Determination of site-specific natural background values for soil potentially toxic elements supported by partial...

The natural background value (NBV) is crucial in determining anthropogenic inputs of soil potentially toxic elements (PTEs). However, due to the fact that almost no soil is completely unaffected by human activities, and NBV should be site-specific and spatially distinct, the determination of NBV, especially at site scale, is often challenging and highlights the urgent need for new methods. Therefore, this study attempted to establish a method to determine site-specific NBVs, and accurately identify anthropogenic inputs of soil PTEs. To achieve this, 1330 surface soil samples were collected from Jieyang for chemical composition analysis, and regression relationships supported by partial least squares regression (PLSR) method were established between eight PTEs and twenty-five chemical compositions. Results showed that PLSR model had excellent predictive ability, obtaining 1330 site-specific NBVs for each of these eight PTEs. Average proportions of anthropogenic inputs for Zn, Cr, Cu, Hg, Cd, As, Ni, and Pb were 7.95 %, 11.56 %, 25.63 %, 37.79 %, 17.35 %, 67.77 %, 6.82 %, and 8.09 %, respectively, and 90.76 % of sites had medium-to-high environmental capacity. Moreover, the study area was at moderate ecological risk level (125.54), with Hg and Cd being the main risk factors. However, 6.84 % and 0.45 % sample sites reached higher and high ecological risk levels, respectively, which needed further investigation and monitoring. These findings indicate that PLSR model can effectively improve the measurement accuracy of NBVs of soil PTEs, and realize accurate point-to-point environmental assessment.