The study area is located in Wanzhou District (30°3′50″-310′18″N, 107°2′22″-108°3′52″E), in the northeastern part of Chongqing, at the heart of the Three Gorges Reservoir region. The study area has an approximate area of 3,456.41 km², with an administrative division that includes 14 sub-districts, 27 towns, and 11 townships, and a total population of roughly 1.55 million people. The district experiences a subtropical monsoon climate, with distinct seasons, an average annual temperature of 15–24°C, and an average annual precipitation of 1,000–1,200 mm, with the flood season concentrated from May to September. By 2022, the region had a total agricultural acreage of 83,908 hectares, with paddy fields accounting for 30,622 hectares (36.49%) and dry land making up 53,286 hectares (63.51%). The main soil types include purplish soil, paddy soil, and yellow soil. The research focused on fragmented cropland in the district. Based on preliminary soil monitoring data, the sampling sites covered five representative towns: Lihe, Longju, Longsha, Yanshan, and Luotian town. In 2023, 50 crop samples, including four crop types (26 grains, nine vegetables, three leaf vegetables, 12 tubers) and their rhizosphere soil samples, were collected from the study area. Figure 1 shows the distribution of sampling spots. Soil sample collection was conducted in accordance with the “Soil Environmental Monitoring Technical Specifications” (HJ/T166-2004) (41). Soil samples were collected from a depth range of 5–20 cm using the plum blossom grid method to ensure uniform sampling over an area of 1 m². Subsampling was performed using the quartering method, retaining 1 kg of soil per sample. Additionally, we also collected samples of the corresponding crops grown on the soil. After mixing, we used the quartering method for subsampling, retaining 500 g as one sample, and properly recorded the sampling process.

In the laboratory, soil samples were dried, ground, and sieved through a 100-mesh sieve. Subsequently, an HCl-HNO₃-HF-HClO₄ solution was added, and the samples were heated at 150°C for 48 h to perform digestion (42–45). Grain samples were rinsed with purified water, dried, threshed, dehulled, and ground to a 40-mesh size and then microwave-digested to determine the concentration of HMs. Vegetable, leaf vegetable, and tuber samples were cleaned with purified water, homogenized, and microwave-digested to determine the concentration of HMs. A programmed temperature ramp was applied to gradually increase the temperature to 185°C, followed by 40-min microwave digestion. The soil pH value was determined using the potentiometric method as specified in the NY/T 1377-2007 (46). The Hg content in soil was measured using the catalytic thermal decomposition-cold atomic absorption spectrophotometry method (HJ 923-2017) (47). The concentration of HMs such as Pb, Cd, Cr, As, Cu, Zn, and Ni in soil were measured using the aqua regia extraction-inductively coupled plasma mass spectrometry (ICP-MS) method (HJ 803-2016) (48). The presence of Hg in crops was assessed using direct mercury measurement (Method II of GB 5009.17-2021) (49), while Pb, Cd, Cr, As, Cu, Zn, and Ni were measured using the inductively coupled plasma mass spectrometry (ICP-MS) method (Method I of GB 5009.268-2016) (50).

Soil and crop detection indicators both adopt standard methods. Instruments used are the Inductively Coupled Plasma Mass Spectrometer NexION 2000 (PerkinElmer, USA), the Fully Automatic Direct Mercury Analyzer MAX-S (Leiberte), and the Multi–parameter Tester pH Meter (Mettler–Toledo). The main reagents used are nitric acid [Thermo Fisher Scientific (China) Co., Ltd.] and hydrochloric acid (Chengdu Kelong Chemicals Co., Ltd., chemically pure). All instruments are verified/calibrated, and all standards are within the validity period.

For quality assurance (QA) and quality control (QC), all sampling and testing personnel were uniformly trained, all equipment was calibrated and verified, all standard materials were within their valid period, and quality control measures such as laboratory blanks, parallel samples, and recovery tests were employed during testing. The results of all quality control measures were within the acceptable range.

Health risk assessment method proposed by the United States Environmental Protection Agency can be used to assess the health risks posed by HMs. Since the carcinogenic and non-carcinogenic effects caused by environmental factors are different in risk characterization, we divided them into carcinogenic risk and non-carcinogenic risk when conducting quantitative risk estimation. The hazard quotient (HQ) and cancer-risk index (CR) were adopted to evaluate the potential non-carcinogenic risks and carcinogenic risks from HMs in crops, respectively. Multiple HMs may have a synergistic effect, so the total hazard quotient (THQ) and total carcinogenic risk (TCR) of crops were determined based on the sum of the HQ and CR, respectively. These were calculated through the following equation (56–58):

where C_api was the HM content in crops (mg·kg⁻¹), IR was the ingestion rate of crops (kg·day⁻¹), EF was the exposure frequency (day·year⁻¹), ED was the exposure duration (year), BW was the body weight (kg), AT was the average exposure time for carcinogenic (non-carcinogenic) effects (day), RfD (oral reference dose) was the corresponding reference toxicity threshold dose (mg·kg⁻¹·day⁻¹), and SF was slope factors (kg·day·mg⁻¹).

HQ and THQ values < 1.0 were considered safe, but values exceeding 1.0 suggested potential adverse effects on human health. CR and TCR values above 1.0 × 10⁻⁴ indicated unacceptable risks, values below 1.0 × 10⁻⁶ implied no obvious hazard, and values between 1.0 × 10⁻⁶ and 1.0 × 10⁻⁴ were acceptable. In order to avoid the uncertainties from using single-point deterministic parameters, which may result in either overestimation or underestimation of risks, Monte Carlo simulation with 10,000 iterations was employed to evaluate the population health risk. The probability distributions and values for the mentioned parameters are detailed in Supplementary Table S1.

All spatial distribution maps were generated using ArcGIS 10.2. Statistical analyses were performed using Excel 2016 and IBM SPSS 27. RStudio and Origin 2024 were used for data visualization. Additionally, Monte Carlo simulation was performed using Crystal Ball software.

The physicochemical properties and amounts of potentially toxic elements (PTEs) in soil samples are shown in Table 1. Among the eight HMs detected, only Hg (20%), As (48%), Cd (28%), Cu (8%), Ni (4%), and Zn (70%) exceeded the background values in Chongqing, with Zn showing the most severe exceedance (59). According to the Soil Environmental Quality—Risk Control Standard for Soil Contamination of Agricultural Land in China (GB 15618-2018) (60), only 14% of the soil samples for Cd exceeded the risk screening values in the study area, indicating an enrichment of soil Cd in the region (Supplementary Table S2). This enrichment is likely associated with rapid urbanization and industrialization, which causes pollutants to enter the soil through atmospheric deposition, leading to elevated levels (61). The coefficient of variation (CV) is used to assess whether HMs are evenly distributed and can reflect the influence of anthropogenic activities (62). The CVs for Pb, Hg, As, Cr, Cd, Cu, Ni, and Zn were 16.8, 103, 29.8, 13.9, 43.6, 36.0, 16.9, and 34.3%, respectively. The CV for Hg was >1, indicating significant spatial variability in its distribution, suggesting anthropogenic activities are the primary source of this variation.

The spatial distributions of the contents of HMs (Pb, Hg, As, Cr, Cd, Cu, Ni, Zn) in the study area were obtained using the Kriging interpolation method (Figure 2). The spatial distribution of Pb and As showed a clear similarity, mainly concentrated in the northern and southwestern parts of the study area. Sampling points with high Cr and Ni concentrations were mainly distributed in parts of the northwest, southwest, and southeast, with their spatial distribution showing similar patterns. The areas with high Cu and Zn values were both concentrated in the eastern part of the study area, with their spatial distribution showing similar patterns. Areas with high Hg and Cd concentrations were relatively few, occurring only in sporadic spots. The area with a high value for Hg was located in the northwest, while the area with a high value for Cd was in the southwest.

Figure 3 and Supplementary Table S3 show the I_geo of eight HMs in 50 soil samples. The mean and standard deviation of the I_geo of Pb, Hg, As, Cr, Cd, Cu, Ni and Zn were −1.08 ± 0.25, −1.15 ± 0.91, −0.61 ± 0.45, −0.91 ± 0.24, −0.80 ± 0.59, −1.22 ± 0.45, −0.91 ± 0.28, and −0.42 ± 0.50, respectively. The max I_geo values of Pb, Cr, and Ni were below 0, suggesting that these metals did not contaminate the soil in the study area. As, Cd, Cu, and Zn showed mild contamination in 12, 6, 2, and 12% of the samples, respectively. Mild and moderate Hg contamination was found in 4 and 6%, respectively. These findings indicate that the soil in the study area was mainly contaminated by Hg, As, Cd, Cu, and Zn, with Hg having a higher contamination risk.

The characteristics of HMs contents in crops are shown in Table 2. The content of Pb in crops was within the safety limits specified by the Chinese National Food Safety Standards for Contaminants in Food (63), and the contents of Cu and Zn in crops were below the thresholds set by the Chinese Health Standards for Cu and Zn limits in food (64, 65). In crops, a small number of samples for Hg, As, and Cd showed standard exceedances, with exceedance rates of 6, 6, and 8%, respectively. However, Cr and Ni exhibited significant exceedances, with more than half of the samples exceeding the standards for these elements, which requires attention. Compared with other crops, the concentrations of As, Cu, Zn, and Hg in grains were significantly higher (P < 0.05), which may be related to differences in HM accumulation patterns among crops.

The BCF serves as a valuable indicator for assessing the ability of crops to uptake HMs from the soil, providing a quantitative evaluation of their cumulative characteristics and potential harm to crops (66). The study showed significant differences in the enrichment capacity for eight HMs among different crops (Figure 4 and Supplementary Table S4). These differences were largely attributed to the unique growth habits of each crop and their varying tolerance levels to different HMs. Additionally, the absorption of different HMs by crops may involve synergistic or antagonistic effects, further influencing the enrichment capacity of different crops for HMs. Among all crops, grains demonstrated a higher enrichment capacity for eight HMs, which may be related to their longer growth cycles. The enrichment capacity for As, Cu, Zn, and Hg in grains was significantly higher than in other crops (P < 0.05). Furthermore, this study found that the average enrichment coefficients for Cd across four types of crops were relatively high, likely due to the higher content of Cd in the local soil (61).

The average IICQ value for the soil-crop system was 2.97, with pollution levels classified as clean, slight, light, moderate, and heavy in 6, 18, 40, 26, and 10% of the samples, respectively. The mean IICQ_s value for soil HMs was 1.15, with no heavy pollution in all samples. The soil was clean, slightly polluted, moderately polluted, and heavily polluted in 54, 30, 6, and 10% of the samples, respectively. For crops, the mean IICQ_ap value was 1.82, with no samples showing heavy pollution. The majority of the samples had light pollution (52%), followed by clean (26%), moderate pollution (12%), and slight pollution (10%; Supplementary Table S5). Crops HMs contributed to approximately 61.28% of the overall IICQ for the soil-crop system, explaining crops as the main source of HM exposure. However, soil still contributed 38.72% of HM exposure. The study found differences in the average IICQ_ap values among various types of crops. The average IICQ_ap value for leaf vegetables was 0.83, indicating no pollution. The average IICQ_ap values for vegetables and tubers were 1.38 and 1.20, respectively, indicating slight pollution levels. The average IICQ_ap value for grains was 2.36, indicating light pollution levels (Supplementary Table S6). The spatial distribution of the IICQ values (Figure 5) revealed higher IICQ_s in the northwest and southwest regions, while higher crop IICQ_ap values were mainly observed in the southeast. The spatial correlation between IICQ_s and IICQ_ap was low, which may be influenced by the types of crops cultivated and their growth habits (67), causing the spatial distribution of IICQ_s and IICQ_ap to be inconsistent.

Pearson's correlation analysis results of eight HMs in the soil-crop system are shown in Figure 6 and Supplementary Table S7. HMs in soil and crops were significantly correlated at the 0.01 or 0.05 levels. The correlation analysis results showed that the pH value had a weak positive correlation with Cu in soil (P < 0.01), a moderate positive correlation with Zn in soil (P < 0.01), and a weak negative correlation with Hg in crops (P < 0.05). Pb in the soil had a moderate positive correlation with Cr and Ni in soil (P < 0.01), a strong positive correlation with As in soil (P < 0.01), and a weak positive correlation with Cd (P < 0.01), Cu (P < 0.05), and Hg (P < 0.05) in soil, and a weak negative correlation with Ni in crops (P < 0.05). Cd in the soil had a weak positive correlation with Cr and Ni in soil (P < 0.05) and a weak negative correlation with Ni in crops (P < 0.05). As in soil had a weak positive correlation with Cr (P < 0.05), Cu (P < 0.05), and Ni (P < 0.01) in soil, and a weak negative correlation with Cr (P < 0.01), Ni (P < 0.01), and Zn (P < 0.05) in crops. Cr in soil had a very strong positive correlation with Ni in soil (P < 0.01). Cu in the soil had a strong positive correlation with Zn in soil (P < 0.01). Hg in the soil had a weak negative correlation with Ni in crops (P < 0.05). Pb in crops had a weak positive correlation with Cu in crops (P < 0.05). Cd in crops had a moderate positive correlation with Cr, Cu, Zn, and Ni in crops (P < 0.01). As in crops had a moderate positive correlation with Cr and Ni in crops (P < 0.01), a weak positive correlation with Cu in crops (P < 0.01), and a strong positive correlation with Zn in crops (P < 0.01). Cr in crops had a moderate, strong, and extremely strong positive correlation with Cu (P < 0.01), Zn (P < 0.01), and Ni (P < 0.01) in crops, respectively. Cu in crops had a strong and moderate positive correlation with Zn (P < 0.01) and Ni (P < 0.01) in crops, respectively. Zn in crops had a strong positive correlation with Ni in crops (P < 0.01).

The results of the Mantel test for the eight HMs in the soil-crop system are shown in Figure 7. The right half of the graph shows the results of the Mantel test analysis, displaying the correlation between the concentrations of HMs in soil and crops. The left half of the graph shows the correlation of the concentrations of HMs in soil. The figure displays the correlations among different factors within the same matrix and between different matrices. There was a significant correlation (P < 0.01) between As in soil and eight HMs in crops, but the strength of the correlation was weak (r < 0.2). Other HMs in the soil did not show significant correlations with HMs in crops (P > 0.05, r < 0.2).

In our study, the concentration of Zn in soil samples was higher than the background values in Chongqing in many cases and the average concentration of Zn in a study on HM pollution in Chinese agricultural and urban soils over the past 20 years; however, it was lower than the concentration of Zn in the HM risk assessment of the soil-crop system along the Yangtze River in Nanjing (53, 68). Greater attention is required for the enrichment of Cd in soil. This enrichment is related to the widespread occurrence of carbonate rocks rich in Cd in the southwestern provinces of China and the impact of urbanization, which leads to the irrigation of sewage and excessive use of chemical fertilizers and pesticides. Unlike the concentrations of HM in the soil, the exceedance of Cr and Ni in crops is more severe. In this study, the content of Cr in crops was higher than that in crops in Gansu, China, and lower than that in crops near non-ferrous metal mining areas in Yunnan (69, 70). The content of Ni in crops was higher than that in grains in Anhui Province, China, and lower than that in vegetables planted in roadside soils in Pakistan (71, 72). This finding indicates that the concentration of HMs in crops is related to the type and growth characteristics of the crops and correlates with the pollution levels of HMs in soil (73–76). Further analysis revealed a significant link between soil arsenic (As) and its accumulation in graminous crops, particularly rice. This may relate to the structure of the iron plaque on rice roots. In As-contaminated soils, this plaque might enhance arsenic uptake (77). Under flooded conditions, soil As (V) is easily reduced to As (III), which has weaker adsorption to soil minerals than As (V). This leads to higher soil As concentrations and easier uptake by rice (78).

HMs in crops can be transferred to the human body through the food chain, leading to several health risks. The accumulation of various HMs in crops is a relatively complex process influenced by external environmental and anthropogenic factors. HMs enter crops through contaminated agricultural soil, sewage, atmospheric deposition, and fertilizers containing HMs (79, 80). Through the absorption mechanism of crop root systems, HMs are transferred to various parts of the plant, such as tubers, leaves, fruits, and seeds, by specific transport proteins (81). Long-term consumption of crops contaminated with HMs poses health risks to human health, including kidney damage, osteoporosis, neurological damage, cardiovascular diseases, and increased cancer risk (82–86). The unique physiological characteristics of children predispose them to the toxic effects of HMs. Long-term intake of crops contaminated with HMs can have a negative impact on their cognitive development and physical growth (87, 88). In this study, the average THQ for adults and children were 2.365 and 1.176, respectively, which were greater than the standard value of 1.0, suggesting that consuming crops is associated with a higher non-carcinogenic risk. The average THQ in this study was much higher than the values of 0.0441 for adults and 0.324 for children near the coal mines in Jinzhong City, Shanxi Province (89). Compared to the THQ values in this study, a study on HM contamination in crops in Iran reported a lower value for adults (1.67) but a slightly higher value for children (1.28) (90). Only As had an average HQ value exceeding the standard value of 1.0 in 59.90% of samples for adults, while As had an average HQ value close to the standard value of 1.0 in 29.88% of samples for children. This finding indicates that As poses a significant health risk to adults, but the health risk to children should not be ignored. Cd (adults: 0.5%, children: 0.05%), Cr (adults: 0.34%, children: 0.06%), Zn (children: 0.05%) had average HQ values below the standard value of 1.0, suggesting that Cd, Cr, and Zn may not pose a health risk to the population at most exposure levels, but the cumulative effect still warrants attention (91, 92).

The average TCR in this study was 2.28 × 10⁻³ for adults and 1.11 × 10⁻³ for children, both significantly higher than the unacceptable level of 1.0 × 10⁻⁴, indicating that HM pollution in crops is prone to carcinogenic risk. The TCR values for both adults and children were lower than the values of 2.92 × 10⁻³ for adults and 4.96 × 10⁻³ for children near the Leng-shuikeng mine in Jiangxi (93). However, the TCR for children in this study was higher than that reported in agricultural soil in Lanzhou (94). For individual HMs, As (adults: 72.61%, children: 45.47%), Cr (adults: 95.72%, children: 85.22%), Cd (adults: 83.15%, children: 61.83%), and Ni (adults: 96.79%, children: 87.20%) had average CR values greater than the standard value of 1.0 × 10⁻⁴, indicating that these metals pose a higher carcinogenic risk to the population and require attention (25, 95–97).

Several studies show that adults are at greater health risk than children due to higher crop consumption and longer exposure time (98–100). However, other studies demonstrate that children are at greater health risk than adults due to their lower body weight and poorer tolerance to toxic HMs (101–103). HMs affect public health through various routes, such as crops, drinking water, soil, and air pollution (104). Therefore, efforts to identify and control the key exposure factors in crops will reduce the population health risks.

Health risk assessments for HM hazards based on Monte Carlo simulation involved many uncertainties. Differences in regions, ethnicities, and ages often lead to variations in model parameters (e.g., exposure frequency and body weight), which can cause deviations in actual health risk levels (105–107). Identifying exposure parameters that significantly impact health risk assessments is essential for reducing uncertainties in the risk assessment process. Reporting the influence and sensitivity of the assessment results can lead to a more accurate evaluation of the impact of heavy metals on population health risks. This study focused on the differences in health risks between adults and children and selected parameters related to human exposure for Monte Carlo simulation and sensitivity analysis. The findings showed no significant differences between the parameters. For both non-carcinogenic and carcinogenic risk assessments, crop intake was found to be the most sensitive parameter. This finding is consistent with previous studies on rice HM pollution in the Minxi region and vegetable HM pollution in Chongqing (56, 61). Sensitivity analysis can help identify key exposure parameters, and when combined with Monte Carlo simulation, it can effectively assess the health risks of HM pollution in different population groups.