HYDROLOGICAL PROCESSES Hydrol. Process. 17, 1887–1897 (2003) Published online 1 April 2003 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/hyp.1216 Statistical descriptions of channel networks and their shapes on non-vegetated hillslopes in Kemmerer, Wyoming D. A. Raff, 1 * J. L. Smith 2 and M. J. Trlica 3 1 Department of Civil Engineering, Colorado State University, Fort Collins, CO 80523, USA 2 Department of Civil and Architectural Engineering, University of Wyoming, Laramie, WY, USA 3 Department of Rangeland Ecosystem Science, Colorado State University, Fort Collins, CO 80523, USA Abstract: Erosion prediction tools are essential for many different aspects of land management. In this paper we examine the applicability of scaling relationships developed for river networks for the description of hillslope erosion channel networks. Eight erosion channel networks were studied on mine spoil sites in Kemmerer, Wyoming. We analysed these networks using well-known geomorphological relationships such as Horton’s laws and Melton’s law. An analysis of the erosion channel hydraulic geometry described by a series of power functions was also conducted. We found that the network geometry resembles larger river networks, and that the cross-sectional shapes of individual erosion channels exhibit statistical characteristics of their larger river counterparts. Copyright  2003 John Wiley & Sons, Ltd. KEY WORDS hillslope erosion; sediment yield; overland flow; channel network INTRODUCTION Our scientific ability to predict hillslope erosion is severely impaired by our ability to interpret the spatially and temporally distributed processes involved. One specific gap in scientific knowledge is a description of the scaling characteristics of erosion, specifically erosion channels. Channel development on actively eroding hillslopes has many implications for sediment detachment and delivery (Nearing et al., 1997). Linse et al. (2001) suggested that the degree of channel network development on rangelands strongly influences sediment yield. Goldman et al. (1986) described erosion through five separate processes: splash, sheet, rill, gully and channel erosion. The definition of a gully relies on whether a tractor can drive over a channel. However, there is no direct physical relationship between the diameter of a tractor tyre and the erosion potential of a hillslope. In fact, recent research has shown that many hydraulic processes that occur in rills are the same as those found in gullies (Bennett et al., 2000). A recent erosion prediction model, the Water Erosion Prediction Project (WEPP) groups these into rill and interrill processes (Weltz et al., 1998). Channelized flow paths (erosion channels) do not occur equally spaced and parallel across a hillslope, as the WEPP model suggests. The pathways by which drainage networks develop on a hillslope (e.g. Glockian or Hortonian) may have an immature stage where channels appear to be parallel and equally spaced, but as time progresses this network is dynamic and eventually results in a dendritic structure characteristic of large river systems (Glock, 1931). We hypothesize that channel networks existing at the hillslope scale are scaled dendritic structures that resemble larger river networks and are interdependent of the interchannel area, just as river network response and development is interdependent with hillslopes. * Correspondence to: D. A. Raff, Department of Civil Engineering, Colorado State University, Fort Collins, CO 80523, USA. E-mail: raff@engr.colostate.edu Received 26 November 2001 Copyright  2003 John Wiley & Sons, Ltd. Accepted 3 July 2002