AREAL EXTENT: NORTHERN CALIFORNIA
INFO ITEMS:
INPUT OUTPUT DATA DECI DESCRIPTION
ITEM NAME WIDTH WIDTH TYPE MALS
---------- ----- ------ ---- ---- --------------------------------------
BASREG 7 7 I - Basin (1st 5 numbers) and region (6th number)
SENS 4 4 I - Problem watersheds (SENS = 1) identified through survey
to watershed professionals
BASINREG 10 10 I - Contains no data
IDNUM 8 8 N 2 Unique watershed identifier. See item descriptions below
for hierarchical codes.
BASIN 5 5 I - Regional Water Quality Control Board Basin number
REGION 1 1 I - Regional Water Quality Control Board Region number
SUBREGION 1 1 I - Super Planning Watershed number
SENSNO 4 4 I - Sensitive watershed number
GEOLLSMN 4 4 F 1 Geologic landslide factor
GEOLDFMN 4 4 F 1 Geologic debris flow factor
GEOLSEMN 4 4 F 1 Geologic surface erosion factor
SLOPLSMN 4 4 F 1 Slope landslide factor
SLOPDFSEMN 4 4 F 1 Slope debris flow and surface erosion factor
PPCTLSMN 4 4 F 1 Precipitation landslide factor
PPCTDFMN 4 4 F 1 Precipitation debris flow factor
PPCTSEMN 4 4 F 1 Precipitation surface erosion factor
LSRATING 5 5 I - Landslide rating
DFRATING 5 5 I - Debris flow rating
SERATING 5 5 I - Surface erosion rating
LSCMPRATS 4 4 I -
DFCMPRATS 4 4 I -
SECMPRATS 4 4 I -
TOTRATING 4 4 I - Total erosion rating
TOTCMPRATS 4 4 I -
ITEM DESCRIPTIONS:
IDNUM: Is a heirarchical code containing codes for
the Regional Water Quality Control
Board region, hydrologic unit, hydrologic area, hydrologic subarea,
super planning
watershed, and planning watershed. Given the code 12345.67, the breakdown
is as
follows:
1 - RWQCB region
23 - Hydrologic unit
4 - Hydrologic area
5 - Hydrologic subarea
.
6 - Super planning watershed
7 - Planning watershed
BASIN: The first 5 numbers from IDNUM (12345 from example above).
REGION: The super planning watershed number (.6 from example above).
SUBREGION: The planning watershed number (7 from example above).
BASREG: An aggregation of BASIN and REGION (123456 from example above).
SENS: Contains a value of "1" if the watershed was identified
as a sensitive
watershed through a survey of watershed specialists. Not all watersheds
identified through the survey are part of this data set (some did
not
meet the basic requirements, i.e. >25% private ownership).
SENSNO: Unique identifier of SENS watersheds. Some numbers are missing
if the
watershed is not part of this data set.
GEOLLSMN: Area weighted average rating for effect of geology on Landsliding.
Values are real numbers between 0 and 3.
GEOLDFMN: Area weighted average rating for effect of geology on Debris
Flows.
Values are real numbers between 0 and 3.
GEOLSEMN: Area weighted average rating for effect of geology on Surface
Erosion.
Values are real numbers between 0 and 3.
SLOPLSMN: Area weighted average rating for effect of slope on Landsliding.
Values are real numbers between 0 and 3.
SLOPDFSEMN: Area weighted average rating for effect of slope on Debris
Flows
and Surface Erosion.
Values are real numbers between 0 and 3.
PPCTLSMN: Area weighted average rating for effect of precipitation
on Landsliding.
Values are real numbers between 0 and 3.
PPCTDFMN: Area weighted average rating for effect of precipitation
on Debris Flows.
Values are real numbers between 0 and 3.
PPCTSEMN: Area weighted average rating for effect of precipitation
on Surface Erosion.
Values are real numbers between 0 and 3.
LSRATING: ???
DFRATING: ???
SERATING: ???
LSCMPRATS: Landslide rating. A 3 column integer, where each column
is equal
to the rounded values of GEOLLSMN, SLOPLSMN, and PPCTLSMN respectively.
DFCMPRATS: Debris flow rating, same derivation as LSCMPRATS.
SECMPRATS: Surface erosion rating, same derivation as LSCMPRATS and
DFCMPRATS.
TOTRATING: Total erosion rating, derived by rounding down (truncating)
the
mean of LSRATING, DFRATING, and SERATING.
TOTCMPRATS: ???
Probably the best field to use for mapping or analysis is TOTRATING.
This is
essentially an aggregation of all the factors that go into defining
a highly
erodible watershed."
Erosion Potential in Private Forested Watersheds
in Northern California
A GIS Model
Prepared for the California Department of Forestry and Fire Protection Under
Interagency Agreement 8 CA17097#2
by
Mary Anne McKittrick
Department of Conservation
Division of Mines and Geology
with assistance from the U.S. Geological Survey
Under Standard Agreement #1093-214
March 1994
DISCLAIMER
The California Department of Conservation makes no warranties as to the suitability of this product for any particular purpose. If you have obtained this data from a source other than the California Department of Conservation, be aware that electronic data can be altered subsequently to original distribution. Contact the Department of Conservation, Division of Mines and Geology (801 K Street, MS 14-33), Sacramento, California 95814) to obtain an original copy of this product.
TABLE OF CONTENTS
INTRODUCTION..............................................1
MODEL INPUT...............................................3
Selection of Geomorphic Variables.....................3
Study Limitations................................5
Data Preparation......................................6
Watershed Boundaries.............................6
Slope............................................7
Geology..........................................8
Precipitation....................................9
Analysis.........................................9
MODEL OUTPUT.............................................11
Landslides......................................11
Debris Slides...................................13
Surface Erosion.................................14
Spatial Relations....................................16
Erosion Potential....................................17
Landslide Potential.............................17
Debris Slide Potential..........................18
Surface Erosion Potential.......................18
Total Erosion Potential.........................19
DISCUSSION...............................................19
ADDITIONAL WORK..........................................23
Land Use.............................................24
Soil Properties......................................24
Transient Snow Zone..................................25
Inner Gorge..........................................25
Pleistocene Glaciation...............................26
Watershed-Scaled Analysis............................26
Conclusion...........................................27
REFERENCES CITED.........................................28
FIGURES
Figure 1 Hillslope Stripped of Colluvium............35
Figure 2 Rainfall Intensity-Duration Thresholds.....35
Figure 3 Slope Calculation..........................36
Figure 4 Model Slope............................... 36
Figure 5 Lithotectonic belts of Northern and
Central California.........................37
Figure 6 Landslide Rating...........................38
Figure 7 Debris Slide Rating........................39
Figure 8 Surface Erosion Rating.....................40
Figure 9 Total Erosion Rating.......................41
Figure 10 Observed Problem Watersheds................42
Figure Captions......................................43
TABLES
Table 1 List of Contributors to the Geologic Unit
Ratings......................................44
Table 2 Geologic Rating Table........................45
Table 3 Data Input...................................58
Table 4 Relative Ratings of Erosion Potential........59
Table 5 Field/Observations of Erodible Watersheds ...68
INTRODUCTION
Timber harvesting is a major land management practice that can affect hillslope
stability. Erosion resulting from timber harvesting activities can increase
sediment yields 4 to 78 times that of natural forest conditions (Megahan and
others, 1978; Bishop and Stevens, 1964; Morrison, 1975). Erosion and sediment
production can have long-term impacts on timber site productivity, fish habitat,
reservoir storage capacity, and domestic water supplies (Reid and Dunne, 1984;
Brown, 1975). With these concerns in mind, the California Department of Forestry
and Fire Protection (CDF) contracted with the Department of Conservation's Division
of Mines and Geology (DMG) to develop a semi-quantitative method to delineate
those forested watersheds which are most susceptible to erosion when hillslopes
are disturbed by timber harvest operations.
Intrinsic erosion potential was modeled on private and state-owned commercial
timberlands regulated by CDF. The goal was to select the most significant factors
controlling erosion that could be delineated consistently over large areas.
The effects of forest land management on erosion vary spatially because of differences
in climate, geologic materials, vegetative cover, and topography. Therefore
all areas are not equally sensitive to a particular forest practice.
A Geographic Information System (GIS) model was developed to prioritize the
relative susceptibility of forested watersheds to erosion. A combination of
the most significant geomorphic factors which contribute to the driving and
resisting forces controlling landscape denudation -- material strength, slope,
and precipitation -- were used to delineate areas most prone to increases in
sediment yield.
Most erosion and sedimentation studies heretofore have been limited to the evaluation
of local conditions at individual harvest sites or within small hydrologic basins
(Sommarstrom and others, 1990; California Department of Water Resources, 1979;
Kelsey, 1977). In contrast this study is a regional evaluation of erosion potential
based on semi-quantitative analyses.
The purpose of the investigation was to develop a quantitative method for ranking
erosion risks based on available geomorphic data and check this system against
the experience of field personnel. The study consisted of two parts: 1) development
of a GIS based model derived from the physical properties within 530 designated
watersheds on private and state-owned commercial timberlands in California.
The details of this system are discussed below, and 2) preparation and distribution
of a questionnaire requesting information on suspected highly erodible watersheds.
The questionnaire was completed by CDF Forest Practice Inspectors; engineering
geologists with the DMG Timber Harvesting Plan Review Project; earth science
professionals involved with reviewing Timber Harvesting Plans with the Regional
Water Quality Control Board (RWQCB); and wildlife biologists with the California
Department of Fish and Game (CDFG) involved with evaluating Timber Harvesting
Plans. The watersheds identified
from the questionnaire and review of existing data were cross checked with the
information developed in the GIS model.
MODEL INPUT
Selection of Geomorphic Variables
The principal factors that have been shown to contribute to erosion in forested
terrane are slope steepness, horizontal concavity, high groundwater, cohesionless
soils, and weak bedrock (Durgin and others, 1989; Lewis and Rice, 1989; Peters
and Litwin, 1983; Campbell, 1975). A total of 23 erosion-contributing variables
were explored for potential use in the model during the design of this study.
These include soil consolidation, soil permeability, soil depth, soil plasticity,
colluvium depth, presence of surficial deposits, geology, vegetation, slope,
slope length, slope aspect, land use, rainfall intensity, rainfall duration,
seasonality of rainfall, temperature, landscape maturity dissection density,
dissection depth, horizontal curvature, stream inner gorges, ground water table
depth, and areas of potential rain-on-snow.
Some factors, such as changes in vegetation, areas and dates of land-use impact,
rainfall duration, and rain-on-snow events are difficult to depict in map form
due to their temporal variability. Many other of the above factors are not easily
depicted in map form at regional scales, such as inner gorge development, stream
dissection density and depth, ground water table levels, and horizontal curvature.
Because all the areas
in this study were forested, variability in vegetation type was considered to
be minimal in terms of providing significant differences in ground coverage.
Therefore, the project focuses on regionally consistent and available information
over the entire study area. This includes geology (with the susceptibility of
each geologic unit to landslide, debris slide, or surface erosion processes),
slope steepness, and rainfall intensity (including mean-annual, 12-hour, and
2-hour precipitation).
Approximately 530 watersheds throughout northern California, each about 20,000
hectares, were evaluated in this study. Each watershed contains at least 25%
private or state-owned commercial timberland. In each watershed, the physical
attributes of slope, precipitation, and lithologic susceptibility to failure,
were stratified into low, moderate, and high categories based on the relative
contribution of that factor to erosion potential. These data layers were entered
as separate digital coverages in an ARC/INF0-based GIS. Rated polygons were
area-weighted and additionally combined for each hydrologic basin. Although
the relationships between these geomorphic factors no doubt are complex and
non-linear, a simple linear additive relationship was used to combine data sets.
The highly generalized nature of the input data, the averaging of data over
watershed areas, and the lack of established empirical-relationships between
the data cause the use of complex algorithms to give an improper impression
of precision. The use of the additive data combination produces an array of
ranked watersheds depicting those basins
which are theoretically most susceptible to accelerated hillslope degradation.
Separate erosion-potential maps were generated for three general types of hillslope
erosion: landslide, debris slide, and surface erosion.
Study Limitations
Any perturbation to a hillslope system may result in either a large or small
erosional response depending on the balance of opposing tendencies preexisting
at the site. For example, a hillslope may be steep and may be underlain by easily
erodible regolith, but previous evacuation of material from the slope may have
left little material available for transport (Figure 1). Similarly, an area
of gentle slope may be deeply weathered and have excess material for transport,
but land use practices may incur little erosion because the slope is gentle.
Thus there is a complex interaction of multiple geomorphic variables which are
in a constant state of apparent balance. This concept was summarized by Hack
(1960), "A landscape is an open system which is in a steady state of balance
with every slope and every form adjusted to every other." The large number
of geomorphic variables controlling the evolution of a natural landscape creates
a system which is difficult, if not impossible to evaluate empirically (Leopold
and Langbein, 1962; Shreve, 1966, 1975; Smart, 1968, 1972). It is also difficult
to predict how close a landscape is to a threshold condition before system disturbance
(Bull, 1991).
For these reasons there are no established quantitative relationships between
the factors controlling erosion. Although it is difficult to predict the response
of a system to a change in land use, many scientists have noted qualitative
cause and effect relationships. For instance a disturbance to a slope of moderate
steepness on a specific geologic unit will tend to have accelerated erosion.
Thus, an essentially qualitative model based largely on the field experience
of numerous geologists and other resource specialists has been developed. The
design and component input of the model is therefore limited to that of a highly
simplistic and qualitative conceptual model.
Data Preparation
Watershed Boundaries
Approximately 53O watersheds, each about 20,000 hectares, were subdivided from
preexisting sub-basins defined by the Hydrologic Basin Planning Areas of the
RWQCB (1986). Land ownership maps were combined with forest coverage maps compiled
by the CDF in order to define areas of private or state ownership with commercial
timber. Only watersheds containing 25% or more private or state-owned commercial
timberland were selected for this study. Tierra Data Systems provided GIS coverage
of the 20,000 hectare watersheds used in this study. Previously existing watershed
boundaries were used where possible, causing many watersheds to be administratively
defined. Therefore, watershed boundaries are inconsistently represented across
the region and do not always reflect individual hydrologic basins. Such watershed
boundaries should not be used for scientific analysis; however, if aggregated
into Hydrologic Sub-Basin Areas as defined by the RWQCB (1986), complete hydrologic
basins are represented.
Slope
A digital slope map was produced for northern California using a derivative
of a 3-arc second raster data set that was developed for the United States by
the Army Map Service (now Defense Mapping Agency). The original data set was
resampled to 150 X 150m pixel resolution by the U.S. Geological Survey (USGS)
in Flagstaff, Arizona. This digital elevation model (DEM) was projected to a
Lambert Conformable Conic projection with a central meridian of 119o W longitude.
Although this data was carefully edited by the USGS, it retains numerous scanning
artifacts contained within the original 3-arc second data.
To evaluate the accuracy of DEM slopes, slopes from selected locations were
compared from measurements made in the field on topographic maps and from digital
data with a 150m pixel resolution. These comparisons show that:
1. Field measurements are difficult to correlate accurately to topographic maps
due to the complex microtopography in the field.
2. Although a comparison of digital slope measurements with topographic slopes
shows a discrepancy at individual sites, a consistent relationship exists between
digital slopes and topographic slopes when digital slopes are averaged over
an entire drainage basin.
3. Digital slopes from random locations are up to 7% to 10% lower on average
than topographic slopes (Figure 3).
4. Slopes are calculated in eight directions surrounding a central elevation.
The steepest slope is assigned to a 150 x 150m pixel. This tie-limited spatial
resolution of the digital slope calculations cause slopes less than about 300m
in length to be measured inaccurately (Figure 4).
Geology
Qualitative evaluations of geologic material strength were developed from personal
interviews with 21 professional geologists who have extensive experience with
erosion in timber harvest areas in northern California (Table 1). The geologists
were asked to classify the geologic units with which they were most familiar
in terms of susceptibility to 1) landsliding, 2) debris sliding, and 3) surface
erosion. Relative ratings of low, moderate, and high were assigned to each geologic
unit (Table 2). To insure regional validity for the erosion values, geologists
were asked to consider erosion responses on equivalent slopes, and to use their
statewide knowledge of erosion susceptible lithologies when drawing comparisons
for their specific area of expertise.
The geology data layer was digitized for generating the digital material-strength
field for erosion analyses using a vector scan of scribed linework of the 1:750,000
scale geologic map of the State of California (Jennings, 1977). Line editing
and polygon labeling were performed by USGS and DMG personnel at the USGS Western
Regional Center, Menlo Park, California. Coregistered hydrologic features (coastlines,
lake shorelines, and river channels) were also scanned to enable accurate coregistration
with the DEM.
Precipitation
The precipitation data were manually digitized by DMG personnel from 1:1,000,000
scale blue line maps prepared for the CDF (1984). Digitized precipitation maps
include, 1) mean annual precipitation (Rantz, 1969), 2) 12-hour intensity with
a 50 year recurrence probability, and 3) 2-hour intensity with a 50 year recurrence
probability. The location of isohyetal lines is highly generalized: rainfall
data in the study area were-extrapolated from only 150 rain gages -- about one
for every 1OO,OOOsq-km. In addition, few of these gages are in mountainous areas,
so orographic effects have been estimated using manual techniques (U.S. National
Weather Service, oral communication, 1993).
Analysis
To preserve the integrity of the DEM slope data, the Lambert Conic Conformable
projection (119 central meridian) of the DEM data was adopted as the map projection
for this project. The vector data layers (geology, precipitation, and watershed
boundaries) were reprojected to this coordinate system. The Lambert Conic Conformable
projection approximates the Albers Equal Area projection and therefore is well
suited for applications of regional spatial analysis. The vector data layers
(geology and precipitation) were rasterized to the level of precision of the
digital elevation raster layer (150 X 150m pixels). To estimate the relative
erosion potential of each pixel area, the raster data layers were added together
within each pixel area according to the ranking scheme summarized in Tables
2 and 3. Separate analyses were calculated for each of the three major slope
erosion categories - landslides, debris slide, and surface erosion. For each
of these analyses the pixel values within each watershed were summed and averaged
to calculate a rating value for each watershed. The results of these three erosion
ratings were then added to estimate the total erosion potential for each watershed
(Table 4). The highest possible theoretical rating is 9 where 100% of the watershed
contains high geology, precipitation, and slope ratings. The highest rated watershed
in our analyses is 7, and the average relative erosion rating is 4.
Systematic visual inspection of a digital overlay of stream channels, derived
from a digital scan of hydrologic features coregistered with the geology scan
and the 150 X 150m pixel DEM, indicates that the general locational precision
of the various data layers falls within +300m or 2 DEM pixels (+0.6mm at 1:500,000
scale).
MODEL OUTPUT
Sources of hillslope sediment include landslides, debris slides, and surface
erosion. Because the physical controls on failure for these three potential
source types differ, they were modeled independently.
Landslides
Landslides modeled in this study include mass failures that have planes of failure
that are relatively deep, generally greater than 3m, and have a fairly low width
to depth ratio. The types of failures included in the landslide model are rotational
and translational landslides (rock slumps, earth slumps, rock block slides,
and earth block slides) and earth flows (Varnes, 1978). Many scientists have
attempted to correlate deep-seated mass movement with the amount of precipitation;
however, these studies show that the relationship is complex (Iverson and Major,
1987; Swanson and Swanston, 1977; Swanston, 1981; Campbell, 1975). Although
the complex movement of subsurface water flow has thwarted attempts to use rainfall
as a systematic predictor of landslide movement, most geomorphologists agree
that the occurrence of deep-seated hillslope failure is related to seasonal
precipitation (Brunsden, 1993; Jahns, 1969; Keefer and Johnson, 1983; Swanson
and Swanston, 1977; Swanston, 1981).
The regional nature of this study, combined with the lack of empirical relationships
between rainfall and landslide movement requires the use of a highly generalized
rainfall distribution to model spatial patterns of relative ground saturation.
Mean annual precipitation was chosen to provide a pattern of general rainfall
for estimation of relative landslide potential. Low, moderate, and high ratings
have been assigned to precipitation values to reflect resultant, and highly
generalized landslide susceptibility categories (Table 3).
Likewise, the relationship between slope gradient and mass wasting processes
is highly generalized due to the complexities of geologic, climatic, and land-use
factors. For purposes of this model, steeper slopes are assumed to have a greater
driving force with slopes from 10% to 30% being assigned a low value, 30% to
50% moderate, and steeper than 50% as high.
Geologic structure and lithology are significant factors predisposing certain
terrane to mass movement. This trend is observed on geologic maps in California
that show the majority of mapped landslides to be concentrated on a few geologic
units. In some regions the overriding influence of lithology has created the
basis-of landslide classification (Takada, 1964). The area of highest landslide
propensity in the study area occurs in the central and eastern belts of the
Franciscan Complex. Here melange and highly sheared and faulted sedimentary
and metamorphosed sedimentary rock dominate. The mineralogic composition of
the rocks causes them to be conducive to weathering and alteration to clay-rich
material, becoming subject to extensive landslide and earthflow movement (Relsey,
1977). In fact, landsliding may be the dominant erosion process in the northern
Coast Ranges.
Debris Slides
Debris slides modeled in this study include mass failures that have surfaces
of failure that are relatively shallow, generally fewer than 3m, and have a
fairly high width to depth ratio. The types of failures included in the debris
slide model are rock, debris, earth falls and toples, debris slides, and debris
flows (Varnes, 1978). Other terms used include debris torrents, mudflows, debris
avalanches, soil flows, and soil slips (Cannon and Ellen, 1985; Campbell, 1975;
Keefer and Johnson, 1983; Wieczorek, 1987; Ellen and others, 1993; Caine, 1980;
Wentworth, 1943). Debris slides commonly occur where thin colluvial deposits
blanket less permeable bedrock or soil material (O'Loughlin, 1972; Swanston,
1974; O'Loughlin and Pearce, 1976; Ellen and others, 1993). Once saturated,
these deposits exceed the resisting forces and fail. Many studies have documented
the predictive relationship between rainfall intensity and shallow debris flows
(Campbell, 1975; Cannon and Ellen, 1985; Caine, 1980; Wieczorek, 1987). These
studies note that antecedent water storage followed by a high intensity storm
systematically triggers debris flows (Canon and Ellen, 1985).
Campbell (1975) and Wieczorek and Sarmiento (1983) indicate that 10 to 15 inches
of antecedent seasonal rainfall is sufficient to set the stage for debris slides.
Once field capacity of the soil mantle has occurred, a high intensity storm
with extreme 1 to 24-hour precipitation can cause saturation and failure (Figure
2). For a 12-hour duration storm, a failure threshold has been shown to occur
at a rainfall intensity of 0.2 to 0.4 inches/hour (Cannon and Ellen, 1985) and
0.25 inches/hour (Campbell, 1975; Caine, 1980). For this study rainfall intensities
below 0.2 inches/hour were considered to be low and above 0.4 inches/hour were
designated high. Isohyetal locations were obtained from a rainfall intensity
map of California with a 12-hour duration and 50-year recurrence probability
(CDF, 1984).
Many researchers have attempted to correlate debris slide occurrence with hillslope
gradient. Shallow mass wasting typically occurs on steeper slopes than do deep-seated
slides, with debris slides commonly occurring on slopes between 40% and 100%
(Campbell, 1975; Sidle and others 1985; Durgin and others, 1989; Corbett and
Rice, 1966; Rice and Foggin, 1971; Kesseli, 1943; Johnson and Sitar, 1990).
In northern and central California, low-cohesion material formed from weathered
granite or sandstone bedrock shows the highest tendency for debris slide failure.
In the Coast Ranges, the Redwood Creek and South Fork Mountain schists of the
Franciscan terrane are highly susceptible to debris slides; in the Sierra Nevada
and Klamath province, granitic plutons are commonly susceptible. However, mass
wasting is limited to areas where hillslope detritus is available; steep slopes
(over 100%) may be covered by little colluvium (Campbell, 1975)
Surface Erosion
In this study surface erosion includes sheetwash, ravelling, rilling, and gullying.
An undisturbed forest in its natural pristine condition usually yields very
little surface runoff (Dissmeyer and Foster, 1980). The forest ground cover
(litter, logs, and rock) protects the soil from raindrop impact and surface
runoff, creating infiltration rates which usually exceed rainfall intensity.
However, land use impact resulting from mechanical site disturbance, (including
road building, tractor yarding, site preparation, and fire) destroys vegetative
cover, locally compacts the soil and exposes bare soil to the erosive energy
of rainfall and runoff.
A few attempts have been made to quantitatively model approximations of surface
erosion controlling factors in forested regions (Dissmeyer and Foster, 1980;
California Soil Survey Committee, 1989). U. S. Department of Agriculture (USDA)
(1978) identified six factors contributing to surface erosion in agricultural
fields. However, the empirically derived relationships known as the Universal
Soil Loss Equation (USLE) shows a poor correlation to sediment yield on forested
hillslopes (Dodge and others, 1976). Part of this disparity results from the
derivation of the USLE on gently sloping, finely textured agricultural fields,
whereas forested landscapes are topographically, botanically, and lithologically
diverse and thus difficult to model over large areas.
Three factors were chosen to approximate regiona1 susceptibility of forested
hillslopes to surface erosion: a 2-hour high intensity rainfall storm with a
50 year recurrence probability slope, and lithologic potential for surface erosion.
Soil loss per unit area generally increases in proportion to a power of hillslope
gradient. In this study, surface erosion potential was rated low on 10% to 30%
slopes, moderate on 30% to 50% slopes, and high on greater than 50% slopes.
Spatial Relations
Areas of steepest slope include the Klamath physiographic province and the deep
canyons draining the west flank of the Sierra Nevada.
Mean annual precipitation is highest in northwestern California north of Eureka.
Annual rainfall is moderate in the Coast Ranges north of Santa Rosa, in the
Sierra Nevada Mountains, and in the Klamath province. Two-hour and 12-hour precipitation
intensities show a generally similar distribution to mean annual precipitation
with particularly high intensity rainfall over Arcata, Mount Shasta, and the
Santa Cruz Mountains.
The general lithologic patterns in California roughly coincide with geomorphic
provinces (Jenkins, 1938) and with tectonostratigraphic terranes (Auboiun and
others, 1980; Irwin, 1966) (Figure 5). These include 1) the coastal Franciscan
Complex, composed chiefly of Mesozoic and Cenozoic sedimentary and metavolcanic
rocks which are highly sheared and deformed, 2) the Klamath crystalline basement
complex consisting of highly metamorphosed Mesozoic and Paleozoic rock intruded
by Mesozoic plutons, 3) the Cascade and Modoc provinces comprised of late Cenozoic
and Quaternary volcanics, 4) and the Sierra Nevada Mountains cored by Mesozoic
granite and granodiorite intruding metamorphosed Mesozoic and Paleozoic igneous
and sedimentary roof pendants of the foothill region.
Erosion Potential
Watershed erosion ratings are high where one data layer is high and two out
of three are either high or moderate. Therefore, although the Klamath province
and the deep canyons of the Sierra Nevada are lithologically resistant, high
precipitation and slope values give these areas high erosion ratings.
Of course, estimates of erosion potential generated for each pixel show greater
geographic detail than estimates averaged over 20,000 hectare watersheds. This
apparent detail, however, is somewhat misleading in that the detailed breaks
between data units cause sharp contrasts which are not representative of actual
field conditions. The generalized nature of individual data layers, combined
with the uncertainty of geomorphic response furthermore causes imprecision of
erosion values at specific locations. Only when erosion values are averaged
over drainage basins do they accurately reflect the erosion potential.
Landslide Potential
The area of highest landslide potential exists in the Coast Range province,
specifically in the eastern and central belts of the Franciscan terrane north
of Clear Lake (Figures 5 and 6). Here, melange, clay-rich soil, and moderately
steep slopes combined with moderate to high precipitation (100 to 250cm/yr)
create unstable hillslopes. Landslide potential is generally low in the Sierra
Nevada, with a few landslide-prone watersheds on the more clay-rich weathered
metamorphosed Mesozoic and Paleozoic roof-pendant rock of the northern foothill
region. In the Klamath province, landslide potential is highly variable, ranging
from low to high with the highest potential occurring on the western side of
the province where serpentinized ultramafic rock, steep slopes, and high precipitation
create unstable hillslope conditions. Landslide potential in the Modoc and Cascade
provinces is low, with only a few localized problem sites.
Debris Slide Potential
Debris slide potential is modeled as low to moderate in the Coast Ranges (low
from the Santa Cruz Mountains to Santa Rosa, and moderate north of Santa Rosa)
(Figure 7). In the Klamath province, debris slide potential is highly variable,
ranging from low to high in a scattered pattern, while in the Cascade and Modoc
plateaus debris slide potential is low. In the Sierra Nevada Mountains the potential
is generally low with a few scattered watersheds having a moderate potential.
Surface Erosion Potential
Surface erosion ratings are low to moderate in the Coast Ranges, low to moderate
in the Klamath province, and low in the Cascade, Modoc Plateau, and Sierra Nevada
Mountains (Figure 8).
Total Erosion Potential
Total erosion potential combines landslide, debris slide, and surface erosion
ratings within each watershed (Figure 9). The geographic distribution of relative
erosion susceptibility shows a high potential in the northern coast ranges,
moderate in the Klamath province, moderate to low in the Sierra Nevada Mountains,
and low in the Cascade and Modoc Plateau physiographic provinces. This pattern
of general erosion susceptibility is similar to that of the modeled landslide
potential (Figure 6).
DISCUSSION
How well these modeled watershed values represent actual erosion potential is
important if the watershed-rating maps are to be used for planning purposes.
By understanding the limitations and uncertainties of the data used in this
analysis, a more realistic use of the model can be facilitated. This will also
aid in evaluating the accuracy of the model.
Four limitations contribute to the uncertainty of this analysis. First, geomorphic
processes range widely, even in similar physiographic settings. In northern
California, the inner-gorges of steeply sloping streams contribute significantly
to total sediment yield within local watersheds (De la Fuente and Haessig, in
review). Here steep stream canyon walls are cut into the toes of broad hillslopes.
The undermined colluvium at the toe of the slope creates a continual cascade
of weathered hillslope material into the streams. The break in hillslope gradient
in these canyons suggests that the stream and hillslope systems are not in equilibrium,
and that there has been a change in the rate of stream degradation at some point
in the recent past. This allows a reservoir of hillslope material to be available
for erosion.
Sediment transport from hillslopes to streams in many areas of the Sierra Nevada,
however, operates quite differently. Here stream gorges are often bedrock walled
and little or no colluvium is available for downslope transport. The process
of grussification of granitic rock requires moisture to be retained at depth
for prolonged periods of time. On steep slopes weathered material is rapidly
removed, and a self-enhancing feedback loop occurs where bare rock does not
retain moisture long enough to form crystalline detritus or gruss. Thus, in
some of the deep, steep-sided gorges of the Sierra Nevada, little colluvium
exists and, unlike the Coast Ranges, little sediment is available for stream
transport. Slope steepness, then, by itself may not be a reliable indicator
of erosion potential if no colluvial material is available for erosion.
Second, many workers have attempted to relate rates of sediment transport to
landscape variables, yet understanding the relationship between geomorphic variables
is incomplete. For this analysis a simple linear relationship between these
variabilities was used. In reality these relationships are highly complex and
involve the interactions of many variables not utilized in this analysis. Furthermore,
the highly generalized nature of the input data, combined with the lack of empirically
derived relationships between data sets, creates large uncertainties in the
accuracy of the model. To structure the analysis in such a way as to assume
precision between data sets would be to misrepresent the large uncertainties
involved in the data relationships. It is therefore appropriate, and has been
the consistent intent of this study, to keep all aspects of the model as simple
as possible including the analysis, as it is based largely on the qualitative
observations of landscape processes
Third, despite the unlimited number of factors affecting rates of hillslope
erosion, this model uses only three factors: slope, precipitation, and, lithology.
Although they are generally the most important under natural conditions, these
three factors account for only a part of the variability in erosion potential.
Fourth, the data input into this analysis is highly generalized. The geologic
units used at the 1:750,000 scale are amalgamations of several map units from
larger scale maps. These units, in turn, commonly include a variety of lithologic
types. The slope estimates are likewise inaccurate at less than a 300m grid.
Precipitation data is furthermore derived from one gaging station every 1OO,OOOsq-km.
Developing ways to quantify, map, and integrate the variables that influence
rates of erosion is a major challenge facing natural resource scientists. Although
this model of relative erosion susceptibility is highly qualitative, it attempts
to synthesize the physical attributes at a regional scale in California. To
appraise the reliability of the model for identifying problem watersheds, three
sources of data are examined: questionnaires inviting identification of known
areas of erosion, published information on specific watershed studies, and suspended
sediment yield data from large drainages.
Natural resource specialists working in timber-harvest- related activities were
asked to identify watersheds that had potentially high rates of erosion (Table
5). Agencies responding to this questionnaire included the CDF, the RWQCB, the
State Water Resources Control Board (SWRCB), DMG, and DFG. Nineteen responses
were received identifying 121 watersheds distributed throughout most of the
study area. To augment these observations, 32 studies of erosion in northern
and central California were reviewed. These data are summarized in Figure 10
and Table 5.
The observed problem watersheds generally correspond with modeled watersheds
having moderate to high ratings. Furthermore, the problem watersheds specifically
correspond with modeled landslide potential rather than debris slide or surface
erosion potential. In contrast, however, many of the erodible watersheds selected
by modeling are not recognized as susceptible to accelerated erosion by field
observations. Also, about 20% of those considered problem watersheds are in
areas of extremely low ratings based on the model. Several reasons could account
for this discrepancy. First, field observations may be incomplete or inconsistent.
Second, land use impacts are not part of the intrinsic erosion potential model
but may well be the primary factor creating observable erosion problems. Third,
additional important factors such as inner gorge development, glacial history,
and topographic maturity could add accuracy and detail to the resultant analysis.
Suspended sediment data support the general trends observed between the geomorphic
provinces found in the erosion model. These data indicate that the Coast Range
province in northwestern California is the most rapidly eroding area in the
conterminous United States (Holeman, 1968; Curtis and others, 1973). These especially
high erosion rates have been attributed to the lithologically unstable Franciscan
terrane, geologically recent tectonism, high and distinctly seasonal precipitation,
and major land use disruption. In contrast, the crystalline rock of the Klamath
province is generally characterized by substantially lower mean annual suspended
sediment yields than the Coast Ranges (Jones and others, 1972; De la Fuente
and Haessig, in review). However, in the Sierra Nevada Mountains, suspended
sediment yields are 30 times lower than averages from watersheds in the Coast
Ranges (Nolan and Hill, 1991).
ADDITIONAL WORK
Additional data layers could enhance the erodible watershed inventory model
by including information on land use history, soil properties, transient snow
zone boundaries, geomorphic limits of inner gorge development, and areas of
Pleistocene glaciation. The applicability of these data layers is described
below.
Land Use
Land use activities can result in substantial increases in soil erosion. Clear
cut timber harvesting and roading resulted in sediment yields 17 times higher
than those in comparable unharvested basins in the Redwood Creek area of northwestern
California (Janda, 1978). Although activities that directly increase erosion
rates have been substantially reduced by Forest Practice Act regulations, the
long term effects of past activities undoubtedly continue to influence sediment
yields.
Digital land use data have been developed for California by the EROS Data Center,
National Mapping Division, USGS and the University of Nebraska, Lincoln, using
Advanced Very High Resolution Radiometry Imagery (AVHRP). Older land use files
have also been published by the EROS Data Center as Land Use-Land Cover maps.
These files could be used to identify where temporal changes in land use activities
have occurred.
Soil Properties
The availability of weathered material is a critical component to erosion-potential
mapping. The depth and grain size distribution of the regolith, however can
be difficult to measure and varies widely even on homogeneous rock (Wahraftig,
1965). The USDA, Soil Conservation Service, has compiled digital coverage of
soils for the state (STATSGO). Although highly generalized, it could be a useful
data layer to supplement the lithologic component. Delineation of low cohesion
soils may be useful for modeling surface erosion and debris slide potential,
and high cohesion soils may provide further accuracy for modeling landslide
potential. These models could then be compared to and adjusted to those defined
by the geologic data. In addition, soil depth might be used for defining areas
of hillslope sediment availability.
Transient Snow Zone
Within a given drainage basin, large storm events can mobilize material equivalent
to many times the mean annual sediment yield (Janda and Nolan, 1979). In northern
California, extreme runoff events typically result from high intensity tropical
storms melting snow pack. It may be possible to delineate those areas where
this phenomenon, known as the rain- on-snow-zone, could impact drainage basins.
By using the methodology developed by the State of Washington (Green and others,
1993), a model could be developed that would define the boundaries of those
areas most susceptible to rain-on-snow events.
Inner Gorge
Steep inner gorges contribute significantly to total sediment yield in many
streams (De la Fuente and Haessig, in review). Areas containing these landform
features could be outlined and added as another data factor.
Review of published literature, interviews with geologists and geomorphologists,
aerial photographic interpretation, and field mapping will be required to identify
boundaries separating regions where inner gorges are common from those areas
where they are rare. Because of the complex nature of the tectonic processes
that result in rapid watercourse base level changes, separating these areas
will be time consuming and controversial.
Pleistocene Glaciation
Glacial scour has removed weathered hillslope material from areas of high elevation
in the Sierra Nevada, Klamath, and Cascade physiographic provinces. Delineating
areas of glaciation would further define watersheds with limited availability
of weathered material.
Watershed-Scaled Analysis
A GIS-based model of erosion potential could be developed for a relatively small
watershed to attempt to quantify erosion controlling factors. Detailed geomorphic
mapping at a scale of 1:24,000 by DMG is in progress for three watersheds in
northern California. A geology and geomorphology data layer combined with digital
elevation data or a digital terrane model developed from aerial photographs
could be used in conjunction with sediment yield data to define more detailed
algorithms for sediment yield modeling.
Conclusion
The understanding of the relationship between geomorphic variables is complex
and incomplete. As research on sediment yield, sediment transport, and geomorphology
advances, the understanding of the relative significance of individual factors
controlling erosion will be improved. The results of current and future research,
in conjunction with the suggested further work, can be used to improve the quality
of the erodible watershed inventory model.
REFERENCES CITED
Auboiun, J., Blanchet, R., and Rangin, C., 1980, Reunion extraordinaire de la societe geologique de France en Californie, Bull. Soc. geol. France, vol 7, no. 4, p.511-553
Bedrossian, T.L., Geology and slope stability in selected parts of the Geysers Geothermal Resources area: A guide to geologic features indicative of stable and unstable terrane in areas underlain by Franciscan and related rocks: Division of Mines and Geology Special Report 142, 65p.
Bishop, D.M. and Stevens, M.E., 1964, Landslides on logged areas in southeast Alaska; Res. Pap. NOR-1, Forest Service U.S. Dep. of Agric., Juneau, Alaska, 18p.
Brown, W.M., 1975, Sediment transport, turbidity, channel configuration and possible effects of impoundment of the Mad River, Humboldt County, California; U.S. Geological Survey Water Resources Investigations 26-75, 63p.
Brunsden, D., 1993, Mass movement; the research frontier and beyond: a geomorphological approach; Geomorphology, v. 7, p. 85-128
Bull, W. B., 1991, Geomorphic Responses to Climatic Change; Oxford University Press, New York, 326 p. Campbell, R. H., 1975, Soil slips, debris flows, and rainstorms in the Santa Monica Mountains and vicinity, southern California: U.S. Geological Survey Professional Paper 851, 51p
Caine, N., 1980, The rainfall intensity control of shallow landslides and debris flows; Geografiska Annaler, 62A, p.23 -27
California Department of Forestry and Fire Protection, 1984, California precipitation intensity maps: 50 year return period with 8 different durations. Maps prepared by California Department of Conservation from California Department of Water Resources data, 22p
California Department of Water Resources, 1979, South Fork Trinity River watershed erosion investigation, 8lp
California Soil Survey Committee, 1989, Erosion hazard rating system for sheet
and rill erosion, unpublished report, 8p
Campbell, R.H., 197S, Soil slips, debris flows, and rainstorms in the Santa Monica Mountains and vicinity, southern California: U.S. Geological Survey Professional Paper 851, 51p
Cannon, S.H. and Ellen, S.D., 1985, Abundant debris avalanches; Calif. Geology, p.267-272
Carson, M.A., and Petley, D.J., 1970, The existence of threshold hillslopes in the denudation of the landscape; Trans. Inst. British Geogr., v.49, p.71-95
Corbett, E.S. and Rice, R.M., 1966, Soil slippage increased by brush conversion; Res. Note PSW-128, For. Serv., U.S. Dept. Agric., 8p
Curtis, W.F., Culbertson, J.K., and Chase, E.B., 1973, Fluvial sediment discharge to the oceans from the conterminous United States: U.S. Geological Survey Circular 670, 17 p
Dodge, M., Burcham, L.T., Goldhaber, S, McCulley, B., and Springer, C., 1976, An investigation of soil characteristics and erosion rates on California forest lands; Calif. Div. Forestry, 105p
De la Fuente, Juan, and Haessig, P.A., in review, Landslide and surface erosion rates in the English Peak Batholith and Ashland Pluton, central Klamath Mountains, California and Oregon
Dissmeyer, G.E. and Foster, G.R., 1980, A guide for predicting sheet and rill erosion on forest land: U.S. Dept. of Agriculture, Forest Service, State and Private Forestry, Southeastern Area, Technical Publication SA-TP 11, 40p
Durgin, P.B., Johnston, R.R. and Parsons, A.M., 1989,Causes of erosion on private timberlands in northern California: Observations of the Interdisciplinary Team: Critical Sites Erosion Study, California Department of Forestry and Fire Protection, Forest Practices Section
Ellen, S.D., Mark, R.K., Cannon, S.H., Xnifong, D.L., 1993, Map of debris-flow hazard in the Honolulu District of Oahu, Hawaii; U.S. Geol. Surv., Open-File report 93-213, 25p
Ernst, W.G., 1984, map of lithotectonic belts of northern and central California; Proceedings of the Circum-Pacific Terrane Conference, Stanford University, p.87
Green, K, Bernath, S, Lackey, L, Brunengo, M., Smith, S. 1993, Analyzing cumulative effects of forest practices: Where do we start; Geo. Info. Systems, p.31-41
Hack, J.T., 1960, Interpretation of erosional topography in humid temperate regions: American Journal of Science (Bradley Volume) 258-A, p.80-97. Leopold, L.B., and Loooangbein, W.B., 1962, The concept of entropy in landscape evolution: U.S. Geological Survey Professional Paper 252
Irwin, W.P., 1966, Geology of the Klamath Mountains Province; in: Geology of Northern California, California Division of Mines and Geology, Bull. 90, p.19-38
Holeman, J.N., 1968, The sediment yield of major rivers of the world: Water Resources Research, V. 4, p.737-747
Iverson, R.M. and Major, J.J., 1987, Rainfall, ground-water flow and seasonal movement at Minor Creek landslide, northwestern California: Physical interpretation of empirical relations; Geol. Soc. Amer. Bull., v. 99, p.579-594
Janda, R.J., 1978, Summary of watershed conditions in the vicinity of Redwood National Park, California: U.S. Geological Survey Open-File Report 78-25
Janda, R.J. and Nolan, K.M., 1979, Stream sediment discharge in northwestern California; in: Geol. Soc. Amer. Field Trip Guidebook, Cordilleran Sect., San Jose, v. 4, 26p
Janda, R.J., Nolan, K.M., Harden, D.R., and Colman, S .M., 1975, Watershed conditions in the drainage basin of Redwood Creek, Humboldt County, California, as of 1973; U.S. Geological Survey, Open-File Report 75-568, Menlo Park, CA, 267p
Jahns, R.H., 1969, California's ground-moving weather; Engineering and Science Magazine, Calif. Inst. of Tech., Pasadena, 8p.
Jenkins, O.P., 1938, Geomorphic map of California; California Division of Mines and Geology.
Jennings, C.W., 1977, Geologic map of California; California Division of Mines and Geology, scale 1:750,000
Johnson and Sitar, 1990, Hydrologic conditions leading to debris- flow initiation; Canadian Geotechnical Journal, v. 27, p.789-801
Jones, B.L., Hawley, N.L., and Crippen, J.R., 1972, Sediment transport in the western tributaries of the Sacramento River, California: U.S. Geological Survey Water-Supply Paper 1798-J, 27p
Keefer, D.K. and Johnson, A.M., 1983, Earth flows; morphology, mobilization and movement; U.S. Geol. Surv., Professional Paper, no. 1264, 56p
Kelsey, H.M., 1977, Landsliding, channel changes, sediment yield and land use in the Van Duzen River basin, North coastal California, 1941-1975; Ph.D. Dissertation, Univ. California, Santa Cruz, 357p
Kesseli, J.E., 1943, Disintegrating soil slips of the Coast Ranges of central California; Journ. of Geol., v. 51, p.342-352
Kilbourne, Richard, 1985, Geology and geomorphic features related to landsliding in the Fortuna 7.5' Quadrangle, Humboldt County, California
Larson, K.R. and Sidle, R.C., 1980, Erosion and sedimentation data catalog of the Pacific Northwest; Cooperative agreement no. 227, U.S. Dept. of Agriculture, Forest Servlce and Oregon State University, 64p
Leopold, L.B., and Langbein, W.B., 1962, The concept of entropy in landscape evolution; U.S. Geological Survey Professional Paper 252
Lewis, J. and Rice, R., 1989, Site conditions related to erosion on private timberlands in northern California: Final Report: Critical Sites Erosion Study, California Department of Forestry and Fire Protection, Forest Practices Section
Mathewson, C.C. and Clary, J.H., 1977, Engineering geology of multiple landsliding along I-45 road cut near Centerville, Texas; ln Reviews in Engineering Geology, v. 3, Landslides, Geol. Soc. Amer., Boulder, Colo., p.213-233
Megahan, W.F., Day, N.F., and Bliss, T.M., 1978, Landslide occurrence in the western and central Northern Rocky Mountain physiographic province in Idaho; in Proceedings of the 5th North American Forest Soils Conference, Colo. State Univ., Fort Collins, p.116-139
Morrison, P.H., 1975, Ecological and geomorphological consequences of mass movements in the Alder Creek watershed Univ. of Oregon, Eugene, 102p
Nolan, M.K., and Hill, F.R., 1991, Suspended-sediment budgets for four drainage basins tributary to Lake Tahoe, California and Nevada, 1984-87
Nolan, K.M. and Janda, R.J., 1981, Use of short-term water and implications for forest land management; B.A. thesis, and suspended-sediment discharge observations to assess impacts of logging on stream sediment discharge in the Redwood Creek basin northwestern California, U.S.A.; in Erosion and Sediment Transport in Pacific Rim Steeplands. I.A.H.S. Publ. No. 132, Christchurch, 1981
O'Laughlin, C.L., 1972, The stability of steepland forest soils in the Coast Mountains, southwest British Columbia, Ph.D. dissertation, Univ. of British Columbia, Vancouver, 147p
O'Laughlin, C.L., and Pearce, A.J., 1976, Influence of Cenozoic geology on mass movement and sediment yield response to forest removal, North Westland, New Zealand; Bull. Int. Assoc. Eng. Geol., v. 14, p.41-46
Palmer, L., 1977, Large landslides of the Columbia River Gorge, Oregon and Washington; in Reviews in Engineering Geology, Geol. Soc. Amer., Boulder, Colo., v. 3, Landslides, p.69-83
Peters, J.H. and Litwin, Y., 1983, Factors influencing soil erosion on timber-harvested lands in California; WESCO, unpublished report, 94p
Rantz, S.E., 1969, Mean annual precipitation in the California region; U.S. Geological Survey, Water Resource Division
Reid, L.M. and Dunne, T., 1984, Sediment production from forest road surfaces; Water Resources Research, v.20, no.11, p.1753-1761
Rice, R.M. and Foggin, G.T., 1971, Effect of high intensity storms on soil slippage on mountainous watersheds in southern California; Water Resour. Res., v.7, no.6, p.1485-1496
Rice, R.M., Tilley, F. B., Datzman, P.A., 1979, A watershed's response to logging and roads: south fork Casper Creek, California, 196701976; U.S. Forest Service, Southwest Pacific Research Station, PSW-146
Scott, Ralph, Buer, Koll, and James, Steve, 1979, South Fork Trinity River watershed erosion investigation; California Department of Water Resources report, 70p
Shreve, R.L., 1966, Statistical law of stream numbers; Journ. of Geology, v. 74, p.17-37
Shreve, R.L., 1975, The probabilistic-topologic approach to drainage-basin geomorphology; Geology, v.3, p.527-529
Sidle, R.C., Pearce, A.J., O'Loughlin, C.L., 1985, Hillslope Stability and Land
Use; Amer. Geophysical Union, Wash., D.C., 140 p
Smart, J.S., 1968, Statistical properties of stream lengths; Water Resources Research, v. 4, p.1001-1014
Smart, J.S., 1972, Channel networks; Advances in Hydroscience, v. 8, p.305-346
Sommarstrom, Sari, Kellog, Elizabeth, Kellog, Jim, 1990, Scott River watershed granitic sediment study; Report for Siskiyou Resource Conservation District, 116p
Spittler, T.E., 1982, Geology and geomorphic features related to landsliding; California Division of Mines and Geology, Watershed Mapping Program, scale 1:24,000
Spittler, T.E., 1983, Geology and geomorphic features related to landsliding; California Division of Mines and Geology, Watershed Mapping Program, scale 1:24,000
State Water Resources Control Board, 1986, Hydrologic basin planning area maps, scale 1:500,000
Swanson, F.J., and Swanston, D.N., 1977, Complex mass-movement terranes in the western Cascade Range, Oregon; in Reviews in Engineering Geology, v.3, Landslides, Geol. Soc. Amer., Boulder, Colo. p.113-124
Swanston, D.N., 1974, Slope stability problems associated with timber harvesting in mountainous regions of the western United States; Gen. Tech. Rep. PNW-21, For. Serv., U.S. Dept. Agric., Oregon, 14p
Swanston, D.N., 1981, Creep and earthflow erosion from undisturbed and management impacted slopes in the Coast and Cascade Ranges of the Pacific Northwest, U.S.A.; IAHS AISH Publ. 132, p.76-94
Takada, Y., 1964, On the landslide mechanism of the Tertiary-type landslide in the thaw time; Bull. Disaster Prevention Res. Inst. Kyoto Univ., v.14, p.11-21
U.S. Department of Agriculture, 1978, Predicting rainfall erosion losses, Handbook number 537, p.57
U.S. Department of Agriculture, U.S. Soil Conservation Service, 1986, Grass
Valley Creek sediment study, 49p
Varnes, D.J., 1978, Slope movement and types and processes; in: Landslides;
Analysis and Control; Transportation Research Board; National Academy of Sciences,
Washington D.C., Special Report 176, ch.2
Waharaftig, C.A., 1965, Stepped topography of the southern Sierra Nevada; Amer. Geol. Soc. Bull., v.76, p.1165-1190
Wieczorek, G.F. and Sarmiento, J., 1983, Significance of storm intensity-duration for triggering of debris flows near La Honda, California; Geological Society of America Abstracts with Programs, v. 15, No.5, p.289
Wentworth, C.K., 1943, Soil avalanches on Oahu, Hawaii: Geological So. Amer. Bull., v. 54, p.53-64
Wieczorek, G.F., 1987, Effect of rainfall intensity and duration on debris
flows in central Santa Cruz Mountains, California: Geological Society of America,
Reviews in Engineering Geology, Volume VII, p.93-104
Generated by mp
version 2.7.3 on Wed Oct 02 11:20:27 2002