Near-Channel Versus Watershed Controls on Sediment Rating Curves

Abstract

Predicting riverine suspended sediment flux is a fundamental problem in geomorphology, with important implications for water quality, land and water resource management, and aquatic ecosystem health. To advance understanding, we evaluated environmental and landscape factors that influence sediment rating curves (SRCs). We generated SRCs with recent total suspended solids (TSSs) and discharge data from 45 gages on 36 rivers throughout the state of Minnesota, USA. Watersheds range from 32 to 14,600 km2 and represent distinct settings regarding topography, land cover, and geologic history. Rivers exhibited three distinct SRC shapes: simple power functions, threshold power functions, and peaked or negative-slope functions. We computed SRC exponents and coefficients (describing the steepness of the relation and the TSS concentration at median flows, respectively). In addition to quantifying watershed topography, climate/hydrology, geology, soil type, and land cover, we used lidar topography to characterize the near-channel environment upstream of gages. We used random forest models to analyze relations between SRC parameters and attributes of the watershed and the near-channel environment. The models correctly classify 78% of SRC shapes and explain 37%–60% of variance in SRC parameters. We find that SRC steepness (exponent) is strongly related to near-channel morphological characteristics including near-channel relief, channel gradient, and presence of lakes along the local channel network, but not to land use. In contrast, land use influences TSS concentrations at moderate and low flow. These findings suggest that the near-channel environment controls changes in TSS as flows increase, whereas land use drives median and low flow TSS conditions.