Decision Support for Road Decommissioning and Restoration by Using Genetic Algorithms and Dynamic Programming
| Authors: | Elizabeth A. Eschenbach, Rebecca Teasley, Carlos Diaz, Mary Ann Madej |
| Year: | 2007 |
| Type: | General Technical Report |
| Station: | Pacific Southwest Research Station |
| Source: | In: Standiford, Richard B.; Giusti, Gregory A.; Valachovic, Yana; Zielinski, William J.; Furniss, Michael J., technical editors. 2007. Proceedings of the redwood region forest science symposium: What does the future hold? Gen. Tech. Rep. PSW-GTR-194. Albany, CA: Pacific Southwest Research Station, Forest Service, U.S. Department of Agriculture; p. 359-370 |
Abstract
Sediment contributions from unpaved forest roads have contributed to the degradation of anadromous fisheries streams in the Pacific Northwest.Efforts to reduce this degradation have included road decommissioning and road upgrading. These expensive activities have usually been implemented on a site specific basis without considering the sediment contributions from all roads within a watershed.
This paper describes results from optimization models developed for determining road removal management plans within a watershed. These models consider the tradeoffs between the cost and effectiveness of different treatment strategies to determine a treatment policy that minimizes the predicted sediment erosion from all forest roads within a watershed, while meeting a specified budget constraint.
Two optimization models are developed using dynamic programming and genetic algorithms. Each model accepts road survey data from the Redwood National Park’s (RNSP) GIS layers for a watershed with approximately 700 road segments and stream crossings. The models also require treatment effectiveness data, which are derived from previous published studies for the same area. The output from the model is the treatment level for each road segment and crossing and the total cost of the road removal management plan. The output is then exported to the GIS.
The models currently consider only road removal, but could be expanded to include additional road modifications or watershed restoration projects. Our approach is portable to other watersheds.