Estimating Erosion Risks from Road Networks

Roads have been widely acknowledged as a major source of sediment to headwater streams.  While a number of tools have evolved to evaluate individual road segments, many questions exist about the interaction of road networks and stream networks at larger scales.  A number of ad-hoc hypotheses have emerged to use GIS layers to roughly estimate road contributions to sediment for watersheds, but validation of such models is difficult without detailed observations of as-built condition and performance of forest road drainage systems.  We present analysis of detailed measurements from three watershed scale road inventory projects using the Geomorphic Road Analysis and Inventory Package (GRAIP, http://www.fs.fed.us/GRAIP/).  The common outcome of all three analyses is that a small fraction of the road network and their associated drainage is responsible for the great majority of fine sediment inputs.  Stream crossings not only represented a dominant location of sediment delivery, but nearly a quarter had a high risk of becoming plugged and overtopped or a potential for flow diversion down the road after plugging.  In some locations, gully risks showed threshold-like behavior relative to a combination of contributing road segment length and the slope of the discharge hillslope.  Further analysis of the outcomes of several road restoration treatments has revealed several important lessons for design of treatments to reduce the impacts of forest roads: 1) decommissioning through simple ripping or tillage of the road surface may yield residual risks on either steep roads or steep hillslopes, 2) there may be negative consequences in attempting to reduce road-stream connectivity by placing additional drainage in steep landscapes, and 3) post-restoration channel readjustment at stream crossings may introduce substantial sediment loading if extra width is not allowed or any vertical adjustment is allowed.