Surface sediment samples concurrently collected in the catchment of the Ningxia-Inner Mongolian reach of Yellow River (NIMYR) were analyzed to determine the phosphorus and correlated physiochemical properties of sediments. Samples were obtained from three main areas: the riverbed surface sediment (RSS), the floodplain surface sediment (FSS) and the desert surface sediment (DSS). The sequence of phosphorus-contamination level in this catchment, determined by a cumulative distribution function and multivariate statistical analysis, were FSS>RSS>DSS. Moreover, because of the impacts of anthropogenic activities, the sampling site with the highest and lowest phosphorus concentration (mgkg-1) of this catchment appeared in RSS (749.40) and DSS (200.10) respectively. In addition, this is the first study to present a qualification of the effect of the sediment's physicochemical properties on phosphorus by the multivariate regression tree analysis. Co-precipitation of phosphate with calcite [i], phosphate absorbed onto ferric hydroxides [ii] and grain-size effect [iii] were the three main mechanisms for phosphorus distribution in the sediment of NIMYR. The contributions of these three mechanisms to RSS and FSS were, respectively, [i]>[iii]>[ii] and [i]>[ii]>[iii]. The heaviest phosphorus-contaminated group of RSS was primarily controlled by [i], whereas the heaviest phosphorus-contaminated group of FSS was controlled by [i] and [ii], indicating that the FSS had a higher potential risk of releasing phosphorus from the sediment to overlying water.
Keywords: Multivariate regression tree analysis; Phosphorus; Surface sediment; Yellow river.
Copyright © 2016 Elsevier B.V. All rights reserved.