Assessment of river bank erosion in Southern Minnesota rivers post European settlement

River bank erosion is one of the major sources of sediment for many rivers around the world. With the current emphasis on developing total maximum daily loads (TMDLs) for impaired waters in the United States, there is heightened interest in quantifying background sediment levels in rivers. In this study, we assessed variations in river bank erosion over time using a combination of 1855 Public Land Survey System plats, aerial photographs from 1938 to 2009, and light detection and ranging (lidar) data from 2005 to 2009 for sediment impaired rivers in Southern Minnesota. Results showed that bank erosion was episodic, making comparisons of erosion rates from dissimilar time intervals unreliable. For comparable time intervals, average river bank retreat rates (0.51myr^-1 from 1855 to 1938 vs. 0.37myr^-1 from 1938 to 2009) were statistically similar (t=2.13, p=0.14) suggesting that bank erosion rates have remained stable since European settlement. Comparisons over shorter time intervals of 1938-1971 and 1971-2009 also showed similar statistical trends (t=0.76, p=0.45) with average river bank retreat rates of 0.57 and 0.50myr^-1, respectively. However, additional 145 observations of bank retreat were found in the period 1971-2009 relative to 1938-1971, indicating that the number of actively eroding river banks may have increased over time. Contrary to assumptions made in the literature, bank erosion measurements using lidar data showed a poor relationship (r²=0.01 to 0.36) with river bank physical features (face area, inclined surface area, length, slope, height, and aspect), thus suggesting that extrapolating a limited number of bank erosion observations to the whole length of a river will lead to erroneous predictions. This lack of relationship was expected considering that most of these bank physical features do not fully represent bank erosion processes such as seepage, freeze-thaw, river migration, under cutting and sapping. We conclude that, in assessing conservation measures or developing TMDLs to manage river sediment loads, (1) background levels of suspended sediments from river bank erosion should be established using comparable time intervals, and (2) up scaling of discrete volume loss measurements to an entire reach should be avoided.