SPECTRUM OF BANK EROSION ALONG POWDER RIVER
Annual bank erosion was temporally and spatially episodic. Therefore, for statistical purposes plus the lack of replicates at any annual peak daily discharge, the erosion data were grouped into four discharges classes (22.7-50, 50-100, 100-200, and 200-800 m 3 s-1). Of the 367 measurements, 71 % represented bank erosion, and 29 % zero erosion. Zero erosion was characteristic of all discharge classes suggesting that discharge was not the only controlling factor. The lowest two discharge classes were not statistical different (p=0.6), whereas the larger two classes were statistical different from themselves and from the lower two classes (p<0.008). Within each discharge class, the bank erosion had a weak dependence on R/w (R2<0.35), which is commonly accepted as a major controlling factor of bank erosion. South-facing banks thaw out sooner than banks with other aspects and are possibly more vulnerable to erosion. Therefore, the data were regrouped into four aspect classes (north, east, south, and west). For discharge classes >50 m 3 s-1, south-facing bank erosion was significantly greater (p<0.034) than erosion on north-facing and east-facing banks. This tendency of increased erosion for south-facing banks may explain previous observations in the literature of bend migration along the lower reach of Powder River that consists predominately of downriver translation and negligible lateral migration.