Evaluation and development of predictive streambank erosion curves for northeast Kansas using Rosgen's "bancs" methodology



Journal Title

Journal ISSN

Volume Title


Kansas State University


The original purpose of this investigation was to develop streambank erosion prediction curves for Northeast Kansas streams. Rosgen's (2001, 2006) methods were employed and eighteen study banks were measured and monitored over a four-year period, summer 2007 through summer 2010. At each study bank, a toe pin and two to three bank pins were set at a recorded longitudinal profile station of the stream. Vertical and horizontal measures from the toe pin to the bank face were taken each summer, 2007 as the baseline measure and 2008 - 2010 as bank change years. Bank profiles were overlaid to gain insight into bank area lost or gained due to erosional or depositional processes. A Bank Erosion Hazard Index (BEHI) and Near Bank Stress (NBS) combination rating was assessed and calculated for each study bank during the initial survey of 2007. The streambanks experienced varied erosion rates for similar BEHI/NBS combinations producing R2 values from 0.43 as the High/Very High BEHI rating and 0.80 as the Moderate BEHI rating. In addition, Moderate BEHI ratings provided higher erosion rates than the High/Very High BEHI rating and curves intersected at lower NBS ratings, suggesting a discrepancy in the fit of the model used in the Northeast Kansas region and conditions. In this light, modification of the BEHI model was evaluated and variables were assessed in the model for additional influence exerted in the Northeast Kansas region. Vegetation seemed to provide the most influence to bank resistance and was more closely evaluated. Banks with and without woody riparian vegetation were then plotted against BEHI and NBS values, as banks lacking woody vegetation eroded at higher rates. This study's findings can allow us to calibrate the BEHI model according to woody vegetation occurrence levels along streambanks in the Black Vermillion watershed. Modifications regarding vegetation occurrence of the BEHI model was completed and the results of these modifications generated R2 values of 0.78 for High/Very High BEHI and 0.82 for Moderate BEHI ratings. High/Very High ratings provided higher predicted erosion rates than Moderate ratings, while the curve slopes did not intersect at lower NBS ratings.



Bank erosion, Predictive erosion curves, Rosgen bancs, Bank erosion hazard index

Graduation Month



Doctor of Philosophy


Environmental Design and Planning Program

Major Professor

Timothy D. Keane