The study was designed using a two-component approach, with intensive and extensive project areas. The Skookum watersheds, with 11 years of pre-treatment data, provided a rare opportunity to examine erosion budgets at multiple scales, where the effects of management prescriptions to treat fuel loadings would be set within the context of longer-term water and sediment yield. Work within the Skookum Watersheds was the “intensive” component of this study. We recognized the possibility that the watersheds would not be treated as planned due to their large size, lack of roads, heavy fuel accumulations, and the fact that they are contiguous with a large roadless area. The risks and consequences of an escaped prescribed fire would be substantial. Therefore we considered other planned prescribed fire projects within the Blue Mountains – an extensive geographic area that encompassed a great variety of site conditions. Of the planned projects, two had a high probability of taking place as planned, had relatively easy access, and contained planar hillslopes with nearby controls. These projects, Red Fir and Lick Creek, were the “extensive” component of this study. They are not located within gauged, experimental watersheds and therefore lack the long period of pre-treatment monitoring available at the Skookum watersheds.
Variable-area erosion plots were used to measure surface erosion in all project areas. In the Skookum project area, plots were established at both the control and proposed treatment watersheds in July 2002. In the Red Fir extensive site, erosion plots were established in the control in July 2002, and one week after the sites were burned in October 2002. Control plots were installed in the Lick Creek project area in July 2003. Treatment plots at the Lick Creek site were installed one week after the sites were burned in October 2004. Erosion plots on the treated hillslopes at both Red Fir and Lick Creek were located and staked with steel fence posts at the time that the control plots were established to avoid unintentionally biasing plot placement to sample areas of greater (or lesser) burn intensity.
Silt fence sediment collection:
Sediment fences were located on planar hillslopes (neither concave nor convex across the contour) so that calculated erosion rates would not be confounded by convergent or divergent patterns of overland flow. Fences were laid out in transects consisting of three replicates, spaced 30 m apart, with transects located at upper-, mid-, and toe-slope positions along the length of the hillslope. Sediment fences were designed following the methods described by Robichaud and Brown (2002). Sediment fences were made from black silt-fence fabric supported with light-weight metal fence posts. Each fence is a total of 5-m wide (including a 0.5-m wide curve at each end to prevent flow around the fence) and oriented perpendicular to the hillslope (or parallel to the contour). Two features of our design differ from the design described by Robichaud and Brown (2002). First, we used 1.83 m wide fabric, using the excess width to form a contiguous apron, covering the soil surface and extending 0.5 to 0.8 m upslope of the actual sediment fence which allowed collection of small amounts of sediment without accidentally collecting any of the underlying soil. Secondly, all plots are “unbounded” so the contributing area of each plot is defined by a variable area of hillslope that contributes runoff, surface erosion and dry ravel over the period of measurement.
Sediment in the fences was collected in late spring (after snow melt) and again in the fall (around September 30th). All large branches, sticks and cones were first removed and discarded. Litter that was not mixed in with sediment was also removed and discarded. Large accumulations of dirt were collected with a trowel and then the fence apron was swept with a whisk-broom and fine sediment was collected with a dust pan. Collected materials were bagged, labeled, and transported to the laboratory.
Plot description/characterization:
Slope: measurement taken from the base of the silt fence to 10 m upslope.
Aspect: measurement taken looking upslope from the base of the silt fence.
Lattitude and Longitude: taken at the silt fence with a Garmin GPS 12.
Visual ground cover estimate:
ocular estimates were taken in a 10 m by 10 m plot with the silt fence as the downslope boundary. Ocular estimates of the percent cover of ash, charcoal, bare ground, gravel, rock, duff, wood, and vegetation were made. These estimates focused on the ground surface exposed to direct raindrop impact so that the sum cannot exceed 100%. Measurements were collected from both control and treatment plots during the summer prior to the prescribed burn, directly after the burn, and every summer after treatment.
Vegetation cover estimate:
Further ocular estimates of vegetation cover were made in the same 10 m by 10 m plot to break total canopy cover into the following growth forms: grasses, forbs, sub-shrubs, shrubs, and trees. Measurements were collected from both control and treatment plots during the summer prior to the prescribed burn, directly after the burn, and every summer after treatment.
Longitudinal and Cross-slope concavity profile:
Visual determination of slope concavity profile at the plot. Longitudinal determination was taken from downslope to upslope, with a choice of concave, convex, or flat. Cross-slope determination was taken looking upslope from left to right across the slope. Concave = convergent flow, Convex = divergent flow, and Flat = parallel flow.
Duff thickness:
Duff thickness in cm was measured at 0.5 m intervals (9 measurements total) along the length of the upslope, vertical face of the 0.1 m wide trench into which the upslope end of the silt-fence material was buried.
Fuel load:
Four fuel load sample points were spaced across each hillslope position, alternating with the silt fences, for a total of 12 fuel-load sample points per hillslope. Two 50 ft. long planar intercept transects were established at each sample point. To determine the random bearing for each transect, we glanced at the second hand of a watch when ready to string the transect. The seconds are multiplied by 6 (60 second intervals X 6 = 360 degrees possible). Fuel load transects were established at each hillslope in the Skookum Project Area in 2004, the Lick Creek Project Area in 2003, and the Lane Creek Project area in 2003 following methods described by Brown (1974).
Silt fence sediment samples were oven dried to remove moisture from the samples to prevent molding. The samples were sieved into size-fractions (great than 12.5 mm, 12.5 – 6.3 mm, 6.3 - 4 mm, 4 - 2 mm, 2 – 1 mm, and less tahn 1 mm) then oven dried (96 hr at 55°C), and weighed. Since the oven dried size fractions of each sample included substantial amounts of fine particulate organic material, they were combusted in a muffle furnace (10 hr at 600°C) and then re-weighed.
Fuels: Measurements are entered into Fuels Management Analyst Plus (2003 version) to calculate tons/acre fuel loading. The results from the 8 transects at each slope position (upper, middle, and lower) are averaged to report a tons per acre at the slope position.
Silt fence sediment samples: The weight of the remaining sample was corrected for the residual mineral content from the combusted organic materials, assuming that the mineral content of organic matter averaged 5.05 percent. The results were multiplied by a constant (0.050468) to account for the mass of ash remaining in the sample.
When fields have a missing value, it usually means the parameter was intentionally not sampled on that date.
References:
Brown, JK. 1974. Handbook for inventorying downed woody material. Gen. Tech. Rep. INT-16. Ogden, UT: U.S. Department of Agriculture, Forest Service, Intermountain Forest and Range Experiment Station. 24 p.
Robichaud, PR, RE Brown. 2002. Silt Fences: an economical technique for measuring hillslope soil erosion. Gen. Tech. Rep. RMRS-GTR-94. Fort Collins, CO: U.S. Department of Agriculture, Forest Service, Rocky Mountain Research Station. 24 p.
Publications: Harris, Robin M., Caty F. Clifton, and Steven M. Wondzell. In Review. Evaluating the Effects of Prescribed Fire and Fuels Treatment on Water Quality and Aquatic Habitat. In Proceedings, Advancing the Fundamental Sciences: A Conference for Forest Service Physical Scientists, October 18-21, 2004. San Diego, CA.
