A survey on soil samples was conducted to study the heavy metal pollutions and their potential sources in Wuqing District, Tianjin, China. A total of 578 topsoil samples were collected and the concentrations of eight heavy metals, namely, Cd, Pb, Cu, Zn, Cr, Ni, As (metalloid) and Hg were analyzed. A summary of descriptive statistics, principal component analysis, geostatistical analysis and stochastic forest regression models were applied to study the spatial and temporal variation and identify proportional contribution from either natural or anthropogenic sources for the eight heavy metals in topsoils of the study region. The results indicated that the average concentrations of all the heavy metals except for Cr in the topsoils exceeded their corresponding natural-background values. As, Ni and Cr were mainly contributed by natural sources (i.e., soil parent materials). Cu and Zn originated from both the soil parent materials and sewage irrigation. Pb and Cd originated mainly from non-point source pollution and partially from point source. Hg originated from sewage irrigation. It was proved that combination of multi-technologies provides an effective way to delineate multiple heavy metal pollution sources.
Keywords: geostatistics; heavy metals; soil; source apportionment; stochastic forest regression.