Relation of surflcial erosion on hillslopes to profile geometry

Differences in erosion rates in relation to the geometry of hillslope profiles were hypothesized by R. E. Horton in 1945. Experimental testing of this hypothesis is described for natural hillslopes in western Colorado using a rainfall simulator to apply five "storms" each having an intensity of 1.85 inches per hour (47 millimeters per hour). Erosion from straight hillslope segments was about twice the amount from convex and concave segments along nine hillslope transects. Standard deviations of erosion measurements for each segment show that the greatest variation occurs in concave slope segments. This is probably because the concave segment is the only place on the profile where deposition occurs. Analysis of variance indicates that the amount of erosion is significant at the 1-percent level of confidence between convex and straight slope segments and at the 5-percent level of confidence between straight and convex slope segments.