Guizhou is a typical karst area located on the Yunnan–Guizhou Plateau and has a subtropical humid climate with annual average temperatures of approximately 12°C–25°C; the January temperatures range from −2°C to 12°C and July temperatures range from 20°C to 29°C. Changshun County (106°12′32″–106°38′48″E, 25°39′28″–26°17′17″N) is a typical karst region in Guizhou Province in southwestern China and covers an area of 1554.6 km². The study area belongs to a humid zone in the mid-subtropical monsoon climate region and has obvious three-dimensional climatic characteristics. The mean annual temperature ranges from 13.4°C to 19.5°C and mean annual precipitation ranges from 1250 mm to 1400 mm, with four clearly evident seasons. The vegetation in this area includes broad-leaved mixed forests, evergreen broad-leaved forests, and montane elfin forests. Sufficient availability of water in this region promotes conditions for karst erosion and general erosion, and the desert area accounts for 77.9% of the total area. No obvious point pollution sources are noted in the study area.

From March to June of 2022, eighteen soil profiles consisting of 120 soil samples were collected from natural forestlands (NFLs), natural brush lands (NBLs), natural pasture lands (NPLs), artificial forestlands (AFLs), artificial brush lands (ABLs), and artificial pasture lands (APLs) in Changshun County (Figure 1). These sampling areas have the same soil formation processes and similar geochemical backgrounds. All artificial areas (AFLs, ABLs, and APLs) were transferred from farmlands that were used for food production to the currently used types over a decade ago. The detailed information for each sampling site is listed in Table 1. Each soil profile was sampled at seven layers (0–10, 10–20, 20–30, 30–40, 40–50, 50–70, and 70–90 cm) if the soil depth was equal to or greater than 90 cm. Otherwise, sampling was carried out to the actual depth of the soil. For example, four samplings were performed if the soil depth was between 40 cm and 50 cm. The soil samples were transferred to the lab and dried at room temperature (ca. 28°C) before being ground and passed through a sieve (0.22 mm). The collected samples were stored in Ziploc bags for the analysis of heavy metals.

For analysis of Pb, Cd, and Cu, about 1 g (accurate to 0.0001 g) of each soil sample was weighed and digested with 30 mL of a digesting mixture composed of HNO₃, HClO₄, and HCl (HNO₃:HClO₄:HCl = 4:1:5 v/v/v). The Pb, Cd, and Cu contents were then determined with inductively coupled plasma atomic emission spectrometry (5300V, Perkin Elmer Corporation, Waltham, MA, USA) (Wang et al., 2020; Wang et al., 2021). For determination of As and Hg, about 0.3 g (accurate to 0.0001 g) of the soil sample was weighed and transferred to a 50-mL tube, followed by addition of 10 mL of a digesting solution composed of HCl, HNO₃, and H₂O (HCl:HNO₃:H₂O = 3:1:4 v/v/v) (Tokumoto et al., 2014). The tubes were then placed in a boiling water bath for 3 h and shaken at intervals of 30 min. The solutions were next transferred to 50-mL volumetric flasks and 5 mL of a reducing agent (10% thiourea solution) was added to each flask, followed by dilution to volume with 5% HCl (v/v). The As and Hg levels were determined using atomic fluorescence spectroscopy (AFS-933, Jitian Corporation, Shanghai, China). The quality of the analytical procedure was controlled using the certified reference material GBW 07403, which was analyzed in triplicate while the soil samples were replicated at a 20% level. The uncertainty of the analytical procedure was thus found to be within 10%. Analytical blanks were provided for all determinations (Huang et al., 2018; Huang et al., 2017). In addition, the other soil chemical indicators, such as total phosphorus (TP), total nitrogen (TN), and total potassium (TK), were tested simultaneously (Wang et al., 2022; Huang et al., 2018).

The heavy metal pollution at each study site was evaluated by gray clustering analysis, and the weight of each heavy metal in the clustering was determined. The basic theory of this analytical method can be found in literature (Pan et al., 2009; Fan et al., 2016). In the present study, five pollution indices (Pb, Cd, Hg, As, and Cu) were selected as the clustering indices to evaluate the pollution status at each studied site. The evaluation criteria were based on those recommended by the Ministry of Ecology and Environment of the People’s Republic of China (GB15618-2018) (Wang et al., 2021; Fan et al., 2016; Fan et al., 2012; Zhang et al., 2019) (Table 2).

Statistical analyses were conducted using Microsoft Excel 2003, SPSS 13.0 for Windows, and Origin 6.1 software.

The levels of heavy metals observed in the soil profiles from each land use are presented in Figure 2 and Table 3. The profile characteristics of these heavy metals clearly vary significantly among different land uses. As shown in Figure 2, the concentrations of Pb in the soil samples ranged from 2.99 to 84.44 mg kg^-1, with a mean value of 38.73 mg kg^-1. The concentrations of Cd in the soil samples ranged from 0.03 to 10.75 mg kg^-1, with a mean value of 1.33 mg kg^-1. The concentrations of Hg in the soil samples for different studied land uses ranged from ND (not detected given a detection limit of 0.005 mg kg^-1) to 0.71 mg kg^-1, with a mean value of 0.35 mg kg^-1. The As concentrations in the soil samples from the NFLs showed a similar distribution as that of Pb for this type of land use; the mean concentration decreased from 31.42 mg kg^-1 in the top 10 cm of the surface soil to 24.07 mg kg^-1 in soil samples at 20–30 cm depth after a small increase (33.44 mg kg^-1 in soil samples at 10–20 cm depth) and then increased gradually to 45.03 mg kg^-1. In the soil samples from the NBLs, the As concentrations increased generally with depth, except for a weak decrease at a soil depth of 50–70 cm. For the NFLs and NPLs, the mean concentrations of As fluctuated irregularly, similar to those of Pb, with increase in soil depth. For the ABLs, AFLs, APLs, and NPLs, no obvious patterns of variation were found with changes in the soil depth. The concentrations of Cu in the soil samples ranged from 7.90 to 29.60 mg kg^-1, with a mean value of 18.94 mg kg^-1.

Overall, there were no regularity or significant distributions of all heavy metals from the surface to subsoil layers, and the distribution trends of each of the heavy elements in the same soil layer varied for different land uses (Table 3). The average values of Pb for different land uses decreased in the order of ABL > AFL > NBL > APL > NPL > NFL, among which the values for ABL and AFL were significant higher than the others. The distribution characteristics of Cd were in the order of ABL > AFL > NBL > NPL > NFL > APL, where the values in ABL and AFL were distinctly higher than in other land uses. There were minimal differences in the Cd content in NFL, NBL, NPL, and APL in the soil layers of depths 0–50 cm. The distribution of Hg decreased in the order of NPL > APL > ABL > AFL > NBL > NFL, and the discrepancies in the Hg values in each of the soil layers among the different land uses were large. The distribution of As decreased in the order of AFL > ABL > NFL > NPL > NBL > APL, among which the values in AFL and ABL were significantly higher than those in the NPL, NBL, and APL. The average values of Cu for different land use cased decreased in the order of ABL > APL > AFL > NFL > NBL > NPL, with only slight differences in soil layers at depths of 0–30 cm. However, there were large differences in Cu values in soil layers at depths of 30–90 cm. In summary, each of the heavy metals were present at higher concentrations in the artificial land use cases than under natural land use; this implies that human interference maybe an important factor in heavy metal pollution.

According to the Environmental Quality of Soil Standards of China, the limits for the studied heavy metals are listed in Table 2 (Fan et al., 2016). There are three limiting values for each element based on soil acidity. In the present study, all soil samples had pH values less than 7.5. Therefore, two groups of the class II limit values were applied.

In the soil samples from the NFLs, all concentrations of Pb and Hg met the limits under class I. About 36% of the Cd concentrations met the limits under class I, 31.58% met the limits under class II, and the rest met the limits under class III. All of the Hg concentrations in the 0–40 cm depth range were very low and met the limits under class I. However, the Hg concentrations at depths greater than 40 cm exceeded 0.15 mg kg^-1 and sometimes exceeded 0.30 mg kg^-1. All of the As concentrations exceeded the limits under class I but were within the limits of class II, except the soil samples collected at depths of 40–90 cm, whose concentrations exceeded 25 mg kg^-1. All concentrations of Cu met the limits under class I.

In the soil samples from NBLs, about 36.84% of the Pb concentrations met the limits under class I, and the rest were within the limits of class II. In both NPLs and APLs, most of the Pb concentrations were below 35 mg kg^-1. It is worth noting that all concentrations of Hg exceeding 0.30 mg kg^-1 (with some sample concentrations greater than 0.5 mg kg^-1) were beyond the limits of class II. In the AFLs, all of the Pb concentrations exceeded the limits under class I, and all of the As concentrations exceeded the limits under class II. Most of the Hg concentrations exceed the limits under class II. In the soils from both AFLs and ABLs, all Cd concentrations exceeded 1.00 mg kg^-1, which was beyond the limits of class III.

According to the Equations 1–6, the heavy metals clustering weights under the different land use types was calculated. This results shown in Figure 3, Pb and Cu contributed little to the heavy metal concentrations under all kinds of land use based on the clustering weights. In the NBLs, AFLs, and ABLs, the clustering weights of Cd were greater than 50% of the total. In the NFLs and NPLs, the clustering weights of Cd were still greater than those of the other heavy metals. However, the clustering weights of Hg were largest among all the heavy metals studied in the APLs. Generally, the clustering weights of Hg in the pasture lands were greater than those in the brush lands, followed by forestlands. It was also found that the clustering weights of As accounted for a considerable portion of the heavy metal concentrations under all types of land use in this study and especially in the NFLs. The pollution degree of each sampling site is listed in Table 4. The soils from the NFLs, NPLs, and APLs belonged to class II, while those from NPL-3 belonged to class I. The soils from the NBLs, AFLs, and ABLs belonged to class III.

The environmental fates of pollutants are determined by many factors. However, the physicochemical properties, geochemical baselines, anthropic activities, and microcircumstances are critical for determining the pollution levels and environmental behaviors of heavy metals (Teng et al., 2022). Based on field investigations and heavy metal distribution characteristics under different land uses, we hypothesized that the heavy metals in soils under different land uses would have different pollution sources. The baseline values (soil formation), application of fertilizers, and atmospheric deposition were considered as the critical sources in study area since there were no local mining or refining industries that would cause significant heavy metal pollution. Based on the Pearson correlation analysis, there were significant correlations between Pb and TK, Cd and TP, as well as As and TN; however, there were no significant correlations among Cu, Hg, and the soil nutrient elements (Table 5).

The Cu concentrations were very low in all soil samples; however, in the deeper layers, Cu concentrations in soils from APLs and ABLs were obviously higher than those from the other samples under study. In addition, the background value of Cu in the karst mountain area has been reported to be high (Wang et al., 2021). This finding indicates that Cu pollution is mainly from soil formation. The ABLs and AFLs were developed for growing different types of fruits; to promote good production, the land owners were reported to have applied different fertilizers that could cause heavy metal pollution. Previous studies have suggested that Cd, Pb, and As concentrations are higher in cultivated soils owing to fertilizer application (Atafar et al., 2010; Wang et al., 2012); these results indicate that fertilizer application is an important source of heavy metal pollution (Cd, Pb, and As) in ABLs and AFLs. According to the information listed in Tables 2, 3 and the GB15618-2018 standards, there were some samples in which the Hg contents were higher than the limiting value for paddy lands but lower than those for other land types. The production of Hg in Guizhou is known globally. Hg pollution in Guizhou Province has been a severe problem for a long time (Feng et al., 2008; Streets et al., 2005). It is well known that karst landforms have unique water ecological systems with dual above ground and underground structures; moreover, the underground water system is highly developed. Xu et al. (2020) noted that there was a significant positive correlation between Cd and Hg levels in underground water in karst mountain areas. In addition, the Cd content is relatively stable in the underground water in karst areas related to mining areas (Liao et al., 2023). Although there were no Hg mining or refining industries in the studied region, the groundwater could also be an important factor in the presence of heavy metals. Guo et al. (2024) showed that atmospheric deposition contributes to accumulation of heavy metals in soil. These results indicate that Hg pollution in this area may be caused by the underground water and atmospheric deposition (Feng et al., 2008).

Land use change is a significant factor affecting the sources and environmental behaviors of heavy metals. Atafar et al. (2010) reported that fertilizer application resulted in significantly higher concentrations of Cd, Pb, and As in cultivated soils (p-value <0.05). The clustering analysis results indicate that heavy metal pollution in the AFLs, ABLs, and NBLs are significantly higher than in other types of land use.

The Cd concentrations in the soil samples from the AFLs and ABLs were obviously higher than those from other types of land use in this study. This suggests that fertilizer application increased the concentrations of Cd, Pb, Cu, and As in the cultivated soils. The concentrations of Hg in the soil samples from NPLs were generally higher than those from the other types of land use. As these lands share the same soil formation processes and similar geochemical baselines, we believe that pollution caused by fertilizer application was the main reason for the increased Hg concentrations in soils from ABLs, which were higher than those from NBLs.

The results of this study show that there is a strong correlation between the soil Hg concentration and application of manure or compost. In forestlands, large tree canopies reduce Hg pollution caused by atmospheric deposition, which may be the main reason why Hg concentrations in the soil samples from NFLs were much lower than those from NBLs, APLs, and NPLs. Furthermore, Hg concentrations in soil samples from NBLs were slightly higher than those from APLs. We inferred that these were possibly the effects of management activities. The local farmers spread grass seeds to increase the grass quantity, and high levels of grazing also promote grass growth rates. Therefore, Hg concentrations in soils from APLs were generally lower than those in NPLs. This may be due to the fact that artificial grass planting has a certain alleviating effect on Hg pollution (Pacyna and Pacyna, 2001).