back to TMDL Implementation PlansCompleted Projects

Incorporating Stream Erosion and Restoration into Watershed

Principal Investigator: Tess Wynn. Other project researchers: Gene Yagow and Brian Benham, Biological Systems Engineering Department, Virginia Polytechnic Institute and State University, Virginia


The final project report is available on our website for download:

Modeling Channel Erosion at the Watershed Scale: Model Review and Case Study.  Center for TMDL and Watershed Studies. BSE Document Number 2006-0009.

While considerable effort has been directed toward reducing erosion from agricultural and urban lands, a major source of sediment - stream bank erosion - has received little attention. Studies have shown that sediment from stream banks can account for as much as 85% of watershed sediment yields and bank retreat rates of 1.5 – 1,100 m/year have been documented. In watersheds where stream bank erosion is an issue, the quantity of sediment contributed by stream bank erosion and the potential effects of stream restoration efforts on sediment yield need to be assessed to develop and implement accurate, effective TMDLs. 

To improve the accuracy and efficiency of watershed-scale sediment models, this project will develop a guidance document and web site on stream bank erosion and restoration modeling by gathering information on existing methods and parameters used to calculate stream bank erosion and to evaluate the impact of stream restoration. A case study for an existing impaired stream will compare the input requirements and output sediment load estimates of a process-based model (AGNPS-CONCEPTS) to those of less detailed models (GWLF and SWAT). This project will also identify data sources and test methods for model parameter estimation. 

Stream bank erosion is a complex cyclic process involving erosion of the bank toe followed by geotechnical bank failure. Methods used to model bank erosion vary widely. The relatively simple watershed-scale, lumped-parameter GWLF model, calculates channel and bank erosion based on an empirical relationship that is a function of stream discharge, percent of developed land in the watershed, watershed animal density, area-weighted average curve number, area-weighted average K-factor from the USLE equation, and average watershed slope. Another commonly used watershed-scale model, SWAT, is more process-based and incorporates sediment transport capacity, the erodibility of the channel materials, and vegetative cover. Recently, the USDA Agricultural Research Service (ARS) developed a reach-scale model (CONCEPTS) that calculates toe scour, bank failure, and sediment transport to determine changes in stream form. The combination of the reach-scale CONCEPTS and the watershed-scale AGNPS model provides a detailed evaluation of stream bank erosion, including the effects of riparian vegetation on bank stability. AGNPS-CONCEPTS is currently the most advanced model of stream bank erosion and, as such, requires the greatest amount of input data.

While these models vary in complexity, guidance in evaluating related model parameters, particularly stream bed and bank erodibility, is commonly absent. The only process-oriented methods available to determine erodibility are laboratory flume studies or in situ jet testing. Little information is available on how to incorporate the influence of increased riparian vegetation from riparian buffers and stream restoration in model parameters. If water quality data are available, model parameters such as channel erodibility are frequently calibrated, rather than being evaluated from field measurements.