Use of artificial neural network to evaluate cadmium contamination in farmland soils in a karst area with naturally high background values

Machine learning-based prediction of cadmium pollution in topsoil and identification of critical driving factors in a mining area

Mining activities have resulted in a substantial accumulation of cadmium (Cd) in agricultural soils, particularly in southern China. Long-term Cd exposure can cause plant growth inhibition and various diseases. Rapid identification of the extent of soil Cd pollution and its driving factors are essential for soil management and risk assessment. However, traditional geostatistical methods are difficult to simulate the complex nonlinear relationships between soil Cd and potential features. In this study, sequential extraction and hotspot analyses indicated that Cd accumulation increased significantly near mining sites and exhibited high mobility. The concentration of Cd was estimated using three machine learning models based on 3169 topsoil samples, seven quantitative variables (soil pH, Fe, Ca, Mn, TOC, Al/Si and ba value) and three quantitative variables (soil parent rock, terrain and soil type). The random forest model achieved marginally better performance than the other models, with an R2 of 0.78. Importance analysis revealed that soil pH and Ca and Mn contents were the most significant factors affecting Cd accumulation and migration. Conversely, due to the essence of controlling Cd migration being soil property, soil type, terrain, and soil parent materials had little impact on the spatial distribution of soil Cd under the influence of mining activities. Our results provide a better understanding of the geochemical behavior of soil Cd in mining areas, which could be helpful for environmental management departments in controlling the diffusion of Cd pollution and capturing key targets for soil remediation.

Uptake of zinc from the soil to the wheat grain: Nonlinear process prediction based on artificial neural network and geochemical data

Trace elements in plants primarily derive from soils, subsequently influencing human health through the food chain. Therefore, it is essential to understand the relationship of trace elements between plants and soils. Since trace elements from soils absorbed by plants is a nonlinear process, traditional multiple linear regression (MLR) models failed to provide accurate predictions. Zinc (Zn) was chosen as the objective element in this case. Using soil geochemical data, artificial neural networks (ANN) were utilized to develop predictive models that accurately estimated Zn content within wheat grains. A total of 4036 topsoil samples and 73 paired rhizosphere soil-wheat samples were collected for the simulation study. Through Pearson correlation analysis, the total content of elements (TCEs) of Fe, Mn, Zn, and P, as well as the available content of elements (ACEs) of B, Mo, N, and Fe, were significantly correlated with the Zn bioaccumulation factor (BAF). Upon comparison, ANN models outperformed MLR models in terms of prediction accuracy. Notably, the predictive performance using ACEs as input factors was better than that using TCEs. To improve the accuracy, a two-step model was established through multiple testing. Firstly, ACEs in the soil were predicted using TCEs and properties of the rhizosphere soil as input factors. Secondly, the Zn BAF in grains was predicted using ACE as input factors. Consequently, the content of Zn in wheat grains corresponding to 4036 topsoil samples was predicted. Results showed that 85.69 % of the land was suitable for cultivating Zn-rich wheat. This finding offers a more accurate method to predict the uptake of trace elements from soils to grains, which helps to warn about abnormal levels in grains and prevent potential health risks.

Potential risk assessment and occurrence characteristic of heavy metals based on artificial neural network model along the Yangtze River Estuary, China

Pollution from heavy metals in estuaries poses potential risks to the aquatic environment and public health. The complexity of the estuarine water environment limits the accurate understanding of its pollution prediction. Field observations were conducted at seven sampling sites along the Yangtze River Estuary (YRE) during summer, autumn, and winter 2021 to analyze the concentrations of seven heavy metals (As, Cd, Cr, Pb, Cu, Ni, Zn) in water and surface sediments. The order of heavy metal concentrations in water samples from highest to lowest was Zn > As > Cu > Ni > Cr > Pb > Cd, while that in surface sediments samples was Zn > Cr > As > Ni > Pb > Cu > Cd. Human health risk assessment of the heavy metals in water samples indicated a chronic and carcinogenic risk associated with As. The risks of heavy metals in surface sediments were evaluated using the geo-accumulation index (Igeo) and potential ecological risk index (RI). Among the seven heavy metals, As and Cd were highly polluted, with Cd being the main contributor to potential ecological risks. Principal component analysis (PCA) was employed to identify the sources of the different heavy metals, revealing that As originated primarily from anthropogenic emissions, while Cd was primarily from atmospheric deposition. To further analyze the influence of water quality indicators on heavy metal pollution, an artificial neural network (ANN) model was utilized. A modified model was proposed, incorporating biochemical parameters to predict the level of heavy metal pollution, achieving an accuracy of 95.1%. This accuracy was 22.5% higher than that of the traditional model and particularly effective in predicting the maximum 20% of values. Results in this paper highlight the pollution of As and Cd along the YRE, and the proposed model provides valuable information for estimating heavy metal pollution in estuarine water environments, facilitating pollution prevention efforts.

Differences in the activities of six soil enzymes in response to cadmium contamination of paddy soils in high geological background areas

East Yunnan province in southwest China is a region with elevated natural abundance (high geological background levels) of Cd due to high metal (loid) contents in the soils. Enzyme activities are useful indicators of metal (loid) toxicity in contaminated soils and whether Cd inhibits enzyme activities in paddy soils in high geological background areas is of considerable public concern. A pot experiment combined with field investigation was conducted to assess the effects of Cd on six soil enzymes that are essential to the cycling of C, N, and P in soils. Inhibitory effects of Cd fractions on enzyme activities were assessed using ecological dose-response models. The impact of soil properties on the inhibition of sensitive soil enzymes by Cd were assessed using linear and structural equation models. Cadmium was enriched in the paddy soils with 72.2 % of soil samples from high geological background areas exceeding the Chinese threshold values (GB 15618-2018) of Cd. Enzyme responses to Cd contamination varied markedly with a negative response by catalase but a positive response by invertase. Urease, β-glucosidase, and alkaline phosphatase activities were stimulated at low Cd concentrations and inhibited at high concentrations. The average inhibition ratios of β-glucosidase, urease, and catalase in high Cd levels were 19.9, 38.9, and 51.9%, respectively. Ecological dose-response models indicate that catalase and urease were the most Cd-sensitive of the enzymes studied and were suitable indicators of soil quality in high geological background areas. Structural equation modeling (SEM) indicates that soil properties influenced sensitive enzymes through various pathways, indicating that soil properties were factors determining Cd inhibition of enzyme activities. This suggests that Cd concentrations and soil physicochemical properties under a range of environmental conditions should be considered in addressing soil Cd pollution.

Using machine learning to predict selenium and cadmium contents in rice grains from black shale-distributed farmland area

Cadmium (Cd) and selenium (Se) are widely enriched in soil at black shale outcropping areas, with Cd levels exceeding the standard (2.0 mg/kg in 5.5 < pH ≤ 6.5) commonly. The prevention of Cd hazards and the safe development of Se-rich land resources are key issues that need to be urgently addressed. To ensure safe utilization of Se-rich land in the CdSe coexisting areas, 158 rice samples, their corresponding rhizosphere soils, and 8069 topsoil samples were collected and tested in the paddy fields of Ankang City, Shaanxi Province, where black shales are widely exposed. The results showed that 43 % of the topsoil samples were Se-rich soil (Se > 0.4 mg/kg) wherein 79 % and 3 % of Cd concentrations exceeded the screening value and control value, respectively, according to the GB15618-2018 standard. Meanwhile, 63 % of the rice samples were Se rich (Se > 0.04 mg/kg) and the Cd content exceeded the prescribed limit (0.2 mg/kg) in Se-rich rice by 26 %. There was no significant positive correlation between the Se and Cd contents in the rice grains and the Se and Cd contents in the corresponding rhizosphere soil. The factors influencing Se and Cd uptake in rice were SiO2, CaO, P, S, pH, and TFe2O3. Accordingly, an artificial neural network (ANN) and multiple linear regression model (MLR) were used to predict Cd and Se bioaccumulation in rice grains. The stability and accuracy of the ANN model were better than those of the MLR model. Based on survey data and the prediction results of the ANN model, a safe planting zoning of Se-rich rice was proposed, which provided a reference for the scientific planning of land resources.

Machine learning-based prediction of cadmium bioaccumulation capacity and associated analysis of driving factors in tobacco grown in Zunyi City, China

Tobacco grown in areas with high-geochemical backgrounds exhibits considerably different cadmium (Cd) bioaccumulation abilities due to regional disparities and environmental changes. However, the impact of key factors on the Cd bioaccumulation ability of tobacco grown in the karst regions with high selenium (Se) geochemical backgrounds is unclear. Herein, 365 paired rhizospheric soil-grown tobacco samples and 321 topsoil samples were collected from typical karst tobacco-growing soil in southwestern China and analyzed for Cd and Se. XGBoost was used to predict and evaluate the Cd bioaccumulation ability of tobacco and potential influencing factors. Results showed that regional geochemical characteristics, such as soil Cd and Se contents, soil type, and lithology, have the highest influence on the Cd bioaccumulation ability of tobacco, accounting for 46.5% of the overall variation. Moreover, soil Se contents in high-geochemical background areas considerably affect Cd bioaccumulation in tobacco, with a threshold for the mutual suppression effects of Cd and Se at a soil Se content of 0.8 mg/kg. According to the results of bivariate local indicators of spatial association analysis, tobacco cultivated in the central, northeast, and southeast regions of Zunyi City carries a lower risk of soil Cd contamination. This study provides new insights for managing tobacco cultivation in karst regions.

A critical review on the migration and transformation processes of heavy metal contamination in lead-zinc tailings of China

The health risks of lead-zinc (Pb-Zn) tailings from heavy metal (HMs) contamination have been gaining increasing public concern. The dispersal of HMs from tailings poses a substantial threat to ecosystems. Therefore, studying the mechanisms of migration and transformation of HMs in Pb-Zn tailings has significant ecological and environmental significance. Initially, this study encapsulated the distribution and contamination status of Pb-Zn tailings in China. Subsequently, we comprehensively scrutinized the mechanisms governing the migration and transformation of HMs in the Pb-Zn tailings from a geochemical perspective. This examination reveals the intricate interplay between various biotic and abiotic constituents, including environmental factors (EFs), characteristic minerals, organic flotation reagents (OFRs), and microorganisms within Pb-Zn tailings interact through a series of physical, chemical, and biological processes, leading to the formation of complexes, chelates, and aggregates involving HMs and OFRs. These interactions ultimately influence the migration and transformation of HMs. Finally, we provide an overview of contaminant migration prediction and ecological remediation in Pb-Zn tailings. In this systematic review, we identify several forthcoming research imperatives and methodologies. Specifically, understanding the dynamic mechanisms underlying the migration and transformation of HMs is challenging. These challenges encompass an exploration of the weathering processes of characteristic minerals and their interactions with HMs, the complex interplay between HMs and OFRs in Pb-Zn tailings, the effects of microbial community succession during the storage and remediation of Pb-Zn tailings, and the importance of utilizing process-based models in predicting the fate of HMs, and the potential for microbial remediation of tailings.

Prediction of cadmium contents in rice grains from Quaternary sediment-distributed farmland using field investigations and machine learning

The Quaternary sediment-distributed regions of South China are suitable for rice cultivation, which is crucial for ensuring food security. Spatial correlations between soil cadmium (Cd) and rice Cd contents are generally poor, making the evaluation of rice quality and associated health risks challenging. In this study, we developed machine learning algorithms for predicting rice Cd contents using 654 data pairs of soil-rice samples collected in Guangxi province, China. After a comprehensive comparison, our results showed that the random forest (RF) had the better performance than artificial neural network (ANN) based on all the data (RMSEtesting 0.066 vs. 0.099 and R2testing 0.860 vs. 0.688). The feature importance analysis showed that soil CaO, Cd, elevation, and rainfall were the four most important features affecting the rice Cd contents in the study area. Using the established RF-predicated model, the rice Cd contents were predicted at the provincial level with an additional dataset of 1176 paddy soil samples. The prediction result revealed about 23 % of farmland cultivated rice with Cd content over 0.2 mg kg-1 in the study area. Therefore, it is recommended to implement strict measures by local agricultural departments to reduce rice Cd contents and ensure food safety in these areas. Our study provides valuable insights into the prediction of rice Cd contents, thus contributing to ensuring food safety and preventing Cd exposure-associated health risks.

Carbonate rocks as natural buffers: Exploring their environmental impact on heavy metals in sulfide deposits

Carbonate rocks are closely related to the genesis and spatial distribution of polymetallic sulfide deposits. The natural buffering of carbonate rocks can reduce the ecological impact of heavy metals produced by mining and smelting. Ignoring the buffering effect of carbonate rocks on the heavy metals in the mine environment leads to inaccurate ecological risk assessment, wasting land resources and funds. This study investigates Cd, Zn, and Pb distribution and speciation in the water and soil-rice system in the polymetallic sulfide deposit at Daxin, Guangxi. The study aims to reveal the effects of the natural buffering of carbonate rocks on the migration and transformation of heavy metals. The results show that the water Zn and Cd concentrations decreased from 1857.0 to 0.9 mg L-1 to 0.16 and 0.001 mg L-1, respectively, from the mining area to 4 km downstream. The natural buffering of carbonate increases the water pH from 2.80 to 7.64, resulting in a tendency for Cd, Zn, and Pb to separate from the aqueous phase and enrich the sediments. Soil Cd content in the mining area reached 110.0 mg kg-1 (mean value 55.88 mg kg-1), and rice Cd seriously exceeded the maximum limit. However, the weathering of carbonate reduces the migration ability and bioavailability of Cd. Soil Cd is mainly in the Fe-Mn bound and carbonate-bound fractions, resulting in lower Cd content in downstream soils (mean value 2.73 mg kg-1). Soil CaO, tFe2O3, and Mn hindered the uptake of soil Cd by rice rendering a lower exceedance of Cd in downstream rice. Therefore, this study recommends a farmland management plan under the premise of rice Cd content and integrated soil Cd content, which ensures food safety and fully utilizes farmland resources. This result provides a scientific basis for ecological risk assessment, mine environmental protection, and management in the carbonatite sulfide mine environment.

Factors influencing the distribution of woody plants in tropical karst hills, south China

The seasonal rainforests distributed across the tropical karst hills of south China are of high biodiversity conservation value and serve many important ecosystem functions. However, knowledge surrounding distribution patterns of woody plants in tropical karst hills remains limited. In this study, we surveyed the distribution of families, genera and species of woody flora at four slope positions (depression, lower slope, middle slope, and upper slope), and analyzed the influence of topographic and soil variables on the distribution of woody plants in the tropical karst hills of south China. Forty forest plots (each 20 m × 20 m) contained 306 species of woody plants with a diameter at breast height (DBH) ≥1 cm, representing 187 genera and 66 families. As slope increased, the number of families increased slowly, and the number of genera and species followed a concave-shaped trend, with the lowest number of genera and species in the lower slope position. Differences in species composition were significantly stronger between slope positions than within slope positions. The topographic and soil variables explained 22.4% and 19.6%, respectively, of the distribution of woody plants, with slope position, slope degree, soil potassium and soil water content as the most significant variables. The results of generalized linear mixed model analysis showed that total R2 of fixed effects on variation of woody species richness was 0.498, and rock outcrop rate and soil total phosphorus were the best fitting effects. Our results help to explain the community assembly mechanism and to inform management and protection strategies for species-rich seasonal rainforests in the karst area.

Cadmium accumulation in paddy soils affected by geological weathering and mining: Spatial distribution patterns, bioaccumulation prediction, and safe land usage

The abnormal enrichment of cadmium (Cd) in soil caused by rock weathering and mining activities is an issue in southern China. Although the soil Cd content in these regions is extremely high, the bioavailability of Cd in the soils differs significantly. The carbonate area (CBA) and tin-mining area (TIA) in Hezhou City were investigated to determine the primary features of soil Cd mobility in these regions and improve environmental management. Lateral and vertical spatial distributions revealed different accumulation and migration mechanisms of soil Cd in the CBA and TIA. Further analyses revealed that mining activities and geological weathering resulted in different soil geochemical parameters, thus yielding significantly lower levels of Cd in rice grains in the CBA than in the TIA. The random forest (RF) model predicted the bioaccumulation factor (BAF) (R2 = 0.69) better than the support vector machine (SVM) model (R2 = 0.68). Subsequently, a novel land management scheme was proposed based on soil Cd and the prediction of Cd in rice to optimize the spatial resources of agricultural land and ensure the safety of rice for consumption. This study provides a novel approach for land management in Cd-contaminated areas.

A coupled geostatistical and machine learning approach to address spatial prediction of trace metals and pollution indices in sediments of the abandoned gold mining site of Bekao, Adamawa, Cameroon

Trace metals present in high amounts in aquatic systems are a perpetual concern. This study applied geostatistical and machine learning models namely Ordinary Kriging (OK), Ordinary Cokriging (OCK) and Artificial Neural Network (ANN) to assess the spatial variability of trace metals and pollution indices in surface sediments along the Lom River in an abandoned gold mining site at Bekao (Adamawa Cameroon). For this purpose, thirty-one (31) surface sediment samples are collected in order to determine the total concentrations of As, Cr, Cu, Fe, Mn, Ni, Pb, Sn and Zn. These trace metals are used to compute pollution indices as the sediment pollution index (SPI), the Nemerow index (NI), the modified contamination degree (mCD), and the potential ecological risk assessment (RI). OK, OCK and ANN models are compared to determine the best model performance. The best models are selected based on the values of the root mean square error (RMSE), the coefficient of determination (R2), the scatter index (SI) and the BIAS. Results showed that the sequence of trace metal mean concentrations in the sediments is Fe > Mn > Cu > Ni > Sn > Cr > Zn > Pb > As. The mean concentrations of Ni, Cu, Zn and Sn are above the average shale values (ASV) and the pollution status is globally moderate to significant with a low potential ecological risk. The spatial dependency obtained with semivariogram models are moderate to weak for Mn, Fe, Ni, Pb, SPI, NI, mCD, RI As, Cr, and Sn and strong for Cu and Zn. According to cross-validation parameters, ANN model is the best method for the prediction on trace metal concentrations and pollution indices in surface sediments along the Lom River in the abandoned gold mining site of Bekao.

Soil heavy metal(loid) pollution and health risk assessment of farmlands developed on two different terrains on the Tibetan Plateau, China

The quality of farmland soils on the Tibetan Plateau is important because of the region's ecological vulnerability and their close link with local food security. Investigation on the pollution status of heavy metal (loid)s (HMs) in the farmlands of Lhasa and Nyingchi on the Tibetan Plateau, China revealed that Cu, As, Cd, Tl, and Pb were apparently enriched, with the soil parent materials being the primary sources of the soil HMs. Overall, the farmlands in Lhasa had higher contents of HMs compared to those in the farmlands of Nyingchi, which could be attributed to the fact that the former were mainly developed on river terraces while the latter were mainly developed on the alluvial fans in mountainous areas. As displayed the most apparent enrichment, with the average concentrations in the vegetable field soils and grain field soils of Lhasa being 2.5 and 2.2 times higher compared to those of Nyingchi. The soils of vegetable fields were more heavily polluted than those of grain fields, probably due to the more intensive input of agrochemicals, particularly the use of commercial organic fertilizers. The overall ecological risk of the HMs in the Tibetan farmlands was low, while Cd posed medium ecological risk. Results of health risk assessment show that ingestion of the vegetable field soils could pose elevated health risk, with children facing greater risk than adults. Among all the HMs targeted, Cd had relatively high bioavailability of up to 36.2% and 24.9% in the vegetable field soils of Lhasa and Nyingchi, respectively. Cd also showed the most significant ecological and human health risk. Thus, attention should be paid to minimize further anthropogenic input of Cd to the farmland soils on the Tibetan Plateau.

Identification of soil parent materials in naturally high background areas based on machine learning

Recently, farmlands with high geological background of Cd derived from carbonate rock (CA) and black shale areas (BA) have received wide attention. However, although both CA and BA belong to high geological background areas, the mobility of soil Cd differs significantly between them. In addition to the difficulty in reaching the parent material in deep soil, it is challenging to perform land use planning in high geological background areas. This study attempts to determine the key soil geochemical parameters related to the spatial patterns of lithology and the main factors influencing the geochemical behavior of soil Cd, and ultimately uses them and machine-learning methods to identify CA and BA. In total, 10,814 and 4323 surface soil samples were collected from CA and BA, respectively. Hot spot analysis revealed that soil properties and soil Cd were significantly correlated with the underlying bedrock, except for TOC and S. Further research confirmed that the concentration and mobility of Cd in high geological background areas were mainly affected by pH and Mn. The soil parent materials were then predicted using artificial neural network (ANN), random forest (RF) and support vector machine (SVM) models. The ANN and RF models showed higher Kappa coefficients and overall accuracies than those of the SVM model, suggesting that ANNs and RF have the potential to predict soil parent materials from soil data, which might help in ensuring safe land use and coordinating activities in high geological background areas.

Spatial distribution characteristics and migration of microplastics in surface water, groundwater and sediment in karst areas: The case of Yulong River in Guilin, Southwest China

Karst regions, due to their unique topography, may be more susceptible to contaminants such as microplastics from other ecosystems. However, few studies reported the occurrence of microplastics in karst areas. Here, we investigated the abundance of microplastics in surface water, sediments and groundwater. In addition, their spatial distribution characteristics and migration were also analyzed in a typical karst area, Yulong River, Guilin, China. Microplastic pollution was found in the sediments, surface water and especially groundwater in Yulong River. The abundance of microplastics was 0-4 items/L, 247-1708 items/kg and 0-4 items/L in surface water, sediments and groundwater, respectively. Microplastics in surface water and groundwater were fiber-based, while those in sediments were fragment-based. Polypropylene (PP), Polystyrene (PS) and Polyethylene terephthalate (PET) are dominant microplastic types in Yulong River. Moreover, the abundance of microplastic pollution in different functional areas ranked as follows: living area > agricultural area > landscape area. Clustering analysis showed that disposable tableware and plastic bags used in tourism activities might be the main source of microplastics. More importantly, our findings suggested that microplastics in groundwater could be the result of hydraulic exchange between groundwater and surface water in karst areas, rather than soil infiltration. These findings provided us with a further understanding of the pollution of microplastics in karst rivers.

Probability of cultivating Se-rich maize in Se-poor farmland based on intensive field sampling and artificial neural network modelling

Selenium (Se) is a necessary micronutrient for humans, and its supplementation from crop grains is important to address the ubiquitous Se deficiency in people worldwide. Se uptake by crops largely depend on soil bioavailable Se rather than soil total Se content, which provides possibilities to explore the Se-rich crops in Se-poor area. Here, the possibility of cultivating Se-rich maize grains in Se-poor farmland was tested based on intensive field sampling and mathematical modelling. Sampling was conducted at county scale, and a total of 7779 topsoil samples and 109 maize samples with paired rhizosphere soils samples were collected. Results showed that although the soil Se content in the study county from southwestern China was at a low level (0.01-2.75 mg kg^-1), 54.1% of the maize grain samples satisfied the standard for Se-rich products (0.02-0.30 mg kg^-1). Soil organic matter, iron oxide, and phosphorus levels were correlated negatively with Se bioconcentration factor (BCF) of maize grain. Compared with the multivariate linear regression model, the artificial neural network (ANN) model was more accurate and reliable in predicting maize Se BCF. Prediction using the ANN model showed that 22.7% of the county's farmland was suitable for cultivating naturally Se-rich maize, which increased 21.3% growing areas than that from cultivation based on simply soil total Se. This study provided a new methodological framework for natural Se-rich maize production and verified the probability of cultivating naturally Se-rich maize in Se-poor farmland.

Heavy Metal Pollution and Source Contributions in Agricultural Soils Developed from Karst Landform in the Southwestern Region of China

Heavy metal pollution of soil in agricultural areas is the most prominent environmental pollution problem in China, seriously affecting human health and food security. It has become one of the environmental problems to which all sectors of society attach great importance. Soil heavy metals in the weathering area of hazardous geological bodies in southwest China have naturally high background attributes. Therefore, ecological risk assessment and analysis of potential sources of soil heavy metals in southwest China is of great significance for soil health management, soil heavy metal pollution control and territorial spatial planning. In this study, we collected 787 soil samples (0-20 cm) in Xuanwei County in China and analyzed the concentrations of As, Cd, Cr, Cu, Hg, Ni, Pb and Zn in soils. Igeo, RI, HI and CR were used to calculate the pollution levels, ecological risks and human health risks. Additionally, the PMF model and one-way ANOVA were used to identify the potential sources and discuss the factors affecting the enrichment of heavy metals. The results showed that the mean contents of the surface soils were 1.190 (Cd), 139.4 (Cr), 96.74 (Cu), 0.081 (Hg), 56.97 (Ni), 46.66 (Pb) and 130.1 (Zn) mg/kg. All heavy metals exceeded the background values of the A layer soil in Yunnan Province. The I_geo showed that Cd was the most hazardous element in the study area, followed by Cu, Cr, As, Ni and Pb. The RI showed that low ecological risks, moderate ecological risks, considerable ecological risks and strong ecological risks accounted for 3.81%, 55.27%, 37.74% and 3.18%, respectively, of the total samples, and Cd was the main dominant element. The HI values of the As element in children were greater than 1, indicating a non-carcinogenic risk, and other elements' risks were acceptable. The CR values of Cr and Ni were higher than their limits (1 × 10^-4), and both had carcinogenic risks in children and adults, as did As in children. According to the PMF model, four heavy metals sources were identified: geological sources (32%), sources from mining activities (19.38%), atmospheric deposition sources (17.57%) and agricultural sources (31.05%). Thereinto, As and Pb were mainly derived from agricultural sources, Cd and Cr were mainly associated with geological sources, Cu was largely from mining activity sources, Hg was mainly from atmospheric deposition sources and Ni and Zn were mainly from geological sources, mining activities and agricultural activities. The parent material has a significant influence on the enrichment of heavy metals in the soil, and the heavy metals are significantly enriched in the carbonate parent material and quaternary parent material. Topography also plays a role in heavy metal accumulation; Cd, Cr, Cu, Ni and Zn gradually decreased with the increase in altitude, and As and Pb increased with the increase in altitude. Mn-oxide played a crucial part in the enrichment of Cu and Zn, while SOC, K₂O and pH had little influence on the accumulation of heavy metals.