Previous studies within forested watersheds in northern California have shown that the location and condition of the road within a watershed can have a significant effect on the amount of erosion associate with the road system. As a result many forests are developing road management plans to efficiently reduce road related sediment. A road sedimentation model, SEDMODL, was applied to the Caspar Creek watershed on the Jackson Demonstration State Forest. The model was developed by Boise Cascade for forest lands in Idaho and Washington. The structure of the model is flexible enough to easily adopt it to forest lands in northern California.

The model was run on Caspar creek to evaluate the contribution of roads as part of the basin's sediment budget and to assist in identifying roads that produce relatively high amounts of sediment. Model results suggest that while the ratio of road sediment to background sediment is fairly high (2x), less than half of the roads are expected to contribute significant sediment to streams and an even smaller portion of road segments would be considered chronic sources. The results of two model runs are shown here. The first run used the entire road network to estimate sediment yield. The second model run took into account the removal of over 4.5 miles of roads that have been slated for decommission. Removal of 7% of the total road network reduced predicted road sediment by 9%.

The model has yet to be validated with field observations. A comparison of current sediment yield estimates from SEDMODL with estimates from independent studies indicates that the model results are close to field based estimates. When used as a screening tool the model can provide first level estimates of road sedimentation and can assist in identifying roads that are currently chronic sources of sediment, and guiding a cost effective plan to reduce road based sediment.