Forest Fire Effects on Hillslope Erosion: What We Know
Peter R. Robichaud
USDA- Forest Service, Rocky Mountain Research Station, Moscow, Idaho
Increased awareness of the role of fire in healthy ecosystems has focused attention on some of the effects of fires, wild and prescribed, on watershed condition and health. Precipitation events after forest fires may cause high sediment inputs, destruction of aquatic habitat and downstream flooding, all which may be part of the natural ecosystem response. However, if the fires are more severe due to past fire suppression activities, then the fire effects may be greater than natural. Fire and erosion are both natural processes that have been impacted by forest management activities such as fire suppression, logging, and road building during the last century. Management activities may contribute to increased streamflows and increased sediment supplies to streams and rivers. Additional sediment places streams and rivers at a higher risk for degradation. Sediment adversely affects spawning and rearing sites for anadromous and resident fish species, mobilizes in-stream sediment, and destroys aquatic habitat. Therefore, various management and mitigation strategies are often devised to reduce the threat of increased sediment. This paper reviews the effects of fire on hillslope erosion and the associated risks on watershed health.
Fire is a natural and important part of the disturbance regime for forested terrestrial and aquatic systems, especially in the western USA (Agee 1993). However, much uncertainty exists in quantifying fire effects on ecosystem components such as watershed condition and health.
Wildfires, which burn both small and large land areas, are often associated with lightning strikes from thunderstorms during the dry seasons and human-caused ignition (Agee 1990). Fire severity is a qualitative term used to measure the effect of fire on ecosystem components (Walstad et al. 1990) and is often used to describe fire effects on soil (Simard 1991). Ryan and Noste (1983) used ground char (burnt organic matter) classes to quantify fire severity. Fire effects on erosion are related to the effects of ground cover destroyed by fire. Ground cover usually consists of duff, grasses and debris on the ground surface. During fire, the consumption of ground cover (i.e. duff) exposes mineral soil which can be subject to overland flow and raindrop impact. The amount of vegetation, residue, and forest floor consumed and the soil heating caused by burning determines the extent to which soil properties are altered. The effects of fire on the forest floor can range from removing just the litter to total consumption of the forest floor and alteration of the mineral soil structure (Wells et al. 1979). The depth of the forest floor (litter layer and humus layer above mineral soil), the moisture content, and the amount of woody residue determine forest floor consumption during fire. When the forest floor is shallow or moisture content is low, fires consume more of the forest floor and have the potential to alter mineral soil (Reinhardt et al. 1991).
High severity burn areas experience higher rates of soil loss from erosion (McNabb and Swanson 1990), increased peak flows of runoff, greater duff reduction, loss in soil nutrients (Harvey at al. 1989), and soil heating (Hungerford et al. 1991). Water and sediment yields may increase as more of the forest floor is consumed (Robichaud and Waldrop 1994, Soto et al. 1994, Wells et al. 1979). If the organic layers are consumed and mineral soil is exposed, soil infiltration and water storage capacities are reduced (Robichaud 1996). Such impacts may last weeks or decades, depending on the fire' s severity and intensity, any remedial measures, and the rate of vegetative recovery (Baker 1990). Burning also reduces the amount of rainfall interception by the forest canopy and reduces evapotranspiration by the forest vegetation.
The use of prescribed fire has increased tenfold over the last decade, as land managers are trying to restore fire suppressed landscapes. For example, logging residue is often burned after timber harvesting. Burning is used alone and in combination with other treatments to dispose of slash, reduce the risk of insects and fire hazard, prepare seedbeds, and suppress plant competition for both natural and artificial regeneration. The effect of prescribed burning on the forest floor varies greatly, depending on fire severity and duration, forest floor consumption, and soil heating.
Surface erosion is the movement of individual soil particles by a force, either by uniform removal of material from the soil surface (sheet erosion) or by concentrated removal of material in the downslope direction (rill erosion) or gravity inducted (dry ravel) or by mass movement as landslides and debris flows (Foster 1982). Inherent erosion hazards are defined as the site properties that influence erosion. They include the ease with which the individual soil particles are detached (soil erodibility), slope gradient and length. Forces required to initiate and sustain the movement of soil particles can be from many sources, such as raindrop impact (Farmer and Van Haveren 1971), overland flow (Meeuwig 1971), gravity, wind, and animal activity. Protection is provided by all material on or above the soil surface, such as vegetation, surface litter, duff, and rocks that reduce the impact of the applied forces (Megahan et al. 1986; McNabb and Swanson 1990).
Soils are critical to the functioning of hydrological processes. Within a watershed, sediment and water responses to wildfire are often a function of fire severity and the occurrence of hydrologic events. For a wide range of fire severities, the impacts on hydrology and sediment loss can be minimal in the absence of precipitation. However, when a precipitation event follows a large, high-severity fire, impacts can be substantial. Increased runoff, peak flows, and sediment delivery to streams can affect fish populations and their habitat (Rinne 1997).
Fire can destroy the forest floor and vegetation, altering infiltration by exposing soils to raindrop impact or creating water repellent conditions (DeBano et al. 1998). Loss of soil from hillslopes produce several significant ecosystem impacts. Soil movement into streams, lakes, and riparian zones may degrade water quality and change the geomorphic and hydrologic characteristics of these systems and soil loss from hillslopes may alter future site productivity.
Two types of water repellency are common in forest environments: the first occurs when the soils and organic material are very dry, and the second occurs when the soils are heated due to fires (Figure 1). Combustion of surface fuels and the forest floor vaporizes hydrophobic organic substances which may move downward and condense at cooler underlying soil layers (DeBano 1981; DeBano et al. 1998). Water repellency in the mineral soil can contribute to reduced infiltration of water into the soil and increased erosion (Robichaud 1996).
Figure 1: Water repellent soils below wettable soils after a high severity wildfire shortly after a summer thunderstorm
DeBano and Krammes (1966) and Robichaud and Hungerford (In press) found that water repellency was dependent on the heating temperatures. At typical wildfire soil profile temperatures (less than 500 F, 260 C) and when the soil was dry, water repellency occurs at shallow depths (less than 1 inch, 25 mm). With wet soils, i.e. conditions that commonly occur during prescribed fire in the spring and fall, water repellency was less pronounced and only occurred after long heating times which, under field conditions, would typically only occur during smoldering fires. Therefore, water repellency after prescribed fire would probably be minimal (Robichaud and Hungerford In press).
Infiltration and erodibility
We have used rainfall simulations and concentrated flow for the past decade to measure infiltration, interrill and rill erodibility and effects of various surface conditions. There are four hydrological surface conditions which are important to characterize hillslope erosion potential in forest environments. These are unburned/undisturbed areas, low severity burn areas, high severity burn areas and skid trails or other highly disturbed areas (Robichaud et al. 1993).
To obtain infiltration and interrill erodibility estimates, simulated rainfall is applied to 11 ft2 (1 m2) plots (Figure 2). Rainfall intensities usually were 4 in hr-1 (100 mm hr-1) for three 30-minute events. Timed bottled samples are collected at the base of the plots. The samples are weighed and dried for flow volumes and sediment yields. Infiltration and erodibility are then calculated. Values depend on surface conditions and inherent soil variability. For example, unburned infiltration rates vary from 1.4 to 3.1 inches hr-1 (35 to 80 mm hr-1), while high severity rates vary from 0.8 to 2.4 inches hr-1 (20 to 60 mm hr-1). Infiltration rates following high severity burns often increase with time, due to water repellent conditions breaking down (Robichaud In press).
Figure 2: Rainfall simulator used to obtain infiltration and interril erodibility values on the Idaho Panhandle National Forest
Rill erodibility has been measured using concentrated flow down hillslopes (Robichaud and Brown 1999a) (Figure 3). Rill erosion is one of the dominant mechanisms of hillslope erosion. Various flow rates were used from 1.8 to 12 gal minutes-1 (7 to 45 l min-1) for 12 min with timed bottled samples used to collect runoff. These results were used to calculate sediment concentrations and rill erodibility. Sediment concentrations vary from 0.008 to 0.8 lb gal-1 (0.1 to 100 g l-1) which also vary according to surface condition and slope.
Figure 3: Concentrated flow being used to determine rill erodibility values on the Wenatchee National Forest
Fire severity is often variable, making erosion potential from burnt hillslopes also variable (Robichaud 1996). Spatial variability is an important characteristic of burned hillslopes. Geostatistical methods may be used to describe the spatial variability and topographic effects (Robichaud and Miller In press). The importance of variability observed in the field has been verified with erosion prediction models examining various arrangements of high- and low-severity fires on a hillslope (Robichaud and Monroe 1997). For example, for a 100 m hillslope with ' low- above high-severity' burn and 'high- above low-severity' burn condition arrangement, the high-severity burn condition above the low-severity burn condition produced about 50 percent more sediment since the rilling initiated in the upper portions of the hillslope continued down throughout the lower portion. When two thirds of the upper portion of the hillslope is in high-severity burn conditions, it produced twice as much sediment as compared to when the upper two-thirds were in low-severity burn conditions. The arrangement of high-severity burn conditions above the low-severity burn condition on a hillslope is common. As a fire burns, the heat generated can dry-out the upper portions of a hillslope and cause it to burn more severely.
Total water yields across the western U.S. vary considerably depending on precipitation, evapotranspiration, soils, and vegetation. The magnitude of measured water yield increases the first year after fire. This magnitude can vary greatly within a location or between locations depending on fire severity, precipitation, geology, topography, vegetation, and proportion of the vegetation burned (DeBano et al. 1998). Increases in water yield are primarily due to elimination of plant cover, with subsequent reductions in the transpiration component of evapotranspiration (Anderson et al. 1976). Water repellent soils and cover loss will cause flood peaks to arrive faster, rise to higher levels, and entrain significantly greater amounts of bedload and suspended sediments. Elevated streamflows decline as both woody and herbaceous vegetation revegetate during a recovery period ranging from a few years to decades.
Increases in water yield from wildfires and prescribed fires are highly variable. The first-year increase in water yield after a prescribed burn in a Texas grassland was 1,150 percent of the unburned control watershed (Wright et al. 1982). In Arizona chaparral burned by wildfire, the first year water yield increase exceeded 1,400 percent mainly due to water repellent soils.
The effects of disturbance on storm peakflows are highly variable and complex. Wildfires generally increase peakflows. For example, the Tillamook burn in 1933 in Oregon increased the total annual flow of two watersheds by 9 percent and increased the annual peakflow by 45 percent (Anderson et al. 1976). A 310 ac (127 ha) wildfire in Arizona increased summer peakflows by 500 to 1,500 percent, but had no effect on winter peakflows (Anderson et al. 1976).
Fire-related sediment yields vary, depending on fire frequency, climate, vegetation, and geomorphic factors such as topography, geology, and soils (Swanson 1981). In some regions over 60 percent of the total landscape sediment production over the long-term is fire-related. Much of that sediment loss can occur the first year after a wildfire (Agee 1993, DeBano et al. 1998, DeBano et al. 1996, Robichaud and Brown 1999b). Suspended sediment concentrations in streamflow can increase due to the addition of ash and silt-to-clay sized soil particles in streamflow which can adversely affect fish and other aquatic organisms.
Sediment yields one year after prescribed burns and wildfires range from very low in flat terrain and in the absence of major rainfall events to extreme in steep terrain affected by high intensity thunderstorms (Figure 4). Erosion on burned areas typically declines in subsequent years as the site stabilizes, but the recovery rate varies depending on fire severity. Soil erosion after fires can vary from under 0.4 to 2.6 t ac-1 yr-1 (0.1 to 6 Mg ha-1 yr-1) in prescribed burns and 9 to over 49 t ac-1 yr-1 (21 to over 110 Mg ha-1 yr-1) in wildfires (Megahan and Molitor 1975; Noble and Lundeen 1971; Robichaud and Waldrop 1994; Robichaud and Brown 1999b). For example, Radek (1996) observed erosion of 0.1 to 0.8 t ac-1 (0.3 to 1.7 Mg ha-1) from several large wildfires that covered areas ranging from 375 to 4,370 ac (200 to 1,770 ha) in the northern Cascades mountains. Three years after these fire, large erosional events occurred from spring rainstorms, not from snowmelt. Robichaud and Brown (1999b) reported first year erosion rates after a wildfire from 9 to 22 t ac-1 (21 to 49 Mg ha-1) decreasing by one to two orders of magnitude by the second year and to no sediment by the fourth in an unmanaged forest stand in eastern Oregon. Erosion rate reduction was due to recovery of natural vegetation. First year growing season shrubs, forbs and grasses accounted for 28 percent of the total ground cover whereas after the second growing season, total ground cover was 82 percent.
Figure 4: Cleaning debris from a sediment trap at the base of hillslope on the Wenatchee National Forest
DeBano et al. (1996) demonstrated that following a wildfire in ponderosa pine, sediment yields from a low severity fire recovered to normal levels after three years, but moderate and severely burned watersheds took 7 and 14 years, respectively. Nearly all fires increase sediment yield, but wildfires in steep terrain produce the greatest amounts. Noble and Lundeen (1971) reported an average annual sediment production rate of 2.5 t ac-1 (5.7 Mg ha-1) from a 900 ac (365 ha) burn on steep river breaklands in the South Fork of the Salmon River, Idaho. This rate was approximately seven times greater than hillslope sediment yields from similar, unburned lands in the vicinity.
Potts et al. (1985) indicated that wildfires increased water yield and sedimentation. Post-burn sediment increases were severe only on sites with both steep slopes and large fires. They found maximum annual sediment production of 1.9 t ac-1 (4.3 Mg ha-1), an increase of 284 percent over natural yields. These estimates were based on large-scale regional estimates on metamorphic parent material.
Hillslope Erosion Modeling
The Water Erosion Prediction Project (WEPP) model can be used to predict hillslope erosion from disturbed forest environments (Elliot et al. 1999). The approach is to predict the probability of erosion occurring after a disturbance by running WEPP model for 50 to 100 years of stochastic climates. Thus, the results will emphasize the risk of various erosion events occurring immediately after a fire and in the following years, when revegetation has caused the area to be hydrologically recovered. Field data collected over the last ten years is being used to populate and validate our modeling efforts.
Hillslope erosion processes can dominate landscape shape, especially after wildfires. Rill erosion is often the dominant mechanism for delivering sediment to the base of the hillslopes. The often denuded landscapes allow for direct impact of precipitation events and overland flow. Sediment may adversely affect aquatic habitat and water quality. Since most of our land management activities have increased sediment loads to rivers and stream, any additional sediment due to the fires could likely be detrimental.
When analyzing hillslope erosion, especially after fire, we should remember that erosion potential is not equal everywhere, erosion will only occur if a precipitation or snowmelt event occurs, and annual sediment yields generally decrease rapidly as natural vegetation reestablishes itself. -
You can reach Pete at email@example.com
Agee, James K. 1990. The historical role of fire in Pacific Northwest forests. In: Walstad, John D.; Radosevich, Steven R.; Sandberg, David V., eds. Natural and prescribed fire in Pacific Northwest forests. Corvallis, OR: Oregon State Univ. Press. chapter 3.
Agee, J.K. 1993. Fire ecology of Pacific northwest forests. Washington, DC: Island Press. 493 p.
Anderson, H.W., M.D. Hoover, and K.G. Reinhart. 1976. Forests and water: effects of forest management on floods, sedimentation, and water supply. Gen. Tech. Rep. PSW-18. Berkeley, CA: U.S. Department of Agriculture, Forest Service, Pacific Southwest Forest and Range Experiment Station.
Baker, M. B., Jr. 1990. Hydrologic and water quality effects of fire. In: Krammes, J. S., technical coordinator. Effects of fire management of Southwestern natural resources: Proceedings; Tucson, AZ. Gen. Tech. Rep. RM-191. Ft. Collins, CO: U.S. Department of Agriculture, Forest Service, Rocky Mountain Forest and Range Experiment Station. 293 p.
DeBano, Leonard F. 1981. Water repellant soils: a state-of-the-art. Gen. Tech. Rep. PSW-46 Berkeley, CA: U.S. Department of Agriculture, Forest Service, Pacific Southwest Forest and Range Experiment Station. 21 p.
DeBano, L. F. and Krammes, J. S. 1966. Water repellent soils and their relation to wildfire temperatures. Bulletin of the I.A.S.H., XI(2):14- 19.
DeBano, L.F., P.F. Ffolliott, and M.B. Baker, Jr. 1996. Fire severity effects on water resources. Pp. 77- 84. In: Ffolliott, P.F., L.F. DeBano, M.B. Baker, Jr., G.J. Gottfried, G. Solis-Garza, C.B. Edminster, D.G. Neary, L.S. Allen, and R.H. Hamre (tech. coordinators) Effects of Fire on Madrean Province Ecosystems‹A Symposium Proceedings. Gen. Tech. Rep. RM-GTR-289. Ft. Collins: U.S. Department of Agriculture, Forest Service, Rocky Mountain Forest and Range Experiment Station. 277 p.
DeBano, L.F., D.G. Neary, and P.F. Ffolliott. 1998. Fire' s Effects on Ecosystems. New York: John Wiley & Sons. 333 p.
Elliot, W.J. R. B. Foltz and P.R. Robichaud. 1999. Measuring and modeling soil erosion processes in forest. Presented at International Forestry Engineering Conference, Edinburgh, UK. 13 p.
Farmer, E. E.; Van Haveren, B. P. 1971. Soil erosion by overland flow and raindrop splash on three mountain soils. Gen. Tech. Rpt. INT-100. Odgen, UT: U.S. Department of Agriculture, Forest Service, Intermountain Research Station. 14 p.
Foster, G. R. 1982. Modeling the erosion process. In: Haan, C. T.; Johnson, H. P.; Brakensiek, D. L., eds. Hydrologic modeling of small watersheds. St. Joseph, MI: American Society of Agricultural Engineers. chapter 8.
Harvey, A.E., M.F. Jurgensen, and R.T. Graham. 1989. Firesoil interactions governing site productivity in the Northern Rocky Mountains. In: D. Baumgartner, ed. Prescribed fire in the intermountain region: forest site preparation and range improvement. Pullman, WA: Washington State University: 9- 18.
Hungerford, R.D., M.G. Harrington, W.H. Frandsen, K.C. Ryan, and G.J. Niehoff. 1991. Influence of fire on factors that affect site productivity. In: F.L. Neuenschwander and A.E. Harvey, comps. Proceedingsmanagement and productivity of westernmontane forest soils. Gen. Tech. Rep. INT280. Ogden, UT: U.S. Department of Agriculture, Forest Service, Intermountain Research Station: 3250.
McNabb, David H. and Swanson, Frederick J. 1990. Chapter 14. Effects of fire on soil erosion. In: Natural and prescribed fire in Pacific Northwest forests. Walstad, John D., ed., et al. Corvallis, OR: Oregon State University Press, pp. 159- 176.
Meeuwig, Richard O. 1971. Soil stability on high-elevation rangeland in the Intermountain area. Res. Paper INT94. Ogden, UT: U.S. Department of Agriculture, Forest Service, Intermountain Research Station. 10 p.
Megahan Walter F.; Molitor, Delbert C. 1975. Erosional effects of wildfire and logging in Idaho. In: Watershed Management Symposium; 1975 August. Logan, UT: American Society of Civil Engineers Irrigation and Drainage Division: 423- 444.
Megahan, Walter F.; Seyedbagheri, Kathleen A.; Mosko, Timothy L., and Ketcheson, Gary L. 1986. Construction phase sediment budget for forest roads on granitic slopes in Idaho. In: Hadley, Richard F., ed. Drainage basin sediment delivery, proceedings. Albuquerque, NM: 1986. IAHS Publication 159. Wajlingford, Oxon, UK. pp. 31- 39.
Noble, Edward L.; Lundeen, Lloyd. 1971. Analysis of rehabilitation treatment alternatives for sediment control. In: Symposium on forest land uses and stream environment: Proceedings; 1970 October; Corvallis, OR: Oregon State University: 86- 96.
Potts, Donald F. 1985. Water potential of forest duff and its possible relationship to regeneration success in the northern Rocky Mountains. Can. J. For. Res. 15:464- 468.
Radek, K.J. 1996. Soil erosion following wildfires on the Okanogan National Forest - initial monitoring results. In: International Erosion Control Association Symposium, February 27- March 1, 1996, Seattle, WA, 7 p.
Reinhardt, E.D., J.K. Brown, W.C. Fischer, and R.T. Graham. 1991. Woody fuel and duff consumption by prescribed fire in northern Idaho mixed conifer logging slash. Research Paper INT443. Ogden, UT: U.S. Department of Agriculture, Forest Service, Intermountain Research Station. 22 p.
Rinne, J.N. 1997. Shortterm effects of wildfire on fishes and aquatic macroinvertebrates in the southwestern United States. N. Am. J. Fish. Manage. 16:653- 658.
Robichaud, P. R. 1996. Spatially-varied erosion potential from harvested hillslopes after prescribed fire in the Interior Northwest. Ph.D. dissertation Moscow, ID: University of Idaho
Robichaud, P. R. [In press]. Fire effects on infiltration rates after prescribed fire in Northern Rocky Mountains. Journal of Hydrology.
Robichaud, P. R. and R. E. Brown. 1999a. Measuring rill erosion rates in a steep forest environment. Presented at the 1999 American Geophysical Union Spring Meeting, Boston, MA. 2 p.
Robichaud, P. R. and R. E. Brown. 1999b. What happened after the smoke cleared: onsite erosion rates after a wildfire in Eastern Oregon. Presented at the Wildland Hydrology Conference, Bozeman, MT. American Water Resources Assoc. 419- 426.
Robichaud, P. R. and S. M. Miller. Spatial interpolation and simulation of post-burn duff thickness after prescribed fire. [In press] International Journal of Wildland Fire.
Robichaud, P. R. and T. M. Monroe. 1997. Spatially varied erosion modeling using WEPP for timber harvested and burned hillslopes. Presented at 1997 Annual International Meeting, Paper No. 97-5015. ASAE, St. Joseph, MI. 8 p.
Robichaud, P.R. and T.A. Waldrop. 1994. A comparison of surface runoff and sediment yields from a low- and high-severity site preparation burns. Water Resources Bulletin 30 (1): 27- 34.
Robichaud, P. R. and R. D. Hungerford. [In press]. Water repellency by laboratory burning of four Northern Rocky Mountain forest soils. Journal of Hydrology.
Robichaud, P. R., C. H. Luce and R. E. Brown. 1993. Variation among different surface conditions in timber harvest sites in the Southern Appalachians. In International Workshop on Quantitative Assessment of Soil Erosion, Proceedings. Moscow, Russia. 231- 241.
Ryan, Kevin C. and Noste, Nonan V. 1983. Evaluating prescribed fires. Wilderness Fire Symposium; November 15-18; Missoula, Mont. 230- 238.
Simard, Albert J. 1991. Fire severity, changing scales, and how things hang together. Int. J. Wildland Fire 1(1):23- 34.
Soto, B.; Basanta, R.; Benito, E.; Perez R.; Diaz-Fierros, F. 1994. Runoff and erosion from burnt soils in northwest Spain. In: Sala, M.; Rubio, J. L., eds. Soil erosion and degradation as a consequence of forest fires: Proceedings; Barcelona Spain: 91- 98.
Swanson, F.J. 1981. Fire and geomorphic processes. Pp. 410- 420. In: Mooney, H. et al. (eds.) Proceedings of the Conference on Fire Regimes and Ecosystem Properties, Gen. Tech. Rep. WO-26. Washington, DC: US Department of Agriculture, Forest Service.
Walstad, John D.; Radosevich, Steven R.; Sandberg, David V. 1990b. Glossary. In: Walstad, John D.; Radosevich, Steven R.; Sandberg, David V., eds. Natural and prescribed fire in Pacific Northwest forests. Corvallis, OR: Oregon State University Press. appendix 3.
Wells, Carol G.; Campbell, Ralph E.; DeBano, Leonard F.; Lewis, Clifford E.; Fredriksen,
Richard L.; Franklin, E. Carlyle; Froelich, Ronald C., and Dunn, Paul H. Effects of fire on soil, a state-of-knowledge review. 1979. Gen. Tech. Rep. WO-7. Washington, DC: U.S. Department of Agriculture, Forest Service. 34 p.
Wright, H.A., F.M. Churchill, and W.C. Stevens. 1982. Soil loss and runoff on seeded vs. non-seeded watersheds following prescribed burning. Journal of Range Management 35:382- 385