Estimating Nutrient Loads to Falls Lake, North Carolina from Streambank Erosion
Layla El-Khoury
North Carolina State University
Raleigh, NC
Authors: Layla El-Khoury, Barbara Doll, Jack-Kurki Fox, Dan Line, Krissy Hopkins, Karl Wegmann
As part of a large study by the NC Policy Collaboratory and funded by the NC General Assembly to analyze water quality and investigate nutrient management strategies for Falls Lake, we evaluated the potential nutrient inputs that could be arriving to the Lake from streambank erosion. Streambank erosion conditions ranging from stable to severely eroding and erosion rates were assessed and monitored throughout the watershed. Soil samples at cross-sections were collected and analyzed for nutrient content and bulk density. NCSU also measured flow, turbidity, total suspended solids (TSS) and total phosphorus (TP) at five subwatersheds to generate TSS and TP loads. Field-based assessments of streambank condition and erosion rates were combined with detailed geospatial mapping and modeling of land use and landforms to develop three models to 1) estimate potential locations where erosion was occurring, 2) the height of the streambank and 3) the rate of streambank erosion at 100 feet increments for all the streams in the Falls Lake watershed. Results of all models were combined with measured soil densities to generate a range of predicted sediment loading for each catchment in the watershed. Soil TN and TP concentrations were also used to generate predictions of nutrients for streambank erosion.
UNRBA and SPARROW both estimate approximately 30% of all sediments delivered to the lake are coming from unstable stream reaches and that these streams are contributing between 14.5 to 16% of the total TP load but only 0.8% of the TN load (UNRBA only). Our estimates of the presence of erosion combined with measured erosion rates, testing of sediment nutrient concentrations, and resulting modeling efforts indicate that the estimates from these models are reasonable. Further, by leveraging terrain data, our models provide desktop procedures for indicating locations where potential stream restoration and enhancement activities could be implemented to target reductions in turbidity, TSS and associated nutrients. Most of the catchments with the highest loads are closer to the outlet of the watershed. Maps indicating areas of predicted higher sediment and nutrient loading included in this report could be used to target areas for stream restoration and stabilization efforts.
About Layla El-Khoury
Layla El-Khoury is a PhD student in the Biological and Agricultural Engineering department at North Carolina State University (NCSU) working with Dr. Barbara Doll. She conducts research in improving methods for identifying, predicting, and quantifying streambank erosion to better target restoration efforts. She conducted research for her master of science degree, from NCSU, that focused on validating a USGS geospatial data layer where the results indicated it could be used to identity locations of erosion. This enables identification of potential stabilization/restoration sites prior to field assessments, maximizing the use of limited time and resources.
In her PhD research, Layla has examined different methods to quantify streambank erosion, including physical surveys, aerial imagery analysis, and geomorphic change detection using LiDAR data. She has also worked on developing a model to predict streambank erosion rates in the Ridge and Valley and Blue Ridge regions of Virginia as well as for the Falls Lake watershed in North Carolina. Her latest research has been focused on developing models to predict TN and TP loads from streambank erosion that go to Falls Lake.
Layla has also been a member of the State Dance Company (SDC) at NCSU since 2021. Through SDC she has had the opportunity to merge her research with dance. Through dance, Layla is developing new techniques for communicating STEM concepts and research to those outside the field, making education more accessible and engaging.
https://www.linkedin.com/in/layla-el-kho