We took an airplane yesterday, flying north west
of Roseburg. During our short 1.5 hour flight, we saw literally hundreds of landslides
coming from clearcuts and their roads. We saw about 4 natural landslides in forested
areas. We couldn't see every natural slide from the plane - but I know for sure,
there were not hundreds. There were many, many more landslides from clearcuts. The
slopes were just dripping with streaks of mud and slides. The draws and valleys were
ripped wide open, and oozing with the displaced soil and rocks.
We flew over the Hubbard Creek slide, which killed four people last month. It was not the biggest or most impressive slide we saw. It started high on a vertical slope - skinny, small. We saw it widen, saw the path it took through some old-growth, and saw pieces of the house, and scattered wreckage where it went through the rural residential area. There are no words to describe...
We flew a few miles further north, over the
Yellow Creek Timber Sale, once a beautiful forest of huge trees in an ancient forest
preserve (LSR), now completely clear-cut, and dripping landslides down her sides
like tears.
Sadly, we must add Yellow Creek to the 318 "salvage" sales cut last summer, that we now know have produced massive landslides. So far, they were concentrated in unit 1, a dozen smaller slides, and two major ones, hundreds of feet long.
We continued flying north, and just before highway 38 we flew over the Brush Creek timber sale - industry owned land. Sun Studs clear-cut it in 1995, with government cooperation on a "fish restoration" project. We, the public, "cooperated" when BLM cut our LSR logs and gave them to Sun Studs to put in Brush Creek where it flowed next to their new clear-cut. Sun Studs then yarded their own logs away. Our logs were used for 'in-stream structures' for fish habitat. Loopholes in the laws allowed Sun Studs to clear-cut right up to the edge of Brush creek and the in-stream structures, opening the fish bearing stream to full sunlight.
It was one of the most landslide devastated places we
saw. The VERY steep slopes, with new side-hill roads, were pouring their soils down
the mountain. The in-stream structures have either been washed away or covered with
mud and silt. Governor Kitzhauber called this a 'model' project of industry and government
cooperation to save fish.
Then we turned West - following Highway 38 and the main stem of the Umpqua, through the heart of industrial forest land in the coast range. We soon stopped oohing and ahhing and pointing at the landslides - there were too many of them. We saw old cuts, and new cuts with their new riparian buffers ripped through with slides. Obviously, no amount of riparian buffer will hold back a land slide if it comes from a steep enough slope.
We were amazed (I don't know why) to see active logging shows on steep mountains, with the trees freshly cut laying on the ground. The yarder was sticking up at the ridge top, dragging the logs up the hill. We saw the loaders loading up logging trucks. We saw one active logging show on a very steep slope that ended just a few hundred yards up the hill from a nice old farmhouse.
The flight was a moving and alarming experience. Thank you Lighthawk, for providing the service.
francis
Scientific studies for more than twenty five years have proven over and over that clearcutting causes very large increases in landslides. The American Fisheries Society is the independent professional society for fisheries scientists. The Society's 1991 publication Influences of Forest and Rangeland Management on Salmonid Fishes and their Habitats [William R Meehan ed, AFS 1991] reviewed all the scientific literature, and concluded:
"The frequency of mass erosion [landslides, debris flows, earthflows, etc] is strongly linked to the type and intensity of land treatment in the basin. Although most mass movements are associated with roads and their drainage systems, many originate on open slopes after logging has raised soil water tables and decreased root strength." [Meehan, 1991 page 194]
"The increase in mass movement due to clear cutting varies widely, ranging from 2-4 times in Oregon and Washington...to 31 times in the Queen Charlotte Islands... An increase of 6.6 times...is probably closer to the norm." [Meehan, 1991, page 194]
In other words, scientific studies show that clearcutting, at a minimum, doubles or quadruples the number of landslides, and the average effect is much greater yet.
The 1996 Storms
Intense winter storms in February and November/December of 1996 have triggered thousands of landslides in the Pacific Northwest, concentrating in the Coast Range of Oregon, the Western Cascades of Oregon, in SW Washington, and most recently in Umpqua and South Coast Regions of Oregon. Similar patterns of landslides occurred in the Skagit Basin in Washington in the early 1990s, and throughout the region in the large storms of 1964 and 1955.
The 1996 storms affected areas that have been subject to very high levels of logging and roadbuilding in recent decades, particularly the 1980s. The absence of really severe storms over the last fifteen years or so means that many logged hillsides and roads are only now being hit for the first time. And the resulting picture is devastating. Aerial reconnaissance surveys documented more than 650 landslides in the February storms alone. More than 70% of these landslides were associated with recent clearcuts.
The report titled Aerial Reconnaissance Evaluation of 1996 Storm Effects on Upland Mountainous Watersheds Oregon and Southwest Washington, [Weaver and Haggans, 1996] is attached. Weaver and Haggans found that:
"Within any particular area there was an obvious and visible association between roads and landsliding, and between recent harvesting (clearcutting) and landsliding." [Weaver and Haggans, 1996, page 7].
A study, which says that clearcuts are almost three times more likely to produce landslides, was published by USDA Forest Service scientists in the July, 1975, Geology Journal. Authors F.J. Swanson and C.T. Dyrness write:
"Slide erosion from clear-cut areas in the unstable zone has totaled 2.8 times the level of activity in forested areas of the unstable zone." The study was conducted in the H.J. Andrews Experimental Forest. The study further reports:
"Deforestation of hillslopes results in a number of changes that may increase the probability of shallow failures of the soil mantle:
These factors have been cited as contributing to a period of increased slide frequency after deforestation, especially between the time of decomposition of root systems of killed trees and establishment of stabilizing roots by incoming vegetation. This temporal relationship between deforestation and slide activity as also been observed in the Andrews forest, where most hillslope failures in clear-cut areas occured in the first 12 years after cutting."
INTRODUCTION
The Oregon Coast Range, especially the Mapleton Ranger District of the Siuslaw National Forest, is notoriously prone to landsliding during infrequent (1 in 10- to 20-year) heavy rainfall events. Since the mid-1970s, it has been the Forest Service's standard practice to inventory landslides following a major slide-initiating rainfall event. These studies have found that the great majority of slides result from logging roads, clearcut logging, or both. Further, slides from logging roads are generally larger by an order of magnitude or more than slides from clearcut units.
Though the rate of logging and roadbuilding on federal lands in the Coast Range has lessened over the past ten years as a result of court orders to protect streams from sliding and the President's Northwest Forest Plan, a large infrastructure of roads and cut-over hillsides remains. Like a time bomb waiting to go off, these roads and clearcuts are a legacy from the past that, without rehabilitation and obliteration, threaten the future health of salmon streams, water quality, and, in some locations, residential dwellings.
METHODS AND MATERIALS
Two days following the Storm of 1996, AFSEEE sponsored an aerial inventory of landslides triggered by the massive rains in the Mapleton Ranger District area. Following the classification system used in previous inventories, AFSEEE preliminarily assigned each landslide to one of three types -- road-related, clearcut-related (referred to here as "in-unit"), and natural. Some landslides that appear to be in-unit might have been triggered by road drainage that is not apparent from the air. Thus, a more accurate determination of causality awaits a ground-based survey.
The aerial survey was done using a two-person Cessna airplane provided by Lighthawk, the "environmental Air Force." The survey was performed in six transects over the course of 2.5 hours from the north boundary of the Mapleton District to the Umpqua River at the district's south boundary. Weather conditions for the survey were excellent. Photographs of several landslides were taken.
RESULTS AND DISCUSSION
A total of 185 landslides from the February 1996 storm was recorded. Of these, 114 were in-unit slides, 68 were road-related slides, and 3 were natural, in-forest slides. Consistent with earlier inventories, we believe the number of natural slides to be somewhat greater than recorded given the difficulty of seeing such slides, especially small, riparian slides, from the air. However, even if natural slides are several times more prevalent, the vast bulk of sliding is logging-related.
On average, road-related slides appeared substantially larger than in-unit slides, which is also consistent with previous studies. Road-related slides also appeared to cause more damage to streams; several large debris torrents were triggered by road failures.
The reduction in slide frequency in the southern-most portion of the district appears related to the path of the rainstorm. There was no evidence of sliding on several very steep and dissected hillsides that were recently logged between the Smith and Umpqua Rivers -- an area that is very unstable -- suggesting that the path of the rainstorm was north of this location. News reports of the storm's path and damage appear to support this conclusion.
SUMMARY
A legacy of logging on steep slopes in Oregon's Coast Range threatens salmon streams, water quality, and even people's homes. Notwithstanding past efforts to stabilize roads, large landslides from roads still cause widespread damage to coastal streams. It appears that only a program of road obliteration that returns hillsides to their original contour will prevent sliding from occurring during periodic high rainfall events.
Clearcut logging on steep hillsides triggers most landslides in the Coast Range. Notwithstanding efforts by the Forest Service during the past decade to protect landslide- prone sites, in-unit slides are still frequent across the landscape.
Landsliding during infrequent, catastrophic events is often excused as an "act of God." This survey shows that it is acts of man that create the conditions nature exploits. The Forest Service should develop training materials to ensure that its land managers thoroughly understand the consequences of roading and logging in erosive areas subject to unpredictable, episodic high rainfall events. We cannot afford to have managers learn by their own experience; they need to learn their predecessors' lessons.
 |
Prepared by William Weaver, Ph.D.
Pacific Watershed Associates
Geomorphic Studies - Erosion & Sedimentation Processes - Wildland Hydrology -
Erosion Control
Box 4433, Arcata, California 95521 (707)-839-5130
Purpose
On February 14-15 Pacific Watershed Associates of Arcata, California, conducted an aerial survey of storm damage in the Middle Oregon Cascades and Middle Coast Range. The purpose of the reconnaissance was to provide an immediate post-storm assessment of the nature, magnitude and spatial distribution of watershed erosion and impacts to stream channels which serve as habitat for anadromous salmonids, resident trout and other riparian resources. A second objective was to visually determine the association between watershed variables, land use practices and landslide activity triggered by the storm. A total of 9 hours of aerial reconnaissance was conducted during which I viewed a portion of a number of watersheds and smaller sub-basin areas. This review was not intended to be a complete or thorough inventory of storm-related landsliding and watershed impacts.
Areas reviewed
Rivers and watershed areas reviewed in the Cascades included portions of the North Fork Middle Fork Willamette River, the McKenzie River and the Santiam River basin. In the Coast Range, a number of mountainous areas were also reviewed, including portions of the Smith River basin, the Suislaw River, the Alsea River, the Yakina River and several small coastal watersheds which drain directly to the Pacific Ocean between Reedsport and Waldport
Summary Observations
- Watersheds and stream channels in many drainage basins of the both the middle Cascades and the middle Coast Range have been hard hit by the February 1996 storm. Sub-basins which were observed during this reconnaissance to be especially impacted include Quartz Creek and Blue River (tributaries to the McKenzie River), Quartzville Creek (a tributary to the Middle Santiam River), Hadsall and Knowles Creeks (tributaries to the Suislaw River), and Lobster Creek (a tributary to the Alsea River), among others....
- The cause of variations in the extent of watershed damage from basin to basin is under investigation. The greatest concentration of landsliding and watershed damage occurred or originated in recently clear cut areas and in areas with logging roads built on steep slopes. However, not all managed areas were heavily impacted
- It appears that the greatest damage occurred in watersheds with a combination of steep slopes and/or unstable bedrock geology, recent harvesting high road densities (or roads built on steep slopes), and within an altitude range where precipitation intensities were probably the greatest (1500 feet for the Coast Range and 3000 feet for the Cascades). In the absence of recent land management (roads and clear cutting) it appears that similar watershed areas experienced much less severe damage.
The most heavily impacted watersheds in the Cascades ranged in elevation between 2,500 to 3,500 feet Watersheds lower than 2,500 feet and higher than 3,500 feet elevation showed substantially less damage. Higher watershed areas are less managed and probably retained their snow cover during the storm while lower watersheds do not contain as many managed steep, unstable slopes. In the Coast Range, most damage occurred on managed lands near the 1500 foot elevation, level where slopes are steepest and precipitation was probably the greatest
Landsliding is the most visible erosion process that has been triggered and dramatically accelerated by the storm. Approximately 60 landslides were documented in the Cascades and twice that many in the Coastal watersheds. An estimated total of 500 landslides were observed during the limited reconnaissance.
There were many landslides in the 100 to 1,000 cubic yard size range, a large number in the 1,000 to 5,000 cubic yard size range and a few very large slides estimated to exceed 10,000 cubic yards in volume.
Most landslides delivered large quantities of sediment to stream channels and many traveled as "torrents" or mud flows for hundreds or thousands of feet down hillslopes and stream channels before coming to rest in larger stream beds and river channels. It is estimated that over 80% of the observed landslides delivered sediment directly to stream channels.
The frequency of landslides within recently clear-cut areas and along forest roads was much higher than for comparable watersheds in middle elevation wilderness or unmanaged areas. Increased rates of landsliding was clearly associated with forest land use activities.
- Some watersheds displayed only localized landsliding while others showed widespread landsliding activity. In both areas, landsliding was almost always associated with managed areas (clear cuts and/or roads).
- Several 40 to 50 year old harvest areas in the Coast Range also showed considerable debris torrenting and landsliding in steep stream channels.
- Logging roads in both the Cascades and the Coast Range were common sites for landslides triggered by the storm. Most of these landslides also delivered sediment to stream channels. Many forest roads have been heavily damaged and will require substantial efforts if they are to be rebuilt. Many probably should not be rebuilt, but will still require considerable preventive treatment to stabilize them against future erosion. Roads located on steep slopes and near stream channels were much more prone to landsliding and failure than roads built on moderate or gentle upland terrain.
In some watersheds, stream channels have been heavily impacted with sediment and logs.
- Debris torrents in steep channels have scoured the stream bed down to bedrock while large log jams and sediment accumulations have been deposited at the mouths of tributaries.
- A number of main tributary streams have experienced considerable sediment deposition as material is washed in from the steeper channel systems.
- Many of the riparian stands of hardwood trees along major streams and rivers have not been uprooted or destroyed. Flood flows were often high enough to move and rearrange large logs and organic debris within the channel system, and to introduce new material (sediment and logs) from the steeper tributaries, but not enough to strip flood plains of their vegetative cover.
Wilderness areas and unmanaged slope areas showed comparably little storm damage and impacts.
- The mid-elevation portions of six different wilderness areas and several unmanaged watersheds were also reviewed to compare storm impacts against managed areas. A few recent landslides were observed in steep stream channels in wilderness slopes but the rate of landsliding (landslide frequency) was very low compared to managed areas especially compared to those which had been recently clear cut. Some wilderness areas contained no visible landslides or stream channel damage.
- In the most impacted wilderness basins observed in the coast range less than 10% of the debris torrent tracks were active this storm. Midslope landslides in wilderness areas were not observed. In contrast, in some recently clear cut basins in the nearby Siuslaw River basin, up to 70% of the headwater swales and torrent tracks were activated by this storm, and midslope landslides were common.
Conclusions
Some tentative conclusions can be drawn from these observations:
- Land use (clear cutting and road building) in some areas and sub-basins has a high risk of resulting in landsliding and stream channel damage. Some types of land use activities may be inappropriate in these areas
- Removal of vegetation from steep swales in watersheds where debris torrenting is a common landsliding process will greatly increase the number of such landslides which occur during a large storm.
- Road construction by side casting on steep slopes is hazardous and will result in greatly increased rates of landsliding and stream sedimentation.
- When the next storm occurs (whether or not it is of lesser or greater magnitude), additional landsliding and erosion will occur in these watersheds. Many existing logging roads show signs of pending slope failure and active clearcutting in these watersheds is occurring on terrain shown to be sensitive to increased landsliding.
- Stream channels will continue to be impacted by the erosion and landsliding which was triggered by this storm event as bare soil areas gully, remaining unstable landslide material continues to fail and move down slope, and sediment deposited in headwater streams is re-eroded and moved down into the larger streams and river systems.
- The impacts of coarse sediment introduced into streams and rivers by landsliding and road failures may persist for decades, depending on transport rates.
- The impacts to lower, larger streams and rivers may actually increase over the near term (next several years) as sediment from the headwater areas is moved downstream and deposited in lower gradient reaches.
- Restoration measures for sediment already delivered to the stream system by landslides and road failures will be impractical and not cost-effective
- Many road systems and road segments that have not yet failed can still be pro-actively treated (upgraded or decommissioned) so they do not fail during future storms. This type of treatment will be most effective and cost-effective. In areas that are most at risk (roads on steep slopes, roads built near stream channels, roads built by side casting, and abandoned and unmaintained roads).
- Prevention of landslides from harvested (clear cut) slopes is dependent on the recognition and avoidance of sites with high risk characteristics. For example, clear cutting some watershed areas exhibiting steep headwater swales appears to have dramatically increased landslide activity. Once cut, the slopes are vulnerable for a period of years until they are well vegetated. In sensitive watersheds or where downstream aquatic resources are threatened, clear-cut harvesting on these sensitive slope locations is best avoided.
- Future restoration and prevention work should be prioritized based on the "value" of down slope and downstream aquatic resources. It is likely that working within stream channels in an attempt to restore productive habitat and conditions will be futile until future up slope sediment sources are addressed through control and prevention.
The Oregon Department of Forestry (ODF), in charge of enforcing rules on industrial forest land, has refused to acknowledge that the recent flooding and landslides is exasperated by clear-cuts. However, in 1995, it appears they said something different.
In 1995, the ODF stated that clearcuts could increase landslides (or 'mass movement') by up to 40 times. The 1995 ODF publication "Cumulative Effects of Forest Practices in Oregon: Literature and Synthesis" as follows:
6.5.4.2 Timber Harvest, page 71:
"Timber harvest in sensitive areas has also been associated with an increased incidence of mass movement. Clearcut harvest and/or slash burning on steep slopes may increase failure rates by two to forty times over rates on undisturbed sites.""With regard to mass soil movements, harvest appears to affect soil and slope stability primarily in two ways: (1) through reduction in evapotranspiration on harvested sites, and (2) through eventual death and decay of root systems of harvested conifer trees. Soil drainage in harvested areas may also be interrupted or concentrated."
6.5.6 Significance of Mass Movement, page 73:
"The occurrence of mass movements results in various kinds of on- and off-site damage. On-site damages include: loss of the current timber stand, road blockage or failure, and loss of soil and associated productivity. The most significant off-site damage is the sedimentation of stream channels and its ensuing problems for water quality and aquatic habitat"."Mass movement rates as the most significant means through which sediment is delivered into stream channels for most mountainous areas."
7.11 Conclusions- C. Water Quality
"In general, forest practices have been shown to have the potential to increase sediment delivery to stream channels. The magnitude and timing of the delivery depends on the source (i.e. mass wasting, debris torrents, or fine sediment), and the type of forest practices (e.g. roading, yarding, site preparation) causing erosion and conditions between the source and receiving waters".