Spatial Variation Of Combined Heavy Metal Pollution In Red Soils Of Hill Region

A case study area around the Jiangxi Copper Smeltery in red soil hill region with an area about 7.7 km2 was selected in Guixi county, Jiangxi province, P. R. China. The spatial variability of combined heavy metal pollution in soils was analyzed by a geostatistical method based on GIS.97 soil samples and 63 plant samples was collected by a grid from 50m to 100m with DGPS, and soil in three layers (0-15cm depth, 20~40cm depth and 40~60cm depth) was collected at each site. The total heavy metal (Cu, Cd, Zn, Pb and Cr) contents and other properties (pH, organic matter Cl- and SO42-) in soils, and Cu and Cd contents in rice straw and brown rice were analysed.The analysis method sequence was as follows: a) explore the data's distribution pattern and spatial trend, then transform the date to get rid of the trend; b) calculate the semivariogram, select the best model and regulate the appropriate parameters; c) select the best kriging methods base on cross validation; d) interpolate the soil heavy metal contents and draw the heavy metal pollution map, then analyze the spatial distribution characteristics and their controlling factors, and the above method based on the geostatistical module in ArcGIS can exactly predict the spatial distribution pattern of combined heavy metal pollution in the hill region.According to the result of spatial analysis, the nugget-to-sill ratio showed a strong spatial dependence for soil pH, a poor spatial dependence for soil Cr, and a moderate one for other properties in surface 0-15 cm soil, their range were between 517 to 2134m; in 20~40cm depth soil, there was a strong spatial dependence for soil Pb, a poor spatial dependence for soil Zn and organic matter, and a moderate one for other properties, their range were between 218 to 345m; in 40~60cm depth soil, there was a strong spatial dependence for soil Cd and Pb, a poor one for soil Cu and a moderate one for other properties, there range were between 355 to 2015m. There was a moderate spatial dependence for the Cu and Cd in both paddy straws and brown rice, and their range was between 592 to 2015m. There was a negative relationship between one element's spatial dependence and its range, the element had a strong spatial dependence would with a small range and a poor spatial dependence with a large range.The result showed that there is a combined Copper and Cadmium pollution in the case study area. In surface soil, the Cu's concentration was between 40.6 to 565 mg-kg-1, and about 99% of which was above the environmental threshold and polluted; the Cd's concentration was between 0.36 to 2.85 mg -kg"1, all was above the environmental threshold and polluted, the Zn was partial polluted and the Pb and Cr was never polluted. In the deeper layer, the heavy metal elements expect Cd were never polluted and their concentration were just around the environmental background. The Cd's concentration was 95% above the environmental threshold in 20~40 cm depth soil and mostly was between 0.3 to 1.2 mg kg"1, in 40~60cm depth soil, the concentration of Cd was still 95% above the threshold and mostly between 0.3 to 1.0 mg-kg"1. The Cu concentration ranges from 28.0 to 488.9 mg kg"1 in rice straw and from 3.87 to 22.6 mg/kg in brown rice; with an average of 146.1 mg-kg"1 and 10.9 mg-kg"1 respectively. The Cd concentration was between 0.35 to 28.1 mg-kg'1 in rice straw and 0.02 to 5.11 mg -kg"1 in brown rice, with an average of 6.439 mg'kg"1 and 1.72 mg-kg"1 respectively. There was 51% of Cu contents and 83% of Cd contents in the brown rice above the national threshold of food quality. Cu was mostly concentrated in the surface layer (0-15cm) and little moved down to the deeper layers. However, Cd moved more easily than Cu and caused soil polluted in three layers. The concentrations of Cu and Cd were much higher in rice straw than that in brown rice.The ordinary cokriging interpolation of soil Cu with Cd contents (the correlation coefficient between predicted and measured value, R = 0.810) was more precise than the ordinary kriging interpolation (R =0.