| Heavy metals contamination of arable soil is of concern due to the adverse effects it causes on agro-food products safety and subsequent human health due to transmission in the food chain.In order to control heavy metal pollution,guarantee agro-food products safety and reduce hazard probability of human health,study on the source of the contamination and bio-availability of the heavy metals in soil as well as associated human health risk is of great significance.The current Soil Environmental Quality Standard of China(GB 15618-1995)could lead to over-estimate or under-estimate of metal contamination and the associated risk for a particular crop,which needs improvement and upgrading.Based on pot experiments,many previous researches have detailedly studied the behavior of metals in soil-crop systems.Many clear models of predicting metal transfer from soil to plants have been obtained through multiple regression analyses and then were used to derive soil metal quality benchmark for food safety.However,such clear regression-based models are generally hard to establish under true field conditions due to the complexity of the interactions between soil,metals,and plants.To overcome this limitation,the potential of modern statistical techniques to predict metal uptake by plant tissues should be explored,which have,unfortunately,been rarely studied.Derivation of health risk-based soil quality guidelines/standards has been an international trend.So far,limited research work has been done in China in relation to health risk assessment of polluted soil.Some previous studies have conducted human health risk evaluations of heavy metals through 'soil-crop-human' exposure pathway,but little research on'soil-human' exposure route has been performed.Therefore,it is very necessary to study soil metal benchmark based on multi-pathway health risk assessments,and on this basis gradually improve the national soil environmental quality standard.In the present study,we sampled paired soil and crop samples from a large number of field sites that encompassed a wide range of soil metal levels within Anhui Province.Our objective was to investigate enrichment characteristics of heavy metals in soil-crop systems and uncover their likely influence factors,with the aim of obtaining algorithms that could be used to predict metal levels in rice grains under field conditions.We also conducted multi-pathway health risk assessment of heavy metals in soils.This allowed us to derive soil metal benchmark values based on food safety standard and health risk assessment,which provide technology support to determine soil quality benchmark using the data from field investigation.The main findings are as follows:1.Mining was proved to be one of the most significant sources of metal contamination of cultivated soils.It was estimated that the average contribution of anthropogenic Cu?Zn?Pb?Cd and As caused by mining activities in the mining associated soils was 46.2%?45.1%?50.4%?64.4%and 42.7%,respectively.The proposed classification and regression tree method(CART)is suitable to estimate the classes of soil Cu,Pb and Cd with a high accuracy of>80%.There was no significant difference of Ni concentrations between the mining associated soils with the non-mining associated soils,implying mostly of natural origins.2.Significant differences existed in heavy metal concentrations in different crop tissues.The mean accumulation of the studied heavy metals(Ni,Cu,Zn,Pb,Cd,As and Hg)in rice decreased in the order straw>grain.Ni and Zn showed higher accumulation in rape seed,whereas other heavy metals had higher accumulation in rape straw.Heavy metals contents of the different crops also showed significant differences.The mean content of As,Cd and Hg in rice straws and grains was higher than that in the same tissue of rape,whereas Ni showed the opposite trend.The content of Zn and Pb in rice straw was higher than that in rape straw,but the content in polished rice was lower than that in rapeseed.The bioconcentration factor(BCF)of heavy metal varied widely in polished rice and rapeseed,and the BCF of Cd and Zn was much higher than other heavy metals.3.The concentrations of Cu,Pb and Cd in polished rice and Cu,Zn,Cd and As in rapeseed were found to be significantly positively correlated with that in soils.However,no correlation was found between other heavy metals in crop grains and in soils.The result indicated that apart from the concentrations of the heavy metal themselves,other factors also exhibit noticeable effects on grain metal accumulation including soil pH,organic carbon content,coexisting elements content and atmospheric deposition.Cd-Zn antagonistic interactions played an important role in determining grain Cd accumulation.The total accumulation of Cd in grains decreased with growing Zn/Cd ratio in soil,showing a generally negative exponential trend of the grain Cd content with the increase of the soil Zn/Cd ratio.The reason that some high Cd level mining-associated soils did not lead to excessive levels of Cd in polished rice can be largely attributed to its high Zn/Cd ratio.4.Multiple linear regression analysis showed that the combination of the soil pH and total soil Cd was a reasonable predictor of Cd in polished rice.Based on food safety standard,the derived soil Cd benchmark was 1.99,2.98,4.45 and 6.65 mg kg-1 for soils with a pH of 5,6,7 and 8,respectively,all higher than the Grade II guideline of Environmental Quality Standard for Soils(GB 15618-1995).To overcome the technological limitation of linear regression analysis,the classification and regression tree method(CART)was proposed to predict Cd pollution levels using both categorical and numerical independent variables.This tree-fitting method had high accuracy up to 92.8%.Based on food safety standard,the derived soil Cd benchmark from CART model was 0.61 mg kg-1.Moreover,CART model further uncovered that soil Cd content,soil Zn/Cd ratio and type of parent soil material were the most significant variables contributing to Cd content in polished rice.5.Carcinogenic and non-carcinogenic risks of soil heavy metals exceeded the acceptable levels for local residents.The estimated risk mainly came from the pathway involving the ingestion of locally produced rice.For the inhalation pathway,the exposure risks of heave metal can be ignored.The highest non-carcinogenic risk happened to the local children.In the view of human health risk,As and Cd were selected as the priority control heavy metals.The potentially carcinogenic and non-carcinogenic risks of As were prevalent in the study area,while non-carcinogenic risk of Cd mainly occurred in the polluted mining areas.6.Based on human health risk assessment,the derived soil Cd benchmark value was 0.57 mg kg-1.The Monte Carlo technique was used to obtain the probabilistic distribution of soil Cd benchmark based on the risk of sensitive exposure population through rice consumption.The derived soil Cd benchmark value was 0.24,0.41 and 0.59 mg kg-1 at the risk probability levels of 10%,20%and 30%,respectively.According to the proportion of rice ingestion risk to total non-carcinogenic risk(94%),soil Cd benchmark value based on total risk assessment was calculated,being 0.23,0.39 and 0.55 mg kg-1 at the risk probability levels of 10%,20%and 30%,respectively. |
PDF Full Text Request