0199-B1

Comparing Sediment Arising from Forest Operations to Sediment from Wildfire

William J. Elliot^[1]

Abstract

Wildfire and forest operations remove vegetation and disturb forest soils. Both of these effects can lead to an increased risk of soil erosion. Operations to reduce forest fuel loads, however, may reduce the risk of wildfire. This paper presents research and modelling results that show that under many conditions, carefully planned operations with adequate buffers, result in lower long-term erosion rates than experienced following wildfire, which is inevitable if fuel loads are not reduced. The effects of reducing fire-induced sedimentation on forest stream systems, however, are unknown.

Introduction

Forests provide numerous benefits for society, including fiber, wildlife, and recreation. Forest managers are challenged to balance ecosystem health and societal needs. During the first half of the last century, public forest management emphasized the harvesting of forest resources. In recent years, the emphasis in public forest management has been shifted to long-term sustainability.

Fuel management is another issue that has recently become important. During most of the last century, fire suppression and timber harvest were the main fuel management practices. This has resulted in forests with an oversupply of fuels and diseased trees, leading to an increased risk of severe wildfires (Duncan 2001).

Soil erosion is one of the major concerns in current forest management. Fires, timber harvest, and roads increase soil erosion and sediment delivery from forest watersheds. Soil erosion reduces upland forest productivity and eroded sediment adversely affects water quality in forest streams. Managers are seeking to minimize erosion by applying improved management practices for forest operations and roads. One of the questions to consider when seeking to minimize erosion is whether more frequent forest operations (and the associated road network) cause more or less erosion than less frequent wildfires.

The purpose of this paper is to compare predicted erosion rates following forest disturbances associated with fuel management to erosion rates following wildfires.

Figure 1. Erosion rates measured following a wildfire in Eastern Oregon. Note log scale on y-axis (Robichaud and Brown 1999).

Forest Erosion Processes

In forests, soil erosion occurs from disturbances such as forest roads, timber harvesting, or fire. These disturbances have major affects on both the vegetation and the soil properties. Soil erodibility depends on both the surface cover and the soil texture (Elliot and Hall 1997, Elliot et al. 2001). The soil erodibility on a skid trail is greater than in the areas between skid trails, and the erodibility following a wildfire is much greater than in an undisturbed forest (Robichaud et al. 1993).

Forests are highly susceptible to erosion in the year following a fire or a forest operation. They do, however, recover quickly as vegetation regrowth is rapid when smaller plants do not have to compete with trees for sunlight, nutrients, and water. For example, figure 1 shows the reduction in erosion rates following a wildfire in Eastern Oregon dropped about 90 percent the first year, with no erosion observed in year 4 on any of the slopes.

Erosion in forests is highly variable, driven by a few extreme events each decade, and highly influenced by recent disturbances. Sediment production during years without large runoff events or disturbances are likely to be well below “average” erosion rates, and the production during a year with a major runoff event or following a major disturbance is likely to be well above the “average” rate (Kirchner et al. 2001). For example, figure 2 shows the variation in road erosion as influenced by quality of aggregate, weather, and presence of traffic. Figure 2 shows that there is more than a magnitude difference in erosion rates due to differences in precipitation, and that in the final year of the study, erosion rates were much less without traffic than in the previous two years, even though it was the wettest year of the study. It also shows the effects of a single improved management practice, the use of high quality aggregate on the surface at reducing sediment production.

Figure 2. Effect of aggregate quality, weather, and traffic on road sediment yields. Years 1992-1994 had traffic. 1995 had no traffic (Based on Foltz (1998)).

Eroded sediments are frequently deposited in stream channels where they may remain for years to decades, slowly moving through the stream system in response to high runoff events (Trimble 1999). Thus, the scale at which sedimentation is measured becomes important. Smaller scales will show large variations in erosion rates as disturbed sites recover, whereas watershed scale observations will tend to reflect long-term trends in erosion rates, with large sedimentation events associated with infrequent watershed disturbances or flood events. Managers and the public tend to focus on the erosion that may occur immediately following a disturbance, and fail to consider medium- or long-term impacts of short-term disturbances.

Erosion Prediction

Prediction of soil erosion by water is a common practice for natural resource managers for evaluating impacts of management activities on upland erosion and offsite water quality. Erosion prediction methods are used to evaluate different management practices and control techniques. One of the prediction tools recently developed is the Water Erosion Prediction Project (WEPP; Flanagan and Livingston 1995). WEPP is a physically-based soil erosion model, and is particularly suited to modeling the conditions common in forests. Forest templates were developed for the model (Elliot and Hall 1997) and later, a user-friendly suite of Internet interfaces called FS-WEPP (Elliot et al. 2000). Included with these interfaces is a database of typical forest soil and vegetation conditions. These databases were populated with values determined from rainfall simulation and natural rainfall field research by scientists within our organization and elsewhere. Validation of the ability of the WEPP model to forest conditions has been encouraging (Elliot and Foltz 2001).

Table 1. Assumptions for two example harvesting systems

* LeDoux and Butler 1981

Comparing Management Strategies

The WEPP model was used to compare the sediment yields from fuel management estimated for wildfire for two different sites. The two ecosystems considered for this comparison were from the Northern Rocky Mountains in Montana, and from the Cascade Mountains in Western Oregon (Table 1). The Montana conditions assumed a relatively dry forest with a 40-year fire cycle, and an 80-year harvest cycle, with 30 percent slopes and tractor logging. The Oregon conditions assumed a 200-year fire cycle, a 40-year harvest cycle, with 60 percent slopes and a cable logging system. The FS-WEPP interface has the ability to describe many other management scenarios, but we felt these two systems would demonstrate the utility of the prediction tool and the erosion risks associated with fuel management and harvest activities.

Figure 3 shows the results of these simulations assuming an average climate. Figure 3a is for the wetter climate in the Oregon Cascade range, and figure 3b for the drier climate in the Bitterroot Mountains in Montana. There are several striking features on these two figures. On both graphs, the vertical axis is logarithmic. The erosion following wildfire is more than 2 magnitudes greater than before the fire, and more than a magnitude greater than following a major forest operation with a buffer. Also, the erosion rate in the Cascades is about two magnitudes greater than the erosion rate in the Bitterroots, even though the difference in precipitation is only about a factor of 2. The majority of the precipitation in the Bitterroots comes as snow, and snowmelt rates (typically 1 mm h^-1) are generally much lower than rainfall rates (typically up to 25 mm h^-1). Figure 3a shows that if we assume that thinning exposes about 15 percent of the mineral soil on a site, then the erosion rate due to thinning is similar to the erosion rate due to harvesting. Figure 3b shows that the erosion rate following a prescribed fire, assuming a 15 percent mineral soil exposure, is similar to the erosion rate following a harvest operation.

Figure 3. Hillslope sediment yield for an "average" weather pattern versus year for different management conditions for the two scenarios described in Table 1. Note the log scales on the vertical axes, and the difference in scales between the two graphs.

3a) Cascade Range, OR

3b) Bitterroot Range, MT

Figure 3 shows the sediment yield values assuming that every year had average weather. The year following a fire or other disturbance may be wetter than normal, increasing sediment yields considerably, or may be drier than normal, so that sediment yields are low to none. Table 2 shows the probability that the sediment yield will be nonzero, the sediment yield for an average year’s weather, and the sediment yield that may occur if the year following the disturbance is the wettest year in 5. There is a much greater likelihood that there will be sediment delivered in the Cascade scenario, and sediment delivery rates are much higher, as shown in figure 3.

In the Bitterroot Range, we predict that there is only an 8 percent chance that there will be any sediment yield in the year following a forest operation. The average sediment yield is greater than the 5-year sediment yield because the “average” value include the large rates predicted for very wet years, skewing the distribution of possible sediment yields. This means that if a study is set up to measure sediment yields following a forest operation, there is only an 8 percent chance that any sediment will be collected, and a 92 percent chance that there will be no observed sediment yield. In the wetter and steeper Cascade Range Scenario, there is an 80 percent chance that there will be sediment delivered across a buffer in the year following the disturbance.

Table 3. Average annual sediment delivery rates over 200 years for the scenarios presented in table 2 and figure 2.

Sediment from the road network depends on a number of watershed attributes including climate, network design and use, and topography. Elliot and Miller (2002) estimated road contributions for a wide range of western U.S. ecoregions. Each estimate included different levels of traffic, and considered three different topographies. Table 3 provides an average estimate of sediment yields from their study for road networks for the two sites, assuming a road density of 3 km km^-2.

Discussion

Figure 3 and table 2 raise a number of issues for further discussion. It is clear that erosion following wildfire is much greater than erosion due to forest operations, even though those operations may be more frequent than wildfire. However, erosion from wildfire is a natural phenomena, which has driven the development forest ecosystems. Occasional high upland erosion rates and large sediment yields have played an important role in shaping landscapes and introducing fresh material into our stream systems (Kirchner et al. 2001). In the last century, scientists found that fire was important for ecosystem health, and that fire exclusion resulted in a decline of the health of many forests. Will we find that the exclusion of wildfire and the large runoff and erosion events that follow will also lead to a decline in the health of our hydrologic and aquatic ecosystems? This question requires some significant interdisciplinary research to answer.

Forest roads contribute sediment to stream systems most years, and in many years, will generate more sediment than the forested hillsides. Forest operations, as shown in figure 3, may contribute low levels of sediment to stream systems more frequently than the natural wildfire cycle. Sediments from both roads and operations are likely to be finer, and less likely to contribute to the cobbled stream beds that are preferred by many aquatic organisms. Further research is necessary to evaluate the importance of large runoff events following fires for flushing these fine sediments through the stream system.

Another issue arises in figure 3 and table 2 is about the validity of using an average erosion rate. Following a forest disturbance, the overwhelmingly greatest amount of sediment is delivered in the first year, and after several years, the delivery is near zero. The amount of sediment delivered is highly dependent on the climate that first year, as shown in table 2. Watershed managers need to understand risks associated with different levels of sediment yield. They must then use that understanding to develop management strategies.

Table 3 shows the sediment yield rates presented in table 2 and figure 2 averaged over a 200-year period. If a manager’s goal is to simply reduce the sediment delivery to the stream systems, then the management strategies analyzed may achieve that goal. However, one must also include the erosion from roads to estimate the total impact of management activities (Conroy 2001; Elliot and Miller 2002). Table 3 shows that sediment from roads may represent a greater portion of the sediment budget in drier climates. As previously discussed, managers need to exercise caution when dealing with average values, as variability and outliers frequently dominate hydrologic processes.

Summary

We used the WEPP model to compare the sediment yields from forested hillslopes following wildfire to sediment yields from the same slopes following forest operations. Sediment yields following forest operations are much lower than following wildfire both the year following the disturbance, and when averaged over two centuries. We are not sure, however, if reducing the large sediment yields that follow wildfire will result in improved watershed health in the long term.

Conclusions

From our field work and our modeling results, we conclude the following:

• Sediment delivery following forest operations and prescribed fire with forested buffers are a magnitude or more lower than following wildfire.

• Roads can be a significant source of sediment, particularly in drier climates.

• Increased research is needed to determine the long term effects of forest operations on watershed health.

References

Conroy, W. J., 2001. Use of WEPP modeling in watershed analysis and timber harvest planning. Paper No. 01-8005. Presented at the ASAE Annual International Meeting, Jul 20 - Aug 1, Sacramento, CA. St. Joseph, MI: ASAE. 12 p.

Duncan, S., 2001. Benefits of hindsight: Reestablishing fire on the landscape. Science Findings. Portland, OR: USDA Forest Service Pacific Northwest Research Station. (36): 1-5.

Elliot, W. J., and D. E. Hall, 1997. Water Erosion Prediction Project (WEPP) forest applications. General Technical Report INT-GTR-365. Ogden, UT: USDA Forest Service, Rocky Mountain Research Station. 11 p.

Elliot, W. J., and I. S. Miller, 2002 Estimating erosion impacts from implementing the National Fire Plan. Paper no. 02-5011. Presented at the ASAE Annual International Meeting, Jul. 28 - Jul. 31, 2002, Chicago, IL. 26 p.

Elliot, W. J., and M. Foltz, 2001. Validation of the FSWEPP interfaces for forest roads and disturbances. Paper No. 01-8009. Presented at the ASAE Annual International Meeting, Jul 30 - Aug 1, Sacramento, CA. St. Joseph, MI: ASAE. 13 p.

Elliot, W. J., D. L. Scheele, and D. E. Hall, 2000. The Forest Service WEPP interfaces. Paper No. 005021. Presented to the ASAE Annual International Meeting, Jul 9-12, Milwaukee, WI. St. Joseph, MI.: ASAE. 9 p.

Elliot, W. J., P. R. Robichaud, D. E. Hall, C. O. Cuhaciyan, F. B. Pierson and P. M. Wohlgemuth, 2001. A probabilistic approach to modeling erosion for spatially-varied conditions. Paper No. 01-8006. Presented at the ASAE Annual International Meeting, Jul 30 - Aug 1, Sacramento, CA. St. Joseph, MI: ASAE. 14 p.

Foltz, R. B., 1998. Traffic and no-traffic on an aggregate surfaced road: Sediment production differences. Proceedings of the seminar on environmentally sound forest roads and wood transport. 17-22 June, 1996. Food and Agriculture Organization, Rome.

Flanagan, D. C., and S. J. Livingston (eds.), 1995. WEPP User Summary. NSERL Report No. 11, W. Lafayette, IN.: National Soil Erosion Research Laboratory. 131 pp.

Kirchner, J. W., R. C. Finkel, C. S. Riebe, D. E. Granger, J. L. Clayton, J. G. King and W. F. Megahan, 2001. Mountain erosion over 10 yr, 10 k.y., and 10 m.y. time scales. Geology 29(7): 591-594.

LeDoux, C. B., and D. A. Butler, 1981. Simulating cable thinning in young-growth stands. Forest Science 27(4): 745-757.

Robichaud, P. R., and R. E. Brown, 1999. What happened after the smoke cleared: Onsite erosion rates after a wildfire in Eastern Oregon. (Revised) In: Olsen, D. S., and J. P. Potyondy (eds.). Proceedings AWRA Specialty Conference: Wildland Hydrology, Nov. 2000, Bozeman, MT. 419-426.

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 Larionov, J. A. and M. A. Nearing (eds.). Proceedings from the Russia, U.S. and Ukraine international workshop on quantitative assessment of soil erosion. Moscow, Russia. Sep. 20-24. West Lafayette, IN: The Center of Technology Transfer and Pollution Prevention, Purdue University. 231-241.

Trimble, S. W., 1999. Decreased Rates of Alluvial Sediment Storage in the Coon Creek Basin, Wisconsin, 1975-1993. Science 285: 1244-1246.