WATER POLLUTION CONTROL RESEARCH SERIES
13010 EGA 02/71
STUDIES ON EFFECTS OF WATERSHED
PRACTICES ON STREAMS
U.S. ENVIRONMENTAL PROTECTION AGENCY
-------
WATER JOLLUnON CONTROL RESEARCH SERIES
The Water Pollution Control Research Series describes the
results and progress in the control and abatement of pollution
in our lation's waters. They provide a central source of
information on the research, development and demonstration
activities in the Environmental Protection Agency, through
inhouse research and grants and contracts with Federal,
State, and local agencies, research institutions, and
industrial organizations.
Inquiries pertaining to Water Pollution Control Research
Reports should be directed to the Chief, Publications Branch
(Water), Research Information Division, B&M, Environmental
Protection Agency, Washington, D.C. 20460.
-------
STUDIES ON EFFECTS OF WATERSHED PRACTICES ON STREAMS
by
School of Forestry
School of Engineering
Oregon State University
Corvallis, Oregon 97331
for the
ENVIRONMENTAL PROTECTION AGENCY
Grant No. 13010 EGA
February 1971
-------
EPA. Review Notice
This report has been reviewed by the Environmental Protec'
Agency and approved for publication. Approval does not
signify that the contents necessarily reflect the vievs ai
policies of the Environmental Protection Agency nor does
mention of trade names or commercial products constitute
endorsement or recommendation for use.
ii
-------
ABSTRACTS
Chapter I - Effects of Clearcutting on Stream Temperature
The purpose of the substudy was to describe the long-term effects of
clearcutting timber on two small streams in the Oregon Coast Range.
One watershed contained three small clearcuts; the edges of the
clearcuts were generally 100 feet from the main stream, thus providing
shade. The second watershed was completely clearcut and most of the
main stream was exposed. A third watershed and some uncut subwatersheds
were retained as uncut controls. The three watersheds range from 175 to
750 acres.
The diurnal temperature regime was not altered after logging on the
watershed with three small clearcuts. The fully clearcut watershed had
a maximum diurnal change in temperature of 8°F. before it was logged,
but 28°F. after logging.
The maximum temperatures recorded in the 2 years after logging on the
patchcut watershed was 60 and 61.5°F.--little different from the control.
The maximum temperature on the fully exposed stream after logging was
85°F., a 28°F. increase. All temperature regimes had a trend toward
the prelogging condition in subsequent years.
The principal conclusion was that temperature change in these small
streams is associated with the degree of exposure to sunlight. The
changes in temperature lessen as the area along the stream revegetates.
Streamside strips will minimize temperature change after logging. The
decision to leave such strips depends on timber and aquatic values, the
degree of temperature change anticipated if timber is removed, and
re-establishment rate of streamside vegetation.
111
-------
Chapter II - Predicting the Effect of Clcarcutting on Stream Temperature
The purpose of this study is to present a practical technique for
estimating the maximum change in temperature that would result after
the shade has been removed from forest streams by clearcutting.
Previous prediction studies by Brown had shown that accurate hourly
predictions could be achieved on small streams using energy budget
techniques. Solar radiation accounted for over 95 percent of the heat
input during the midday period of midsummer and led to the conclusion
that a simple technique could be applied in estimating temperature
change.
The components of prediction arc
AT = A * H x 0.000267
where AT is the predicted change in temperature; A = surface area in
square feet of the stream exposed by clearcutting, excluding isolated
pools; II = rate of heat absorbed by the stream in British Thermal
Units; D = minimum discharge rate in cubic feet per second.
Application of the method is described in detail.
IV
-------
Chapter III - Heat Loss from a Thermally Loaded Stream
The purpose of this study was to ascertain the way heated streams lose
heat as they flow into forests from a clearcut.
Energy budget components were measured using a track and trolley
system especially developed for making spatially integrated radiation
measurements above small streams. Other measurements included air and
water temperature, vapor pressure of the air, and wind speed.
The study site was located downstream from a clearcut in the Cascade
Mountains near Roseburg, Oregon. The water temperature reached 80 to
85°F. during midday in the areas of the clearcut.
Energy disposition by radiation, evaporation, convection, and
conduction and storage are given for a fully shaded reach below a
clearcut, and for a buffer strip within a clearcut. Heat was added to
the streams in both cases; amounts were larger where the buffer strip
shaded the stream, but in neither case did stream temperature change.
This study confirmed the earlier work on temperature prediction by
Brown (1969) that temperature change is closely associated with shade.
The hypothesis that shaded zones below clearcuts will cool heated water
is questionable. Further, the study showed the effectiveness of narrow
buffer strips in controlling water temperature.
-------
Chapter IV - Heat Flow in Stream Beds
The role of the stream bed in heat exchange was studied by measuring
thermal conductivities of various bed materials. Thermal gradients wer<
measured in stream beds and cores were obtained for laboratory analysis
The study was terminated before these both could be evaluated.
VI
-------
Chapter V - Clearcut Logging and Sediment Production in the Oregon
Coast Range
The purpose of this paper was to discover the effect of road building,
clearcut logging, and slash burning on suspended sediment production
from three forested watersheds in Oregon's Coast Range where
precipitation averages 100 inches annually and topography is steep.
Sediment-yield characteristics were monitored for seven years prior to
road building and clearcut logging in two of three watersheds; and for
four additional years during and after treatment.
On Deer Creek and Needle Branch, mainline roads were built in 1965. In
1966, Deer Creek was clearcut in patches at three locations (about 25
percent cut) and Needle Branch was completely clearcut. Slash in one
unit of the patch-cut watershed and in all of Needle Branch was burned.
A third watershed was left as a control.
Road building significantly increased sediment yield in both of the
treated watersheds. Annual sediment yield increased markedly after
logging and burning, and although it declined in subsequent years, was
still significantly higher than normal for four years after treatment
had been initiated.
During low flow periods (below 5 csm), even the most severe treatment
on Needle Branch resulted in only a few days with concentration above
10 ppm.
The results indicate that clearcut logging does not cause major impact
on sediment concentrations, but that road building and perhaps burned
areas are important sources of sediment.
VII
-------
Chapter VI - Evaluation of Bed Load and Total Sediment Yield Processes
on Small Mountain Streams
A facility was developed to study bed-load processes in small gravel-
carrying, mountain streams. One portion had a reach subject to natural
runoff and one portion had a controllable channel where gravel beds coul
be placed. The facility provided data on bed-load transport rates for
gravelly mountain streambeds, on limits of scour during storms and on
interrelations between the streambed and the suspended sediment load of
the stream.
An "armor" layer of coarse gravels at the top surface of the streambed
was examined. The layer appeared to protect the streambed from scour
until discharges became sufficiently great, after which the bed was set
in motion to a depth of several particle diameters. As discharges
receded after storms, the armor layer acted as a silt reservoir for the
stream.
A major contribution of this research is the development of an
effective tool for sampling the bed load of a stream in sediment-yield
studies of water quality, turbidity, and watershed practices and their
relationship to stream hydraulics.
In large peak discharges, the bed-load at the study area approached 70
percent of the suspended-load transport.
Basic research on bed-load transport rates during storm runoff should
continue to extend the tentative results obtained to date, by this
study, so as to better assess the effect of logging practices and other
watershed practices upon streamflow, sediment transport, and stream
channel changes.
This report was submitted in fulfillment of Project 13010EGA under the
partial sponsorship of the Environmental Protection Agency.
VI11
-------
CONTENTS
Page
Chapter
T~ EFFECTS OF CLEARCUTTING OH STREAM TEMPERATURE 1
Introduction 1
The Study 3
Results 6
Discussion 12
References 14
II. PREDICTING THE EFFECT OF CLEARCUTTING ON STREAM
TEMPERATURE 15
Introduction 15
Temperature Prediction 16
Discussion 22
References 24
III. HEAT LOSS FROM A THERMALLY LOADED STREAM 25
IV. HEAT FLOW IN STREAM BEDS 31
References 34
V. CLEARCUT LOGGING AND SEDIMENT PRODUCTION IN THE
OREGON COAST RANGE 35
Introduction 35
The Study 38
Changes in Annual Sediment Load 41
Changes in Sediment Concentration 46
Discussion 52
References 55
VI. EVALUATION OF BED LOAD AND TOTAL SEDIMENT YIELD
PROCESSES ON SMALL MOUNTAIN STREAMS 58
Introduction 58
Objectives 61
Experimental Approach 63
Development of Oak Creek Research Facilities 66
Experimental Procedures and Methodology 80
liydrologic Characteristics of Oak Creek Study Reach S7
Changes in Channel Topography 102
Suspended Sediment Load 124
Bed-Load Transport 128
Bed Measurements, Alsea Experimental Watersheds 165
References 169
Appendix 171
Acknowledgments 173
IX
-------
LIST OF FIGURES
CHAPTER I
NUMBER TITLE 3E_
1 The watersheds of the Alsea Basin Logging - Aquatic 4
Resources Study.
2 The maximum diurnal change in temperature recorded on 7
the clearcut and uncut control watersheds before
(1965), during (1966), and after (1967-1969) logging.
3 A frequency distribution of diurnal temperature changes 9
on the clearcut, patchcut, and uncut control watersheds
before (1965), during (1966), and after (1967-1969)
logging.
4 Mean monthly maximum temperatures for the clearcut and -0
uncut control watersheds before (1965), during (1966),
and after (1967-1969) logging.
CHAPTER II
Hourly values for net solar radiation above water 20
surfaces in clear days between latitudes SON and SON
for several solar paths.
Average net solar radiation absorbed by streams between 20
latitudes SON and SON on clear days during several
periods of exposure to different solar paths.
CHAPTER III
1 A system for spatially integrating solar radiation. 26
2 The digital data acquisition system. 27
3 An energy balance on a shaded reach of Cedar Creek 28
during a clear day in July, 1969.
4 Net all-wave radiation measured in a shaded reach of 30
Cedar Creek .and within a shaded buffer strip on Little
Rock Creek during clear days in July, 1969.
-------
CHAPTER IV
NUMBER TITLE PAGE
1 Cutaway view of frozen core sampler in operation 32
Arrows indicate direction of acetone circulation
(After Ringler, 1969).
2 The thermal conductivity chamber. 33
CHAPTER V
The study watersheds and sediment sampling stations 39
in the Alsea Watershed Study.
A comparison of normalized annual sediment yield from 43
Flynn Creek (unlogged) and Deer Creek (patchcut) for
seven years prior to treatment. Comparative yields
after road building (1966) and logging (1967-1969) are
shown in relation to the 95 percent confidence limit
about the pretreatment regression.
A comparison of normalized annual sediment yield from 44
Flynn Creek (unlogged) and Needle Branch (clearcut)
for seven years prior to treatment. Comparative yields
after road building (1966) and logging and burning
(1967-1969) are shown in relation to the 95 percent
confidence limit about the pretreatment regression.
CHAPTER VI
1 Schematic view of research facilities at instrumented 67
reach of Oak Creek.
2 Broad-crested weir and vortex-type sediment sampling 70
system (weir/trap structure).
3 Discharge capacity of vortex bed-load sampler 73
4 Flow across weir/trap structure during operation of 74
vortex bed-load sampler.
XI
-------
NUMBER TITLE PAGE
5 Variation of Froude number with discharge during 75
operation of the vortex bed-load sampler.
6 Schematic views of instrumented concrete channel. 77
7 Detail views of concrete channel. 78
8 Location map for Oak Creek sediment research facilities. 83
9 Rainfall and streamflow at Oak Creek weir/trap 92
structure in mid-January, 1970.
10 Rating curves for Oak Creek at the weir/trap structure. ^
11 Mean daily discharge at Oak Creek weir/trap structure 95
during 1969-1970 field studies.
12 Variability of velocity with discharge at Oak Creek 97
streamgaging sections.
13 Variability of hydraulic parameters with discharge 99
at Oak Creek streamgaging section upstream of weir/
trap structure.
14- Locations of cross-sectioning stations, scour ball 103
devices, and staff gages.
15 Channel topography in October 1969, before winter 106
bed-load transport.
16 Channel topography in late February 1970, after 107
winter bed-load transport.
17 Net s-easonal changes in channel topography, October 108
1969 to late-February 1970, due to winter bed-load
transport.
18 Sequential changes in cross-sectional shape at Hi-
sections 1-4, winter 1969-70.
19 Sequential changes in cross-sectional shape at 112
sections 5-6A, winter 1969-70.
20 Sequential changes in cross-sectional shape at 113
sections 7-8A, winter 1969-70.
21 Sequential changes in cross-sectional shape at 114
sections 9-11A, winter 1969-70.
XI1
-------
NUMBER TITLE PAGE
22' Sequential changes in cross-sectional shape at 115
sections 12-13A, winter 1969-70.
23 Sequential changes in cross-sectional shape at 116
sections 14-15A, winter 1969-70.
24 Sequential changes in cross-sectional shape at 117
sections 16-18, winter 1969-70.
25 Suspended sediment concentration and unit transport 127
rate as functions of discharge.
26 Sediment delivery ratio as a function of basin size. 132
t
27 Typical particle gradation curves for the bed surface 135
in Oak Creek study reach, fall 1969.
28 Variation of mean size and coarse fraction of streambed 136
surface layer with location along Oak Creek study reach,
fall 1969.
29 Variation of size gradation of bed material with depth 159
below bed surface, Oak Creek, fall 1969.
30 Unadjusted bed-load transport curves, Oak Creek, winter 145
1969-70.
31 Superimposed unadjusted bed-load transport curves, I4"7
Oak Creek, winter 1969-70.
32 Bed-load transport for sand-size material, Oak Creek 148
winter 1969-70.
33 Trap efficiency for sand-size bed-load material, Oak 149
Creek, winter 1969-70.
31 Adjusted total bed-load transport rate, Oak Creek, winter 152
1969-70.
35 Superimposed adjusted bed-load transport curves, Oak 153
Creek, winter 1969-70.
36 Unit bed-load transport rate as function of discharge, 154
Oak Creek, winter 1969-70.
37 Estimated critical discharge to initiate bed-load 155
transport near Oak Creek weir/trap structure.
XI11
-------
NUMBER TITLE PAGE
38 Laboratory and field data on critical shear stress 157
required to initiate movement of particles.
39 Variations in characterizing particle sizes for bed- 158
load samples as function of bed-load transport rate,
Oak Creek, winter 1969-70.
40 Distance moved and probability of movement in 161
cainted-gravel experiment, Oak Creek, 16-17
February 1970.
41 Bed-load/suspended-load relation as function of 162
discharge, Oak Creek, winter 19G9-7L .
4-2 Hydrographs used to estimate sedinwi yield, Oak 163
Creek, winter 1969-70.
43 Streambed material characteristics, -lected 167
alluvial oortions of Deer Creek.
xiv
-------
LIST OF TABLES
CHAPTER I
NUMBER TITLE PAGE
1 A time series analysis of the difference between 11
daily maximum temperatures of streams before and
after patchcutting, for the period of June 1 to
October 1.
CHAPTER II
1 Incoming shortwave and diffuse radiation on a IB
horizontal surface, Btu/ft2-min.
2 Percent reflection from water surfaces, after 18
Dirmhirn (1964).
CHAPTER V
Total annual sediment yield in tons per square 42
mile computed from U.S. Geological Survey records.
An analysis of simultaneous sampling of suspended 48
sediment concentration and streamflow at U.S.
Geological Survey stations during rising stages and
discharges greater than 5 csm.
An analysis of simultaneous sampling of suspended 49
concentrations and streamflow at stations within
Deer Creek during rising stages and discharges
greater than 5 csm.
A frequency distribution of mean daily suspended 5]
sediment concentrations during days with mean
flow less than 5 csm.
xv
-------
CHAPTER VI
PAGE
1 Sieve series and equivalent particle diameters for 83
analysis of sediment samples in study.
2 Monthly rainfall totals during field studies. S°
3 Daily rainfall totals during winter 1969-70. yi
'i Mean hydraluic gradients for study reach. 100
5 Times of collecting data on scour and changes in Iu4
channel topography, 19G9-70.
6 Survey data for scour ball devices, winter 1969-70. 120
7 Scour and deposition indicated by scour ball devices, 122
winter 1969-70.
8 Oak Creek suspended sediment concentrations, winter 12.-^
1969-70.
9 Particle size distribution for Oak Creek bed material 134
forming the surface layer i the study reach, fall
1969.
10 Variation of streambe" particle size distribution with 138
depth below bed surface •<•. k Creek, fall 1969.
11 Uniformity of Oak Creek bee material at selected 140
locations, fall 1969.
12 Variation of particle weights given size ranges for 141
Oak Creek armor layer, fall 1969.
13 Particle size distributions for bed- load samples, Oak 143
Creek, winter 1969-70.
14 Summary of bed- load transport data, Oak Creek, winter 144
1969-70.
15 Adjustment of bed-load transport data for trap 151
efficiency of sand-size material, Oak Creek, winter
1969-70.
16 Painted-gravel experiment, Oak Creek, 16-17 February 159
1970.
xvi
-------
NUMBER TITLE PAGE
17 Bed load, suspended load, and total sediment yield 164
for storm-runoff periods, Oak Creek, winter 1969-70.
18 Deer Creek streambed materials in selected alluvial 166
areas.
xvn
-------
CHAPTER I
EFFECTS OF CLEARCUTTING ON STREAM TEMPERATURE
SECTION I
INTRODUCTION
Timber, water, and sport and commercial fish are the principal
resources in the Oregon Coast Range. The need for delineating
the areas of conflict between logging and utilization of the other
resources led to the establishment of the Alsea Logging-Aquatic
Resources Study in 1958. The purpose of this broadly interdisciplinary
study was to determine the effect of logging on the physical,
chemical, and biological characteristics of small coastal streams.
The purpose of this paper is to describe the long-term effects
of two clearcuttings on the temperature regime of two small streams
in Oregon's Coast Range. One watershed contained three small
clearcuts; the edges of the clearcuts were at least 100 feet from
the stream. The second watershed was completely clearcut. An
earlier report (Brown § Krygier, 1967) described the first-year
effect of clearcutting only during the logging operation on the
completely clearcut watershed. This report reviews results from
a network of eighteen thermograph stations distributed through
the watersheds. The observation period extends from two years
before logging through the fourth summer after logging.
Temperature is a significant water-quality parameter. It strongly
influences levels of oxygen and solids dissolved in streams.
Temperature changes can induce algal blooms with subsequent changes
in taste, odor, and color of a stream. Warm water is conducive
to the growth and development of many species of aquatic bacteria,
such as the parasitic columnaris disease. Increased populations
of these bacteria may cause fish mortality [Brett, 1956). The
growth of fish may be directly affected by water temperature as
demonstrated on juvenile coho salmon (Hall, 1968). In short, water
temperature is a major determinant of the suitability of water
for many uses.
Research has been limited on temperature changes in small streams
from land use, although fisheries biologists have long been concerned
with the effects that deforestation can produce on water temperature.
Meehan, et^ al., (1969) studied the effects of clearcutting on the
salmon habitat of two Southeast Alaska streams. They noted a
statistically significant increase in mean monthly temperatures
after logging. The maximum increase in average monthly temperature
was about 4°F. The increase in maximum temperatures was about
9°F. during July and August.
-------
During a study of logging and southeastern trout streams, Greene
(1950) reported that maximum weekly temperatures recorded during May
on a nonforested stream were 13°F. higher than those recorded on a
nearby forested stream. He noticed also that the maximum temperature
dropped from 80 to 68°F. after the nonforested stream meandered through
400 feet of forest, and brush cover.
Levno and Rothacher (1967) reported large temperature increases in
two experimental watersheds in Oregon after logging. The shade
provided by riparian vegetation in a patchcut watershed was eliminated
by scouring after large floods in 1964. Subsequently, mean monthly
temperatures increased 7-12°F. from April to August. Average monthly
maxima increased by 4°F. after complete clearcutting in a second . .
watershed. The smaller increase in the completely clearcut watershed
was the result of shade from the logging debris that accumulated in
the channel.
Patric (1969) compared the effect of two clearcutting patterns on water
quality. Temperatures were unaffected by clearcutting the upper half
of one watershed. Clearcutting the lower half of the second watershed
increased temperatures up to 7 F.
-------
SUCTION II
TJIE STUDY
Three experimental watersheds are included in the Alsea Logging-
Aquatic Resources Study (Figure 1). These watersheds, which vary
in size from 175 to 750 acres, are located in Oregon's Coast Range
about 10 miles from the Pacific Ocean. Each stream is an important
rearing area for coho salmon (Oncorhynchus kisutch Walbam) and
coastal cutthroat trout (Salmo clarki clarki Richardson). In
its natural condition, the study area was densely forested with
Douglas-fir (Pseudotsuga menziesii [Mirb.] Franco) and red alder
(Alnus rubra Bong). The study streams were overgrown with
salmonberry (Rubus spectabilis Pursh.), vine maple (Acer circinatum
Pursh.), and other species. Although annual precipitation is about
100 inches, the summer months are generally hot and dry. Summer
streamflow regularly drops below 0.20 cubic feet per second or 0.17
csm (cfs) on Deer Creek, the largest watershed, and to 0.01 cfs
or 0.04 csm on Needle Branch, the smallest watershed.
The low summer flows described above may seem insufficient to support
salmon. The adults, however, spawn in these streams during the
high flows of the winter months. Salmon fingerlings live in these
streams during the summer. The fingerlings inhabit pools, many
of which become nearly isolated during the late summer. In the
fall, the yearling fish migrate to the sea when rains again increase
streamflow. Before logging, the number of yearling fish passing
through the fish trap to the sea ranged from 1809 to 3175 in Deer
Creek, and from 166 to 630 in Needle Branch (Hall and Lantz, 1969).
The study was designed to permit comparison of two different logging
patterns. One watershed, Needle Branch (175 acres), was fully clearcut.
A second watershed, Deer Creek (750 acres), was patchcut; 25 percent
of this area had several clearcut units. The remainder of the
watershed was unlogged. In Deer Creek, strips of vegetation were
left along the perennial streams. A third watershed, Flynn Creek
(502 acres), was left unlogged as a control. Two small subwatersheds
in Deer Creek also served as unlogged controls.
Eighteen 7-day thermographs, accurate to 0.5°F., were installed
in the three watersheds to evaluate the effect of the cutting
(Figure 1). Thermographs were placed below each proposed logging
unit and at the junction of each major tributary in Deer Creek so
that effects within the watershed could be determined. In Needle
Branch, thermographs were distributed within the clearcutting to
evaluate the spatial temperature changes occurring in a fully exposed
stream. Thermographs were installed in March, 1964. Probes were
placed in flowing water deep enough to insure complete coverage
throughout the year. The years 1964 and 1965 served as control
periods, and 1966-1969 as treatment periods.
-------
DEER CREEK
ALSEA WATERSHED STUDY
FLYNN CREEK
NEEDLE BRANCH
Figure 1. 'Hie watersiieds of the Alsca Basin logging
a^'HKitic resources stud..'.
-------
Road building was completed in 1965, but, because the roads were
built on ridges, little cha^' occurred that could be interpreted
as having any influence on water temperature. Logging began in
March, 1966, in both watersheds and was completed in August on Needle
Branch and in November on Deer Creek. In October, 1966, the stream
in the clearcut watershed was cleared of logging debris. Following
clearing, a well-distributed burn removed most logging debris and
streamside vegetation. Data reported extend through September, 1969.
-------
SECTION III
RESULTS
The data from this study have been analyzed in two ways. The large
changes occurring after clearcutting are presented graphically. The
small changes occurring after patchcutting required a statistical
analysis to ascertain the significance of these changes. The standard
statistical technique of regression could not be used because of the
nonrandom effects of climate, the lack of independence between successive
daily maxima, and the potential alteration of the variance of seasonal
temperature distributions by logging. A stationary time series was
developed to circumvent these difficulties (Beck, 1968). Time series
techniques are commonly used for analysis of weather data where similar
difficulties abound. Jenkins and Watts (1968) describe the application
of this technique to several such problems. Our time series compared
daily maximum temperatures recorded June 1 - October 1 of the
pretreatment years with the daily maxima for the same period during
each treatment year. The analysis was applied to data from the control
as well as the patchcut watershed to ascertain the effects of climate.
Diurnal Temperature Regimes
The temperature patterns recorded on the days of the annual maximum
on the clearcut watershed from 1965 to 1969 are illustrated in
Figure 2. The values for 1965 and 1966 occurred at the watershed
outlet. The values for 1967, 1968, and 1969 occurred within the
cutover unit. A thermograph was installed at 1,000 feet above the
outlet in the spring of 1967, after intensive sampling showed that
the maximum occurred at this location. This inconsistency in the
temperature pattern was the result of incomplete removal of shade from
the lower portion of the stream channel after logging and burning.
The variation in temperatures recorded at the outlet of the unlogged
watershed for the same days is also shown. Minima recorded on the
clearcut watershed are about the same as the maxima recorded on the
unlogged control. This occurs because travel time through the clearcut
watershed is greater than 24 hours during the low flow period.
Convection and nocturnal basic radiation are insufficient to cool the
water to the same minima measured on the control. The maximum diurnal
fluctuation recorded on the clearcut watershed was 28°F. during 1967.
The maximum temperature, 85°F. during 1967, represents an increase
of 28°F. over the prelogging maximum of 57°F. for 1965. The decline
of maximum temperatures after 1967 represents the rapid return of
streamside vegetation in this watershed.
The maximum temperatures recorded on the patchcut watershed were 60
and 61.5°F. for 1965 and 1967, respectively. The maximum diurnal
fluctuation during both years was 10°F.
A cumulative frequency distribution of the'diurnal fluctuations in
temperature at the outlet of all three streams from June 1 to
-------
(f)
UJ
UJ
cr
o
UJ
o
cr
or
UJ
o_
5
UJ
I-
UJ
DC
H
CO
if \ •!• P !•••«»
CONTROL
965-1969
60
50|-
i I I I I I I L_
6 12 18
HOUR (PST)
Figure 2. The maximum diurnal change in temperature recorded
on the clearcut and uncut control watersheds before (1965),
during (1968), and after (1967-1969) logging.
-------
October 1 for the years 1965-1969 is shown in Figure 3. The
temperature stability of natural, forested streams is illustrated again
in this figure. The maximum fluctuations in temperature before logging
in 1965 were 6, 8 and 10°F. on Flynn Creek, Needle Branch, and Deer
Creek. These maximum fluctuations were not exceeded on Flynn Creek
or at any station within Deer Creek during or after logging. On Needle
Branch, the maximum fluctuation of 8°F. was exceeded 28 percent of
the time in 1966 and 82 percent of the time in 1967, the year
immediately following burning and stream clearance. This percentage
dropped to 46 percent in 1968 and 36 percent in 1969, again reflecting
the regrowth of streamside vegetation.
The outlet stations are representative of the changes that occurred
within each watershed. Temperature fluctuations generally decreased
with distance upstream in the patchcut watershed. The most remote
station (15) in the clearcut performed similarly to the outlet of
Deer Creek.
Monthly Maximum Temperatures
Clearcut. Maximum daily stream temperatures averaged by month for
one year before and four years after logging are shown for the outlets
of the clearcut and unlogged watersheds in Figure 4. Except for the
most remote station (15), the changes recorded at the outlet station
are representative of those occurring at the other stations within
the clearcut watershed. The highest temperatures again are shown to
occur during 1967, the year after stream clearance and slash burning.
The mean monthly maximum for July increased from 57°F. in 1965 to
71°F. in 1967. The trend toward the prelogging condition is again
shown in this figure.
Patchcut. The frequency diagram illustrates the nearly constant pattern
of daily temperature fluctuation recorded on the patchcut watershed
throughout the study (Figure 3). A time series was required to
determine whether the small increases observed initially were the
result of logging or climatic differences between years. The results
of the time series are presented in Table 1. Significant changes in
the summer maximum temperatures were observed at the outlets on the
control and patchcut watershed one year after logging. The larger
changes observed on the control indicate that climatic factors, and
not the patchcutting, were responsible for this increase. During
1966 and 1968, summer maxima in the stream of the patchcut were nearly
the same as those observed before logging. The ten internal stations
exhibited smaller changes in temperature than the outlet station.
These data show that patchcutting which leaves streamside strips of
brush and trees, did not alter temperature patterns of the adjacent
stream.
Other inferences about the temperature patterns may be drawn from these
data. Clearly, the patterns in summer temperatures of forested streams
are relatively constant from year to year. The small differences
-------
AFTER LOGGING
1969
AFTER LOGGING
967
DURING LOGGING
966
80
60
40
20
BEFORE LOGGING
1965
D UNCUT CONTROL
o PATCH CUT
A CLEAR CUT
10
20
DIURNAL TEMPERATURE CHANGE, DEGREES F
Figure 3. A frequency distribution of diurnal
temperature changes on the clearcut, patchcut,
and uncut control watersheds before (1965), during
(1966), and after (1967-1969) logging.
-------
CONTROL
1965- 1969
J J
Figure 4. Mean monthly maximum temperatures for the clearcut and
uncut control watersheds before (1965), during [1965], and after
(1967-1969) logging.
10
-------
between years listed in Table 1 and the statistical significance
of a 2-degree change illustrate little variability in average
maximum temperature for a summer.
Table 1. A time series analysis of the difference between daily
maximum temperatures of streams before and after patchcutting, for
the period of June 1 to October 1.
Temperature Change , °F.
Watershed
Flynn Creek
(unlogged)
Deer Creek
(patchcut)
1966
0.8
0.4
1967
2.3*
1.9*
1968
0.3
-0.4
"Significant at the 5% level of probability.
11
-------
SECTION IV
DISCUSSION
A detailed description of the hydrologic and atmospheric factors
affecting water temperature has been given earlier (Brown, 1969). The
most important environmental factor governing temperature change is
solar radiation received at the stream surface.
Temperature differences between watersheds and all of the temperature
anomalies within the clearcut watershed can be explained in terms of
shade differences. The patch cuts on Deer Creek did not produce any
significant changes in temperature in the main stream. One-hundred-foot
strips of timber were left beside each perennial stream; the amount of
shade on the stream surface was essentially unchanged. On Needle Branch,
little shade remained after the clearcutting and burning were completed.
As a result, large changes in annual and daily patterns of temperature
were observed.
The principal cause of high temperature of water following logging
is exposure of the stream surface to direct insolation, not the increased
soil temperature on the clearcut slopes as has been suggested by
Eschner and Larmoyeux (1963]. Satisfactory predictions of temperature
have been made on Deer Creek and Needle Branch with net radiation as
the primary parameter. On these same streams, we found that the
maximum rate of net thermal radiation added to the unshaded stream
in the clearcut was more than ten times that added to the shaded stream
in the patchcut watershed. These differences are reflected in the
postlogging temperature patterns of each stream and help explain the
temperature stability of most forested streams.
For a given level of solar radiation or heat, stream temperature is
inversely proportional to volume. As a result, the temperature
patterns of small, shallow streams typical of headwaters regions may
be increased significantly by any changes in the solar radiation.
Discharge of the stream in the clearcut watershed regularly drops to
0.01 cfs during the hot summer months. Thus, the large changes in
temperature recorded after the shade was removed from this stream were
to be expected. The flow regime of this small stream is typical of
many western Oregon streams that must support salmon and trout during
the period of low flow.
Can the results of this study be classified as typical of western
Oregon conditions or were they merely caused by unusually hot summers?
The maximum temperature of 85°F. recorded for the clearcut and burned
watershed is undoubtedly close to the maximum temperature that could
occur for this size of clearcut and stream. Because of its small size,
the stream responded to each clear day with high temperatures,
regardless of the previous day's weather. Even if only one clear,
hot day had occurred, this temperature would have been observed. The
12
-------
records of mean monthly temperatures and the frequencies of diurnal
change present data for longer periods. The number of days of sunshine,
overcast, fog, and rain influence such results. Long-term records
of sunshine are not available for the study area. Records at nearby
stations, however, indicate that the period of study was not abnormal.
Duration of temperature effects following clearcutting is the subject
of continued observation in these experimental streams. The
amelioration of temperature is related to the development of shade.
As stream bank vegetation becomes re-established, temperatures drop
accordingly. In many of the watersheds of the Pacific Northwest,
regrowth occurs very rapidly on moist sites. On the basis of our
data, it seems that summer maxima may approach prelogging levels
within six years after logging has completely exposed the stream
if vigorous invasion of such moist-site species as alder, salmonberry,
elderberry, and vine maple occurs. The decline of high temperatures
noted in Figures 2, 3, and 4 after the 1967 maximum on Needle Branch
illustrate this recovery. The stream's temperature patterns are
well on the way to returning to the prelogging condition.
Our research reported here and a companion study (Brown, 1969) have
illustrated the effect of two patterns of logging on the temperature
of small coastal streams and the amelioration of the effect with
time. These results, however, have raised several questions about
temperature control in the management of fishery, forest, and water
resources. Foremost among them is the effect of high temperatures,
such as those recorded during 1967 in the clearcut watershed, on
fish. Earlier work (Brett, 1952) indicated that, at 27.5°C. (81.5°F.),
the time required to induce 50 percent mortality in a population
of coho salmon fry is only 70 minutes. But no unusual mortality
in coho was observed during this study even though streajii temperatures
were often above this limit for four to six hours.
The study illustrated the benefit of strips of vegetation alongside
small streams for temperature control. The width, density, species
composition, and costs incurred when planning streamside strips are
only a few of the questions posed by forest managers.
Finally, this study should encourage water resource agencies to engage
in further studies of the aquatic habitat in small streams and its
relation to land use. Although the temperature changes recorded
are defined as thermal pollution in Oregon's current water quality
standards, the effect on the coho fishery would suggest that a more
precise definition is required. This, of course, will require a
better understanding of the response of the aquatic system to
temperatures in the lethal and sublethal ranges.
13
-------
SECTION V
REFERENCES
Beck, L. C. Basic concepts and theory of stationary time series and
their application to the problem of determining effects of logging
on water temperature. M.S. report in statistics, Oregon State
University, Corvallis. 1968.
Brett, J. R. Temperature tolerance in young Pacific salmon, genus
Oncorhynchus. J_. Fish Res. Bd_. Canada 9, 265-323, 1952.
Brett, J. R. Some principles in the thermal requirements of fishes,
Quart. Rev. Biol. 31, 75-81, 1956.
Brown, G. W. Predicting temperatures of small streams. Water
Resources Research, 5, 68-75, 1969.
Brown, G. W., and J. T. Krygier. Changing water temperatures in small
mountain streams J_. Soil and Water Conserv., 22, 242-244, 1967.
Eschner, A. R., and J. Larmoyeux. Logging and trout: four experimental
forest practices and their effect on water quality. Prog. Fish Cult,
25, 59-67, 1963.
Greene, G. E. Land use and trout streams. J_. Soil and Water
Conserv., 5, 125-126, 1950. ~
Hall, James. D. Biological effects of thermal pollution. Unpublished
Report, Oregon State University, Corvallis, 1968.
Hall, James D., and Richard L. Lantz. Effects of logging on the
habitat of Coho salmon and cutthroat trout in coastal streams.
In: A_ symposium on_ salmon and trout in_ streams. T. G. Northcote,
Ed. University of British Columbia, pp. 353-375,, 1969.
Jenkins, G. M., and D. G. Watts. Spectral analysis and its application.
Holden-Day, San Francisco. 1968.
Levno, A., and J. Rothacher. Increases in maximum stream temperatures
after logging in old-growth Douglas-fir watersheds. Pac. NW Forest
and Range Exp. Sta.,Res. Note PNW-65. 12 pp. 1967.
Oregon Sanitary Authority. Standards of quality for public waters of
Oregon and disposal therein of sewage and industrial wastes,
Administrative Order SA 26, June 1, 1967.
Patric, James H. Changes in streamflow, duration of flow, and water
quality on two partially clearcut watersheds in West Virginia.
Trans. American Geophysical Union, 50(4), 144 (Abstract only), 1969.
14
-------
CHAPTER II
PREDICTING THE EFFECT OF CLEARCUTTING ON
STREAM TEMPERATURE
SECTION I
INTRODUCTION
Effects of logging on water yield, peak discharge, and sediment
have long been measured and predicted by forest hydrologists.
Only recently have researchers and land managers become concerned
about the effect of logging on other characteristics of water
quality. Considerable attention is now being focused upon water
temperature, because of its potential effect on fish populations.
In an earlier paper (Brown § Krygier, 1967), the general effect
of removing all of the shade from forested streams has been described.
How may this information be extended to estimate the effects of
clearcuttings of various sizes on other streams? The purpose of
this paper is to present a practical technique for estimating the
maximum change in temperature that would result after the shade
has been removed from forest streams by clearcutting.
15
-------
TEMPERATURE PREDICTION
The adverse effect of elevated temperature of water on fish and other
aquatic organisms led to studies designed to predict changes in
temperature. The first efforts were made by engineers concerned with
regulation of temperature below large dams (Burt, 1958; Delay § Seaders,
1966) . The techniques developed for prediction of temperature on large
rivers were subsequently modified by Brown (1969) for use on small
streams.
My previous studies (Brown, 1969) showed that accurate, hourly
predictions of temperature could be made on small streams using energy
budget techniques and on-site meteorological measurements. I found that
air temperature and the cooling effect of evaporation were much less
important than solar radiation in controlling temperature on small,
unshaded streams in the Oregon Coast Range. Solar radiation accounted
for over 95 percent of the heat input during the midday period at
mid-summer.
Predicting the hourly change in temperature is often unnecessary,
however, if the maximum change produced by clearcutting can be
assessed. Estimates of the maximum change would be sufficient to permit
a prediction of any major change in stream ecology. Hourly prediction
of temperature requires experience in micrometeorology and considerable
expenditure for equipment.
The simplified temperature prediction technique has been published
(Brown, 1970). It also appears below.
The potential effect of a clearcutting on maximum temperatures of a
stream may be estimated without extensive meteorological expertise or
equipment. Uniform exposure of the stream surface to direct sunlight
can often be assumed. If the stream surface is not dappled with spots
of sunlight and shade, elaborate systems for sampling solar radiation
are not required.
The dominance of direct sunlight also simplifies the prediction
technique. Because the maximum change in temperature will no doubt
occur during the midday hours on a clear day, predicted rather than
measured solar radiation can be used to estimate the heat input for an
exposed stream. Topographic shading is also likely to be unimportant
during this midday period.
Once this heat input is assessed, the predicted change in temperature
(AT) can be computed as the ratio of the heat added to the volume of
water heated. The heat added to the system is computed as the product
of the rate of heat (H) absorbed by the stream, in British thermal
units per square foot per minute, and the surface area (A) in square
feet of the stream exposed by the clearcutting. Water in isolated
pools should not be measured. Neither should moving water in the
16
-------
backwater sections of large pools in the main stream where velocity
is zero or directed upstream. The maximum change in temperature
will occur when the greatest amount of heat warms the smallest
volume of water. This volume is represented by the minimum discharge
rate in summer (D) in cubic feet per second. Rates can be used
in this ratio because the times during which the rates apply are
the same for numerator and denominator. In equation form, this
ratio becomes:
AT = A * H x 0.000267 [1]
The constant, 0.000267, converts discharge in cubic feet per second
to pounds of water per minute. The temperature change is then
in units of Btu per Ib of water and thus is equivalent to temperature
in degrees Fahrenheit.
The stream surface to be exposed by a proposed clearcutting and
the lowest discharge in summer can be measured the year before
logging. How, then, may the heat input be estimated?
The amount of heat received from the sun on a-'clear day depends
upon solar angle. Solar angle, in turn, depends upon season, time
of day, and latitude. The symmetry of the solar path, however, permits
construction of a series of curves that provide hourly values for
solar angle on the basis on the sun's angle at solar noon. This
information may be obtained with a solar ephemeris (List, 1966,
pp. 500-502).
Raphael (1962) converted Moon's (1940) standard curves to determine
incoming solar radiation.in Btu/ft2-hr as a function of solar angle.
Raphael's values have been multiplied by the appropriate solar angles
to provide, in Table 1, an estimation of the total (direct and diffuse)
incoming shortwave radiation throughout the day as a function of
midday solar angles.
Part of this incoming radiation is reflected. The amount reflected
is3 again, determined by the sun's angle. Dirmhirn (1964) has presented
reflection values from water surfaces for several solar angles. These
data are included in Table 2.
The incoming energy in Table 1 is reduced by considering the appropriate
solar angle and the reflectivity given in Table 2. The product of
this computation is the approximate hourly net solar radiation for
given sun angles at solar noon (Figure 1).
lieat may be added to the stream by incoming long-wave radiation.
This input, however, is about equal to the back radiation from the
water. Subsequent computations recognize these sources of heat gain
and loss, but because their net effect is approximately zero, they are
not considered directly.
17
-------
Table 1. Incoming shortwave and diffuse radiation
on a horizontal surface, Btu/ft2-min
Hour
A.M. P.M.
12
11 1
10 2
9 3
8 4
7 5
6 6
Midday solar angle
30°
2.
2.
1.
1.
0.
0.
0.
42
37
86
31
52
04
00
140°
3.
3.
2.
2.
1.
0.
0.
27
10
77
05
12
29
00
4
3
3
2
1
0
0
50°
.02
.87
.42
.77
.77
.58
.13
60°
4.63
4.51
4.02
3.27
2.33
1.31
0.45
70°
5.10
4.91
4.45
3.65
2.68
1.49
0.51
80°
5.36
5.18
4.72
3.95
3.02
1.77
0.58
Table 2. Percent reflection from water surfaces, after Dirmhirn(1964).
Radiation
Sun angle , degrees
5 10 15 20 25 30 40
50 60 70
Percent reflection
Solar 67 40 26 17 12 12 8.5 5 3 2.7
Diffuse sky 17 15.5 14.5 13.2 12.5 11.2 9.3 8 7.4 7.0
Total 84 55.5 40.5 30.2 24.5 23.2 17.8 13 10.4 9.7
18
-------
'Die rate of incoming energy is constantly changing (Figure 1].
To determine the appropriate rate of energy for the temperature-
change equation, we need to know how long the stream will be exposed
to the direct rays of the sun. This necessitates measuring the time
required for the stream to flow through the proposed clearcutting
during the low-flow period—that is, the travel time. Travel time
can easily be estimated by observing the time required for a slug
of dye to move through the area.
To predict the maximum change in water temperature as the stream
flows through the clearcutting, an estimate of the maximum input
of heat must be made. This estimate is made by averaging the net
radiation about the noon maximum for a given travel time (Figure 2).
A 2-hour travel time, for example, would require averaging the incoming
radiation from 11:00 a.m. to 1:00 p.m.
A forester may now estimate the maximum change in temperature that
a clearcutting would produce. The procedure may be summarized as
follows:
1. Mark the upstream and downstream boundaries of the proposed
clearcutting.
2. Determine the lowest discharge during the summer and the dates
during which it occurs.
3. Determine the surface area of the stream in the proposed
clearcutting during the low-flow period.
4. With dye, determine the travel time of the stream through the
proposed clearcutting during the low-flow season.
5. From a solar ephemeris, determine the highest sun angle at
solar noon for the period of low flow.
6. Enter figure 2 with the appropriate travel time. Move up to
the correct curve for the sun angle at solar noon and read the
average radiation in Btu/ft2-min.
7. Compute the predicted maximum change in temperature using
•Equation 1.
19
-------
MIDDAY SOLAR ANGLE
80
TIME OF DAY
Figure 1. Hourly values for net solar radiation above water
surfaces in clear days between latitudes 30N and SON for several
solar naths.
20
-------
CVJi
00
•»
z
o
H
5
or
or
< 2
o
Ld
CD I
cr
o
CO
CD
1 T
MIDDAY
SOLAR
ANGLE
80
30'
0246
TRAVEL TIME, HOURS
Figure 2. Average net solar radiation absorbed by streams
between latitudes 30N and SON on clear days during several
periods of exposure to different solar paths.
L
8
21
-------
SECTION III
DISCUSSION
On a single stream without tributaries, this method will provide estimates
of changes in temperature within about 3°F of the true value. The
technique was tested on two streams in Oregon bounded by clearcuttings.
Temperature changes of 16°F were predicted within 1°F.
This simplified method for predicting stream temperature does have
some limitations, however. First, it is assumed that no surface
tributaries flow into the section of stream exposed by clearcutting.
If such flow occurs, we must estimate the temperature of the inflow
and adjust the predicted temperature of the main stream. This
adjustment is made with a mixing ratio of the predicted temperatures
of the tributary stream (T ) and the main stream (T ), weighted by
their respective discharges (D and D ). This may oe written as:
t m
CD )(T ) + (D )(T )
Adjusted temperature = —-— —— [2]
*• t' ^ nr
The second limitation is that low-flow discharge is assumed constant.
Changes in summer discharge could occur because of climatic change,
or, if the area logged is large enough, the clearcutting might increase
base flow. In some areas, base flow has been increased by 50-75
percent after the entire watersheds have been clearcut. An increase
in discharge may be partially compensated by a corresponding increase
in surface area that depends upon channel configuration, which would
reduce the error in Equation 1. The principal error comes from the
low temperature of this unknown volume of groundwater. In areas where
hydrologic data indicate the extent of this increase, the effect of
groundwater addition may be estimated by treating it as a tributary
(Equation 2). Groundwater temperatures may be obtained by making
midsummer measurements in auger holes placed in the saturated.zone
near the stream.
A final limitation is imposed by the assumption of complete exposure
of the stream surface to direct sunlight. If considerable overhanging
vegetation remains after clearcutting, or if topographic shading occurs,
the predicted change in temperature will be too large.
In practice, these limitations are not likely to restrict the application
of this method, even though their presence makes the predicted change
in temperature too high. Watersheds are seldom clearcut in their
entirety today. The most common occurrence probably is exposure of
2,000-3,000 feet of stream, which may eliminate much of the error
induced by increasing groundwater discharge. Noting the duration
of any topographic shading will permit adjustment of the exposure
22
-------
interval used In Figure 2. Finally, any cooling effect of residual
shade after logging may be taken as a "safety factor" in making
the temperature-change estimate.
The "quality of our environment" is a popular phrase often used
by politicians, conservationists, and others interested in pollution.
Responsibility for the quality of the environment extends to everyone.
Flood prevention and erosion control have often been listed as a
primary justification for our national forests, but forests and
foresters have recently been assigned broader roles in maintaining
environmental quality. Floods and sediment are no longer the only
measure of a forester's effectiveness in controlling water quality.
For the first time, national legal criteria have established several
characteristics to define water quality.
The Federal Water Quality Act of 1965 required that each state prepare
a set of water-quality standards and a plan for their enforcement.
The standards include a large number of water-quality characteristics
over which the forester has some influence. They include not only
sediment, but also dissolved oxygen and water temperature.
The implications of such standards are clear. Foresters have been
given legal responsibility for maintaining the quality of the aquatic
environment. Furthermore, a yardstick has been adopted that will
judge the effectiveness of their efforts.
Temperature prediction has increased meaning under these circumstances
This technique now becomes a tool for the solution of a practical
problem. It permits the forester to do an improved job of managing
all the resources of the watershed. Temperature prediction will
also help him avoid violation of water-quality criteria established
in his state. In short, this predictive model will help the forester
to meet his growing responsibility in maintaining the quality of
our environment.
23
-------
SECTION IV
REFERENCES
Brown, George W. Predicting temperatures of small streams. Water
Resources Research. 5(l):68-75. 1969.
. and James T. Krygier. Changing water temperatures in
small mountain streams. J. Soil and Water Cons. 22(6):242:244. 1967.
Burt, W. V. Heat budget terms for Middle Snake River reservoirs. In:
Water Temperature Studies on the Snake River.U.S. Fish and WildTife
Service Technical Report No. 6. 23 pp. 1958.
Delay, W. H. and John Seaders. Predicting temperatures in rivers and
reservoirs. Proc., Amer. Soc. Civil Eng., Jour. San. Eng. Div.
92:115-134. 19661.
Dirmhim, Inge. Das Strahlungsfeld im Labensraum. Akad. Verlags.
Frankfurt/Main. 426 pp. 1964.
Greene, G. E. Land use and trout streams. Jour. Soil and Water Conserv.
5:125-126. 1950^
List, R. J. Smithsonian Meteor. Tables. Smithsonian Inst., Washington,
D.C. 1966- ~~
Moon, P. Proposed standard solar radiation curves for engineering use.
Jour. Franklin Inst. 230:583-617. 1940.
Raphael, J. M. Prediction of temperatures in rivers and reservoirs.
Proc., Amer. Soc. of Civil Eng., Jour. Power Div. 88:157-181. 1962.
Titcomb, J. W. Forests in relation to fresh water fishes. Trans. Amer.
Fisheries Soc. 56:122-129.1926.
24
-------
CHAPTER III
HEAT LOSS FROM A THERMALLY LOADED STREAM
iinergy balance studies proved invaluable for understanding the
processes governing heat exchange on small streams. This technique
was also utilized to ascertain the way heated streams lost this
heat as they flowed into the forest from a clearcut.
Our earlier work (Brown, 1969) illustrated the importance of the
solar flux in temperature prediction. Measuring this flux beneath
a highly varied forest canopy required development of some new
techniques. A system was developed to spatially integrate solar
radiation penetrating the canopy in both time and space.
The system developed for spatially integrating solar radiation is
shown in Figure 1. It consists of an aluminum track, track supports,
a trolley for moving the sensor along the track, cabling for moving
the trolley and transmitting the sensor signal, and a recorder.
A strip chart recorder with integrator measures sensor output and
integrates it with time and thus space. Aspirated psychrometers,
thermocouples, and anemometers were also used to measure components
of the energy balance.
Data collection was accomplished with a 25-channel digital data
acquisition system (Figure 2). This system permitted us to sample
the environment quickly and accurately. All 25 channels were measured
within a 30-second period at intervals of one to ten minutes. Sensor
output could be measured to within one microvolt. A teletype provided
both typewritten and punch-tape records of the measurements. Punch-
tape records were later fed directly into a computer for interpretation
and analysis.
Data were collected during the summer of 1969. Energy balance
measurements were made in the Umpqua National Forest. The study
site was located downstream from a clearcut in which water temperature
reached 80-85°F. during midday. Additional measurements were made
along a stream protected by a buffer of trees separating the stream
from a large clearcut.
The result of the energy balance measurements below a clearcut during
a typical day is shown in Figure 3. Positive values indicate energy
additions to the stream. Negative values represent energy losses.
The curve labeled "storage" is the algebraic sum of net radiation,
evaporation, convection and conduction.
What this figure tells us is that even in the shade, relatively
small amounts of heat are added to the stream throughout the day
since the storage function is always positive. Further, we could
anticipate that the stream will not cool as it flows into the shade.
but continue to increase in temperature at a very slow rate.
25
-------
Figure 1. A system for spatially integrating solar radiation-
26
-------
Figure 2. The digital data acquisition system,
27
-------
1.4
1.2
CM
2
O
N»
-J
<
u
0.8
GLOBAL
RADIATION
°6
U.
0.4
cr
LU
Z 0.2
UJ
0
-0.2
i r
T r
T r
NET RADIATION
CONDUCTION
EVAPORATION
J L
10
12 14
HOUR (POT)
16
Figure 3. An energy balance on a shaded reach of Cedar Creek
during a clear day in July, 1969.
28
-------
This conclusion is strengthened by thermograph measurements obtained
within shaded reaches and studied elsewhere in the Cascades.
The second study evaluated the net radiation absorbed on Little
Rock Creek. At this site, the water flowed through a reach shaded
by a strip of trees left during a clearcut operation adjacent
to the stream. Again this measure of net radiation was made with
the device that gave a time and space integration.
The integrated net radiation measured on Little Rock Creek is
shown in Figure 4. The net radiation on Cedar Creek observed the
previous day and the measured global radiation are also shown. The
curve labeled "expected net radiation" is the net radiation that
would have been absorbed by both streams had the trees not been
present.
The curves show us that the streamside strip on Little Rock Creek
is less efficient in reducing net radiation than the uncut block
of timber through which Cedar Creek flowed. Considering the potential
net radiation, however, its efficiency is remarkable. Thermograph
measurements made by Dallas Hughes along this same stretch of stream
confirm this efficiency. As he illustrates in another section of
this report, the temperature increase through this section was nil.
In very practical terms, then, these measurements have confirmed
an earlier hypothesis: as long as the stream surface can be shaded,
temperature can be controlled. Further, extremely wide buffer strips
are not necessarily required to do the job. The strip on Little
Rock Creek was only two trees deep, about 50-75 feet. The volume
contained in this strip is estimated at 75 M bd feet or roughly 0.1
percent of the sale volume, from the 55-acre unit.
The hypothesis that shaded zones below clearcuts will cool heated
water is questionable. The energy balance and thermograph measurements
lead to the very practical conclusion that if cool water is desirable,
it should not be heated in the first place.
29
-------
CM
<
(J
X
D
_J
CL
LJ
z
1.4
1.2
0.8
0.6
0.4
0.2
0
RADIATION
GLOBAL
\
\
\
\
\ -I
NET
LITTLE ROCK CREEK
/
NET
CEDAR CREE
10
12 14
HOUR(PDT)
16
Figuj'e A. Net alJ-wave radiation measured in a shaded
reach of Cedar Creek and within a shaded buffer strip on
Little Rock Creek during clear days in July, 1969.
30
-------
CHAPTER IV
HEAT FLOW IN STREAM BEDS
The role of the stream bed in heat exchange was attacked through
analysis of thermal conductivities of various bed materials. Cores
were extracted from selected streams for laboratory analysis of thermal
conductivity.
These cores were frozen using a heat-pump technique developed by
the Canada Department of Fisheries and utilized successfully by Ringler
(1969). Acetone is cooled using dry ice, causing circulation within
a pipe penetrating the gravel. The temperature of the acetone is
about -79°C. and freezes a core around the pipe. This core can then
be extracted and taken to the laboratory for analysis. Importantly,
the position of the gravel particles and the composition of the mass
is fixed. The device is illustrated in Figure 1.
Thermal conductivity measurements were begun using a device much
simpler than that proposed in the original research proposal. A
system devised by Wang and Knudsen (1958) and successfully used in
predicting thermal conductivity in two phase systems was substituted.
It was much simpler in design and much less expensive.
The thermal conductivity device is shown in Figure 2. A heat source
is created in the top chamber and a heat sink created in the bottom
chamber. A substance of known thermal conductivity (usually water)
is introduced into the chamber immediately below the heat source.
Heat flow through this chamber is calculated by:
Q - K A £ [1]
where q is the heat flow, K is the thermal conductivity of water,
A is the cross sectional area of the chamber, dT is the temperature
gradient across the chamber, and dL is the chamber height.
Under conditions of steady state, q out of the chamber containing
water will equal q into the chamber containing the material with
unknown K. Thus, this unknown K can be determined with the equation
above after measuring dT, dL, and A.
Laboratory analysis of cores together with evaluation of thermal
gradients measured in_ situ was scheduled for Year 3. Termination
of this grant during Year 2 precluded further work.
31
-------
Surface Water
i-i^urc 1. Cutaway view of frozen core sampler in operation.
Arrows indicate direction of acetone circulation (After Rinjiler,
1969).
-------
! Lgure 2. The thermal conductivity chamber
-------
REFERENCES
Ringler, Neil II. Effects of logging on the spawning bed environment
in two Oregon coastal streams. M.S. thesis, Oregon State University.
1969.
Wang, R. 11. and James G. Knudsen. Thermal conductivity of liquid-
liquid emulsions. Industrial and Engineering Chemistry 50(11):
1667-1670, 1958.
34
-------
CHAPTER V
CLEARCUT LOGGING AND SEDIMENT PRODUCTION
IN THE OREGON COAST RANGE
SECTION I
INTRODUCTION
In the Pacific Northwest, commercial forests cover much of the
headwaters landscape. Timber harvest in these forests is the
predominant land use activity. In Oregon alone,, between 500,000 and
700,000 acres are logged annually. Logging is preceded by road
construction. Clearcutting is the principal logging technique,
followed by slash burning. The change in the appearance of the
landscape is dramatic and abrupt. The crucial question from a water
quality standpoint is: how does clearcut logging affect erosion and
sedimentation in headwater streams? In Oregon, this question is crucial
since salmon, steelhead and trout utilize these small streams for
spawning and rearing sites.
The purpose of this paper is to describe the effect of road building,
clearcut logging, and slash burning on suspended sediment production
from three forested watersheds in Oregon's Coast Range where
precipitation averages 100 inches annually and topography is steep.
The effects of sediment on fish have been summarized by Cordone and
Kelley (1961). Excessive concentrations of suspended sediment
(20,000 ppm) can cause gill injury to fish or alteration of behavior
patterns. The most significant sediments effects at more typical
concentrations occur because of alteration and destruction of bottom
organisms, and from indii-ect influences of sediment deposition on
intragravel flow and aeration. Mineral and organic sediments in water
or deposited in spawning gravels may cause mortality, delayed
development, or poorer condition in salmon and trout (Brannon, 1965;
Kramer, 1965; Koski, 1966; Shelton, 1966; Servizi, et al_. , 1969).
Gravel size has been related to the interchange of dTssolved oxygen
(Cooper, 1965; Oregon Game Comrn. , 1966; Ringler, 1970). Cooper found
that deposition will occur in spawning gravels at moderate concentrations
even though velocities are too high to permit deposition on the surface.
Public concern for pollution has led to the establishment of water
quality standards. Oregon has now gone beyond interstate standards
(Water Quality Act of 1965), and has applied water quality standards
to sub-basins with direct intent to control quality of upstream waters
(Oregon Dept. of Environmental Quality, 1969). These standards have
been set without a full understanding of sediment concentrations or
sediment production rates from mountain streams under both natural
conditions and conditions influenced by logging operations.
35
-------
Forest hydrologists have often related sediment production to timber
harvest operations in headwater areas. Road construction preceding
logging is often the most serious cause of erosion. In the volcanic
formations of the Oregon Cascades, sediment yields from three small stee;
watersheds tributary to the McKenzie River seldom exceeded 200 parts per
million (ppm) prior to treatment (Fredriksen, 1965, 1970). Immediately
after roads were constructed across one watershed, a peak sediment
concentration of 1,780 ppm was observed, 250 times that recorded in a
control. This initial effect subsided after two months, but
concentrations remained two to three times the level predicted from the
control. These results did not include samples from landslide events.
In 1961 and 1964, landslides from the roads produced average
concentrations about 34 times greater than expected from the
pre-treatment relationship. Mean annual sediment yield including bed
load was 8,000 tons/mi2-yr in a nine-year period, 109 times the loss
from an undisturbed control watershed.
At Castle Creek, in California, where the primary influence was roads,
average sediment concentrations and loads from a 4-mi2 watershed
increased five-fold in the first year, from 64 ppm to 303 ppm, or from
935 tons/mi2-yr to 4,600 tons/ini2-yr. Concentrations and yield
declined to twice the normal rate in the second year (Rice and Wallis,
1962; Anderson and Wallis, 1963). In the Idaho granitic batholith,
roads associated with jammer logging (a high-density road system),
produced highly variable sediment yields from three logged watersheds
of 12,400, 8,900, and 89 tons/mi2 in one season (Copeland, 1965).
Neighboring drainages without roads in this area of highly erodible
granitic soil produced no sediment; high yields in watersheds with
roads were attributed to inadequate cross drains.
The effects of the logging operation are often difficult to separate.
In many erosion studies, the sediment contribution caused by road
construction, skid trails, and logging are measured together. One such
study at Fernow Experimental Forest in West Virginia reported an
average turbidity of 490 ppm (Jackson Turbidity Units) during tractor
logging. One year later average turbidity dropped to 38 ppm; two
years later it was only 1 ppm. Another study at this forest
illustrated the importance of planning logging operations. On a well
planned logging operation, the maximum turbidity was only 25 ppm. An
adjacent watershed was logged without any plan or direction and
maximum turbitities of 56,000 ppm were recorded (Reinhardt and
Eschner, 1962).
Sediment sampling before, during, and after clearcut logging was
conducted in Maybeso and Harris River Valleys in southeastern Alaska
(Meehan et_ aj_. , 1969). No significant change could be detected in the
concentrations, possibly due to inadequate sampling. Sheridan and
McNeil (1968) found small increases in the percentage of fine
sediments deposited in the stream gravels after logging in this same
36
-------
area. 'Hie probable source of this sediment was debris avalanches
which were common in clearcut areas.
Fredriksen (1970) reported that on a watershed that was clearcut over
a period of three years with a sky-line system, and thus without
roads, concentrations were only modestly affected during logging.
Mean concentration during storm periods remained below 10 ppm until
slides, triggered by the record storms of 1965, brought about 800 tons
of soil and rock material into the channel. Most of this material
remained trapped by logging debris.
Controlled slash burning is a common practice following clearcutting
in the Pacific Northwest. Very little information exists about the
effect of controlled burning on sediment production from forests.
Burning following logging with the sky-line system described above
was also reported by Fredriksen (1970). Resulting sediment
concentrations during two subsequent years ranged from 100-150 ppm and
were 67 and 28 times those recorded on an undisturbedwatershed during
the same period. Fredriksen noted that sediment had been trapped in
the logging debris and was released only after burning.
37
-------
SECTION II
THE STUDY
In 1958, Oregon State University began a cooperative study of the
effects of logging on water quality and fishery resources of three
small watersheds in Oregon's Alsea Basin. These watersheds are located
about 8 miles south of Toledo and about 10 miles from the Pacific Ocean
[Figure 1). The watersheds were forested with Douglas-fir and alder.
Mean elevations are between 740 and 1,000 feet. Mean slopes range from
25 to 50 percent. The maritime climate produces a mean annual
precipitation of about 100 inches. Summers are dry, however, and most
of the rainfall occurs between November and April. The soils are
derived from the Tyee sandstone formation. Over 80 percent of the soils
are from either the Slickrock or Bohannon series. The Slickrock soils
are derived from sandstone colluvium and are fairly deep. The
Bohannon series, a shallow, stony soil, is derived from the sandstone
residuum. Both series are moderately stable.
The sediment yield characteristics of the watersheds were monitored
for seven years prior to treatment, from 1958 to 1965. Suspended
sediment was measured at the mouth of each watershed and at six small
stream gages in Deer Creek. The gages at the mouth of each watershed
are operated by the U.S. Geological Survey and integrate the effects of
land use on each watershed. The small gages in Deer Creek are operated
by Oregon State University's School of Forestry. These gages were
installed in 1963 to evaluate the effect of each cutting unit on the
tributaries within the Deer Creek watershed.
Logging roads were constructed into Deer Creek and Needle Branch betweei
March and August, 1965. Flynn Creek, a 500-acre watershed, served as a
control and remained in its natural condition throughout the study.
Sediment samples were collected during the winter of 1965-66 to
evaluate the effect of roadbuilding. Logging began in March, 1966, and
ended in November of that year. Needle Branch, a 175-acre watershed,
was fully clearcut; Deer Creek, a 750-acre watershed, was 25 percent
clearcut with three small units (Figure 1). The effects of these three
units were measured at four weirs. The 138-acre watershed above weir
II was 30 percent clearcut. The 100-acre watershed above weir III was
65 percent clearcut. Weir IV measures the runoff from a 39-acre
watershed that was 90 percent clearcut. Weir VI measures the combined
effect of the upper watersheds; its 572 acres were 25 percent
clearcut.
The slash on Needle Branch was burned in October, 1966. The upper
units of Deer Creek remained unburned; the lower unit was lightly
burned in October, 1966. Sediment sampling on Needle Branch the
following winter reflected the combined effect of roadbuilding, logging
and slash burning. Sediment measurements at the six stations within
Deer Creek permitted evaluation of the first two effects; the U.S.
38
-------
DEER CREEK
TOLEDO
7
MILES
ALSEA WATERSHED STUDY
0
I
SCALE, MILES
LEGEND
FISH TRAP
ROAD
LABORATORY
STREAM
STREAM GAGE
LOGGING UNIT
FLYNN CREEK
NEEDLE BRANC
UNEWPORT
TOLEDO
STUDY AREA
Figure 1. The study watersheds and sediment sampling stations
in the Alsea Watershed Study.
39
-------
Geological Survey samples included the effect of burning the lower
unit. Sediment sampling continued through the 1967-68 and 1968-69
storm seasons. Sediment yields will be monitored for several more
years.
Routine suspended sediment concentrations in parts per million were
obtained daily at the U.S. Geological Survey weirs by Oregon Game
Commission personnel. During storms, samples were taken at more
frequent intervals to ascertain sediment loads as stream levels
changed. At the small weirs within Deer Creek samples were taken
only during storm periods.
40
-------
SECTION III
CHANGES IN ANNUAL SEDIMENT LOAD
Two analyses were run on the suspended sediment data; analysis of total
annual load and analysis of suspended sediment concentration. Changes
in annual sediment load for the three watersheds were estimated using
an averaging technique designed to reduce the variation in sediment
associated with a changing streamflow regime. It has been well
documented in streamflow studies conducted concurrently with these
sediment studies that timber harvest produced significantly increased
volumes of streamflow on both treated watersheds (Harper, 1969; Hsieh,
1970).
A long-term flow-duration curve was used to estimate annual sediment
yields based on the flow regime during an average year. The procedure
used in this averaging technique was as follows. First, six years of
data from the calibration period were used to estimate the long-term
flow-duration characteristics of each stream. Next, a relation between
mean daily sediment concentration and mean daily discharge at each of
the three weirs was developed for each year of the study. Finally, the
sediment concentration-discharge relationship was combined with the
mean flow-duration curve for each weir to obtain the mass of sediment
carried in each flow class. Summing these values provided estimates
of total annual sediment yield from each watershed. Thus, this
technique assumed that the flow each year was equal to the long-term
mean in volume and distribution. This normalized the effect of abnormal
years by reducing the variation in sediment yield associated with
annual differences in discharge. The sediment yields thus attained
provide an indication of the average expectancy of a change associated
with the treatments. The normalized annual sediment yield from each
of the treated watersheds was then compared to that of the control with
regression analysis. A similar analysis technique has been described
by Anderson (1954).
Annual sediment yields are shown in Table I for each watershed.
Included are both normalized, or weighted, yields and "actual" yields
provided from the annual sediment-hydrograph analyses reported by the
U.S. Geological Survey. Regressions comparing normalized annual
sediment yield on Flynn Creek, the control, with that of each treated
watershed are illustrated in Figures 2 and 3. Only the upper 95%
confidence limits (CL) are calculated, since there is no reason to
suspect that treatment would reduce sediment yield.
Annual sediment yields were highly variable during the pretreatment
period. There was a three-fold difference between the minimum and
maximum annual yields on each watershed during this period, even using
normalized values.
Roadbuilding significantly increased sediment yield in Deer Creek,
41
-------
Table I
Total Annual Sediment Yield in Tons Per Square Mile
Computed from U.S. Geological Survey Records
Water Year
Flynn
Creek
Normalized Actual
1959
I960
1961
1962
1963
1964
1965
1966
1967
1968
1969
92
105
172
136
212
223
337
246
136
92
123
66
50
258
84
127
209
1237
300
137
59
139
Deer
Creek
Normalized Actual
114
81
193
178
285
231
308
577
251
101
162
82
77
286
97
160
199
1040
740
213
84
162
Needle Branch
Normalized
74
96
98
201
161
181
129
270
570
372
279
Actual
49
40
180
115
115
187
422
365
904
490
517
-------
LU
0 100 200 300 400 500
FLYNN CREEK SEDIMENT, TONS/SQ MILE
Figure 2. A comparison of normalized annual sediment yield from
Flynn Creek (unlogged) and Deer Creek (patchcut) for seven years
prior to treatment. Comparative yields after road building (]966)
and logging (1967-1969) are shown in relation to the 95 percent
confidence limit about the pretreatment regression.
43
-------
LU
600
2
O
LU
§300
LU
200
o:
m
o
LU
100
0
e
1967
e
1968
1969
A
95% CL
I
I
I
I
1
0 100 200 300 400 500
FLYNN CREEK SEDIMENT, TONS/SQ MILE
Figure 3. A comparison of normalized annual sediment yield from
Flynn Creek (unlogged) and Needle Branch (clearcut) for seven years
prior to treatment. Comparative yields after road building (1966)
and logging and burnlnj1, (1967-1969) are shown in relation to the
95 percent confidence linit about the pretreatnicnt regression.
44
-------
the patchcut watershed during the 1966 water year. One road slide
produced a sediment yield of 347 tons, or 40 percent of the
non-normalized yield, for the 1966 water year.
The increase in annual sediment yield after roadbuilding in Needle
Branch was also statistically significant at the 95% level. Road
drainage or erosion of side cast materials along roads seem the most
likely sources of this increase. No large road slide occurred in
Needle Branch.
The annual sediment yield observed during the first year after logging
in Deer Creek (1967) was significantly higher than during the control
period. This yield may include materials deposited by the large slide
the previous year. During the two post-logging water years of 1968 and
1969, sediment yields returned to prelogging levels.
Annual sediment yields in Needle Branch increased markedly immediately
after the watershed was logged and burned. The normalized sediment
yield increased four-fold over the pretreatment mean. The normalized
yield on the control watershed dropped to three-fourths of its
pretreatment mean during this year. The annual sediment yield declined
during the following years as vegetation returned, but yields remained
higher than before logging and burning.
45
-------
SECTION IV
CHANGES IN SEDIMENT CONCENTRATION
A second analysis compared the instantaneous concentrations of sediment
during storms with the streamflow at which the samples were obtained.
This was done both before and after roadbuilding and logging.
Understanding how sediment concentration varies with watershed treatment
is of great significance, since most of the new water quality standards
for sediment are related to this value, and not annual yield.
Regressions comparing instantaneous sediment concentration with the
streamflow observed when the sediment sample was taken were prepared for
each sampling station. It was necessary to segregate the concentration
data into two groups for these analyses since the sediment-streamflow
relationship for rising stages was significantly different from that
observed for falling stages at all sampling sites. A similar procedure
has been used by Fredriksen (1970). The correlation coefficient (r)
was generally much higher for rising stage data. All of the
concentration analyses therefore utilize only rising stage data.
Rising stage data were further segregated to include only storm data
with discharges greater than 5 csm.
Simultaneous samples of sediment concentration and discharge were fitted
to a regression equation of the form.
log S = a + b log D [1]
where S is the sediment concentration, D is the discharge measured
when the sediment sample was obtained, and a and b are regression
constants. ~~
Evaluating the differences in sediment concentration resulting from
watershed treatment proved a difficult task. Sample sizes tend to be
unequal. Variations in the sediment concentration-discharge equation
can also occur because of annual changes in runoff pattern. Two
additional types of variation may be imposed by treatment: clearcutting
may not only change the variation in sediment concentration, but the
variation in discharge as well. Thus, the assumption of orthogonality
in the individual degrees of freedom test may not apply. This means
that the standard test for interaction between regression equations
may not be appropriate.
The test statistic selected to circumvent this difficulty was:
SSB - (SSBj + SSB2)
2 SSA -
where SSp is the error sum of squares obtained by combining or pooling
46
-------
the sediment concentration-discharge data for the pretreatment period
(1) and each treatment year (2) for each watershed (A or B), SSi is
the error sum of squares for the pretreatment period, 882 is the error
sum of squares for the post-treatment period. Watershed A represents
the control watershed, Flynn Creek, and watershed B represents either
of the treated watersheds, Deer Creek or Needle Branch.
Mean streamflow and sediment concentrations are shown in Table II for
the U.S. Geological Survey weirs and in Table III for the small weirs
within beer Creek for each year of the study. Only two years of
pretreatment data were used in this analysis since more data were not
available from the small weirs in Deer Creek. Notation of a significant
increase is the result of testing regressions with equation (2) above
and not by simple tests on the mean values.
Some interesting differences appear in the conclusions which can be
drawn from analysis of separate samples and from analysis of annual
yield. In Deer Creek, the road slide in 1966 produced an increased
annual yield that was significant at the 95% level. The relative
significance dropped in the analysis of sediment concentration. The
increase was not significant at 95%, but rather 90%. The following
water year, 1967, annual yield was still significantly higher than the
pretreatment period at the 95% level. The comparison of sediment
concentrations during the 1967 water year with those during the control
period revealed no. significant differences. The reason for this
discrepancy is that there was a significant shift in the sediment-
discharge relationship of the control watershed during this year.
It is quite likely that the shift in the sediment concentration-discharge
relationship of the control occurred as a result of the 1964-1965 floods,
and was produced by residual materials deposited during those major
events. Prior to the floods, the maximum suspended sediment
concentration recorded on Flynn Creek, the control, was 682 ppm
compared to 969 ppm recorded on Needle Branch. During the flood, the
maximum Flynn Creek concentration was 2,050 ppm compared with 476 ppm
on Needle Branch. Most of the pools in Flynn Creek were filled with
sediment (Williams, 1965). Thus, the control watershed responded
differently to the same storm event than did the other two watersheds
which were treated the following year. Anderson (1968, 1970) has noted
the dissimilar responses of other watersheds to the same large event.
He has also shown that materials deposited during these events provide
a sediment reservoir for many subsequent years. Thus, the classical
concept of a single "control" watershed for sediment studies of this
type may not always be valid.
The same pattern appears on Needle Branch. The difference during the
first water year after logging, 1967, is even more profound. A
five-fold increase in mean sediment concentration was not significant
at the 90% level because of the upward shift in the sediment-discharge
regression of the control watershed during this same year. The
47
-------
Table II
An Analysis of Simultaneous Sampling of Suspended Sediment Concentration
and Streamflow at U.S. Geological Survey Stations During
Rising Stages and Discharges Greater Than 5 CSM
Station
Flynn
Creek
(control)
Deer
Creek
(patch
cut)
i
Needle
Branch
(clear-
cut)
Water
Year
1964-65
1966
1967
1968
1969
1964-65
1966
1967
1968
1969
1964-65
1966
1967
1968
1S6S
Sample
Size
72
64
28
18
17
71
66
32
20
49
88
68
89
44
38
Sediment Concentration (ppm)
Range Mean
1- 205 194
1- 718 128
38- 439 148
32- 256 109
1- 200 57
1-1610 267
1-6960 337*
53- 670 233
35- 345 115
5- 381 90
1- 969 116
1- 892 179*
1-6300 589
20-7670 640**
70- 738 280**
Streamflow (cfs)
Range Mean
4.2-148 32
4.2- 66 22
4,2- 69 32
11 - 44 26
7-44 20
6 -204 58
6 -115 32
20 -105 59
1C - 46 40
8-76 37
1.6- 45 13
1.6- 27 9
1.6- 25 11
l.S- 32 10
3.1- 24 12
Significant increase at 90% level.
** Significant increase at 95% level.
48
-------
Table III
An Analysis of Simultaneous Sampling of Suspended Sediment Concentrations
and Streamflow at Stations Within Deer Creek During
Rising Stages and Discharges Greater than 5 CSM
Station
Deer
Creek II
Deer
Creek III
Deer
Creek IV
Deer
Creek VI
Water
Year
1964-65
1966
1967
1964-65
1966
1967
1964-65
1966
1967
1964-65
1966
1967
Sample
Size
55
15
14
54
15
14
20
14
13
33
21
19
Sediment Concentration (ppm)
Range Me an
1-716 70
2-178 32
3-242 67
1-793 79
1-410 117
4-194 72
1- 99 15
1- 10 2
1-8 4
3-462 88
1-720 127*
1-366 122
Streamflow (cfs)
Range Mean
1.0- 34 0.8
1.0- 9 3.7
1.0- 30 11.3
0.8-19 3.7
1.2- 18 4.9
0.8-18 7.1
0.3-10 1.7
0.3- 8 1.8
0.3- 7 2.0
4.5-136 16
4.5-108 20
6-84 33
•e Significant increase at 90% level.
49
-------
increase on Needle Branch is significant at about 87%, but the difference
in statistical significance is still surprising.
The comparison of changes in sediment concentration for the small weirs
in Deer Creek is shown in Table' III. Roadbuilding produced significant
changes only at Station VI. Station III, immediately below the road
.'••lide, showed no increase in sediment concentration because samples were
not taken during this event. Logging did not produce significant
increases in sediment concentration at any of the sampling stations in
Deer Creek.
A frequency distribution of mean daily sediment concentration during
low flow periods is shown in Table IV. Mean daily flow less than 5 csm
occurred during 60-70% of the year on these coastal watersheds. Even
with the severe treatment given Needle Branch during the 1967 water
year, mean daily concentrations were less than 10 pprn during about 97%
of these low-flow days. These data substantiate the fact that in
mountain watersheds the majority of the sediment load is carried during
a few large storms. The best indication of treatment effect is shown
in the increased maximum concentration at low flow on Needle Branch
from water years, 1966 through 1969. The pattern is similar to that
of sediment yield shown in Table I.
50
-------
Table IV
A Frequency Distribution of Mean Daily Suspended Sediment Concentrations
During Days with Mean Flow Less Than 5 CSM
Station
Flynn
Creek
Deer
Creek
Needle
Branch
Water
Year
1959-65
1966
1967
1968
1969
1959-65
1966
1967
1968
1969
1959-65
1966
1967
1968
1969
Concentration Class (ppm)
0-5 5-10 10-20 20-30 >30
% Days <5 CSM By Class
91.9 5.4 2.1 0.3 0.3
84.4 11.0 4.2 0.4
97.4 2.2 0.4
98.1 1.1 0.4 0.4
92.8 6.8 0.4
89.3 7.7 2.5 0.3 0.2
84.7 11.5 3.4 0.4
77.1 16.7 1.2 3.8
79.3 17.9 0.8 0.4 1.6
89.4 4.6 4.2 0.9 0.9
94.4 4.7 0.8 0.1
89.8 9.0 0.4 0.4 0.4
96.6 0.8 1.4 0.4 0.8
93.2 2.4 2.0 2.4
93.4 3.5 1.3 1.4 0.4
Maximum Cone
(ppm)
26
21
13
26
12
28
28
52
53
37
15
74
220
413
230
Days /Year
With Flow
<5 CSM
231
263
237
267
251
237
261
240
257
216
235
256
238
250
23C
51
-------
SECTION V
DISCUSSION
The results of this intensive study of sediment yield and land use
clearly illustrate the effect of several forest management practices
on water quality. The influence of roads on sediment yield has again
been demonstrated by this study, substantiating the conclusions drawn
by Fredriksen (1970). The road system in Deer Creek, for example, was
conservatively located, constructed, and used. The roads were located
near the ridges. They entered the watershed from the back of the
ridges, thus minimizing the road mileage within the watershed. The
roads were well gravelled and were not used during the winter months.
Even with these precautions, one slide occurred and its effect was
quite significant. A large volume of material still remains trapped
behind a log jam in the upper part of the watershed, providing a
potential source for additional sediment yield at some later date.
About 1.5 miles of road were constructed within Needle Branch. This,
together with the landings for logs, exposed mineral soil over about
7% of the watershed. This undoubtedly was the source of sediment in
water year 1966.
High-lead logging alone did not produce amounts of sediment significantl;
different from those in the calibration period. This result also
coincides with those observed elsewhere (Fredriksen, 1970; Lull and
Satterlund, 1963; Meehan, et^. al_. , 1969; Packer, 1967). The maximum
sediment concentration observed at weir IV in Deer Creek was less than
20 ppm. The watershed was 90 percent clearcut, but unburned. This
can be compared with a maximum sediment concentration of over 7,000 ppm
observed after logging and burning in Needle Branch.
Slash burning is a common management practice in the Pacific Northwest
following clearcutting. Arguments both for and against burning are
numerous. A review of this controversy is clearly beyond the scope of
this paper. The sediment data collected during this study, however,
indicate the effect of this practice on water quality.
The slash fire in Needle Branch was extremely hot; mineral soil was
exposed through most of the watershed. High sediment yields could
be expected where mineral soil is subjected to over 100 inches of rain
during a six-month period.
The cause of increased sediment yield after logging in Deer Creek is
somewhat obscure and may be the result of interacting factors. The
sediment contribution of the two upstream clearcuts, as indicated
by changes in their sediment concentration-discharge regressions, was
not significant. Downstream, two possible sources of sediment exist.
Materials which accumulated in the stream channel as a result of the
52
-------
road slide the previous year may have accounted for part of the
noted increase. The mean sediment concentration at station VI
in Deer Creek, though not statistically significant, remained high
implicating some residual source. A second downstream source of
sediment may have been the clearcut furthest downstream, which
lies between Station VI and the U.S. Geological Survey station.
This clearcut was lightly burned after logging, which could have
contributed to an increase in sediment yield.
The data indicate that sediment yields should approach pretreatment
levels 5-6 years after complete clearcutting and burning. Two
important concepts must be understood about this recovery. First,
these results pertain to a case study conducted in an area where
vegetation grows rapidly and returns to unoccupied sites very quickly.
Exposure time for mineral soil is thus minimized. Second, it is
essential to remember that erosion is significant for both terrestrial
and aquatic habitats. Although the supply of sediment from the
slopes may decline rapidly, the presence of this material in the
stream gravels may persist. Such an accumulation of fine materials
in spawning gravel can significantly reduce the emergence of salmonid
fry (Hall and Lantz, 1969).
What inferences can be drawn from the data about sediment sampling
or monitoring? This question is a crucial one, not only from the
point of view of studying sediment transport processes, but from
a water quality control aspect as well. The Oregon water quality
standards (Oregon Department of Environmental Quality, 1969) for
example, specify that no activities will be permitted which cause
"any measureable increases in natural stream turbidities when natural
turbidities are less than 30 Jackson Turbidity Units (JTU) or more
than a 10 percent cumulative increase in natural stream turbidities
when stream turbidities are more than 30 JTU..."
The important question that must be answered is how to obtain the
best standard of comparison. What, in other words, is a "natural"
sediment concentration for small streams? Our data indicate that
rather large annual variations in the sediment-discharge relationship
can occur on undisturbed watersheds. Variations between watersheds
may also be large. Variation in the sediment-discharge relationship
is stage-dependent. A much better correlation between sediment
concentration and discharge was observed during rising stages.
This leads to the conclusion that a great deal of experience, together
with an intensive, rigidly standardized sampling scheme based on
flow regimes is required before a judgment with a precision of
10% can be made.
Our ability to make accurate judgments about changes in the sediment
concentration-discharge relationship in this study would have been
greatly improved by replicating the control. The assumption that
neighboring watersheds respond in similar fashion to similar events,
53
-------
regardless of magnitude, is certainly questionable. It would seem then,
that any sediment monitoring system would require more than one control
for comparison.
A further constraint in sediment sampling or monitoring is imposed by
the influence of a few large storms on annual sediment yields. If these
events are not sampled adequately or are missed, conclusions about the
treatment effect are likely to be erroneous. This problem is compounded
by the treatment itself, which imposes a greater variation on the
sediment-discharge relationship, particularly at high flows. Thus,
monitoring a specific stream to detect a 10% change in sediment
concentration will require more than just a few random samples.
We have shown that clearcut logging may produce little or no change in
sediment concentrations in small streams. The greatest changes were
associated with the roadbuilding operation that preceded logging and
the controlled slash burning afterward. We have also shown that unless
these changes are large, it may be very difficult to separate
man-caused changes in sediment concentration from those imposed by
natural variation, particularly if very large events occur within the
measurement period.
54
-------
SECTION VI
REFERENCES
Anderson, Henry W., Suspended sediment discharge as related to
streamflow, topography., soil, and land use, Trans, Amer.
Geophys Union 35: 268-281, 1954.
Anderson, Henry W. and James R. Wallis., Some interpretations of
sediment sources and causes, Pacific Coast Basins in Oregon
and California, in Proc, Fed. Interagency Sedimentation
Conf., 1963. U.S. Dep. AgrT, Misc. Pub. No. 970, pp. 22-30.
1965.
Anderson, Henry W., Major floor effects on subsequent suspended
sediment discharge. 49th Ann. Meeting Amer. Geophys. Union,
Washington, D.C., Apr. 8-11, 1968, (Abstract) Trans. Amer.
Geophys. Union 49 (1):175, 1968.
Anderson, Henry W., Relative contributions of sediment from source
areas, and transport processes, in Proc. Symposium on Forest
Land Uses and Stream Environment, Oregon State University"
Corvallis,~T970T
Brannon, E. L., The influence of physical factors on the development
and weight of sockeye salmon embryos and alevins, Int. Pac.
Salmon Fish Comm. , Progress Report No. 12. 26 pp. 1965"!
Copeland, 0. L., Land use and ecological factors in relation to
sediment yields, in Proc. Fed. Interagency Sedimentation
Conf., 1963. U.S. DepT~Agr77 Misc. Pub. No. 970, pp 72-84.
1965.
Cordone, A. J. and D. E. Kelley., The influence of inorganic
sediment on the aquatic life of streams, Calif. Fish and
Game 47:188-288. 1961.
Fredriksen, R. L. , Sedimentation After Logging Road Construction
in a Small Western Oregon Watershed, in Proceedings of the
Federal Interagency Sedimentation Conference , 1963. U.S. Dep.
Agr., Misc. Pub. 970. pp. 56-59. 1965.
Fredriksen, R. L. Erosion and sedimentation following road
construction and timber harvest on unstable soils in three small
western Oregon watersheds. U.S. Dep. Agr., Forest Serv.,
Pac. Northwest Forest and Range Expt. Sta., Res. Paper PNW-104,
TT~pp. 1970.
Hall, James D. and Richard L. Lantz, Effects of logging on the
habitat of coho salmon and cutthroat trout in coastal streams,
55
-------
in A_ Symposium on_ Salmon and Trout in_ Streams, T. G. Northcote,
Ed. University of British Columbia, Vancouver, 388 pp. 1969.
Harper, Warren C., Changes in Storm Hydrographs Due to Clearcut
Logging in Coastal Watersheds, M.S. Thesis, Oregon State
University, Corvallis. 1969.
Hsieh, Frederic S., Storm Runoff Response from Roadbuilding and Logging
on Small Watersheds in the Oregon Coast Range, M.S. Thesis, Oregon
State University, Corvallis. 1970.
Kramer, Robert H., Effects of suspended wood fiber on brown and rainbow
trout eggs and alevins, Trans. Ainer. Fish Soc. 94:252-258. 1965.
Koski, K. V. , The Survival of Coho Salmon (Oncorhynchus kisutch) from
Egg Deposition to Emergence in Three Oregon Coastal Streams, Masters
thesis, Oregon State University, Corvallis, 1966.
Lull, H. W. and D. R. Satterlund, What's new in municipal watershed
management. Proc. Soc. Amer. Foresters. pp. 171-175. 1963.
Meehan, W. R., W. A. Farr, D. M. Bishop and J. H. Patric, Effect of
clearcutting on salmon habitat of two southeast Alaska streams.
U.S. Dep. Agr., Forest Serv., Pac. Northwest Forest and Range
Expt. Sta. Res. Pap. PNW^ST, 45~pp. T35W.
Oregon Department of Environmental Quality, Water Quality Standards for
the Umpqua River Basin, Oregon. 34 pp. 1969.
Oregon Game Commission, A study of the effect of logging on aquatic
resources, 1960-66. Research Div. Progress Memorandum, Fisheries,
No. 3, 28 pp. 1967.
Packer, Paul E., Forest treatment effects on water quality, in Int.
Symposium of_ Forest Hydrology. William E. Sopper and Howard W. Lull,
EdlTPergambn Press, pp. 687-689. 1967.
Reinhart, K. G. and A. R. Eschner, Effect on streamflow of four
different forest practices in Allegheny Mountains, Journal of
Geophys. Res. 67_, 2433-2445, 1962.
Rice, R. M. and J. R. Wallis, How a logging operation can affect
streamflow. Forest Industries 89(11)."38-40, Nov. 1962.
Ringler, Neil, Effects of Logging on the Spawning Bed Environment in
Two Oregon Coastal Streams, M.S. Thesis, Oregon State University,
96 pp. 1970.
Servizi, J. A., R. W. Gordon and D. W. Martens, Marine disposal of
sediments from Bellingham Harbor as related to sockeye and pink
56
-------
salmon fisheries, Int. Pac. Salmon Fish. Comm., Progress Report
No_. 23. 38 pp. 1969.
Shelton, J. M. and R. D. Pollock, Si Itat ion and egg survival in
incubation channels, Trans. Amer. Fish. Soc. 95:183-187. 1966.
Sheridan, W. L. and W. J. McNeil, Some effects of logging on two
salmon streams in Alaska. Jour, of Forestry 66(2):128-134. 1968.
Williams, R. C., Report to Annual Alsea Study Meeting, Oregon State
University, Corvallis. 1965. U.S. Geological Survey (Unpublished).
57
-------
CHAPTER VI
EVALUATION OF BED LOAD AND TOTAL
SEDIMENT YIELD PROCESSES ON SMALL
MOUNTAIN STREAMS
SECTION I
INTRODUCTION
Scope and Purpose
The purpose of this phase of the research has been to develop a
fundamental grasp of the physical processes involved in sediment
movement in small mountain streams. Particular emphasis has been
given to the bed-load processes which may cause great quantities of
gravel to be disturbed and transported during periods of storm
runoff. At such times the amount of silt also carried by the stream
may become much higher, partly attributable to what might be called
a "silt reservoir" effect of streambed gravels—whereby silt is trapped
and collected in the gravel bed when the bed is undisturbed and is
released by the bed when the protective gravels are dislodged and
moved by the flowing water.
It is common in studies of watershed practices to develop correlations
between stream discharge and various modes of sediment yield, such as
suspended sediment yield or total sediment yield. Development of
such correlations depends upon field sampling of the flow and
measurement of trapped sediments, both of which have deficiencies of
many types, including infrequency of sampling. Because the transport
rate of suspended particles is principally a function of the
availability of particles for transport, rather than any physical
relationship with streamflow, the correlation diagrams typically
exhibit a great deal of scatter. Nevertheless, if sufficient data
are obtained the effect of a watershed practice upon suspended load
can often be reliably detected. The total sediment yield, however,
is much less reliably and more difficultly determined in most
instances and therefore is often omitted from studies of watershed
practices. The total load consists of both the suspended load
and the bed load. Considerable research on bed-load transport has
shown a definite physical relationship between rate of transport and
stream discharge (unlike the situation for suspended load).
Unfortunately, the application of known relationships becomes
questionable for small mountain streams with coarse gravel beds,
shallow flow depths, and frequent riffles and pools. Consequently,
the alternative approach to bed-load measurement by trapping sediment
in pools above weirs (or similar techniques) is often resorted to.
The results are unsatisfactory for several reasons, which include
loss by flushing out during very high runoff, compaction of sediment,
and an overly long time increment between measurements — such that
the transport may be estimated for a particular runoff hydrograph,
58
-------
but is not sufficiently related to the individual discharges which
make up that hydrograph (i.e., the transferability of results remains
uncertain).
With an awareness of the difficulties in assessing watershed practice
effects upon total sediment yield, this phase of the research was
established to seek a clearer understanding of the bed-load processes
for small mountain streams. The immediate scope of the study has
been to develop an instrumented reach of stream to conduct detailed
investigations of bed-load and total-load transport for natural or
controlled streamflow conditions. The broader scope of the study
has been to apply the research findings to identification of the
effects of watershed practices upon the stream environment downstream
of the disturbed watershed.
Relation to Water Quality Problems
Turbidity, along with temperature, dissolved oxygen, and other key
parameters, is an important determinant of water's quality. Hence,
use of the term "silt pollution" has become commonplace where the
turbidity of streams by unnatural causes has had any adverse (or
suspected adverse) effects on use of that water for whatever purpose,
including its aesthetic use, by man or other organisms.
Sedimentation and increased turbidity of streams are common
consequences of road building, logging, agricultural land use, mining,
sand-and-gravel operations, and urban development in all parts of the
nation. These problems are especially severe in regions of steep
topography, such as the mountainous states of the West. They are
perhaps most critical in the Pacific Northwest, where abundant
rainfall and extensive forests result in widespread logging and
severe watershed erosion—where the observer in an airplane can see
miles of uncut forests drained by clear blue-green streams or see
patches of logged forests drained by chocolate-colored sediment-laden
streams--where clear and muddy tributaries join and flow together to
the sea in a growing river of silt-polluted water.
In time, nature may have a chance to heal the scars on the watershed.
But meanwhile, any changes in sediment production alter the substrate
of streams and estuaries and, consequently, the biology and fishery
resources of those areas. Similarly, water withdrawals from silt-laden
streams for municipal, industrial, and agricultural uses are
adversely affected, which leads to greatly increased costs for water
treatment. Water-oriented recreational activities also suffer from
muddy waters.
The role of the streambed in controlling the quality of water flowing
over it may be appreciable. A watershed practice could have a
significant effect upon the composition of material in the streambed
and thus exert its influence upon the stream environment long after
the initial removal of sediment or organic material from the watershed
by storm runoff. At times, a gravel streambed may provide in nature
59
-------
much the same filtration effect upon fine particulate matter as
does a sand filter in a water treatment plant. Subsequent storm
runoff might be sufficient to disturb the gravel bed and permit
a return of those fine particles to the flowing stream. This could
compound any problem already present due to other material carried
from upstream by the runoff. Furthermore, the watershed practice
might increase flood runoff and cause greater disturbance of the
streambed.
Although fine sediments and organic matter may be the chief offenders
of water quality in streams draining timbered watersheds, sorbed
toxic materials might also occur downstream because of applications
of herbicides. Toxic materials likewise may be found downstream of
some mining and agricultural operations. Also., an extensive literature
deals with analogous problems with radionuclides on large streams.
Hence, the types of pollutant material which may be temporarily
stored in the gravel pores of a streambed are quite varied and not
merely confined to fine sediments.
60
-------
SECTION II
OBJECTIVES
Broad Objectives
The broad objectives of this phase of the research were to:
1. develop a better understanding of the sediment transport
mechanisms in small rough-bottomed streams;
2. relate sediment transport to watershed conditions and land use;
and
3. relate sediment transport to stream water quality.
These objectives were complementary to other phases of this research
in developing a fuller understanding of the effects of watershed
practices upon streams.
Specific Objectives
The detailed objectives of this phase of the research were to:
1. obtain measures of bed-load movement for a variety of steady and
transient flow conditions in order to better evaluate:
(a) the relationship between stream hydraulics and bed-material
movement for coarse bed materials, especially for transient conditions.
(b) the applicability of existing theories for evaluation of
bed-load movement of coarse gravels subject to steady and variable
discharges.
(c) the effect of heterogeneity of bed-material on the transport
of smaller particles present in the bed.
(d) the depth of scour of coarse bed materials in mountain
streams and the influx or efflux of silt particles from the gravel
pores.
(e) the factors influencing gravel bar formation in mountain
streambeds.
2. evaluate the relationships between total sediment yield,
suspended-sediment yield, bed-material disturbance, and land use
based upon information obtained in achieving the first objective.
3. apply the knowledge gained under the first and second objectives
to evaluate the total sediment load from the Alsea Experimental
Watersheds and thus to evaluate the effects of land use on water
quality.
Timetable to Achieve Objectives — Curtailment of Project
The objectives for this phase of the research were established for
accomplishment with a three-year program. Principal activities
scheduled for the first project year were to be literature reviews,
facilities design, and facilities construction. The second and third
years were to be devoted to concurrent field experiments and data
analyses, and the third project year was scheduled to conclude with data
evaluation and interpretation and the preparation of a detailed
61
-------
report. The work was to be conducted by a faculty principal
investigator, a Ph.D. candidate developing his thesis along the
general lines of the project, and several aides to assist with field
office work.
Curtailment of the project at the end of the second project year
has made it impossible to complete several elements of the research
included in the original objectives. However, in connection with
some research objectives, a sufficient amount of data and/ or
insights have been obtained to permit at least some tentative
conclusions to be drawn. Therefore, such conclusions have been
made where sufficient sound evidence is available, together with
adequate time for careful evaluation and interpretation.
62
-------
SECTION III
EXPERIMENTAL APPROACH
General Aspects
This phase of the research was designed to provide answers to some
critical questions on sediment yield processes and related stream
hydraulics for rivers transporting a broad range of particle sizes.
Emphasis was placed upon both the coarse particles (gravels), which
determine the general streambed features and provide a "silt
reservoir" in their intragravel void spaces, and the fine particles,
which often constitute some type of pollutant through adverse effects
upon water quality. Here the term "silt" is being used rather
loosely, for descriptive purposes, to indicate fine particles in the
stream which usually are carried in suspension when in motion and
which may be found in the intragravel void space of an undisturbed
streambed.
Previous research on suspended sediment yield from logged areas has
indicated the need for a sediment model that defines the following:
1. the relationship of suspended sediment load to bed-load transport
and to total sediment yield;
2. the relationship of total sediment yield to land use;
3. the hydraulic factors influencing movement of gravels and boulders
of various sizes along the streambed;
4. the magnitude of bed-material transport during
(a) steady-state stream conditions
(b) transient (storm) stream conditions;
5. sediment yield related to a range of logging practices.
The present research was planned to develop such a model during a
three-year project period. Because the frequent storm-flood event is
crucial in introducing large quantities of fine sediment into the
streamflow, this transient condition was given particularly detailed
examination.
The total sediment load of a stream consists of (1) the bed load of
material which moves at the streambed and is similar in size to
particles making up the bed, and (2) the suspended or wash load of
smaller particles carried throughout the entire water profile of the
stream. The suspended-load particles do not represent an important
percentage by weight of the streambed's composition. Nevertheless,
such particles occupy the pore spaces among the coarse particles
which comprise the bed and thus significantly affect the pore water
quality of the bed. The streamflow may easily carry large
quantities of this suspended load; hence, the wash load is limited
only by the availability of small particles in the watershed,
streambanks, and streambed, all of which are, in turn, significantly
affected by watershed practices. Conversely, the amount of bed-material
load is governed by the ability of the flow to transport it, rather
63
-------
than by availability of particles in the watershed. Here, too,
watershed practices may have a significant effect by altering flood
discharges (as has been demonstrated on the Alsea watersheds).
Thus, the total sediment load of a stream is influenced by watershed
practices, but the influence may differ for the wash load and bed-
material load which make up this total load.
In most watershed studies, only the suspended load has been measured.
The measured load was then related to land use. This approach has,
of necessity, been utilized for studies on mountain streams because
classical theories of bed-load movement were derived principally for
streams with sandy bottoms or for equilibrium flow conditions.
Bed-load theories for equilibrium transport of coarse gravels and
boulders have many limitations, and none is universally applicable.
In short, the effect of land use on the erosion process and on
water quality was assumed previously to be reflected in suspended
sediment yields, without consideration of the effect of land use on
the hydraulic processes ultimately responsible for sediment transport
in mountain channels. Particularly concerning movement of large
gravel, such assumptions were the only recourse in studies of the
effect of land use on water quality.
Therefore, the general approach taken in this research regarding the
problem of total sediment yield was to measure total yield in such a
manner that the two distinct sediment-transport modes could be
separately determined. Further, the process of sediment transport was
analyzed in terms of its controlling hydraulic factors and the degree
of disturbance of the streambed.
Planned Approach for Three-Year Project Period
Detailed examination of bed-load transport, sUspended-sediment
transport and total sediment yield was to be conducted at the Oak
Creek facilities especially developed for this purpose. The bed-load
process was to be analyzed in terms of its controlling hydraulic
factors. Both transient and steady-state conditions were to be used.
The degree of disturbance of the stream bed during transport was to
be established.
Completion of this part of the project was expected to permit
consideration of a wide range of sediment-related questions which
comprise the remaining objectives. For example, a realistic
(bed/load)/(suspended-load)/total-load) sediment model might then be
formulated and related to the changed hydraulic characteristics
below logged areas. This would identify the effects of logging on
the sediment processes. The findings would also permit an
examination of the utility of predictive techniques for relating
land-use effects to water quality.
64
-------
'Die effect of logging on sediment production from forested watersheds
was to be determined from sediment data obtained at the Alsea
experimental watersheds, where different logging practices were
applied on several sub-basins for this purpose.
Extensive data on streamflows and sediment transport (suspended
load and approximate total load) have been collected over several
years, up to the present. Logging on two of the Alsea experimental
watersheds occurred in 1966] the third watershed serves as a "control"
for comparative studies. The analyzed data, together with information
on streambed scour and deposition, was to be related to sediment
transport processes. Hydraulic characteristics of the Alsea
streams and composition of the streambeds were to be determined for
use in determining sediment transport conditions below different
logged areas.
The fate of released sediment and the nature of resulting problems
also were to be determined at the Alsea watersheds, as well as at
the Oak Creek watershed. Inferences were to be drawn in part from
data on streambed siltation and in part from other sediment data
obtained at the Alsea. Inferences were also to be made from
relevant published research, so as to expand the applicability of
this research.
Changes in hydraulic characteristics of streams downstream of
logged areas were to be determined from data and observations at the
Alsea watersheds. Supporting information published by other
researchers was to be utilized also.
Finally, the research findings were to be evaluated from the viewpoint
of suggesting possible watershed practices to minimize or avoid
adverse water-quality problems with silt and related pollutants so as
to protect the stream environment.
Modified Approach Due to Curtailment of Project
No major modifications in experimental approach were made at the Oak
Creek facilities concerning those experiments dealing with sediment
transport during naturally varying streamflows. However, the number
of experiments conducted under controlled streamflows was greatly
reduced to permit more time for data work-up and analysis.
Only a limited amount of data regarding bed-load materials in the
Alsea Experimental Watershed was gathered because of the time reduction
for this research.
65
-------
SECTION IV
DEVELOPMENT OF OAK CREEK RESEARCH FACILITIES
Oak Creek, draining McDonald Forest in the Coastal Range just five
miles west of the Oregon State University campus, offers several
hundred feet of stream channel suitable for research purposes. The
vegetation, terrain, stream profile, streamflows, and streambcd
composition are typical of many Pacific Northwest streams which
provide spawning gravels for anadromous fish and which drain watersheds
subject to logging. Access roads conveniently parallel the stream
over most of its length. Because of these suitable stream
characteristics and the great convenience for conducting research, a
portion of Oak Creek was developed to facilitate study of sediment-
transport processes in small mountain streams. Facilities were
developed at two nearby reaches of the creek; the downstream reach was
left essentially in its natural state and the upstream reach modified
to permit water diversion into a concrete channel.
Instrumented Natural Stream
The downstream reach provides a 500-foot instrumented portion of the
natural channel for study of sediment transport as a result of natural
storm runoff. A schematic diagram of the reach and facilities is shown
in Figure 1. Both the bed load and the suspended load of the stream
can be measured. Depth of streambed scour and deposition can be
determined as well as any significant changes in the features of the
streambed and stream banks. Most of this reach has been improved only
by clearing overhanging bank vegetation, channel debris, and some
overbank vegetation to improve access and permit the use of surveying
techniques in data acquisition.
Along this natural test reach, cross-sectioning stations were
established at 12.5-foot intervals to permit study of changes in
streambed and bank topography. Metal pins embedded in concrete mark
the far ends of each cross-section.
Staff gages were mounted at seven locations along the reach to allow
determination of water surface slopes and energy gradients over segments
of the reach in conjunction with other experimental observations.
At the downstream end of the study area are several devices to
permit the collection of streamflow and sediment transport data.
Progressing in the downstream direction (Figure 1) these include:
(1) an access bridge for work during major floods; (2) a plank bridge
for stream-gaging moderate flows and for obtaining suspended sediment
samples with a hand sampler (DI1 48) ; (3) a support structure for the
intake tube to an automatic sampling system for determining the
integrated suspended sediment transport over extended periods; (4)
intakes to a stilling well and the stilling well and water-level
recorder (Type F) for obtaining a continuous record of river stage;
66
-------
; •. ' , • .'• .'Gravel Roact ••'•
::.;-.ire 1. Schematic view of researcli facilities at instrumented reach of Oak Creek.
-------
(5) the instrument house for the automatic suspended-load sampler
(connected to the stilling well so that sampling frequency is a function
of stage); and (6) a broad-crested weir into which a bed-load
sampler has been incorporated. The broad-crested weir acts as a control
for water level at the stilling well to provide a constant stage-
discharge relation for the reach. The bed-load sampler is of unique
design, using a vortex principle to remove the bed load from the
channel into a sampling box, after which the withdrawn water (only
a small fraction of the total streamflow) is returned to the stream
via a bypass line. Control gates for the sampler permit continuous or
discrete sampling of the entire bed load passing this section of the
stream. Detailed information on the features of the weir/trap
structure is presented in the next section because of the potential
utility of such a device for other sediment research.
The Vortex Bed-Load Sampler
During the literature review on sediment sampling, before the design
of the Oak Creek research facilities, it was thought that a bed-load
trap for in-stream sediment collection might be devised with features
similar to those used in some of the large flumes in various hydraulic
laboratories. However, a vortex tube sand trap described by Robinson
(1962) for excluding unwanted sediment from irrigation and other canals
appeared to have possibilities for adaptation as a bed-load sampler.
Because there was not much hydraulic and sediment information for Oak
Creek upon which to base a careful design, a rough correspondence to
Robinson's design criteria was attempted instead. Subsequent operation
of the bed-load sampler indicated no major difficulties (although
several minor changes might be incorporated into any future
sampler). The resulting vortex-type bed-load sampling system is shown
in plan view and cross-sectional views in Figure 2.
In designing the vortex bed-load sampler, doubts existed as to the
capability of the vortex flume for handling coarse gravel and cobbles
up to six inches in diameter (major axis). Therefore, a second
trough was placed two feet douTistream of and parallel to the vortex
flume to act as a backup trough. Large material escaping from the
vortex presumably would fall into this trough, where it could be
collected. By use of the two troughs, the hope was to have a 100
percent efficient bed-load trap. Hindsight indicates that the second
trough was unnecessary insofar as the coarse bed-load material was
concerned. All particles larger than No. 4 sieve size (U.S. Standard
Series) in diameter (3/16 inch) were trapped and held by the vortex
for all streamflows--none reached the backup trough when the vortex
trap was operated. As particle size became smaller in the sand range,
the trap efficiency decreased and became dependent upon streamflow,
according to inferences made from the data on bed-load transport.
Thus, at flows greater than 40 cubic feet per second (cfs) the trap
efficiency became very low for material of less than the No. 50
sieve size (0.297 millimeter) and as the flow decreased the trap
68
-------
1 f
T
» I
' t»
»
•
:-.'
KL.AN VltW
v •. . • '.-•• .•••:>.••.. I h'Vrv
<^Pr~rmM A A
:'•
•;
;r
...
•«**
P
^
H^otf A*c
*">
1
i
a
'^^t."*j
7'4»--
SECTION B-B
SECTION
se*.t,
'••';c"ire 2. Rro-.'1-crcste;.: <;eir ".mi v~:-i.nx : 7? sediment s^»rli.'ir systcr (!-cir/tr'-.r structure).
-------
efficiency increased. This same tendency was exhibited in various
degrees for other particle sizes ranging from those passing the No. 4
sieve to those retained on the No. 200 sieve (0.074 millimeter).
Particles finer than 0.074 millimeter were trapped in such small
amounts that the trap efficiency for silt-sized and smaller particles
was believed to be virtually zero. The trap efficiencies of less
than 100 percent indicated for the finer fraction of bed load are
believed to be explained by the turbulence of flow in the vortex,
such that the finer particles were temporarily placed in suspension
and thrown back out of the vortex flume or carried past the trap
and into the bypass line.
The vortex flume and backup trough have an angle of orientation of
almost 60 degrees to the direction of flow. The orientation was
determined as much by tree roots in the streambanks as by the
criterion of 45 degrees recommended by Robinson.
The vortex flume is placed horizontally and has its upstream and
downstream edges at the same level. In cross-sectional shape the bottom
is flat and 12 inches wide, but the sidewalls are curved and have a
maximum width of 18 inches. The top opening is 12 inches wide.
Total depth of the flume is 12 inches. The shape was selected for
easy fabrication. The total flow length of the vortex flume is
19.5 feet, and the length of opening in the concrete channel floor
is 14.5 feet. A vortex develops readily at all stream stages when the
control gates are opened.
The backup trough is horizontal with upstream and downstream edges
at the same level. It has a square 12-inch by 12-inch cross-sectional
shape.
The concrete channel floor, in addition to holding in place the two
sampling troughs, acts as a broad-crested weir and helps stabilize the
stage-discharge relation at the stilling well. However, operation of
the vortex causes a backwater curve that changes the stages for a
short distance upstream. Consequently, either a correction curve or
a dual rating curve is required to convert the water-level data to the
corresponding discharges—both when the vortex bed-load sampler is
open for use and when it is closed. Placing the stilling well a
greater distance upstream could avoid or minimize this problem.
The vortex flume leads from the stream to a work pit having a concrete
floor at the same level as the channel floor. This pit greatly
increases the ease of collecting samples. Flat metal plates with
handles serve as control gates in the work pit to regulate the vortex
flow. The vortex flume opens into a deep metal box, or sampling trap,
in the floor of the work pit. A smaller sample box can be placed in
position at the vortex exit within the sampling trap to catch the bed
load as it decelerates upon leaving the vortex flume.
71
-------
Sample boxes are raised from and lowered into the sampling trap by
means of a chain hoist attached to a pulley and supported by a hoist
frame. The hoist frame also permits shifting the sample box to
higher ground outside the work pit, where the sample can be stored or
transferred to containers for subsequent laboratory analyses.
After the bed load has been deposited in the sampling box, the
water drawn into the sampling area is returned to the stream by a
return pipe. Because of local topographic features, a 100-foot line
of 12-inch-diameter culvert pipe was used. Under different circumstances
a shorter "bypass" or return line would be equally effective. The
difference in energy head across the bypass culvert depends upon river
stage and has a mean value of approximately three feet.
Operational experience has shown that most of the bed load drops into
the sample box. However, sufficient turbulence occurs in the sampling
trap so that some of the fine sand deposits in the bottom of the trap
instead of collecting in the sample box. Subsequent collection of
this sand poses no special problems other than time loss and some
inconvenience. An improvement would be the use of a sampling trap
large enough to use larger sampling boxes and to achieve a greater
reduction in water velocity and turbulence.
Performance characteristics of the vortex bed-load sampler are
illustrated by Figures 3, 4, and 5. Figure 3 shows the varying
capacity of the vortex system as a function of river discharge (and
hence indirectly of upstream energy head). No accurate data were
obtained on the tail-water level at the bypass outlet, however, to
relate vortex flow to differential energy head. But the system
appears to be regulated principally by "inlet control" at the entry
to the culvert from the sampling pit. At streamflows of less than
2.35 cfs., the vortex diverted the entire creek [Figure 4). The
strength of the vortex increased considerably with increasing
streamflow and water stage. At intermediate stream depths, a distinct
breaker of white water occurred directly over the downstream edge of
the vortex flume. At highest stages, the stream surface was generally
wavy (suggesting critical-flow conditions) and the breaker was no
longer visible, although a strong vortex could be felt if one stood
in the flume in wading boots. The vortex always exhibited sufficient
strength to transport any material carried into it by the streamflow
(except when the vortex flume was completely filled with gravel
after occasional periods of non-operation during unexpected storm
runoff). From Figure 4 it may be seen that the vortex handled an
increasing quantity but a decreasing proportion of the total flow as
the river discharge increased. For example, at 40.9 cfs (the largest
flow at which streamgagings were conducted both upstream and downstream
of the weir/trap structure for the purpose of developing a rating
curve of the vortex system) 8.1 cfs or 20 percent of the streamflow
was diverted through the vortex. During discharges exceeding 150 cfs,
the vortex flow was estimated to not exceed 15 cfs (i.e., 10 percent
72
-------
top
Figure 3. Discharge capacity of the vortex bed-load sampler.
73
-------
Fif.ure 4. Flow across weir/trap structure during operation of vortex
bed-load sampler.
74
-------
J L_!_l_L-H
I i : i r n
• -I I '"t~T
L_J I—H-L
i^3-BjJ4,J_-U
-1 p- y^ _ 1
! I- ' ; ! Mi U
r I . ''""I " i M i :
j...u:-J... -
Figure 5. Variation of Froude number \s'ith discharge during operation of the vortex
bed-load sampler.
-------
8~ co^crftm floor ^^> •$»•/ cnomtll Iv -^ ' \
6
6'conemf* fieor Q
V
!
5-0- 1-
Ayytyatt
„... .1, t .... ,l|
3?QfO
.1
_i
- — ro" —
,£!•« //«5.
— •• T-r .'•;—. -
u
C. PROFILE OF CHANNEL EXIT
Figure 7. Detail views of concrete channel.
-------
A. PROFILE VIEW
Basic, dafo -fo
=7',, depth - ^'} slop
if protc.cf-ion
coo, c Gr-ei vey
" O Cl O • —
fc n g° 0° f O
B. PLAN VIEW
Figure 6. Schematic views of instrumented concrete channel.
-------
or less of the total river flow). For comparison, Robinson (1962)
indicates a flow removal of from 5 to 15 percent of the total flow as
a criterion for successful operation with sand.
In Figure 5, the variability'of Froude number (the mean flow velocity
divided by the square root of the product of gravitational
acceleration (g) and mean depth of flow) with streamflow is shown.
The Froude number is calculated for the cross-section just upstream of
the vortex flume. A shift in the location of the control point for
the stage-discharge relationship occurred during the night of January
16 and accounts for the different lines in Figure 5. During periods
of bed-load transport, the Froude number ranged from 0.6 to slightly
less than 1.0, which indicated sub-critical flow approaching the
vortex trough. The shift in control point and corresponding reduction
of Froude number at a given discharge caused a change in trap
efficiency for sand-sized particles, as will be discussed later.
Instrumented Concrete Channel
A short distance upstream of the instrumented "natural stream" research
area, a 157-foot concrete research channel lies across a large meander
loop of Oak Creek. The channel is three feet wide by three feet deep.
A diversion structure at the upper end of the reach permits control
of the flow entering the concrete channel. This facility permits
controlled experiments of bed-load transport and related phenomena for
placed streambeds of selected material and for simulated storm runoff
and other regulated flows.
Schematic representations of the research facilities in this reach of
Oak Creek are shown in Figure 6. Details of the concrete channel are
presented in Figure 7.
The diversion structures consist of wooden plank dams supported
between pairs of steel channels embedded in concrete. The structure
nearest the channel (see Figure 6) rests on a low dam of broken
concrete riprap sealed with a plastic sheet. Low ground between the
diversion structures is similarly dammed by broken concrete and a
plastic sheet. A trash rack upstream of the diversion structures
prevents branches and logs from entering the concrete channel.
Stoplogs may be placed near the upstream entrance to the concrete
channel to exclude or regulate streamflow through the channel. (The
diversion structures also provide such regulation, but with much less
control.) A short distance downstream, additional stoplog slots
accomodate a sharp-edged weir of adjustable height. A stilling well
alongside the concrete channel indicates water levels upstream of the
weir by means of a pipe connection through the channel wall. A point
gage in the stilling well allows measurement of water levels for use
with a rating curve to obtain the discharge rate through the channel.
Gravel beds of selected material sizes are placed in the channel over
78
-------
lengths that may be varied by end walls at different stoplog slots
along the channel length (e.g., see the profile view in Figure 6).
The tailwater level and depth of flow over the gravel also can be
regulated with stoplogs in the slots near the downstream end of the
gravel bed. Although the slope of the concrete channel is fixed at
0.027 feet/foot, the initial slope of a gravel bed can be varied
considerably by its manner of placement in the channel.
Bed-load transport is measured at a sediment trap at the downstream
end of the gravel bed. A sampling box fits snugly into a trap area
between two sets of stoplogs to accomplish this. The sample box
can be removed by a chain hoist and pulley attached to a hoist frame,
as at the downstream reach, and the sample collected for subsequent
analyses.
Gravel can be fed to the streambed at its upstream end at a steady,
adjustable rate. This is done by a V-shaped supply trough above a
power-driven conveyor belt. A gravel storage area alongside the
trough permits replenishing the supply material as required.
Elevations of the gravel bed and water surface are determined with
movable point gages. Stations of known elevation at 1-foot intervals
along the full length of the channel facilitate these measurements.
79
-------
SECTION V
EXPERIMENTAL PROCEDURES AND METHODOLOGY
The general procedures and methods followed in conducting
experiments on sediment-yield processes at Oak Creek are described
briefly in the following paragraphs. Where applicable, standard
techniques were followed and are, therefore, not explained in great
detail.
Streamflow
Continuous records of stream discharge were needed to convert individual
observations of sediment transport rate into estimates of sediment
yield. A water-level recorder installed in a stilling well near the
weir/trap structure gave a continuous record of water stage, which
was related to discharge through a series of streamflow gagings by
conventional measurement techniques using standard and small current
meters (Gurley Price, Gurley Pygmy, Ott Cl Small, and Neyrpie Midget)
at gaging sites near the stilling well. From these observations a
rating curve was developed so that continuous streamflow data could
be obtained. Streamgagings were conducted over as full a range of
discharges as possible and with sufficient frequency to detect any
changes in the stage-discharge relationship attributable to shifting
control points for the flow.
Supplemental measurements were obtained concurrently with several of
the stream gagings. Among these were observations of the effect upon
water stage of the operating status of the vortex bed-load sampler
(whether open or closed). Also, simultaneous or nearly concurrent
measurements were made at the downstream end of the weir/trap
structure and at the gaging station just upstream to determine the
vortex discharge at various streamflows.
Hydraulic gradients along the instrumented reach of Oak Creek were
determined at the times of most Streamgagings. This was accomplished
from observations of water stages at seven staff gages along the
channel spanning sub-reaches with differing hydraulic characteristics.
The energy gradients along the instrumented reach could also be
determined from the data on water surface slope and from estimates
of velocity head based on streamflow data.
Suspended Load
Suspended sediment loads were sampled by two techniques--periodic
hand sampling and composite automatic sampling. Individual samples
were collected near the stilling well by means of a DH 48 hand
sampler. These samples were integrated with respect to depth and
lateral position in the stream cross-section, but were discrete with
respect to time. For composite automatic sampling, a specially
constructed sampler was used, patterned after a similar sampler
developed at the Forestry Sciences Laboratory, U.S. Forest Service,
80
-------
Corvallis. A sampling tube extended from a framework, which held it
in the stream, to an instrument house, which contained a. pump,
solenoid-activated valves, sample container, timer, power supply,
and switches for controlling sampling frequency. Upon activation,
the pump removed all "old" water from the sampling tube before the
sample was taken. A float in the stilling well was connected to a
counterweight in this instrument house so that different switch
circuits were opened or closed as water stage changed, thus varying
the sampling interval between 10 minutes at highest stages and 12
hours at lowest stages. All samples for any period of days were
composited in a 20-gallon bottle.
The collected samples were analyzed for suspended solids concentration
using filtration with 0.45 micron Millipore filters. Standard
procedures were followed.
Bed Load
The field procedures followed to obtain bed-load samples with the
vortex bed-load sampler evolved with experience in operating the
sampler. At first, the sampler was operated continually. However,
the samples occasionally could become completely filled at times of
abrupt streamflow increase if no operator was present (e.g., during the
middle of the night). On such occasions, the rate of sediment
transport into the trap was uncertain. Furthermore, considerable
effort was required to completely clean the sampler before further
sampling.
The procedure eventually followed was to keep the vortex gates closed
at all times except when obtaining bed-load samples. At such times,
the vortex flume and backup trough first were checked and any sediment
deposits cleaned out. The trap gates then were opened for a variable
sampling period. When the streamflow was high, several minutes of
sampling were sufficient to nearly fill the sample box, at which time
the gates again were closed. Times and water stages at the stilling
well were noted upon opening and closing the gates. As soon as the
partly filled sample box had been removed and an empty box put in its
place in the sampling trap, the gates would be opened again. Any
material deposited in the vortex trough during the brief period when
the gates were closed would be included in this subsequent sample, with
respect to both the sediment amount and the time period for the
subsequent sample. In this respect, then, the bed-load sampling could
be done continuously, if desired. When the bed-load transport rate
was low, the material collecting in the sample box would be checked
periodically for size and amount before deciding whether to close the
gates for sampling. Ideally, samples were sought which would
correspond to a fairly constant stream discharge during the sampling
interval, and the sampling interval desired would be a small segment of
time on the rising or falling limb of a runoff hydrograph.
Samples of bed load were carefully transferred from the sample boxes
81
-------
to 5-gallon buckets and given identity numbers before being taken to
the laboratory. On one occasion where the vortex had been left open
overnight and the entire sampling trap had filled with sediment, the
material was randomly split and only a known fraction of the sample
was taken to the laboratory for analysis.
In the laboratory, the entire sample first was ovendried at about
150°C (the controls on this large oven could not be adjusted to lower
temperatures). Then the dried sample was processed through a series
of sieves, using mechanical vibration, to determine both the total
dry weight of the sample and the particle size distribution of the
sample (based on dry weight of sample retained between successive
sieves). The sieve screens conformed to the U.S. Standard Series in
the size of the mesh openings. A listing of the complete set of
sieves used is given in Table 1, together with the equivalent
diameter of the smallest particle retained on each sieve. Sieving
was facilitated by first passing the dried sample through the sieves
ranging from No. 4 to 2 inches in mesh. The fraction retained on the
2-inch sieve was then checked for material coarser than 3-inch and
4-inch sizes. Similarly, the fraction passing the No. 4 sieve was
analyzed with the series of screens from No. 8 to No. 200. If the
amount of material passing the No. 4 sieve was too great to be held in
the smaller screens, this fraction of the sample was first split before
resieving.
Bed Materials in Instrumented Reach
Samples were collected of surface and subsurface sediments in the
streambed to characterize these materials. The majority of samples
were obtained in several locations where channel cross-sectioning
stations were established. Others were obtained away from the stations
if some distinct feature existed, such as the rapid local down cutting
of a gravel bar, which left a vertical cut through the bar upon
recession of streamflow. Many of the samples consisted of collecting
the surface layer of sediment. However, in some locations sets of
samples corresponding to adjacent vertical zones were obtained.
Laboratory analyses of the samples included particle-size distributions
of the bed material, based upon sieving the ovendried samples as
already described, and weight distributions of individual particles
within each sieve-size range. The shapes of individual particles
within each sieve-size range were determined by measuring the major,
minor, and intermediate diameters of each particle with calipers.
Specific gravity tests were made on a composite sample of streambed
particles to determine the representative specific gravity of bed
material from Oak Creek and any possible variability in this value
as a function of particle-size range. Specific gravities of the
smaller particles were determined by standard procedures with a
pycnometer, or volumetric flask. Gravel coarser than No. 4 sieve in
82
-------
Table 1
SIEVE SERIES AND EQUIVALENT PARTICLE
DIAMETERS FOR ANALYSIS OF SEDIMENT SAMPLES
IN STUDY
Sieve
Size
4 -inch
3 -inch
Z-inch
1 1/2-inch
3/4-inch
3/8 -inch
No. 4
No. 8
No. 16
No. 30
No. 50
No. 100
No. 200
Approx.
Inches
4
3
2
1.5
0. 75
0.375
0. 1875
0. 0937
0. 0469
0. 0234
0. 0117
0. 0059
0. 0029
size of sieve opening
Millimeters
101.6
76.2
50.8
38. 1
19. 05
9.52
4.76
2. 38
1. 19
0. 595
0. 297
0. 149
0. 074
(1) U.S. Standard Series sieve sizes used.
83
-------
size was analyzed for specific gravity using a wire basket suspended
in water and following accepted procedures,
Changes in Channel Topography
During the initial field surveying and mapping before the development
of sediment research facilities, several control points of known
elevation (arbitrary datum) were established, some of which were
adjacent to the instrumented reach of Oak Creek. During development of
the facilities, pairs of cross-sectioning stations were established at
25-foot intervals along the axis of the stream. Metal pins were
embedded in the ground and held with concrete at the outer ends of
each cross-section for permanent reference. Subsequently, additional
cross-sections were established half-way between the existing ones,
giving a 12.5-foot average spacing of sections. The purpose of the
cross-sections was to provide fixed locations for periodic resurveying
of the channel to map the streambed and banks, and determine any changes
in shape because of erosion and sedimentation.
Such mapping of the channel was conducted on several occasions using
conventional surveying techniques. Elevations were obtained at all
"break-points" in shape across each cross-section. A surveyor's
level, leveling rod, and steel measuring tape were used. At the time
of the first detailed survey, elevations were obtained at numerous
points between cross-sections for the purpose of completing a detailed
map.
Depth of Scour and Deposition
The depths of streambed scour and deposition during floods were studied
by means of buried "scour" balls as well as through the cross-sectional
surveying already described. The scour balls were of two types:
punctured ping pong balls of neutral buoyancy and buoyant stryofoam
balls of the same size as ping pong balls. Each type was installed in
vertical arrangement in the streambed and the point of burial
referenced to the metal pins of nearby cross-sections for later
relocation.
To place the balls in the streambed, a solid rod and concentric
hollow pipe were first driven into the bed to a depth of over one foot.
The rod was then removed, leaving only the pipe sleeve. For the
punctured ping pong balls, a fishing weight with attached colored
line was dropped down the pipe such that the line would trail out in
the water to facilitate later relocation of the buried balls. The
numbered balls were then sunk into the pipe until they extended from
the top of the fishing weight to the top of the streambed. The
depths of the bottom and top balls were noted so that the positions
of all balls in the column (about a dozen balls, typically) could be
determined. The pipe sleeve was then removed from the bed and a
small amount of sand placed over the top ball at the streambed surface.
During floods, any scour that occurred would remove one or more of the
ping pong balls, depending upon the depth of scour, which would
84
-------
then move downstream toward the vortex trap along with other bed-load
material.
The styrofoam balls, because of their buoyancy, would rise to the
stream surface at that instant when scour developed to the depth of
burial of such a ball. Each ball had a separate fishing weight as
its anchor. The anchors were all placed in the bottom of the pipe
sleeve and covered by some sediment. Individually numbered balls
were buried with some gravel and much slack fishing line between
them so that one ball rising to the water surface would not pull the
other balls out of the hole. During floods, if scour was sufficient
to disturb any of the styrofoam balls, they would pop to the surface
and an observer making a periodic inspection of the test reach
would, hopefully, see the balls and note their presence and
identification number.
The depth of maximum scour during flood periods and any subsequent
redeposition as flood strength weakened were both determined. To
accomplish this, the collection of scoured balls as described above
was supplemented by relocation of the burial points during low water
between floods and excavation down to the top ball. The depth of
excavation indicated the depth of redeposition and the identity
number of the top remaining ball indicated the depth of maximum scour.
The balls were then carefully covered to the pre-excavation level.
Painted-Gravel Experiments
Painted gravel was used to study the distance-of-travel and time-of-
travel of large bed-load material. The gravel used for these experiments
was taken from the analyzed material collected from Oak Creek during
early bed-load sampling. The sizes of the particles ranged from 3
inches to 3/16 inch in nominal diameter. A yellow paint similar to
that applied for highway marking was applied by hand brush. The
individual rocks then were numbered and weighed. The prepared gravel
then was dropped into the stream from above the water surface so that
the particles would come to rest on the streambed in as natural an
orientation as they could be placed. During and after periods of
storm runoff, inspections of the stream were made to locate the
painted gravel, and the positions and rock numbers were .recorded.
Some problems were experienced in making these inspections: it was
necessary to wait until storm runoff receded and the water cleared
before the streambed became visible again; the paint tended to wear
off rapidly because of abrasion at the time of bed-load transport; and
painted gravel tended to become buried beneath the surface of the bed
as a result of bed-load transport of these and other particles.
Period of Intensive Field Work at Instrumented Reach
In the fall of 1969, preparatory work was completed for winter
sediment-yield observations. Detailed mapping of the channel bed and
banks was followed by installation of several scour balls. All
sediment sampling apparatus was set up, as well as related equipment
for obtaining precipitation and streamflow information.
85
-------
Sediment was sampled during December, January, and February when
streamflows were sufficient to disturb the streambed. Depth of scour
and deposition and of changes in channel topography also were
determined periodically during this period. By April, spring streamflows
clearly were no longer sufficient to transport gravel, so no further
field work was done at the instrumented reach.
Research Activities at the Concrete Channel
Graded river-run gravel from the Willamette River was placed in the
concrete channel to create a plane gravel bed of fairly uniform
particle size. The commercial gravel used ranged from 3/4 to 1 1/2
inches, although some finer and coarser material was present.
Movement of this artificial streambed was studied under a range of
transient and steady-state flow conditions. Particular focus of
these initial experiments was upon incipient motion of the particles,
scour patterns, and tendencies for any development of a natural
protective "armor" layer at the surface of the bed. This information
was sought to give greater insight to the bed-load observations made
in the natural study reach and the basic mechanisms involved.
The data obtained during the late winter and early spring of 1970
are rather tentative in their applicability at the moment. Several
additional experiments are required when streamflow again becomes
sufficient to transport gravel. Therefore, no firm conclusions have
been drawn from the data obtained in the concrete channel. Because
of this, any further discussion of experimental procedures and
methods at the concrete channel is omitted from this report.
86
-------
SECTION VI
HYDROLOGIC CHARACTERISTICS OF OAK CREEK STUDY REACH
The Oak^Creek Drainage Basin
The Oak Creek Basin used for this study lies almost entirely in
McDonald Forest, which is managed by Oregon State University's
School of Forestry. The forest lands lie on the eastern edge of the
Coastal Range. The drainage basin above the weir/trap structure is
outlined in Figure 8, which also gives the general location of the
basin with respect to Corvallis and Oregon State University.
The drainage area tributary to the weir/trap structure is approximately
2.8 square miles. This land ranges in elevation from 480 feet (MSL)
at the structure to 2,178 feet at the highest ridge fringing the
basin.
Most of the drainage basin lands are covered with young-to-mature
Douglas fir. Various other tree species occur, including Oregon
white oak, grand fir, western yew, western redcedar, western hemlock,
Oregon ash, red alder, and bigleaf maple. Some clearcutting, thinning,
and replanting have taken place on the watershed in the past. Natural
open meadows also occur in the basin.
The soils of the drainage basin vary in depth from a few inches or
less in steep areas to several feet in the alluvial deposits along the
lower portions of Oak Creek near the research facilities. Infiltration
rates are quite high because of a well-developed forest litter and
root system in all vegetated areas.
The subsoil zone consists of fractured and weathering bedrock, whicli is
principally sedimentary and basaltic in origin. Outcroppings of this
fractured and weathered bedrock along the channel provide the source
material for bed-load transport in Oak Creek. (The alluvial portions
of the lower part of the basin allow long-term storage of this material
as the channel changes in position over the years.)
Precipitation During Field Studies
During sediment sampling, Oak Creek streamflow was very responsive to
changes in rainfall intensity at times when the watershed was thoroughly
wet from antecedent rainfall. This sensitivity to rainfall was
particularly evident during part of December 1969, most of January 1970,
and part of February 1970, because of the heavy rainfall over several
days in those periods. These long "steady" rains exhibited considerable
variation in intensity, which caused the stream level and discharge
to fluctuate considerably during many 24-hour periods.
The seasonal variation in rainfall at the Oak Creek drainage basin
can be seen from Table 2. Data from the Hyslop farm northeast of
Corvallis have been included in this table to extend the Oak Creek
data back into early fall, 1969. The two precipitation stations are
87
-------
KEY
r
r
r J
I
i
"~l
Oov.wDAC.x-
CORVKUJS
Figure 8. Location map for Oak Creek sediment research facilities,
-------
Table 2
MONTHLY RAINFALL TOTALS DURING FIELD STUDIES
Month
Aug 1969
Sept
Oct
Nov
Dec
Jan 1970
Feb
Mar
Apr
May
Corvallis
Precipitation
Trace
3.62
3.91
2.86
11. 59
15.51
5.97
2. 29
2.66
1. 12
Oak Creek
Precipitation
12. 08
17.45
5.28
2.46
3.91
1.62
{1} Measured at Hyslop farm near Corvallis; elevation 225'.
(2) Measured adjacent to research facilities; elevation 490'.
89
-------
generally similar in broad rainfall characteristics, as can be noted by
a comparison of the concurrent records from December through May.
Most of the rainfall through early December was intercepted by
vegetation or utilized on the basin to recharge soil moisture.
Similarly, much of the rainfall on the basin after the end of February
was used to recharge soil moisture or lost to the atmosphere by
evapotranspiration.
Daily precipitation values near the basin from December through
February are shown in Table 3. Data from the Hyslop farm were used
because of the consistent time of daily observation. The daily
rainfalls at Hyslop farm did not differ importantly from those
observed at the Oak Creek gages .
Two major storm periods occurred in December 8-14 and 20-23. In the
first period, 5.64 inches of rain fell at Oak Creek rain gages from
noon on the eighth until 5 p.m. on the fourteenth, and considerable
sediment transport took place in the stream. During the second
storm, an inch of rain fell at Oak Creek during the night of the
twentieth and morning of the twenty-first and also caused considerable
sediment movement in Oak Creek.
January precipitation in Corvallis was the greatest in 89 years of
record. By inference, this also was true at Oak Creek. Bed-load
transport occurred in the stream on several occasions because of
particularly heavy, extended rainfall. The more important periods
included 8-9, 12-14, 15-20, and 22-27.
In February, the principal storm resulting in bed-load transport in
Oak Creek caused heavy rainfall on 15-16.
The variability of rainfall intensity during a storm period and its
resulting effect upon Oak Creek discharge are shown in Figure 9. The
period 12-20 January is illustrated because of the large streamflows.
The rate of change of cumulative rainfall at the recording rain gage
(the slope of the curve at any point of tangency) gives the intensity
of rainfall. Rainfall intensity was quite changeable, and during
low-intensity rainfall the streamflow tended to recede. Each period
of sustained high-intensity rainfall was marked by an abrupt rise in
streamflow with corresponding effect on sediment movement. Rainfalls
of about 3/4 inch in less than four hours on January 16 and 19 led to
the two maximum streamflow peaks of the winter season.
Streamflow During Field Studies
Rating curves were developed for water stage and discharge at the Oak
Creek weir/trap structure, based on numerous streamgagings at differing
flows. Instead of a single rating curve, two rating curves were
required, because of a shift in control point during the flood on
January 16-17 (Figure 9). At that time, the bed-load transport past
the vortex sampler was so great that continuous operation of the
90
-------
Table 3
DAILY RAINFALL TOTALS DURING WINTER 1969-70
Date
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Total rainfall, inches,
December
-
-
-
- 19
. 01
. 12
.09
.29
.57
.22
.42
2.37
.35
1. 17
.02
.22
.24
.41
. 14
.35
1.30
. 02
1. 18
.32
. 11
.59
. 20
.67
.02
-
-
for 24 hrs
January
-
_
_
. 13
_
-
.
. 04
.86
.84
. 03
. 52
.47
.50
.50
.45
1.80
.79
.54
.98
. 57
. 75
1.51
. 74
1.73
.54
1. 02
. 08
_
. 10
. 02
to 8 am of date shown
February
.29
.01
.03
_
. 02
1. 14
. 25
_
_
_
_
_
.20
_
. 05
1.93
1. 73
.21
_
_
_
_
_
_
_
_
_
. 11
(1) Data from Hyslop farm near Corvallis used because of consistent
time of daily observation. Source: U.S. Weather Bureau, E.S.S.A.,
"Climatological Data, Oregon" December, 1969; January, 1970 -
February, 1970.
91
-------
Figure 9. Rainfall and streamflow at Oak Creek weir/trap
structure in mid-January, 1970.
92
-------
sampler was not possible. After the vortex flume and sampling trap
had completely filled with gravel, large quantities of material
moved past the structure and deposited immediately downstream at an
emergency bed stabilization structure. (This structure had been
placed during the December 11-12 storm runoff because of excessive
scour from interception of all bed load by the vortex sampler.) The
numerous storms in January made reshifting the control point back
to the weir/trap structure impractical, and hence two rating curves
were required. Rating curve No. 1 covers the period through January
16, 9 p.m., and rating curve No. 2 covers the study period since
that time. The rating curves are shown in Figure 10.
Before January 16, the vortex trap was frequently open, even when no
bed-load transport occurred. Therefore, a correlation curve was
developed to evaluate the effect of vortex operation upon water stage
at the stilling well. This effect is portrayed in Figure 10 by use of
a dual rating for curve No. l--one line for the vortex operating (open)
and another for the vortex trap closed.
Potential extreme floods were of great interest and concern, both for
sampling and for protection of the research facilities. Therefore,
a flood frequency analysis for the drainage basin was made according
to U.S. Geological Survey procedures and information on Western
Oregon given in Water Supply Paper 1689 (U.S.G.S., 1964). The
estimate is approximate only, as the average annual runoff was unknown
and local orographic effects could cause this and other parameters to
differ in value from those at nearby stations used by the Geological
Survey in its regional analysis. The estimated once-in-50-years flood
is shown in Figure 10 for comparison with the rating curves at the Oak
Creek facilities. Also shown is the estimated mean-of-annual-floods
(or mean annual flood). Both are indicated by a range of discharge
values rather than a single value because of the approximations
involved in their estimate. The largest observed discharge during
the study period, 225 cubic feet per second (cfs) on January 19,
is also noted in Figure 10. This runoff would appear to be on the
order of a once-in-20-years event. This runoff was influenced by very
wet antecedent conditions, which included a peak discharge of 210 cfs
three days earlier, on January 16 (Figure 9).
The hydrograph of mean daily discharges spanning the winter field
study period at Oak Creek is shown in Figure 11. Fall runoff until
December was low and quite stable, as response to fall rainfall was
very slight. Streamflow during the period of no record was similar to
that before and after (the water stage recorder and other equipment were
removed from the field for protection during a brief hunting season).
The periods of major runoff corresponded to those of major storms, as
described earlier. Bed-load sampling occurred during these periods and
streambed surveying took place between these periods. Spring flows
continued to recede with only minor rises after March (not shown in
Figure 11).
93
-------
| LS oo
River discharge, cfs
Figure 10. Rating curves for Oak Creek at the weir/trap structure.
-------
-I
UJ
..'-•
5
>
j
U
loo
feo
o
TO
»*•' 1 1
•»*-r-r
—
rf
....
....
; ;
O
•rf
•-
I »
CT
-«•• ' *
o a
B£J
• I...
in::
-Li*
f
....
-K
1"!
f*l
» (
rftj
f*fi
0 1
" N
q
t . ....
» -2
0 ••
£
P
rm
• • • i
o a
,-'
a
5
Hp
l^fJrt
i
,.
U'.:
':
1 •
t
I
1 "•
1
. 1. . .
• .' -
-r
1
- -
~ 1
D
- 1- ,
fjTJ
* a
CC
" "t
*
'..:
I
3
a
|
:*
;
;
-
.'
n
|p;;:r
L '
{^
r
P
S t
0 1
••7
I
^rl
T'"
S 1
^N
i
i
"
1-
.-
4
1
o ;
b^rr
5}li
g
rt
""" 'U.
Tfr
j
|
u...-
i
W-J
S
t...*
O 1
f
11
1..
fi
•K"
T*
9 a
E£
TC
1
*4"'
0 2
/
i
i
g
tr*
8 1
O
V
£
• i
MS(
5s-
e i
L
:::]
• • i •*
mw
B
Figure 11, Mean daily discharge at Oak Creek weir/trap structure during 1969-70 field studies
-------
The hydrograph of mean daily discharge at Oak Creek has certain
limitations in portraying streamflow variations for so small a basin.
For example, the hour at which peak discharge occurs may greatly
influence the average daily runoff value. More importantly, the
maximum dailydischarges at Oak Creek during floods are much smaller
than the maximum instantaneous discharges at peak flow. Therefore,
the relative magnitudes of flood peaks can be disguised in hydrographs
that are based on daily values. Thus, from Figure 11, the three
greatest daily peak flows of the winter occurred on January 22 and 25
and on February 16, respectively (in particular, they were larger than
the daily flows on January 17 and 19). However, the instantaneous
peak flows on the first three dates mentioned (158, 182, and 175 cfs,
respectively) were smaller than for the latter two floods (210 and
225 cfs, respectively). Hence, daily flow values can be misleading
regarding instantaneous flood peaks.
Segments of the instantaneous discharge hydrographs during periods
of bed-load sampling will be presented in a later section and are not
given here.
Hydraulic Characteristics of Reach
Some idea of the velocities experienced in the Oak Creek study can be
obtained from velocities measured during streamgagings. During large
flows, the velocity was frequently 7-8 feet per second (fps) at 0.2
depth below the water surface and 5-6 fps at 0.2 depth above the
strearabed in water of about 2-foot depth. The largest point velocity
observed was 8.71 fps. Velocities in the test reach away from the
gaging station were smiilar in magnitude, although some pools had
slower flow and one steep portion of the channel had velocities that
may have exceeded 10 fps on occasion. Velocities also were locally
high in the vicinity of tree-root obstructions at the channel banks
during times of flood runoff.
The mean velocities at the gaging station cross-sections are shown as
a function of discharge in Figure 12. By extrapolation, the largest
average velocity experienced at these sections during the study period
is estimated to have been about 7.5 fps, for which the point velocities
may have approached 10 fps near the channel center and water surface.
In Figure 12A, the natural channel upstream of the weir/trap structure
exhibits a flatter velocity-discharge curve than at the structure
(Figure 12B). This is attributable to the widening of cross-section
with increasing stage at the upstream location, whereas the sidewalls
of the weir/trap are vertical. The shift in control to a more
downstream position on January 16 is reflected in Figure 12B by lower
velocities at a given discharge after that date. This is because of a
greater depth of flow at the structure afterwards caused by the new
backwater conditions extending upstream from the control point. The
effect of vortex operation upon nearby velocities can also be seen from
Figure 12B. With the vortex sampler operating, the flow becomes
96
-------
-t-wp
Figure 12. Variability of velocity with discharge at Oak Creek streamgaging sections.
-------
shallower and swifter near the vortex trough. Froude numbers for this
flow were presented earlier, in Figure 5.
Additional hydraulic properties at the gaging section upstream of the
weir/trap structure are presented in Figure 13. Because channel width
only increases slightly with discharge at the gaged flows, the mean
flow depth and the cross-sectional area are related by a nearly
constant factor. At flows greater than gaged at this section, the
cross-section widens abruptly so that the logarithmically linear
relationships of Figure 13A and 13B would become curved at higher
discharges.
Manning's "n" at this cross-section (a measure of channel roughness)
was about 0.035 and appeared to decrease slightly as water stage and
discharge increased. This variation is attributable to differences in
bed and bank composition. The local streambed consists of gravel
across the full bottom width of the channel, but the banks are a
mixture of finer soils with some gravel and offer a smoother surface to
the flow than does the streambed.
The overall channel slope for a 2,600-foot reach of Oak Creek that
includes the study area was 1.81 percent, or 0.0181 feet/foot.
Considerable local variability was noted over short distances, because
of numerous riffles and pools. The stream profile in this part of
the creek did not exhibit any concavity, but instead had irregular
variations. The stream profile was determined by means of a survey
line along the main thread of low-water flow.
The average hydraulic gradient across the instrumented test reach during
the study period was 1.39 percent (0.0139 feet/foot), or somewhat less
than that for the 2,600 feet of channel spanning this reach. This
figure was based upon observations on 20 occasions during streamgagings
which covered a large range of discharge (3-164 cfs) and included
rising, falling, and steady flow conditions. Observations were made
at seven staff gages located along the channel. The spacing of staff
gages and the mean hydraulic gradients observed between staff gages
are summarized in Table 4. The mean hydraulic gradient for the steep,
straight subreach between staff gages 6 and 7 was more than twice as
great as the mean gradients of flatter subreaches 1-2 (straight) and
5-6 (winding somewhat around tree roots along the subreach). Extreme
variations in hydraulic gradient were generally within 30 percent of
the mean value for the subreaches. Over the entire reach, the
hydraulic gradients observed did not deviate by more than five percent
from the mean value. At very low flows, numerous riffles were exposed
along the test reach and caused local drops in the hydraulic gradient.
As noted earlier, velocities were variable along the channel. Therefore,
the energy gradients along the reach for subreaches were not identical
with the hydraulic gradients given. Nevertheless, mean velocities
were sufficiently similar so that the hydraulic gradients offer
98
-------
: A.
so
oo
10
: B.
So
w ^
11
v
0- xo
*
al
.04-
.OZ -
0 I
-i—I . I rm
I C. MANNING^ r\
-4 , 1—|- ) I I
MEAVi fLO'(V
-I h
i i i t i i
-\ 1 1-
-3. 3
K>
•30 To 5o loo
CSr S
i i
300 5^50
ul
U-
0,3
Figure 13. Variability of hydraulic parameters with discharge
at Oak Creek streangaging section upstream of weir/trap structure.
99
-------
Table 4
MEAN HYDRAULIC GRADIENTS FOR STUDY REACH
Staff
Gage
Number
1
2
3
4
5
6
7
Flow Distance
to weir/trap
feet
0
95
160
211
281
362
434
Incremental
Length,
feet
95
65
51
70
81
72
Mean Incremental Mean Hydraulic
Hydraulic Gradient, Gradient
feet/foot feet/foot
0.0087 ""
0.01 24
0. 01 71
\ 0. 01 39
0. 01 66
0. 00 92
0. 02 17
100
-------
reasonable estimates of the energy gradients for the subreaches and a
very good estimate of the mean energy gradient for the entire study
reach.
101
-------
SECTION VII
CHANGES IN CHANNEL TOPOGRAPHY
Organization of Data Collection
Changes of channel shape caused by scour and deposition in the study
reach of Oak Creek were detected through a sequence of surveys at
cross-sectioning stations and intermediate points. The surveying
included streambed elevations and bank elevations extending well
above the highest water level during the winter, so as to note any
bank caving that might have occurred. Survey data on net changes of
channel topography were supplemented by information on maximum depth
of scour during the interval between surveys, as determined from
buried scour-ball devices.
The locations of metal reference pins and cross-sections used to
evaluate changes in stream topography are given in Figure 14. Stations
identified by number only were established in early fall, 1969, and
station numbers followed by the letter A were added at the end of
December.
The positions of buried scour balls are also shown in Figure 14. These
positions were chosen randomly to avoid any human bias in their selection
(a coin-flipping system was developed for selection of position along
and across the channel). Most of the scour balls were installed in
fall, 1969. A few additional installations were made in February, 1970.
Figure 14 also shows the locations of the seven staff gages used for
determining water levels and hydraulic gradients for the subreaches.
Approximate edge-of-water lines for frequent high streamflows are
indicated, although at the highest streamflows observed there was some
overbank flow along the left bank from station 11A to 16A (up to eight
feet farther from the stream center than shown in Figure 14) and flow
in a higher overflow channel behind the right bank from Station 12A to
15.
At low streamflows, numerous riffles were exposed. These changed in
position somewhat as a result of intervening storm runoff, as did the
location of the low-flow channel. The low-flow channel as it appeared
in early fall, 1969, is shown in Figure 14.
Data collection was organized to span periods of storm runoff. The
specific dates and the intervening runoff-conditions are summarized in
Table 5. Initial topographic surveying and mapping and the installation
of scour-ball devices was completed in the fall, before stream rises
of any significance occurred (Figure 11). A considerable amount of
sediment transport took place during the storm runoff of December
11-12, and a topographic survey was carried out on December 21, after
recession of streamflow from that storm period and as rains began to
cause a new rise in streamflow. After this new storm runoff and a
lesser storm immediately thereafter, another topographic survey and a
102
-------
«« ftftun.ee Pm,LiM<9 urn u«a
D >~>««* FOB Bu«'i
-------
Table 5
TIMES OF COLLECTING DATA ON SCOUR AND
CHANGES IN CHANNEL TOPOGRAPHY, 1969-70
Date of Date of
Cross-Section Scour Ball Remarks
Survey Survey
5 Oct Nov Fall baseflow period
21 Dec Survey after single
large storm-runoff
event
31 Dec 31 Dec Survey after two
storm-runoff events
31 Jan Survey after numerous
storm-runoff events
14 Feb Survey after numerous
storm-runoff events
Zl Feb 21 Feb Survey after single
large storm-runoff
events
24 Mar Supplemental survey
104
-------
survey for scour-ball devices was conducted on December 31. The
continual storm runoff during January prevented any topographic
surveying until January 31, and was followed on February 14 with a
survey for scour balls buried in the strearobed. The storm runoff of
February 16-18 was followed by additional topographic and scour-ball
surveys. Supplemental topographic details were obtained on March 24,
by which time it was clear that no additional bed-load transport
could be expected for the season.
Net Seasonal Changes
The net changes in channel topography caused by storm runoff during
the winter bed-load transport season can best be examined by a
comparison of maps prepared from the October and February-March
topographic surveys. Contour maps for these surveys are presented in
Figures 15 and 16, respectively. The contour interval selected for
presentation is 0.1 foot (working maps of survey data were first
prepared with a 0.5-foot contour interval).
Pools or "holes" in the streambed can be identified in Figures 15 and
16 near channel bends and constrictions caused by tree root systems.
One pool also occurs near section 2A in an otherwise wide, straight
reach (Figure 15). This is caused by a large fallen branch wedged
against the upstream edge of a tree and partly buried in the streambed,
so as to act like a weir and also to cause debris to lodge occasionally.
Riffles and shallow flow occur along the central portion of the channel
between the pools except at higher stages, when even these shallow areas
are covered by over two feet of water.
In Figure 17, the topographic maps of Figures 15 and 16 have been
superimposed and the net seasonal changes determined. Supplemental
data from cross-sectioning surveys and scour-ball devices also have
been incorporated in Figure 17. The approximate channel width over
which streambed disturbance and bed-load transport took place has
been estimated and indicated in this figure. Zones where net scour or
net deposition occurred over the winter are shown, as well as zones
where bed load was transported but no net change in streambed elevation
was detected.
The importance of tree root systems as stabilizing agents for the
channel platform can be noted from Figure 17. Near sections 6-7, tree
roots have maintained an outside edge of the low-flow channel in spite
of high-velocity flood flows that strike the trunks and send "bow-waves"
downstream. Even more notable is the narrow, deep channel between
tree-stabilized banks in the vicinity of sections 12-14, where no
bank scour was evident. The limited deposition that occurred in the
pools here may have been caused by a raising of backwater because of
accumulation of gravel at the bend downstream (particularly near
sections 11A-12).
Cohesive banks offered a great deal of channel stability in a manner
analogous to tree root systems. For example, upstream of section 16
105
-------
o
cr>
c»csi-
• TkEc. TKUNI^ OK
icxTCCR LINC A\NAT
PIN »WD SECTICH
MA.N C
O Zo 4O
. I , .„
I/I /I-
f7/?<>
"*t /-
" i.'Vv! »'» • ..
J««°
J««,
O l»«
V
••
-N V
* I. ,a\ • •• *'. i* A
'•""AM\.- .•
>}w^-
"XxX'^ rs,
W-
Figure 15. Channel topography in October 1969, before winter bed-load transport.
-------
'c* CRPiS -SECTIONING RE.t'GKENCt. PIN AUD SECTION NUMBEB
• TKSE TK.UNT; OR -STUMP
let, CC^TOUt LINE ALON6 CH^NNEL
LINE Awfty re.cs^ MAI
15 fllll - r. .* \ I
fy ^v.vls
*-iS rJ1\V. ,.,
AV\ \!*\ v\ '
1 *\\ X'. V V
Tigure 16. Cha.r:nel topography in late February 1970, after winter bed-ioad transport
-------
L\Mn Of StOMENT
'•?& ZONE OF NET 'oEhSONM- DetcSrrc
\>s>- ZONE OF NET •scA-bowAt. scooir
AND TR/W5POCT
SCALI. rtrr
ill •"
"»• !;v.£»
Figure 17. Net seasonal changes in channel topography, October 1969 to late-Februarv 1970,
due to winter bed-load transport.
-------
the banks are more cohesive than they are from section 16 downstream.
As a consequence, bank scour was minimal between sections 16 and IS,
even though considerable scour of the gravel streambed took place
(streambed degradation appeared to take place in an upstream-progressing
direction here). The bank scour that did occur took place adjacent to
the bed, rather than higher on the steep bank, and was presumably
caused by abrasion by gravel rather than shear forces exerted by the
water. Upstream of the study reach, local zones of cohesive material
extend across the streambed and appear to be quite stable in spite of
frequent scouring and abrading by storm runoff and bed-load transport.
Near section 14A-15 a weak right bank (looking downstream) which had
been subject to erosion before the winter of 1969-70 was subject to
severe caving during this winter as a result of additonal erosion.
This resulted after debris clogged a trash rack (1/2 inch pipes of
1-foot spacing) near section 14 in mid-December and was subsequently
washed out. When the debris dam formed, some local scour occurred at
the trash rack and more than a foot of gravel deposited a short
distance upstream in the backwater of the debris dam. (The trash rack
had been installed to keep leaves, bark, and branches from entering the
vortex bed-load sampler, with the intention of cleaning the trash
rack frequently during early winter storm runoff to prevent its
clogging. However, the overnight runoff of December 11 caused
clogging when no one was present.) The new gravel deposit forced
erosive flows against the already weak bank. The trash rack was then
removed and subsequent local scour after mid-December removed not only
the recent gravel deposit but also much of the caved bank.
The channel bend between sections 8 and 12 was subject to considerable
modification during the extensive bed-load transport of January. The
point bar on the inside of the bend built outward as well as upward at
this time (over a foot of fill—see data on scour balls at location
11-1, in Table 6). Some bank scour on the inside bank occurred but
the most pronounced scour took place at the outside of the bend, as
deposition on the point bar forced the current to erode the outer
bank. The effects of deposition and scour at the bend apparently
extended downstream for a short distance, causing an altered flow
alignment. This is shown by some deposition in the main channel and
some scour in an "overbank" portion of the stream within the higher
banks between sections 6 and 8A.
'[he lower straight section of the study reach reflected some tendency
toward non-uniformity of cross-sectional depth above section 3A and
a tendency toward greater uniformity of cross-sectional depth
downstream of section 3A. The actual depths of net scour or
deposition were not very great (see cross-sectional figures, in the
following section).
Sequential Changes at Cross-Sections
The cross-sectional surveys at each of the sections on the dates
109
-------
indicated in Table 5 give a description of the progressive changes in
channel topography. By comparison of the surveys at a section,
additional insights to the net seasonal change and perhaps some further
idea of the processes involved can be obtained.
The cross sections have been grouped for presentation, based upon the
observed net seasonal changes indicated in Figure 17. Figure 18
shows changes in shape at sections 1 through 4, all in the straight
lower end of the study reach. Figure 19 and 20 show changes at
sections 5 through 8A, where the flow winds slightly among trees.
The cross sections in Figure 21 (section 9 through 11A) describe
changes at the large bend midway in the study reach. Sections 12
through 13A, in Figure 22, show the cross-sectional changes which
took place in this deep, narrow pool zone. In Figure 23, the changes
are given for sections 14 through ISA, near some extensive bank
caving. Finally, Figure 24 indicates the changes of shape at
sections 16 through 18, in the steep, narrow, upper portion of the
study reach.
The accuracy of the cross-sectional survey work presented in these
figures should be mentioned at this point. The accuracy was about
0.1 foot for horizontal and vertical positioning. Occasional
difficulties in identifying breaks in side slopes because of vegetation
were compensated for by comparison of successive field survey notes.
The elevations of adjacent points in the streambed varied by as much
as 0.2 foot, depending upon whether the leveling rod rested upon or
alongside some of the larger gravel. Therefore, the rod was placed at
what appeared to be the mean level of the bed surface whenever such
problems arose.
Sections 1 (Figure 18) and 1A (not shown) exhibited little change in
bed elevation between surveys, although a slight net seasonal accretion
occurred (as also noted in Figure 17). Sections 2 (Figure 18) and 2A
(not shown) experienced localized scour and deposition which, at a
given spot, seemed to alternate from one survey to the next. As at the
downstream sections, the total net change was not very large. Section 3
(Figure 18) underwent scour over its full bed width during the early
December storm runoff and lesser alternating scour and deposition
thereafter, for a modest net seasonal scour. The sequential changes
observed at section 3A (not shown) were even slighter. Sections 4
(Figure 18) and 4A (not shown) exhibited alternating scour and fill
from one survey to the next, with heaviest scour during the January
runoff.
In Figure 19, the changes of shape at section 5 were small in each
period. At section 5A the central part of the channel had a net
January scour of up to 1.2 feet and February fill of 0.9 feet as part
of a net seasonal scour of only a few tenths of a foot. Appreciable
(1-foot) scour at the left edge of section 6 (Figure 19) in December
and some bank scour at the right bank in January suggested some
110
-------
~r— — i
. AVERAGE SHKPES OF CROSS-SECTIONS
SECTION 1 SECTION 2
UJ
LU
L.
O
>
UJ
6>. NET CWNGES 'sOCT-Zl DEC
C NET CHANGES ^1 DEC-31 DEC
D. NET CHANGES 31 DEC-31 JAN
E. NET CHANGES 3UAN-£IF£E>
U——-^
i i I L.
SECTION 3
|O 15 20
SECTION 4
5 10 IS 10 O 5 10 IS 20 O
DISTANCE FROM LEFT BANK PIN, FEET
K> 15
Figure 18. Sequential changes in cross-sectional shape at sections 1-4, winter 1969-70.
-------
AVERAGE SHAPES Of CROSS-SCCTIONS
SECTION 5 \ SECTION 5A
u.
B NET CHANGES 5 OCT-£| 06C
C NET CHANGES 2\ DEC -31 DEC
Z
o
f- ,o.r D NET CHANGES 3IDEC-3IJAN
E. NEr CHANGES
O S VO IS 20 O 5
10 IS ZO 2S O S 10 IS
DISTANCE FROM LEFT BANK. PIN, FEET
SECTION 6A
SO 2S O 5 10 IS 20
Figure 19. Sequential changes in cross-sectional shape at sections S-6A, winter 1969-70.
-------
A AVERAGE SHAPES Of CROSS-SECT IONS
SECTION 7 / \ SECTION 7A
s
I-
LU
O
§
B NET CWvNSETS 5 OCT-El DO.
C NET CHfvNGES £rC£C-3IOEC
NET CHANGES 3IDEC-5IX
NET CHANGES 3ij*N-2ir
O 5 0
25. O 5
SECTION
DISTANCE.
20 25 O 5
LEFT BAN^ PlN3 FEET
>o 13 K>
SECTION a*,
~,.y *•*««.*
O 5 IO IS JO
Figure 20. Sequential changes in cross-sectional shape at sections 7-8A, winter 1969-70.
-------
AVERAGE SHAPES or CROSS-SECTIONS
SECTION 9
s
CHANGES 5 OCT. 21 OEC
t
Ul
C -.NET ChftNSES 2IDEC-3DCC
D |\N£T CHANetS 3ID6C-3UAN
E. iNET CMANSES 3IJAN-£|FEB
SECTION 10
\
o s K> t& 2> 2s o s ro is
SECTION U
SECTION 11A
V
DISTANCE FROM L£PT
»O IS X> ZS O 5
FEET
Figure 21. Sequential changes in cross-sectional shape at sections 9-11A, winter 1969-70.
-------
A AVERA6E SHAPES OF Cf?OSS- SECTION 5
104
SECTION 12
B. NET CHANGES 5 OCT^I DtC
u.
i
c NET CHANSETS
D NtT CHANGES 3IOEC-5UAN
NET CHANGES
io is ao Fs
SECTION I2A
SECTION 13
3 IO IS 2O O S 10 IS
DISTANCE FKOM LEFT BANK PIN, f=EE.T
Figure 22. Sequential changes in cross-sectional shape at sections 12-13A, winter 1969-70.
-------
A CVERASE SHAPES OP CROSS-SECTIONS
SECTION 14 (' • SECTION I«K
-^
B. MET CHANfiES 5
LJ C. NET CHNNSES
D. NET CHANGES SIOEC-IUAN
r E. NET
3IJAN-ZIFEB
10 IS 20 25 JO S 10
SECTION 15
to 71 » * IQ 15 20 2
DISTANCE TROM UtrT BANK piNj FE£T
S 1° IS
SECTION ««.
O S 10 IS IO
Figure 23. Sequential changes in cross-sectional shape at sections 14-15A., winter 1969-70.
-------
A AVERAGE SHAPES OF rRoss-S£-CTioN5
SECTION 16 SECTION 16 A.
uj ^
I
NET CWN6ES SOCT^IDEC
C NET CHANGES 21 DtC-31 DEC
D NET CHANGES 'il DEC- 31 JAN
E. NET CHANGES
o S 10 is 20 o s 10
SECTION 17
SECTION I-7A
SECTION 13
v_ > i»*iim
IS 2O O 5 10 IS 2O ^> *O
DISTANCE FROM LEFT &AN K PIN, FEET
is to o s 10 is io zs
Figure 24. Sequential changes in cross-sectional shape at sections 16-18, winter 1969-70.
-------
realignment of the main thread of storm runoff past this section.
This realignment was also reflected by slight local scour and fill
at section 6A and heavy deposition in an eddy zone where flow
separates from the left bank-just upstream of 6A (Figure 17).
Sections 7 through 8A (Figure 20) experienced alternating scour and
fill between surveys, which resulted in net deposition near the left
bank as a result of flow realignment because of growth of a point bar
just upstream. The flow realignment also caused local net scour at
these sections.
The growth of the point bar, illustrated in Figure 21, initially
occurred with deposition at the lower end, near section 9, and so?ne
erosion of the outer bank over its full length (sections 9-11A). The
January storm runoff caused the bar to build most notably at its upper
end, near sections 11 to 11A. February runoff then caused additional
deposition at the lower end of the bar.
Upstream of the bend, the narrower, deep sections 12 and 12A (Figure 22)
experienced minor scour and fill between surveys and practically no net
seasonal change. The greatest incremental change observed at section
12 was about 0.3 feet of scour in late December near the left bank. At
section 13(Figure 22) the net seasonal deposition was small compared
to the fill in early December, because of subsequent alternating scour
and fill. Section 13A (Figure 22) received a deposition of about 0.5
foot in January as part of the net fill there.
At section 14 (Figure 23), in the midst of a fairly mild-sloped reach
(Table 4), a foot of deposition occurred in early December near the
trash rack whereas little net change took place at section 15.
Subsequently, some erosion of the right bank occurred at and between
these sections; also, alternating scour and fill took place at the bed.
The incremental scour at section 14A in January amounted to one foot
in some places as the deposit behind the trash rack site was eroded.
Much of this material was probably redeposited at the upstream end of
the point bar (Figure 21).
In the upstream portion of the study reach, both scour and fill
occurred at given locations during different parts of the winter. At
section 16 (Figure 24) the incremental changes were minimal until
January, during which scour occurred up to a foot in depth. Some
refilling took place here during the February storm runoff. Similar
January and February events occurred to a lesser degree at section ISA
(Figure 23), but at section 16A (Figure 24) slight scouring continued
into February. The early December flood caused up to a foot or more
of scour at section 18 (Figure 24), near a log buried in the streambed,
and up to a foot of deposition at section 17 Thereafter, the cross-
sectional shape began to alter from section 16A to 18 and become
relatively deeper along the right bank. This change was not as
progressive at section 18 as further downstream, as may be noted from
118
-------
the foot of scour over the left half of section 18 during January,
which tended to give the cross section a less one-sided appearance.
The sequential changes of shape at the several cross sections shown
in Figures 18-24 illustrate the dynamic nature of this gravel
streambed. Short-term changes often differed from seasonal changes.
In fact, at some sections the magnitude of short-term change far
exceeded the net seasonal change, because of somewhat compensatory
changes from flood to flood. As was particularly evident near the
point bar, some changes in channel topography set off a "chain-reaction"
of events. Build-up of the bar caused greater scour of the opposite
bank, which led to further buildp-up of the bar and modified flow
aligment downstream of the bar. Bank vegetation, most notably tree
trunks and roots, held the banks against this modified flow alignment
wherever it grew. In the absence of such vegetation, the channel
banks and streambed downstream of the point bar underwent scour and
deposition in response to the modified flow alignment. In steeper
reaches, the flow energy was directed more at reworking the streambed
than at attacking the channel banks (sections 16-18), but in flatter
reaches, bank attack was noted (near sections 14-15).
Maximum Depth of Scour and Redeposition
The extreme depths of scour during a period at selected points in the
streambed were detected by relocating previously buried ping pong
balls at the end of that period, a technique similar to that used by
Leopold and others with buried chains (Leopold, Wolman and Miller,
1964). The depth of excavation needed to find the topmost of the
remaining ping pong balls and determine its position number (the basis
for measuring maximum scour) also gave the amount of redeposition
that followed this maximum scour. The technique required shallow,
slowmoving, clear flow at the time of relocation surveying for its
successful application in Oak Creek.
Table 6 summarizes the survey data on scour-ball devices, indicating
the top ball found at each location and its amount of streambed
cover at the installation date and at up to three subsequent dates
(Table 5). Data for 19 of 33 scour-ball locations are given. At
the other 14 locations (and sometimes at the 19 locations as well) it
was not possible to relocate the scour devices at surveys made
after installation because of adverse streamflow conditions, scour of
all devices (but no recover)' at the bed-load trap), or for other
reasons. The 19 scour-ball locations are shown in Figure 14.
Table 7 presents an interpretation of maximum scour and redeposition
at the 19 scour-device locations, based upon the survey data from
Table 6 and information on installation depths of balls at each
location.
The styrofoam balls at locations 1-3A, 2-1, 3-1, and two other locations
(not given in Tables 6 and 7) were of only limited value. Abrasion
119
-------
Table 6
SURVEY DATA FOR SCOUR BALL DEVICES,
WINTER 1969-1970.
November
Location
Number
1-LA
1-3A(1)'(2)
2-l(1)
3-l(1)>(3)
3-2
3-3
4-1
5-1
5-2
6-l(4)
6-2
7-1
8-1
8-2
8-3<3)
11-1
11-2
15-1
17-1
Top
Ball
--
5
5
5
11
--
12
18
18
12
13
14
17
14
13
12
13
11
14
Cover,
Feet
0.5
0.1
0.2
0.2
— -
0.1
0.4
0.05
0.05
0.05
0.05
0.2
0.05
0.05
0.1
0.05
0.2
0.05
31 Dec 14
Top
Ball
--
3
--
11
18
14
17(7)
10
7
10
10
11
Cover, Top
Feet Ball
13
(?)
9
14
0.1
O.OS 18
16
13
0.4
0.2 14
0.3
0.1
0.2 10
0.2
10
0.3
Feb 21 Feb
Cover, Top Cover,
Feet Ball Feet
0.05 8 0.4
none
0.4 9 0.3
0.05 10 0.4
0.05 18 O.OS
0.1 16 0.1
0.8 13 0.9
0.2 14 0.35
1.3
0.4 10 0. S
6 O.S
(1) Styrofoam "pop-up" balls buried at this location; punctured ping pong balls buried at all other
Locations.
(2) Found number 5; date of scour unknown.
(3) Found styrofoam ball here on 7 Feb. with number abraded; date of scour unknown.
(4) Found number 10; date of scour unknown.
(5) Found number 13 in bed-load trap on 12 Dec; found number 4 downstream of bed-load trap
on 7 Feb.
120
-------
of the styrofoam by the bed load tended to remove the identity numbers
as scour reached each ball. Debris caught the strings holding some
of the balls, occasionally (together with abrasion from the bed load)
causing the lines to snap and the styrofoam balls to be carried
downstream. Consequently, it was not always possible to know which
styrofoam ball had been scoured out of the bed or at what time and
portion of the hydrograph such scouring took place.
The ping pong balls at locations 1-1A and 3-3 were not installed until
February 14. Hence, they only give data on one storm runoff event,
that of February 16.
At location 1-1A, the scour during the February 16 storm runoff
extended 0.6 foot below the pre-flood surface of the bed, followed by
0.4 foot of redeposition. Thus, a net degradation of the streambed
of 0.2 foot occurred, although triple this depth of the bed actually
was reworked near this location by the storm runoff. The net changes
during this runoff at the adjacent cross sections 1 and 2 (Figure 14
and IS) were also slight, although it is reasonable to believe that a
similarly greater degree of bed reworking occurred there. Hence,
the net seasonal changes in this portion of the stream, shown in
Figure 17 and from the overall trends in Figure 18, can be assumed
to greatly underestimate the extent of extreme scour and deposition--
because of the compensating nature of these two processes during
individual runoff events.
A styrofoam ball from location 1-3A was recovered to indicate scour
of at least 0.5 foot in that area. The date of this scour is unknown.
However, this depth of scour supports the observations made regarding
the general vicinity of location 1-1A that a fairly stable bed position
may still permit a considerable depth of bed disturbance during storm
runoff.
The styrofoam balls at location 2-1, near section 2, were placed to a
total depth of 1.1 foot, with the top two balls near the surface
(their position was disturbed during installation). The net changes
based on seasonal (Figure 17) and storm-runoff (Figure 18) periods
were modest and suggested a slight net deposition. Yet the bed was
disturbed to a total depth of at least 1.1 foot during the season. This
further supports the conclusion that bed disturbance may greatly
exceed net bed changes during a runoff period.
Data from location 3-2 provide additional evidence of bed disturbances
to greater depths than indicated by net seasonal changes in bed
position. At location 3-3 this was also true during the February 16
runoff, when little net change in bed level occurred (about 0.1 foot
of scour in mid-channel between sections 3 and 4; Figure 18), yet
over one-half foot of scour took place before redeposition began or
the bed load ceased moving and began to rebuild a stationary bed.
121
-------
Table 7.
SCOUR AND DEPOSITION INDICATED BY
SCOUR BALL DEVICES, WMTER 1969-1970
Location
Number
December runoff
Scour, Refill,
Feet Feet
January Runoff
Scour, Refill,
Feet Feet
February 16 Runoff
Scour, Refill,
Feet Feet
1-1A — — — — 0.6 0.4
] -3A (at least 0. 5 ft scour on unknown date)
2-1 (at least 1.1 ft scour below initial bed over season)
3-1 (at least 0.2 ft scour by 7 Feb.)
3-2
3-3
4-1
5-1
5-2
6-1
6-2
7-1
8-1
8-2
8-3
11-1
11-2
15-1
17-1
...
0.22
0.35
0
(no net
o.s
0.65
0.34
0.41
0.29
(scour 0.
---
0.1
0
(scour 0.
(at least
(net fill
0.35
change)
0.3
0.1
0.2
0.2
(scour 0.
0.3
42 ft; refill 0.4ft)
...
0 0
27ft; refill 0.1 ft)
0. 3 ft scour on unknown date)
0. 75 ft)
0.53 0.2
(at least 0. 5 ft additonal scour
(net fill 1. 1 ft)
3 ft; refill 0. 4 ft)
(scour 0.9 ft; refill 0.8 ft)
(net scour 0. 1 ft)
0.53 0.4
0 0
0 0
(net fill 0. 1 ft)
(net fill 0. 15 ft)
by 7 Feb.)
(net fill 0. 1 ft)
(1) Styrofoam "pop-up" balls buried at this location; punctured ping pong balls buried at
all other locations.
122
-------
The depth of bed disturbance 'during December storm runoff was slight
at locations 4-1, 5-1, and 5-2. The channel bar at locations 5-1
and 5-2 appeared to be quite stable in spite of regular submergence
during storm runoff, as indicated in Figure 17 by very minor net
seasonal changes (see also Figure 19).
The amount of scour in the deeper flow near this stable bar was
uncertain, but at least 0.3 foot of scour took place at location 6-1 at
some unknown date.
At the protected eddy area of location 6-2, net deposition was
evidenced over the winter runoff season (Figure 17). Occasional
scour may have occurred after initial deposition, but it was never
sufficient to reach the topmost buried ball. The deposition during
December and January consisted of sand and fine gravel, but the net
additional deposit in February was coarser. Such areas might well
serve as storage zones for transported material until such time as
flow alterations upstream reduce the protected nature of the area and
permit further disturbance and transport of the deposited material.
At location 7-1, in the main flow channel, about one-third foot of
deposition took place as part of a more extensive seasonal aggradation
of the streambed in this zone, below the large bend and point bar
(Figure 20).
Streambed disturbance was quite noticeable at locations 8-1, 8-2, and
8-3, situated in a zone subject to net seasonal scour (Figure 17).
Data indicated that extreme scour extended to greater depths than
reflected by net short-term changes (Figure 20), again supporting
conclusions drawn from scour ball observations made downstream.
The large amount of deposition on the point bar began in January
(Figure 21) and was reflected by 1.1 foot of accumulated gravel at
location 11-1 in that month. Previously, a small amount of net scour
took place at 11-1 and 11-2. The gravel that deposited on the point
bar was generally coarse; it was probably of nearly the same size as
transported throughout the study rock. The pore space below the top
gravel was partially filled with sand, and a great deal of filtering
of fine particles probably occurred there during recessions of storm
runoff.
As elsewhere, the dual processes of scour and refill during flood
periods are illustrated by data from location 15-1. Net seasonal
deposition did not preclude scour at this point.
Perhaps the greatest relative differences between extreme scour and
net short-term change of bed level are shown by data from location
17-1. Here, the bed appeared to undergo a net scour of 0.1 foot
during January-February storm runoff, although the bed was actually
disturbed and reworked to almost 10 times that depth. This offers
further evidence of the dynamic nature of a seemingly stable streambed.
123
-------
SECTION VIII
SUSPENDED SEDIMENT LOAD
Observed Concentrations
Several discrete samples (with respect to time) of the suspended load
were obtained with the DH 48 hand sampler during January and
February runoff. The suspension concentrations and stream discharges
at the times of sampling are shorn in Table 8. A broad range of flows
was sampled, although no samples were obtained at the highest stages
because of the press of time to conduct bed-load samplings and stream
gagings during the brief periods of highest discharge.
The individual samples obtained with the hand sampler were originally
intended to supplement the continuous automatic sampling. However,
failure of the system's timer in cold weather rendered the samples
obtained with the automatic system completely unreliable until a
suitable replacement timer was obtained in February. Thereafter, no
important stream rises occurred and the data reported in Table 8 from
the automatic system represent only the spring recession flows with
their minimal suspended load.
Four storm runoff events are represented by the data in Table 8.
During the first, second, and fourth of these periods, only the runoff
crests and recession limbs were sampled. The first two samples
collected on the afternoon of January 9 actually straddled a slightly
larger peak discharge (36 cfs). The first sample collected on
January 14 was obtained about three-quarters of an hour after the
stream crested at 60.3 cfs. However, the sample for February 16 was
obtained 4 1/2 hours after the stream had crested at 175 cfs, and
hence did not reflect the maximum suspended solids concentration of
that runoff event. (It is assumed here that the maximum suspended
solids concentrations occurred on the rising limb near crest stage
or at about crest stage and not on the falling limb of the hydrograph.)
Sampling of the third runoff event of Table 8 included the rising limb
as well as the falling limb of the hydrograph. The concentration of
the second sample of January 17 appears to be in error, but no
justification could be found for discarding it, so it has been included,
(As will be seen in Figure 25A, this data point does not contribute
worse scatter than do other data points.) The 1 a.m. sample on
Jnauary 17 was obtained an hour after the peak discharge of 210 cfs and
perhaps may roughly indicate the maximum turbidity of the stream under
present watershed conditions of land usage, because of the relative
infrequency of occurrence of such a large peak (on the order of
a once-in-twenty-years event).
When discharge is plotted versus the corresponding suspended
sediment concentration, considerable scatter may be expected because
of the lack of a precise physical law to govern this relationship.
124
-------
Table 8
OAK CREEK SUSPENDED SEDIMENT
CONCENTRATIONS, WINTER 1969-1970
Date
9 January
9 January
9 January
10 January
14 January
14 January
14 January
1 5 January
16 January
16 January
17 January
1 7 January
1 7 January
1 7 January
16 February
17 February
19 Feb - 16 Mar(1)
16 Mar - 8 Apr(1)
Time
13:43
14:50
20:18
9:30
15:42
17:00
23:15
11:15
15:25
16:45
1:00
9:15
11:00
12:55
17:00
12:10
Discharge,
cf s
35
35
28
13
60
50
34
21
45
40
138
52
46
44
80
35
8
3
Suspended Sediment
Concentration, nig /I
204
166
49
9
147
111
28
7
25
96
417
75
52
195(?)
86
23
5
< 2
(1) Obtained with composite automatic sampler; all other samples
obtained with DH 48 hand sampler.
125
-------
Figure 25A shows this scatter for the discrete and composite samples of
Table 8. The data have been fitted with a somewhat arbitrary line of
relation (the fit was made by eye rather than by a least squares
analysis), to show the general behavior of suspended load concentrations
as a function of streamflow. At about 30-40 cfs, the suspended
sediment concentrations begin to increase rapidly as discharge
increased. Interestingly, an independent analysis of bed-load
transport as a function of discharge in Oak Creek suggested that
significant transport was initiated at about this same range of flows.
This strongly suggests the close relation of the suspended load to the
behavior and condition of the streambed.
Suspended Sediment Transport Rate
The estimated line of relation of suspended load concentration versus
discharge can be converted into various equivalent relations to describe
the rate of suspended sediment transport. The parameter used to
describe this rate in this study is shown in Figure 25B and might be
termed a "unit" suspended sediment transport rate--i.e., the transport
rate of suspended sediment in pounds per hour for a unit of stream
discharge (for one cubic foot per second). This is almost a
dimensionless parameter and would become one upon division by the
specific weight of water and use of a seconds-to-hours conversion
factor. This parameter was originally selected in analysis of the
bed-load transport rate and is used here to provide an analogous
relation for the suspended load.
From Figure 25B, note that a unit of discharge carries approximately
70 Ib/hr of suspended solids at 100 cfs. For comparison, about
30 Ib/hr of bed load are moved by a unit of discharge in Oak Creek
at this streamflow (to be discussed in a following section).
Once a relationship such as that in Figure 25B has been developed,
it may be readily applied to storm runoff hydrographs to estimate the
total suspended sediment yield as a result of a runoff event. This
has been done for Oak Creek with three storm runoff periods having
differing magnitudes of peak discharge. Similar calculations were
also made for the bed load, and the results were combined to
estimate the fraction of total sediment load that is transported by
bed-load processes. The results conclude the following section on
bed-load transport, so no further discussion will be presented here.
126
-------
C-
o.
I
2oo
o
U
oi
f.
120
\°
P 8
1 I
f\.
B. U W\T
1 T
Q J I U—
1 2. 3
L^J-1-4-4-4
._L L-u-
;>o~ ix 5tx>
il. t>iS<:rU\iil.4E y
Figure 25. Suspended sediment concentration and unit
transport rate as functions of discharge.
-------
SECTION IX
BED-LOAD TRANSPORT
The primary purpose of this research has been to obtain a usable
understanding of the changes in total sediment yield from a drainage
basin with changes in land use. As stated previously, a study of the
bed-load transport process in a typical mountain stream [Oak Creek)
was made to obtain insights and data describing the hydraulic and
morphological factors that influence the movement of bed material.
This section of the report gives the pertinent results of a review of
the literature on sediment yield., of measurements made on bed movement
and bed materials at the Oak Creek instrumented study reach, and of
measurements made of the bed material in one of the Alsea Experimental
Watersheds (Deer Creek). A comparison of bed-load discharge and
suspended load discharge is also made.
Sediment Yield From Watersheds
At the start of this study, a review of the literature on sediment
yield from watersheds was made to obtain information on the changes
of sediment input to a stream channel with changes in land use in the
watershed.
The yield of sediments from a watershed can be considered to involve
two processes: one on the land surface that transfers sediment to a
stream channel, and one in the stream that transports the sediment
out of the watershed. The transfer of sediment to a stream consists
of three elements, which can act in combination or alone: (1} local
and sheet erosion by overland flow; (2) mass movement into the
stream channel; and (3) direct erosion by the stream. Roehl (1962)
reports that in the Southeastern United States sheet erosion represents
66-100 percent and channel erosion 0-34 percent of the total erosion.
No information was found in the literature that quantifies the
importance of mass movement of earth materials in the delivery of
sediment to stream channels. However, the geologic literature suggests
that mass movements are quite important in mountainous areas and
that infrequent landslides can deliver very large quantities of
sediment to streams. Therefore, mass movements might be important in
some instances even though no data on the relative importance of mass
movements in comparison to water erosion were found in the literature.
Numerous studies have been made to determine watershed factors that
affect the total sediment yield process. One equation relating
watershed variables to erosion is that developed by Smith and Wischmeier
(19S7) for agricultural areas in the Midwestern United States. The
Smith and Wischmeier equation is of the form
X = (I) (JO (LS) (C) (P), Eq. 1
128
-------
where X = annual soil loss (erosion), I = erosive potential rainfall
factor, K = soil credibility, LS = topographic factor, C = cropping-
management factor, and P = conservation practice factor.
This equation suggests that sediment yield is a function of: (1)
amount of rainfall and the time distribution of rainfall; (2) soil
type and geology; (3) topography and geomorphology; and (4) land use.
Hence, the equation could be rewritten as
X = x (rainfall, soil, geology, topography, land use). Eq. 2
Anderson (1954) developed a somewhat analogous equation for sediment
yield from forested watersheds in western Oregon. His is a regression
equation with many empirical terms but, in essence, it can be restated
as :
Y = y (runoff, drainage area, bank erosion, soil type, geology,
slope, land use, snow-rain factor), Eq. 3
where Y is the total sediment yield from a watershed. The snow-rain
factor takes into consideration the effect of snow or a rain-snow
situation upon the runoff process. The runoff factor is a function of
a rainfall, drainage area, soil type, geology, topography, and snowpack.
Hence, Equation 3 can be rewritten as:
Y = y (rainfall, soil type, geology, topography, land use,
snow-rain factor, drainage area, bank erosion). Eq. 4
A comparison of Equations 2 and 4 suggests that erosion in forested
mountain areas is a function of (1) rainfall, (2) soil type and geology,
(3) topography, (4) land use, and (5) snow-rain factor, and that the
transport of the sediment from the watershed is a function of (1)
drainage area and (2) bank erosion. For a given watershed, the
factors that can be influenced by land management practices are land
use, bank erosion, and over a long period of time, soil type.
Equation 4 can be considered to include a set of erosion terms
(identified by the symbol X from Equation 2, but with rainfall
assumed to include the snow-rain factor) and a set of sediment transport
terms and can be written as:
Y = y (X, drainage area, bank erosion) Eq. 5
Furthermore, it is convenient to regard the sediment yield problem
as one where material, once eroded, must be either redeposited in the
watershed or transported from the watershed.
Roehl (1962) has presented a procedure to estimate the "sediment
delivery ratio" based on watershed parameters. He defines the
sediment delivery ratio as "the percentage relationship between the
sediment yield at a specified measuring point in a watershed and the
gross, or total, erosion occurring in the watershed upstream from the
129
-------
point." The sediment delivery ratio was found to be a function of
drainage area, ratio of basin relief to length, and a factor that is a
function of the density of drainage channels. The density of drainage
channels is strongly influenced by the soil type and geology. From the
definition of sediment delivery .ratio, it would appear reasonable to
include bank erosion as one of the influencing factors. Hence, we can
write
SDR = f (soil type, geology, topography, drainage area,
bank erosion), Eq. 6
where SDR is the sediment delivery ratio.
Equation 6 may be combined with Equation 4 and 5, giving
Y = y (X, SDR) Eq. 7
The resulting equation expresses total sediment yield from a watershed
as a function of the total watershed erosion and the transport
capability for removing these erosion products (as expressed by the
sediment delivery ratio).
The equations as presented above obviously give only qualitative
relationships based on quantitative data—yet, there are certain
unexplained factors. One of these is the method of measuring the total
sediment yield from a watershed. The total load of material moving in
a stream usually is divided into bed load and suspended load and the
only part measured is that in suspension. On smaller watersheds, the
measurement technique often used is to induce turbulence into the flow
such that part of the bed load will be suspended at the sampling point.
If the bed material is fine sand, this technique may be reasonably good.
But if the material is coarse sand or gravel, the material either will
not be suspended by the turbulence or will not be sampled--because
errors in sampling increase as the particle size increases (Love and
Benedict, 1948). In larger streams, the bed load is often not
measured in any way. This means that the process of erosion will
place material in a stream that will be transported by the stream
but not included as part of the sediment yield by the measurement
techniques.
In determining his equation for sediment delivery ratio, Roehl used
data on sedimentation in a reservoir to determine the sediment yield
from a watershed. The amount of sediment in a reservoir was measured
and divided by the trapping efficiency of the reservoir to obtain
the estimated total sediment yield. Unfortunately, trap efficiencies
of reservoirs are based on sediment measurements that usually do not
represent the total sediment transported from a watershed. Consequently,
neither reservoir surveys nor sediment sampling will give reliable
estimates of the sediment yield from a watershed unless the bed load
is measured.
As an illustration, assume that the sediment delivery ratio varies
130
-------
with drainage area as shown by Roehl and as reproduced in Figure 26.
As is shown, the sediment delivery ratio decreases rapidly as the
drainage area increases. For an illustrative drainage area of 6 square
miles, only 20 percent of the material eroded will be carried out of
the watershed, which means that 80 percent is redeposited in the
watershed to be eroded again when there is another storm. If the
assumed gross erosion has been accurately determined, then this
example suggests three possibilities:
1. once eroded, sediment will move on the land surface until reaching
a stream or be redeposited on the land surface as overland flow
diminishes;
2. deposition occurs in the stream channel and results in net
aggradation of the channel (but much of the available literature,
including Roehl's paper, indicates that channel erosion is common in
upland areas); and
3. the estimate of total sediment yield is too low because the bed
load usually is not known.
Both sheet erosion and channel sediment transport will include
material that is being carried at the "bed" of the flowing water. For
a stream, this is the bed load and for sheet flow, this is really
bed load, although no such distinction of the separate components of
sediment transport is made for sheet flow. In each instance, as the
flow recedes the "bed" material will redeposit and be at rest. If this
redeposition material is transported again with the next storm runoff
and is classified again as part of the gross erosion, this results in
only a low percentage of the gross erosion products being carried
out of a watershed in any one storm.
The main reason for discussing the ideas presented above is to suggest
that an understanding of the process of bed-load transport is important
in understanding erosion on a watershed and in quantifying the amount
of sediment yield from a watershed.
Oak Creek Bed Material
In the fall of 1969, before any storm runoff, the bed materials were
sampled in the instrumented natural study reach of Oak Creek. The
samples were taken at previously established cross sections (Figure
14) . All but one of the samples were of armor material at the
surface of the streambed and were made with the object of obtaining
data on the size of material that forms the boundary of the flowing
water. One sample was also taken of the material just below the
surface layer.
The technique used in sampling the bed material in the study reach was
to obtain representative surface material at the center of the stream
and at the quarter points of a given cross section. These samples
were then combined to form a sample for the cross section. To do this
at each point, a bucket with the bottom cut out was set on the bed and
all material within it exposed at the surface was collected. The
131
-------
loot
O.Q
\00
|0oo
Figure 26. Sediment delivery ratio as a function of basin size (adapted from Roehl, 1962, Figure 2)
-------
diameter of the bucket opening was 8+ inches, for an area of about
0.4 square foot per sample point and 1.2 square feet per cross
section. Some fines [small sand and silt-size particles) were washed
off the larger particles and lost if the sample point was located in
the water.
Particle-size analyses were made with this bed material. The results
of these analyses are presented in Table 9. The table gives the
fraction (by weight) of each cross-sectional sample retained on each
of several sieves. (For oblong particles, the size measured by a
sieve analysis is the intermediate dimension of the particle.)
The mean particle size (size for which 50 percent of the sample by
weight is smaller—designated as D50) for the surface layer at each
cross section is also given in Table 9. These values were obtained
from particle gradation curves for each sample (i.e., from graphs of
percentage of sample passing each sieve versus sieve opening size,
the latter being assumed equivalent to particle diameter). The
gradation curves for the samples with the smallest and largest mean
particle sizes, those for samples at sections 9 and 12 respectively,
are presented in Figure 27. These illustrate the typical gradation
of the surface (armor) layer throughout the study reach, because of
little variation in the mean values of particle size from section to
section. If the mean particle sizes at the sections are averaged, a
mean size of surface material for the entire reach can be estimated.
This value was found to be 2 1/4 inches.
Of some interest is the variation of mean grain size of the streambed
surface with distance along the study reach. A plot of the mean
particle size for each cross section versus location along the channel
is given on Figure 28A. As shown, there is some irregular variation
with distance, but there may be a slight tendency for the mean
particle size to decrease with distance downstream. In this reach,
the smallest observed mean size at a cross section was 1 3/4 inches
at section 9, located in a pool (Figure 15) and the largest was 2 3/4
inches at section 12, located near the upstream edge of a bar. For
comparison, a sample of the armor material located near a debris
dam about 1,000 feet downstream of the study reach had a mean
diameter of 1 3/4 inches, which tends to add support to the idea that
the mean particle size of streambed particles decreases with distance
downstream.
Figure 28B shows the variation of the surface layer fraction 1 1/2
inches or coarser in size with distance along the channel. The
variation is considerable. Yet, the suggested limits of this variation
indicate a decrease in the coarse fraction in the downstream direction,
again supporting the idea that particle size of streambed materials
decreases with distance downstream.
Another important aspect is the variation of the size gradation of
133
-------
Table 9
PARTICLE SIZE DISTRIBUTIONS FOR OAK CREEK BED MATERIAL
FORMING THE SURFACE LAYER IN THE STUDY REACH, FALL 1969
Sieve
Size
3 - in.
2 - in.
1 1/2 - in.
3/4 - in.
3/8 - in.
No. 4
No. 8
No. 16
No. 30
No. 50
No. 100
Pan
Mean
Particle size
at
Cross -section
inches
Fraction of sample retained
3
0.200
---
0.782
0.926
0.962
0.977
0.983
0.989
0.994
0.997
0.999
1.000
2 1/2
4
0.085
0.552
0.699
0.920
0.979
0.996
0.999
0.999
0.999
0.999
0.999
1.000
2 1/4
5
0.111
0.686
0.919
0.978
0.992
0.996
0.997
0.998
0.999
0.999
1.000
2 1/4
6
0
---
0.561
0.886
0.970
0.987
0.994
0.998
0.999
0.999
0.999
1.000
2 1/4
by indicated
7
0.070
---
0.787
0.931
0.967
0.980
0.988
0.992
0.996
0.999
0.999
1.000
2 1/2
sieve at the
8
0
---
0.683
0.885
0.965
0.986
0.994
0.996
0.998
0.999
0.999
1.000
2 1/4
given sampled cross -section
9
0
---
0.611
0.897
0.970
0.986
0.989
0.993
0.996
0.998
0.999
1.000
1 3/4
10 1/2(1)
0.098
---
0.679
0.858
0.935
0.964
0.980
0.988
0.993
0.996
0.998
1.000
2
' 12
0.119
---
0.863
0.964
0.991
0.998
0.999
0.999
0.999
0.999
0.999
1.000
2 3/4
13
0
...
0.625
0.942
0.986
0. 994
0.997
0.999
0. 999
0.999
0.999
1.000
2 1/2
14
0.279
---
0.625
0.846
0.926
0.963
0.969
0.990
0.996
0.998
0.999
1.000
2 1/2
15
0
---
0.862
0.971
0.997
0.999
0.999
0.999
0.999
0.999
0, 999
1.000
2 3/4
(1) Midway between sections 10 and 11.
-------
&
| oo
loo
2
iL
\-
1
Hi
V
D-
|OO
loo i ao 5o
0.2. '
Figure 27. Typical particle gradation curves for the bed surface
in Oak Creek study reach, fall 1969.
135
-------
3oo
iNCe uP>sT£tf\r"A OF \\c
Figure 28. Variation of mean size and coarse fraction of streambed surface
layer with location along Oak Creek study reach, fall 1969.
-------
the bed material with depth below the bed surface. One sample was
taken of the material below the surface layer at cross section 15.
The data are given in Table 10 and on Figure 29A. In addition, a
sample was taken from a bar that had deposited behind a fallen tree
and debris dam about 1,000 feet downstream of the instrumented channel.
The debris dam subsequently failed abruptly during a flood and part of
the bar was eroded, which lowered the water level and exposed a cut
cross section of the bar to a depth of over a foot below its surface
layer. Samples of this surface layer and of the material below it were
taken without losing the fines (because there was no flow to contend
with). The results of the sieve analysis are given in Table 10 and on
Figure 29B.
As is shown in Table 10 and Figure 29, there is a significant difference
between the surface layer and the material below. Furthermore, there
is little difference in material composition for the first and
second six inches of bed below the surface at the downstream bar.
Similar observations by other investigators of coarse surface layers
overlying vertically well-mixed (but finer) streambeds have given
rise to the terms "armor layer" and "protective pavement" to describe
such surface layers. Their existence appears to be quite common when
dealing with gravelly streams. They offer an added measure of
protection against bed-load transport until the flow increases
sufficiently to disturb them.
Streambed materials, like other earth materials, can be described by an
index or coefficient of uniformity, derived from the particle size
gradation curve. This index is defined as the ratio of that particle
size such that 60 percent of the sample is finer to that particle size
such that 10 percent of the sample is finer (i.e., U = D60/D10). Data
on the uniformity of selected samples from Oak Creek are given in Table
11. The index of uniformity for the armor is between 2.3 and 2.8, but
that for the material below the armor layer is between 5 and 8. Hence,
the armor layer is considerably more uniform (a lower index value) in
its size range than is true for the material beneath the surface.
Another important consideration is the variation of particle weights
within each of the size ranges. Data on this are given in Table 12.
The data indicate a wide range of weight for a given sieve size. There
is also considerable overlap of particle weights from the larger
grains of one size range to the smaller grains of the next-larger
size range. The weight of a "representative sphere" also was calculated
for each size range, based upon the geometric mean (representative)
particle diameter for each size range. These calculated weights of
the equivalent spheres compare favorably with the measured median
particle weights of each fraction, which suggests the equivalent
weight may be as good as any other single weight in representing the
weight of any given set of particles. However, the variance of
particle weight is very large and probably quite important in
understanding the transport and armoring processes. The calculated
weight for the No. 4 sieve size range is much less than the median
137
-------
Table 10
VARIATION OF STREAMBED PARTICLE SIZE
DISTRIBUTION WITH DEPTH BELOW BED SURFACE,
OAK CREEK, FALL 1969
Fraction of Sample Retained by
Indicated Sieve at Sampling Location
Sieve
Section 15 , ^
Size Surface
3 in.
2 in.
1 1/2 in.
3/4 in.
3/8 in.
No. 4
No. 8
No. 16
No. 30
No. 50
No. 100
Pan
Mean
Particle
Size at
Location,
mm. (inches
0
0.862
0.971
0.997
0.999
0.999
0.999
0.999
0.999
0.999
1.000
68
(2 3/4)
Subsurface"'
0
0.426
0.695
0.860
0.893
0.928
0.992
0.996
0.998
1. 000
16
/T\
Downstream Bar
Surface 1st 6 inches
0 0
0.390 0.076
0.624 0.197
0.958 0.576
0.998 0.785
1.000 0.905
0.951 .
0.973
0.986
0.995
0.997
1. 000
45 22
(1 3/4)
2nd 6 inches
0
0. 135
0.263
0.623
0.813
0.933
0.978
0.990
0.995
0.998
0.999
1. 000
24
(1) This bar is about 1, 000 feet downstream of the study area.
(2) Approximately the first three inches of streambed below the
layer.
surface
138
-------
OJ
10
Mill
8.
.—"DO WNSTfcE W»VP ftl
log 5o
Figure 29. Variation of size gradation of bed material with depth below bed surface, Oak Creek, fall
1969.
-------
Table 11
UNIFORMITY OF OAK CREEK BED MATERIAL
AT SELECTED LOCATIONS, FALL 1969
Property of
particle size Section
gradation curve 9
(11
D ( ' 52
60
(2)
D n 19
10
(31
UV J 2.74
Values at Indicated Locations
Section
12
73
31
2.36
Section 15
Surface
70
30
2.33
Subsurface
20
2.5
8.0
Downstrc-am Bar
Surface
51
21
2.43
. 1st 6 indies 2nd 6 indies
25 30
4.8 6
5.2 5.0
(1) Size such that 60 percent of sample particles are smaller.
(2) Si/c such that 10 percent of sample particles are smaller.
<3> U = V °10
(Note: if U > U , then sample 1 is less uniform than sample 2)
140
-------
Table 12
VARIATION OF PARTICLE WEIGHTS WITHIN GIVEN
SIZE RANGES FOR OAK CREEK ARMOR LAYER, FALL 1969
Sieve
Retained
on
3 - in.
2 - in.
1 1/2 - in.
3/4 - in.
3/8 - in.
No. 4
Representative
Particle
Diameter,
Millimeters
88
62
44
27
13.5
6.7
Weight of
Equivalent
Sphere(2)
grams
1,020
360
127
29.2
3.7
0.45
Particle
Weights
Maximum, Median,
grams grams
1,465
766
234
70.6
12.9
2.03
936
318
134
19.3
2.8
0.99
Minimum,
crams
552
169
69
3.8
0.89
0.77
(1) The representative particle diameter is the geometric mean of the sieve on which the particles
are retained and the next larger sieve (which the particles were able to pass through).
(2) The equivalent sphere is based upon the representative particle diameter for that size range mid
its weight is calculated assuming a specific gravity of 2. 85 for the grains (based on laboratory
measurements).
(3) Data for this size range were based upon a single sample at an exposed bar downstream of the
study reach; all other data represent composited samples.
141
-------
weight, probably because the lighter particles had been removed from
the armor layer so that the measured weights are not representative
of the full range of particles between No. 4 and 3/8-inch., even
though the calculated weight is.
Oak Creek Bed-Load Transport
This section reports the results of bed-load measurements made at Oak
Creek in the winter of 1969-70 using the vortex bed-load sampler.
Considerable experience had to be gained in operating the installation
and at the same time collecting samples for use in the sediment yield
analysis. Therefore, the data reported here are the results of
measurements made in the process of learning how to operate the
installation and represent only those data taken at time intervals
when the discharge was reasonably constant. Measurements made over
other time intervals when the discharge varied widely were not
considered usable because the effective discharges for such situations
were not known.
Mechanical analyses of the 14 bed-load samples for which the effective
discharges could be determined reliably are summarized in Table 13.
Also shown are several useful measures of particle size distributions,
obtained from the gradation curves for each sample. These include the
average (D50) particle diameters, the indices of uniformity (based on
D&oand Dio), indications of bed roughness (Des), and indications of
maximum particle size (D95). In each instance the subscript refers to
the fraction of the sample finer than the indicated size. Considerable
variation in all of these measures is evident in this table. These
values can be related to stream discharge and bed-load transport rate,
as may be seen by comparison with data in Table 14 and as discussed
in subsequent paragraphs.
Bed-load transport data are summarized in Table 14. The indicated
discharges are the mean or effective values for the sampling intervals.
A descriptive column in the table indicates the manner in which the
hydrograph was varying during each sampling period. The total bed-load
transport rate at each sampling period is reported, followed by a
breakdown giving transport rates for various size fractions of the
bed-load.
Data from Tables 13 and 14 are presented in Figures 30A, 308, and 30C
to show the variation with stream discharge of total bed load, bed
load coarser than 1 1/2 inches, and bed load coarser than No. 4 sieve
size (4.76 mm.), respectively. In Figure 30, no data adjustments were
made for bed-load sampler efficiency. Lines of "visual best fit"
were drawn through the unadjusted data points. The total bed-load
transport, in Figure 30A, at 130 cfs was estimated during the large
storm runoff event that transported more bed load than the sampling
installation could handle. The estimate was made on the basis of the
rate at which the vortex trough filled after being partly cleared and
is considered to be a good estimate of the possible range of the
142
-------
Table 13
PARTICLE SIZE DISTRIBUTIONS FOR BED-LOAD SAMPLES,
OAK CREEK, WINTER 1969-1970
Sieve
Size
3 - in.
2 - in.
1 1/2 - in.
3/4 - in.
3/8 - in.
No. 4
No. 8
No. 16
No. 30
No. 50
No. 100
Pan
V
Millimeters
D
60,
Millimeters
D
10,
Millimeters
D
0=^
D
10
D6S,
Millimeters
°95
Millimeters
Fraction of sample retained by indicated sieve for sample of given identity number
1
0.0
0.054
0.070
0.116
0.238
0.454
0.762
0.919
1.000
0.52
0.74
0.17
4.35
0.78
11.0
2
0.00
0.006
0.028
0.117
0.194
0.285
0.419
0.677
0.887
0.965
0.966
1.000
2.00
2.70
0.54
5.00
3.40
33.0
3
0.001
0.074
0.195
0.555
0.739
0.818
0.870
0.930
0.976
0.992
0.996
1.000
22.5
26.0
1.80
14.4
30.0
56.0
4
0
0.034
0.060
0.093
0.176
0.432
0.682
0.832
0.923
1.000
0.96
1.20
0.18
6.66
1.40
12.7
5
0
0.121
0.141
0.252
0.358
0.428
0.520
0.660
0.810
0.897
0.951
1.000
2.78
6.80
0.28
24.3
10.0
70.0
6
0
0.02S
0.081
0.301
0.468
0.509
0.563
0.707
0.845
0.914
0.963
1.000
6.50
13.0
0.32
40.6
17.0
43.0
7
0
0.097
0.272
0.358
0.414
0.500
0.666
0.820
0.919
0.927
1.000
2.40
6.00
0.38
15.9
12.0
44.0
8
0
0.080
0.194
0.512
0.697
0.792
0.860
0.933
0.978
0.992
0.996
1.000
20.0
24.0
1.50
16.0
28.0
56.0
9
0
0.046
0.136
0.442
0.646
0.762
0.837
0.922
0.976
0.993
0.997
1.000
16.3
21.0
1.40
15.0
23.0
51.0
10
0
0.065
0.101
0.230
0.421
0.584
0.699
0.850
0.947
0.981
0.992
1.000
7.20
10.0
0.90
11.1
14.7
55.0
11
0
0.022
0.056
0.239
0.432
0.535
0.641
0.821
0.944
0.982
0.992
1.000
7.20
11.0
0..84
13.1
13.0
40.0
12
0
0.023
0.072
0.333
0.560
0.672
0.748
0.849
0.931
0.965
0.983
1.000
12.5
18.0
0.84
21.4
20.0
42.0
13
0
0.007
0.036
0.279
0.516
0.647
0.740
0.867
0.964
0.991
0.996
1.000
10.4
14.0
0.98
14.3
17.0
37.0
14
0
0.006
0.010
0.285
0.532
0.652
0.773
0.871
0.921
0.955
1.000
5.20
7.20
0.42
17.1
8.00
18.0
-------
Table 14
SUMMARY OF BED-LOAD TRANSPORT DATA,
OAK CREEK,- WINTER 1969-1970
Sample
Identity
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Stream
Discharge
cfs
19.
49.
71.
23.
33.
32.
20.
54.
50.
39.
44.
46.
45.
34.
0
4
0
3
6
9
3
0
0
4
0
5
0
8
Water
, Stage
Trend
Varying
Rising
Falling
Rising
Rising
Falling
Falling
Falling
Falling
Falling
Falling
Falling
Rising
Falling
Bed -Load Transport,
Ib. /hr.
For Fraction Coarser than Indicated Sieve
Total
0.050
26.05
441.36
1.22
7.06
3.90
1.01
100. 65
82.93
5.67
77.26
466.06
173.05
14.11
1 1/2
0
0.
85.
0.
0.
0.
Q.
19.
11.
0.
4.
33.
6.
0.
- inch
73
88
0
88
36
10
50
26
58
32
32
17
08
No. 4
.0.004
7.42
360. 09
0.11
2.66
2.26
0.42
79.70
63.19
3.31
41.36
313. 36
112.03
7.50
No. 16
0.012
17.63
410. 66
0.52
4.40
2.98
0.67
93.92
76.50
4.82
63.45
395.72
150.08
10.90
No.
0.
23.
430.
0.
5.
3.
0.
98.
80.
5.
72.
433.
166.
12.
30
023
10
69
83
50
40
83
46
93
37
98
80
80
28
No. SO
0.038
25.13
437. 88
1.01
6.22
3.62
0.93
99.88
82.33
5.56
75.87
449.66
171.48
12.99
144
-------
T—I—Mill
120
<*>
K
-------
bed-load discharge when the flow rate is 130 cfs. For comparative
purposes, the three curves of Figure 30 are superimposed., with data
points omitted in Figure 31.
Figure 30 shows considerable scatter of the data. Furthermore, the
convergence, as discharge increases, of the lines of best fit for
total bed-load transport and transport of particles coarser than No. 4
sieve size implies that the fraction of the total bed load smaller
than No. 4 sieve size (represented by the gap between the converging
lines) decreases toward zero with increasing transport rate, which is
not reasonable. Consequently, the bed-load data were critically
examined to find an explanation for this seeming inconsistency
regarding the transport of sand-size material.
To obtain some idea of the transport rate of the sand-size material,
graphs of the bed-load transport rate for grains smaller than No. 4
sieve size and No. 30 sieve size (0.595 mm.) were made (Figure 32).
As shown, there are two possible lines through the data--one that
includes the low-flow samples plus data for the early-winter samples
and another through the low-flow data and the late-winter samples.
The probable reason for this difference is that the stage-discharge
relationship had changed on January 16 and the Froude number was
lower for the later samples (Figure 5). The trap .efficiency of the
bed-load sampler for small particles appears to be related to the
Froude number. Field observations indicated a higher degree of
turbulence as the Froude number increased, which would tend to cause
suspension of an increasingly higher percentage of the sand and silt-size
particles. If it is assumed that the lower lines in Figure 32A and
32B represent higher trap efficiencies than do the upper lines (since
the lower lines indicate greater transport at a given discharge than
do the upper lines), then it appears that trap efficiencies were higher
for samples 11, 12, 13, and 14 than for other samples collected at
high streamflows. In fact, sample 11 was obtained on January 16 as the
rating curve was shifting toward lower Froude numbers. Using Figure 32,
the data can be adjusted to obtain a better estimate of the true bed-load
transport and also some idea of the efficiency of the sediment trap
for the sand-size particles. To do this, Froude numbers were calculated
for each sample, based on Figure 5. Relative trap efficiences were also
calculated with the assumption that the lower curves of Figure 32
correspond to maximum trap efficiency. From these calculations, a
diagram of trap efficiency versus Froude number was prepared (Figure
33).
The curve of Figure 33 shows a marked decrease in the trapping
efficiency for small particles as the Froude number increases much
above a value of 0.83.
Based on Figure 32, adjustments of the bed-load tranport rates for
samples 2, 3, 8, 9, and 10 were made to obtain a better idea of the
"true" transport rate. The revised estimates of the total bed-load
146
-------
2.90
IbO
1ZO
uP
r
« 80
ui
>
cc
O
I (MM
I i t i t t
~f—i—r i i i n
J 1 t I I 11
1 1—I M I IT
1 1 I t I I I
T 1—I MM)
t L 1 i I
4-J
) 50 loo Soo
T**NSPofcT fc*TE, LB./M-R.
T 1—I Mill
Figure 31. Superimposed unadjusted bed-load transport curves, Oak Creek, winter 1969-70.
-------
feo
1C
vo
h- a«_T OF P
8.
O, I 0.1. 0.3 0.5
i i i i i
13
a. 3 5
NO. 4-
Si2,E
SIEVE sire
i i .11111
(4/76
13.
"So So \oo
Sou
Figure 32. Bed-load transport for sand-size material, Oak Creek, winter 1969-70.
-------
SOO
I
60
o
2
ul
a
^ -»rx _
o
NO. 4- "Sieve
PA*2.n
-------
transport rate for the five samples are given on Table 15 and the
adjusted total bed-load transport rates are plotted on Figure 34.
In comparison with the unadjusted data of Figure 30A, the scatter is
much reduced. Sample 12 still appears to have a bed-load transport
rate that is too high for the discharge. Sample 12 was collected
just after a period when the vortex and backup troughs were filled
and then cleaned out just before opening the sampling gate. Opening
the gate probably caused material temporarily stored upstream of the
trap (deposited in the backwater of the filled troughs] to scour out
because of increased shear velocities at the streambed. (Normally,
material would not be stored upstream of the trap in such a manner, but
deposition of material in the troughs and on the downstream edge of
the weir/trap structure had raised the effective bed level. These
deposits were removed, which lowered the bed level before the gates
were opened.) If the plotted location of sample 12 is assumed wrong
on Figure 34, the appropriate correction to make, based upon the
argument of excessive scour because of adjustment of bed level, would
reduce the bed-load transport and draw the plotted point closer to
the curve.
The curve for the adjusted total bed-load transport rate is
superimposed on the curves for bed-load transport of material
coarser than 1 1/2 inch and No. 4 sieve size in Figure 35. It is
evident that the convergence of the unadjusted data of Figure 31
has been accounted for by the correction for trap efficiency of sand-
size particles.
The data for adjusted total bed-load transport in Oak Creek have been
rearranged in Figure 36 to present the "unit" bed-load transport rate,
or transport rate per unit of discharge (1 cfs), as a function of
streamflow. This curve can be conveniently used to estimate the
total bed-load transport during a storm-runoff period (see discussion
in a following section).
The adjusted data on total bed-load transport are presented on an
arithmetic scale in Figure 37, in comparison with the logarithmic
scale used in Figure 34. This form of plotting permits an estimate to
be made of the critical discharge for initiation of bed-load transport.
This was done on Figure 37 by extending the linear portion of the
bed-load transport curve to the ordinate axis at zero bed-load
transport. The critical discharge for incipient motion was found to be
43 cfs.
This critical discharge would be relative to the armor layer, which
has a fairly uniform particle size (low index of uniformity). Using
this discharge with a channel slope of 0.009 ft/ft from Table 4,
channel width and depth from Figure 13, and approximate mean particle
diameter of 2 1/4 inches (57 mm) from Table 9, an estimation of the
critical shear stress was made and compared to data given in Figure 6-
11 (p. 170) of the text by Leopold, Wolman, and Miller (1964). A
150
-------
Table 15
ADJUSTMENT OF BED-LOAD TRANSPORT DATA FOR
TRAP EFFICIENCY OF SAND-SIZE MATERIAL,
OAK CREEK, WINTER 1969-1970
Sample
Identity
Number
Total Bed-Load Transport,
Ib / hr
Measured
Adjusted
Bed-load Transport for
Material Smaller than No. 4
Sieve Size, Ib / hr
Measured
Adjusted
2
3
8
9
10
26.0
441
101
82.9
5.7
127
1059
331
183
17.7
18.6
81.3
21.0
19.7
2.4
120
699
Z51
120
14
151
-------
200
IO.OOO
Figure 34. Adjusted total bed-load transport rate, Oak Creek, winter 1969-70.
-------
Cn
O-l
Figure 35. Superimposed adjusted bed-load transport curves, Oak Creek, winter 1969-70.
-------
It"
U
D 6
w 4c U
tv
20 t-
J L
' t \ 1 I 1 t
J
J L
5 10
UWIT
Figure 36. Unit bed-load transport rate as function of discharge, Oak Creek, winter 1969-70,
-------
)OO
SO
C/1
Ul
u
til
i..
6C.
1 r
1 r
I T
i <_
iOO
60O
goo
^ v_8./HR.
Figure 37. Estimated critical discharge to initiate bed-load transport near Oak Creek
weir/trap structure.
-------
portion of that figure has been reproduced in Figure 38 with the
calculated point for Oak Creek armor material added (a calculated
critical shear stress of about 0.6 pound per square foot). The
plotted point from the Oak Creek data lies well within the range of
other data shown in the figure.
Several measures of particle size were given in Table 13 for the bed-
load samples. These characterizing diameters and derived parameters
are plotted in relation to their corresponding bed-load transport
rates in Figure 39. The diagrams are based on the measured data, and
are in error because they lack any trap efficiency correction. Even
with these errors, certain tentative conclusions for Oak Creek bed-load
can be drawn. These are:
1. The mean size of the bed-load (Dso) increased with transport rate
to a limiting value of about 14 mm (0.55 inch). (The points above
the line would shift to the right and down after correction for trap
efficiency.) In comparison, the mean size of the material below the
armor layer was estimated to be about 20 mm (Table 10). Hence, the
bed load was "finer" than the streambed from which it derived.
2. Except at low transport rates, the index of uniformity was in
the range 10 to 20. In comparison, measurements of the bed material
indicated the armor layer had a uniformity index of 2.3 to 2.8 and
the bed material a uniformity index of 5 to 8. Hence, the bed load
was less uniform than the streambed.
3. The D&5 size also approached a constant value of from 20 to
30 mm. Although D6s is considered an index of bed roughness, its
exact significance for moving material is uncertain.
4. The maximum size as indicated by DBS did not increase with transport
rate but reached a constant value of from 40 to 60 mm.
At this point, no further conclusions can be made without further
analysis (of the type planned for the third year of the original
project).
Oak Creek Painted Gravel Experiment
To obtain some idea of the distance that individual particles move in
Oak Creek during a period of high flow, a collection of rocks was
painted yellow and placed in the stream. They were located on a bar at
the upper end of the instrumented reach (at section 18). In this
experiment, five nominal sizes were used: 2 inch, 1 1/2 inch, 3/4 inch,
and 0.19 inch (No. 4 sieve size). Each particle was weighed and all
were placed in a group on the surface of the streambed. Upon placement,
the smaller particles tended to fall between the particles of the
natural armor layer.
On 16-17 February 1970, an isolated storm-runoff event took place that
moved bed material. After the flow subsided, a reconnaissance was
made to locate the yellow rock. A summary of the results of the
experiment is presented in Table 16. Graphs of the distance moved
156
-------
1,000
500
200
100
t/t
Jj
f 20
1 10
o
=6 5
c
I
2
1
05
0.2
0.1
A
-
-a
i
a
A
A *
' a <
0 Q^
CO
&
>
>s
CMC
KYJiv
A A
• •>
« 0
>
e*
GK
6
* »0
'
**~
.+
0
o
o
0 1
0 °
0
w
a
o • U.5.W.E.5.
0 Chang
O Not! Bur Stds.
* Kramer
• Indri
^ Chitty Ho
* Krey
* Prussian Exp. Inst.
O Engels
o Fahnestock, 1963,
White River, Washington
i i i —
D.OOI 0.002 0.005 0.01 0.02 0.05 O.I 0.2 0.5 1.0 2 5 10
IC; critical shear stress, pounds per square foot
Figure 38. Laboratory and field data on criticalshear stress
required to initiate movement of particles (after Leopold,
Wolman and Miller, 1964, Figure 6-11, p. '170).
157
-------
i
jl.0
I
5
r
-41
&.vi\Ji»rrvjN of MriMe IbtJ) Plaints int o* 8tO toMJ
W ITV1 6tO*U9At) TtAWiftA
4o»
0V IMOW 0» Ul»m»«T
-1 h
1.*
i i-
*
S°-sr *
<. -1
to.o-
1 b
0.5-
_J 1 I l L_
\OW ZOO "*-JO
BtD -uoAD T»iANSPOi2.T R.ftTfT _ LB- /«*•
Figure 39. Variations in characterizing particle sizes
for bed-load samples as function of bed-load transport
rate, Oak Creek, winter 1969-70.
158
-------
Table 16
PAINTED-GRAVEL EXPERIMENT, OAK CREEK,
16-17 FEBRUARY 1970
Particle
Size,
Inches
2
1 1/2
3/4
3/8
0.19
Fraction of Sample
Moved
0.83
0.75
0.92
0.92
1.00
Recovered
0.42
0.33
0.33
0.25
0
Mean Weight, gtn
Placed
370
146
44
6.6
0.7
Recovered
276
144
50
6.6
—
Distance
Moved,
feet(1>
160
160
112
125
?
Particle
Velocity,
ft/hr<2)
6.7
6.7
4.7
5.2
?
(1) The mean distance moved for all recovered particles of the indicated size which actually
did move.
(2) Calculated assuming that particles only moved at discharges exceeding 40 cfs, based on
critical discharge for incipient motion discussed in connection with Figure 37. The time
interval for such movement was 24 hours.
159
-------
and the probability of movement are given in Figure 40 as a function
of particle weight.
Because of the low percentage of particles recovered, very little
information on the distance moved can be extracted from the data. The
data suggest that the larger particles move farther and faster than the
smaller particles. The smaller particles were found on the bottom of
pools and the larger particles on bars or transitions between bars
and pools. This fact suggests that the larger particles are carried
rapidly through pool areas, but the smaller particles may deposit in
such pools and be protected from further disturbance. The grain-size
data for the streambed support this observation, as the bed was
somewhat finer in composition in pools than on bars.
Another interesting result shown in Figure 40 is that the probability
of movement increases as the particle size decreases (for the flow
condition at the site where the rocks were placed), even though some
of the smaller particles were "protected" by the larger particles.
Importance of Bed Load in Total Sediment Yield
Sufficient information on sediment transport at Oak Creek was obtained
to develop some rather tentative relations for suspended-load and bed-
load transport. These were presented in Figures 25B and 36,
respectively. From these two curves, a third relationship can be derived
that expresses the ratio of bed-load to suspended-load transport rate
as a function of streamflow. This relationship is presented in
Figure 41. The ratio increases with discharge toward some limiting
value, which suggests that at quite large flows the bed load approaches
but does not quite match the suspended load in magnitude. At very
small flows, the ratio becomes very small as the bed load approaches
zero.
To obtain some idea of the importance of bed load during a storm
runoff event, three runoff events were selected that exhibited low,
intermediate, and high peak discharges. The curves in Figures 25B and
36 were used to estimate the total amount of sediment transported past
the gaging station during the runoff event. The hydrographs used are
shown in Figures 42A, 42B, and 42C. A summary of the results obtained
is 'given in Table 17.
The data in Table 17 suggest that the bed load is of minor importance
for low-peak runoff events, but that the importance of the bed load
increases as the peak discharge increases. This relation suggests
that a change in land use that would increase only the peak flow would
also increase the importance of the bed load. An increase in bed-load
transport, even for a couple of years, could result in stream-channel
changes that would persist long after the peak flows had returned to
the previous levels.
160
-------
(U
u.
e
Si
v>
o
Loo
Ul
•>
»
A.
B.
BX
MQVE1WEVJT.
uoc51% or
Figure 40. Distance moved and probability of movement in
painted-gravel experiment, Oak Creek, 16-17 February 1970.
161
-------
200
10
4? ISO
8
3
too
*
5O-
I III
J I I ( I I 11
IIIF!)'
I I I I I 1 I I
1 \ I
I I I
Figure 41. Bed-load/suspended-load relation as function of
discharge, Oak Creek, winter 1969-70.
162
-------
•701
Figure 42. Hydrographs used to estimate sediment yield, Oak Creek, winter 1969-70.
-------
Table 17
BED LOAD, SUSPENDED LOAD, AND TOTAL SEDIMENT YIELD
FOR STORM-RUNOFF PERIODS, OAK CREEK, WINTER
1969-1970
Peak
Discharge,
cfs
Bed-Load
Transport,
Ib,
Suspended-Load
Transport,
Ib.
Total
Sediment
Yield,
Ib.
Percent
of Total
Yield as
Bed-Load
Bed-Load
suspended -Load
Ratio
36
(low peak)
60
(intermediate
peak)
175
{high peak)
24
3,070
50,800
2,400
18,430
142,900
2,424
21,500
193, 700
<1
14
26
0.01
0.17
0.36
164
-------
SECTION X
BED MEASUREMENTS, ALSEA EXPERIMENTAL WATERSHEDS
Dear Creek Streambed Materials
Because considerable work has been done to measure suspended sediment
discharges in the Alsea experimental watersheds, it was thought
desirable to obtain some idea of the bed materials in an alluvial area
on one of the basins. Therefore, a portion of Deer Creek was sampled.
Sampling was done where the creek flows through an open meadow area
and in a reach of the stream where the sediment sizes and type (except
at one sampling location) are controlled by the stream itself and not
by the movement of material into the channel by mass movements. The
one exception (the sampling section where hydraulic flows do not
control the sediment properties) was just upstream of a "channel-
control" reach where mass movement into the stream is possibly
important. This section (18+00) had numerous boulders in the stream
and the bordering valley slopes are steep close to the stream. All
of the samples were obtained on low bars at bends in the stream, with
similar locations used at each bar to provide consistent sampling for
later comparisons of data.
The sampling technique was to use an apparatus similar to the one
described by McNeil and Ahnell (1964). The procedure was to set a
6-inch-diameter pipe on the bed surface, remove the armor layer and
save it in a separate sample bag, and then sample the bed material to
a depth of about eight inches. The material was removed from the
pipe into a concentric holding barrel welded to the pipe and the pipe
was advanced into the streambed as the material was removed. Using
this procedure, little material was lost and any lost material was
probably smaller than No. 100 sieve size. Because the sample of
armor material collected at each point with the 6-inch pipe was too
small for reliable analysis, additional samples were composited to
increase the sample size, all such samples being obtained from the
armor layer adjacent to the main sampling location.
Particle-size data for the samples collected from Deer Creek are
presented in Table 18. Graphs of mean particle size and of the
variation of the fraction of bed material coarser than 3/4 inch
as a function of distance are presented in Figure 43.
The sampled bed materials have an upper size limit of less than three
inches. A distinct difference is evident between the materials in the
surface layer of the streambed and those beneath. Except at station
18+00, there is a distinct tendency for particle size to decrease in
the downstream direction. Presumably this trend is attributable to
particle breakdown during periods of bed-load transport—the
sedimentary-type rock at the Alsea basins does not appear to produce
gravel as durable as that from the sedimentary and basaltic rock in the
Oak Creek watershed. The increase in particle size at station 18+00
165
-------
Table 18
DEER CREEK STREAMBED MATERIALS IN SELECTED ALLUVIAL AREAS.
Fraction
18+00
Sieve
Size
3 -in.
2 -in.
1 1/2 -in.
3/4-in.
3/8 -in.
No. 4
No. 8
No. 16
No. 30
No. 50
No. 100
Pan
DSO,
MM
MM'
D60,
MM
MM'
U<2)
(1) Sampled
(2) U=D60/
Armor
Layer
0
.1645
.2751
.7718
.9660
.9928
.9963
.9970
.9974
.9980
.9988
1.0000
30
36
35
14
2.5
to a depth
Dio
Sub-
SurfaceU)
0
.0969
.0969
.3411
.5559
.7167
.7948
.8346
.8692
.9429
.9872
1.0000
4
9.6
7.6
0.24
32.6
of Sample Retained by Indicated Sieve at Sampling Location
28+00
Armor
Layer
--
--
0
.3756
.5990
.7304
.8191
.8664
.9133
.9751
.9939
1.0000
16
20
18
0.88
20.5
of 8 inches after removal
Sub-
Surface
--
--
0
.0907
.3092
.4807
.5871
.6541
.7228
.8924
.9670
1.0000
4
8
6.8
0.27
25.2
of armor
36+76
Armor
Layer
0
.1501
.3564
.8840
.9636
.9852
.9937
.9962
.9972
.9985
.9994
1.0000
38
39
39
18
2.16
(surface)
Sub-
Surface
—
0
.0903
.2635
.4649
.6457
.7586
.8183
.8733
..9618
.9885
1.0000
8.5
17
14
0.55
25.4
layer.
37+OO
Armor
•Layer
--
0
.1422
.6483
.8751
.9469
.9749
.9832
.9872
.9937
.9973
1.0000
27
31
30
8
3.75
Sub-
Surface
—
—
0
.2928
.5145
.6519
.7359
.7916
.8477
.9543
.9859
1,0000
10
16
15
0.38
39. S
43+34
Armor
Layer
--
—
0
.8241
.9229
.9568
.9763
.9858
.9911
.9951
.9974
1.0000
25
30
28
11
2.55
Sub-
Surface
--
0
.1769
.3856
.5290
.6506
.7532
.8315
.8959
.9528
.9812
1.0000
11
21
18
0.58
31.0
44+00
Armor
Layer
--
0
.3020
.7786
.9233
.9652
.9790
.9843
.9890
.9941
.9973
1.0000
30
35
34
12
2.84
Sub-
Surface
0
.1641
.2135
.4906
.6720
.7952
.8408
.8608
.8878
.9558
.9855
1.000
18
28
22
0.65
33.8
-------
30 -
20
2
<
•>
10 -
or NVEAW pf\fL-pc.Le si-ze ALON g
0.8
0
!••'
a
2 .. i
o °- *
of
o
IN
Figure 43. Streambed material characteristics, selected
alluvial portions of Deer Creek.
167
-------
may be attributable to mass movements of sediment adjacent to the
channel, which locally contributes large particles to the streambed.
Implications of'Oak Creek Data for Deer Creek
Deer Creek has an armored streambed but with somewhat smaller
surface material than for Oak Creek. The fact that the Deer Creek
streambed has such an armoring indicates that in such reaches of the
creek a critical discharge must be reached before general bed-load
transport of an appreciable magnitude. Comparatively more of the bed
material at the sampled Deer Creek reaches is in the sand-size range.
Therefore, the armor layer in Deer Creek may exert a stronger
controlling effect upon the suspended load at high streamflows, when
the finer bed material may actually move in suspension. Further
examination is required of the Alsea experimental basins and of data
already collected before other implications of Oak Creek research
can be offered with adequate confidence.
168
-------
SECTION XI
REFERENCES
Anderson, H. W., Suspended sediment discharge as related to stream
flow, topography, soil, and land use, Transactions of
American Geophysical'Union, Vol. 35, p. 268, 1954.
Leopold, L. B., M. G. Wolman, and J. P. Miller, Fluvial Processes
in Geomorphology , W. H. Freeman and Co., San Francisco, pp.
170, 235, 1964 .~~
Love, S. K. and P. C. Benedict, Discharge and sediment loads in the
Boise River drainage basin, Idaho, 1939-1940, U. S. Geological
Survey Water Supply Paper 1048, Washington, D. C., 1948.
McNeil, W. J. and W. H. Ahnell, Success of pink salmon spawning
relative to size of spawning bed materials, U. S. Fish and
Wildlife Service Special Scientific Report, Fisheries No. 469,
January 1964.
Robinson, A. R., Vortex tube sand trap, Transactions of American
Society of Civil Engineers, Vol. 127 Part III, Paper 3371,
391-433. 1962.
Roehl J. W., Sediment source areas, delivery ratios, and influencing
morphological factors, Publication 59, International Association
of Scientific Hydrology, 1962.
Smith, D. D. and W. H. Wischmeier, Factors affecting sheet and rill
erosion. Transactions of American Geophysical Union. Vol. 38,
pp. 889-896, 1957.
U. S. Geological Survey, Magnitude and frequency of floods in the
United States: Part 14, Pacific slope basins in Oregon and lower
Columbia River basins, Water Supply Paper 1689, Washington, D.C.,
1964.
U. S. Weather Bureau, Climatological Data, Oregon, December 1969, also
January 1970, February 1970, Environmental Sciences Service
Administration, Washington, D.C.
169
-------
APPENDIX
The Alsea Watershed Study was launched as a broadly interdisciplinary
undertaking and formed the backdrop for most of the work reported here.
The Alsea Project is a case study utilizing experimental watersheds
located in the Oregon Coast Range. Many investigators, agencies, and
members of the University staff attempted to pool resources to study
not only land use but also the physical, chemical, and biological
effects of logging operations in an area where timber, water, and
fisheries are all important resources.
The effort reported here is only part of a progression of work already
reported from the Alsea and related studies. Some of this published
research is to be found referenced at the end of the chapters.
This report clearly shows the shift in research from the case study
approach, such as for temperature and sediment effects of logging, to
that of studying pertinent processes. The type of research contribution
each has made is clearly seen. The case studies on water temperature
provided orders of magnitude of specific treatment effects, while
process studies provided insight for more universal adaptation and
prediction. Both were complementary. Water temperature related to
cover conditions is currently being studied by Dr. Brown under another
grant from the Environmental Protection Agency.
A similar research approach was undertaken with sediment production
and land use. First the case study was undertaken, then a shift
to a study of processes to provide predictions in small streams and
watersheds. This proved to be a far more difficult undertaking. The
original proposal for the work contained in this report was for a
three-year period; hence, the report should be examined in this
context. Dr. Hall's work on biological effects of water temperature
changes was terminated and then transferred from the conjunctive effort
and therefore will be reported elsewhere. Discontinuance of financial
support for the sediment project at the end of the second year also
prevented accomplishment of all original objectives, yet substantial
contributions are reported.
Chapters I and II have been published elsewhere and Chapter V will
appear in a national journal in the near future.
The project leaders of the studies undertaken, Drs. Brown, Hall, and
Klingeman, undertook to meet their objectives diligently. The results
reported here along with corollary efforts should contribute in a
major way to our understanding of environmental change from logging the
watersheds of small streams.
James T. Krygier
Principal Investigator
171
-------
ACKNOWLEDGMENTS
The research was sponsored by the Federal Water Quality Administration,
under Grant WP-423, and Oregon State University. Cooperators in
those parts associated with the Alsea Logging-Aquatic Resources
Study include: School of Forestry, Department of Civil Engineering,
Department of Fisheries and Wildlife, and Water Resources Research
Institute, all of Oregon State University; Oregon Game Commission;
U.S. Geological Survey; U.S. Forest Service; Georgia-Pacific Corporation;
Stokes Lumber Company; and F. W. Williamson.
The authors are particularly indebted to L. C. Beck and Dr. F. L.
Ramsey, Department of Statistics, Oregon State University, for their
assistance in the time series analysis of a portion of our temperature
data; to Dr. Henry Anderson and Dr. Scott Overton, who provided
assistance in reduction and analysis of the sediment data; and to
Wayne Hug of the Oregon Game Commission, who gave five years of effort
in intensive sediment sampling. The U.S. Geological Survey was very
cooperative in making their records available.
Chapters I and V were authored by Dr. George W. Brown and Dr. James T.
Krygier, School of Forestry, Oregon State University; Chapters II, III,
and IV were by Dr. George W. Brown; and Chapter VI was authored by
Dr. Peter C. Klingeman, School of Engineering, Oregon State University.
173
-------
1
At'ccstiioti Number
w
5
/^ Sttbjvct F/e/d &. Group
5A
SELECTED WATER RESOURCES ABSTRACTS
INPUT TRANSACTION FORM
Organization
School of Fores frv ('also Dent, of Civil En«O
Forest Research Laboratory
Oregon State University
7'if/e
Studies on Effects of Watershed Practices on Streams
10
Authors)
Krygier, James T.
Brown, George W.
Klingeman, Peter C.
1 z Project Designation
EPA, WQO Grant No. 13010 EGA 02/71
21
Note
22
Citation
23
Descriptors (Starred First)
*Thermal Pollution, *Sediment Yield, Sediment Transport, Watershed Management.
25
Identifiers (Starred First)
*Water Pollution Sources
27
Abstract
A number of studies were undertaken related to effects of clearcut
logging on water quality and the process affected in small streams.
Water temperature studied before and after logging was increased
significantly where stream cover was removed. Energy balances of small
streams were measured and predictive models were developed.
Road building significantly increased sediment yield in clearcut and
patch cut watersheds. Logging itself was not an important sediment
contributor.
Methods for sampling bed load and suspended sediment were developed.
Bed load constituted 70 percent of suspended load during peak discharges.
Abstractor
James T. Krygier
Institution
Oregon State University
WR: 102 IRE v.
WRSIC
SEND WITH COPY OF DOCUMENT. TO: WATER RESOURCES SCIENTIFIC INFORMATION CENTER
U.S. DEPARTMENT OF THE INTERIOR
WASHINGTON. D. C. 20240
«U.S. GOVERNMENT PRINTING OFFICE: 1972-484-'82/« 1-3
-------
ENVIRONMENTAL PROTECTION AGENCY
Publications Distribution Section
Route 8, Box 116, Hwy. 70, West
Raleigh, North Carolina 27607
Official Business
i POSTAGE AND FEES PAID
ENVIRONMENTAL PROTECTION AGENCY
------- |