干旱区生态环境知识资源中心

Fine sediment inputs into river systems can be a major source of nutrients and heavy metals and have a strong impact on water quality and ecosystem functions of rivers and lakes, including those in semiarid regions. However, little is known to date about the spatial distribution of sediment sources in most large scale river basins in Central Asia. Accordingly, a sediment source fingerprinting technique was used to assess the spatial sources of fine-grained (<10 mu m) sediment in the 15 000 km(2) Kharaa River basin in northern Mongolia. Variation in geochemical composition (e.g. in Ti, Sn, Mo, Mn, As, Sr, B, U, Ca and Sb) was used for sediment source discrimination with geochemical composite fingerprints based on Genetic Algorithm (GA)-driven Discriminant Function Analysis, the Kruskal-Wallis H-test and Principal Component Analysis. All composite fingerprints yielded a satisfactory GOF (>0.97) and were subsequently used for numerical mass balance modelling with uncertainty analysis. The contributions of the individual sub-catchment spatial sediment sources varied from 6.4% (the headwater sub-catchment of Sugnugur Gol) to 36.2% (the Kharaa II sub-catchment in the middle reaches of the study basin), generally showing higher contributions from the sub-catchments in the middle, rather than the upstream, portions of the study area. The importance of river bank erosion is shown to increase from upstream to midstream tributaries. The source tracing procedure provides results in reasonable accordance with previous findings in the study region and demonstrates the applicability and associated uncertainties of the approach for fine-grained sediment source investigation in large scale semi-arid catchments. (C) 2015 Elsevier B.V. All rights reserved.