centrally processed.

[0021] Furthermore, eigenvalue decomposition is performed on the covariance matrix to obtain eigenvalues ​​λ1, λ2, λ3...λ n and the corresponding feature vectors ν1, v2, v3...v n .

[0022] Furthermore, the first eigenvector can be represented as v1 = [v 11 ,v 21 ,v 31 ...,λ n1 ] T The eigenvectors correspond to the principal component directions in the dataset, and the eigenvalues ​​represent the variance of the data explained by each principal component.

[0023] Furthermore, the contribution rate of each principal component PC is:

[0024] λ1 / sum(λ1, λ2, λ3...λ) n )

[0025] Furthermore, sorting by eigenvalues ​​from largest to smallest, we select the eigenvectors corresponding to the top k eigenvalues; each principal component summarizes the original variable as follows:

[0026] PC K =v 1K X1+v 2K X2+ν 3K X3+...+v nK X m

[0027] Furthermore, principal components can explain the maximum variance in the dataset. Selecting the first few principal components effectively preserves the main information of the data. The original water sample data is projected onto the principal component coordinate system to obtain the principal component scores of different samples.

[0028] Preferably, the specific steps for screening river water mixing end-members in step S3 are as follows:

[0029] Step S301: Perform mixed analysis on the river water samples;

[0030] Step S302: Perform terminal member analysis using conserved species and calculate the mixing ratio for each sample.

[0031] Compared with the prior art, the beneficial effects of the embodiments of this application are as follows:

[0032] (1) This application can effectively identify the mixing end-members of river water on the one hand, and further screen out representative physicochemical parameters on the other hand, realize the quantitative estimation of the mixing ratio, and the physicochemical parameter data of the mixing end-members; at the same time, it can effectively identify the groundwater end-members in the mixing process of different water bodies, intuitively and quantitatively assess the impact of river water infiltration on groundwater, and improve the accuracy and effectiveness of groundwater mixing end-member identification. Attached Figure Description

[0033] In the attached diagram:

[0034] Figure 1 This is a flowchart illustrating an embodiment of this application;

[0035] Figure 2 This is a schematic diagram showing the PCA analysis results of river water data and the contribution of each variable to the eigenvector in the embodiments of this application;

[0036] Figure 3 This is a scatter plot of PC obtained from river water samples in the embodiments of this application;

[0037] Figure 4 This is a schematic diagram showing the estimated and measured concentrations in river water samples in the embodiments of this application;

[0038] Figure 5 This is a schematic diagram showing the PCA analysis results of groundwater data and the contribution of each variable to the eigenvector in the embodiments of this application;

[0039] Figure 6 This is a schematic diagram of the PCA results of groundwater samples in the embodiments of this application;

[0040] Figure 7 This is a schematic diagram showing the measured and estimated concentrations of groundwater samples in an embodiment of this application. Detailed Implementation

[0041] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some embodiments of this application, but not all embodiments. The components of the embodiments of this application described and shown in the accompanying drawings can be arranged and designed in various different configurations.

[0042] Example:

[0043] like Figure 1 As shown, a method for identifying mixed end-members in groundwater that takes into account the influence of seasonal variations in river water includes the following steps:

[0044] Step S1: Screening data from river and groundwater samples;

[0045] Step S2: Analyze the principal components of the river water sample;

[0046] Step S3: Perform river water end-member mixing analysis to screen typical river water mixing end-members that can represent the influence of water samples with different seasonal characteristics.

[0047] Step S4: Analyze the main components of the groundwater sample;

[0048] Step S5: Screening groundwater mixed end-members.

[0049] The screening data parameters for water samples in step S1 include: field index, cations, anions, nitrogen, total organic carbon, permanganate index, and total dissolved solids;

[0050] The screening method used was direct elimination.

[0051] It should be noted that the field indexes are pH and EC values, and the cations include K+, Na+, Ca2+ and Mg2+; the anions include Cl-, SO42- and HCO3-; and the nitrogen includes NH4-N, NO3-N and NO2-N.

[0052] It should also be noted that during the elimination process, the physicochemical parameters of river water and groundwater samples were screened and evaluated. Samples that clearly exceeded the reasonable range and could not be reasonably explained were directly eliminated.

[0053] The specific steps for analyzing the principal components of the river water sample in step S2 are as follows:

[0054] Step S201: Set a dataset X containing m samples and n chemical and physical parameters, where X = [X1, X2, ..., Xm], and each sample Xi is an n-dimensional vector;

[0055] Step S202: Calculate the mean vector of the samples, center the dataset to obtain the centered dataset, and then calculate the covariance matrix of the data.

[0056] Step S203: Perform eigenvalue decomposition on the covariance matrix to obtain eigenvalues ​​and corresponding eigenvectors;

[0057] Step S204: Sort the feature values ​​according to their size;

[0058] Step S205: Obtain the principal component score based on the eigenvectors corresponding to the sorted eigenvalues.

[0059] The formula for calculating the covariance matrix in step S202 is as follows:

[0060]

[0061] X centered It is data that has been centrally processed.

[0062] Eigenvalue decomposition of the covariance matrix yields eigenvalues ​​λ1, λ2, λ3...λ n and the corresponding feature vectors v1, v2, v3...v n .

[0063] The first eigenvector can be represented as v1 = [v 11 ,v 21 ,v 31 ...,λ n1 ] T The eigenvectors correspond to the principal component directions in the dataset, and the eigenvalues ​​represent the variance of the data explained by each principal component.

[0064] The contribution rate of each principal component PC is:

[0065] λ1 / sum(λ1, λ2, λ3...λ) n )

[0066] Sort the eigenvalues ​​by magnitude and select the eigenvectors corresponding to the top k eigenvalues ​​from largest to smallest; each principal component summarizes the original variable as follows:

[0067] PC K =v 1K X1+v 2K X2+v 3K X3+...+ν nK X m

[0068] Principal components can explain the maximum variance in a dataset. Selecting the first few principal components effectively preserves the main information of the data. Projecting the original water sample data onto the principal component coordinate system yields the principal component scores for different samples. For example, the score of the k-th principal component for the i-th sample is Fik.

[0069] F ik =v 1K X 1i +v 2K X 2i +v 3K X 3i +...+v nK X ni

[0070] Quantitative analysis of the mixing effects of different endmembers and screening of mixed endmembers is an optimization process for estimating endmember components and mixing ratios. For example, existing sample data X is a dataset containing m samples and n physicochemical parameters, where X = [X1, X2, ..., Xm], and each sample Xi is an n-dimensional vector.

[0071] When it is necessary to quantitatively analyze the mixing ratio of m samples from w end-member water bodies, it is necessary to estimate the concentration or value of n physicochemical parameters in each end-member water body. When all the physicochemical parameters used are conservative, the multi-end-member mixing process can be simply summarized as follows:

[0072] X = YΔ + ε

[0073] Where X represents an n×m dimensional matrix of all physicochemical parameters of all samples:

[0074]

[0075] Y represents an n×w dimensional matrix with w mixed endmembers and n physicochemical parameters:

[0076]

[0077] Δ represents the mixing ratio of different endmembers and is a w×m dimensional matrix:

[0078]

[0079] ε represents the measurement error and is an n×m dimensional matrix.

[0080] In addition, the mixing ratio needs to satisfy [0,1], and for any column, its sum is 1.

[0081] Assuming the concentration of endmember components is known, the least squares method can be used to calculate the mixing ratio matrix. Under the assumption of a certain measurement error (e.g., 5%), the problem of estimating the mixing ratio and the mixing endmember components becomes the problem of solving for the minimum value, min{X-(YΔ+ε)}. Considering the [0,1] condition that the mixing ratio needs to satisfy, and the condition that the sum of any column of Δ is 1, the minimum value problem is iteratively calculated until convergence.

[0082] The specific steps for screening river water mixing end-members in step S3 are as follows:

[0083] Step S301: Perform mixed analysis on the river water samples;

[0084] Step S302: Perform terminal member analysis using conserved species and calculate the mixing ratio for each sample.

[0085] In areas affected by river water infiltration, groundwater in aquifer systems can be considered a mixture of river water and initial groundwater. To differentiate these effects, it is necessary to consider the PCA scores in groundwater samples and the loadings of multiple variables. By quantitatively comparing the loadings of physicochemical parameters in different principal components (PCs), the influence of different processes (such as mineral dissolution and river water infiltration) on the chemical properties of groundwater can be analyzed.

[0086] The selection of principal components and the calculation of principal component scores for samples are described in step S2.

[0087] Based on the multiple river water end-members identified as the final components in step S3, and excluding groundwater samples that show anthropogenic influence, the principal component analysis results of the groundwater are used to create a two-dimensional scatter plot. The principal component scores of the river water end-members are then superimposed, and groundwater samples are selected as new initial groundwater mixing end-members based on the distribution of the two-dimensional scatter plot.

[0088] When quantitatively analyzing the mixing effects of groundwater samples, it is necessary to select multiple physicochemical parameters as conserved species and set the standard deviation of groundwater mixing endmember EM. In this invention, it is assumed that the species concentration in river water in the river water mixing endmember EM remains constant in the multi-endmember mixing model.

[0089] Based on the multi-end-member hybrid analysis process in step S3, the estimated values ​​of different physicochemical parameters of the EM groundwater end-members are obtained by performing least squares analysis on different physicochemical parameters. Based on the multi-end-member hybrid calculation results, the degree of influence of river water infiltration on groundwater samples at different depths can be analyzed. Comparing the results with known field geological conditions can evaluate the EM of river water in different seasons. 河道水 and groundwater EM 地下水 The effectiveness of the hybrid mode.

[0090] In summary, the above methods were applied in the following experiments, and experimental data were recorded during the application process, as detailed below:

[0091]

[0092] Table 1

[0093] Table 1 lists the in-situ parameters such as pH and the descriptive statistics of major ions in groundwater and river water samples. The average concentrations of K+, Na+, Cl-, SO42-, NO3-N, COD / DMn, and TOC in river water samples were higher than those in groundwater, while the average concentrations of Ca2+, Mg2+, and HCO3- in the latter were higher than those in the former. Looking at the standard deviation values ​​of groundwater and river water samples, the concentration variations of Ca2+, HCO3-, Na+, Cl-, SO42-, and TDS were greater than those of other parameters.

[0094] Reference Figure 2 Principal component analysis (PCA) was performed on river water samples. The 11 species selected by PCA were K+, Na+, Cl-, SO42-, NO3-N, Ca2+, Mg2+, HCO3-, CODMn, TDS, and TOC, while pH and NH4-N were excluded due to their relatively small data variations (Table 1). Three principal components were selected. The first component (EG1, accounting for 63.9% of the variance) had high positive loadings on Cl-, Mg2+, Na+, K+, and TDS. The second component (EG2, accounting for 17.1% of the variance) had high positive loadings on NO3-N, Ca2+, and HCO3-, and high loadings on CODMn and TOC. The third component (EG3, accounting for 8.6% of the variance) had a high positive loading on SO42-.

[0095] Reference Figure 3 During the experiment, river water is affected by seasonal factors such as evaporation, river replenishment, and rainfall. The first principal component distinguishes river water during the rainy season (September sample) and the dry season (May sample), while samples collected from replenishment sources (such as SP1-09 and SP1-11) and locations near the outlet have higher scores on the second principal component. Based on the scatter plots of the first and second principal components, the three samples (SP1-09, SP1-11, and SP6-05) can be defined as the initial endmembers for the endmember pooling analysis.

[0096] The first terminal molecule (EM1) corresponds to sample SP5-09, and its TDS value (310.93 mg / L) is the lowest among all river water samples. The concentrations of Ca2+, Mg2+, Na+, K+ and Cl- (17.6 mg / L, 12.9 mg / L, 40.1 mg / L, 8.2 mg / L and 53.1 mg / L, respectively) are also the lowest among all river water samples.

[0097] The terminal molecule may be river water diluted by rainwater. The second initial terminal molecule (EM2) is initially defined as sample SP1-11, which has the lowest CODMn and TOC values ​​(3.11 mg / L and 3.30 mg / L, respectively). The third initial terminal molecule (EM3) is selected as sample SP6-05, which has the highest Cl- value (104.0 mg / L) in the river water sample. Given the conservative nature of Cl-, the high concentration of Cl- in sample SP6-05 can be attributed to the water sample collected during the dry season.

[0098]

[0099] Table 2

[0100] Reference Figure 4 A series of pooling analyses were performed on river water samples using a least squares optimization method. First, pooling analyses were performed using the same dataset selected for the river water PCA. This step was performed to qualitatively assess the conservatism of all selected species, as the estimated concentrations of conserved elements tend to coincide with the measured concentrations at a 1:1 pooling line. Then, end-membership analyses were performed using only conserved species to calculate the pooling ratio for each sample.

[0101] In the estimation process using the least squares optimization method, the standard deviation values ​​need to be used as inputs for all end-member species and water samples to represent the uncertainties related to unknown true end-members and seasonal and spatial variations. The standard deviations of non-conservative species are generally higher than those of conservative species; therefore, we set the standard deviation of Cl- in the final members and river water samples to 1%, and the standard deviation of other species to 5%.

[0102] Figure 4 A comparison plot of the estimated and measured concentrations of the river water sample and its final components was then created. The scatter plot of the estimated and measured concentrations shows that NO3-N, HCO3-, and SO42- in the water sample deviate significantly from the 1:1 mixing line, indicating that other processes were involved besides simple mixing.

[0103] Referring to Table 2, the deviation analysis results between the estimated and observed values ​​of all parameters for all samples are listed. The correlation coefficients for parameters Cl-, Ca2+, Mg2+, Na+, TDS, CODMn, and TOC are all above 0.9, and their root mean square errors are significantly lower than those of the other parameters. These results demonstrate the effectiveness of the least squares optimization method in identifying endmembers.

[0104] In addition, the selected endmembers can well characterize the spatial and seasonal variations of river water samples. In subsequent analysis, since the parameters CODMn and TOC are easily affected by hydrogeochemical reactions during the infiltration of river water, they were not included in the calculation of the mixing ratio.

[0105] This step of the analysis fully demonstrates the usability of using Cl-, Ca2+, Mg2+, Na+, and TDS as conserved species parameters for molecular analysis of river water at the end of the channel.

[0106] It should be noted that, Figure 4 The estimated concentrations of end-member components in the middle channel water are shown in Table 3:

[0107]

[0108] Table 3

[0109] Reference Figure 5 Principal component analysis was performed on the groundwater samples. For the groundwater, 12 species were selected for PCA analysis, and three principal components were retained.

[0110] The first component (EG1, accounting for 44.4% of the variance) has high positive loadings for Ca2+, Mg2+, HCO3-, and TDS. The second component (EG2, accounting for 23.0% of the variance) has high positive loadings for Cl-, Na+, and K+. The third component (EG3, accounting for 12.6% of the variance) has high positive loadings for NO3-N, NH4-N, and TOC.

[0111] The above Figure 5 and Figure 2 The different contribution patterns of the variables to the principal components indicate that the hydrochemical control factors differ between river water and groundwater samples.

[0112] Reference Figure 6 At the sampling points in the demonstration area, the groundwater in the multi-layer aquifer system can be regarded as a mixture of river water and environmental groundwater. The PCA scores of all groundwater samples are shown. In the first PC, the positive loadings of Ca2+, Mg2+, HCO3- and TDS are relatively high, indicating that mineral dissolution has affected the chemical properties of groundwater. In the second PC, the positive loadings of K+, Na+ and Cl- are relatively high, which clearly indicates that the infiltration of river water has affected the chemical properties of groundwater.

[0113] exist Figure 6The final composition also includes three river water samples (SP6-05, SP1-11 and SP5-09) that were identified as final components in step S2. Groundwater samples (such as 13(30), 13(50), 46(30) and 46(80)) are shown to be affected by anthropogenic sources and are therefore not included in the final composition estimation.

[0114] according to Figure 6 In addition to the three end members of the river channel, samples 32 (50) and 31 (80) were selected as two new end members EM4 and EM5 for mixed analysis.

[0115] Reference Figure 7 When using the least squares method to optimize the analysis of groundwater samples, Cl-, Ca2+, Mg2+, Na+, and TDS were considered as conserved species. The standard deviations of Cl- and other species in groundwater samples and terminal molecules EM4 and EM5 were 1% and 5%, respectively. The species concentrations in EM1, EM2, and EM3 remained constant in the five terminal molecule mixture models.

[0116] The mixed analysis was also performed by incorporating data from five categories—Cl-, Ca2+, Mg2+, Na+, and TDS—into the least squares optimization method. The results of the mixed analysis are as follows: Figure 7 As shown.

[0117] from Figure 7 The scatter plot of estimated concentrations versus measured concentrations for all mixed endmembers and groundwater samples shows that the estimated values ​​in the water samples are very close to the 1:1 mixing line, which indicates the effectiveness of the mixing model.

[0118]

[0119] Table 4

[0120] Table 4 lists the results of the analysis of the deviation between the estimated and measured concentrations in the groundwater samples. The correlation coefficient and root mean square error (normalized) indicate that the five-endmember mixture model can well characterize the variation characteristics of the five physicochemical parameters of the groundwater samples.

[0121]

[0122] Table 5

[0123] Table 5 lists the estimated and measured concentrations in the groundwater samples.

[0124] Generally, the measured concentrations of groundwater samples from aquifers at depths of 30 meters, 50 meters, and 80 meters are in very good agreement with the estimated concentrations, indicating that using these species for mixed analysis is reliable.

[0125] from Figure 4The Cl- ion mixing lines also show that the groundwater sample from the 80-meter aquifer is more influenced by the groundwater end-member EM5 compared to samples from depths of 30 and 50 meters. Based on the fact of vertical groundwater infiltration, the groundwater sample from the 80-meter aquifer should be less affected by the infiltration of river end-members (EM1, EM2, EM3) than samples from depths of 30 and 50 meters. The mixing analysis results obtained in this invention are in good agreement with known field geological conditions, indicating that the mixing mode of the five end-member components is effective.

[0126] In summary, the groundwater mixing end-member identification technology proposed in this invention, which combines principal component analysis and least squares optimization, can fully consider the influence of seasonal variations in mixing end-members, effectively elucidate the groundwater mixing process caused by the infiltration of surface water or reclaimed water from river channels, and quantitatively identify multiple end-members and their estimated concentrations.

[0127] The embodiments described above are merely preferred embodiments of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this application. It should be noted that those skilled in the art can make various modifications, improvements, and substitutions without departing from the concept of this application, and these all fall within the protection scope of this application.

Claims

1. A method for identifying groundwater mixing end-members considering the influence of seasonal variations in river water, characterized in that, Includes the following steps: Step S1: Screening data from river and groundwater samples; Step S2: Analyze the principal components of the river water sample; Step S3: Screening river water mixing end-members; Step S4: Analyze the main components of the groundwater sample; Step S5: Screening groundwater mixing end-members; The screening data parameters for the water sample in step S1 include: field index, cations, anions, nitrogen, total organic carbon, permanganate index, and total dissolved solids; The screening method uses a direct elimination method; The specific steps for screening river water mixing end-members in step S3 are as follows: Step S301: Perform mixed analysis on the river water samples; Step S302: Perform terminal membership analysis using conserved species and calculate the mixing ratio for each sample; Based on the multiple river water end-members identified as the final components in step S3, and excluding groundwater samples that show anthropogenic influence, the principal component analysis results of the groundwater are used to create a two-dimensional scatter plot. The principal component scores of the river water end-members are then superimposed, and groundwater samples are selected as new initial groundwater mixing end-members based on the distribution of the two-dimensional scatter plot. Because river water is affected by seasonal factors such as evaporation, river replenishment, and rainfall, principal component analysis was performed on the river water to extract seasonally relevant factors and calculate principal component scores to distinguish the seasonal influence. The first principal component distinguishes river water in the rainy and dry seasons, while samples collected from the replenishment source and near the outlet have higher scores on the second principal component. Based on the scatter plots of the first and second principal components, the three samples were defined as the initial endmembers for the endmember mixture analysis. The three samples are SP1-09, SP1-11, and SP6-05. The first terminal molecule corresponds to sample SP5-09, which has the lowest TDS value among all river water samples, and the lowest concentrations of Ca2+, Mg2+, Na+, K+, and Cl- among all river water samples; therefore, this terminal molecule represents river water diluted by rainwater. The second initial terminal molecule is defined as sample SP1-11, which has the lowest CODMn and TOC values. The third initial terminal molecule is selected as sample SP6-05, which has the highest Cl- value among river water samples; therefore, the high Cl- concentration in sample SP6-05 can be attributed to water samples collected during the dry season. When using the least squares method to optimize the analysis of groundwater samples, Cl-, Ca2+, Mg2+, Na+, and TDS were considered as conserved species. The standard deviations of Cl- and other species in groundwater samples and terminal molecules EM4 and EM5 were 1% and 5%, respectively. The species concentrations in the first terminal molecule, the second initial terminal molecule, and the third initial terminal molecule remained unchanged in the mixed model of the five terminal molecules.

2. The groundwater mixing end-member identification method considering the influence of seasonal river water variations according to claim 1, characterized in that, The specific steps for analyzing the principal components of the river water sample in step S2 are as follows: Step S201: Set up a dataset X containing m samples and n chemical and physical parameters; Step S202: Calculate the mean vector of the samples, center the dataset to obtain the centered dataset, and then calculate the covariance matrix of the data. Step S203: Perform eigenvalue decomposition on the covariance matrix to obtain eigenvalues ​​and corresponding eigenvectors; Step S204: Sort the feature values ​​according to their size; Step S205: Obtain the principal component score based on the eigenvectors corresponding to the sorted eigenvalues.

3. The groundwater mixing end-member identification method considering the influence of seasonal river water variations according to claim 2, characterized in that, The formula for calculating the covariance matrix in step S202 is as follows: ; It is data that has been centrally processed. .

4. The groundwater mixing end-member identification method considering the influence of seasonal river water variations according to claim 3, characterized in that, Perform eigenvalue decomposition on the covariance matrix to obtain the eigenvalues. and the corresponding feature vector .

5. The groundwater mixing end-member identification method considering the influence of seasonal river water variations according to claim 4, characterized in that, The first feature vector can be represented as The eigenvectors correspond to the principal component directions in the dataset, and the eigenvalues ​​represent the variance of the data explained by each principal component.

6. The groundwater mixing end-member identification method considering the influence of seasonal river water variations according to claim 5, characterized in that, The contribution rate of each principal component PC is: 。 7. The groundwater mixing end-member identification method considering the influence of seasonal river water variations according to claim 6, characterized in that, Sort the eigenvalues ​​by magnitude and select the eigenvectors corresponding to the top k eigenvalues ​​from largest to smallest; each principal component summarizes the original variable as follows: 。 8. The groundwater mixing end-member identification method considering the influence of seasonal river water variations according to claim 7, characterized in that, Principal components can explain the maximum variance in the dataset. Selecting the first few principal components effectively preserves the information in the data. The original water sample data is projected onto the principal component coordinate system to obtain the principal component scores of different samples.

Citation Information

Patent Citations

  • Water body tracing characteristic index system and water body tracing method

    CN116223751A

  • Method for calculating karst underground river basin surface karst water reserves

    CN116227178A