Introduction

The rapid development of economy and the continuous growth of population have led to the rapid expansion of domestic waste landfills. The leachate produced by landfill contains high concentration of organic matter, inorganic matter and heavy metals1,2,3. And groundwater polluted by the landfill leakage has become an important ecological and environmental problems1,2,3. Most developing countries such as China1,4,5, India6,7, South Africa8, Vietnam9, and Brazil10 have reported leakage pollutions in landfill. It may be closely related with the development stage of these countries11. In order to identify the pollution degree, existing researches on landfills mainly focus on the pollution migration process, water quality detection, hydrogeochemical analysis and other methods2,12,13. Danley et al.14 measured conventional parameters such as chemical oxygen demand (COD), total organic carbon (TOC), total dissolved solids (TDS), pH and heavy metals in groundwater contaminated by leachate through sampling and geochemical techniques. During the operation of a landfill or after its closure, the increase in number and area of liner system defects caused by chemical aging of geomembrane is the basic reason for long-term leakage of leachate and groundwater pollution15. However, few studies have paid attention on the infiltration of leachate from reservoir basin to groundwater. And it is very important for preventing leachate infiltration and controlling groundwater pollution.

Studies show that the infiltration of leachate into underground aquifers was a complex and variable process16. In simple landfills without liner system, various organic and inorganic compounds may pollute the unsaturated and saturated area through vertical infiltration of leachate, reaching to the groundwater17. Even if the landfill has a liner system, the long-term impact of landfill leakage on the environment would be seriously underestimate18,19. The performance of anti-seepage system degraded significantly for the hidden leakage points formed by ground settlement, puncture, weld cracking, material aging and other factors20,21,22. As a non-invasive technique, geophysical technology can preliminarily detect the geological environment characteristics of study area23, and the pollution plumes produced by landfill leachate infiltration24,25. Based on the difference of physical properties between landfill bodies, unsaturated zones, polluted and natural groundwater, geophysical methods were widely applied to identify groundwater pollution24,26,27. In addition, with the development of new technologies and the improvement of people’s awareness of environmental risk, health risk assessment had gradually become one of the hot research fields in China20,28,29. The health risk assessment relies on monitoring single point data, while the spatial distribution characteristics of risks would be obtained by combining with numerical simulation method28,30,31. Methods used in the previous researches of landfill are relatively simple (Table S1), and hard to accurately locate leakage points and assess health risk from the leachate pollution, which is of great significance for the effective disposal of the landfill and control of human health risk. Valley-type landfills are the most typical landfills in South China, but few researches focus on them12. Surface water in hilly areas is quite limited by the topography and hydrogeological conditions, while groundwater is relatively abundant and easy to obtain. Therefore, the groundwater is often used as a distributed water source and the main source of drinking water for local residents. The study of valley landfills is particularly important for the protection, management and governance of regional water resources. In view of this, A valley-type landfill in central and southern China was taken as an example in this study for the following purposes: (1) geophysical technology (transient electromagnetic method) were conducted to detect and determine the pollution range and depth produced by landfill leachate infiltration; (2) based on the traditional data collection method, the groundwater and landfill leachate were collected to analysis pH and the contents of TDS, total hardness, CODMn, NH3-N, SO2−, Cl−, F−, NO3-N, NO2-N, heavy metals (Hg, Fe, Mn, Cu, and Zn), and the NPI was used to assess groundwater quality and pollution levels; (3) the CA and PCA were applied to analyze the environmental factors of groundwater polluted by seepage; (4) RBCA model and numerical simulation method were used to delimit unacceptable risk areas, hoping to provide necessary theoretical support for the effective management and treatment of leachate pollution and local health risk in the landfill.

Materials and methods

Overview of the study area

The valley-type domestic waste landfill in South China was put into operation in 2014 and is still running. Its covering area is about 10,000 m2, with the storage capacity of 1.22 million m3. The designed daily capacity is 150 tons of domestic waste with the service life of 19 years. and the current daily disposal of garbage is about 60 tons. The actual used storage capacity of the existing landfill is about 450,000 m3, and the remaining storage capacity is about 770,000 m3. The studied area has a subtropical monsoon humid climate with abundant sunlight, rainfall, and the annual average temperature of 17.8 °C. Rainfall concentrates in April to October, with an average precipitation of 1660 mm. The catchment area of the landfill area is about 100,000 m2.

The landfill is located in a small arc valley surrounded by low mountains on the north, west and south sides, only opens to the northeast (Fig. 1). The barrier system for leakage prevention from the waste body is arranged at bottom and slope of the landfill with thickness ranging from several meters to more than ten meters. The surface water system is undeveloped. The strata of the landfill mainly include quartz sandstone and siltstone of Middle Devonian Tiaomajian Formation (D2t), artificial fill (Q4ml) and residual silt formed by weathered quartz sandstone (Q4eld) of Quaternary system. None magmatic rocks are found in this region. Fold structure and rock fracture are developed in the study area, which control the groundwater flow direction from southeast to northeast. Two descending springs outcrop at the upper and central landfill. The shallow metamorphic quartz sandstone with weathered fissure is the main aquifer. It is recharged by atmospheric precipitation along the rock structures and weathered cracks in process of hillsideoverflow32.

Fig. 1
figure 1

Location of the study site. Note: The satellite imagery was obtained from the Ovital map (https://www.ovital.com/) and the map was drawn using ArcGIS 10.2 (https://www.esri.com/).

Full size image

Geophysical research

According to the field investigation, the landfill was covered with a thin impermeable layer composed of geomembrane and a small amount of clay on the surface. The domestic waste layer (1–18 m thick) is buried below. The bottom liner system (1 m thick) is an impermeable layer located at the bottom and surrounding of the landfill. The lower bedrock is mainly composed of Quaternary silty clay and bedrock (quartz sandstone), with a thickness of 10–20 m. According to the previous data combined with field detection, the anti-seepage system was a relative insulation layer with the high resistance due to the characteristics of the material, and the resistivity value was as high as thousands to tens of thousands of Ω·m. In addition, the resistivity of the quartz sandstone, siltstone and other bedrock outside the site was several hundred Ω·m. The obvious difference of electrical properties between landfill body, anti-seepage system and underlying bedrock provides the premise of electromagnetic exploration.

From March to May 2021, transient electromagnetic method (TEM) was used to detect the impervious layer in this landfill (Fig. 1). 25 measured lines were set, and data collection adopted the positioned measurement mode, with 5 m interval of measuring points, 6.25 Hz of transmitting frequency, 10 A of transmitting current, 50 µS of turn-off time, and 400 times of periodic superposition. The HPTEM system had its own processing software for data processing. In field work, it was essential to strictly adhere to the instrument operation manuals and relevant technical specifications, and attention was given to improving electrode grounding conditions and overcoming random interference. Furthermore, different operators were selected to detect the representative section repeatedly at the same position for self-inspection to ensure the quality of measurement results.

Samples collection and analysis

Sampling of groundwater and leachate were carried out in the summer of 2022, as the water yield and water level would be greatly reduced in the dry season due to the climate and topography. According to the geological environment conditions, hydrogeological characteristics, sewage seepage location and regional catchment relationship in the study area, a total of 7 groundwater sampling points were set up to investigating the groundwater environment (Fig. 1) including 5 boreholes (ZK01 ~ ZK05), 1 existing groundwater monitoring well (LS01) and 1 background well (LS02) at the upstream of studying area. All the water samples were taken from the quartz sandstone reservoir to effectively cover the entire landfill site. Untreated leachate was taken from the entrance of the leachate equalization tank located at the northeast part of the landfill. All groundwater and leachate samples were stored and transported to laboratories for further analysis.

Before the analysis, the samples were filtered with a 0.45 μm microporous filter membrane and sealed in a 500 mL polyethylene bottle that was washed with deionized water and fully moistened with water sample. The samples for the heavy metal concentrations tests should be acidified to the pH value ≤ 2 by adding the nitric acid, then stored in a light-free refrigerator at 4 °C. 19 chemical indexes were measured including pH, TDS, TH, CODMn, NH3-N, SO42−, Cl−, F−, NO3-N, NO2-N, and heavy metals (Hg, Fe, Mn, Cu, Zn, Cr, Pb and Cd). The test method, execution standards and detection limits for these indexes were shown in Table S2. Blank sample was set in each group of the samples, the relative standard deviation was guaranteed to be less than 5%, and the measurement results were controlled within the range required by the quality standard.

Analytical method

A strategy to map pollution and risk of groundwater is depicted in Fig. 2, in which the main topic comprised: (i) the groundwater quality and pollution assessment; (ii) the health risk assessment for main pollutants.

Fig. 2
figure 2

Flowchart for the methodology.

Full size image

Nemerow pollution index

The Nemerow pollution index (NPI) is a comprehensive pollution evaluation method33, and the calculation process is as follows:

$$\:NPI={\left[\left\{{\left({M}_{i}/{I}_{i}\right)}_{mean}^{2}+{\left({M}_{i}/{I}_{i}\right)}_{max}^{2}\right\}/n\right]}^{\frac{1}{2}}$$
(1)

where, \(\:{M}_{i}\) is the measured concentration of pollutant i, \(\:{I}_{i}\) is the highest desirable concentration of pollutant i, and \(\:n\) represents the number of pollutants. \(\:{\left({M}_{i}/{I}_{i}\right)}_{mean}\) and \(\:{\left({M}_{i}/{I}_{i}\right)}_{max}\) are the mean and maximum values of \(\:\left({M}_{i}/{I}_{i}\right)\) for pollutant \(\:\:i\) in water. \(\:NPI\) is divided into four grades: no pollution (\(\:NPI\)< 1), slight pollution (1≤\(\:NPI\)< 2.5), moderate pollution (2.5≤\(\:NPI\)< 7), and severe pollution (\(\:NPI\)≥7).

Multivariate statistical analysis

Pearson correlation analysis can provide important information for identifying the relationship between variables and soil metal sources. On this basis, principal component analysis (PCA) can be applied to further explain the variance-covariance structure of multiple variances by transforming variances into a few comprehensive indicators34.

As a commonly used dimensionality reduction method, the PCA can summarize the few uncorrelated components and retain more information of the original variables. Descriptive statistical analysis, correlation analysis, and principal component analysis (PCA) were carried out using SPSS24.0 software. For more information, see Supplementary Material I (SM.I).

Groundwater modeling

The Visual Modflow is commonly used to simulate the groundwater pollution transportation, helping to predict the impact of different management plans on pollutant transport under different boundary conditions in different saturated heterogeneous regions35. Visual Modflow uses a modflow module to perform a three-dimensional finite-difference numerical simulation of groundwater in a porous medium to establish a flow model.MT3DMS (three-dimensional modular pollutant transport model) module is generally used to simulate solute transport in polluted aquifers. The transport model(MT3DMS) generated by the ordinary flow model is used in order to determine the pollution plume. For more information, see Supplementary Material II (SM.II).

Health risk assessment

Health risk assessment generally includes hazard identification, dose-effect relationship evaluation, exposure assessment, risk perception. Hazard identification and dose-effect relationship assess the relationship between the population exposure degree to contaminants and the possibility of negative effects. For the more details of the health risk assessment, see Supplementary Material III (SM.III) and Table S3.

According to the formula of risk calculation (SM.III), the control variable affecting risk value is the concentration of pollution factor Cw in drinking water ingested by human. After detecting the instantaneous concentrations of pollutants, the health risk can be evaluated and characterized, but the characterization form is instantaneous and discrete. The numerical simulation method coupled with the above assessment method has the function of predicting and continuously output the concentration of observation points, which can realize the continuous characterization of groundwater risk index HI in time and space dimension, and carry out risk zoning which is more conducive to the prevention and management of health risks of groundwater drinking water sources.

Data analysis

The location map was drawn using ArcGIS 10.2 (https://www.esri.com/). The descriptive statistics and CA/PCA of the test data were analyzed by SPSS24.0 software (https://www.ibm.com/cn-zh/spss). And the risks of groundwater pollution on human health were predicted through numerical modeling using theVisual Modflow 4.0 (https://www.waterloohydrogeologic.com/products/visual-modflow-flex).

Results and discussion

Determination and characteristic of the leachate leakage

The overall decline in resistivity can be influenced by soils formed from altered or broken rocks and cracks and other geological structures10,42. As the landfill body is mostly domestic garbage, it is easy to corrode and produce leachate. The ion concentration in the leachate was high, and had a low electrical resistance, strong conductivity with the resistivity value below 100 Ω·m. Thus, the appearance of low resistivity could be explained by high iron concentrations in leachate43,44. The electrical characteristics in the studied landfill showed obvious longitudinal three-layer structure of low-medium-high, corresponding to the upper waste stacking layer, the medium impervious layer, and the quartz sandstone at the bottom (Fig. 3).

The resistivity of the inversion section was more sensitive, and 8 inversion sections showed funnel-shaped low resistivity anomalies. Low resistance zone appeared under the landfill body in survey lines of S022 ~ S024, and two abnormal areas with low-resistance were found in the line S024. It was inferred that the anomalies were caused by the leakage of landfill leachate. On the basis of field investigation, it was confirmed that leakage areas were existed in the impervious layer of the landfill, which were respectively named as leakage area I and II. Areas with low resistance were also measured in the deep lines of S005 ~ S007 and the lines S008-S009 (at depth of 60–70 m), which named as the leakage area III and IV, respectively. By Comparing the inversion sections of survey lines S005-S009 and S022-S024, the low-resistance areas were narrow and extended deeper down (lines of S005-S009), and the undulating contact boundary areas between the waste body and the quartz sandstone were unclear (lines of S022-S024). It can be clearly seen that the leachate diffused vertically, and the landfill impervious layer was broken.

Through analyzing of the multiple anomalies with low resistance occurred in different survey lines, it is inferred that leakage areas are existing in the middle of the landfill site with a depth of 15–20 m (lines of S005 ~ 009), the middle area near the anti-seepage dam with a depth about 17–22 m (lines of S022 ~ 024), and the south area of the dam with a depth about 15–22 m (line of S024). These low-resistance anomalies were drafted on the study area plan (Fig. S1), and four leakage areas were inferred. Leakage area I was located at the middle of landfill zone near the dam body, with the polluted area about 132 m2 and the leaked depth of 17–22 m. Leakage area II was at the south of landfill zone, with the polluted area about 26 m2 and the leaked depth of 15–22 m. Leakage area III and IV were also at the middle of the landfill, with the polluted range of 97.1 m2 and 81.7 m2 respectively, and depth both about 15–20 m.

Fig. 3
figure 3

Transient electromagnetic method 2D inversion profile of different survey lines. (a)–(e) represent five lines of S005 to S009 in the middle of the landfill site, respectively; and (f)–(h) represent three lines of S022 to S024 near the dam, respectively)

Full size image

Characteristics of groundwater quality and pollution

In order to further understand the possible influence of landfill leachate on surrounding environment, untreated leachate from the regulating tank was analyzed specially and the results are shown in Table 1. In order to evaluate the quality of leachate, we used the standard for pollution control on the landfill site of municipal solid waste45, which was issued by China’s Ministry of Ecology and Environment to regulate pollution control and supervision of landfill. Except As, all the measured values of parameters exceeded the emission limits of water pollutants, and NH3-N, CODMn which reflect the main pollution factors of household waste far exceeded the standard by dozens or hundreds of times. The NPI method was used to evaluate the leachate pollution in the landfill site. The NPI of leachate of landfill site was 83.6 (Table 1). The maximum values of leachate pollution indexes in two landfills from Poland (Lubna) and the Czech Republic (Zdounky) were 39.41 and 14.25, respectively, and the alkaline characteristics of both landfills were also evidenced by CODMn, NH3-N and Cd46. The formation of leachate is mainly due to the percolating of rainwater and the physical, chemical, and biological processes through the landfill body46. The leachate gathers through the horizontal and vertical block by impervious barrier47, causing a potential huge threat to the surrounding environment.

Table 1 Composition and quality assessment of landfill leachate (mg/L).
Full size table

According to the Groundwater Quality Standard48, tested parameters of groundwater from the upstream to the downstream in the studied area are shown in Table S4. The concentrations of As, Cr, Pb and Cd in all samples, as well as fluoride, Cu and nitrite in some samples were below the detection limits. The pH values in groundwater ranged from 5.80 to 6.70, with the rate of 28.57% exceeding the standard value of Class III water in the Groundwater Quality Standard48. And the concentrations of NH3-N, Fe and Mn ranged from 0.06 to 0.74 mg/L, 0.002–2.763 mg/L, and 0.003–1.181 mg/L, with the rate of 28.57%, 4.29%, and 14.29% exceeding standard, respectively. The maximum concentrations were 1.48, 9.21, and 11.81 times of the values of Class III water quality standard, respectively. In addition, the groundwater quality at ZK02, ZK03 and ZK04 sampling points reached the Class IV standard, and the main pollutants were pH, NH3-N and Mn. While SK01 reached Class V water quality standard, its pH and Mn reached Class IV standard and Fe reached to Class V standard, which showed the extremely poor groundwater quality.

Generally speaking, leachates generated from upstream site may infiltrate and pollute downstream groundwater, and furthermore lead to the concentrations of Fe, NH3-N, NO3-N and other compositions exceeding the standard values13. As shown in Fig. 4, all parameters of borehole samples (ZK01 to ZK05) were higher than that of background wells (LS01 and ZK02), suggesting that the groundwater system in the landfill has been polluted. According to the locations of each sampling point, the linear distance to landfill site from far to near was ZK03 < ZK04 < ZK02 < ZK01 < ZK05, and the sampling depth from shallow to deep was ZK03 < ZK01 < ZK05 < ZK04 < Zk02. But the concentration of indicators did not change regularly with the distance and depth, indicating that there were other contributing factors in the pollution process. The impact of landfill on groundwater quality depends on waste composition, soil accumulation, rainfall, hydraulic gradient and the geological environment of the landfill49. An investigation reveal that the radial diffusion of pollutants in the clay liner after leakage from an geomembranes is 5.5–20.0 times higher than the theoretical value, reaching to 58 m1. Due to the performance degradation of liner systems, the impact on groundwater changes over the time of leachate leakage, with the impact extent expanding from 50 to 1000 m18.

Fig. 4
figure 4

Concentrations of pollutants in groundwater. Note: (a) contains high-value parameters like pH, TDS(total dissolved solids), TH(total hardness), CODMn(potassium permanganate index), NO2-N(nitrite), SO42−(sulfate) and Cl−(chloride); and (b) contains low-value parameters like F−(fluoride), NH3-N(ammonia nitrogen), NO3-N(nitrate), and and heavy metals (Hg, Fe, Mn, Cu, and Zn).

Full size image

The highest concentration of Fe was found at ZK01, secondly at LS01 background well. Contents of Fe and Mn in quartz sandstone in the study area were 4.47% and 0.21%, respectively. The excess of Fe in groundwater may be related to the underlying strata. The highest concentration of Mn was found at ZK04, followed by ZK01 to ZK03, but the content of Mn was very small in the background monitoring points. ZK04 was sampled beneath the damaged anti-seepage dam and closest to the waste body, indicating that Mn pollution in groundwater possibly mainly originated from landfill leakage rather than underlying bedrock. The concentrations of NH3-N were higher in samples of ZK02 to ZK04, followed by ZK05 and ZK01, and a small amount was detected in the background wells. NH3-N mainly originated from the leakage of landfill, and spread to the surrounding pollution with the landfill as the center. It mainly accumulated downstream controlled by rock fractures, hydraulic gradient and dam body damaged.

In order to further understand the groundwater quality characteristics, the NPI method was used based on the parameter concentrations of background well at LS01 and the Class III standard in the Groundwater Quality Standard48. The results (Table 2) showed that compared with the local background value, groundwater of all monitoring sites in the landfill were heavily polluted (NPI > 3), with a descending order of ZK04 (NPI = 23.19) > ZK01 (NPI = 19.46) > ZK03 (NPI = 8.75) > ZK05 (NPI = 8.65) > ZK02 (NPI = 5.31). While compared with the Class III standard (GB/T 14848 − 2017)48, only ZK01 and ZK04 were severe polluted condition (NPI > 3), samples of ZK02 and ZK03 were in slight polluted risk (1.0 < NPI ≤ 2.0), and ZK05 was non-polluted (NPI ≤ 0.7). The order of NPI for samples in descending order was ZK04 (NPI = 8.39) > ZK01 (NPI = 6.61) > ZK03 (NPI = 1.68) = ZK02 (NPI = 1.68) > ZK05(NPI = 0.55). The two calculated results showed a great difference about the comprehensive pollution degree of groundwater. But groundwater at ZK01 and ZK04 was seriously polluted in either case. The NPI method would highlight the impact of maximum value on pollution degree47,50, Fe and Mn in the sites of ZK04 and ZK01 seriously exceeded the standard, which results in that the indexes of samples ZK04 and ZK01 were much higher than that of others.

Correlations between groundwater parameters

Correlation analysis (CA)

The correlation analysis result of each parameter in groundwater of landfill site was shown in Fig. 5. pH was negatively correlated with most of the parameters, especially with Mn (r= − 0.897)and Zn (r= − 0.778). Podlasek et al. also reported a similar relationship in the leachate parameters of different landfills46. A large number of studies have confirmed that pH could affect the solidification of heavy metals in soils, and the solubility of heavy metals would increase with the decrease of pH value51,52,53. There were significant positive correlations between total mineralization, total hardness, NH3-N and SO42− in groundwater (r = 0.834–0.982). Similar results were also reported by Zhao et al.54 and Zhou et al.55, that the total dissolved solids in groundwater was significantly related with SO42− and Ca2+. Moreover, some studies believed that EC was highly associated with NH3-N56, and sensitive to changes of TDS in leachate57. The high values of these parameters indicated high contents of dissolved inorganic and organic compounds57.

Remarkably, positive correlation appeared in CODMn and Fe (r = 0.949). CODMn reflects the pollution degree of water by reducing organic or inorganic matter, which can activate metals such as Fe and Mn in soil and enhance the migration ability of these metals into the groundwater by rainfall, resulting in serious Fe and Mn pollution58. In addition, weak correlations between HMs should be noted, possibly due to their low concentration in leachate59. However, Cl− was positively related with Mn (r = 0.839) and Zn (r = 0.822), suggesting they may have the same source60.

Fig. 5
figure 5

Correlation analysis of various indicators in groundwater. Note: NO3-N: nitrate; NO2-N: nitrite; Cl−: chloride; F−: fluoride; SO42−: sulfate; NH3-N: ammonia nitrogen; CODMn: potassium permanganate index; TH: total hardness; TDS: total dissolved solids; depth: sampling depth of groundwater. * means that the 0.01 < P < 0.05 level is correlated, ** means P < 0.01 level is significantly correlated.

Full size image

Principal component analysis (PCA)

The relevance of parameters can be revealed by PCA to supplement the obstacles by CA, for minimizing the loss of data analysis16,61. Because the sample size was small for source analysis in this study, PCA analysis aimed to reduce the variables that characterize groundwater quality53.

PCA extracted five principal components from the parameters of groundwater, which accounted for 98.89% of the total variance (Table 2). Among them, PC1 accounts for 29.31% of the total variance loaded strongly by the total salinity, total hardness, NH3-N and SO42−, followed by Hg and sampling depth. Researches have pointed out that inorganic parameters such as EC, TDS, COD and NH4+ were dominant in PC1, resulting in great data differentiation46,62. The landfill formed with domestic waste and construction waste was characterized by high sulfate, high calcium ion and high TDS63. Ca and SO42− contents were important components in concrete block, engineering mud and gypsum board components of construction waste, and high correlation existed between adhesion mortar and SO42− 64. Waste batteries, fluorescent lamps, electronic products and medical devices in landfills were also important sources of Hg65.

PC2 accounts for 26.85% of the total variance, which exhibits strong loading of Cl−, Mn and NO3-N. Cl− was the main ion in natural groundwater, and the formation mechanism of chloride in groundwater was mainly the dissolution of various Cl-containing rocks, such as albitite, apatite, mica and hornblende formed by long-term geological deposition66,67. These Cl-containing minerals were rich in calcium and easy to be isotropic substitutes with trace elements such as Zn, Mn and Cr.

The CODMn and Fe have extremely strong loads on PC3, accounting for 17.23% of the total variance. The Fe content of the underlying quartz sandstone in the study area was high, and researches also showed that Fe in natural groundwater could be affected by local hydrogeological conditions and rock characteristics68,69. The CODMn and Fe content in the leachate were far more than the standard values (Table 1), indicating the landfill contributed greatly to Fe and CODMn in groundwater.

PC4 is strongly correlated with Cu and F−, and accounts for 13.39% of the total variance. The contents of Cu, F− and Zn were higher in sample ZK05, which may be influenced by both nature and human activities. PC5 is related to strong negatively loading of NO2-N and account for 12.11%. NO2-N was an intermediate product of the nitration reaction in the downward migration process of NH3-N, its content increased slightly in the middle layer, and gradually decreased as the depth continues to increase70.

Table 2 The PCs distinguished from the chemical characteristics of groundwater.
Full size table

Health risk assessment

In order to obtain the change of the evaluation control variable (\(\:{C}_{w}\)) and characterize the risk of difference areas in continuous space and time, a three-dimensional numerical model was built with the Modflow software based on the hydrogeological model. And the current monitoring values of groundwater level and water quality were used as the reasonableness verification basis, the change trend of groundwater pollutant concentration could be predicted and calculated. After reasonable verification, the model would output the change trend of the evaluation control variable Cw, and accurately depict thescope of various risk areas in different periods coupled with the above-mentioned health risk assessment method.

Model construction and calibration

  1. (1)

    Model grid and parameterization

The main type of groundwater type in simulated area is shallow weathering fissure water of bedrock. The occurrence medium is typical clastic rocks, mainly composed of interbedded quartz sandstone and siltstone, and groundwater mainly exists in the weathered fracture grid. Based on the hydrogeological drilling and physical exploration data, a groundwater model with the drainage surface of Hanchong River and its tributaries and an area of 1.1 km2 was established. According to the difference of permeability coefficient, it was vertically divided into 2 layers, corresponding to the strong regolith and medium-weak regolith respectively. The thickness of each layer was generalized based on the drilling. The drilling data and the corresponding vertical division of the model can be found in Supplementary Fig. S2 and S3 online. The depth of strong regolith was 15–25 m. The grid size in the model is subdivided into 10 × 10 m. The main controlling factors of mode accuracy were hydrogeological and solute transport parameters, which were obtained through in-situ hydrogeological experiments and data collection. The values of various parameters were input to a fitted reasonable model, as shown in Table 3.

Table 3 Model parameter values used in numerical experiments.
Full size table
  1. (2)

    Boundary conditions and pollution source setting

Boundary conditions reflect the process and intensity of water and material exchange between the model and external entities, and which is a necessary condition for solving the simulation calculation71. Its setting should be based on the correct understanding of the hydrogeological conditions in this simulated area to ensure that the objects is reflected by the mathematical model. The hydrodynamic conditions of shallow weathering fissure water were significantly controlled by local topography and adjacent waterbody. Therefore, the boundary conditions were set as follows: (a) the surface watersheds in the west, south and east were set as Neumann boundary with a flow of 0 m3/d; (b) the north side was the downstream of simulate area and set as Neumann boundary with a flow of 40 m3/d through the calculation of the outgoing model flow based on Darcy’s law; (c) The controlling water system of Hanchong River in the middle model was set as Dirichlet boundary; (d) the 7 monitoring wells around the landfill were set as monitoring wells for water level and water quality to fit and check flow field and pollution status. The model domain with the boundary conditions, positions of the observation wells and pumping wells is shown in Fig. S4.

  1. (3)

    Flow-model calibration

According to the hydrogeological parameters measured in this study, the parameters of hydraulic conductivity, recharge and specific yield in the numerical model were set. After calculating the output flow field data, the rationality of the model was verified by the measured water level data. As shown in Figs. 6 and 19 actual observed wells were used in flow model validation. Table 4 shows that the mean square root error and maximum root mean square error are very small, only 0.43 m and 0.33 m, respectively. The calibration model matches the calculation results of observation head.

Table 4 Residual analysis of head values.
Full size table
Fig. 6
figure 6

Equipotential lines and vectors that indicate the flow direction and magnitude of underground flow.

Full size image
  1. (4)

    Transport model calibration

The purpose of numericalsimulation is to obtain the spatial distribution characteristics and trend over timeof pollutants to assess human health risk. After calibrating the flow field model, the solute transport model needs to be further verified. The trial-and-error method was suitable for the calibration of seepage field model and solute transport model72. According to the mathematical equation of solute transport, the uncertainty parameters affecting the solute transport model include the dispersion coefficient of pollutants and pollutant flux which is equal to the product of solvent infiltration and pollutant concentration. The dispersion coefficient in the flow field model determined and verified by hydrogeological parameters of aqueous medium was no longer regarded as a variable, and the parameters that affected the output of concentration value were solvent infiltration and pollutant concentration. The calculated model value was compared with the actual monitoring values of indicators to complete calibration by adjusting the pollutant flux.

As shown in Fig. S5, 10 actual concentration observed wells were used for the transport model verification. After calibrating the solute-transport model, the RMS errors of Mn and NO3-N were 0.01 mg/L and 0.521 mg/L, respectively. The absolute residual mean values were 0.008 mg/L and 0.344 mg/L, respectively (Fig. S6).

The calibration model fitted according to the current situation shows that the landfill leakage has had an impact on the groundwater. The affected area was concentrated in the surrounding area and downstream of the groundwater runoff direction. The contents of Mn and NO3-N at pollution point reached 2.0 and 15.2 mg/L, respectively. The corresponding hazard index of Mn and NO3-N were 2.50 and 1.66, respectively, which is greater than the standard risk value 1. The scope of HI greater than 1 for Mn was 28,020 m2, with the longest distance of 70 m downstream on the northeast side of the landfill. The scope of HI greater than 1 for NO3-N was 19,470 m2, the farthest distance is 50 m downstream of the northeast side of the landfill.

Based on the verified fitted model, the HI values indicating non-carcinogenic health risk of Mn and NO3-N were calculated assuming that no pollution remediation were taken, and risks after 5a, 10a, and 20a were evaluated and characterized, respectively.

Result analyze

If the pollution source is not cut off and other measures are not taken, the maximum value of pollutants in groundwater continues to rise and the migration range of pollutants continues to expand. At the same location within the polluted area, the concentration of pollutants would continue to rise over time, and the health risk gradually accumulated. The HI after 5a, 10a, and 20a were shown in Fig. 7. If the landfill continues operating for 5a, maximum HI value for Mn and NO3-N would increase to 2.75 and 1.88, respectively. The scope of HI greater than 1 for Mn would increase to 29,345 m2, extending to 92 m below the landfill, and increasing by 4.7% compared with the covering area. And for NO3-N, it would increase to 21,015 m2, extending to 73 m below the landfill and increasing by 4.7%. After 10 years, the maximum values of HI calculated for Mn and NO3-N was 3.25 and 2.10, with the range of HI greater than 1 increased by 12% to 31,405 m2 and 23% to 23,930 m2, respectively.

The predicted results showed that the average increase rate of HI extreme value for pollutants in the groundwater within the past 20 years was between 0.04 and 0.08/a, and the average expansion rates of the risk area were 341–432 m2/a. The expansion rates of risk area along the groundwater runoff were 1.6–3.8 m/a. The upstream and lateral risk area characterized with the diffusion as the migration mode extent by 0.59–0.63 m/a. The safety of drinking water downstream of main pollution source runoff is the focus of protection and pollution prevention and control. Therefore, it is urgent to take active measures for soil and groundwater remediation.

Fig. 7
figure 7

Distribution maps of the HI of Mn and NO3-N for human health after 5a, 10a, and 20a. The satellite imagery was obtained from the Ovital map (https://www.ovital.com/) and the map was drawn using ArcGIS 10.2 (https://www.esri.com/).

Full size image

Uncertainty analysis

This study contains some uncertainties due to limitations in technical approaches and data availability. For instance, the detection process of transient electromagnetic method (TEM) is susceptible to site-specific environmental factors. Therefore, the delineated low-resistivity anomalies may exhibit interpretive ambiguities, requiring further verification during practical application. Additionally, parameters in the numerical simulation models, such as hydraulic conductivity coefficients and rainfall recharge rates, were approximated through interpolation techniques to represent the complex heterogeneous characteristics of surface water and groundwater in the study area. Such simplifications propagate uncertainties into simulation outcomes. Although the results reflect the current groundwater flow and contamination states, limitations persist in the prediction of contaminant transport and transformation processes. Seasonal hydrogeological variations exert significant influence on contaminant dynamics, yet the single-phase groundwater level and water quality monitoring data used for model calibration failed to capture these temporal fluctuations. To reduce uncertainty impacts and improve the model reliability, it is recommended that details of model should be optimized by improving the hydrogeological surveys precision, or implementing continuous groundwater monitoring systems to collect long-term, multi-phase observational datasets, to help accurately simulating and predicting groundwater flow and contaminant transport processes.

Conclusion

In this study, geophysical exploration, chemical sampling analysis and model simulation were conducted on the groundwater of a valley-type landfill in South China, providing a strong basis for control, repair and management of leachate pollution. The main conclusions are as follows:

  1. (1)

    The detection results of transient electromagnetic method showed that there are four damage leakage areas of anti-seepage layer in the reservoir area of the landfill site, the area and vertical impact depth ware 132 m2 and 17–22 m (leakage area I), 26 m2 and 15–22 m (leakage area II), 97.1 m2 and 15–20 m (leakage area III), and 81.7 m2 and 15–20 m (leakage area IV), respectively.

  2. (2)

    The results of NPI showed that ZK01 and ZK04 were in the state of heavy pollution (NPI > 3), ZK02 and ZK03 had a slight pollution risk (1.0 < NPI ≤ 2.0), and only ZK05 was non-polluted (NPI ≤ 0.7). Water quality of points ZK01 ~ 04 exceeded the standard value of Class III water in the Groundwater Quality Standard (GB/T 14848 − 2017), with the excessive parameters of pH, NH3-N and Mn.

  3. (3)

    The CA results showed that pH was strongly negatively correlated with Mn (r = 0.897) and Zn (r = 0.778). There were significant positive correlations between TDS, TH, NH3-N and SO42− (r = 0.834 ~ 0.982), and a significant positive correlation between CODMn and Fe (r = 0.949). The results of principal component analysis showed that the five main components accounted for 98.89% of the total variance. The mainpolluted sources of groundwater were landfill leakage, rock weathering dissolution and water-rock interaction.

  4. (4)

    Numerical simulation based on the RBCA health risk model predicted that the values of hazard index for Mn and NO3-N were respectively 2.50 and 1.66, posing a non-carcinogenic risk to human health. Numerical simulation predicted that the average increase rate of HI for pollutants within the past 20 years was between 0.04 and 0.08/a, and the average expansion rates of the risk area were 341-432m2/a. The expansion rates of risk area along the groundwater runoff were 1.6–3.8 m/a. The upstream and lateral risk area with the diffusion as the migration mode extent were 0.59–0.63 m/a.