Introduction

Soils in vegetable fields are of great significance to sustain public health through vegetable production. The rapid development of urbanization and industrialization as well as increasing dietary requirements has attracted much more attention of trace element contamination in vegetable soils (Li et al. 2012a; Mapanda et al. 2005; Rehman et al. 2017; Zhao et al. 2010), especially in the urban-rural conjunction areas due to the proximity to large residents (Chen et al. 2008; Sharma et al. 2007). Moreover, megacities have been the center of resource consumption and then the major sources of pollutant emissions, which elevate the concentrations of trace elements in the soils surrounding the city. Toxic trace elements in soils pose potential threat to human health through various absorption pathways such as diet through soil-food chain, direct ingestion, and dermal contact (Khan et al. 2008; Luo et al. 2011b). Consequently, some serious systemic health issues may develop as a result of excessive accumulation of trace elements (e.g., As, Cd, Pb), which are associated with diseases like breast and gastrointestinal cancer, kidney damage, nervous system, memory deterioration, and bone diseases (Itoh et al. 2014; Kabata-Pendias and Mukherjee 2007; Luo et al. 2012; Zhao et al. 2014).

Identifying the sources of trace elements in the vegetable land of urban-rural conjunction areas is not easy due to the complex human activities. Multivariate statistical analysis (e.g., principal component analysis, cluster analysis) with geographic information and temporal variation provides the possibility to identify the sources of trace elements in soils (Huang et al. 2007; Li and Feng 2010; Zhao et al. 2010). Moreover, as an effective fingerprinting method, stable Pb isotopic ratios have been effectively applied to distinguish its sources in soils (Bing et al. 2014; Komarek et al. 2008; Li et al. 2012a; Kelepertzis et al. 2016). Thus, the integration of these different methods and techniques will help to better understand the characteristics and anthropogenic sources of trace elements in the soils of vegetable fields.

Many environmental models/indices have been developed to assess and classify the contamination of trace elements in soils. Single geochemical contamination indices (e.g., geoaccumulaiton index, pollution index) generate the contamination assessment of individual element and decipher the effect of single element on soils (Chen et al. 2005b; Li et al. 2014, 2015; Tian et al. 2017). However, integrated pollution index considers the general contamination state of all trace elements focused in soils, which is a better indicator for the general assessment of soil contamination (Chen et al. 2008, 2015; Luo et al. 2012). Human health risk assessment proposed by USEPA connects the contamination levels of trace elements with a probability of toxic effect for individual population and has been widely used to assess the exposure rate of trace elements in soils on human health (Chen et al. 2015; Jiang et al. 2017; Li et al. 2014; Niu et al. 2013; Ordóñez et al. 2011). Ingestion and dermal contact is the main pathway of population exposure to toxic elements in soils and may cause adverse effects on immune systems.

Chengdu city, the capital city of Sichuan province, is the economic, cultural, education, and high-tech industry center of Southwest China. Since last few years, Chengdu city has developed rapidly, and it is building a nationally advanced city. The gross domestic product (GDP) of Chengdu city has been increasing since 2011 and reached 1080 billion in 2015 according to the Statistical Bulletin for Chengdu Economy and Society Development in 2015. With the urban expansion and economic increment, the land layout has been altered in the urban-rural conjunctions, and the pollutant emissions tend to deteriorate the environment. Consequently, diverse industries, heavy traffic, and persistent agricultural production could increase the loadings of trace elements in agricultural soils. More importantly, Chengdu city as a popular tourism city, its food (Szechuan cuisine) is world-famous. A large amount of vegetables are ingested by local residents and tourists every day. The quality of the vegetable soils from trace element contamination is thus vital in keeping the public health.

In this study, seven potential toxic elements (As, Cd, Cr, Cu, Ni, Pb, and Zn) in the suburban vegetable soils of Chengdu city were selected to analyze their concentrations. The objectives are to (1) investigate their spatial and temporal variation in 2012 and 2016; (2) identify the possible sources and factors controlling their distribution through statistical analysis, geographic information system, and Pb isotopic ratios; and (3) assess the degree of trace element contamination subject to human health risk in local populations.

Materials and methods

Study area

Chengdu city (30° 05′–31° 26′ N, 102° 54′–104° 53′ E) is located at the western Sichuan plain (Fig. 1). The total area of Chengdu city is 14,605 km2, and the permanent resident population is 15.7 million in 2015. The climate of Chengdu city is typical subtropical humid monsoon climate with annual mean temperature and precipitation of 15.2–16.6 °C and 900–1300 mm, respectively. The soils dominate by paddy soils, purple soils, fluvo-aquic soils, and yellow brown soils according to the China soil classification system. The study area including eight counties in the suburban areas has been interfered with intense urbanization and industrialization over last few decades. The industries, constructions, and agricultures currently coexist in these areas. Nearly every county has its industrial or economic development zones, whereas the high and new technology industries developed in recent years are concentrated in the southern and western regions (Table S1).

Fig. 1
figure 1

The study area and the sampling sites in Chengdu city

Full size image

Sampling and preparation

Soil samples were collected from the vegetable field in the Chengdu suburban area (outside the three-ring road) in summer of 2012 and 2016, respectively (Fig. 1). With the fast variation of land use in Chengdu city in recent years, the sampling sites in the two campaigns were not absolutely uniform. Thirty-three sites at each campaign were selected to collect the surface soils (0–20 cm) by the same investigators. At each sampling site, soils from five plots were collected to form a composite sample. The specific information about the sampling sites, vegetable types, and human activity around the sites as sampling is presented in Table S2. In the laboratory, the samples were air-dried and sieved to 2 mm. A portion of soils was used to measure the pH and organic matter content, and the other one was grounded passing into a nylon sieve of 200 mesh for the analysis of element concentrations.

Chemical analysis

Soil organic matter (SOM) was determined by the method of LOI (loss on ignition). Soil samples were combusted at 450 °C for 4 h in a muffle furnace after 105 °C oven-dried, and the LOI was calculated through the mass difference of before and after ignition. Soil pH was measured in CO2-free deionized water by a pH meter with a soil-water ratio of 1:2.5 (w/v).

The soil samples were digested with a concentrated acid mixture (HNO3, HF, HCl, and HClO4), and then treated by a variety of analytical methods (Liu et al. 2007). The concentrations of As, Cd, Cr, Cu, Ni, Pb, and Zn were determined by inductively coupled plasma mass spectrometer (ICP-MS), and those of Al, Ca, Fe, K, and Mg were measured by inductively coupled plasma-atomic emission spectrometer (ICP-AES). Standard solution SPEX™ was used as the standard. Quality control was assured by the analysis of blanks, duplicate samples, and reference materials (GSD-9 and GSD-11). The results of both precision and accuracy were below 5% through the relative standard deviation (RSD) of the repeated samples and standard reference materials, respectively.

Lead isotopes (208Pb, 207Pb, and 206Pb) in the soils were determined by ICP-MS (Agilent 7700×). A standard reference material from the US National Institute of Standards and Technology-SRM 981 was used for instrument calibration, and the standard material of GBW04426 from China was selected for the quality control. The maximum deviations of 208Pb/206Pb and 207Pb/206Pb related to the repeated measurements of the Pb standard (GBW04426) were less than 0.002, and the RSD were 0.09 and 0.17%, respectively.

Contamination indicators

Geoaccumulation index (I geo) was calculated to evaluate contamination level of individual element in the soils (Muller 1969):

$$ {I}_{\mathrm{geo}}={\mathit{\log}}_2\left({C}_i/1.5{C}_0\right) $$
(1)

where C i is an element concentration in samples, C 0 is the background, and the constant 1.5 is used to analyze natural fluctuations of the element concentration and very small anthropogenic influences (Loska et al. 2004). In this study, the background concentrations of trace elements in the soils of Sichuan province were chosen as the reference (Chen et al. 2015). The I geo consists of seven classes: uncontaminated (I geo ≤ 0), uncontaminated to moderately contaminated (0 < I geo ≤ 1), moderately contaminated (1 < I geo ≤ 2), moderately to heavily contaminated (2 < I geo ≤ 3), heavily contaminated (3 < I geo ≤ 4), heavily to extremely contaminated (4 < I geo ≤ 5), or extremely contaminated (I geo > 5).

Nemerow integrated pollution index was applied to assess the overall contamination of trace elements in the soils (Nemerow 1985):

$$ NIPI={\left[ 0.5\times \left({PI_{\mathrm{Avg}}}^2+{PI_{\mathrm{Max}}}^2\right)\right]}^{1/ 2} $$
(2)

where pollution index (PI) is defined as: PI = C i /C 0. The PI values are categorized into three ranges to indicate the element contamination degree: low (PI ≤ 1), middle (1 < PI ≤ 3), and high (PI > 3) (Chen et al. 2005b). PI Avg and PI Max are the mean and maximum values, respectively. The classification of Nemerow integrated pollution index (NIPI) is the following: safe (NIPI ≤ 0.7), precaution (0.7 < NIPI ≤ 1), slight (1 < NIPI ≤ 2), moderate (2 < NIPI ≤ 3), or heavy (NIPI > 3).

Health risk assessment indicators

Exposure assessment

Individual exposures to soil elements are generally through three main pathways: ingestion, dermal contact, and inhalation. Ingestion and dermal absorption are the most common exposure pathways for As, Cd, Cr, Cu, Ni, Pb, and Zn in contaminated soils (Ordóñez et al. 2011; Li et al. 2014). Thus, these two exposure ways were considered in this study, and the exposure dose was calculated by the following functions (USEPA 1989):

$$ {\mathrm{ADI}}_{\mathrm{I}}=\frac{C_s\times SIR\times EF\times ED}{BW\times AT} $$
(3)
$$ {\mathrm{ADI}}_{\mathrm{D}}=\frac{C_s\times SA\times AF\times \mathrm{ABS}\times EF\times ED}{BW\times AT} $$
(4)

where ADI I is the mean daily intake of elements from soil ingestion (mg/kg-day), ADI D is the mean daily intake of elements from dermal absorption (mg/kg-day), C s is the soil element concentration (mg/kg), SIR is the ingestion rate (mg/day), SA is the exposed skin surface area (cm2), AF is the adherence factor (mg/cm2-day), ABS is the dermal absorption factor (unitless), EF is the exposure frequency (day/year), ED is the exposure duration (year), BW is the body weight of the individual exposed (kg), and AT is the time period over which the dose is averaged (day).

Non-carcinogenic risk assessment

The hazard quotient (HQ) defined as the quotient of chronic daily intake or the dose divided by a toxic threshold (reference dose, RfD, mg/kg-day) of an element is calculated to characterize the non-carcinogenic hazards. A hazard index (HI, Eq. 5) method was applied to assess the overall potential for non-carcinogenic effects by much more elements (USEPA 1989). It is unlikely to cause adverse health risk if HI < 1; if HI > 1, the adverse health risk may occur (Man et al. 2010; Li et al. 2014).

$$ HI=\sum {HQ}_i=\sum \frac{ADI_i}{RfD_i} $$
(5)

Carcinogenic risk assessment

The carcinogenic risk (Eq. 6) is the probability of an individual developing cancer in the whole lifetime by exposure to potential carcinogen. The slope factor (SF) converts the daily intake of an element to the incremental risk of an individual developing cancer (USEPA 1989). Risks ranging between 10−6 and 10−4 are considered as acceptable or tolerable.

$$ Risk= ADI\times SF $$
(6)

In this study, the populations were divided into three groups (adult males, adult females, and children) to determine the exposure risk to the targeted trace elements due to the behavioral and physiological differences. All probabilistic parameters used in Eqs. (3)–(6) can be preferentially collected from reports in China (Chen et al. 2015; Xu et al. 2014).

Statistical analysis

One-Way ANOVA (Fisher Test, p < 0.05) was applied to characterize the significant difference of element concentrations. Multivariate statistical analysis including factor analysis and Spearman correlation analysis were used to identify the relationships between elements and other parameters in the soils. The statistical analysis was performed by the software package SPSS 16.0 and Origin 8.0. Spatial distributions of trace elements were produced by geographic information system (GIS) mapping software (ArcGIS 9.3). The Kriging method was used to interpolate the element concentrations between sites.

Results and discussion

Soil properties

A descriptive statistical summary of the soil properties is presented in Table 1. The soils were generally acidic in both surveys. Compared with the soil pH in 2012, it was relatively lower in 2016 with a higher coefficient of variation (CV) spatially. The contents of SOM did not show significant difference in both years, whereas its CV was much higher in 2016 than in 2012. Compared with the soil pH and SOM in Chinese farmland soils (Niu et al. 2013), the much lower soil pH (especially in 2016) and higher SOM were observed in the Chengdu suburban vegetable soils, which suggested that the disturbance of human activity around the city might affect the soil properties.

Table 1 The statistical characteristics of element concentrations, pH, and organic matter contents in the soils
Full size table

The major elements (Al, Ca, Fe, K, and Mg) in the soils, normally the main constituents of soil parent minerals, did not vary significantly in 2012 and 2016. The spatial variation was very low for Al and Fe, which indicated that the soils of Chengdu city had a similar natural geological background. The spatial distributions of Ca, K, and Mg in the soils showed nearly moderate variation according to their CV values. Except for soil heterogeneity, this was attributed to different fertilization ways and intensity due to the nutrition effect of them for vegetable production or the lime addition for neutralizing the soil acidification.

Concentrations of trace elements

In general, the concentrations of trace elements in the soils did not show significant difference in the two surveys, except for Pb and Zn which were slightly higher in 2016 (Table 1). Moreover, the maximum concentrations of As, Cd, Cr, Cu, and Zn were clearly observed in the soils investigated in 2016. According to the values of CVs, the spatial variation degrees of trace elements in the soils decreased in the order of As > Cd ≈ Zn > Pb > Cr ≈ Cu ≈ Ni in 2012 and Cu > As ≈ Zn > Cd > Pb > Cr > Ni in 2016. Arsenic, Cd, Cu, and Zn presented higher spatial heterogeneity in 2016 than in 2012. The concentrations and CV values inferred that the accumulation of trace elements in the suburban vegetable soils was site-specific in 2016, although it did not increase remarkably in 2016 compared with that in 2012.

Compared with the concentrations of trace elements in the agricultural/vegetable soils from different regions (Table S3), the difference was clearly observed among the cities and/or countries. This was to a large extent attributed to the spatial heterogeneity of the soils. On the other hand, the anthropogenic inputs have been considered as a major factor especially in the undeveloped countries (e.g., Bangladesh) and the industrial cities (e.g., Daye city in China). The concentrations of trace elements in the Chengdu suburban vegetable soils were comparable to those in the Chinese farmland soils and in the soils of Sichuan Province (Table S3). However, they were distinctly higher than those in the soil background of Sichuan Province. The sites with the concentrations of As, Cd, Cr, Cu, Ni, Pb, and Zn below the Chinese soil guidelines (Grade I) reached 42.4, 24.2, 97.0, 78.8, 97.0, 72.7, and 60.6%, respectively in 2012, and 45.5, 30.3, 97.0, 78.8, 100, 39.4, and 36.4% respectively in 2016. The percentages of the sites with these element concentrations between Grade I and Grade II were 57.6, 75.8, 3.0, 21.2, 3.0, 27.3, and 39.4% respectively in 2012, and 54.5, 69.7, 3.0, 18.2, 0, 63.6, and 60.6% respectively in 2016. Furthermore, one site in 2016 showed the concentrations of Cu and Zn exceeding the Grade II standard. Based on the comparison, it was clear that As, Cd, Cu, Pb, and Zn in the vegetable soils had higher concentration grade than Cr and Ni. Moreover, more sites showed the concentrations of As, Cd, Pb, and Zn above the Grade I in 2016 compared with that in 2012.

Factors influencing element distribution

Multivariate statistical analysis including factor analysis and Spearman correlation analysis was performed on the concentrations of trace elements to decipher their relationships. Trace elements in vegetable soils surrounding urban or suburban areas were susceptible to natural and anthropogenic influences (Chen et al. 2008; Liu et al. 2005; Luo et al. 2011a; Micó et al. 2006), which included the soil weathering and anthropogenic inputs from various sources.

Three main factors for the soil elements investigated in 2012 and 2016 were respectively extracted with total accumulative variances of 82.2 and 81.7% (Table 2, Fig. S1). In 2012, the Factor 1 included the positive loadings of Al, Ca, K, Mg, and Zn, and the negative loadings of As and Pb. The elements of Al, Fe, Cr, Cu, Ni, and Zn were positively correlated to the Factor 2. The remaining elements of Cd, Pb, and Zn positively grouped the Factor 3. In 2016, the Factor 1 corresponded positively with Al, Ca, K, and Mg, and negatively with Fe, As, and Pb. The Factor 2 included the positive correlation loadings of Cd, Cr, Cu, Pb, and Zn. The Factor 3 consisted of Al, Fe, Cr, and Ni with positive loadings. The results of Spearman correlation analysis were generally consistent with those of factor analysis (Table S4). In 2012, As correlated positively with Pb but negatively with Zn, and Cd showed positive correlation with Cu and Zn (p < 0.05). The interrelationships of Cr, Cu, Ni, and Zn were significant (p < 0.01), and these elements also showed more significant correlations with most of major elements compared with As, Cd, and Pb. The correlation of trace elements with each other and major elements presented marked difference in 2016 from that in 2012 (Table S4). Specifically, As correlated positively with Pb (p < 0.05) but negatively with Cu (p < 0.05) and Zn (p < 0.01). Cadmium was positively related to all trace elements, Cr had a close relationship with Cu, Ni, and Zn (p < 0.01), but Cu did not correlate with Ni. Moreover, As showed negatively correlation with Al, Ca, K, and Mg (p < 0.01); Cd correlated positively with Ca and Mg (p < 0.01); and Pb did not correlate with all major elements. The remaining trace elements of Cr, Cu, and Ni displayed significant correlations with at least three major elements.

Table 2 Characteristics of the total explained variance and rotated component matrix for the elements
Full size table

The results of statistical analysis above classified clearly the trace elements in the soils, which indicated their different sources. Considering the various human disturbances in the suburban areas of Chengdu city as well as the insignificant or negative relationships with crustal elements (e.g., Al, K, Mg), As, Cd, and Pb in the vegetable soils investigated in 2012 were mainly derived from non-crustal sources, Cr, Cu, and Ni were mostly from natural sources, and Zn indicated the heterogeneity of its origins, from both natural and non-crustal sources. In 2016, As, Cd, and Pb were still mainly from non-crustal sources, Cu, Cr, and Zn were notably influenced from the anthropogenic sources, whereas Cr and Ni were mainly from the natural sources.

Besides the source effect, soil properties also played a vital role in the spatial distribution of trace elements in soils (Niu et al. 2013). Compared with the soil properties analyzed in 2012, the soil pH in 2016 decreased observably (Table 1). The direct correlation of soil pH with trace elements was not significant except for Cd in 2012 according to the correlation analysis (Table S4). However, the significantly positive correlation of pH with Ca (p < 0.01) and Mg (p < 0.05) was clearly observed. The variation of soil pH could alter the distribution of other major elements, especially for the alkali metals (e.g., Ca, Mg) which commonly formed carbonates in soils. As the soil pH decreased, the trace elements bound with the carbonates might be released from the soils. On the other hand, SOM showed different correlations with individual element in the soils (Table S4), and its larger spatial variation was observed in 2016 than in 2012 (Table 1). The SOM has capacities to adsorb and/or complex trace elements in soils (Zeng et al. 2011a), which might be another factor resulting in the heterogeneous distributions of race elements.

Spatial distribution of trace elements

The spatial distribution of trace elements in the vegetable soils in 2012 and 2016 is shown in Fig. 2 and Table S5. In 2012, the distributions of As and Pb showed a similar spatial pattern with a higher accumulation in the southern and eastern regions of Chengdu city (Jinjiang, Chenghua, and Longquanyi County). The lower concentrations of soil As were observed in Wenjiang and Pidu County, and those of Pb existed in Wenjiang, Qingyang, and Shuangliu County. Compared with the distribution in 2012, we observed the similar spatial patterns of As and Pb in 2016, whereas their concentrations increased generally at respective county. In 2012, the distribution of soil Cd was mainly concentrated in the southern and western areas (Shuangliu and Qingyang County) as well as partial eastern sites (e.g., Chenghua County). However, its highest concentration was observed in Longquanyi County where the Cd concentrations were generally in a low level. In 2016, Chenghua, Shuangliu, and Qingyang County exhibited higher Cd concentrations in the soils.

Fig. 2
figure 2figure 2

The spatial distribution of individual element (unit: mg/kg) in the vegetable soils in 2012 and 2016. The vegetable lands in the urban-rural conjunction areas were considered as mapping the spatial distribution of trace element concentrations

Full size image

The concentrations of soil Zn investigated in 2012 were relatively higher in the Xindu and Qingyang County compared with those in the Jinjiang, Longquanyi, and Chenghua County. In 2016, the concentrations of soil Zn were relatively high in the Pidu County, northwestern Chengdu, with the highest value of 297 mg/kg. Moreover, Qingyang County also displayed relatively high concentrations of Zn in the soils, whereas Jinjiang County showed the lowest level. Similar to As and Pb, the concentrations of soil Zn increased markedly at respective county in 2016 compared with that in 2012.

The remaining trace elements (Cr, Cu, and Ni) in the soils generally presented a similar spatial distribution. In 2012, their concentrations were clearly higher in the Xindu County of northern Chengdu, although the value of Ni was highest in Pidu County. The concentrations of Cr, Cu, and Ni exhibited a relatively low level in the soils of the Shuangliu and Chenghua County. In 2016, the higher concentrations of these elements were observed in the Pidu County of northwestern Chengdu, whereas the lower values were observed in the Chenghua and Xindu County for Cr, Longquanyi County for Cu, and Chenghua and Qingyang County for Ni. Except the Pidu and Xindu County, the concentrations of Cr, Cu, and Ni at respective county did not vary significantly in 2012 and 2016.

Source identification of trace elements

Statistical analysis with geographic information

The results of statistical analysis (Tables 2, S4) showed that the sources of As, Cd, and Pb in the vegetable soils were mainly derived from non-crustal sources, and Zn as well as Cu in 2016 was from mixed sources of natural and anthropogenic origins. Therefore, the specific sources of As, Cd, Cu, Pb, and Zn in the soils from different regions were further identified combined with their spatial and temporal variations.

The hot spots of As and Pb in the soils were markedly distributed in the eastern and southeastern regions of Chengdu city, such as the Jinjiang, Longquanyi, and Chenghua County; those of Cd were observed in the Chenghua County, and its highest existed in the Longquanyi County (Fig. 2, Table S5). Trace elements like As, Cd, and Pb are important components of pesticides, herbicides, and fungicides, and they are also well-known elements related to the application of phosphate fertilizers (Acosta et al. 2011; Marrugo-Negrete et al. 2017; Nezat et al. 2017). In the suburban Chengdu, the farmlands are mainly located at the eastern areas including many vegetable production bases (Table S2). Agricultural activities were a major reason for the high accumulation of As, Cd, and Pb in the soils of these regions. Compared with As and Pb, the hot spots of Cd tended to be more local specific, especially in 2016 (Fig. 2), which was probably associated with varied fertilizer inputs. The high accumulation of Cu and Zn in the vegetable soils of southeastern areas could be related to the utilization of fertilizers, pesticides, fungicides, and livestock manure.

The spatial layouts of industrial and mining enterprises and transportation industries are another major reason for the accumulation of trace elements in the vegetable soils. The industrial and mining enterprises with coal consuming as the main energy in Chengdu city are mainly distributed in the suburban areas, especially in the northeastern, eastern, and southeastern regions (Table S1). The emissions from coal combustion and metal smelting could directly increase trace element accumulation in the soils. Traffic effect on soil trace elements has been widely reported, and the elements from vehicle emissions included As, Cd, Cu, Pb, and Zn through leaded gasoline, wear of tires, road abrasion and brake linings, leaked fluids and fuel combustion (Nezat et al. 2017; Trujillo-Gonzales et al. 2016; Wang et al. 2017a; Zhong et al. 2012). According to Chengdu Yearbook, the vehicle quantity has been increasing in last few decades. The automobile volume in Chengdu reached 4.29 million in 2015, and the private car quantity accounted for 77.2%. Consequently, the traffic volume per day increases significantly, but the travel speed decreases markedly. This variation would increase toxic element emissions due to the longer operation time and more frictions between the tire and the road. Moreover, the road dust containing trace elements could also deposit in the vegetable soils under the effect of heavy traffic (Werkenthin et al. 2014). Many researches have observed the negative relationship between trace element concentrations and the distance to road (Clarke et al. 2015; Wang et al. 2017a; Yan et al. 2013; Zhang et al. 2015). However, we did not find this relationship based on our data (Table S2). This was attributed to the complex road conditions in the suburban areas, and our analysis only focused on the distance of the sampling sites to the main roads (e.g., highway).

The hot spots of Cd, Cu, Pb, and Zn (especially in 2016) in the soils were also observed in the northern and northwestern regions (mainly in Pidu County) (Fig. 2). Besides the industrial and agricultural sources, these elements were probably from domestic pollutants and sewage irrigation that contained large amounts of toxic elements (Lu et al. 2015; Sharma et al. 2007). When collecting the soil samples, we observed lots of municipal solid waste, household garbage, and construction waste surrounding the vegetable lands in this region. The open burning of municipal solid waste are an important source of toxic elements, especially in urban-rural conjunctions (Wiedinmyer et al. 2014), and the atmospheric emission volumes of Cd, Cu, Pb, and Zn from open burning in Sichuan province have been reported ranking the first in Southwest China (Wang et al. 2017b). Moreover, the long-term use of wastewater irrigation for agricultural production increased the loadings of toxic elements in the soils. Among all trace elements, only Cd showed the hot spot in the southwestern regions in both surveys (Fig. 2). The automobile industry of Chengdu city is concentrated in this region indicating the soil Cd sources from the automobile service and maintenance. Cadmium as an important alloy material is commonly used to metal plating from corrosion. In addition, the Chengdu airport, the largest in Southwest China, is also located in this region, which suggested that the emissions from the air transportation were a potential source of Cd in the vegetable soils.

Lead isotopic ratios

The Pb isotopic ratios of 206Pb/207Pb versus 208Pb/206Pb were applied to trace the specific sources of Pb in the soils (Table S6), which would be also conducive to reveal the probable sources of other toxic elements that correlated significantly with Pb (e.g., As, Cd). The 206Pb/207Pb and 208Pb/206Pb showed significant relationships not only in the soils of individual year but also of both years (Fig. 3), which indicated that two end-members of Pb including natural and anthropogenic sources existed in the soils (Klaminder et al. 2011; Komarek et al. 2008). Compared with the isotopic ratios (Fig. 3a), the values in the vegetable soils were close to those in the emissions of coal combustion in Chengdu city, Pb-containing ores, and smelting. However, the ratios in the Chengdu exurban soils were generally biased from those in our investigated soils. This difference suggested that the main Pb sources in current China indeed contributed to its accumulation in the surface vegetable soils (Flegal et al. 2013; Li et al. 2012b). Since the reports of Pb isotopic ratios in auto parts (e.g., tire, brake linings) were limited, the ratios in the vehicle exhausts in Chengdu were compared. The results showed that the Pb isotopic ratios in the vehicle exhausts were different from those in the soils suggesting the low contribution of vehicle emissions to the soil Pb. This further supported that the banned application of leaded gasoline since 2000 in China has reduced the inputs of Pb in the soils (Chen et al. 2005a). Compared with aerosols and precipitation, the Pb isotopic ratios in the air dust were closer to those in the vegetable soils, which suggested that the emissions of air dust from local roads, industrial plants, and so on could be a more important source of Pb in the soils due to the faster deposition of coarser particles.

Fig. 3
figure 3

The Pb isotopic ratios in the vegetable soils and other materials (a) and their spatial and temporal distribution in different counties (b). JJ Jiangjin County, LQY Longquanyi County, CH Chenghua County, PD Pidu County, WJ Wenjiang County, XD Xindu County, QY Qingyang County, SL Shuangliu County. Refs: Exurb soils (Shang 2007); vehicle exhausts (Gao et al. 2004); coal combustion in Chengdu (Shang 2007); air dust in Chengdu (Gao et al. 2004; Shang et al. 2006); aerosols in Chengdu (Bollhofer and Rosman 2001; Gao et al. 2004); rainwater in Guiyang (Li et al. 2012); smelting (Bi et al. 2007); Pb-containing ores in SW China (Xiao et al. 2012; Xue et al. 2007); and Chengdu population blood (Zeng et al. 2011b)

Full size image

As illustrated in Fig. 3, it was clear that the Pb isotopic ratios showed a marked variation in 2012 and 2016. The 206Pb/207Pb in the soils of 2016 decreased and tended to reach the ratios in anthropogenic emissions including air dust, coal combustion, ores mining, and smelting, indicating the increasing anthropogenic contributions to the soil Pb. Furthermore, the Pb isotopic ratios showed a distinct spatial variation (Fig. 3b). The ratios in the southeastern, eastern, and northeastern regions of Chengdu (Jinjiang, Longquanyi, Chenghua, and Xindu County) were similar to those in the emissions of coal combustion and ores-mining-related activities. However, the sources of soil Pb in the rest counties (Pidu, Wenjiang, and Shuangliu County) through its isotopic ratios tended to be more complex, which was also related to the air dust inputs.

Assessments of trace element contamination and health risk

Contamination assessment

According to the concentrations of trace elements in the vegetable soils (Table S3), the values of some elements (e.g., As, Cd, Pb) exceeded the Chinese soil guidelines (Grade I), and in some sites, their concentrations reached the Grade II level. This indicated that the soil quality of Chengdu suburban vegetable lands was influenced by these elements. The geochemical indices of I geo and PI for individual element were calculated to assess their contamination state (Fig. 4). Both I geo and PI decreased in the order of Cd > As ≈ Zn > Cu ≈ Pb > Cr ≈ Ni in 2012 and of Cd > Zn > As ≈ Cu ≈ Pb > Cr ≈ Ni in 2016. In general, the contamination of As, Cd, Cu, Pb, and Zn in the soils reached the uncontaminated-moderate level through the classification of I geo, especially in 2016, and Cr and Ni showed very low contamination level. Compared with the level in 2012, the contamination of Cd, Pb, and Zn increased in 2016, and that of Cu showed a large variation; otherwise, the contamination of rest elements did not show a significant difference. The spatial distribution of trace element contamination in the soils corresponded to that of their concentrations (Fig. 2). In 2012, except Cd that showed moderate contamination in nearly 50% sampling sites, the contamination of other elements in most sites did not reach moderate level (Table 3). However, the regions with moderate contamination of soil Cd, Pb, and Zn increased in 2016, and one site in the Pidu County showed the heavy Cu contamination.

Fig. 4
figure 4

The characteristics of I geo and PI of trace elements in the soils

Full size image
Table 3 Percentage of the sampling sites with different contamination levels of trace elements in the vegetable soils based on I geo and PI
Full size table

The NIPI was applied to assess the soil contamination by all trace elements (Fig. 5). The NIPI values varied between 1.4 and 2.7 (mean: 1.8, median: 1.7) in 2012, and between 1.2 and 4.0 (mean: 1.9, median: 1.8) in 2016, which generally showed a slight contamination level. However, there were 12.1 and 21.2% soil sites in 2012 and 2016 respectively reaching the moderate contamination level, and 6.1% sites in 2016 were heavily contaminated by these elements. The higher contamination of trace elements in the suburban vegetable soils was mainly distributed in the eastern and western regions of Chengdu city in 2012, whereas it changed markedly in 2016 (Fig. 5). The soils in the southeastern, southern, and northwestern regions presented the relatively higher contamination by the trace elements in 2016, although the characteristic of sporadic hot spots was clearly observed. The spatial variation of soil trace element contamination in 2012 and 2016 confirmed that the intensive human interference deteriorated the vegetable soils, which might threat to the health of local people. On the other hand, our data indicated that the trace elements in the suburban vegetable soils could be an effective index to reflect the progress of modern urbanization with the fast changes of land layout and living standards.

Fig. 5
figure 5

The spatial variation of Nemerow integrated pollution index (NIPI) of trace elements in the soils in 2012 and 2016. The vegetable lands in the urban-rural conjunction areas were considered as mapping the spatial characteristics of trace element contamination

Full size image

Health risk assessment

As shown in Table 4, the trace elements in the Chengdu suburban vegetable soils had low non-carcinogenic risk to local residents in both surveys due to the HI far lower than 1.0 (Man et al. 2010; Li et al. 2014). In general, the non-carcinogenic risk for local populations showed the same order of As > Cr > Cd > Ni > Pb > Cu > Zn in 2012 and 2016, and for the public considered, it decreased in the order of children > adult females > adult males. The similar findings have been reported from the whole Chinese soils (Chen et al. 2015), Chinese mine soils (Li et al. 2014), and other agricultural soils (Khan et al. 2008; Rehman et al. 2017). Children are easily susceptible to soil trace elements due to their physiological features and poor health consciousness; for example, the hand-mouth oral ingestion and dermal adsorption are their critical exposure pathways (Guney et al. 2010; Li et al. 2014). Except the HI of Zn increased slightly in 2016 than in 2012, the HI values of other elements did not vary significantly.

Table 4 Non-carcinogenic risks of trace elements and carcinogenic risk of As for males, females, and children in the vegetable soils of Chengdu city
Full size table

Since the carcinogenic slope factors are not available for other trace elements, only the carcinogenic risk of As in the soils was assessed (Table 4). The carcinogenic risk of As in the soils was generally acceptable for adults according to the risk ranking of 10−6–10−4 (Chen et al. 2015). The carcinogenic risk posed by As was markedly higher for children than for adults. In 2012, the values of carcinogenic risk for children varied between 5.48 × 10−5 and 1.59 × 10−4, and the sites with the risk index higher than 10−4 reached 60.6% in the study area. In 2016, the carcinogenic risk of As varied between 4.40 × 10−5 and 1.82 × 10−4, and the percentage of concerned sites were 48.5%. Although the region contaminated by As decreased in 2016 compared with that in 2012, its carcinogenic risk in some sites was much more above acceptable levels. The health risk of As in the vegetable soils, especially in the regions with high As concentrations (Fig. 2), should be paid more attention in future research.

Conclusions

The concentrations of As, Cd, Cr, Cu, Ni, Pb, and Zn in the Chengdu suburban vegetable soils did not vary significantly in 2012 and 2016. The spatial distribution of As, Cd, Cu, Pb, and Zn presented much more hot spots in 2016 than in 2012. Anthropogenic sources including agricultural activities, industrial emissions, road dust with heavy traffic, and open burning of solid waste contributed to the soil element accumulation. The contamination of trace elements in the soils decreased in the order of Cd > As ≈ Zn > Cu ≈ Pb > Cr ≈ Ni in 2012 and Cd > Zn > As ≈ Cu ≈ Pb > Cr ≈ Ni in 2016, and the soils were slightly to moderately contaminated by these elements through integrated contamination index. The sites with the contamination of trace elements increased markedly in 2016 compared with that in 2012. The non-carcinogenic risk of the trace elements was in the acceptable level, whereas children showed a higher health risk than adults. There were 60.6% and 48.5% of studied sites in 2012 and 2016 respectively displaying the carcinogenic risk of As for children.