Border polje soil in Southwest China originated from non-karst areas: Evidence from geochemical fingerprint data

The purpose of this study was to provide scientific evidence to inform decisions on sediment mitigation strategies. A sediment fingerprinting approach was applied to identify soil sources for a border polje (BP) between karst (K) and non-karst (NK) areas in Southwest China. Three composite geochemical fingerprints, composed of basic, conservative, and high variability fingerprints, were used to quantify source contributions from two original groups (geology and land use), and three new classifications (two-, three-, and four-clusters) obtained using the k-means cluster analysis. The bi-plot, range test, tracer variation rate, and virtual mixture were used to test the reliability of each model configuration. The five classifications enabled the identification of > 85.7 % of the source samples. The two-cluster classification was consistent with the geology; the three-cluster divided the K into slope and depression, and the four-cluster further partitioned the NK into hydro-fluctuation belt of reservoirs (HFLRs) and highlands. The goodness of fit for these models was generally > 75 % and uncertainties were < 7.1 %. The HFLRs contributed > 90 % of soil parent material in the BP. The contribution of HL to the dry land near the NK in the BP was lower than that of HFLRs. Depressions contributed more than half to the dry land near the foothills of the limestone in the BP rim, followed by HFLRs. These findings highlight that the fluviatile sediments from NK areas significantly contribute to the BP. Hence, the identification of soil sources in the BP requires an evaluation from the perspective of hydrology, geology, and anthropogenic activities.