Managing Forest Road Removal Using Dynamic Programming: A Pilot Study

Since the late 1970’s the U.S. Geologic Survey (USGS) has led a program to remove abandoned logging roads in Redwood National Park. Because abandoned logging roads contribute large amounts of sediment to local fish bearing streams, the ecosystem health of these waterways suffer. Recent research has identified the effectiveness of preventing sediment from reaching the streams for different road treatments after significant storm events. However, road removal is expensive and time consuming. This research reported in this paper was part of that pilot study, and specifically reviews the feasibility of the optimization algorithm Dynamic Programming (DP), using data from recent research on road removal effectiveness. The DP sought to determine the road removal treatment that maximizes the amount of sediment saved from erosion, while meeting a budgetary constraint. The results indicate that DP is an effective tool for developing a road removal management plan. However, the order in which roads and stream crossings are treated has a large effect on the solution, indicating that the DP formulation has room for improvement. The USGS is supporting further research to reformulate the DP.