Project Title: Developing Strategies for Effective Watershed Restoration

Investigator:	Mary Ann Madej USGS-BRD Redwood Field Station Arcata, CA USA 707-464-6101 Ext. 5490 FAX: 707-822-8904 mary_ann_madej@usgs.gov
Partners:	Redwood National and State Parks Arcata, CA USA 707-464-6101	Abandoned logging road in Redwood National Park (left) still presents an erosional threat to downslope resources. The same road following road decompaction and mild outsloping (right), techniques used to accelerate revegetation and decrease erosion.

BACKGROUND

Many anadromous fisheries streams in coastal California have been damaged by various land use activities, including timber harvest, road construction and grazing. Coho (Oncorhynchus kisutch) and chinook salmon (Oncorhynchus tshawytscha) and steelhead trout (Oncorhynchus mykiss) populations are in serious decline in Redwood Creek, north coastal California, and coho are now listed as threatened. Loss of habitat is a major contributing cause of this decline.

In 1978 Redwood National Park was expanded to include 48,000 acres of recently logged lands and more than 300 miles of abandoned logging roads. Public Law 95-250, the Redwood National Park expansion legislation, directed that a rehabilitation program be developed for the Redwood Creek basin. The goals of the program are to preserve and restore the natural ecosystem and landscape to a condition similar to what would have existed before disturbance by humans. The assumption underlying the watershed restoration program is that road removal will decrease human-induced sediment loads to perennial streams, and result in more favorable stream habitat, such as an increase in the frequency, distribution and size of pools. To date, the watershed restoration program has removed 200 miles of roads. Much interest and effort is directed at road removal in California and the Pacific Northwest, yet very little quantitative information exists to evaluate the effectiveness of such work.

In January, 1997, the first large flood to ‘test' the restoration efforts in Redwood National Park occurred. This was the largest flood in Redwood Creek since 1975. The project is evaluating the effectiveness and cost-effectiveness of different road removal techniques, and comparing amounts of post-flood erosion on treated and untreated roads in basins that support steelhead, coho and chinook salmon.

OBJECTIVES

The objectives of this study are two-fold. The first is to compare the amount of material eroded from treated and untreated roads on four terrain types as a result of the 1997 flood. The second objective is to compare the effectiveness of different techniques of road removal. Based on the results of these studies, guidelines for future restoration work through the most cost-effective strategies will be developed for federal and private land managers.

LOCATION

Both treated and untreated road segments are located in tributary basins in the Redwood Creek watershed which represent four geologic terrains. Bridge, Tom McDonald and McArthur Creeks are representative of schist terrain; Coyote Creek is underlain by an incoherent sandstone unit, roads in Emerald Creek are located on a coherent sandstone unit, and roads in Prairie Creek are underlain by unconsolidated coastal plain sediments. These four terrain types are typical of landscapes throughout north coastal California, and represent a range of susceptibility to erosion.

PROJECT DESCRIPTION

The project consists of field mapping all road crossings, intervening road reaches and landings on both treated and untreated abandoned haul roads in the study basins. Each road crossing, road reach and landing in the road segments was evaluated using the attached form. Much of the background site information is contained in Redwood National Park's annual rehabilitation reports or is available from U.S.G.S. topographic maps. Revegetation information is included because of 1) the concern of conifer regrowth in riparian zones and the recruitment potential for large woody debris in streams, and 2) the effect of revegetation on increased shading of the stream channel, with possible effects on stream temperature. The cause of erosion is documented (by fluvial processes - Section I, or mass movement - Section II, or both). The volume of material eroded will be measured, and the amount of sediment that was delivered into a stream channel was estimated. Erosion potential (the amount of sediment likely to fail and enter a stream in the next large (50-year) flood) was also estimated in the field.

The type of rehabilitation treatment was also documented (on untreated roads, questions regarding rehabilitation on the assessment form would obviously be left blank). Rehabilitation techniques range from minimal treatment, such as ripping and draining a road surface, to full landform reconstruction and recontouring of the road bench (maximum treatment). Differences in post-flood erosion on minimal, moderate and maximum treatment sites were computed.

In Coyote Creek, some roads were not removed, but instead had rolling dips constructed at stream crossings for erosion prevention while leaving the roads drivable. Elimination of diversion potentials by the construction of rolling dips is considered a fourth category of treatment that will be compared to similar roads without rolling dips.

Based on the preliminary results of this project, strategies for the most cost-effective treatment on a watershed scale will be developed. Dr. Elizabeth Eschenbach and her students at Humboldt State University will run optimization programs (dynamic programming and genetic algorithms) to help design the most effective erosion control strategy. Cost-effectiveness of the techniques will also be compared--for example, a low cost technique may not save as much sediment at a particular site, but because more road mileage can be treated for a given amount of funding, on a basin-wide basis it may be the more cost-effective approach to use. Recommendations regarding future road decommissioning activities will be developed.

RESULTS TO DATE

Since 1978, Redwood National Park has removed about 300 km of abandoned logging roads. Treatment involves removing culverts, excavating road fill from stream crossings, and decompacting or recontouring road benches between crossings. To date, 500,000 m3 of road fill has been excavated from 550 stream crossings.

In 1997 these decommisioned roads experienced their first large storm (10-year return interval). During the 1997 storm, untreated roads contributed four times the sediment to streams than treated roads on a per road length basis. The 1997 storm erosion accounted for 62% of the total post-treatment (1978-1997) erosion on road benches, whereas it accounted for only 5% of the post-treatment erosion at stream crossings. In Bridge Creek, a 30 km2 basin in which most roads have been removed, treated roads contributed 230 t/km2 to the sediment yield in 1997. In contrast, the sediment yield in Panther Creek, an adjacent basin with untreated roads, was 10 times greater in 1997.

Road benches in lower hillslope positions had a higher frequency of failure, and a greater overall volume of erosion per road length than road benches in upper and middle hillslope positions (even though individual landslides were smaller on lower hillslope roads). Initial results show that minimally treated roads did not contribute significantly more sediment to streams than more completely recontoured roads. Results from these studies can be used in formulating the most effective strategy for sediment reduction on a watershed scale.

CITATIONS

Baker, K. M, E. A. Eschenbach, and M. A. Madej. 2004. Forest road decommissioning policies determined using deterministic and stochastic dynamic programming. American Geophysical Union Annual Meeting.

Eschenbach, E. C. Diaz, R. Teasley, and M. A. Madej. In press. Decision support for road decommissioning and restoration using genetic algorithms and dynamic programming. Proceedings of the Redwood Region Science Symposium. Rohnert Park, California. March, 2004.

Madej, M. A. 2004.Restoration strategies for streams in steep, forested terrain. Joint International Geomorphology Conference. Glasgow, Scotland. Madej, M. A. 2001. Erosion and sediment delivery following removal of forest roads. Earth Surface Processes and Landforms. Vol. 26. No. 2 p. 175-190.

Ozaki, V., J. Wartella, and M.A. Madej. 1997. Environmental Tradeoffs in Management Decisions. pp. 444-449 in Making Protection Work: Proceedings of the Ninth Conference on Research and Resource Management in Parks and on Public Lands. D. Harmon, ed. The George Wright Society. Hancock Michigan. 493 pp.

AVAILABLE DATA

Data file (all-fields.xls 432 kbytes)
Data form (roadinventoryform.wpd 48 kbytes)
Definitions of fields in the data file (rehabdef.wpd 40 kbytes)
Report (includes analysis, photos and bibliography) (project-doc.pdf 2,657 kbytes)

Keywords: watershed restoration, roads, sediment sources, landslide, gully, stream crossings, erosion control