The Impact Of Water Management On Cadmium And Arsenic Accumulation In Rice Grain: Implications For A Regional Prediction Model For Cadmium In Rice Grain

In recent years,the rapid industrialization process in China has led to the pollution of some farmland soils by heavy metals and metalloids.According to the"Chinese National Soil Pollution Survey Bulletin",the amount of soils polluted by cadmium（Cd）and arsenic（As）are more than any other inorganic elements.Both Cd and As are easily translocated and accumulated in rice grains,threateningthe safety of agricultural products and human health.Due to the contrary biogeochemical cycles of Cd and As,it is difficult to effectively reduce the accumulation of cadmium and arsenic in rice grains simultaneously using the existing agricultural engineering methods.Meanwhile,many studies have shown that the correlation between rice grain Cd and soil Cd is poor.As a result,there are cases that soil with Cd below the threshold value produced rice with Cd above the threshold value or soil with Cd above the threshold value produced rice with Cd below the threshold value,that will cause inconvenience in agricultural and environmental supervision.In order to solve the two existing main problems above,this study firstly investigated the key rice growth period for the accumulation of Cd and As in rice grains by conducting pot experiments,and proposed a water management strategy to simultaneously reduce Cd and As accumulation in rice grains by taking advantage of the difference in the key stages between Cd and As accumulation.Meanwhile,a soil Eh continuous monitoring experiment was designed to determine whether the soil redox potential（Eh）is the key factor that affect the accumulation of Cd and As in rice grain.Furthermore,a field survey was conducted in the typical Cd contaminated region,Xiangtan,Hunan Provence and the accuracy of rice grain Cd prediction in a large scale was assessed through incorporating fresh soil moisture content as a parameter.The specific results are as follows:（1）Analysis of the key periods of Cd and As accumulation in rice grains.Both Cd and As are loaded into rice grain via two pathways:i)root uptake from the soil and then translocation to the grain,and ii)remobilization of Cd and As previously accumulated within the vegetative tissues to the grain.However,the relative contributions of the two pathways are not well understood in soil-grown rice plants.In this experiment,we used eight different water management regimes applied at different growth periods to manipulate the concentrations of Cd and As in porewater and then established a mathematical model to estimate the relative importance of the two pathways.Different water management regimes had dramatic and opposite effects on the solubility of Cd and As in soil,and their subsequent accumulation in both straw and grain.Water management applied at different growth periods had markedly differentiated different impacts on the grain Cd and As accumulation.Water management during grain filling had a much greater impact on grain Cd than on grain As concentration,whereas water treatment during the vegetative growth stage had a larger effect on grain As concentration.Under the normal water management practice（i.e.flooding during the vegetative stage followed by drainage during grain filling）,grain filling is the key period for the accumulation of Cd in the grain,with 98%of the grain Cd being derived from root uptake during this period and the contribution of remobilization being very limited.In contrast,95%of the grain As was remobilized from that accumulated within the plant prior to the grain filling,with the tillering,jointing,and heading periods contributing almost equally to the grain As,whereas root uptake during grain filling contributed minor.These differences can be harnessed to design a segmented water management strategy to control grain Cd and As accumulation simultaneously.（2）Study on the relationship between enrichment factor of Cd and As in rice grains and soil average Eh.In order to reveal the relationship between Cd and As in rice grains and soil redox potential before and after rice filling,a pot experiments was conducted in an acidic paddy soil collected from Beishan Town,Changsha,Hunan.Five different water management regimes were set up in this pot experiments,including T1 continuously flooded,T2 continuously aerobic,T3 aerobic after heading,T4 aerobic one week after heading,and T5 circular flooded and drained.The results showed that soil Eh,p H and As/Cd content in soil pore water changed differently under different water management regimes.Similarly,water management also significantly affected the concentrations of Cd and As in rice straw and grains,and the Cd concentration in grains(2.44 mg kg^-1)of T3（aerobic after heading）was 46.7%lower than that of T4（aerobic one week after heading）,suggesting that delayed drainage during grain filling period could effectively reduce the accumulation of Cd in grains.By calculating the average Eh of soils under different treatments during the vegetative growth stage and filling stage of rice,we concluded that the transport coefficient of As in rice grains（i.e.,the ratio of As concentration in grains to As concentration in soil,representing the capacity of grains to absorb arsenic）presented a linear regression with the average Eh value of soil during the vegetative growth stage of rice.However,the transport coefficient of Cd in rice grains was linear regression to soil average Eh during rice filling period.The results show that the grain filling stage was the key period for Cd uptake and accumulation in rice grains.（3）Investigation and analysis of the causes of Cd pollution in soil-rice grains in Xiangtan County.In this field survey,Xiangtan County,a typical Cd contaminated area in Hunan Provence,southern China,was selected.201 and 200 pairs of soil-rice corresponding samples were collected and analized in the late rice filling stage（about 1 week before rice harvest）in 2019 and 2020,respectively.The results showed that over 81.5%and 85.0%soil samples had Cd concentration above the Chinese national soil standard in 2019 and 2020,respectively.However,only 37.8%and 36.5%rice samples in 2019 and 2020 respectively had the Cd concentration above the Chinese national food safety standard.The sampling point with soil Cd above the risk screening value but grain Cd below the threshold was 46.7%and 52.0%for 2019 and 2020,respectively.There were no correlations between soil Cd and rice Cd,which could lead to an extremely high error rate evaluating rice Cd.There was a significant correlation between soil Cd,lead（Pb）and zinc（Zn）based on analysis of correlation and principal component.Meanwhile,a strong spatial correlation was found that concentrations of soil Cd,Pb and Zn was interrelate with the distance from Xiangjiang River,indicating that the Cd contribution was probably related to irrigation from Xiangjiang river and the historical Pb and Zn mining activities.Compared with the previous researches in this area,it was found that soil Cd did not vary significantly,but rice grain Cd decreased significantly,that was probably caused by several effective treatments to limit contamination of rice Cd,such as increase of soil p H,replacement of low Cd accumulation rice varieties and fallow of farmland with high soil Cd.These treatments initiated by local government has greatly alleviated the Cd contamination of rice grain in Xiangtan.（4）Regional prediction model of Cd accumulation in rice grains.Considering that soil moisture has an important impact on rice grain Cd during grain filling stage,a prediction model of rice grain Cd was established with soil properties.The results showed that there was no significant correlation between rice grain Cd and soil Cd（R²=0.14）,even adding other parameters such as soil p H and soil manganese（Mn）could not significantly improve the R²,result in a high error rate（52.6%）evaluating rice Cd.The performance of rice grain Cd prediction increased（R²=0.58）,which was established by extractable soil Cd and Mn（extracted by hydroxylamine hydrochloride）,along with soil p H and moisture contents.Further more,machine learning（support vector machine）method could enhance the accuracy of prediction（R²=0.88 for test set）,and the model could be used to predict regional rice Cd distributions.Based on the prediction model,three scenarios were simulated to evaluate the impact on rice Cd,and the results showed that adjusting soil p H to 6.5 would be the best treatment because of the lowest cost and best performance of Cd reduction.This study identifies the critical period during which the majority of Cd and As accumulates in rice.According to this result we propose a segmented water management strategy to simultaneously control Cd and As accumulation in rice grains from the perspective of agricultural engineering.A county scale prediction model of Cd in rice grain was established and the accumulation of Cd in rice grains under several scenarios was evaluated.These results provide theoretical basis and technical support for reducing Cd and As accumulation in rice grian to ensure safe utilization of Cd-contaminated soils.