Sediment Source Investigation
and Sediment Reduction Plan for
the Bear Creek Watershed,
Humboldt County, California



prepared for
The Pacific Lumber Company
Scotia, California
Pacific Watershed Associates
Arcata, California
(707) 839-5130
April, 1998
Sediment Source Investigation and Sediment Reduction Plan
for the Bear Creek Watershed, Humboldt County, California
Table of Contents

I. Introduction 

II. Methods  III. Results and Discussion  IV. Implementation Plan for Erosion Prevention and Control V. Bear Creek Monitoring Plan

VI. Summary and Conclusion

VII. References

VIII. Appendices A, B



Sediment Source Investigation and Sediment
Reduction Plan for the Bear Creek Watershed,
Humboldt County, California


The Bear Creek watershed is an 8.0 mi2 tributary to the Eel River located approximately 8 miles upstream from Scotia, California in Humboldt County (Figure 1). In late December 1996 and early January 1997 a series of intense rain storms triggered widespread erosion and sediment delivery to Bear Creek and other watersheds of the northcoast. Bear Creek and several adjoining drainages exhibited significant geomorphic change, including debris landsliding, debris torrents, channel aggradation, bank erosion and loss of the established hardwood riparian canopy.

At the request of the Pacific Lumber Company (PL), Pacific Watershed Associates (PWA) prepared an analysis of sediment production and yield in the Bear Creek watershed. The inventory and analysis was to identify sources of erosion, and to distinguish between natural, uncontrollable sediment sources and management-related sediment sources which are amenable to prevention or control. The goal of the assessment was to identify remedial measures and practices which could reduce future sediment production and delivery to streams in the watershed.

This analysis and report describes the effects of the 1996/7 storms in the context of changes which have occurred during previous storms and floods in the watershed. Old aerial photos from 1947 document conditions in the basin at one point-in-time largely before land management activities had occurred. Stereo aerial photography from a number of later years and decades was used to identify watershed changes that have occurred in Bear Creek over the subsequent 50 years of storms and land management. Analysis of historic photos is useful for identifying the nature, magnitude and significance of the changes which occurred in Bear Creek in the 1996/97 storm.


Field inventories were conducted in the Bear Creek watershed to provide ground-truthing of measured landslide areas and stream channel changes which were documented in the aerial photographs. Field surveys were also conducted to determine past sediment production and yield from the road network (including both landsliding and fluvial erosion), as well as the location and volume of future preventable road-related sediment sources. A database of these inventoried sites contains information on past and future erosion as well as recommended erosion control and erosion prevention treatments.


Finally, a brief monitoring plan to track the biological and physical recovery of Bear Creek has been proposed. This plan consists of existing monitoring strategies and protocols as well as additional channel monitoring activities to identify the rate and nature of future changes in channel morphology, aquatic habitat and biological productivity. These permanent monitoring stations can be re-occupied, and variables remeasured, at periodic intervals in the future to document trends and patterns of recovery.



Figure 1. Location map of the Bear Creek watershed.

The Bear Creek watershed is an extremely steep, relatively short fourth order basin draining north into the Eel River. Ridge-top areas around the margins of the watershed are fairly gentle but slopes quickly become steep in the interior of the basin. High rate of tectonic uplift and rapid channel down cutting have resulted in a deeply dissected terrain. Steep and long inner gorge slopes occur along a large number of tributary streams in the basin. The watershed is located in a geologically complex and active region and this complexity is reflected in the rock types and geologic structures within the region.


Geologic setting of the Bear Creek watershed

The geologic and geomorphic development of the region has been controlled by Neogene uplift and deformation of the Coast Ranges associated with plate tectonics (the convergence of the North American plate margin with the Juan de Fuca plate at the south end of the Cascadia subduction zone), the migration of the Mendocino triple junction, and the San Andreas fault system. Recent interpretation has placed the Mendocino triple junction in the vicinity of Honeydew, located approximately 10 miles southwest of Bear Creek (McLaughlin and Clark, 1992). During the past 28 million years most of the bedrock underlying the Bear Creek watershed has undergone extreme degrees of folding, faulting, fracturing and shearing resulting in highly deformed rocks and structures (Atwater and Molnar, 1973). Uplift rates have resulted in deeply incised streams. Shear zones appear to be localized but were not studied in detail and may or may not be active. A mapped fault crosses the lower portion of the watershed and is associated with the eastern end of the Russ Fault (Clark, 1992).


The majority of the Bear Creek watershed is underlain by the Coastal belt of the Franciscan Complex. Only a small area in the lower portion of the watershed, at the confluence with the Eel River, is underlain by undifferentiated units of the Wildcat Group. Exposed undifferentiated units of the Wildcat Group consisted of conglomerates, soft sandstones and shales which were in fault contact with rocks of the Franciscan Complex.


The Coastal belt of the Franciscan Complex is divided into three units, referred to by some as tectonostratigraphic terranes, two of which are found in the Bear Creek watershed (McLaughlin and Clark, 1992). The Coastal terrane of Late Cretaceous, Paleocene and Eocene age underlies most of the upper watershed and is in fault contact with the fault bound Yager terrane of Paleocene and Eocene age which underlies the area between the Wildcat Group of Late Miocene to Middle Pleistocene age and the Coastal terrane.


The Yager terrane of the coastal belt, initially named by Ogle (1953), is composed mainly of sandstone, mudstone, shale and conglomerate and comprises approximately one tenth of the watershed. Rocks of the Yager terrane are often difficult to distinguish from the Coastal terrane rocks. Yager rocks are generally somewhat softer than the Coastal terrane rocks and rock outcrops are less common. In near surface exposures, such as in less than 5-foot high roadcuts, the Yager rocks usually show highly fractured and weathered rock fragments in a soil matrix where rock structure such as bedding planes are typically absent.


The Coastal terrane is composed of predominantly sandstone and argillite (derived from mudstone and shale). Generally, outcrops consist of thick beds of sandstone inter-bedded with thin beds of argillite. Blocks of all sizes (up to map-scale), of sandstone and argillite are found enclosed in a fragmented and sheared matrix. The sandstone is commonly white streaked with laumonite veins. Other rocks found in the Coastal terrane are conglomerate and less frequent occurrences of basaltic intrusive rocks, breccias and flows as well as limestone and chert. The Coastal terrane has undergone extensive disruption as evidenced by pervasively sheared, tightly folded and broken rocks. Outcrops observed in road cuts along the mid to upper hillslope position, exhibited evidence of shearing and extensive deformation. Polished and striated rock surfaces (slickensides), off-set bedding planes, fault gouge, and generally disrupted and broken features were common. These features influence slope stability. Focused geologic mapping, not within the scope of this study, would be required to determine the orientation and possible structure and connectivity of these zones.


Many landforms throughout the region exhibit characteristic form of ancient deep-seated and shallow mass movement processes. Some percentage of these ancient landforms were probably initiated by earthquakes (seismic shaking) and others by intense winter rain storms. Colluvium and deep ancient landslide deposits mantle many hillslopes throughout the watershed, chiefly in swales and at lower slope and streamside positions. Colluvium is potentially unstable in most slope positions where slope gradients exceed 20 degrees (about 40%)(Hunt, 1994).


Slope stability

The Coast Range watersheds of Northern California display some of the most unstable terrain in the Pacific Northwest. Factors influencing slope stability on the northcoast have not been quantified, but include:


1) Geologic structure: dip of beds (dipping down slope nearly parallel to or less than the slope inclination); fractures; weak beds inter-bedded with competent beds; faults; shear zones; and surfaces of weakness are all contributing factors present on the northcoast and each lends to decreased slope stability.


2) Material strength: decomposition of materials over time reduces strength; chemical weathering; diminished rooting strength of vegetation following vegetation removal (by fire or harvesting) is important in shallow soils overlying bedrock; widespread, deep colluvium deposits located on steep slopes are factors lending to inherently unstable conditions. Likewise, road and landing construction can change the distribution of mass in steep swales and on potentially unstable hillslopes.


3) Seepage forces reduce resisting forces and increase driving forces. Rainfall and rising water tables, influenced by type and density of vegetation, drainage characteristics of the soils and geology, slope gradient, and rainfall events (intensity and duration) all influence the amount of infiltration and, in turn, groundwater levels and seepage forces.


4) Weather: intense rainfall events are common on the northcoast, and intense and prolonged precipitation events are known to be associated with increased rates of landsliding, especially when high antecedent ground water conditions exist. The area can frequently experience 5 inches of precipitation in 24 hours. During the 1996/97 storm, it is estimated that Bear Creek received around 10 inches of rainfall in a 36 hour period.


5) Slope gradient:

A) Geologic materials have characteristic slope inclinations at which they are stable and just barely stable: residual soils - 30 to 40 degrees; colluvium - 20 to 30 degrees (Hunt, 1994). Slope gradients exceeding these limits are common in Bear Creek, especially along the network of incised stream channels.


B) Slope inclination is increased by road cuts and sidecasting, as well as by processes associated with stream bank erosion, and the redistribution of material during landsliding events.


C) Tectonic uplift rates can cause both subtle and rapid increases in slope inclination over time. Streams incise in response to lowering base levels, producing steep slopes and steep inner gorges along drainages. Dumitru (1991) estimated uplift rates in the King Range, 15 miles to the west of Bear Creek, to be as high as 12 feet per 1,000 years. The extremely rapid uplift rates in the Bear Creek watershed are estimated to be on the order of 5 to 10 vertical feet per 1000 years.


6) Seismic activity, especially the proximity to major earthquake sources, can be an extremely important factor affecting slope stability and the occurrence of landslides. The Bear Creek watershed has experienced MMI VI or greater earthquakes 10 to 14 times since 1900 (Dengler and McPherson, 1993). MMI VII or greater earthquakes can cause widespread landsliding in mountainous terrain (Keefer, 1984)). Earthquakes also have a profound effect on the movement of subsurface water through deep colluvial soils, such as those found in Bear Creek, by cutting off macro-pores and increasing pore water pressures in potentially unstable, steep hillslope areas. The effects of large earthquakes, such as those of 1991-1992, are most likely to express themselves during the first high magnitude storm following severe seismic shaking (ie., the winter of 1996/97).





The sediment source assessment for Bear Creek was comprised of three elements: 1) an aerial photo analysis of mass wasting, 2) field sampling of stream side debris slides, torrent tracks, bank erosion and channel deposits, and 3) a complete field inventory of all road related sediment sources.


Aerial photo analysis

To understand the relationship between landslide occurrence, storm/flood events, geomorphic/geologic conditions and land use, we analyzed landslide occurrence in the Bear Creek watershed from five different sets of vertical aerial photography: 1947, 1966, 1974, 1994 and 1997. Each new landslide or erosional feature which appeared on the photographs was assigned a unique site number and characterized using a variety of factors. These factors included:

1. Feature type (debris landslides, debris torrent source areas, deep seated landslides, debris torrent tracks, bank erosion, enlarged channels, stream crossings and gullies),

2. Certainty of interpretation (definite, probable, questionable),

3. Feature dimensions (length, width),

4. Aspect (compass direction),

5. Sediment delivery (estimated <25%, 25-50%, 50-75%, 75-100%),

6. Type of stream receiving deposits (perennial, intermittent, ephemeral),

7. Land use history at initiation point (road, skid trail, tractor clearcut (<15 and >15 years old), cable clearcut (<15 and >15 years old), partial cut, advanced second growth, unmanaged),

8. Geomorphic association (inner gorge, swale, break-in-slope, headwall, etc.)

9. Vegetation classification (conifer forest, mixed conifer, oak woodland, grassland),

10. Bedrock type at point of initiation,

11. Elevation at point of initiation, and

12. Hillslope steepness passing through initiation point.


Debris torrent source areas were classified as debris landslides which turned into debris flows and scoured some length of natural stream channel downstream from the origination point. Debris torrents typically scoured the bed and banks of the channel in the higher gradient reaches (called torrent tracks) and then deposited their load of sediment at stream junctions or in lower gradient reaches (enlarged channels). During the analysis phase of the project, landslide lengths, torrent channel lengths, bank erosion lengths and enlarged channel lengths measured from the aerial photography were corrected using a multiplier based on slope gradients measured from topographic maps. Landslide depths were applied based on a field sampling of over 40 slides.


Field sampling

Approximately 5 miles of stream channel were surveyed in the field to identify and quantify dimensions (length, width and depth) of stream side landslides, bank erosion, torrent track scour and the volume of accumulated sediment deposited in the tributaries and the main stem of Bear Creek. Landslides were located on aerial photos and measured using tape and range finder. Average landslide depth, volume and sediment delivery (%) were estimated for each feature by at least two people.


Torrented stream channels (channels scoured by debris flows) were evaluated for bank scour (length, height and depth) as well as erosion of the channel bed. In lower gradient channels, the volume of recent sediment deposits was estimated (length, width and depth) based on field evidence, as was the subsequent remobilization (erosion) volumes of the 1996-97 flood deposits. In these aggraded areas, bank erosion caused by lateral channel migration impinging against natural hillslopes was also quantified (length, height, depth, volume).


Results from the field sampling program were used to assign landslide depths, torrent scour volumes (yds3/ft), bank erosion rates (yds3/ft), and aggradation rates (yds3/ft) for features and sites identified on aerial photos but not visited in the field. Approximately 80% of the debris slides which were identified as initiating during the winter of 1996-97 were physically measured in the field. A simple linear regression of landslide surface area against landslide depth for 40 inner gorge and stream side debris landslides measured in the field (D = 4.35x10-5 A + 2.76; r2 = 0.47) provided a mechanism for assigning depth values to each landslide identified on aerial photos from 1947, 1966, 1974, 1994 and 1997.


Road Erosion Inventories

A road construction history map was prepared for the Bear Creek watershed using the same five sets of aerial photographs used in the landslide survey. All roads were then field inventoried for past erosion and sediment delivery, including road and landing fill slope failures, stream crossing washouts, stream diversion gullies and surface erosion. Field crews traced each erosion feature down the slope (sometimes for distances up to 500 feet) to determine dimensions (length, width, depth, volume) and past sediment delivery. The cause and age (decade) of each road-related sediment source was recorded, as were a number of geomorphic and land use associations.


The road inventory also included a variety of site information about future (expected) erosion and sediment delivery from the road system. Information included the identification of unstable and potentially unstable fill slopes, potential erosion at stream crossings (from a variety of sources), active gullying, and road surface erosion. Future erosion sources were identified and quantified only if they had the potential to deliver sediment to a stream channel. Thus, cut bank landslides (and other potential and existing erosion sources) were not included in the survey if they would not result in sediment delivery.


For each future sediment source, a variety of site information was collected. Stream crossings were evaluated for the type and adequacy of drainage structures (usually culverts), the potential for stream diversion, the potential for culvert plugging, and outlet erosion. Field measurements (profile and cross-sections) were entered into a computer program to determine the volume of fill in the crossing, and to estimate the amount of fill that would have to be excavated to upgrade the culvert where it is undersized. Each stream crossing was also evaluated for the contribution of road runoff and fine sediment from the road surface and ditch to the stream. The length of road and ditch draining directly into each stream (typically through the ditch to the culvert inlet) from the adjacent road was measured for all stream crossings and other road drainage locations (ditch relief culverts) which were delivering sediment to stream channels.


For potentially unstable fill slopes (sidecast), the volume of unstable fill was measured, as was the distance to the nearest stream channel and the gradient of the hillslope below the instability. This, and other data, provided information useful for determining the likelihood and potential magnitude of the unstable material that, if it were to fail, would be delivered to the stream.


The data on future sediment sources also included information on the likelihood of the potential erosion, erosion and delivery volumes, recommended erosion-prevention treatments, equipment and labor requirements, estimated treatment costs, and treatment priorities for each identified site. These and other data have been assembled in a computer data base and described in an implementation plan (presented in a later section of this report) for road-related erosion prevention and erosion control.



Results and Discussion


Land use history

Logging in the Bear Creek watershed began in the early 1940s, primarily in the lower watershed, with a minor incursion along ridge-top areas and below grasslands in the western middle portion of the basin (Map 1). The lower one mile of the main access road along the lower main stem of Bear Creek (the same road that is still used today) had been constructed. By 1947, approximately 330 acres, or 6% of the watershed, had been selectively logged (Figure 2 and Map 1).


By 1954, a logging road had been constructed four miles up the basin along the entire length of the main channel of Bear Creek, and extending an additional 0.5 miles up the major eastern tributary stream located at stream milepost 3.0. Adjacent lower hillslope areas within 1000 feet of the main stem were rapidly being tractor logged, with logs yarded downslope to the main truck road along Bear Creek. Remnants of these old roads are still visible along the main channel, but most of the stream crossings and potentially unstable fill materials have long since eroded or failed. In the upper watershed, the main access road on the western side of the watershed had been constructed by 1954. Along the first half of this route (4.75 miles) the slopes above the road had recently been clearcut and tractor logged while steep slopes below the road had been clearcut and cable yarded.


Between 1947 and 1966 virtually the entire remainder of the watershed had been either selectively logged (seed tree) or clearcut (Figure 2 and Map 1) and 90% of today’s road system was in place (Map 2). Over 4,600 acres (90% of the watershed) was cut during this period. At the time of the 1964 storm and flood, substantial portions of the watershed had recently been harvested and recent ground disturbance was at its historic maximum. The extensive ground disturbance is one of the factors which may have contributed to the high rates of landsliding and erosion in 1964.

Road construction during the period from 1947-1966 was equally as intense as timber harvesting during the same period (Map 2). Of the 39 miles of road existing in the watershed today, 26 miles (68%) were built during this period (Figure 3). Numerous landslides and stream crossing failures had closed existing roads in the basin during the 1964 storm, but by 1966 most roads had been reopened and reconstructed. Numerous and substantial segments of the logging road built up the main channel failed or eroded during the 1964 flood and associated debris torrents. Although the entire route had been reconstructed by 1966, it appears to have been finally and permanently abandoned by 1974.


The 30 year period from 1966 to 1997 witnessed considerable re-logging of residual seed tree stands which had been left in the upper and middle portions of the watershed (Map 1). Approximately 1900 acres (37% of the watershed area) was relogged during this period (Figure 2). Most harvesting of slopes over 50% involved limited ground disturbance (compared to the pre-1966 tractor yarding) and involved cable yarding from upper slopes and ridge top areas. Post harvest ground conditions in these areas generally consisted of radiating narrow yarding corridors separated by relatively undisturbed areas of brush and young second growth conifer vegetation. It is estimated that less than 10% of the area was tractor yarded during this period, and these areas were confined to relatively low gradient upslope areas.


To accommodate the conversion to cable yarding methods, a number of ridge-top spur roads were constructed between 1966 and 1997, most within the last 10 years. Approximately 9.5 miles of road (25% of the road network) were built during this period (Figure 3 and Map 2). These roads experienced only minor erosion problems and few landslides during the 1996/97 storm event.


Storm history

As with most other watersheds in north coastal California, flood events in Bear Creek have been the triggering events for large scale and widespread landsliding. Regional flood histories which would be applicable to the lower Eel River basin have been described by Harden (1995), Coghlan (1984) and Helley and LaMarche (1973). A number of large flood producing storms occurred in the late 19th century and these are thought to have been comparable to, or larger than, those of the period 1953 to 1975. These include the "unprecedented" floods of 1861-1862 (Harden, 1995), as well as major events in 1867, 1879, 1881, and 1888. North of the Eel River, the 1890 flood is thought to have exceeded the magnitude of the 1964 event. According to Coghlan (1984) flood peaks on the Eel River during the 1861-62 floods were higher than any flood events in the 20th century. Floods in 1879 and 1888 also exhibited significant flooding in the Eel.


In the 20th century, flood events recorded in the north coast area in 1907, 1915, 1927, and 1937 were locally significant but did not match those in the 1950s and 1960s (Coghlan, 1984). Flood events of 1953, 1955, 1964, 1972, 1975, 1986 and 1996 appear to have been higher and produced greater watershed response than those in the first half of the century. The storms of 1953 and 1972 were centered north of the Eel and produced only moderate runoff events. The 1955, 1964 and 1975 storms tracked over the lower Eel and produced substantial rainfall (16", 15" and 9", respectively, at Scotia). The magnitude of the 1996/97 event on the northcoast and in Bear Creek is estimated to be less than either the 1955 or the 1964 storm, but was clearly intense enough to trigger locally significant landsliding and sediment production.


The erosional response of Bear Creek to the floods of the 1900s appears to be a complex function of both storm intensity (and duration), prolonged antecedent rainfall, as well as land use history. In 1953 and 1955, over 50% of the watershed was still unlogged, and the steepest terrain had not yet been entered. Road construction was confined to the riparian corridor along the main channel as well as several miles in the western headwaters. Landsliding visible on the 1954 aerial photos was largely confined to managed hillslopes, mostly road fillslope failures and one very large inner gorge debris slide. Ten years later, by the time of the 1964 flood event, most of watershed had been logged and 75% of the road network had been built. The 1964 storm triggered widespread landsliding in many watersheds of the northcoast, including Bear Creek, and this high rate of slope failure is thought to be the combined result of high intensity storms and the slope-weakening effects of harvesting (Kelsey et al., 1995). Storms, such as those in 1964 and 1996, are the triggering events for these geomorphic changes.


Sediment sources

Sources of sediment in the Bear Creek watershed include mass wasting (deep-seated landslides, shallow-rapid debris landslides, and debris torrents), fluvial erosion (gullying, channel erosion, and stream bank erosion) and surface erosion. The sediment source investigation included data from three investigation techniques designed to identify the frequency and magnitude of these processes in the basin. Each technique provided unique data that, when combined, produces a relatively complete picture of the major mechanisms of sediment production and yield in the Bear Creek watershed.


Mass wasting

An aerial photo analysis was employed to identify large landslides, debris torrents, bank erosion and channel aggradation that could be identified from 1:12,000 scale images. The minimum measurement resolution for features identified on the photos was approximately 20 feet. The oldest images evaluated were taken in 1947, when virtually the entire watershed (except for a small portion of the lower basin) was still unmanaged (no roads, no harvesting).


The 1947 photos were examined for evidence of past mass movement features, to provide a simple measure of background watershed sediment production. A total of 29 debris landslides were identified, ranging in estimated age from less than five years old to about 50 years old. Landslide age was estimated by the degree of revegetation on the landslide scar, and ages were broken down into the age classes <15 years, 15-30 years and 30-50 years. Older features, older than about 50 years, although visible, were not measured.


Table 1 outlines the data collected on pre-1947 mass movement sites and other erosion features that were identified and measured on the old aerial photos. An estimated total of 265,100 yds3 of sediment was delivered to stream channel from these landslide features. Based on the air photo analysis (and using erosion rates from 1997 torrent tracks and bank erosion sites) three visible torrent tracks and two reaches of bank erosion generated another 19,500 yds3 of yield.
Table 1. Pre-1947 (pre-land use) mass movement features in the Bear Creek watershed.

Estimated age


Number of slides


Average length (ft)


Total yield to streams (yds3)

Geomorphic association (#)

Inner gorge


stream side




<15 years


15-30 years


30-50 years




Torrent tracks

l=4,900 ft

Bank erosion

l=3,600 ft



284,600 ds3



The total yield of 284,600 yds3 (about 960 tons/mi2/yr) represents a minimum background sediment yield estimate for the 50-year period. Unfortunately, air photo analysis identifies only the largest, most visible mass movement features. It does not reveal smaller landslides or isolated sites of bank erosion that are hidden by the old growth canopy. Landsliding frequencies in Bear Creek appear to be highly correlated with major storm events, but none of the five events in the preceding 50 year period appear to have been comparable to the 1955 or 1964 storm and flood. Thus, landslide frequencies (as visible on the 1947 aerial photos) may be relatively low as compared to the succeeding 50 years (1947-1997).


Land use in the Bear Creek watershed was just underway by 1947, and much of the lower hillslopes in the lower half of the watershed had been logged by 1954. This phase of harvesting also involved road construction along the main channel and up alongside several of the larger lower-basin tributary streams. By 1966 (the date of the next available air photo coverage), most of the watershed had been logged and nearly all the currently existing road network had been constructed. Analysis of the 1966 aerial photos revealed 117 new landslides in the watershed, most of which were assumed to have been triggered by either the 1955 storm or, more probably, by the 1964 storm event (Figure 4 and Map 3).


Land use activities, including road construction, was significantly reduced during the next two photo periods (1974 and 1994) and storm events comparable to those of 1964 or 1996/97 were absent. As a consequence, only 29 new landslides were recorded by 1974 (Figure 4), and only 34 new sites were identified during the next 20 year period (1974-1994).


Beginning in the early 1990s, residual stands of seed trees were being harvested and cable yarded from steep slopes in the middle and upper watershed areas. Some new road construction had been completed, but these roads were largely built on upper hillslope or ridge-top slopes where erosion and slope stability problems are minimized. Most logging during the period resulted in the removal of large conifers and the retention of understory vegetation and 20 year old conifer regeneration.


The 1996/97 storm event triggered several large debris landslides in the steep upper hillslope areas of the watershed, several of which torrented as debris flows and carried sediment and organic debris through the main channel. Analysis of 1997 aerial photos revealed 43 new landslides in the basin, and the enlargement of eight pre-existing features (Figure 4 and map 3).


Landslide distribution

Landslides in Bear Creek are generally concentrated on steep streamside slopes throughout the watershed (Map 3). Rapid rates of uplift in the basin, combined with erodible bedrock, has resulted in rapid channel incision and steep headwater areas. Landslides are especially common in inner gorge slopes along the main channel and major tributaries, compared to less steep slopes of upland areas and interfluves (ridges).


Sediment delivery

During the air photo analysis, sediment delivery was estimated for each landslide. Landslides which developed high on the hillslope were less likely to deliver a large proportion of their material to a stream channels. Small landslides were also less likely to yield large percentages of their failure materials. In contrast, large debris landslides, and debris slides which turn into debris flows and torrents are relatively efficient mechanisms of sediment delivery.


Estimated sediment delivery ranged from 65% for the large number of slides appearing on the 1966 photos (triggered by the 1964 storm event), to an estimated low yield of 40% for slides found on 1994 photos (a 20 year non-storm period). For comparison, field measurements of 40 landslides triggered by the 1996/97 storm revealed a delivery rate of just under 70%. Rates of sediment delivery may be related to the magnitude of the triggering storm event, with large storms delivering material more efficiently than small and moderate size events. In general, large storms are more likely to produce larger landslides and debris torrents.


Landform Associations with Mass Wasting

In all years of analysis, the great majority of debris landslides (69%) occurred within steep inner gorge hillslope areas of the watershed (Table 2). This association is perhaps the most significant relationship explaining the location of mass movement features in the watershed. Inner gorge slopes are generally defined as slopes which are steeper than 65% which occur below the last (lowest) break-in-slope next to a stream channel. For this study, streamside slopes occupy the same slope positions, but occur on slopes less than 65%. Table 2 lists the geomorphic associations, and landslide frequencies, for each of the photo periods analyzed.
Table 2. Landform associations with mass wasting, Bear Creek watershed

Photo Year


Inner gorge




Swales and Headwalls




Planar slopes


Total number of slides






1974 +1994






Percent of Total



Average landslide length was remarkably consistent from one time period to the next, although landslide volumes seems to vary considerably (Table 3). The average length of 1997 inner gorge landslides was 220 feet (maximum = 820 feet) while those of 1966 averaged 260 feet (maximum = 1,280 feet). For all four years of analysis, average inner gorge landslide length was 250 feet (slope distance). In contrast, streamside landslides (those slides which occurred on slopes <65% in steepness) averaged 130 feet in length, with a maximum length of only 190 feet.


Table 3. Frequency, length and volume of inner gorge and streamside debris slides



Photo year

Number of debris slides
Length (ft)
Average volume (yds3)
Number of streamside landslides
Length (ft)
Average volume (yds3)












Not only were landslides on the lower gradient hillslopes shorter, they delivered much lower volumes of material to stream channels. Thus, the average volume of sediment delivered by steamside landslides was just over 1,300 yds3. In contrast, inner gorge landslides delivered an average of nearly 12,000 yds3 of sediment. Inner gorge landslides which developed during the 1964 storm were the largest, averaging over 16,800 yds3 each. During the following 29 year period of small to moderate size storm events (1966 to 1994), average landslide delivery volume (as identified on the aerial photography) dropped to about 6,600 yds3.


Land use associations with mass wasting

The distribution of landslides inventoried in the Bear Creek watershed, with respect to land use, is shown in Figure 5. Because the entire watershed was harvested by 1966, comparisons with unharvested areas in the same watershed was not possible. However, data from the 1947 (pre-land use) aerial photos clearly shows that the watershed has experienced periods of increased landslide activity, even before human disturbances. These were likely triggered by large storms of the early 1900s, or by seismic shaking, and involved both large debris landslides as well as debris torrents and enlarged channels.


Large landslides - Data from the aerial photo analysis can be used in several ways to suggest how land management has altered landslide processes in the Bear Creek watershed. Visual comparison of the 1947 (pre-management) aerial photos and those from 1966 and 1997 suggest that landslide frequencies have increased considerably. However, the identification of landslides in a relatively undisturbed forest setting (1947), especially after some recovery time has elapsed, is difficult, at best. Many smaller landslides that do not create openings in the forest canopy are not readily visible. Only the larger landslides are clearly identifiable on all aerial photographs.


This suggests at least two simple comparisons of landslide data relative to land management:

1) What is the relative frequency of large and very large landslides in the pre-management and post-management periods?

2) What is the difference in the frequency of large and very large landslides triggered by storms during the two managed periods (pre-1966 and pre-1997)?


The largest debris landslides which had been identified for each of the three photo dates were divided into volume classes for comparison: 1) large: 5,000 to 10,000 yds3, and 2) very large: >10,000 yds3. A comparison of the size and frequency of these large landslides in the three photo periods (1947, 1966 and 1997) reveals the importance of these large debris slides in relation to total sediment production, and suggests the possible role of land use and other factors in determining their frequency of occurrence and average size.


Several factors, such as storm magnitude, elapsed time since the last storm event, seismic history and watershed condition, are likely to affect the geomorphic response of the landscape to a large storm event. Watershed response would most likely be expressed in landslide frequency and size, as well as torrent track formation and the development of enlarged channels (aggraded channel reaches in which riparian vegetation is damaged or lost and canopy openings along low gradient stream channels become apparent on aerial photos).


The management condition of the Bear Creek watershed was significantly different at the time of each photo set. In 1947 only about 6% of the watershed had been logged and roaded. The few managed areas were located in the lower basin and along the western ridge-line. Watershed disturbance was still negligible, and most of the basin was comprised of old growth conifer vegetation.


In contrast, by 1966, most of the watershed’s forests had been harvested using a combination of both tractor yarding and cable yarding. The management style for this early logging was typical for most areas of the north coast and can be described as "intense," with substantial ground disturbance, little protection of stream channels and riparian zones, extensive road construction, and little or no recognition of the potential influence of harvesting on inner gorge slope stability.


By 1996, substantial portions of the watershed were covered with 20 to 30 year old conifer regeneration. Over 60% of the watershed consisted of vegetation over 15 years old. Areas first harvested 50 years earlier in 1947 were already classified as advanced second growth. Residual stands of seed trees were being harvested, primarily using cable yarding, on the eastern side of the watershed. As of 1997, roughly 37% of the basin had been relogged and consisted of young (<15 year old) regeneration.


Significantly, large and very large landslides occurred in Bear Creek during all three photo periods, and in each period the few landslides in these categories represented the overwhelming majority of sediment yield originating from debris landsliding processes (Table 4). Clearly, sediment production and yield from large landslides is the most important sediment production process in the Bear Creek watershed.

 Table 4. Management condition, large landslide frequencies and large landslide contribution to mass wasting sediment yield for three time periods, Bear Creek 
Photo Date
Watershed condition
Debris Landslides 
5,000 - 10,000 yds3 
Debris Landslides
> 10,000 yds3
Large and very large landslides
(% of total yield from all landslides)
Avg. Vol. (yds3)
Avg. Vol. (yds3)


Largely unmanaged 



Pre-FPR; Heavily managed; Tractor and cable yarding; Many new midslope roads.


Post-FPR; Modest relogging; Mostly cable yarding; Some upslope and ridge roads

 1 One extremely large 1966 landslide has been omitted from the analysis. It was considered an outlier due to its anomalous size (12 acres) and volume (336,000 yds3).

Similar relationships between landslide size and contribution to total landslide sediment yield were published by Kelsey et al. (1995). In an analysis of stream side landslides in the Redwood Creek basin (northern Humboldt County) they found that the largest 10% of the landslides accounted for 60% of the total landslide volume. The largest 15% of landslides, by number, contributed nearly 80% of the landslide contribution to sediment yield. Similarly, the smallest 50% of the inventoried landslides, by number, contributed only 5% of the total landslide volume. This relationship suggests the small landslides that are not visible on the 1947 Bear Creek photos, because of heavy forest cover, probably constitute a relatively minor component of basin-wide sediment production.


In the unmanaged condition (1947), five debris landslides occur in each of the two size classes (Table 4). Most landslides in the 5,000 to 10,000 yard category in the unmanaged terrain (ie. pre-1947) on average were smaller than the large landslides in the management periods (1966, 1997). In 1966, at the height of aggressive harvesting and road building in the watershed, the frequency of large and very large landslides (triggered by the 1964 storm) increased by a factor of 2.5 over the unmanaged condition. Although the sample is small, the average size (volume) of the largest landslides also increased by 20% for this photo date, compared to the landslides measured from the 1947 photos.


As conditions in the watershed improved, through the introduction of more protective harvesting and road building practices (the Forest Practice Rules), debris landsliding also moderated. In 1997, the frequency of large and very large landslides triggered by the 1996/97 storm had dropped to levels close to that seen in the 1947 photography. However, the average volume of both the large and very large landslides was still elevated by about 10%. Watershed response, including the combined effects of landsliding, torrenting and channel aggradation, was still substantial in 1996/97. The magnitude of the impact, especially the substantial aggradation in the main channel, was likely the combined effect of increased landslide size, and the lack of long, whole logs which would otherwise have been incorporated in the debris torrent flows and which would have acted to retard torrent travel distances.


Road-related landsliding - Road-related landslide frequency peaked in the 1966 aerial photo period and has continued to diminish over time (Figure 5). The 1966 storm generated 23 identifiable landslides along roads in the basin, whereas the 1996/97 storm produced only three new road-related landslides. Four factors are likely responsible for this: 1) the 1964 storm, by all accounts, was of greater magnitude than the 1996/97 event, and thereby produced a larger number of failures, 2) a large number of roads had recently been constructed (and had not been storm tested) by 1964, and these susceptible sites failed in the 1964 event, 3) prior to the 1964 storm harvesting and yarding practices were not protective of water quality or slope stability, and 4) roads built following the 1964 storm were confined to upland areas and ridge-top hillslopes which are less susceptible to debris landsliding. With time, as unstable fills and hillslope areas fail, the number of potentially unstable areas which remain along the road system diminish in number. Likewise, much of the pre-1966 road network was upgraded during the 1990s, and this may have contributed to the reduction in landslide risk and sediment production during the 1996/97 storm.


Regardless of the analysis period, hillslope landslides (non road-related) in Bear Creek were much more common than road-related slides. During the 1964 storm period (1966 photos), large landslides (those that show up on 1:12,000 scale aerial photography) were at least four times more common in non-roaded hillslopes than along roads (Figure 5). In 1996/97, hillslope debris slides outnumbered road-related debris slides by a margin of 16:1. The relatively high frequency of large debris slides in Bear Creek can be tied to a number of contributing factors. These include 1) naturally unstable substrate (fractured bedrock and deep colluvium), 2) high rainfall rates, 3) numerous steep and long inner gorge slopes, 4) tectonic activity (frequent, large magnitude earthquakes) and 5) land use.


It is not possible to quantitatively differentiate the importance of each of these factors, but is clear that the only controllable element is land management. The 9.6 miles of new road construction which occurred prior to 1996 did not result in observable increases in landsliding, as seen on the 1997 aerial photography. These newer roads were built on relatively low gradient ridge-top areas and upper hillslope locations which involved very few new stream crossings. Pre-existing roads in Bear Creek date mostly from before 1966, and these roads weathered the latest storm with relatively little new erosional activity (only 3 new road-related sites developed). Smaller scale road-related erosion and sediment delivery (too small to be visible in the aerial photographs) was documented from the field surveys of roads in the watershed, but the volumes generated from these processes were far outweighed by larger mass movement processes occurring on unroaded hillslope areas throughout the watershed.


Harvest associations - Harvesting has occurred in Bear Creek since the 1940s, and harvested areas have been visually associated with mass wasting beginning with the storms of the 1950s. Because of the way in which the 1964 storm interacted with harvesting in the watershed, it was not possible to determine the exact relationship between harvesting and increased rates of landsliding for that event. Thus, because almost the entire watershed had experienced harvesting by the time the 1964 storm occurred, slopes were relatively vulnerable throughout the basin and landsliding was widespread. Some of the slopes had been cable yarded, while others had been tractor logged. There were no substantial or obvious differences in the rates of landsliding between the two yarding methods, but it was clear that most failures occurred on steep inner gorge slopes along stream channels.


By the time of the 1996/97 storm, there was a more divergent range of harvest ages throughout the watershed with which to evaluate the effects of a triggering storm. In 1964, over 90% of the basin’s slopes had been recently harvested when the storm occurred. In 1996/97, over thirty years after the 1964 storm, approximately 37% of the Bear Creek watershed was in a state of "recently" harvested condition (<15 year old harvested slopes)(Map 1). In response to the 1996/97 storm event, 34 landslides (78% of the total) occurred on this recently managed part of the basin (mostly from cable yarding units)(Table 5). Approximately 85% of the 1996/97 landslide sediment delivery came from this 37% of the watershed. The landslide frequency and volumetric data, although not definitive, suggests that harvesting certain slopes increases the rate of debris landsliding, and that the landslides that occur in these areas are larger and deliver more sediment to the stream system than slides found in harvest units older than about 15 years. Importantly, 75% of these slides also occurred on inner gorge hillslopes.
Table 5. Debris landslides associated with recently harvested and older harvested slopes 


Photo year

Slides on >15 year old harvested slopes (yds3)
Slides on <15 year old harvested slopes (yds3)
Slides on all harvested slopes1 (yds3)
Volume (yds3)
Volume (yds3)
Volume (yds3)

Average slide yield

 1 Includes landslides in advanced second growth areas.
Field Sampling of Streamside Landslides and Torrent Channels

Field inventories were conducted along four to five miles of torrented stream channel to determine actual rates of bank erosion, torrent scour (bed and banks) and channel bed aggradation. At the same time, 41 streamside and inner gorge landslides were identified on the ground and systematically measured.

Debris landslides - Forty-one debris landslides were inventoried in the field to determine actual ground measurements of landslide dimensions. These relationships, especially for measurements of slide depth, were then used as ground-truth data to extend to the features which had been mapped on the aerial photography from 1947, 1966, 1974, 1994, and for unvisited sites identified on the 1997 photos.


Average length for the 41 debris slides was 135 feet, with slope distances ranging from a low of 15 feet to a high of 850 feet for the largest slide. Landslide depth varied less widely than either the width or length dimensions. Average slide depth for the 41 debris slides was 3.5 feet, with a range from 2 to 18 feet, but only two slides had measured depths of 10 feet or more. Most slides exhibited average depths in the 2 to 5 foot range.


Landslide volumes and sediment delivery showed considerable range. Total landslide sediment delivery for the 41 measured features was 132,500 yds3. Average sediment delivery for the slides was 4,560 yds3, with a range from 22 yds3 to 78,000 yds3. The latter feature was the largest landslide to occur in the Bear Creek watershed since the 1964 storm. Three landslides larger than this were triggered by the 1964 event, including the single largest documented landslide in the Bear Creek basin with an estimated sediment delivery of over 300,000 yds3.


Most landslide features in the watershed occurred on steep inner gorge and stream side slopes. For this reason, they were fairly efficient at delivering sediment to stream channels. The average estimated sediment delivery from the measured debris landslides was 66%. This means that approximately 34% of the unstable materials involved in the original landsliding still reside on the hillslopes. Much of the remaining landslide material will probably be stabilized on the hillside, but some fraction will be delivered to the stream channel system in the future. Observed sediment delivery rates ranged from a low of 0% to a high of 100% for a single small slide.


Torrent tracks - The field inventory included the two largest and longest 1996/97 debris torrent tracks in the watershed (Map 4). Approximately 7,400 feet of torrent channel was measured and inventoried in the field. When a debris landslide originates along steep stream channels in the upper watershed, and it turns into a debris flow, the viscous mixture travels at a relatively high rate of speed down the channel, scouring previously stored sediment and organic debris from the channel bed and banks. The torrent channels became significant sediment sources in their own right. For this reason, the torrent channels in the main stem and in the unnamed eastern tributary, totaling over 5,900 feet, were measured for scour volumes (bed scour and sideslope soil stripping). Total scour was then computed in "yds3/ft" of torrent channel, and applied to torrented channels which were identified on aerial photos but not visited in the field.


The 1996/97 torrents in Bear Creek appear to have been composed of hyper-concentrated mixtures of water, sediment and organic debris, rather than classical debris flows of viscous mud and debris. That is, they reacted more like stream flows than mudflows. The material scoured from the channels was carried by the flows and deposited wherever channel gradients decreased, or where channel width increased rapidly. Aside from several short reaches of sediment storage in the tributaries, most of the fluid landslide material which traveled through and down torrent channels was carried to the main channel of Bear Creek and deposited. Landslides from inner gorge slopes that failed before the passing of the torrent were largely incorporated in the flows and swept downstream when the flow passed. In contrast, landslide deposits which entered the channel after a torrent had passed often still reside in the channel. In tributaries which experienced large debris landslides, but no torrents, sediment has been stored at the foot slope and in channel reaches immediately downstream from each landslide. This material will likely take a number of years to be mobilized and flushed out of the channel system.


Bank erosion - Within the main stem of Bear Creek, and within aggraded (classified as "enlarged channels") sections of the main tributaries, lateral channel migration during and following the main flood event caused local bank erosion. This process constituted another source of sediment delivered to the channel system. During our field inventory, we documented and measured sites of bank erosion along the streams, and derived unit estimates of bank erosion sediment production that could be used to apply to sites of bank erosion which had been identified during the aerial photo analysis but which were never visited in the field.


Approximately 10,970 feet of bank erosion was measured at 82 sites in the field, generating 14,100 yds3 of sediment delivery to the stream system. Average bank retreat distances were 2.4 feet for the 1996/97 flood event, with a range from 1 foot to 30 feet. A sediment production rate of 1.42 yds3/ft was applied to sites of bank erosion identified on the aerial photography from earlier years. Bank erosion in 1996/97 and in earlier years was most pronounced in low gradient channel reaches and where channel aggradation was the greatest.


Enlarged channels (deposition areas) - Prior to both the 1964 and the 1996/97 floods, the main stem of Bear Creek had a partially closed canopy dominated by young alder. As a result of the 1996/97 storm and flood event, just over 3.0 miles of the main stem of Bear Creek became a sediment "sink" for landslide and torrent debris generated in upstream and upslope areas (Map 4). This has resulted in the local burial of the original stream bed by up to nine feet of sediment. As of April, 1998, we observed a significant portion of this stored sediment had already been remobilized and transported out of the watershed.


During the field inventory, approximately 3.4 miles of aggraded channel were inventoried. Channel cross sections were measured and estimates of fill depth and subsequent scour were made at 25 individual cross section locations. In many locations, 1997-1998 stream flow and scour has exhumed the original pre-1996 channel bed and actual fill depths could be measured. Data collected at the cross section locations was assumed to be representative of the adjacent channel reaches. Thus, width, depth and volume data was then extended upstream and downstream half way to the adjacent cross sections.


Channel deposits in the main stem (and including several short segments of aggraded tributary reaches) averaged 5.5 deep at the time of maximum filling. Depth ranges at the 25 measured cross section sites ranged from 4.0 feet to 9.0 feet. The deepest fills occurred at locations where log jams had temporarily set up in the channel and ponded large accumulations of sediment and organic debris. None of these initial storm- and torrent-formed log jams endured and survived subsequent winter and spring 1997 stream flows, so the stored sediment was flushed farther downstream. Although the main stem channel system has aggraded along most of its length in the lower 3.0 miles, it is unlikely that the documented maximum thickness of sediment deposits occurred at one time. That is, as log jams built up and broke apart during the flood, channel aggradation reached local maximum thicknesses, which then diminished as jams broke apart and material was flushed downstream. Some of this break-up may have occurred during the 1996/97 flood event, and some may have occured later that winter, but the result was a relatively flat, featureless channel bed at the end of the 1997 winter period.


At the time of maximum channel aggradation, a total of 322,900 yds3 of stored sediment resided in the inventoried channel system. Additional stored sediment is also located in channel reaches that were not inventoried, and these have been estimated from the aerial photography. The average width of the aggraded channel reaches, as measured at the main stem cross section sites, was 89 feet. At the time of our field survey (April, 1997), a substantial channel had been scoured through these deposits along almost the entire length of the lower 3 miles of the main stem. This scoured channel averaged 48 feet wide and had remobilized and transported nearly half of the 1996/97 stored sediment deposits out of Bear Creek and into the Eel River. We estimate that approximately 160,100 yds3 of 1996/97 channel-deposited sediment has been eroded and flushed from the main stem by stream flows during the winter, 1997-98.


Road-related sediment sources

As the third phase of the sediment source investigation, a 100% field inventory was conducted to identify and measure sediment production and sediment delivery from road-related sediment sources throughout the Bear Creek watershed. A total of 39 miles of road was inventoried, including both old abandoned routes as well as ridge roads built within the last several years. Every site of past erosion which delivered sediment to a Class I, II or III stream channel was included in the inventory. The width, depth and length of each erosion feature was measured, a volume was calculated, and an estimated sediment delivery rate (%) was assigned. In many cases, gullies and fill slope failures along the roads had to be traced hundreds of feet downslope through dense vegetation to accurately determine their volume and sediment delivery ratio.


Table 6 summarizes road-related sediment yield from the 39 mile road system in Bear Creek, excluding the persistent contribution of fine sediment from road surfaces and ditches. A total of 186 sites were mapped in the field where past road-related sediment yield had occurred. The data has been segregated according to estimated decade in which the erosion was initiated (based on revegetation of the eroded area) and location of the erosion. Because it is difficult to identify sites which have failed more than one time in the field (reconstruction often masks these sites), such as washed out stream crossings, gullied road beds and failed road prisms, these volumes should be considered a minimum estimate.


Likewise, we have not quantified the volume of fine sediment delivered to stream channels associated with surface and rill erosion along cutbanks, road beds, ditches and skid trails throughout Bear Creek. Sediment reduction techniques and prescriptions for controlling the fine sediment discharge from roads and ditches is addressed in the erosion prevention implementation plan. However, in light of the large volumes of sediment delivered by both mass wasting and fluvial erosion outlined in Table 6, as well as inner gorge and stream side landslides described earlier (Table 5), we estimate that surface erosion processes account for no more than 1% to 2% of all sediment production and delivery in the watershed.


A total estimated sediment yield volume of 144,800 yds3 was delivered to the Bear Creek stream network from road-related erosion processes from the 1950s to the 1990s (Table 6). Sediment yield from the road system mirrored the occurrence of the major flood producing storms in the basin. Thus, approximately 44% of the documented road-related yield occurred in the 1960s (presumably as a result of the 1964 storm), and 37% was delivered during the 1990s (mostly from several large road-related failures that were triggered by the 1996 storm).
 Table 6. Road-related sediment delivery (to stream channels), by decade and erosion process, Bear Creek watershed
Stream crossing
(fillslope/ hillslope)
& channel
Total road related  
past yield 
per decade
Total road- related past yield (yds3)   

 As a process, mass movement accounted for 69% of the total sediment delivery identified along the road system (Table 6). A total of 73 past road-related landslides were identified and inventoried for a total yield of 99,900 yds3. Over 60% of the slope failures occurred on inner gorge slopes where roads approached stream crossings along deeply incised tributary channels. This site location accounted for just over 50% of the total sediment delivered from road-related mass wasting processes. The two next-most important landslide sites were slope breaks and headwall areas, both of which are locations where slope gradients change rapidly and ground water commonly emerges from the subsurface. Combined, these sites accounted for approximately 30% of the sites, by number, and 42% of the total sediment yield from road-related mass wasting.


Fluvial processes accounted for the remaining 31% of sediment yield from the road system (Table 6). As has been the case in a number of other inventoried northcoast watersheds, stream diversions at logging road stream crossings was the single largest component of road-related fluvial erosion and yield. Of the 92 existing stream crossings along the road system, 67 exhibit a diversion potential and 35 of these have diverted at least once in the past. These diversions created hillslope gullies that delivered over 31,100 yds3 of sediment to the stream system. Stream crossing failures (washouts) accounted for only 8% (11,500 yds3) of the past sediment delivery from roads in the Bear Creek watershed. Again, stream crossing wash-out volumes are probably underestimated because of the masking effects of subsequent road reconstruction.


Sediment Production and Delivery from the Bear Creek Watershed

The three elements of the sediment source investigation clearly indicate that debris landslides are the dominant sediment production and delivery mechanism in the Bear Creek watershed (Table 7). This relationship holds for the entire study period (approximately 1900 to 1997) as well as for each of the individual aerial photo periods (1947-1966, 1966-1974, 1974-1994, 1994-1997). Overall, 82% of the documented post-management sediment delivery to stream network originated from non-road-related debris landsliding. For the 1996/97 storm, debris slides accounted for approximately 75% of the sediment delivered to the channel system.


Total sediment production and delivery to the Bear Creek channel system from the 51 year post-management period (ie., 1947 to 1997) is estimated at 1,866,900 yds3 from all sources. This represents an average yield of approximately 6,177 ton/mi2/yr. It is clear that the major storms of 1964 and 1996/7 were the triggering mechanism for major sediment inputs over the last 50 years.


 Table 7. Sediment production and delivery from the Bear Creek watershed

Air photo year

Debris landslides
Torrent track scour1 (yds3)
Bank erosion3
Road-related erosion (yds3)
Total sediment production 
Channel deposition2 (yds3)
Total yield
Pct. of total   
1 A rate of 2.91 yds3/ft of torrented channel was assigned to 1966 photo measurements of torrents (rate is based on 1997 field survey measurements) 

2 Initial post-flood deposition, much of which is removed in the years succeeding the storm event 

3 Bank erosion cannot be reasonably identified along channels in a forested environment. A rate of 1.42 yds3/ft of eroded channel was assigned to areas exhibing bank erosion on the aerial photos (rate is based on 1997 field survey measurements) 

4 There were no visible debris torrents in the period from 1966 to 1994.  

5 No newly enlarged (aggraded) channels were identified on the 1974 or 1994 aerial photos.



The post-land management period can be logically broken into two time intervals (1947-1966 and 1966-1997) and sediment yield for these periods can be determined (Table 8). Likewise, sediment production from the pre-1947 period, as determined from landslide measurements, can be viewed as providing an estimate for the minimum natural background sediment delivery rate for the Bear Creek watershed. Background rates are estimated at roughly 1,000 t/mi2/yr for the 50 year period from 1897 to 1947. In contrast, during the period of relatively unregulated logging from 1947 to 1966, sediment delivery rates were approximately 19,400 t/mi2/yr, a 20 fold increase over the pre-management period.


For comparison, the average sediment delivery rate for the post-1966 period was 3,350 t/mi2/yr. This represents a four fold increase over the pre-management period, but a 6-fold decrease in comparison to the accelerated sediment production rates preceeding 1966. It is reasonable to assume that a significant portion of this reduction can be attributed to the increased protection afforded by implementation of the Forest Practice Rules (FPR) in the 1970s, as well as improved practices brought about by increased landowner awareness.


The controlling variable in determining unit sediment yield for Bear Creek, as expressed in the Table 8, is the duration of time between major storm events. The longer the time interval, the lower the rate, all else remaining equal. Because we have photo data for the three year period from 1994 to 1997, we can more closely isolate the actual effect of the individual storm on sediment production. The unit sediment delivery rate for 1994-1997, based on our field and aerial photo analysis, is 21,230 t/mi2/yr. It is clear that the infrequent, large magnitude storms (such as 1964 and 1996/7) are the driving mechanisms for watershed sediment production and change. This model of episodic geomorphic change applies equally well to both the pre-management and the post-management time period.


Stream channel impacts following the 1996/97 flood in Bear Creek were substantial, but not nearly as widespread or as continuous as they were in 1964. Aerial photo evidence shows that many more channels were torrented and much longer segments of tributary streams were buried by sediment deposits following the 1964 flood as compared to the 1996/97 event (Map 4) . Although sediment deposits behind some log jams were deeper in 1996/97 than in 1964, field evidence indicates most 1964 alluvial deposits in the main channel were emplaced at higher levels, and were slightly deeper, than those of 1996/97. Significantly, channel aggradation and riparian impacts were substantial in both events.


It is likely that the differences in stream channel response between the pre-1947 and post-1947 periods was strongly influenced by the amount of large woody debris which was available for entrainment in torrents and flood flows. In the pre-management period, vegetation in the sourceareas probably reduced landslide volumes (see earlier discussion on large landslides). In addition, large numbers of whole trees would have been incorporated in debris slides and debris torrent source areas, and entrained as flows passed through heavily wooded downstream channels. This large organic debris would have acted to strictly limit runout distances of debris torrents in the unmanaged watershed. In contrast, harvesting over the last 50 years has removed most of the large wood which would otherwise have been available for transport. As occurred in 1996/97, torrents now travel farther and channel aggradation extends over long lengths of the channel system than in pre-management times.

 Table 8. Estimated sediment delivery rates for three times periods, Bear Creek watershed

Photo period

Watershed condition
Sediment delivery 
Delivery rate1
 1897-1947 Largely unmanaged 





Pre-FPR; Heavily managed; Unrestricted tractor and cable yarding; Many new midslope roads.




Post-FPR; Modest relogging; Mostly cable yarding; Some new upslope and ridge roads

 1 Assumes a bulk density of 100 lbs/ft3
Although debris landsliding and sediment yields for the watershed are still accelerated over pre-1947 "background" levels, data collected for this investigation strongly suggest that the Forest Practice Rules, as well as recent changes in road location, road construction and harvesting techniques employed by the landowner (eg., increased emphasis on ridge-top road location and the utilization of cable yarding systems), are having a measureable and significant effect on reducing long term sediment yields. Sediment production and yield data suggests that Bear Creeks average annual unit sediment yields (tons/mi2/yr) are closer to estimated background rates than at any time since 1954, even with the occurrence of the extensive erosion which occurred in 1996/97. Additional, important improvements in forest practices, and expected reductions in sediment production, should continue to occur as a result of implementation of the measures outlined in the proposed HCP (see Appendix B - Interim Aquatic Strategy and Mass Wasting Avoidance Strategy). Vegetative retention in the streamside zone can also be expected to have a beneficial long term effect on reducing the magnitude of channel impacts following major floods.


Implementation Plan for Erosion Prevention and Control


A variety of erosion prevention and erosion control measures can be used to limit management-related sediment production and delivery in the Bear Creek watershed. These include measures to control and prevent both road-related and harvest-related sediment sources. Some measures, especially those along the road systems, are proactive projects designed to prevent future erosion. Most harvest-related mitigation measures are employed to avoid or modify land use activities on sensitive terrain.


Roads - Future road-related erosion

A total of 39 miles of road were inventoried in Bear Creek to identify potential (future) sediment sources. Four mechanisms of road-related sediment production were identified: 1) those related to stream crossings (fluvial processes of crossing failure and stream diversion) , 2) those related to mass wasting (usually road and landing fillslopes), 3) hillslope gullies below ditch relief culverts, and 4) persistent sediment sources (ditch and road runoff) (Table 9).


Stream crossings

Ninety-two stream crossings were inventoried (2.4 crossings/mile) in Bear Creek. Most newly constructed roads have been built on upper hillslope areas and along ridge-tops, so these routes have very few streams or crossings. The 92 identified crossings include 63 culverted fills, 17 unculverted fills, 3 Humboldt log crossings, 1 bridge, 1 armored fill, 1 washed out crossing (it was scoured by a 1996/7 debris torrent) and 6 "pulled" crossings on temporary roads. Crossings along the old (pre-1954) Bear Creek road that was constructed up the main channel were not included in the inventory because most are completely washed out and have little or no future yield potential.


The 17 unculverted fills are located on small streams so the erosion at most of these sites has been minimal. Flow at these crossings is generally diverted down the ditch to the nearest stream or ditch relief culvert. Several roads built within the last five years (such as the F27 Road; see Map 5) have been closed by excavating the stream crossing fills, as well as the potentially unstable road and landing fillslope materials.


Serious erosion at stream crossings is often a result of flood flows exceeding the culvert capacity, culvert plugging, and/or stream diversion. Of the 63 culverted stream crossings in Bear Creek, 42 were diagnosed as having a high or moderate potential for culvert plugging, and 37 (some of the same ones) are undersized for the 50-year design flow. Of the 92 stream crossings in the watershed, 67 exhibit a potential for stream diversion (the road slopes away from the crossing). If the culvert or other drainage structure were to plug at these sites, streamflow would be diverted down the road. In the past, 35 separate stream diversions were documented to have occurred along the road system, and these produced over 31,000 yds3 of erosion and sediment yield (Table 6).

 Table 9. Sites of future erosion and sediment delivery along 39 miles of roads in Bear Creek

Sediment source

Number of sites with future delivery
Number of sites (or road miles) recommended for treatment
Estimated future sediment delivery 
(if not treated)

Stream crossings


Mass wasting (fills)


Ditch relief culverts





Persistent surface erosion3  
5.5 miles
5.5 miles
 1 A total of 92 stream crossings were inventoried, with 5 having no expected future delivery. 

2 At stream crossings with a diversion potential, future gully erosion is difficult to predict. A minimum estimate of the stream crossing volume was used as a predicted value for this table.. This value is probably low. If past erosion rates from stream diversions and gullying are applied, future sediment delivery should be increased by an additional 24,000 yds3) 

3 5.5 miles of road ditch currently drain directly into stream crossing culverts.



Mass wasting (road-related)

Potential mass movement features related to the road system were divided into cutbank failures, landing fill failures, road fill failures and others. Of the 65 identified sites of future road-related mass wasting identified along the road system, 47 are road fills, 4 are sites with both cutbank and road fill instability, two are cutbank landslides, 6 are landing fills, and 6 are other types of sites. Most of the potential road fill failures occur where the road crosses steep inner gorge slopes on the approach to a crossing of a deeply incised stream channel.


Sediment delivery from potential road-related mass wasting sites was estimated to range from 10% to nearly 100%, depending on several site variables. In general, those with a higher delivery rate are located closer to a stream and on steeper slopes than those with less delivery. The average natural hillslope gradient for the 76 sites exceed 80%, but the distance to the receiving stream ranges from 0 feet (for those unstable fills which already toe out in the channel) to 350 feet for several potential landing fill failures. Unstable fill slopes that are located 200-350 feet from the nearest stream channel were typically assigned a delivery of only 10% to 20%, at best. The few instabilities with a projected 100% delivery were located right next to the stream channel.


Erosion prevention

A total of 179 sites of potential future road-related erosion were inventoried in the Bear Creek watershed. All ditch relief culverts were mapped on the aerial photo overlays, but only those which could deliver sediment to a stream were included in the database of problem sites. All inventoried sites, by definition, showed future potential for sediment delivery to Bear Creek or its tributaries. Some had already delivered sediment. A total of 156 sites have been recommended for treatment (Map 6). These include those sites most likely to yield sediment to stream channels in the future if erosion prevention work is not completed, and they are at locations where cost-effective work can be accomplished.


Treatment immediacy

Not all sites that display potential for sediment delivery to stream channels have the same need, or urgency (priority), for treatment. This fact led to the utilization of criteria for prioritizing all the potential work sites in Bear Creek. Recognition of site differences, and differences in erosion and sediment delivery potential, led to the development of a rating system based on "treatment immediacy." In the field, erosion features that threatened to deliver sediment to stream channels were designated as having a high, moderate or low "immediacy" of needed treatment. For all inventoried sites recommended for treatment, the Treatment immediacy is indicated on the inventory data sheet.


Site priority - Table 10 outlines the treatment immediacy (ie., priority) for 156 road-related erosion sites recommended for treatment in the watershed. Altogether, 59 sites were listed as having a high or moderate high need for treatment with a potential sediment "savings" of nearly 19,800 yds3. Eighty-one moderate and moderate/low priority sites account for 33,360 yds3 (Table 10). Finally, 16 sites are listed as having a low treatment immediacy.
Table 10. Treatment priorities for all inventoried sediment sources in the Bear Creek watershed.
"Immediacy" or Priority
Future sediment delivery (yds3)












55,190 yds3

Road priority - Another way of looking at treatment priorities is to identify high priority roads for treatment. This manner of treating sites maximizes equipment efficiency and minimizes the need to "jump around" the watershed treating only the high priority sites. Prioritizing roads is the preferred method of establishing watershed work plans for erosion prevention. Table 11 outlines the proposed work according to treatment immediacy by road. Only the most "site-rich" roads have been listed. Sites on all other roads can be treated as equipment is in the area. Regardless of the order in which sites are treated, costs can be minimized by treating all sites in a particular area rather than just attacking sites based on individual site priorities.

Table 11. Top road treatment priorities based on site density and future delivery,  

Bear Creek watershed1 



No. of sites to treat
High Immediacy
High/Moderate Immediacy
Moderate Immediacy
Volume "saved" (yds3)










27,865 yds3

1 All other roads and sites (51 sites and an additional 27,325 yds3 on a number of roads)


Reducing road-related sediment risks

A variety of treatments can be applied to prevent erosion and sediment yield to stream channels from roads and other eroding areas along roads in Bear Creek. These include upgrading and erosion-proofing existing roads and landings, total or partial road decommissioning, and specific erosion control treatments along road surfaces, ditches, eroding stream banks, gullies and other bare soil areas. Sites which are expected to erode and deliver sediment to streams in the future are the only locations where opportunity exists for meaningful erosion control and erosion prevention work. At these locations, a variety of specific treatments can be employed to control and prevent future erosion and sediment delivery to stream channels.


Types of prescribed heavy equipment erosion prevention treatments

Generic specifications for a variety of preventive watershed treatments have been developed for decommissioning and erosion-proofing (upgrading) roads and landings. Recommended treatments may range from no treatment or simple cross-road construction, to full road decommissioning (closure), including the excavation of unstable sidecast materials, road fills, and all stream crossing fills.


Road upgrading involves a variety of treatments used to make a road more resilient to large storms and flood flows. The most important of these include stream crossing upgrading (especially culvert up-sizing, to accommodate the 50-year storm flow and debris in transport, and to eliminate stream diversion potential), removal of unstable sidecast and fill materials from steep slopes, and the application of drainage techniques to improve dispersion of road surface runoff. The road drainage techniques include berm removal, road outsloping, rolling dip construction, and/or the installation of ditch relief culverts. The goal of all treatments is to make the road as "hydrologically invisible" as is possible.


General heavy equipment treatments for road decommissioning or closure are newer and less well published, but the basic techniques have been tested, described and evaluated. Decommissioning essentially involves "reverse road construction," except that full topographic obliteration of the road bed is not normally required to accomplish sediment prevention goals. In order to protect the aquatic ecosystem, the goal is to "hydrologically" close the road; that is, to minimize the adverse effect of the road on natural hillslope processes and watershed hydrology. Several roads in Bear Creek, including the A3966.66 and the last mile of the A39.36 road, have been prescribed for permanent hydrologic closure.


Typically, potential problem areas along a road are isolated to a few locations (perhaps 10% to 20% of the road network to be decommissioned) where stream crossings need to be excavated, unstable landing and road sidecast needs to be removed before it fails, or roads cross potentially unstable terrain and the entire prism needs to be removed. Most of the remaining road surface simply needs permanently improved surface drainage, using decompaction, road drains and/or partial outsloping. Table 12 lists a number of treatments and their typical applications.


Control of persistent sources of sediment yield from roads, ditches and cutbanks

Road cutbanks and road ditches are thought to deliver relatively significant volumes of fine sediment to some watersheds (e.g.., Reid, 1981) and they have been found to significantly affect watershed hydrology (Wemple, 1994). In Bear Creek, 5.5 miles of road side ditch drain directly into stream crossing culverts. This may seem like a relatively unimportant sediment source relative to landslides and gullies, but persistent fine sediment production can impede the recovery of fish-bearing streams.


Relatively simple treatments can be performed to upgrade road drainage systems to significantly reduce or largely eliminate these watershed effects. Fine sediment can usually be prevented from entering culverted stream crossings by installing a series of rolling dips or ditch relief culverts just up-road from stream crossings, and/or by outsloping roads in the immediate vicinity of stream crossings to disperse road runoff.

Table 12. Sample techniques and applications for temporary or permanent road closure
Typical use or application

Ripping or decompaction


improve infiltration; decrease runoff; assist revegetation - used in decommissioning


Construction of rolling dips and cross-road drains


drain springs; drain insloped roads; drain landings. Installation of critical dips is key to road upgrading.


Partial outsloping  

(local spoil site; fill against the cutbank)


remove minor unstable fills; disperse cutbank seeps and runoff


Complete outsloping  

(local spoil site; fill against the cutbank)


used for removing unstable fill material where nearby cutbank is dry and stable - a decommissioning tool.


Exported outsloping  

(fill pushed away and stored down-road)


used for removing unstable road fills where cut banks have springs and cannot be buried


Landing excavations  

(with local spoil storage)


used to remove unstable material around landing perimeter - used in upgrading and decommissioning.


Stream crossing excavations  

(with local spoil storage)


a road decommissioning technique; complete removal of stream crossing fills (not just culvert removal)


Truck endhauling (dump truck)


hauling excavated spoil to stable, permanent storage location where it will not discharge to a stream




A computerized database of all the sites recommended for treatment has been developed in MS Access format compatible with Pacific Lumber Company’s road database. Site numbers, site locations, descriptions, prescriptions, equipment requirements and potential sediment savings are outlined in detail in the database for each site (see Appendix A). The specific location of each site recommended for treatment is shown on Map 6. Inventory sites have been also been plotted on 1:12,000 mylar overlays to 1997 color aerial photos of the watershed. Each site is also flagged in the field.


Site specific treatments - Table 13 summarizes recommended erosion prevention and road upgrading treatments for roads in Bear Creek. The database, as well as the field inventory sheets, provide details of the treatment prescriptions for each site. Most treatments require the use of heavy equipment, including an excavator, tractor, dump truck and/or backhoe. Some hand labor is required at sites needing new culverts or culvert repairs.


Recommended treatments range from upgrading existing roads that are favorably located, to closing (decommissioning) roads which are no longer needed or are located in hillslope areas where high rates of erosion and sediment yield are occurring or can be expected. Most roads in the basin have been recommended for upgrading, but several routes are suggested for closure. These include several old abandoned routes that need erosion prevention work, as well as a dead-end portion of the 3936 road that was permanently "cut-off" but a large debris torrent in 1996/97.


Table 13. Recommended treatments along roads in the Bear Creek watershed.






Critical dip


To prevent stream diversions


Remove berm


Remove 2850 feet of road berm


Install cmp


Install a cmp at an unculverted fill


Outslope road


Outslope 490 feet of road to improve road surface drainage


Replace cmp


Upgrade an undersized cmp


Inslope road


Inslope 150 feet of road


Excavate soil


Typically fillslope excavations; excavate a total of 25,215 yds3


Clean ditch


Clean 710 feet of ditch


Clean cmp


Clean a cmp inlet to prevent plugging


Install ditch relief cmp


Install ditch culverts; use rolling dips if possible


Trash racks


Installed to prevent culvert plugging


Install rolling dips


Install rolling dips to improve road drainage


Down spouts


Installed to protect the outlet fillslope from erosion


No treatment recommended




Wet crossing


Install a rocked rolling dip or armored fill


 A number of culverted stream crossings on designated permanent and seasonal roads were observed to be undersized for the design runoff event (50-year). Plugging or overtopping during high flow events in the past has caused numerous past stream diversions and episodes of stream crossing erosion. On roads recommended for upgrading, we have calculated drainage areas and performed discharge calculations for the 50-year storm runoff peak at stream crossings which appeared to have undersized culverts. The mathematical algorithm, based on the Rational formula for calculating peak flows from small watersheds, has been modified and incorporated into a spread sheet format. This numerical model was used to estimate peak discharges and appropriate culvert sizes. The recommended culvert sizes included in the treatment database shown on the data sheet for each stream crossing site is based on this model, modified as necessary to fit field conditions and observations. Field evidence and discharge calculations from at least 37 sites indicate that a culvert upgrade is warranted.


A common treatment prescribed for upgrading stream crossing sites is for the installation or construction of rolling dips to eliminate the potential for stream diversion. These are called "critical dips" in the treatment database (Table 13). Critical dip installation is a simple, inexpensive, preventive erosion control technique which provides protection to the road system (and downstream areas) during the most extreme winter floods, when culverts may become plugged or their capacity exceeded. In total, 54 critical dips have been recommended for installation at stream crossings on permanent and seasonal roads. The treatment takes a tractor about one hour (plus re-rocking, where necessary) to perform.


Control of ditch and road surface fine sediment yield - There are several ways to effectively reduce the contribution of fine sediment from road ditches. First, road-side ditches near stream channels can be eliminated by outsloping the road bed and dispersing runoff rather than collecting and concentrating it in ditches. In this strategy, not all the road ditches in the watershed would have to be eliminated or shortened, just the ones that drain road and cutbank runoff directly into stream channels. Alternatively, a series of rolling dips or ditch relief culverts could be installed just up-road from each stream crossing so that ditch runoff (and eroded sediment) is diverted and dispersed on the hillslope below the road rather than being discharged through the ditch and into the inlet of the stream crossing culvert. Because installation of ditch relief culverts also increases long term maintenance requirements, rolling dips are the preferred method for road surface drainage.


Along upgraded roads, approximately 5.5 miles of ditch can be disconnected from the stream system by the installation of rolling dips and ditch relief culverts. Culverts or dips (which intercept all ditch flow) would need to be installed approximately 50-75 feet up-road of each crossing to achieve maximum effectiveness at reducing sediment contributions to the stream channel. Specific locations for the placement of rolling dips and ditch relief culverts should be mapped in the field just prior to equipment work so that sediment and water from ditches can be effectively dispersed onto hillslopes below the road with no threat that it will enter the stream channel. Stream crossings requiring this treatment have been identified in the treatment database.


Equipment needs and costs

Treatments in the watershed will require approximately 736 hours of excavator time and 723 hours of tractor time to complete all prescribed upgrading, road closure, erosion control and erosion prevention work at the 156 recommended treatment sites (Table 14 and Map 6). Excavator and tractor work is not needed at all the sites that have been recommended for treatment and, likewise, not all the sites will require both a tractor and an excavator. Approximately 204 hours of dump truck time has been listed for work in the basin for endhauling excavated spoil from stream crossings and unstable road and landing fill where local disposal sites are not available.

Table 14. Heavy equipment requirements for road-related erosion prevention, 

Bear Creek watershed


Treatment Immediacy

Future Yield (yds3)
Dump Trucks
Backhoe (hrs)
Labor (hrs)

High, High/Moderate










Estimated costs for erosion prevention treatments - Prescribed treatments are divided into two components: a) site specific erosion prevention work identified during the watershed inventories, and b) control of persistent sources of road surface, ditch and cutbank erosion and associated sediment delivery to streams. The site-specific work is further divided into road upgrading activities and road closure activities. Total costs for the project are estimated at approximately $201,625, for an average cost-effectiveness value of $3.65 per cubic yard of sediment prevented from entering Bear Creek and its tributaries.


Overall site specific erosion prevention work: Equipment needs for site specific erosion prevention work are expressed in the database, and summarized on Table 14, as direct excavation times, in hours, to treat all sites in the basin which have a high, moderate, or low treatment immediacy. These hourly estimates include only the time needed to treat each of the sites, and do not include travel time between work sites, the time needed to reconstruct or clear roads which have been abandoned, or the time needed for work conferences at each site. These additional times are accumulated as "logistics" and must be added to the work times to determine total equipment costs as shown in Table 15. Costs in Table 15 assume that the work in this watershed is accomplished during a single summer work period employing one or two equipment teams. This minimizes moving and transport costs for equipment and personnel.
Table 15. Estimated logistic requirements and costs for road-related erosion control 
and erosion prevention work along 39 miles of roads, Bear Creek watershed



Cost Category1

Cost Rate2
Estimated Project Times
Total Estim.
Treatment3 (hours)

Move-in; move-out6  

(Low Boy expenses)



Heavy Equipment


D-7 size tractor






Dump Truck




Layout, Coordination, Supervision, and Reporting7 

Total Estimated Costs


1 Costs for tools, new culvert materials, for mulching and related materials (grass seed, fertilizer and straw), and for plant materials have not been included in this table. Culvert costs could be considerable. Costs for administration and contracting are variable and have not been included. Costs and dump truck time (if needed) for re-rocking the road surface at sites where upgraded roads are outsloped, where rolling dips are constructed and where stream crossing culverts are replaced or upgraded have not been estimated.

2 Costs listed for heavy equipment include operator and fuel. Costs listed are estimates for favorable local private sector equipment rental and labor rates.

3 Treatment times include all equipment hours expended on excavations and work directly associated with erosion prevention and erosion control at all the sites.

4 Logistic times for heavy equipment (30%) include all equipment hours expended for opening access to sites on maintained and abandoned roads, travel time for equipment to move from site-to-site, and conference times with equipment operators at each site to convey treatment prescriptions and strategies. Logistic times for laborers (30%) includes estimated daily travel time to project area.

5 Total estimated project costs listed are averages based on private sector equipment rental and labor rates.

6 Lowboy hauling for tractor and excavator, four hours round trip each piece. Costs assume 2 hauls for two pieces of equipment to each side of the Bear Creek watershed (one to move in and one to move out).

7 Supervision time includes detailed layout (flagging, etc) prior to equipment arrival, training of equipment operators, supervision during equipment operations, supervision of labor work and post-project documentation and reporting).

Bear Creek hillslopes - Landslide prevention and avoidance

The second strategy for reducing management-related sediment production in the basin involves efforts needed to minimize the risk of accelerated landsliding. The sediment source inventory has shown that debris sliding is by far the most important (volumetrically) sediment source in the watershed. In addition, landsliding on inner gorge slopes is the most common and obvious natural association in the watershed. Past land management has locally contributed to increased rates of landsliding, especially on steep inner gorge slopes.


To mitigate or reduce the influence of management on slope stability in Bear Creek, measures can be undertaken to better characterize and predict potential slope stability problems before land management is undertaken. This will involve the field inspection and analysis, by a trained geologist, for all harvesting and road construction plans in sensitive portions of the landscape in Bear Creek, specifically on inner gorge slopes and headwall swales. Inspection, analysis, avoidance, vegetation retention and other measures outlined to accomplish this are included in PALCO’s "Interim Aquatic Strategy for Timber Harvest and Roads", including the attached document "Mass Wasting Avoidance Strategy for the Interim Period" (see Appendix B). The Company has committed to employing these measures on all future harvesting plans submitted for Bear Creek.


Landsliding is a natural, episodic and important sediment production process in Bear Creek. As this study revealed, portions of the Bear Creek watershed are especially susceptible to high rates of landsliding, with or without the influence of timber harvesting and road building. Even in the absence of land management, episodic landsliding will continue to be the most important sediment production mechanism in the basin. A number of hillslope areas currently show signs of fairly large, deep seated slope movement. Other inner gorge areas exhibit actively developing debris landslides. Some of these may continue to develop and fail, while others may stabilize over time. Regardless, the increased protection provided by the proposed inspection and avoidance measures should reduce the effects of future land management. Although past land management will continue to have a lingering effect on the frequency of mass wasting, this effect should diminish as inner gorge protection measures and vegetation retention becomes established over the long term.



Bear Creek Monitoring Plan


A variety of techniques are available to document the physical and biological recovery of Bear Creek. The most useful strategy will be one that documents changes in aquatic habitat condition and utilization, as well as the physical recovery of stream channel morphology in the main stem. Pacific Lumber Company already has monitoring stations in Bear Creek and the Company plans to continue to collect data at these sites. Collected information includes water temperature, particle size data, fine sediments, and a number of macro invertebrate "richness" indices at several stations. PALCO has provided some detail on the current sampling and monitoring strategy for Bear Creek in their cover letter to this report.


PALCO will augment the existing monitoring with several measures to document physical channel recovery. Some of these measures have already been initiated, as a part of the sediment source investigation. Once stations have been permanently established, sites will be remeasured only as conditions warrant. This may be annually for the first several years and then every two to five years, or as storms and floods dictate.


Physical channel monitoring will include two main elements:


1. Photo points - Permanent photo point stations will be established along the main channel to document the scour and removal of sediment deposited during the 1996/97 storm and flood. The channel has already been stationed at 300 foot intervals and temporary photo point stations have already been established and flagged in the field. Photos were taken at most of these sites, covering about 2.5 miles of channel, in December, 1997. Additional photos of the channel were taken at selected upstream locations in March, 1998. A select number of these will be permanently established for repeated photography to document channel recovery.


2. Cross section measurements - A number of permanently monumented channel cross sections will be established along the length of the main stem of Bear Creek to document rates of channel incision through the 1996/97 flood deposits. Measurements at 22 temporary cross section sites have already been taken to determine depths of aggradation and subsequent scour (as of 3/98). A selected number of these measurement sites will be permanently established for repeated cross section surveys in the future. These sites will be narrowed-down and selected this summer when stream levels have dropped.


The combined physical and biological monitoring program will provide a evolving picture of the physcial and biological recovery of the main stem of Bear Creek following the 1996/97 storm and flood. The precise location of monitoring stations will be established this summer, and PALCO will permanently identify the type of monitoring and the measurement techniques for each site at that time. The monitoring stations will be incorporated in the Company’s GIS mapping program.



Summary and Conclusion


A variety of processes are responsible for sediment production and yield in the Bear Creek watershed. These include mass wasting (rapid debris landslides, slow moving deep seated landslides), debris torrenting (including torrrent track scouring), and stream bank erosion. Road-related erosion consisted of mass wasting processes (cut bank, road fill and landing fill), fluvial processes (steam crossings washouts, stream diversion gullies, and gullies from ditch relief culverts) and surface erosion (road surface, ditch and cutbank erosion).


Mass wasting is clearly the most important sediment production process in the Bear Creek watershed. Both numerically and volumetrically, sediment production in both the pre-management (pre-1947) and post-management (post-1947) periods has been dominated by debris landsliding from steep (>65%) inner gorge and streamside slopes. Nearly 70 % of the 260 non-road-related landslides in the watershed occured on inner gorge slopes (Table 2). Steep stream side slopes (<65% steepness) accounted for 13% of all landslides. Over 82% of sediment production from all sources, over all photo years (combined), originated from debris landslides (Table 7). The largest debris slides (those >5,000 yds3) accounted for less than 20% of the number of landslides inventoried, but accounted for over 75% of landslide sediment yield.


Both road construction and harvesting have been linked to increased sediment production and yield in Bear Creek. Harvesting began in the watershed in the 1940s, and has been visually associated with increased rates of mass wasting since the 1950s. At the time of the 1964 storm, most of the watershed had experienced intense harvesting, ground disturbance, and road building. Partly as a consequence, sediment production, as measured from analysis of the 1966 aerial photos, reached a maximum during this period. Landslide frequency and volumetric data from the most recent period (1966-1997) suggests harvesting of certain slopes increases both the frequency and size of landslides (Table 5). Regardless of the management associations, it is clear that most landslides from both the 1947-1966 and 1966-1997 periods occurred on steep inner gorge and streamside slopes.


Landslides in Bear Creek are also more commonly associated with unroaded hillslopes than with roads. In 1966, debris landslides visible on the aerial phtoographs were four times more common on non-roaded hillslopes than along roads. In the 1996/97 storm, hillslope debris slides outnumbered road-related slides by 16:1. Much of this difference stems from reduced rates of road-related landsliding visible on the 1997 aerial photography, as compared to the 1966 photo period.


Watershed sediment production for the period of record has been apportioned among four general sediment production mechanisms: debris landslides (82%), torrent track scour (7%), bank erosion (3%) and road-related erosion (8%). Sediment production from surface erosion (not measured for this study) is estimated to account for less than 2% of watershed sediment production. Presently, approximately 52% of the sediment generated during the 1996/97 storm events is stored in low gradient main stem and tributary stream channels of Bear Creek. Approximately 48% of the 1996/97 stored sediment has been remobilized and flushed to the Eel River.


Sediment delivery rates for the three major time periods of this study, pre-1947 (pre-land use), 1947-1966 (pre-FPR, intense and relatively unrestricted land use) and 1966-1997 (post-FPR modified land use), suggest that watershed sediment production and delivery has been influenced by both the intensity and nature of harvesting, road location and construction practices, as well as the frequency, magnitude and intensity of major storms. Extremely high rates of sediment production measured from the 1966 photography (triggered by the 1964 storm) has given way to much lower rates for the 32 year period of record since 1966.


An implementation plan has been prepared for erosion control and erosion prevention along 39 miles of road in the Bear Creek watershed. Roads have been prioritized for upgrading and closure, depending on their location and current status. A data base of treatment prescriptions, equipment and labor requirements and estimated costs has been prepared for each of the 156 sites recommended for treatment. Implementation of the work plan will prevent the future delivery of over 55,000 yd3 of sedment to Bear Creek and its tributaries.


A landslide prevention and avoidance strategy has also been proposed. It is based on minimizing the effects of forest management activities on potentially unstable inner gorge slopes where debris landsliding is a common and important sediment production process. It is expected that the long term downward trend in accelerated sediment production and yield in the Bear Creek watershed that has occurred since the mid-1960s will continue into the next century with the application of newly developed measures outlined in the "Interim Aquatic Strategy for Timber Harvest and Roads" and the associated procedures described in "Mass Wasting Avoidance Strategy for the Interim Period" (see Appendix B).


Finally, a monitoring plan has been proposed to document and track the physical and biological recovery of Bear Creek. The plan calls for the continued sampling of a variety of aquatic parameters at existing monitoring stations, as well as the establishment of a number of long term physcial monitoring stations. Physical monitoring will involve repeated ground-level stereo photographic documentation of channel conditions from monumented photo-point stations (22 temporary stations have already been established) and cross section surveys to determine the location, nature and rate of future channel changes and sediment flushing (or aggradation).


References Cited


Clarke, S. H. Jr., 1992, Geology of the Eel River basin and adjacent region. AAPG Bul. V. 76, No. 2 Feb 1992 P. 199-224.

Coghlan, T.M., 1984. A climatologically-based analysis of the storm and flood history of Redwood Creek. Redwood National Park Technical Report 10, Arcata Calif., 47 p.

Dengler, L. and R. McPherson. 1993. The August 1991 Honeydew Earthquake, North Coast California: a case for revising the Modified Mercalli Scale in sparsley populated areas. Bull of the Seismological Society of America 83 (4):1081-1094.

Dumitru, Trevor A. 1991 . Major Quaternary uplift along the northernmost San Andreas fault, King Range, northwestern California. Geology V. 19 p. 526-529.


Grant, G., 1988, The RAPID technique: a new method for evaluating downstream effects of forest practices on riparian zones. USDA Forest Service, Pacific Northwest Forest Research Station, Gen. Tech. Report PNW-GTR-220, Portland, OR. 36 pages.


Hagans, D.K. and W.E. Weaver, 1987, Magnitude, cause and basin response to fluvial erosion, Redwood Creek basin, northern California. In: Erosion and sedimentation in the Pacific Rim (proceedings of the Corvallis symposium, August, 1987), Eds. R.L. Beschta and others, IAHS Publication No. 165, pp. 419-428.


Harden, Deborah R., 1995. A comparison of flood-producing storms and their impacts in northwestern California. In: Geomorphic processes and aquatic habitat in the Redwood Creek basin, northwestern California. U.S. Geological Survey Professional Paper 1454, pp. D1-D9.


Helley, E.J. and LeMarche, V.C., 1973. Historic flood information for northern California streams from geological and botanical evidence. U.S. Geological Survey Professional Paper 485-E, pp. E1-E16.


Hunt, Roy E. (1994) "Geotechnical Engineering Investigation Manual". McGraw-Hill Inc 674p.


Keefer, D.K., 1984. Landslides caused by earthquakes: GSA Bull V.95, no.4, p.406-421.


Kelsey, H.M., 1980. A sediment budget and an analysis of geomorphic process in the Van Duzen River Basin, northern coastal California, 1941-1975. Geological Society of America Bulletin, v. 91, pp. 1119-1216.


Kelsey, H.M., Mike Coghlan, John Pitlick, and David Best, 1995. Geomorphic analysis of streamside landslides in Redwood Creek basin, northwestern California. In: Geomorphic processes and aquatic habitat in the Redwood Creek basin, northwestern California. U.S. Geological Survey Professional Paper 1454, pp. J1-J12.


Lisle, T.E., 1981, The recovery of stream channels in north coastal California from recent large floods. In: Habitat Disturbance and Recovery Proceedings (K. Hashagen (ed)), Cal Trout, San Francisco, CA, pages 31-41.


McLaughlin, R.J. and Clarke, S.H. Jr., 1992. Neotectonic Framework of the Southern Cascadia Subduction Zone - Mendocino Triple Junction Region. 1992 Pacific Cell FOP, N. Coastal CA p 64-72.


Ogle, B.A., 1953, Geology of the Eel River Valley area, Humboldt County, California:CDMG Bull 164, 128p.


Pacific Watershed Associates, 1994, Handbook for forest and ranch roads, prepared for the Mendocino County Resource Conservation District in cooperation with the California Department of Forestry and the U.S. Soil Conservation Service. Mendocino Resource Conservation District, Ukiah, California. 163 pages.


Reid, L.M., 1981, Sediment production from gravel-surfaced forest roads, Clearwater Basin, Washington. University of Washington, College of Fisheries, Fisheries Research Institute, Publication No. FRI-UW-8108, Seattle, WA. 247 p.


Wemple, B.C., 1994, Hydrologic integration of forest roads with stream networks in two basins, western Cascades, Oregon. Oregon State University. M.S. Thesis.

Zinke, P.J., 1981. Floods, sedimentation, and alluvial soil formation as dynamic processes maintaining superlative Redwood groves. In: Coats, R.N., ed., Watershed rehabilitation in Redwood National Park and other Pacific coastal areas. Center for Natural Resources Studies, John Muir Institute Incorporated, pp.26-49.

Sediment Source Investigation
and Sediment Reduction Plan for the
Bear Creek Watershed, Humboldt County, California


 A. PWA Road Inventory Data Form A

B. PALCO HCP - Draft Iterim Aquatic Strategy for Timber Harvest and Roads
and Mass Wasting Avoidance Strategy for the Interim Period B


Appendix A


  PWA Road Inventory Data Form A1-A2
ASAP____ P W A R O A D I N V E N T O R Y D A T A F O R M (3/98 version) Check_____



Site No: ________










Treat (Y,N):


Photo: ______




Road #: 


Mileage: ___________








Date: ________


Year built:______


Sketch (Y):




















Stream xing


Landslide (fill, cut, hill)


Roadbed (bed, ditch, cut)












Location of problem (U, M, L, S)


Road related? (Y)


Harvest history: (1=<15 yrs old; 2=>15 yrs old) 

TC1, TC2, CC1, CC2, PT1, PT2, ASG, No


Geomorphic association: Streamside, I.G.,  

Stream Channel, Swale, Headwall, B.I.S.


Road fill


Landing fill








Already failed


Pot. failure




Slope shape: (convergent, divergent, planar, hummocky)


Slope (%) ______


Distance to stream (ft) __________














Armored fill






Pulled xing: (Y)


% pulled ______


Left ditch length (ft) ___________


Right ditch length (ft) ___________




cmp dia (in) ______


inlet (O, C, P, R)


outlet (O, C, P, R)


bottom (O, C,P, R)








Headwall (in) ____


CMP slope (%) _____


Stream class (1, 2, 3)


Rustline (in)






% washed out ____


D.P.? (Y)


Currently dvted? (Y)


Past dvted? (Y)


Rd grade (%) ________






Plug pot: (H, M, L)


Ch grade (%) _____


Ch width (ft) _____


Ch depth (ft) ____






Sed trans (H, M, L)


Drainage area (mi2) _________






E.P. (H, M, L)


Potential for extreme erosion? (Y, N)


Volume of extreme erosion (yds3): 100-500, 500-1000, 1K-2K, >2K


Past erosion...


Rd&ditch vol (yds3) 



Gully fillslope/hillslope 



Fill failure volume 

(yds3) _________


Cutbank erosion 


Hillslope slide vol. (yds3)



Stream bank
erosion (yds3)



xing failure
vol (yds3)






Total past erosion (yds) __________


Past delivery 

(%) __________


Total past yield  

(yds) _________


Age of past erosion (decade)_______


Future erosion...


Total future erosion 

(yds) __________


Future delivery 

(%) __________


Total future yield  

(yds) _________


Future width  

(ft) _________


Future depth 

(ft) ________


Future length 

(ft) _______






Immed (H,M,L)


Complex (H,M,L)


Mulch (ft2)






Excavate soil


Critical dip


Wet crossing (ford or armored fill) (circle)


sill hgt (ft) ___


sill width (ft) _______




Trash Rack




D.S. length (ft) ________


Repair CMP


Clean CMP






Install culvert


Replace culvert


CMP diameter (in) _____


CMP length (ft) _______






Reconstruct fill


Armor fill face (up, dn)


Armor area (ft2) _______


Clean or cut ditch


Ditch length (ft) _________






Outslope road (Y)


OS and Retain ditch (Y)


O.S. (ft) ____________


Inslope road


I.S. (ft) _____


Rolling dip


R.D. (#) __




Remove berm


Remove berm (ft) _____


Remove ditch 


Remove ditch (ft) __________


Rock road - ft2 __________




Install DR-CMP


DR-CMP (#) ________


Check CMP size? (Y)


Other tmt? (Y)


No tmt. (Y)










































Total excavated (yds3) _______


Vol put back in (yds3) _______


Volume removed (yds3) ________




Vol stockpiled (yds3) _______


Vol endhauled (yds3) _______


Dist endhauled (ft) ______


Excav prod rate (yds3/hr) _______




Excavator (hrs) ________


Dozer (hrs) ________


Dump truck (hrs) ______


Grader (hrs) ________




Loader (hrs) _________


Backhoe (hrs) ________


Labor (hrs) _______


Other (hrs) ________













Stream Profile Through Crossing Xing type 1, 2, 3, 4
(begin at top of profile) (Circle)
Angle (deg) (dwn=- )
Distance (feet)







in natural channel


























































































































































Computer volumes- 1. Computer erosion volume (1:1): _____________

3. Humboldt excavation volume (1:1): _____________


Site Sketch - Site No._______


Cross Section(s) (begin on left bank)
Angle (deg)
Distance (feet)









Base of cutbank










































































































































































































2. Culvert excavation vol (add/repl - 1:1): _______________

4. Decommission volume (2:1): _______________






Appendix B








PALCO HCP - Draft Iterim Aquatic Strategy for

Timber Harvest and Roads and Mass Wasting

Avoidance Strategy for the Interim Period B