Bioavailability of Cd in Agricultural Soils Evaluated by DGT Measurements and the DIFS Model in Relation to Uptake by Rice and Tea Plants

Abstract

The elevated accumulation of cadmium (Cd) in rice (Oryza sativa L.) and tea (Camellia sinensis L.) grown in agricultural soils may lead to a variety of adverse health effects. This study collected and analyzed crop samples along with paired rhizosphere soil samples from 61 sites in Cd-contaminated regions in Anhui Province, China. The findings revealed that both the diffusive gradients in thin-films (DGT) and soil solution were capable of effectively predicting Cd contents in crops. Conventional chemical extraction methods were inappropriate to evaluate the bioavailability of Cd. However, the effective concentrations (C_E) corrected by the DGT-induced fluxes in soils (DIFS) model exhibited the strongest correlation with crop Cd contents. Except for C_E, various measurement methods yielded better results for predicting Cd bioavailability in tea compared to rice. Pearson’s correlation analysis and the random forest (RF) model identified the key influencing factors controlling Cd uptake by rice and tea, including pH, soil texture, and contents of zinc (Zn) and selenium (Se) in soils, which antagonize Cd. To reduce the potential health risk from rice and tea, the application of soil liming and/or Se-oxidizing bacteria was expected to be an effective management strategy.

1. Introduction

Cadmium (Cd), which has been identified as the primary contaminant in Chinese agricultural land, poses a substantial threat to both food safety and public health due to its widespread presence in China’s agricultural ecosystems [1]. Elevated Cd accumulation in Chinese soils is the result of a complex interplay between natural geological factors and human activities, including industrialization and agricultural practices. While geological bedrock and mineral deposits can naturally contribute to Cd content in certain areas, anthropogenic activities, especially industrial pollution and agricultural practices, have played a significant role in increasing Cd levels in many regions [2]. In comparison to other toxic elements, Cd exhibits a heightened capacity for absorption by edible plants from soils characterized by high transfer rates. Given its status as a noxious and carcinogenic metal for humans, exposure to Cd can induce kidney damage and various other detrimental impacts, and long-term exposure to Cd has been associated with an increased risk of certain cancers, particularly lung cancer [3]. Some studies have suggested that adequate selenium (Se) intake may mitigate the adverse health effects of Cd exposure. Se can also enhance the activity of certain enzymes that are involved in detoxifying and eliminating heavy metals, including Cd, from the body. This is particularly relevant in regions where there is a risk of Cd contamination in the food supply or environment [4].

Rice (Oryza sativa L.), which serves as a pivotal cereal crop in eastern China, holds a propensity to accumulate more Cd than other cereals, establishing itself as a principal contributor to human Cd exposure through dietary intake. Correspondingly, tea (Camellia sinensis L.), a widely consumed beverage worldwide, particularly in Asia, is susceptible to Cd contamination. In some Asian countries like India, China, and Japan, Cd levels in tea infusions may surpass permissible daily limits, thereby giving rise to cumulative chronic and acute health complications through oral ingestion [5,6]. Moreover, the health risks stemming from Cd exposure via crops hinge not only upon its concentrations within edible portions but also on consumption rates [7]. Consequently, an accurate assessment of Cd bioavailability in rice and tea cultivated in Anhui, a typical province in China where the local populace predominantly relies on rice for caloric intake and tea as a favored hot beverage, becomes an exigent necessity.

Soil characteristics such as pH, organic matter content, and soil texture can exert influence over the bioavailability of Cd [8,9,10]. For example, in acidic conditions, Cd ions are less likely to bind to soil particles, making them more mobile and potentially more harmful to plants. The dissolution of Cd-containing minerals can also increase under acidic conditions; for example, Cd sorption on kaolinite reduced significantly at low soil pHs (pH < 6) since Cd is sorbed as outer-sphere complexes. Organic matter can form complexes with Cd ions, reducing their mobility and uptake by plants. This is due to the formation of stable complexes between Cd and organic ligands. Further, Cd tends to have higher retention and lower availability in clayey soils due to the higher cation exchange capacity (CEC) and surface area for adsorption. Cd ions can bind tightly to clay particles, reducing their movement in the soil [11,12]. However, the specific mechanisms by which these influencing factors affect the bioavailability of Cd remain insufficiently elucidated. The classification and regression tree (CART) methodology stands as a widely employed machine learning approach known for its adeptness in unveiling relationships. In tandem, the random forest (RF) technique, an extension of CART, represents a relatively novel method aimed at enhancing prediction accuracy [13,14,15].

Furthermore, it is widely acknowledged that the uptake of Cd by plants is contingent not solely upon Cd speciation within soils but also upon the dynamic interplay of Cd resupply from the solid phase to the soil solution [16]. To gauge labile Cd concentrations, encompassing both dissolved fractions within the soil solution and a portion of resupplied fractions from the solid phase, the diffusive gradients in thin films (DGT) technique is employed. Previous investigations have attested to the efficacy of the DGT technique in evaluating Cd bioavailability in soils [17,18]. Facilitating quantitative descriptions of chemical exchange between the solid phase and solution under DGT-induced perturbation, the DGT-induced fluxes in soils (DIFS) model provides a dynamic framework for the DGT-medium system. Additionally, this model permits the calculation of diffusion coefficients to rectify DGT measurements [19,20].

The evaluation of Cd bioavailability in rice and tea, along with its governing factors within the soil properties, is imperative within this study area. This study encompasses the pursuit of several objectives: (i) to appraise the predictive efficacy of Cd accumulation in rice grains and tea leaves, employing diverse measurement approaches encompassing conventional chemical extraction, the DGT technique, soil solution analysis, and the DIFS model; (ii) to discern the factors that wield influence over Cd uptake by rice and tea; (iii) to offer recommendations that can steer secure local cultivation management.

2. Materials and Methods

2.1. Study Area

The research area is located in Shitai County, Anhui Province, China. The area is about 24 km², with elevation between 50 to 1000 m, and the terrain is generally mountains in the northwest and valley plains in the middle. Limestone, shale, and carbonaceous siliceous rock are the dominant lithology across the study area. The annual precipitation is approximately 1100 mm, and the daily temperature ranges between 5 and 25 °C [10]. The main types of soil-forming parent materials are carbonate rocks, clastic rocks, and alluvial deposits. The major soil types have been identified as Cambisols, Acrisols, and Anthrosols [21]. The soils in Shitai County have been identified as seleniferous and Cd-contaminated [10,22]. Rice and tea are the most extensively grown crops in the region.

2.2. Sample Collection and Pretreatment

In this investigation, a total of 61 agricultural sites were surveyed. We collected 43 samples of rice grains (Indica cultivar) and 18 samples of tea leaves, with principal growth stage of ripening, and their corresponding rhizosphere soil samples were extracted from depths of 0 to 20 cm (Figure 1). The soil samples were collected from 3 to 5 sub-sample points per site and subsequently amalgamated to form a composite representative of each location. Employing a meticulous approach, all collected samples were hermetically sealed within airtight plastic pouches to maintain their integrity during transportation to the laboratory within a time of 48 h. After collection, the soil samples were air-dried at ambient temperature and subsequently sieved through a 10-mesh sieve to eliminate any extraneous matter or pebbles, after which they were finely ground into a powder of uniform consistency. The rice grains, divested of their husks, and the tea leaves were subjected to an oven-drying process at 50 °C for a duration of three days to ensure a consistent weight, following which they too were ground into a finely textured powder.

2.3. Chemical Analysis

Soil pH was measured using the ion selectivity electrode (ISE) method. Soil particle size composition was determined using Malvern Mastersizer 2000, with sand, silt, and clay fractions defined as particles >20 μm, 2~20 μm, and <2 μm, respectively [23]. The oxidation with dichromate procedure is used for the determination of soil organic carbon (SOC). The measurement of soil cation exchange capacities (CEC) was carried out through the ammonium acetate exchange method, as detailed in reference [24]. The determination of total Cd concentrations in soil, rice grains, and tea leaves involved the use of 1.00 g of powdered samples for each matrix. Soil samples underwent digestion with HF, HNO₃, and HClO₄, while the crop samples were subjected to digestion with a concentrated mixture of HNO₃ and HClO₄. Inductively coupled plasma mass spectrometry (ICP-MS) was employed to quantify Cd, Zn, and Se concentrations in both soil and crop digestion solutions. Quality assurance and quality control (QA/QC) measures were applied to ensure accuracy and precision, incorporating blank samples, duplicates, and standard reference materials.

The sequential extraction procedure, adapted from the methodologies of Tessier et al. [25] and Quezada-Hinojosa et al. [26], facilitated the fractionation of Cd into seven distinct forms: F1 water-soluble, F2 exchangeable (extracted for 1 h using 1 mol·L⁻¹ MgCl₂), F3 carbonate-bound, F4 humic acid-bound, F5 amorphous Fe/Mn oxide-bound, F6 crystalline Fe oxide-bound, and F7 residue [27]. The filtrates obtained at each extraction step were subjected to ICP-MS analysis. The reliability of the sequential extractions was verified by comparing the cumulative concentrations of fractions F1 to F7 with the total Cd concentrations.

2.4. DGT Measurements and the DIFS Model

In this investigation, we employed Chelex-100 resin DGT devices procured from Weishen DGT^® Research Ltd. (Nanjing, China). Our procedures adhered to established standards for the deployment and measurement of DGT devices in soil, as outlined in reference [28]. A total of sixty-one rhizosphere soil samples were subjected to measurement via the DGT technique. Preceding device deployment, 40 g of air-dried soil was introduced into a beaker, set at 60% of its maximum water holding capacity (MWHC), for a span of 48 h, followed by a subsequent phase at 80% MWHC for 24 h. After 24 h of deployment at a temperature of 25 °C, the DGT devices were carefully retrieved from the soil, subjected to a thorough rinse with deionized water, and then disassembled. Subsequently, the binding gels were subjected to elution with 1 mL of 1 M HNO₃, allowing for a 24-h interval prior to the commencement of analysis. The eluates were diluted by a factor of ten utilizing HNO₃ before Cd analysis was conducted using ICP-MS. The computed concentration of Cd (C_DGT) through the DGT approach was determined using Equation (1), with a comprehensive elucidation of the calculation provided in reference [20]:

$$C_{\mathrm{D}\mathrm{G}\mathrm{T}}=\frac{M\times \Delta g}{D\times A\times t}$$

(1)

In this equation, the symbol M represents the accumulated mass of Cd (in ng) during the deployment interval, while Δg denotes the thickness of the diffusive layer (in cm). The diffusion coefficient of Cd within the resin is represented by D (in cm²·s⁻¹), A signifies the area of the exposure window (in cm²), and t corresponds to the deployment time (in seconds). Following the retrieval of the DGT devices, soil paste was procured, transferred into 50 mL polyethylene tubes, and then subjected to centrifugation at 5000 rpm for a duration of 20 min, aimed at extracting the soil solution. After centrifugation, the supernatant was subjected to filtration through a 0.45 μm filter, with the resultant filtrate utilized for the quantification of the soil solution’s Cd content (C_solu). To assess the resupply originating from the solid phase, the ratio (R) between C_DGT and C_solu was computed [29]. To comprehend the kinetics of Cd resupply from the soil’s solid phase to the porewater within the rhizosphere zone or to ascertain the effective concentration (C_E) of Cd, the DGT-induced fluxes in soils (DIFS) model for the DGT-soil system was employed. The term C_E pertains to the theoretical porewater concentration necessary for the accumulation of the measured Cd quantity on the DGT resin solely via diffusional supply [18]. Equation (2) is employed to convert C_DGT to C_E:

$$C_{\mathrm{E}}=\frac{C_{\mathrm{D}\mathrm{G}\mathrm{T}}}{R_{\mathrm{d}\mathrm{i}\mathrm{f}\mathrm{f}}}$$

(2)

Here, R_diff denotes the C_DGT to C_solu ratio when supply to the DGT device solely occurs via diffusion. The DIFS model permits the quantification of the interdependency between R_diff and the rate and extent of Cd resupply from the solid phase to the solution, encompassing both the soil-to-interface transition and diffusion through the diffusion layer to the binding layer’s sink [20]. The software 2D DIFS (version 1.2.3-3, 2005, Lancaster, UK) [30], available online, was utilized for the calculation of R_diff values pertinent to the Cd in the DGT-soil system.

2.5. Statistical Analysis

Descriptive statistics and Pearson’s correlation analysis were conducted employing IBM SPSS Statistics version 27 (IBM Corp, Armonk, NY, USA). To ascertain and quantify the significance of potential factors influencing the bioavailability of Cd in agricultural soils, the random forest (RF) technique was employed. Implementation of the RF model was carried out in R version 4.2.3 (R Development Core Team, Vienna, Austria), utilizing the random forest package [31]. For graphical representation, box plots and linear fit analyses were generated using Origin 2021 (Origin Lab Corporation, Northampton, MA, USA).

3. Results and Discussion

3.1. Concentrations of Cadmium in Soils, Rice Grains and Tea Leaves

The box plots illustrating the total Cd content in the plow layer soil, rice grains, and tea leaves within the study area are presented in Figure 2. In these box plots, the internal horizontal line represents the median value of Cd, while the cross symbol denotes the mean value. The top and bottom edges of the box correspond to the 25th and 75th percentiles of Cd, and the lines extending from the top and bottom of the box represent the 5th and 95th percentiles of Cd, respectively. In the Cd-contaminated region of Anhui, the variation in Cd total content in rhizosphere soils ranged from 0.117 to 1.029 mg·kg⁻¹, with an average of 0.564 mg·kg⁻¹ in rice grains. The average Cd content was 0.191 mg·kg⁻¹ in rice seeds and 0.038 mg·kg⁻¹ in tea leaves within the study area (Figure 2b).

We employed the modified Tessier sequential extraction method to investigate the various binding forms of Cd in 61 rhizosphere soil samples [25,26], yielding a total of seven forms (F1~F7). Among these forms, the exchangeable Cd fraction exhibited the highest content, constituting 42.4% of the total, indicating notable Cd mobility (Figure 2a). The amorphous Fe/Mn oxide-bound fraction also exhibited considerable content at 21.9% of the total, signifying the influence of Fe/Mn oxides on Cd distribution. The residual form appeared as the most stable, accounting for 16.0% of the total Cd content. The carbonate-bound and humic acid-bound fractions demonstrated similar contents, representing 7.19% and 7.37% of the total Cd content, respectively. However, the carbonate-bound fraction displayed greater variability among different soil samples. The crystalline Fe oxide-bound and water-soluble fractions exhibited the lowest contents at 3.90% and 1.21% of the total Cd content, respectively. In summary, Cd exhibited a substantial proportion of mobile forms. Previous research suggests that Cd in soil is more prone to migration compared to other heavy metals, thereby necessitating heightened attention to its ecological risk [2].

The bioconcentration factor (BCF) is defined as the ratio of the concentration of heavy metals, such as Cd, in crops to the concentration of these elements in the topsoil where the crops are grown. It is commonly used to indicate the capacity of crops to absorb and accumulate heavy metals. The formula for calculating the BCF is provided in Equation (3) [32]:

$$\mathit{BCF} =\frac{C_{\mathrm{crop}}}{C_{\mathrm{soil}}}$$

(3)

where BCF represents the value of the bioconcentration factor, C_crop is the concentration of the heavy metal in the crops, and C_soil is the corresponding concentration of the heavy metal in the rhizosphere soil. In this study, BCF_rice and BCF_tea represent the BCF for Cd in rice and tea, respectively. The transfer of harmful elements like Cd from the soil to crops is a vital link in the food chain for human exposure to heavy metal pollutants. Investigating the bioaccumulation characteristics of Cd is essential for evaluating the safety risks associated with agricultural products and managing safe crop cultivation. Table 1 presents the Cd content in soil and crops, BCF values, and exceedance rates for the study area as well as other Cd-contaminated regions (Jiangsu and Guizhou) and non-contaminated regions (Henan and Yunnan).

Recent studies have shown that in Cd-contaminated regions in Jiangsu, the average BCF_rice is approximately 0.155 [8]. In comparison, in the agricultural fields of Henan, the average BCF_rice is around 0.145, which is at a similar level [33]. In the study area, the average BCF_rice is higher than those in both the contaminated and non-contaminated regions. This indicates a greater Cd activity in the Cd-contaminated areas of Anhui. As discussed earlier, the sum of the water-soluble, exchangeable, and carbonate-bound forms (F1+F2+F3) exceeds 50% of the total Cd content in the soil, highlighting its potential biological availability [34,35]. However, according to the national standards [36,37], the exceedance rate of rice (32.5%) is at a comparable level to that of Jiangsu (30.6%), whereas it is more than 10 times higher than the non-contaminated area (2.5%). In contrast to rice, the ability of tea leaves to accumulate Cd shows notable differences between contaminated and non-contaminated areas. In Guizhou, the average BCF_tea (0.303) is ten times higher than that in Yunnan (0.038), while in this study, BCF_tea falls between these two values [38,39]. The underlying cause of this divergence in rice and tea lies potentially in the significantly lower pH of the rhizosphere soil for tea plants compared to rice, resulting in a greater release of labile forms for Cd into the soil solution, thereby facilitating their uptake by crops.

Table 1. The median contents of total soil Cd, crop Cd, BCF, and exceedance rate according to national thresholds.

Region	Anhui	Anhui	Jiangsu	Guizhou	Henan	Yunnan
Crop	rice	tea	rice	tea	rice	tea
n	43	18	70	22	40	8
pH	6.33	5.39	6.10	4.30	5.90	4.48
Total soil Cd	0.737	0.417	0.970	0.330	0.124	0.079
Crop Cd	0.191	0.038	0.150	0.100	0.018	0.003
BCF	0.259	0.091	0.155	0.303	0.145	0.038
Thresholds	0.2 a	1.0 b	0.200 a	1 b	0.200 a	1 b
Exceedance rate (%)	32.5	0.0	30.6	0.0	2.5	0.0
Reference	this study	this study	[8]	[38]	[33]	[39]

^a National Standards for Food Safety and Limits of Contaminants for Food [36]. ^b Residue Limits for Chromium, Cadmium, Mercury, Arsenic, and Fluoride in Tea [37].

Table 1. The median contents of total soil Cd, crop Cd, BCF, and exceedance rate according to national thresholds.

Region	Anhui	Anhui	Jiangsu	Guizhou	Henan	Yunnan
Crop	rice	tea	rice	tea	rice	tea
n	43	18	70	22	40	8
pH	6.33	5.39	6.10	4.30	5.90	4.48
Total soil Cd	0.737	0.417	0.970	0.330	0.124	0.079
Crop Cd	0.191	0.038	0.150	0.100	0.018	0.003
BCF	0.259	0.091	0.155	0.303	0.145	0.038
Thresholds	0.2 a	1.0 b	0.200 a	1 b	0.200 a	1 b
Exceedance rate (%)	32.5	0.0	30.6	0.0	2.5	0.0
Reference	this study	this study	[8]	[38]	[33]	[39]

^a National Standards for Food Safety and Limits of Contaminants for Food [36]. ^b Residue Limits for Chromium, Cadmium, Mercury, Arsenic, and Fluoride in Tea [37].

3.2. Evaluation of Cd Bioavailability

In this study, conventional chemical extraction methods were inappropriate to evaluate the bioavailability of Cd in the soils. The MgCl₂ extractable fraction (F1 + F2 from sequential extraction) showed no significant correlation with Cd content in rice grains and tea leaves (

Table 2. Pearson’s correlation coefficients (r) between Cd accumulations in rice grains (n = 43)/tea leaves (n = 18) and measurements of Cd in soils, between BCF_rice/BCF_tea and soil properties.

Measurements	Rice	Tea	Soil Properties	BCFrice	BCFtea	Soil Properties	BCFrice	BCFtea
DGT and DIFS (CE)	0.84 b	0.79 b	pH	−0.42 b	−0.68 b	Zn	−0.49 b	0.09 NS
Soil solution (Csolu)	0.68 b	0.71 b	SOC	−0.12 NS	0.15 NS	Sand	−0.37 a	−0.57 a
DGT (CDGT)	0.61 b	0.66 a	CEC	−0.09 NS	−0.16 NS	Silt	0.20 NS	−0.16 NS
MgCl2 extraction	0.28 NS	0.48 NS	Se	−0.41 b	−0.51 a	Clay	0.39 a	0.49 a

^{NS, a, b} Denote not significant, p < 0.05, p < 0.01, respectively.

). In this study, the DGT-measured Cd concentration (C_DGT) exhibited a positive correlation with Cd content in rice grains (r = 0.61) and tea leaves (r = 0.66), though the predictive performance was superior for tea compared to rice (Table 2). Surprisingly, a strong positive correlation was found between Cd in soil solution (C_solu) and Cd content in crops, with Pearson’s correlation coefficients surpassing those of DGT. However, the predictive performance for rice remained inferior to that for tea (Figure 3a). This phenomenon is primarily attributed to the formation of iron plaque on the root surfaces of many aquatic plants, including rice, within the aquatic rhizosphere microenvironment. This iron plaque is an adaptation mechanism to aquatic environments. Elevated levels of soil iron oxides lead to an increase in the formation of the iron plaque on rice roots. This root-bound iron plaque is predominantly composed of amorphous or crystalline Fe (hydroxy) oxides and tends to immobilize a portion of Cd as oxide-bound forms on rice roots, reducing the Cd uptake by rice plants. Consequently, the DGT measurement of the bioavailable fraction of Cd performs less effectively for rice compared to upland crops like tea [40,41].

The supply capacity of Cd from the solid phase to the soil solution can be quantified using the R value (R = C_DGT/C_solu, 0 < R < 1), where a higher R value (approaching 1) indicates a stronger ability of soil particles to resupply Cd into the soil solution [28]. When the supply of Cd from the solid phase to the solution is limited or approaches negligible supply (mainly diffusion), the R value approaches its minimum possible value (close to 0). In this study, the sequence of different soil R values from high to low is as follows: soils with pH 4.5~5.5 > soils with pH 5.5~6.5 > soils with pH above 6.5. The average R value for soils with pH above 6.5 is 0.39, notably lower than other soils, indicating a continuous supply process of solid-phase Cd into the soil solution in slightly alkaline soils. However, for other acidic soils, especially those with pH 4.5~5.5, the R value is higher (average 0.62), suggesting a relatively abundant supply of solid-phase Cd to the soil solution in more acidic conditions. In these more acidic soils, as plants absorb Cd from the soil solution, the solid phase of the soil can release a greater amount of bioavailable Cd into the liquid phase.

In order to investigate the concentrations of analytes that are effectively utilizable in both the soil solid phase and soil solution, the concept of effective concentration (C_E) is commonly utilized when employing DGT for studying soil environments. The calculation method for C_E is outlined in Equation (2). The diffusion coefficient (R_diff) values calculated using the DIFS model are used to correct C_DGT, resulting in the determination of analyte concentrations that can be effectively utilized within the soil. When the correlation between C_DGT and Cd content in crops is suboptimal, employing C_E for predicting Cd accumulation in crops often yields improved results [18]. In this study, the predictive performance of C_DGT for Cd accumulation in crops was not entirely satisfactory. Consequently, a correction was applied using the C_DGT and R_diff to derive the effective concentration (C_E) for a refined prediction. This approach demonstrated enhanced predictive accuracy (Figure 3b), especially the correlation coefficient (r = 0.84) between C_E and Cd accumulation in rice, surpassing the correlation coefficient (r = 0.79) between C_E and Cd accumulation in tea leaves. This observation underscores the capacity of the combined DGT technique and DIFS model to better simulate the dynamic migration of Cd in the soil-crop system.

3.3. Factors Controlling Cd Uptake by Rice and Tea

To identify the major factors influencing the uptake of Cd by rice and tea from the soil, we conducted a correlation analysis between BCF_rice, BCF_tea, and various soil physicochemical parameters. The Pearson’s correlation coefficients between BCF values and soil properties for rice and tea samples are presented in Table 2. Overall, significant correlations were observed between BCF_rice and soil properties such as pH, Se, Zn, clay, and sand content, while BCF_tea was influenced by soil pH, Se, sand, and clay content. Given the multitude of factors affecting the migration and accumulation of Cd in soil-crop systems, we introduced the random forest (RF) method as an advanced classification approach to further verify the influencing factors on Cd absorption by rice and tea. The RF model, an extension of the classification and regression trees (CART) model, is described in detail by Breiman [13]. Its advantages include sensitivity to noise and reduced bias in error estimation compared to other models [10]. Notably, soil properties including pH, Zn, sand, Se, and clay content account for 23.6%, 19.6%, 14.7%, 13.5%, and 11.4%, respectively, in influencing BCF_rice (Figure 4a). Similarly, sand, pH, clay, and Se account for 22.6%, 20.9%, 19.5%, and 16.5%, respectively, in affecting BCF_tea (Figure 4b).

Soil pH has been identified as a crucial factor influencing the uptake of Cd by rice and tea, displaying significant negative correlations with both BCF_rice and BCF_tea (Table 2). Soil pH could affect the speciation of Cd in soil. As soil pH increases, the reduction in H⁺ leads to a heightened negative surface charge on the soil, thereby promoting the adsorption of Cd²⁺ by various minerals in the soil. Additionally, pH affects the reactivity of Cd, compound solubility, and the extent of hydroxylation of Cd²⁺ on mineral surfaces [42,43]. Soil acidification, on the other hand, reduces the adsorption capacity and affinity of minerals for Cd²⁺ in the soil, converting Cd²⁺ adsorption from specific to non-specific ion exchange adsorption. Therefore, the application of soil liming can reduce the bioavailability of Cd in the soils, making it less prone to uptake by crops like rice and tea.

Soil texture is commonly recognized as a significant factor influencing the uptake of heavy metals such as Cd by crops. Typically, finer soil particles tend to accumulate more Cd compared to coarser soil particles [44]. This outcome can be attributed to the enhanced water retention, nutrient-holding capacity, and support for above- or below-ground organisms provided by finer soil particles. Moreover, finer particles exhibit greater capability in adsorbing key carriers of Cd, such as organic matter and soil minerals. In this study area, there is a notable negative correlation between sand content and both BCF_rice and BCF_tea, while clay content exhibits a significant positive correlation, and the influence of silt content is marginal (Table 2). The impact of sand content on Cd uptake by crops, particularly tea (Figure 4b), is substantial. In soils with high sand content, the ability of rice and tea to absorb Cd is diminished due to their poorer water-holding capacity and weaker adsorption capabilities towards Cd-bearing minerals like iron and manganese oxides [45].

Soil Zn content is an additional factor affecting the uptake of Cd by rice and tea. The soil-crop system constitutes a complex arrangement, wherein alterations in the content of one element can influence the functionality of others, ultimately impacting the migration of heavy metals like Cd within the soil-crop system. There exists a certain antagonistic relationship between Zn and Cd since previous studies have found that applying low doses of ZnO nanoparticles to soil can reduce the toxicity of Cd on crop growth [46]. Given the structural similarity between Zn and Cd ions, higher Zn content in the soil can mitigate Cd toxicity to crops and promote crop growth. As the bioavailable concentration of Zn increases, the bioavailable concentration of Cd in the soil decreases significantly [47].

Furthermore, prior research indicates an antagonistic interaction between Se and Cd. Elevated Se levels in soils can facilitate the deposition of Cd in root cell walls or vacuoles, consequently reducing its transportation to other plant organs, thus diminishing the bioavailability of Cd [48,49]. Soil Se content ranked fourth among the factors influencing the uptake of Cd by both rice and tea (Figure 4). Since Se-oxidizing bacteria can enhance the bioavailability of Se in soils [50]. In this study, the application of Se-oxidizing bacteria in the seleniferous soil was thought to be an effective approach to diminish the uptake of Cd by crops such as rice and tea and its subsequent transport to the aboveground parts.

4. Conclusions

We conducted an analysis of the Cd content characteristics in soil, rice grains, and tea leaves within the study area. Our findings indicate that the cumulative content of water-soluble (F1), exchangeable (F2), and carbonate-bound (F3) Cd fractions collectively accounts for more than 50% of the total Cd content in the soil. This suggests a high potential for Cd bioavailability in the soil. However, this situation is also associated with an elevated rate of Cd accumulation in rice, surpassing 32.5%. When assessing Cd bioavailability using the DGT method, we observed favorable and positive correlations between C_DGT, C_solu, and Cd content in crops. The most accurate predictions were obtained by utilizing the corrected C_E values based on the R_diff calculated using the DIFS model.

More alkaline and sandy agricultural soils are considered more suitable for growing rice and tea plants with low Cd content. In addition to soil pH and texture, Se and Zn levels in the soil play crucial roles in influencing the uptake of Cd by rice and tea crops. Despite the known antagonistic effect of Se on Cd uptake by crops, the co-enrichment of Se and Cd in the soil often results in the concurrent accumulation of both elements in crops. This situation may lead to the production of Se-enriched crops that also exceed permissible Cd levels. Therefore, enhancing the bioavailability of Se in the soil and reducing the bioavailability of Cd through soil liming and/or the application of Se-oxidizing bacteria can facilitate the safe cultivation of locally Se-enriched agricultural products while mitigating the risks of Cd contamination.