GEOMORPHIC AND LAND USE CONTROLS ON SEDIMENT YIELDS IN EASTERN USA

Dams and reservoirs are ubiquitous anthropogenic features across the United States. The Reservoir Sedimentation Database (ResSed), a subset of the National Inventory of Dams (NID), is a catalogue of reservoirs and relevant depositional data maintained by the USGS and the Subcommittee on Sedimentation. It contains detailed data from repeated surveys on 1,823 US reservoirs. The recent digitization of ResSed allows for rapid calculation of sedimentation rates over decadal timescales. This study seeks to gauge the relative importance of geomorphic and land use parameters in controlling sediment yield (Y) of watersheds in the eastern United States. We hypothesize that mean watershed slope, drainage area, and land-use metrics (e.g. percent impervious surface cover, percent agricultural land) should correlate with Y. We develop an ArcGIS-based model that delineates watersheds upstream of ResSed dams using 30-meter digital elevation data obtained from the USGS. Mean watershed slopes and drainage areas are computed for each reservoir. Calculated areas are used to determine Y and erosion rates for 500 watersheds east of the Mississippi River, based on assumed values for reservoir sediment and rock densities. Land use data (impervious surface area and categorical land cover) are obtained from the USGS and quantified within ResSed watersheds. Simple linear regressions, using 99 watersheds within Hydrologic Unit Code 02, show no correlations between geomorphic parameters or land use parameters and Y. However, some of the highest yields are associated with watersheds characterized by atypical land use or high impervious surface area, indicating that there are complex interactions between land use, geomorphic metrics, and Y. Further work on the project will include (1) analysis of ~400 more reservoirs to increase sample size and the power of statistical tests; (2) corrections of Y values for reservoir trap efficiency; (3) an exploration of how differing dam management strategies may influence sedimentation rates; and (4) implementation of statistical tests including multiple regression analyses to test interactions, and principal component analyses to better identify patterns in the dataset.