Sediment Source Discrimination for a Series of Connected Check Dams in an Agricultural Catchment in the Chinese Mollisol Region

Abstract

The composite fingerprinting approach is valuable in sediment source identification at different temporal and spatial scales. However, few studies have estimated watershed sediment sources under a series of connected check dams. This is due to the complexity of particle sorting under flood discharge practices. Here, we attempt to identify sediment sources with three connected check dams in an agricultural Mollisol catchment in Northeast China based on the optimal fingerprints screened from different sized particles (i.e., <10, <63, and <1000 μm). The results showed consistent sediment contributions based on fine particle size ranges (i.e., <10 and <63 μm) but an inconsistent contribution based on a coarse particle size range (i.e., <1000 μm) in all check dams, regardless of the particle size or the organic corrections applied (Figure 1). Sediment deposited in check dams No. 1 and No. 2 was mainly from hillslopes (cropland) (>66.7%), while sediment in check dam No. 3 was from adjacent upstream check dam (i.e., No. 2; >60.5%) based on the optimal fingerprints screened from <10 and <63 μm particle size ranges. However, when fingerprints were screened from a coarse particle size (<1000 μm), sediment in all three check dams was mainly from gully slopes (>79.0%), regardless of the specific surface area (SSA) of particles or the soil organic carbon (SOC) correction factors used in the model except for SOC correction for check dam No. 2 (Figure 2). <fig><graphic xlink:href=23079_files/23079-00.jpg id=ID_99c1b84d-cf65-4e79-ad31-76ff8fbb8ca4></graphic></fig> Conversely, sediment was mainly derived from hillslopes when fingerprint values were corrected using SSA and SOC, except that no optimal fingerprints were selected. This indicated the feasibility of sediment source discrimination based on the composite fingerprinting approach for sediment source identification in catchments with a series of connected check dams under the condition that particle fractions impacted by check dams are relatively weak. However, optimal fingerprints screened from a relatively wide range of coarse particle caused divergent results regardless of the particle size or the organic correction factors used in model estimation or in fingerprint values. Further studies are necessary to explore the reliability of the composite fingerprinting approach for sediment source tracing under a series of connected check dams, where coarse sediment is selectively deposited and fine sediment is massively discharged under flood discharge practices.<underline> </underline>