FORTY YEARS OF LESSONS LEARNED ABOUT THE IMPACTS OF
FOREST PRACTICES ON WATERSHED HYDROLOGY AND WATER QUALITY

George Ice, Principal Scientist
 National Council for Air and Stream Improvement, Corvallis, Oregon

For more than 100 years, forest management in the United States has been evolving to address forest goals for timber and water.  The origins of the Forest Service, the history of forest watershed research and education, adoption of nonpoint source control programs, and classic Journal of Forestry articles such as “The relation of forests to our water supply” (Croft and Hoover 1951) all point to the long-recognized importance of water in forest management.

The Random House Dictionary defines “manage” as “to bring about; succeed in accomplishing.”  Among the tools to “succeed in accomplishing” these forest goals are Best Management Practices (BMPs).  It is sometimes remarkable how little credit we give ourselves for the extensive research that has been conducted to define management practices to meet our forest and watershed goals.  A recent report by the National Academy of Science on Hydrologic Effects of a Changing Forest Landscape (NRC 2008) spends a mere two pages on BMPs and describes them as “…negotiated compromises between parties with economic interests in management activities and those with interests in environmental protection.”  We must ask what environmental management activity does not involve some balance with economic and social realities.  The report goes on to state that “…very little research has investigated whether the current suite of BMPs will be effective in reducing cumulative watershed effects, maintaining viable fish populations, or preserving the integrity of forest and stream ecosystems.”  Here, we will argue that we have learned a lot about the impacts of forest practices on hydrology and water quality (Figures 1 to 3), and ongoing research such as that being conducted by the Watersheds Research Cooperative is addressing cumulative watershed effects and the viability of fish populations.

Key Lessons About Forest Practices Effects on Hydrology and Water Quality

Watershed lessons are often allegorical in nature and one of my favorite watershed stories comes from a follow up to a study by Dr. John Hewlett from the University of Georgia.  Hewlett studied loggers-choice forest harvesting on the Grant Forest, Georgia, in the mid-1970s (Hewlett 1979).  He found that “…sediment pollution of streams by forestry operations appears from this study to be far less than annual levels of export deemed tolerable under agriculture.”  However, he also concluded that if the operations had employed just three additional control practices (well designed and maintained roads, adequate streamside management zones, excluding machine planting near abandoned gullies) the sediment produced from forest harvesting could have been reduced by about 90%.  Williams et al. (2000) and Jackson (personal communication) repeated watershed studies in the Piedmont, the first in South Carolina and the second back at Grant Forest.  In both cases they confirmed Hewlett’s prediction that when BMPs were used sediment losses due to management were reduced by about 90%.

In 2000, with celebration of the 100th anniversary of the Society of American Foresters (SAF), the SAF Water Resources Working Group organized a special session of “grandmasters of watershed management” in Washington, DC.  Papers from that session were collected for publication in the book, A Century of Forest and Wildland Watershed Lessons (Ice and Stednick 2004).  The key watershed lessons herein represent some of the findings from that book and other forest watershed literature.

Forest management practice choices make a difference in watershed impacts

In A Century of Forest and Wildland Watershed Lessons, Beasley et al. (2004) discussed a series of studies in the mid-South during the 1980s designed to assess the impacts of forest management on water quality.  Like Hewlett, they found that alternative practices and different site conditions could affect water quality responses.  In the central study discussed by Beasley et al., nine small basins from 2 ha to 6 ha in size at each of four locations in Arkansas and Texas were instrumented and monitored.  The study design allowed for three replicates of three treatments (including a control) with pre- and post-treatment comparisons.  Results showed that sediment losses were low compared to other land uses but that the highest sediment losses occurred for sites where mechanical site preparation exposed soil (shear, windrow, and burn), slopes were steep, and there were low amounts of rock fragments (to create an armor layer) in the soils.

The basins that experienced the most sediment losses were the sheared, windrowed, and burned basins in the watershed study located near Alto, Texas.  Values were nearly three times the sediment losses experienced at the rest of the study locations.  Several years ago it was recognized that these basins near Alto were once again nearing harvest age and that the concrete approach sections were still in place and able to be re-instrumented.  As a result, a study was begun to test whether contemporary practices under the Texas Forestry BMPs protected water quality and what the impacts of more intensive management using additional soil and chemical treatments would be.  In addition, new questions were asked about things like the cumulative effects of multiple activities in the larger watersheds and the role of roads in overall watershed responses.

Like every good story, the Alto Watershed Study involves numerous unexpected twists and turns.  Tropical Storm Allison hammered East Texas during the pre-treatment phase.  In addition to catastrophic flooding in Houston, it caused extreme runoff in the Alto Watershed and damaged some of the monitoring infrastructure.  Small basins that were clearcut with roller-chopping site preparation in the 1980 generated 2.5 times the sediment during three hours of this storm than for the entire year following clearcutting and site preparation in the original study (McBroom et al. 2008).  The 2002 treatments included a conventional treatment with clearcutting and herbicide site preparation and an intensive treatment that added subsoiling, fertilization, and a release herbicide application.  Both the conventional and intensive treatments employed designated streamside management zones around channels and had reduced exposed mineral soil compared to 1980s treatment with shearing and windrowing.  Sediment production the first year after harvesting in 2002, even from the single largest sediment producing small basin was only one-fifth that experienced on the same watersheds in 1981 after shearing and windrowing.  The scale of sediment losses observed for the larger watersheds were similar to those observed in the small basins.  McBroom et al. concluded that, “these sediment loss rates are much lower than those observed in the 1980’s.”

The importance of streamside management zones in minimizing water quality impacts

Both Jackson et al. (2004), describing watershed lessons from the Southeast, and Ice et al. (2004), describing lessons from the Pacific Northwest, identified research on the important role streamside management zones (SMZs) play in minimizing water quality impacts from forest management activities.  The role is clearly captured in the original Alsea Watershed Study in Oregon, but also in studies from the South such as Swift and Messer (1971), where unbuffered streams in clearcuts increased an average of 4°C.  SMZs also play an important role in minimizing increases in sediment, nutrients, and chemicals.  A recent study in the Oregon Coast Range found that drift from helicopter spray operations can be reduced 90% or more by riparian vegetation (Ice et al. 2008).  SMZs are also looked at to provide a source of large wood for streams and wildlife habitat.  Perhaps this is the reason that various SMZ or buffer practices are universally found in state BMPs.  An example of these is the Montana “seven don’ts” under the Streamside Management Act (Logan and Clinch 1991):

  • Don’t broadcast burn
  • Don’t operate equipment except on established roads
  • Don’t clearcut
  • Don’t construct roads except when necessary to cross
  • Don’t handle, store, apply, or dispose of hazardous or toxic chemicals
  • Don’t side-cast road material
  • Don’t deposit slash in streams or other waterbodies

Some type of SMZ is universally accepted as a tools to reduce water quality impacts, but controversy arises from defining adequate SMZ dimensions.  This problem is further muddled because we tend to think one-dimensionally:  how wide should the SMZ be?  Two other important dimensions are the length of stream or stream type to be managed and the management restrictions to be imposed.  These limitations can range from equipment exclusion zones to full buffer protection.

Given this range of SMZ widths and management restrictions, it is not surprising that different management prescriptions create conflicts and questions about excessive or inadequate protection.  When the debate becomes contentious, remember there is a law of diminishing returns for SMZs.  For stream functions, most benefits come from the area nearest the stream.  An example of this is a study in Maine of riparian buffer effectiveness that looked at alternative widths and their impacts on stream temperature.  Wilkerson et al. (2005) reported:

Streams without a buffer showed the greatest increase in mean weekly maximum temperatures following harvesting (1.4-4.4°C).  Streams with an 11-m buffer showed minor, but not significant increases (1.0-1.4°C).  Streams with a 23-m buffer, partial-harvest treatment, and control streams showed no changes following harvesting.  The mean weekly maximum temperatures never exceeded the thermal stress limit for brook trout (25°C) in any treatment group.

One of the questions that has been raised about the Watershed Research Cooperative is whether alternative riparian prescriptions are being tested in addition to simply testing the effectiveness of the current Forest Practices Act rules.  This testing of alternative riparian prescriptions is occurring at the Trask Watershed.

Forest road impacts on watersheds

It has long been recognized that careful management of forest roads is essential to maintaining high water quality.  The US Environmental Protection Agency (EPA) is currently evaluating whether forest roads should be reclassified under the federal Clean Water Act from nonpoint sources to point sources of pollution, subject to stormwater discharge permits.  Hornbeck and Kochenderfer (2004), in A Century of Forest and Wildland Watershed Lessons, described watershed research from the Northeast on roads:

Studies at the Fernow have resulted in guides for all phases of road construction, including planning, layout, construction, care after logging…, use of gravel to protect against erosion…, sizing of culverts… and drainage structures…

Likewise, Jackson et al. (2004) found that:

The greatest potential for increases in sediment from forest activities is associated with poorly designed or maintained rods and channel disturbance for hilly or mountainous sites.  Practices that reduce these impacts greatly reduce overall changes in sediment related to forestry activities.

For example, Jackson et al. reported a series of watershed studies at Coweeta Hydrologic Laboratory that showed that graveled roads reduced sediment losses eight-fold compared to bare roads.  As part of EPA’s review of roads we recently identified ten key lessons and supporting literature about roads and water quality.

I.                Most of erosion comes from a small fraction of the road network (the 20/80 rule is that 20% of the road network contributes 80% of the sediment).  Surveys and modeling can target these problem spots.  High erosion-risk conditions may need enhanced treatment.

1.      Rice, R.M. and J. Lewis.  1991.  Estimating erosion risks associated with logging and forest roads in northwestern California.  Water Resources Bulletin 27(5):809-817.  Key finding:  Critical road sites with erosion features greater than 100 yd3acre-1 represented just 2% of the roads network but 70% or more of the erosion measured.

2.      Cafferata, P.H., R. Harris, and D.B.R. Coe.  2007.  Water resource issues and solutions for forest roads in California.  Proceedings of the 2007 American Institute of Hydrology annual convention.  Key finding:  5.5% of drainage structures in California had problems; 20% of road-stream crossings were found to have implementation or effectiveness problems; inventory work shows that a relatively small portion of the road network is the source of most of the sediment.

3.      Toman, E.M. and A.E. Skaugset.  2007.  Designing forest roads to minimize turbid runoff during wet weather use.  612-616 in Watershed management to meet water quality standards and TMDLs: 4th Conference proceedings.  St. Joseph, MI: American Society of Agricultural and Biological Engineers.  Key finding:  Sediment losses from the road surface are greatest where ruts form, providing a physical indicator of the sediment source.

II.             With forest roads, as with real estate, it’s location, location, location.

1.      Megahan, W.F., M. Wilson, and S.B. Monsen.  2001.  Sediment production from granitic cutslopes on forest roads in Idaho, USA.  Earth Surface Processes and Landforms 26:153-163.  Key finding:  Slope gradient is by far the most influential site factor affecting erosion.

2.      Brake, D., M. Molnau, and J.G. King.  1997.  Sediment transport distances and culvert spacing on logging roads within the Oregon Coast Mountain Range.  Paper presented at 1997 annual international meeting of ASAE, Minneapolis, MN.  Key finding:  Travel distances for sediment transport below roads can be used to locate roads far enough away from streams to allow for settling of sediment.

III.          The shape and surface of the road influences its ability to handle traffic and its erodability.  High quality rock is often the most expensive part of road construction and it can be difficult to find in some areas, but it and other road surface practices can reduce surface erosion.

1.      Handouts to Forest road surfacing: Basic design principles and applied practices.  Workshop sponsored by the Western Forestry and Conservation Association.  March 5-6, 2007, in Canyonville, OR, and March 8-9, 2007, in Shelton, WA.  Key finding:  Field methods are available to build less erosive roads.

2.      Coe, D.B.R.  2006.  Sediment production and delivery from forest roads in the Sierra Nevada, California.  M.S. Thesis, Colorado State University, Fort Collins, CO.  Key finding:  There is a 16-fold difference in sediment lost from rocked and un-rocked roads.

3.      Swift, L.W. Jr.  1984.  Gravel and grass surfacing reduces soil loss from mountain roads.  Forest Science 30(3):657-672.  Key finding:  Rock (15 cm or more deep) reduced sediment losses from a road by 78%.  Grass reduced sediment by 45% compared to a bare road surface.

IV.           Mulching and seeding can reduce erosion from exposed cut and fill slopes and the road surface.

1.      Megahan, W.F.  Erosion processes on steep granitic road fills in central Idaho.  Soil Science Society of America Journal 42(2):350-357.  Key finding:  Tree planting and mulching reduced erosion for road fills by 44% and 95%.

2.      Rothwell, R.L.  1983.  Erosion and sediment control at road-stream crossings.  The Forestry Chronicle 59(1):62-66.  Key finding:  About five times as much sediment came off road sections without mulch as from road sections with brush mulch.

V.              Dispersing flow off of roads and detaining in once it leaves the road prism allows increased settling of sediment before it reaches a stream.  Dispersing flow can be accomplished by outsloping, rolling dips, relief culverts, and even belt diverters.  Increased roughness of the forest floor will increase sediment detention and settling.

1.      Packer, P.E.  1967.  Criteria for designing and locating logging roads to control sediment.  Forest Science 13(1):2-18.  Key finding:  Effective spacing decreases with steeper road gradients and with lower topographic position.

2.      Mills, K., L. Dent, and J. Robben.  2003.  Oregon Department of Forestry wet season road use monitoring project: Final report.  Forest Practices Monitoring Program Technical Report 17.  Salem, OR: Oregon Department of Forestry.  Key finding:  The two road construction factors found associated with increased turbidity at road/stream crossings during wet weather periods were more than 250 feet of road ditch flowing directly to a stream channel and road surface material with more than 7% silt-sized or smaller particles.

VI.           Direct delivery of sediment from hydrologically connected roads can be a source of fine sediment in forest watersheds.  The forestry community is actively disconnecting legacy conditions and building new roads to avoid these conditions.

1.      Bilby, R.E., K. Sullivan, and S. Duncan.  1989.  The generation and fate of road-surface sediment in forested watersheds in southwestern Washington.  Forest Science 35:453-468.  Key finding:  Most fine sediment from roads gets to streams by direct delivery at stream crossings or where gullies are directly connected to streams.

2.      Ketcheson, G.L. and W.F. Megahan.  1996.  Sediment production and downslope sediment transport from forest roads in granitic watersheds.  Research Paper INT‑RP‑486.  Ogden, UT: USDA Forest Service, Intermountain Research Station.  Key finding:  Sediment is transported much farther in concentrated flows than in dispersed runoff.

3.      Furniss, M.J., S.A. Flanagan, and V. McFadin.  2000.  Hydrologically-connected roads: An indicator of the influence of roads on chronic sedimentation, surface water hydrology, and exposure to toxic chemicals.  Stream Notes (July 2000).  Key finding:  Multiple studies show that a large fraction of forest road networks has historically been hydrologically connected to surface runoff features (streams and gullies).  Actions to reduce impacts include disconnecting roads from streams, decreasing cross drain spacing, and applying treatments to retard flow and allow sediment to settle.

4.      Mills, K., L. Dent, and J.L. Cornell.  2007.  Rapid survey of road conditions to determine environmental effects and maintenance needs.  Paper presented at 9th International Conference on Low-Volume Roads.  Austin, TX.  Key finding:  Road inventories show that the percent of the road network draining to streams is being decreased.  Historic rates were 57 to 75% of the road network, while recently constructed or upgraded roads have rates of 15 to 34%.

VII.        In steep, forested regions flow diversion can be a major source of sediment, but these problems are recognized and are being addressed.

1.      Hagans, D.K. and W.E. Weaver.  1987.  Magnitude, cause and basin response to fluvial erosion, Redwood Creek basin, northern California.  419-428 in Erosion and sedimentation in the Pacific Rim.  Beschta, R.L., T. Blinn, G.E. Grant, G.G. Ice, and F.J. Swanson [Eds.].  Publication 165.  Wallingford, Oxfordshire, UK:  International Association of Hydrological Sciences.  Key finding:  Surface erosion sediment delivered to streams is nearly equal to the sediment from landslides, and most of that comes from gullies created by flow diversion on forest roads.

2.      Cafferata, P.H., R. Harris, and D.B.R. Coe.  2007.  Water resource issues and solutions for forest roads in California.  Proceedings of the 2007 American Institute of Hydrology annual convention.  Key finding:  Storm-proofing of roads is designed to upgrade culverts and crossings to avoid catastrophic failures.  One small company in northern California has upgraded 845 km of roads over the past seven years.

VIII.     Landslides can be another important source of sediment from forest roads in steep terrain, but management practices such as pull-back of unstable sidecast road material can improve road performance.

1.      Ice, G.G.  1985.  Catalog of landslide inventories for the Northwest.  Technical Bulletin No. 456.  Research Triangle Park, NC: National Council for Air and Stream Improvement, Inc.  Key finding:  Inventories of landslides for older road systems built with sidecast road construction methods on steep slopes showed most landslides coming from roads.  With recognition of this problem, more recent landslide inventories show a decreasing percent of slope failures from roads.

2.      Robison, E.G., K. Mills, J. Paul, L. Dent, and A. Skaugset.  1999.  Oregon Department of Forestry storm impacts and landslides of 1996: Final report.  Forest Practices Technical Report 4.  Salem, OR: Oregon Department of Forestry.  Key finding:  Road-associated landslides made up a smaller percentage of the total number of slope failures in this recent study than in most past studies.

IX.           Maintenance practices are faced with the Goldilocks dilemma:  not too much or too little.

1.      Sugden, B.D. and S.W. Woods.  2007.  Sediment production from forest roads in western Montana.  Journal of the American Water Resources Association 43(1):193‑206.  Key finding:  Reducing the frequency of grading can significantly reduce sediment yields from forest roads.

X.              Legacy road conditions, rather than current activities, are often a source of sediment problems.  These existing problems can be reduced as part of active forest management.

1.      Sullivan, K.  2003.  Variation in turbidity at the THP scale.  Abstract to paper presented at A conference on water quality monitoring: Spatial and temporal variability in forest water quality monitoring: Water quality research and regulations.  Key finding:  The source of increased turbidity in streams in northern California was often old legacy road conditions such as failed Humboldt crossings.

2.      Cafferata, P.H., R. Harris, and D.B.R. Coe.  2007.  Water resource issues and solutions for forest roads in California.  Proceedings of the 2007 American Institute of Hydrology annual convention.  Key finding:  Road practices are dramatically changed and the forestry community (federal, state, and private) is conducting road inventories, upgrading roads (storm-proofing), and even decommissioning unnecessary roads.

3.      Rogers, D.  2007.  Montana forestry best management practices monitoring: 2006 forestry BMP audit report.  Missoula, MT: Montana Department of Natural Resources and Conservation, Forestry Division.  Key finding:  61% of harvest sites with roads had reduced sediment delivery as a result of upgrading roads during active management.

Summary and Conclusions

BMPs have been found to be very effective in reducing impacts from forest management.  A synthesis of research on BMP effectiveness shows that these control practices can be as much as 80 to 90% effective in reducing pollution loads to streams.  A substantial amount of literature is available on BMP effectiveness and there are many ongoing research and modeling efforts to further improve our understanding, including the Watersheds Research Cooperative.  This ongoing research is especially important as we consider modern forest practices compared to historic management activities that may have created legacy conditions.  By recognizing hazards from existing road networks and legacy conditions, managers are able to apply BMPs and mitigation measures to further reduce the risk of negative impacts from forest management activities.

References

Beasley, R.S., E.L. Miller, W.H. Bleackburn, and E.R. Lawson.  2004.  Assessing the impacts of intensive forest management: Watershed studies of the mid-South.  113- 128 in A century of forest and wildland watershed lessons.  Ice, G.G. and J.D. Stednick [Eds.].  Bethesda, MD: Society of American Foresters.

A.R. Croft and M.D. Hoover.  1951.  The Relation of Forests to Our Water Supply.  Journal of Forestry 49(4):245-249.

Hewlett, J.D.  1979.  Forest water quality: An experiment in harvesting and regenerating Piedmont forests.  Athens, GA: University of Georgia School of Forest Resources Press.

Hornbeck, J.W. and J.N. Kochenderfer.  2004.  A century of lessons about water resources in northeastern forests.  19-32 in A century of forest and wildland watershed lessons.  Ice, G.G. and J.D. Stednick [Eds.].  Bethesda, MD: Society of American Foresters.

Ice, G.G., P.W. Adams, R.L. Beschta, H.A. Froehlich, and G.W. Brown.  2004.  Forest management to meet water quality and fisheries objectives: Watershed studies and assessment tools in the Pacific Northwest.  239-262 in A century of forest and wildland watershed lessons.  Ice, G.G. and J.D. Stednick [Eds.].  Bethesda, MD: Society of American Foresters.

Ice, G.G. and J.D. Stednick [Eds.].  2004.  A century of forest and wildland watershed lessons.  Bethesda, MD: Society of American Foresters.

Ice, G., H. Thistle, and R. Karsky.  2008.  Reduction in chemical drift by a vegetative buffer.  In Proceedings of the AWRA 2008 summer specialty conference - Riparian ecosystems and buffers: Working at the water’s edge (CD).  Okay, J. and A. Todd [Eds.].  ISBN 1‑882132‑77‑7.  Middleburg, VA: American Water Resources Association.

Jackson, C.R., G. Sun, D. Amatya, W.T. Swank, M. Riedel, J. Patric, T. Williams, J.M. Voes, C. Trettin, W.M. Aust, R.S. Beasley, H. Williston, and G.G. Ice.  2004.  Fifty years of forest hydrology in the Southeast.  33-112 in A century of forest and wildland watershed lessons.  Ice, G.G. and J.D. Stednick [Eds.].  Bethesda, MD: Society of American Foresters.

Logan, B. and B. Clinch.  1991.  Montana forestry BMPs: Forest stewardship guidelines for water quality.  EB0096.  Bozeman, MT: Montana State University Extension Service.

McBroom, M.W., R.S. Beasley, M. Chang, and G.G. Ice.  2008.  Storm runoff and sediment losses from forest clearcutting and stand re-establishment.  Hydrologic Processes 22(10):1509-1522.

National Research Council (NRC).  2008.  Hydrologic effects of a changing forest landscape-Prepublication copy.  Washington, DC: The National Academy Press.

Swift, L.W. Jr. and J.B. Messer.  1971.  Forest cutting raise temperatures of small streams in the southern Appalachians.  Journal of Soil and Water Conservation 26(3):111-116.

Wilkerson, E., J.M. Hagan, D. Siegel, and A.A. Whitman.  2005.  The effectiveness of different buffer widths for protecting headwater stream temperature in Maine.  Forest Science 52(3):221‑231.

Williams, T.M., D.D. Hook, D.J. Lipcomb, X. Zeng, and J.W. Albiston.  2000.  Effectiveness of best management practices to protect water quality in South Carolina Piedmont.  In Tenth Biennial South Silvicultural Research Conference.  General Technical Report SRS-30.  Asheville, NC: USDA Forest Service, Southern Research Station.

Figure 1.   Forestry BMP Effectiveness Studies in the Southeastern United States
(from Dr. Masato Miwa, International Paper Company)

Figure 2.   Draft Map of Forestry BMP Effectiveness Studies in the Western United States
(from Dr. George Ice, NCASI)

Figure 3.   Draft Map of Forestry BMP Effectiveness Studies in the Eastern and
Midwestern United States (from Dr. Erik Schilling, NCASI)