&EPA
EPA 910/9-87-162
United States
Environmental Protection
Agency
Region 10
1200 Sixth Avenue
Seattle WA
Alaska
Idaho
Oregon
Washing M
Water Division
Ai.ril 1987
Development of Criteria for
Fine Sediment in the
Northern Rockies Ecoregion
Final Report
4* x-^-»J • -- -.. f-f 3
-------
DEVELOPMENT OF CRITERIA FOR FINE SEDIMENT
IN THE
ECOREGION
by
D. H. Chapman
and
K, P. McLeod
Final Report
Work Assignment 2-73
Battelle Columbus Laboratories
EPA Contract No. 68-01-6986
Contractor:
Don Chapman Consultants, Inc.
3180 Airport Way
Boise, Idaho 83705
March 10, 1987
-------
ACKNOWLEDGMENT
The U. S. Environmental Protection Agency, Region 10, would like to
extend its appreciation and acknowledge the time and efforts of the Technical
Advisory Committee which participated in the development of this document, EPA
910/9-87-162. The Technical Advisory Committee provided technical review and
comment on the plan of work and draft document. He also gratefully acknowledge
support of the Criteria and Standards Division, U. S. Environmental Protection
Agency in Hashington, D. C. , especially John Maxted.
The Technical Advisory Committee was composed of experts in the area of
salmonld ecology from the Pacific and Intermountain Northwest. It met several
times In Boise, Idaho to discuss the draft document for Its technical merits.
Dr. Chapman and Ms. McCloud were always in attendance to benefit from the
often spirited discussions. The professional manner in which everyone
conducted themselves was exemplary and appreciated. The Technical Advisory
Committee consisted of the following individuals;
the
Susan Martin/Steve Bauer
Idaho Department of Health and Welfare
Pat Murphy
Nez Perce Tribe
Dr. Peter Bisson
Weyerhaeuser Company
Dr. William S. Platts
Intermountain Forest and Range
Experiment Station
U. S. Forest Service
Al Espinosa
Clearwater National Forest
U. S. Forest Service
Dr. Mostafa A. Shirazi
Environmental Research Laboratory
Corval1 is
U. S. Environmental Protection Agency
Donald M. Martin, Co-Chair
Region 10
U. S. Environmental Protection Agency
Virgil Moore/Russ Thurow
Idaho Department of Fish and Game
Dr. Dale McCullough
Columbia River Inter-Tribal Fish
Commi ssion
Dr. Ted Bjornn
Idaho Cooperative Fisheries
Dr. Dave Burns
Payette National Forest
U. S. Fish and Wildlife Service
Roy Heberger
Ecological Services, Boise
U. S. Fish and Wildlife Service
Rick Albright, Co-Chair
Region 10
U. S. Environmental Protection
Agency
-------
CONTENTS
Topic Page
I. INTRODUCTION 1
II. INTRAGRAVEL ENVIRONMENT AND SPAWNING SUCCESS 5
A. MEASURES OF STREAMBED CHARACTER IN SPAWNING AREAS 5
A.I. Percentage of fines 5
A.2. Geometric mean particle size 10
A.3. Stratified analysis of sieved samples 16
A.4. The fredle index 21
A.5. Visual assessment 24
A.6. Permeability and apparent velocity 29
A.7. Timing and location of -samples 34
B. PHYSICAL ENVIRONMENT IN GRAVELS USED FOR SPAWNING 37
B.I. Redd structure 37
B.2. Intrusion of fines into gravel 45
B.3. Porosity, permeability and water movement 52
C. INTRAGRAVEL ECOLOGY OF SALMONID EMBRYOS 54
C.I Apparent velocity 54
C.2 Permeability 56
C.3 Dissolved oxygen 57
C.4 Fines 70
D. EMERGENCE FROM GRAVELS 85
D.I Entrapment by fines 85
D.2 Effects of fines on size of emergent fry 88
D.3 Effects of fines on emergence timing 92
III. SUBSTRATE CHARACTER AND ECOLOGY OF REARING 97
SALMONIDS
A. SUBSTRATE CHARACTERIZATION 97
A.I Visual assessment 97
A.2 Photographic measurement 102
A.3 Core samples 103
A.4 Embeddedness 106
A.5 Sediment traps 114
B. SUBSTRATE UTILIZATION BY REARING SALMONIDS 116
B.I Newly-emergent fry 116
B.2 Parr and fingerlings 116
B.3 Adults 143
-------
C. INSECT ABUNDANCE 145
C.I Insect density 145
C.2 Insect drift 158
D. RUBBLE COVER FOR SALMONID WINTER HIDING 164
D.I Effects of temperature on substrate use 164
D.2 Winter carrying capacity 168
D.3 Habitat and species differences 171
D.4 Conditions in the Northern Rockies 178
IV. FINE SEDIMENTS AND CHANNEL MORPHOLOGY 182
A. AGGRADATION AND DEGRADATION 182
B. ROLE OF BED MATERIAL IN GOVERNING MORPHOLOGY 186
OF STREAMS
C. WOODY DEBRIS AND CHANNEL MORPHOLOGY 188
D. RELATIVE ROLE OF STREAM MORPHOLOGY AND FINES 191
V. TOOLS FOR PREDICTING EFFECTS OF FINE SEDIMENTS 192
ON FISH AND AQUATIC MACROINVERTEBRATES
A. INTRAGRAVEL SURVIVAL OF EMBRYOS TO EMERGENCE 192
OF ALEVINS
A.I Fredle index 197
A.2 Geometric mean particle size 201
A.3 Percentage of fines 203
A.4 Permeability 208
A.5 Electronic probe measurement of pore velocity 210
A.6 Dissolved oxygen assessment 211
A.7 Best available information 211
B. SUBSTRATE FINES AND EFFECT ON REARING DENSITIES 216
B.I Embeddedness 216
B.2 Effects of fines on rearing densities of fish 220
B.3 Substrate score and fish density 221
B.4 Best available information 221
C. MACROINVERTEBRATE RESPONSE TO FINES 222
C.I Insect density as a function of embeddedness 222
and fines
C.2 Insect drift as a function of embeddedness 224
and fines
C.3 Best available information 224
D. EFFECTS OF FINES ON WINTER HABITAT OF SALMONIDS 225
D.I Embeddedness and winter habitat 225
D.2 Best available information 227
ii
-------
E. MONITORING ' 229
VI. GENETIC RISK, BIOLOGICAL COMPENSATION AND 230
LIMITING FACTORS
A. GENETIC STRESS ON POPULATIONS 230
B. ROLE OF BIOLOGICAL COMPENSATION 231
C. LIMITING FACTORS 232
VII. PREDICTIVE TOOLS, MANAGEMENT, AND REGULATORY 239
UTILITY
A. PREDICTORS FOR FISH 239
B. PREDICTORS FOR MACROINVERTEBRATES 241
C. MEASURES OF LAND USE 242
D. ROLE OF JUDGEMENT IN CRITERIA FOR MANAGEMENT 246
PRACTICES
VIII. SUMMARY OF RESEARCH NEEDS 249
A. INTRAGRAVEL ENVIRONMENT 249
B. REARING HABITAT 251
C. WINTERING HABITAT 252
D. EXPERIMENTAL MANIPULATION OF FINES 253
IX. SUMMARY 254
X. LITERATURE CITED 251
111
-------
I. INTRODUCTION
Early work on fine sediment and fish ecology began with
Harrison (1923), who reported low survival of sockeye salmon in
gravels with high percentages of fines. Hobbs (1937) conducted
field studies of sediment effects on salmonids in some New
Zealand streams, evaluating the relationship between embryo
mortality and fines in the substrate. Shapovalov and Berrian
(1940) and Shaw and Maga (1943) reported effects of sediment on
survival in the redd and effects of mining silt, respectively.
Stuart (1953) experimented in the laboratory on effects of silt
on early stages of brown trout in Scotland. Campbell (1954)
combined laboratory and field exercises by placing rainbow trout
embryos in baskets in a stream affected by gold dredging and in a
control stream. Shelton (1955), in simulated field conditions,
showed that fines reduced Chinook salmon emergence success.
These early efforts were soon supplemented by a veritable
flood of laboratory and field studies on effects of fines in
spawning gravel on salmonid survival (McDonald and Shepard
(1955), Wickett (1954), Cooper (1956), Terhune (1958), Alderdice
et al. (1958), Cordons and Kelley (1961), Coble (1961), Phillips
and Campbell (1962), McNeil (1962), Vaux (1962), Bianchi (1963),
Silver et al. (1963), McNeil and Ahnell (1964), Shumway et al.
(1964), Cooper (1965), Mason and Chapman (1965), Koski (1966),
Mason (1969) and Hall and Lantz (1969)).
These early studies led to consensus that low dissolved
oxygen and reduced water exchange in laboratory environments
caused reduced survival. It was also apparent that incubation
success could decline in the intragravel environment in field
situations in which fines accreted. Furthermore, laboratory
studies demonstrated that emergence of alevins declined as
percentage of fines increased in experimental mixtures of
1 /SECTION I
-------
substrate. Field study also demonstrated that emergence was
lower in gravels with high percentages of fines.
From 1969 to the present, research on sediment in the redd
environment tended toward corroboration and refinement of the
relationship between emergence success and percentage of fines
Bjornn et al. (1977), Koski (1975, 1981), Sowden and Power
(1985), Tagart (1976 and 1984), Phillips et al. (1975), Cederholm
et al. (1981), Cederholm and Salo (1979), Tappel and Bjornn
(1983), Irving and Bjornn (1984). In addition to percentage-of-
fines analyses, embeddedness indices (Klamt 1976, Kelley and
Dettman 1980) have been compared to fish abundance (Bjornn et al
1977, Thurow and Burns, unpublished).
Some effort has been devoted to the effects of sediment on
other life history stages. Juvenile salmonid density has been
correlated with living space reductions caused by sedimentation
(Bjornn et al. (1977), Stuehrenberg (1975), Klamt (1976),
Cederholm and Reid (1986), Everest et al. (1986)), not because of
embeddedness but because of lost pool volume. Chapman and Bjornn
(1969), Hartman (1965), Rimmer et al. (1983), and Campbell and
Neuner (1985) noted the importance of winter hiding locations for
juvenile salmonids within the stream substrate. Hillman et al.
(1986) demonstrated that addition of rubble piles to a sedimented
stream increased winter density of Chinook salmon by several-
fold.
Effects of sedimentation on macroinvertebrate density or
drift in streams used by juvenile salmonids have been evaluated
by Bjornn et al. (1977), Martin (1976), Brusven and Prather
(1974), and Konopacky (1984), sometimes with equivocal results,
with environmental features other than sedimentation sometimes
confounding results.
Some work has addressed the difficult task of defining
2 /SECTION I
-------
criteria for healthy and degraded intragravel environments
(Bjornn et al. (1977), Shirazi et al. (1981), Stowell et al.
(1983), Klamt (1976), Platts and Megahan (1975), and Everest et
al. (1986)). Most of the information that underlies these
criteria is derived from laboratory situations and, usually,
single-factor analyses.
A basic confusion in establishing criteria in ways
meaningful to fish ecology has been the difficulty inherent in
establishing which environmental variable(s) act as limiting
factors. [For example, sedimentation in spawning gravels alone
may not reduce populations of adult salmonids if winter hiding
space controls carryover of juveniles from one year to the next].
Although we address this problem in 'our report, we contend that
criteria suitable for evaluation of watershed management prac-
tices may not necessarily relate directly to limiting factors.
That is, a habitat criterion for evaluation of best management
practices may serve solely to indicate streambed composition in
areas used for salmonid spawning, rearing, or winter hiding. Any
criterion that we may suggest should have demonstrable relation-
ships to ecological requirements of salmonids or of the food base
for salmonids, but may not necessarily measure a limiting factor.
A major problem in field situations is that of quantifying
effects of sedimentation in the face of natural variability and
extreme hydrologic or meteorologic events. Another major obstacle
to development of meaningful models is the difficulty of, or lack
of awareness of need for, sampling within the salmonid egg
pocket. In laboratory studies, the principal problem is that of
approximate duplication of field conditions.
In spite of these and other difficulties, it is appropriate
at this time to critically review sedimentation criteria and
progress toward management techniques for instream evaluation of
sedimentation. The following review is divided into sections on
3 /SECTION I
-------
the environments for reproduction and for rearing and winter
ecology, fine sediments and channel morphology, tools for
prediction of effects of fines, a discussion section on genetic
risk, biological compensation and limiting factors, and a
treatment of predictors and management regulation. We also offer
some research recommendations.
/SECTION I
-------
II. INTRAGRAVEL ENVIRONMENT AND SPAWNING SUCCESS
A. MEASURES OF STREAMBED CHARACTER IN SPAWNING AREAS
Measures of streambed character offer an attractive possible
means for assessment of effects of land management practices on
habitat. Percentages by weight of particle sizes smaller than
some specified level, such as <1 mm in diameter, might serve as
useful indices of habitat quality. visual scoring techniques
might prove useful. Geometric mean particle size offers another
measure, as do the fredle index and permeability.
A.I. Percentage of fines
Percentages of fines in stream substrata may be examined by
means of sieving samples withdrawn from the streambed in frozen
cores (Walkotten 1976) or excavated within a cylinder that sur-
rounds a bottom area and core (McNeil and Ahnell 1964). In
either case, extracted samples are wet- or dry-sieved through a
set of wire meshes and weighed by volume or measured by volume of
water displaced.
Adams and Beschta (1980) examined 5 streams in the Oregon
Coast Range to assess temporal and spatial variation in streambed
composition, as well as factors affecting the amount of fine
sediment in the streambed. Adams and Beschta used a freeze-core
system (Walkotten 1976), which froze two cores at each sample
location to 25 cm in depth. The cores for each location were
combined for analysis. These workers sampled at the same sites
over time, maintaining records of earlier sample locations to
prevent repeat sampling of the same specific location. Records
were kept on stream profile to document scour and fill, and
stream gages were used to index size and sequence of storms.
Large particles (>50.8 mm) were removed from samples and
from the analysis of percentage of fines, in order to reduce
variance. Although Adams and Beschta justify this on the basis
5 /SECTION II
-------
that other workers have excluded large rocks from analysis
(McNeil and Ahnell 1964, Wendling 1978), the subsequent treatment
is artificial to an unknown degree. Chapman et al. (1986)
rejected use of the freeze-core method for a spawning area on the
mid-Columbia River because the relatively small-diameter freeze
cores (obtained with single-probes) were excessively influenced
by presence or absence of large particles. They adopted use of
samples excavated within a cylinder 50 cm in diameter.
Ringler (1970) reported that freeze-core and McNeil samplers
yielded somewhat different results, but was unable to explain the
differences. Ringler (1970) compared freeze core samples with 15
cm core samples from redds in Drift Creek, an Alsea River tribu-
tary. He found that the freeze core underestimated the fines
smaller than 0.83 mm and 3.3 mm.
Shirazi et al. (1981) tested freeze-core samples with single
probe against a McNeil sampler in Berry Creek, Oregon, and found
that in this coarse-substrate stream, a single probe tended to
produce data with a smaller geometric mean than did the 30 cm
McNeil core sampler (Table A.I). They also compared results from
Table A.I. (From Shirazi et al. 1981). Comparison of results
from freezecore and McNeil samples.
Samp
d mm
g
Iie * treezecore
~977
2 42
3 42
4
^ 33.9
5 - 37
Mean 33.9
12" Core
32J 1
45.5
45.0
27.0
66
42.4
(should be 43.2)
/SECTION II
-------
a 3-tube freeze core sampler with those from a 30 cm McNeil core
sampler. Although the mean dg of 10 such tests was only about
10% higher for the 30 cm core, there were considerable differ-
ences between specific samples (Table A.2).
Table A. 2. (From Shirazi et al. 1981). Comparison of tri-tube
freezecore and 12-inch (30 cm) manual core samples taken from
Rogue River, Oregon.
Sample ID
Bridge Hole
Hatchery Site
Sand Hole
8lg Butta
Mean
B
C
A
a
c
A
B
C
A
C
d mm
g
Tri-core
30.4
24.0
42.1
18.0
69.6
34.2
22.6
46.0
15.5
19.3
327T
12" core
30.9
17.3
21.8
18.6
48.4
58.7
58.9-
65.3
22.9
19.5
30
Shirazi et al. (1981) concluded, based on case studies, that
by combining the analysis of many single freeze core samples from
a spawning site having relatively fine textured substrate, one
can obtain a reasonable estimate of the mean composition of that
site. They also stated that single freeze cores, particularly
from a coarse substrate with a single tube, may not be a good
representation of gravel composition, but that a 3-tube system
provides a good representation of a gravel patch over the range
of grain sizes up to 100 mm. We note that diameters of particles
in many substrata used by adult Chinook salmon and steelhead
often exceed 100 mm. The egg pocket often contains these larger
particles. Platts et al (1979) found that about 18% of the
particles "in" a Chinook salmon egg pocket consisted of particles
larger than 100 mm. These workers used a definition of the egg
pocket somewhat broader than that used elsewhere in our review.
7 /SECTION II
-------
The principal finding of Adams and Beschta (1980) in their
five-stream study was that the fraction of particles smaller than
1 mm was varied greatly in time and space. During a 19-month
sampling period, flushing of fines from gravels during high flows
caused temporal variability. Percentage of fines varied between
streams, between locations in the same stream, and between
locations within the same riffle. Even in streams with a
relatively high sediment content, channel roughness features such
large organic debris, boulder elements, and channel constrictions
can sort gravels and create "islands" of clean gravel (Everest et
al. 1986).
If percentage of fines (or any other physical measure of
substrate quality) serve to gauge stream guality, the time of
sampling is important (Adams and Beschta 1980). Freshets and
spawning fish can clean fines from the stream bed, and freshets
must be accompanied by disturbance of the gravel bed. In other
words, flushing will not occur in interstices of the substrate
unless that substrate moves (Beschta and Jackson 1979). On the
other hand, reduction of fines does not always result from a
movement of streambed. The net cleaning of fines from any
location during a storm event depends on both disturbance of the
bed and on re-intrusion of fines when the bed stabilizes.
Adams and Beschta (1980) recommend that if the investigator
can sample a stream only once, this should occur during low flow
when beds have stabilized. They suggest that if fines serve to
index stream guality for fish, the bed should be sampled during
the period when eggs of fish lie in the substrate. Following
this line further, we conclude that one should sample in redd egg
pockets during the period when embryos are incubating within the
pockets, in order to evaluate survival and gravels in a pertinent
time stratum. The investigator must decide whether he or she
should evaluate fines midway through the incubation period or at
8 /SECTION II
-------
the conclusion of emergence.
In addition to the 5-stream study, Adams and Beschta (1980}
obtained liquid nitrogen cores from riffle substrata in a total
of 21 streams in the Coast Range in both disturbed and undis-
turbed drainages. Fines content averaged 19.4% (range of 10,6 to
49.3%). In 5 undisturbed streams, fines ranged from 10.6 to
29.4%. In about 75% of all comparisons between plots within the
same stream, the percentages of fines differed significantly
(p =.05). Fines content parallel and normal to flow differed
significantly (p = .05).
In the 21 streams examined by Adams and Beschta (1980),
fines varied with depth in the substrate. In 59 cores, the top
10 cm had significantly (p = .05) less fines than the 10-25 and
25-40 cm zones. Fines for the 0-10, 10-25, and 25-40 cm
respective depths averaged 17.4, 22.3, and 22.2%, The authors
suggest that surface armoring (Milhous and Klingeman 1973) may
have reduced fines in the shallowest stratum.
The authors noted that this spatial variability in bed
composition may prohibit a simple characterization of gravel bed
guality within a given area or an individual channel. They state
that the large number of samples needed would also prohibit
precise estimates of fines content, although the meaning of this
statement is not clear, for one should expect more precise esti-
mates with increased sample size.
Table A.3, which is Table 3 from Adams and Beschta, provides
regressions, developed from various data on 21 streams, and that
relate fines to slope, area, relief, and land use (r2 = 0.66).
In an undisturbed drainage (Flynn Creek) fines were related to
sinuosity and bank-full stage (r2 = 0.820, and to sinuosity alone
(r2 = 0.74).
9 /SECTION II
-------
Table A.3. (From Adams and Beschta 1980). Regressions of fines
content (PF) in percent by weight for streambed cores frozen with
liquid nitrogen. * indicates significance at p = 0.05 after
application of correlation tests of linearity (t-test) and
multiple linearity (F test).
No.
obs«r-
Model
1
2
3
vattona
21
7
7
Regression equation I
PF - — 1 . 23 (Average watershed, slope. %)
+0.03 (Watershed aren.haj
4- 68. 89 (Watershed relief ratio, m/m)
+0.09 (tflnd-us< factor, %)
+41. 94 (Constant, %)
PF - +27.02 (Sinuosity, m/m)
+ 33.22 (Bunk-lull )t«ge, m)
-31.04 (Constant, %)
PF - +38-83 (Sinuosity, m/m}
-24.94 (Conslanl, %)
F R*
• 0.66
• 0.82
• • 0. 74
Adams and Beschta (1980) warn that temporal variations in
percentage of fines may obscure effects of land use, and that
more knowledge of in-stream processes that change gravel
composition is needed,
A.2. Geometric mean particle size
Platts et al. (1979) adapted several works to propose use of
geometric mean particle diameter as a companion measurement to
"percent fines" for a more complete analysis of sediments. They
suggest that geometric mean particle size is a statistic more
amenable to statistical analysis than percent fines. They
defined geometric mean diameter as:
dg = (d16d84>1//2 where
d^g = particle diameter corresponding to the 16th percentile,
d§4 = particle diameter corresponding to the 84th percentile.
Everest et al. (1981) defined dg as:
dg = (d^l x d2w2 x dnWn) ' wnere
dn = mid-point diameter of particles retained on the n*-n sieve,
wn = decimal fraction by weight of particles retained on the
n*-" sieve.
10 /SECTION II
-------
Platts et al. (1979) related sediment porosity to dg by
using Cooper's (1965) data (Figure A.I). They also compared dg
for several areas of the South Fork Salmon River, and found that
those in upper reaches had a higher dg than those in lower
reaches (Table A.4), but were unable to determine if this was a
result of availability of sediments or preferences of fish.
.10
.n
.11
.10
>-• • ,iiw:
nt
>
.01
mil!
Mtt litto fna ti61« t in
tilt* M C«W(I«M
Figure A.I. (From Platts et al. 1979).
sediment porosity and d.
Relationship between
Platts et al. (1979) lifted a complete egg pocket from a
chinook salmon redd by using multiple freeze-cores, and measured
the dg of the strata in the pocket as 39.2, 20.1, and 35.2 mm for
the top to 15 cm, 15-30 cm, and 30-45 cm respective strata. These
data constitute the only available information on composition of
the egg pocket.
11
/SECTION II
-------
Table A.4. (From Platts et al. 1979). Variation of geometric mean
diameter of spawning sediments collected from channel reaches in
the South Fork Salmon River in 1976.
Site d (mm)
Downstream reaches:
Glory Hole 9.6
Oxbow Area 8.5 Average = 8.8
Poverty Flat Area 8.4
Upstream reaches:
Dollar Creek Area 13.5 . _ ld 7
Stolle Meadow Area 15.8 /werage - n. /
Platts et al. (1979) list advantages of using dg as:
1. It is a conventional statistical measure used in sedimentary
petrology and engineering to describe sediment composition.
2. It is a convenient standard measure that enables comparison
of sediment sample results between two studies.
3. It may be calculated from £04 and d^g, two parameters that
may also serve for calculation of the standard deviation.
4. It relates to permeability and porosity of channel sediments
at least as well as percentage of fines.
5. It is a more complete description of total sediment
composition than percentage of fines and sediment composition
evaluations in many cases involve less sampling error using
V
6. Because it relates to porosity and permeability, it is
potentially a suitable unifying measure of channel substrate
condition as it affects embryo survival.
Shirazi et al. (1981) related percentage embryo survival to
substrate composition as expressed in dg (Figures A.2 and A.3).
12 /SECTION II
-------
IVU
90
eo
70
> 60
E
w 50
0
re
3O
20
10
Koill t9«
PMINfif (III. 1979
1978
Platt
Dtllt Cr»k,OR
Lobototwy
CMtoHcfei I
Lllttll* 1974
O SiKltri Ct>op«t 1999
4 Slnftnarf PMItlpi tl al. 1979
A SlMUmrf Ctfftrhohii im4
tlllfltl 1974
__J |
Lobemlorr •
Labatalarf
JL
0 9 '0 19 2O 23 30
GEOMETRIC MEAM DIAMETER, d, (mm)
Figure A.2. (From Shirazi et al. 1981). Relationship between
embryo survival and substrate composition expressed as d.
100
90
9O
7O
60
50
40
3O
2O
10
Ke.H I9fi«
Phlinpt •! of 1979
Togar! 19 7 B
Cld«ho(m I
LIII til* 1974
Phillip. (ta*. 1979
C*Htfholm and
LtXdtt 1974
JL
JL
0 I.O 2.0 3O 1.0 5.0 6.0
GEOMETRIC MEAN DIAMETER/ MEAN EGO DIAMETER,
7.0
-------
We note that Platts et al. (1979) reported dg equaled 23 mm
for a Chinook salmon egg pocket that had been frozen in place and
extracted for laboratory analysis. We used their gravel composi-
tion data to calculate a dg of 31.2 mm. We think the 23 mm cal-
culation is a typographic inversion, and are strengthened in this
by the authors' calculations of dg for 6-inch strata that show a
dg of 39.2, 20.1, and 35.2 mm for the 18-inch deep sample.
A dg of 32 mm falls at the extreme upper end of the model in
Figure A. 2. The authors noted that this dg was somewhat large
compared with similar measurements taken by other means both in
1978 and previously. They attributed the difference to coarsen-
ing of the substrate by the spawning fish, relative to the sur-
rounding substrate, and of course this is true. Average dg in
the Poverty area of the South Fork Salmon River was reported by
Platts et al. (1979) as 8.4 mm without correction for water on
the samples, and 11.9 mm with the correction, which they note
would be a correction of 42%. The dg of the egg pocket was well
over twice that of the Poverty substrate. The egg pocket, as
withdrawn intact, also included the large particles at the
centrum of the pocket.
Even the egg pocket data of Platts et al. probably incorp-
orated portions of the redd outside the actual egg pocket, and
the pocket itself would likely have a makeup different from that
of the sample. We will repeatedly note in this report that egg
pockets differ greatly from the surrounding substrate in gravel
composition.
The data in Figure A. 2 show fairly conclusively that high
survivals tend to occur in association with high dg. However,
Tappel and Bjornn (1983) concluded that the usefulness of the dg
is limited because gravel mixtures with the same geometric mean
can have different size compositions. Lotspeich and Everest
14 /SECTION II
-------
(1981) also noted that dg may not differ in gravels of vastly
different makeup. Tappel and Bjornn noted, after an analysis of
100 samples of spawning gravels from the South Fork Salmon River,
that some samples had substantial deviations from lognormality
and were not accurately represented by the regressions. These
deviate samples curved upward in the upper end of cumulative
distribution plots (Figure A.4). The correlation improved if
particles larger than 25.4 mm were deleted from the plot.
A much more serious problem with the data in figures A.2-A.4
is that dg for the cited studies was calculated for McNeil
corings that may or may not have included egg pockets, even
though taken on redds. Another procedural problem is that McNeil
corings may, in redds or in "spawning gravels" either extend too
deeply, into consolidated substrate, or too shallowly, missing
• 95
I • 90
«••»
2 • 70
30
«. 0
o >
Sc
Mini
I I I tIItl
.51 5 10 50 100
Particle size (mm)
Figure A.4. (From Tappel and Bjornn 1983). Typical deviation
from lognormal particle size distribution in gravels from the
South Fork Salmon River. Note upper end of the cumulative
distribution plot. Solid line is for all particle sizes (r2 =
0.89); broken line is for material smaller than 25.4 mm in
diameter (r2 = 0.97).
15
/SECTION II
-------
the lowest portion of gravels that make up the environment of
salmonid embryos. Use of triple-probe freeze cores would
probably minimize this problem in samples within redds because
the field worker can visually examine the frozen core to
determine whether it probed the relatively consolidated materials
at depth, or failed to reach the bottom of an egg pocket.
A.3. Stratified analysis of sieved samples
Tappel and Bjornn also analyzed data from 126 salmon
spawning areas sampled by Cederholm et al. (1977) in the
Clearwater River drainage in Washington, finding r2 values for
these of 0.97 when particles larger than 26.9 mm were eliminated
from the samples.
Tappel and Bjornn (1983) proposed a regression equation for
particle sizes smaller than 25.4 mm as:
PERCENT = C + K logeSI2E where
PERCENT = inverse probability transformation of % of
substrate smaller than a given sieve size,
C = intercept of regression line,
K = coefficient of variable logeSIZEr
SIZE = sieve size in mm.
Tappel and Bjornn (1983) state that because the distribution of
particles less than 25.4 mm in diameter can be represented with a
regression with r2 close to 1.0, lines passing through data
points for two sieve sizes closely approximate lines determined
by the least squares regression procedure. They showed that the
line passing through data points for the 9.5 and 0.85 mm particle
sizes closely approximated the line from regression procedures.
Plots of gravel samples from the South Fork Salmon River and
from the Clearwater River (Washington) were prepared (Figure
A.5). Points A and B represent two different spawning gravel
samples. The vertical line through A and B represents a
16 /SECTION II
-------
30
• 0
20
• £
ss
0.
• South Fork Salmon Rlv«r, Idaho
* Claarwatar River system, Washington
* ** •
• * * *
.• * **
• • • I
*•* • • i
• ** •
,*. • •* •
• •* « •
. • • •
•»«•.• .
•
*•..•'.*•.•••:"*•••*{' •• :* *
..*•«•***• ;"-r*-:; • '.
•*.*•*.*• • B
JL
J_
J.
J.
10 20 30 40 50
Percentage of substrate smaller than 9.50 mm
v *
A / *
• • •
•
60
Figure A.5. (From Tappel and Bjornn 1983). Range of spawning
gravel sizes for samples from two river systems. Line AB
represents mixtures with the same percent fines (50% smaller than
9.5 mm) and dg but different size compositions.
continuum of gravel size compositions, all with 50% of the
substrate less than 9.5 mm in diameter. If particles less than
9.5 mm are termed "fines", any data points on the line AB
represent samples of spawning gravel with the same percent fines
but different particle size distributions.
Tappel and Bjornn (1983) further showed that although
gravels A and B both had a d« of 9.5 mm, they had very different
16th and 84th percentile diameters. The 16th percentile, for
example, was 0.95 mm for gravel A and 2.2 mm for gravel B.
Tappel and Bjornn (1983) proposed to test embryo survivals
in relation to gravel mixtures on the basis of two substrate
variables; diameters 9.5 and 0.85 mm. They suggested that this
technique removed the need to define exactly which gravel sizes
harm embryos, and it only requires the assumption that gravel
material larger than 25.4 mm is not harmful to incubating
17
/SECTION II
-------
embryos. It also requires the assumption that particles larger
than 25.4 mm are not beneficialf and this again raises the point
that for many salmonids, particularly for anadromous fish, the
egg pocket centrum in natural redds is formed of particles much
larger than 25.4 mm. We will repeatedly note, later in this
report, that laboratory studies simplify of necessity, and that
results obtained from them cannot serve to predict conditions in
natural egg pockets in quantitative fashion.
Tappel and Bjornn (1983) prepared gravel mixes as shown in
Figure A. 6 to cover the range of natural spawning substrata
shown in Figure A. 5. We caution that the range of gravels
referred to the spawning area, not to egg pockets. The results
of survival tests to emergence for 15 mixtures, each test
composed of 2-3 replicates, permitted the development of a model
of survival (Figures A.7 and A.8). Isolines of survivals were
placed through the test results, and second-order equations of
30
m 6
I6
| a 20
• 6
og
• 5
01 **
2 •
S110
2 E
_*• .*
•&§••
e. ''f-t^"f ri 4
^:**?,/i %
^;j
Jfc«*ii'' t r ,. •- 'f' N >
^X^v" 4AW * -:,
i*K/" , x K ^ v i- t f-v •-,.*•! >' '
^^ ^r-r \\<^r SBH* ^
.'T^ioiii'>v,> ^*(i;: .
^*>>t^-apiiz ^;> ,> - i&i*
^r^^^^^r^4?*^:
!••," '
t
10 20 3Q 40 50
Percentage of substrate smaller than 9.50 mm
60
Figure A.6. (From Tappel and Bjornn 1983). Experimental gravel
mixes overlying range of natural spawning substrata. Labels
correspond to respective gravel compositions.
18
/SECTION II
-------
survival were prepared:
For steelhead,
Percent survival = 94.7
and for Chinook salmon,
Percent survival » 93,4
- 0.11639^530^35 + 0.007
3-8<7s0.85
The equations had r2 values of 0.90 and 0.93, respectively.
The authors noted that green steelhead eggs were used in the
tests but that Chinook embryos were eyed at the start of the
experiment. Therefore, and in accord with data from Bjornn
(1969), the survival isolines should be shifted toward the origin
for chinook by an unknown amount.
Figure A. 9 shows that percent survival to emergence was
about 90% when dg exceeded 10 mm, while Shirazi and Seim (1979)
reported survival generally was less than 90% unless dg exceeded
15 mm. Tappel and Bjornn (1983) showed that in gravels with
identical dg, survivals were higher than those reported by
Shirazi and Seim. Tappel and Bjornn thought this was because of
10 20 30 40 SO
Percentage of substrate smaller than 9.SO mm
Figure A.7. (From Tappel and Bjornn 1983). Isolines showing
survival predictions (0-80%) for steelhead embryos under various
gravel combinations. Scattered numbers are empirical results.
19
/SECTION II
-------
30r
2 e
• o
2 1
= 10
2 I
1O 20 30 40 SO
Percentage of iub«lra(* imaller (Nan 9.30 mm
80
Figure a.8. (From Tappel and Bjornn 1983). Isolines showing
survival predictions for Chinook salmon embryos under various
gravel mixes. Scattered numbers are empirical survivals.
differences in gravel composition not described by dg alone. We
submit that dg in the Tappel and Sjornn information actually more
d
survivals were obtained.
some laboratory cells, but include many samples in the immediate
area of the redd or in the redd, not necessarily in the egg
pocket.
closely reflected dg in the laboratory environments in which
The Shirazi and Seim data reflect dg of
In spite of the excellent laboratory work done by Tappel and
Bjornn (1983), we state that their results cannot serve to
predict field survivals of embryos (eg. Talbert 1983, 1985a,
1985b) without field verification of the laboratory results.
Such verification should consist of measurements of survival to
emergence in egg pockets of natural redds.
20
/SECTION II
-------
9 100
g
a. 80
I
o
**
13
60-
40
" 20
c
u
9
CL
10 20 30
Geometric mean (mm)
Figure A.9. (From Tappel and Bjornn 1983). Relation between
geometric mean of each experimental gravel mixture and steelhead
and Chinook salmon embryo survival. Solid line fitted by eye to
laboratory tests. Broken line is survival curve from Shirazi and
Seim (1979).
A.4. The fredle index
A "fredle index" was developed by Lotspeich and Everest
(1981). The index, dg/Sg, where dg is geometric mean particle
size and
is geometric standard deviation, is stated to
indirectly characterize the pore size and permeability of
substrates. The index has been suggested as an integrated
measure of substrate suitability for development of embryos.
Lotspeich and Everest cautioned that the gravel samples used to
develop the fredle index should be taken from locations known as
spawning sites and no deeper than the depth of egg deposition.
This stratification precaution should be applied to other sub-
strate sampling as well, as noted by Burns (1984). We go one
step beyond Lotspeich and Everest by stating that for purposes of
predicting survival in natural redds, the fredle index must be
developed with data from natural egg pockets, and we discuss the
reasons in section II.B,
To date, no evidence from field sampling has been presented
to establish utility of the fredle index in a variety of field
21 /SECTION II
-------
situations. Sowden and Power (1985) correlated the index with
rainbow trout survival in groundwater-fed streams in southwestern
Ontario, and found no significant correlation until redds with
mean dissolved oxygen lower than 5.3 mg/1 and extreme gradients
were removed from the analysis. When this was done (Table A. 5
and Figure A. 10}, correlation improved. The correlations do not
include the emergence phase.
Table A.5. (From Sowden and Power 1985).'
Correlation coefficients between percent survival of preemergent rainbow trout embryos (arcsine-tram-
fortned) and percentage of fine sediments less than 2.0 mm in diameter (fines), the geometric mean particle
siit (VJ, fredle indices of substrate quality fDg&f}, and permeability for two groupings of redds. DO denotes
mean dissolved oxygen concentration and I denotes hydraulic gradient; asterisks denote *P s 0.05; "P <
0.01.
Redd grouping Aresln(% lines) Dr D^st Log^permeability)
AH redds iff - 19) -0.3175 0.4090 0.3620 0.3604
Redd* with mean
DO s 5.3 mg/L
indO,0»s /30.I2 -0.7618 0.8457 0.9049* 0,9826**
Perusal of Table A.5 indicates that correlation coefficients
equalled -0.76, 0.84, 0.90, and 0.98 for independent variables of
percentage of fines < 2.0 mm, geometric mean particle size, the
fredle index, and permeability, respectively. The latter two
variables provided significant regressions.
Sowden and Power explain the need for stratification by
noting that survival in redds was not described by the measures
of substrate composition tested because the oxygen content of
groundwater varied independently of substrate makeup, and
substrate composition accounted for only a limited portion of the
variance in groundwater velocities.
Sowden and Power did their work well. Unfortunately, the
results do not carry survivals to the emergence phase, and,
although a single freeze-core was taken from redds, we have no
way to determine if they were extracted from egg pockets.
22 /SECTION II
-------
40
09
*
O
cc
20
to
Figure A.10. (From Sowden and Power 1985). Relation between
survival of preemergent rainbow trout embryos and the fredle
index for redd substrates. Redds for which mean dissolved oxygen
content was 5.3 mg/1 or greater are noted by dark and split
circles. Open circles denote redds in which dissolved oxygen
content was less than 5.3 mg/1.
Lotspeich and Everest (1981) calculated fredle indices for
laboratory data of Phillips et al. (1975) (Figure A.li). These
data included only the intragravel period of "swim-up" fry to
emerging fry, the period after "button-up". Figure A.li must be
|
M
JO-
to-
* 3 4 • 1 T
Prviil* Index ff|)
Figure A.li. (From Lotspeich and Everest 1981). Survival from
introduction of swim-up fry to emergence from gravel mixtures
with various fredle numbers. The underlying data on survivals
were those of Phillips et al. (1975).
23
/SECTION II
-------
used with caution. Lotspeich and Everest called the models in
Figure A. 11 "preliminary", and identified the next task as
development of relationships between fredle numbers and survival
to emergence in natural gravel mixtures. We go one important
step further, stating that the relationships should be between
embryo survival to emergence and fredle numbers in egg pockets.
Tn the absence of field corroboration and correlation of
embryo survival to emergence with the fredle index, it appears
premature to assess the fredle index, then to predict survival of
species of interest as has been done for steelhead in Pete King
and Deadman creeks in the Idaho batholith, for example (Talbert
1983, I985a, 1985b). However desired are tools for survival
prediction, an extrapolation of this extent should be avoided.
This is not to say that the fredle index does not offer a useful
measure of gravel character. But the next step to predictions of
survivals is premature without field verification.
A.5. Visual assessment
All of the measurements described in A.I through A.4 require
coring and sieving of substrate samples. Shirazi and Seim (1981)
suggested that natural variability of gravels in space and time
makes desirable a rapid visual examination of gravel heterogen-
eity. Such a measure would reduce sampling cost and effort.
These authors worked in the field in Oregon and Alaska to test
utility of the method. At each sample location, the spawning
area was visually divided into three groupings based on apparent
composition of surface gravels: coarse, fine, and intermediate.
A single 30 cm core sample was collected within each area and
field-sieved for analysis.
When mean particle diameters in bed material of neighboring
patches differed by about 10%, based on the visual procedure,
Shirazi and Seim were able to correctly identify the coarser
material 87% of the time. When differences were about 20%, the
24 /SECTION II
-------
visual estimation of relatively coarser material was correct 93%
of the time.
Shirazi and Seim (1981) proposed to use the visual stratifi-
cation system as a means of allocating gravel sampling effort,
not as a method of characterizing the gravels in detail. It is
logical that the surface appearance of gravels may be quite
different in watersheds of different hydrologic pattern (eg.
coastal rain forests or interior plateau or northern Rockies)
with varying degrees of surface armoring, making a unifying
visual separation suspect or impossible. However, if the visual
stratification is used solely as a means of allocating core
samples, risk is minimized.
Use of a visual system for sediment classification requires
familiarity with the stream system and some objective means of
characterizing surface appearance. Platts and Megahan (1975)
assessed long-term trends in sediments in the South Fork Salmon
River by using visual classification of the substrate surface
into groups as noted in Table A.6.
Table A.6. (from Platts and Megahan 1975). Size classification
of riverbed materials in ocular evaluation of substrate.
Particle diameter Classification
12 inches or over (304.8 mm or over) Boulder
3 to 11.9 inches (76.1 to 304.7 mm) Rubble
0.19 to 2.9 inches (4.7 to 76.0 mm) Gravel
0.18 inch and less (less than 4.7 mm) Fines (Sand)
They used groups of channel cross-sections within major spawning
areas, and evaluated composition of bed materials from waterline
to waterline along the sections. They visually projected each
30-cm division of a measuring tape to the bed surface and
assigned the observed sediment to one of the 4 sediment classes.
25 /SECTION II
-------
The visual examination of substrate was reported for 9 years
of data (figures A.12 and A.13), and showed gradual decreases in
fines (and concomitant increases in amounts of larger particles).
Fw»r«Y
— — ;i(rilt
-~ Kimri
Otory
Yf AM Hi YIAM ELAPSED SIMCt IMS
Figure A.12. (From Platts and Megahan 1975). Trends in percent
fines in spawning areas of the South Fork Salmon River as
determined visually.
Although the error bands for each relationship are wide, as
indicated by low coefficients of determination, the fitted
regressions are significant (Platts and Megahan 1975).
Although not available for the same time period as the data
reported by Platts and Megahan (1975), core samples reported by
Corley and Newberry (1982) show a downward trend in proportions
of fines (< 6.3 mm) from 1975-1981 in the main South Fork (espec-
ially since 1979) , but no trends in the Stolle Meadows site in
the upper river. These data somewhat loosely support the
26
/SECTION II
-------
n
so
8
ro
* NubWt • 10.14 - 1104 r* • 0.0]
19M
HI
IMf
tJI
I9M
01
it; i
181
HI i»» m m
VIAM3 «r«* VCAHS IIAPSCD SIHCC 198S
-------
than 4.75 mm were compared with core samples. The photo method
resulted in assessment of a mean of 13.2% fines but failed to
distinguish fines of size less than 0.85 mm. The author
concluded that the photo method was viable. However, the Payette
National Forest has subsequently abandoned photo and visual
assessment techniques (D. Burns, personal communication).
Scrivener and Brownlee (1981) analyzed over 1000 gravel
cores (freeze cores with dry ice and acetone) from Carnation
Creek, Vancouver Island, and found that fines were more abundant
deeper in the substrate than at the gravel surface (Figure A.14).
They also found evidence of seasonal trends in fines, with fines
accumulating in the substrate during low flow periods and being
removed by spawning fish or freshets in the higher flow periods.
This underscores the importance of timing of sampling in sub-
strate evaluations.
Ringler (1970) reported a variable pattern of vertical
stratification of sediments in freeze-cores extracted from coho
salmon redds in three tributaries of Drift Creek, an Alsea River
drainage. Fines smaller than 0.83 mm and smaller than 3.33 mm
tended to be more abundant in the top 1/3 of a 25 cm core segment
than in deeper strata in two streams but not in the third.
The disparate findings of Scrivener and Brownlee and of
Ringler should, at minimum, be carefully considered by workers
seeking to evaluate sediments based on surface conditions.
Evaluation of surface character on the same location each year
and at the same time would at least reduce variability associated
with time and space, although leaving unresolved the fact that
surface condition will not accurately or even approximately
reflect conditions deeper in the substrate, or differences in
depths of intrusions of fines.
28 /SECTION II
-------
l-Tff
(976-78
TIME PERIODS
Figure A.14. (From Scrivener and Brownlee 1981). Mean percent of
particles smaller than 9.55 and 2.38 mm in top, middle, and
bottom layers of gravel cores in Carnation Creek.
A. 6. Permeability and apparent velocity
Pollard (1955) described the theory of flow through spawning
gravels. He defined porosity as the ratio of volume of voids to
the total volume of solids plus voids. Hydraulic gradient is the
pressure head drop, or head loss, per unit length of stream, or
the slope of the hydraulic grade line. Apparent velocity is the
rate of seepage of water expressed as the volume of liquid flow
per unit time through a unit area (solids plus voids) normal to
flow direction. Apparent velocity is sometimes called the super-
ficial or macroscopic velocity. True velocity, or pore velocity,
is the actual velocity of flow through the interstitial spaces,
and differs from pore to pore.
Permeability, which is more properly termed Darcy's coef-
ficient of permeability, is the constant of proportionality, K,
in the function below:
29
/SECTION II
-------
V = KS, where
V is apparent velocity of groundwater and S is hydraulic grad-
ient. K has the dimension of length (or distance) per unit of
time. Permeability is a packing indicator, so that apparent
velocity is proportional to packing and gradient. The looser the
gravel and the higher the gradient, the faster water will flow
through the substrate. Permeability of a gravel can be measured
by forcing water through a sample of it.
Pollard (1955), and Terhune (1958) described a standpipe and
associated equipment for measurement of apparent velocity and
permeability. Apparent velocity measurements required a visual
comparison of water samples withdrawn at various time intervals
from a standpipe "cell" about 25 cm deep in the substrate.
Variability in this technique is often high, and apparent vel-
ocity measurements made by the techniques described by Terhune
(1958) have utility principally for purposes other than for
descriptions of suitability of gravels for spawning (see next
major section for applications of apparent velocity measurement)„
Gravel permeability indices have been described by Cooper
(1965) and Pollard (1955). Cooper shows beta, the gravel perm-
eability, as a * f_g^J\/" ' \ u ^
f =
He defines a permeability function as
Cooper conducted tests with various gravels, demonstrating
the relationship between sediment addition, time, and permea-
bility. He showed a reduction in permeability as gravels trapped
sediment over time. In general, the rate at which sediment was
removed from surface water was directly related to permeability
30 /SECTION II
-------
of the gravel. Less permeable gravels trapped more fines. Silt-
laden water can deposit fines more readily in the voids, where
low velocity permits silt to settle. The egg pocket is more
porous and permeable at the conclusion of redd construction than
is the case outside the pocket. It will trap fines at a rate
different from the retention rate in areas surrounding the redd
outside the egg pocket or in the gravels termed by many workers
"spawning habitat" or "spawning gravels."
Pollard (1955) showed that permeability declines as porosity
declines when gravel is packed more and more tightly (Figure A.
15). Pollard demonstrated the relationship between volume of
water pumped from a standpipe well under one-inch suction head
and gravel permeability in a laboratory permeameter (Figure
A.16). The standpipe, later thoroughly discussed by Terhune
(1958), is an artifice for creating, with minimum gravel
disturbance, a cavity 25 cm below the surface of a streambed, or
roughly at the depth of a salmon egg pocket (Figure A.17).
Terhune provided a plot, developed from a laboratory permeameter,
of permeability (K) in relation to volume of water pumped per
unit of time (Figure A.18).
a «
at
O.I
0.09
0.04
o.oz
O 01
"Jlil'
I i
!£=
id
100
1000
Permeability
(em/tir)
Figure A. 15. (From Pollard 1955). Log-log plot of permeability
against porosity for one gravel bed successively compacted.
31
/SECTION II
-------
u
w
1.0
0.1
-(#
/'
»
-------
10'
w
-
/
/
^
, 1
/
t
^
Ji
^
o n ' • •
O(ml/iM] h.
\J—
P
f
1
I
-
: =1
-"
)
UWTATtON
I/XW
Figure A.18. (From Terhune 1958). Permeability (K) versus rate
of inflow (Q) to Mark VI groundwater standpipe at one inch head.
Platts et al. (1979) extracted data from Cooper (1965) to
relate permeability and dg (Figure A.19). Permeability was
directly related to dg with a high (90%) coefficient of deter-
mination. Platts et al. used this relationship to help explain
why dg offered a unifying statistic for gravel description. One
could also use it in the reverse, to demonstrate that permea-
bility is a measure of gravel composition. We later contend that
permeability offers a useful tool for correlations with survival
and for assessment of fines intrusions in egg pockets, as well as
for evaluation of land mangagement practices.
Beta, the permeability noted on the vertical axis of Figure
A. 19, differs from the permeability function (K) in Pollard
(1955) and Terhune (1958). Beta is defined by porosity, a shape
factor, and a sum of fractions of particles by weight of each
size class, and has a centimeter dimension, while K lumps several
factors and has a dimension of cm/hr.
33
/SECTION II
-------
.n
. .20
c
12
yj. O.OOSMj1**
»
KM PMTICII DIMETER, 4 . unttMtert
Figure A.19. (From Platts et al. 1979). Permeability as a
function of geometric mean particle diameter (dg).
A.7. Timing and location of samples
Burns (personal communication), in discussing embeddedness
(see section III) statistics, made the point that he directed
embeddedness samples to a particular stratum, namely to rubble or
cobble areas under laminar flow conditions. In a technical
sense, true laminar flow would not occur over a rubble substrate,
but the point is that selection of smoothly-flowing areas over a
particular substrate type narrows sampling focus and reduces
variance among embeddedness samples. In evaluations of spawning
34
/SECTION II
-------
gravels, it is pointless to sample deep pools with a boulder
substrate, as an extreme example.
Cederholm et al. (1977) extended this reasoning in the
Clearwater River system in Washington by sampling with a core
sampler only in riffles known to be selected by spawning
salmonids. This approach automatically reduced scope of sampling
and numbers of samples required to estimate parameters of
spawning gravels. McNeil and Ahnell (1964) obtained core samples
only from spawning beds. Lund (1985) core-sampled only in known
spawning areas.
We would carry the stratification process on by noting that
if the study objective is to assess conditions in the substrate
as they control salmonid incubation and emergence, samples should
come only from natural egg pockets. If the objective is to
determine overall condition of "spawning gravels" without
reference to embryo survival, then a somewhat less-restrictive
stratification could suffice.
Whether one uses permeability, coring, or even visual scor-
ing to describe condition of spawning gravels, time of sampling
may also have a major impact on results. For spring spawners,
gravel conditions present after the peak of the hydrograph prob-
ably best represent conditions faced by the female when she
selects a redd site and begins excavating. For fall spawners,
conditions in September or October are more appropriate.
Condition of gravel before redd construction does not indi-
cate that present during incubation or emergence. The spawning
female purges the redd of a fraction of the fines present in
undisturbed gravel, as noted in the next major section. Hence,
coring or permeability data may or may not correlate with
embryonic survival, depending upon when the data were obtained,
35 /SECTION II
-------
quite apart from problems associated with failure to sample in
the egg pocket.
36 /SECTION II
-------
B. PHYSICAL ENVIRONMENT IN GRAVELS USED FOR SPAWNING
We see the structure of salmonid redds as critical to the
role of fine sediments in affecting survival of incubating
embryos and emergence success. We believe that many field and
laboratory researchers have failed to understand redd structure,
with pervasive effects upon utility of data. The following dis-
cussion supports these statements.
B.I. Redd structure
Burner (1951) described the shape and size of redds con-
structed by Pacific salmon. Figure B.I shows development of a
chinook salmon redd over time and locations of ova.
Pfan
\ \ •<>-;—
--.NX -*-
Figure B.I. (From Burner 1951). Diagrammatic views of a fall
chinook salmon redd that was measured daily.
37
/SECTION II
-------
Burner's figure shows the position of the ova within the redd.
They lie well upstream from the crest of the tailspill.
Hawke (1978) excavated chinook salmon redds in New Zealand.
Although his methods section did not explicitly describe the
tailspill, his definition of a redd as the total area excavated
by a fish, and the area in which ova lie, implicitly means that
he did not include the tailspill in the term "redd". His dia-
grams show that deposited embryos lie not beneath the tailspill
crest, but upstream in a series of 5-6 egg pockets within the
disturbed area of the substrate (Figure B.2). Observations of
Hobbs (1937) support this point.
'CSIO
* i i 1 4 i I i »
• Itnglh (m)
Figure B.2. (From Hawke 1978). Positions of egg pockets in 7
redds of Oncorhynchus tshawytscha in a New Zealand river.
38
/SECTION II
-------
The redd begins as an initial pocket from which the female
has removed fine materials by the lifting action created by
turning on her side and vigorously flexing her body. Current
helps carry the lifted fines downstream; the finest particles
travel well downstream and gravels move into a pile or low ridge
below the pocket.
The largest particles in the substrate cannot be lifted by
the female; these form the clean egg pocket centrum, commonly a
grouping of 3 or 4 large gravel or cobble particles. The female
deposits the first group of eggs into this centrum and the male
simultaneously fertilizes them. The eddying currents within the
pocket (Figure B.3) probably help retain sperm in contact with
the eggs. The female quickly begins digging upstream from the
Figure B.3. (From Burner 1951). Drawing of currents within a
Chinook salmon redd.
first pocket, both directly and obliquely upstream. Current
again carries the finer materials downstream below the redd, and
gravels lifted from this newly-excavated area drop into the first
egg pocket or onto the tailspill, depending upon size of the
excavated material. The female then prepares a new egg pocket
with a centrum of several large particles, cleaned of fine
39
/SECTION II
-------
materials, and the egg deposition and fertilization process
continues.
Vronskiy (1972) confirmed characteristics of the egg pocket
centrum by noting: "One interesting structural feature of most
Chinook redds is the presence of one or two large stones (15-23
cm in diameter) lying on the bottom of the redd; the bulk of the
eggs are concentrated around them." He excavated redds to reach
this conclusion.
Hawke (1978) showed that pocket placement progresses up-
stream, but that excavation also extends normal to streamflow and
to the pocket line (because of oblique digging to cover the egg
centrum). Not all digging in the redd serves to cover earlier
egg pockets or even to construct new ones. Females apparently
direct some digging at testing the substrate for suitability.
Not all initial pockets are used for egg deposition, and females
may move to a new area for actual spawning.
When the final, most upstream egg pocket has been prepared,
digging upstream covers the pocket, but no identifiable new
single pocket is excavated. Rather, several shallow excavations
often appear directly and obliquely upstream.
The point of the foregoing description is to clarify that
the redd in longitudinal section is a series of pockets, the
bottoms (or "floors") of which consist of undisturbed streambed.
Several large gravel or rubble particles that the female cannot,.
or does not, move form the centrum of the pocket, and small and
medium gravels lie among, around, and over the centrum (Hawke
1978).
The female removes fines from an area much wider than the
pockets themselves as she obliquely digs to cover deposited eggs
and test for the next pocket location. Thus fine particles of
40 /SECTION II
-------
silt and sand are substantially decreased in the completed redd.
Some sand, probably tending to consist of larger-diameter sands
from upstream pocket preparation and final digging, drops back
into the redd. One might expect, in theory, that these fines
would most likely fall back into the gravel bed in which the
downstream egg pockets lie, rather than into the area occupied by
the most upstream pockets, but turbulence and eddying within the
redd prevent any clear longitudinal stratification. Possibly the
waning energy of the spawning female as she deposits and covers
the last eggs would lead to reduced movement of larger sands out
of the redd, further obscuring stratification. Hawke (1978)
notes that the first egg pockets are deeper than later ones, and
disturbed gravel from the last pockets tends to depo- sit toward
the lower end of the redd. Hobbs (1937) described the "floor" of
the redd by noting that the substratum of undisturbed material
underlying the redd falls steadily from the upstream end of the
redd down to the point where the female commenced work.
Egg pocket depth ranged from 18 to 43 cm for Chinook salmon
redds excavated by Hawke (1978) , and 8-22 cm for brown trout
(Hardy 1963) . Ova tend to be concentrated at the bottom of the
egg pocket (Hawke 1978). Dr. Fred Everest (personal communi-
cation) recorded the vertical stratification of ova in freeze-
core samples taken from chinook salmon redds in the Rogue River.
He found that the eggs lay in a stratum 2-3 cm thick just above
the undisturbed streambed at the bottom of the egg pocket.
Although stray eggs lay higher in the redd matrix, the bulk of
the ova were in the deepest portion of the' redd. A photo of an
egg pocket in Everest et al. (1986), and Hawke's (1978) photo-
graph and schematic of the photo of a section through an egg
pocket support Everest's description (Figure B.4).
Hobbs (1937) stated that chinook (quinnat) salmon eggs
within egg pockets were at a depth of about 25 cm beneath the
surface of the redd. Chapman et al. (1986) reported that the
41 /SECTION II
-------
20.
Figure B.4. (From Hawke 1978). Typical egg pocket in a redd of
Oncorhynchus tshavytscha in section view.
shallowest Chinook salmon eggs lay 10 cm beneath the gravel
surface in the redd, but that 19 cm was the mean depth at which
the first eggs were encountered. In this study of spawning in
the main Columbia River, the mean depth of egg pockets was 29 cm
(range 19-37 cm).
Hobbs (1937) found that although redds of brown and rainbow
trout were smaller than those of Chinook salmon, the redd
structure was similar. Egg pocket number in brown trout redds
ranged from one to four, the number a function of redd size, and
most eggs lay 20 cm beneath the gravel surface. Hobbs stated
that rainbow trout eggs also lay in well-defined egg pockets at a
depth of about 20 cm beneath the gravel surface.
Redds may be constructed in isolation but are frequently
adjacent and often overlap* Chapman et al (1986) found that mean
redd size with tailspill was about 17 m2, and 13 m2 without
tailspill. Considerable overlap can occur before egg pockets are
disrupted (Chapman et al 1986), as is obvious from Hawke's
42
/SECTION II
-------
diagram of egg pocket placement in relation to overall redd
shape.
In areas heavily used by spawners, adjacent redds are often
constructed in a pattern that creates "dunes" of tailspills lying
normal to streamflow (Envirocon 1984). These dunes may persist
from year to year, especially in regulated streams, and fish that
spawn on the upstream faces inevitably "inarch" the peak of each
dune upstream over time. The dune configuration duplicates the
location of redds in runs and tails of pools just above riffles,
with a slight tilt downward in an upstream direction. This shape
facilitates water movement downward into the redd (Cooper 1965).
R. Thurow (personal communication) reported that steelhead
(spring spawners) spawned in some identical areas used in fall by
Chinook salmon in the South Fork Salmon River. Fish spawned in
clusters, and Thurow observed dune formation as a result.
Where spawners utilize the same areas year after year, they
may maintain the area in a coarser condition than surrounding
gravels that remain unused. Presence of large numbers of
spawners in the same area should lead to a "mass cleaning", as
fines removed from one redd may deposit downstream, then be
lifted and passed along by females working downstream. Large
annual spawning escapements have a major impact on maintenance of
high quality spawning habitat. When populations are reduced, the
overall quality of spawning habitat can decline because the
annual cleaning effect exerted by spawners is diminished (Everest
et al. 1986).
McNeil and Ahnell (1964) found that • pink salmon signifi-
cantly reduced the percentage of solids in the substrate that
passed through a sieve of 0.833 mm and of 0.104 mm, and a portion
of the removed materials consisted of light organic material.
Organics in the materials that passed the 0.104 mm sieve amounted
to an average of 3.9% of solids retained by a 0.074 mm sieve, and
43 /SECTION II
-------
12.4% of solids that passed a 0.074 mm sieve. Thus, the highest
organic fraction was in the smallest size fractions, and would
easily be removed by females during redd construction. This
should be typical of other salmon and trout redds as well.
Ringler (1970) demonstrated that new redds contained 32%
less organic material than old redds (from spawning in the
previous year) in Needle Branch, but in Flynn Creek new redds and
old redds contained approximately the same amount of organic
material. Needle Branch had been logged while Flynn Creek served
as a control.
Ringler (1970) also compared gravel composition in new and
old (previous year) redds (Figure B.5). His data demonstrate
considerable reduction of fines during spawning. Removal of
fines from the redd is also demonstrated by Figure B.6 (Everest
et al. 1986), which depicts the extent of reductions in fines of
1
its JUT *» it.
StEVf S1Z£ (mm)
Figure B.5. (From Ringler 1970). Mean size distribution of
gravels in new redds and former redds in Needle Branch, an Alsea
River tributary.
44
/SECTION II
-------
I I ••!•!• *p«»ftl*f
DlHI All*? •M««lii«
Pink* pink* Plflftt Plnki Pink*
I««r»t1 •< ll
(HI)
«II*K
CHS)
(••)
R*ikl *•!!• Metttll •
IMII <1»T»J AKfttll
Figure B.6. (From Everest et al. 1986). Percent fine sediment
before and after spawning for several species of anadromous
salmonids at various sites.
various sizes from the substrate by spawning females. Although
the criterion for fines differed among workers, the evidence for
substantial cleaning is clear.
B. 2 . Intrusion of fines into gravel
A photo in Everest et al. (1986) demonstrates that the egg
pocket is overlain by relatively clean gravels of lesser size.
Hobbs (1937) describes the formation of a "crust" of fines in the
surface layer of the redd as time passes. The rate at which
fines intrude into clean materials is of great interest.
Beschta and Jackson (1979) tested intrusions of fine sedi-
ments (sands with median particle size 0.5 mm) into an initially
clean gravel bed (median particle diameter 15 mm) . They found
that sands were trapped in voids within the upper 10 cm of the
bed, forming a barrier to further intrusions. Intrusion amounts
45
/SECTION II
-------
varied from 2 to 8% of total bed volume. Once the intrusion
"seal" developed, intrusion stopped and additional sands were
transported past the bed.
Froude numbers (Fr) help characterize flow conditions.
Beschta and Jackson (1979) describe this dimensionless variable,
which represents the ratio of inertial to gravitational forces in
fluid flow (from Streeter and Wylie 1975) :
Fr = V/gy, where
V = mean velocity, m/s,
g = acceleration due to gravity, 9.8 in/s2,
y = depth of flow, m.
For subcritical flow (Fr < 1.0), conditions consist of relatively
deep, slow flow. At a critical flow (Fr = 1.0), the specific
energy (E = V2/2g + y) is at a minimum. Standing waves in a
stream indicate critical flow conditions. Supercritical flow
(Fr > 1.0) is typical of relatively shallow, rapid streamflow.
At low Fr, the 0.5 mm sands quickly established a sand seal
in the upper 5 cm of the clean gravel bed as the larger sand
particles bridged the openings between adjacent gravel particles
and prevented downward movement of additional sands. At higher
Fr> flow characteristics began to alter the sealing process, and
most deposition and intrusion occurred within the 5-10 cm depth
zone in the bed. Higher velocities (associated with higher Fr)
led to greater bed shear and "jiggling" of surface gravels,
inhibiting formation of a sand seal near the gravel surface.
•Hence, the sand seal still formed, but deeper within the bed, and
where it would prevent deeper intrusion.
These observations on intrusion of fines parallel those made
regarding brown trout redds nearly 50 years earlier by Hobbs
(1937) , who stated that "Sediment tended to lodge and to cake
firstly amongst the surface material of the redds, forming a silt
"crust" overlying cleaner material. In some cases, but subse-
46 /SECTION II
-------
quently, silt penetrated to egg-pockets, virtually restoring the
bed to its original state." But he stated that it was unusual
for silt to penetrate to the egg pocket while ova or alevins
remained in the pocket.
Beschta and Jackson (1979) used 0.2 mm instead of 0.5 nun
sands in two tests. They found that the sand seal in the upper
level in the bed did not develop. Instead, the finer sands moved
down through the gravels by gravity and began to fill the test
gravels from the bottom up. Particle size appears to be an
important variable that influences depth of intrusion. The
amount of intrusion by 0.2 mm sands decreased as Fr increased
from 0.6 to 1.1. Beschta and Jackson showed that particle size
distribution of 0.5 mm sands 1 cm above the gravel bed was the
same as that at the point of sand introduction to the channel,
but that intruded sands tended to be smaller (Figure B.7).
Beschta and Jackson noted that flow conditions, sediment
transport rates, and sediment particle size all influenced the
amount of fines deposited in initially-clean gravels. They state
that the quantitative results of their study cannot be directly
applied to natural streams, but that their study had several
too
O.O6 O.I O3 1.0 2.0
PflflTICLC DIAMETER (mm)
Figure B.7. (From Beschta and Jackson 1979)
of sands during an intrusion test.
47
Size distributions
/SECTION II
-------
implications. Fine sediments added when the bed is stable will
deposit and intrude into initially clean gravels. If the
particles are large enough, a bed seal will form, and subsequent
deposition will be above the seal. If the fine particles are
small enough, they can fill the bed from the bottom up. As long
as the bed is stable, addition of fines can only result in
intrusion or a blanketing of gravel surface.
Phillips (1971) provided a diagram that showed a typical
salmonid redd. It illustrated that hydraulics within the redd
should tend to pull surface water through the redd. Vaux (1962)
and Cooper (1965) provided the experimental data that showed how
surface waters penetrate the substrate. Cooper showed
penetration to depths as great as 46 cm, much below the average
depth of the "floor" beneath the egg pocket.
Although gravel composition would affect the depth to which
surface water circulates, it is clear that the shape of the
salmon redd leads to greater surface water penetration than would
be the case in a gravel bed with relatively flat surface (Figures
B.8 to B.ll).
Cooper (1965) studied the effect of this "pulling" on
intrusion of fines in gravels in an experimental environment. He
confirmed that deposition of silt occurs within the gravel even
though surface water velocities are too high to permit deposition
on the gravel surface. The intrusion of fines reduced gravel
permeability.
The lowest retention occurred in a very coarse gravel, and
the greatest in a finer gravel such as that in typical spawning
beds. Figure B.12 illustrates this point.
48 /SECTION II
-------
r
MM fl
\
!.•»»• '",*?** \
"—•-—\ ^—~-g^,- -^^ '• .rf.,, jyr .-.j...,: -...-a. A
I . I I . I . I i I . I . I , I . I . I . I . I . I . t . I
M W T«
H •• cfunvt tm
Figure B.8. (From Cooper 1965). Flow through homogeneous gravel
with level gravel surface.
^***» ••>!•
• :,--~_, ]Ir _ h » ... . j ••ujr •
• . I . J I . ^~~?~~^^.t ^ 'i ' F 1 ' I 1 I E r i \ i.\ Ij Elir<"~Tn. >- ier 'j?' ifl" 1- , - I i ~
•• M t»
Figure B.9. (From Cooper 1965). Flow through homogeneous gravel
with level surface and stones on surface.
» <•>•»<• M w r«
H* 114 iro
Figure B.10. (From Cooper 1965). Flow through homogeneous
gravel with surface similar to a new salmon redd.
49
/SECTION II
-------
Figure B.ll. (From Cooper 1965). Flow through homogeneous gravel
with surface similar to a new salmon redd.
Figure B.12. (From Cooper 1965). Rates of sediment removal from
surface waters for different surface conditions, permeability,
and silt sizes.
Meehan and Swanston (1977) found that the rate of retention
of fines <2 mm that were introduced into flow of an artificial
stream channel was greater in test baskets a given distance
downstream from the source of sediment for rounded gravels than
50
/SECTION II
-------
for angular gravels at very low flows, and that the results
reversed at higher flows. The authors attributed these results
to presence of more low-velocity areas in rounded gravels at low
flows, which permitted fines to intrude and settle. In angular
gravels, more tractive force was needed to carry sediments
downstream, but more zones of low velocity were present in
angular particles at high flows, permitting more accumulation of
sediment.
Koski (1975) showed that the percentage of silt (<0.105 mm)
retained in gravels in experimental stream channels is related
inversely to the amount of sand (particles >0.105 and <3.327 mm).
The relationship appears to reflect "space available", in that
voids filled with sand cannot fill with silt.
Although it is perilous to project the findings of Cooper
(1965) and Jackson and Beschta (1979) to the salmonid redd in
detail, some conjecture seems reasonable. From the moment when
fertilized eggs fall into the cobble or large gravel centrum of
the egg pocket, digging by the female spawner results in a
bridging of smaller gravels among egg pocket centrum components,
and then a mix of gravels over the centrum. Finally a seal of
of finer sediments will develop somewhere in the redd, with the
depth and composition of the seal dependent on sediment transport
patterns in the surface flow. The seal may develop partly during
the completion of the redd.
Hawke (1978) described the gravels above the egg pocket
centrum as mainly fine, with a high proportion of coarse sand.
In some pockets a loose core of pebbles ran from the egg pocket
to the surface (an example of bridging, perhaps). The smaller
the spawner, the smaller will be the average particle size in the
redd and the smaller should be the average diameter of the "seal"
components. Of course, egg and alevin size tend to be directly
related to fish size as well. The implications of this in rela-
51 /SECTION II
-------
tion to the "seal" will be covered in another report section.
The complex interaction among Fr, freshet events, and sedi-
ments in transport during the incubation period for salmonids
will strongly influence conditions in the redd during incubation
and emergence.
B.3. Porosity, permeability, and water movement
Porosity of gravels is defined as the ratio of the volume of
the voids to the total volume of solids plus voids. It can be
measured by dividing volume of water in a gravel bed by volume of
water plus gravel.
Permeability is a measure of the ability of gravel to pass
water per unit of time and is reported as a distance per unit of
time. It is a function of hydraulic gradient and apparent velo-
city and temperature. Gravel porosity is embodied in the appar-
ent velocity variable.
Water movement in the substrate is measured by apparent
velocity, the volume of water passing through a unit area of the
gravel bed per time unit. It is a function of gravel permea-
bility, hydraulic head, and temperature. The modifier "apparent"
is used because true velocity at any micro-point in the sub-
strate, say at a point on the surface of an incubating embryo,
varies greatly from point to point. Pollard (1955) describes the
relationships among these variables. An excellent body of labor-
atory research and theory is formed by Pollard and by Cooper
(1965). Terhune (1958) and Wickett (1954, 1958) complemented
this with work on field measurement of in situ intragravel
conditions.
Cooper (1965) showed that intrusion of fines in the
substrate reduced porosity and permeability. He also demon-
strated that gravels with lower permeabilities trapped more
52 /SECTION II
-------
sediments than those with higher permeability (Figure B.12).
Sediment intrusion also reduces apparent velocity.
Theoretically, blanket deposition of fines on the surface of the
gravel might reduce water movement, as reflected in apparent
velocity, without reducing permeability of the gravel if the
surface seal of fines were precise and complete. However,
intrusion would probably occur in cleaner gravels without a seal,
such as those in the redd, until a seal develops.
53 /SECTION II
-------
C. INTRAGRAVEL ECOLOGY OF SALMONID EMBRYOS
C.I. Apparent velocity
Cooper (1965) showed that apparent velocity strongly
influenced survival of eyed sockeye salmon eggs to emergence
(Figure C.I). He also demonstrated that survival declined with
increased fractions of particle sizes smaller than 3.36 mm,
possibly caused by packing of particles around embryos or by
Figure C.I. (From Cooper 1965). Survival of sockeye salmon
embryos as a function of apparent velocity in the gravel.
transfer of soil pressure, and possibly with higher uniformity of
gravel size (Figure C.2) except perhaps in very coarse gravels.
Shumway et al. (1964) established that water velocity past
embryos and dissolved oxygen concentration directly affected
survival of embryos. But these workers showed that the oxygen
requirements of the embryo can be met by very low water velo-
cities when oxygen levels are sufficient. The influence of
54 /SECTION II
-------
apparent velocity is slight in comparison to influence of oxygen
levels. i
Figure C.2. (From Cooper 1965). The effect of gravel size and
uniformity on the survival of sockeye eggs at a flow of 0.0167
cm/s.
Coble (1961) demonstrated that survival of steelhead embryos
was directly related to apparent velocity of intragravel water.
However, when he adjusted his data to normalize dissolved oxygen
level at 6 mg/1, he found that survival was no longer an obvious
function of apparent velocity. Coble noted that it is oxygen
that is essential to incubating embryos, not velocity, and the
function of water movement is mainly to deliver oxygen to the
embryo and to carry'away waste. He prepared a graph (Figure C.3)
to illustrate the relationship between dissolved oxygen and
apparent velocity in artificially-dug redds that contained steel-
head embryos. This showed that when velocities were low, oxygen
concentrations were low; when velocities were high, oxygen
concentrations were higher as well. Coble reported that survival
of embryos is related to apparent velocity, but indirectly
through dissolved oxygen concentrations.
55 /SECTION II
-------
10
SB
I 1 I L I 1 I 1 1 1
O KJ 10 30 10 50 80 TO BO 9O KM 110
MEAN tPPJineiT VttOCITY
IN CCNtlMCTCRS PER HOUR
Figure C.3. (From Coble 1961). Relationship of dissolved oxygen
to apparent velocity in artificially-dug redds.
C. 2. Permeability
Wickett (1958) related percent survival (to emergence) of
pink and chum salmon fry to permeability. McNeil and Ahnell
(1964) showed that permeability was inversely related to the
percentage of substrate particles that passed through a 0.333 mm
sieve (Figure C.4) , and that the more productive pink salmon
spawning streams that they examined had high permeability
coefficients. Wells and McNeil (1970) found that the largest
embryos of pink salmon in Sashin Creek, Alaska, came from a
stream segment with a relatively steep grade and coarse materials
in the bed.
We used water volumes measured per unit of time by McCuddin
(1977) to calculate permeabilities of gravels that he used for
examination of embryo survival to emergence. These data (Figure
C.5) indicate that survival of Chinook salmon and steelhead trout
is positively related to permeability.
56
/SECTION II
-------
3OO
ZOO
<3 100
4"
\
\
\
.\*
Cor»e inn filled
by tyt.
\ .
•v .
\«
\
• v
10
13
ZO
Percent of sample passing a 0.833 mm sieve
Figure C.4. (From McNeil and Ahnell 1964). Relationship between
permeability and percentage of fines smaller than 0.833 mm.
• =
• * cfnfiooK
orty
F ' 23.871
1 + 3 dt
100
1000
PERMEABUTY Icm/hourl
10OOOO
Figure C.5. (Adapted from McCuddin 1977). Survival as a
function of permeability of gravel mixes in cells in laboratory
studies of survival of chinook salmon and steelhead embryos from
green egg to emergence.
C.3. Dissolved oxygen
Alderdice et al. (1958) tested survival of chum salmon eggs
exposed to various constant levels of dissolved oxygen for 7 days
at various development levels. They showed that exposure to low
57 /SECTION II
-------
dissolved oxygen caused premature hatching (Figure C.6), and that
rate of oxygen uptake increased steadily from fertilization to
hatching. They also calculated and plotted critical dissolved
oxygen levels (the oxygen needed for successful incubation) and
median lethal levels of dissolved oxygen (Figure C.6) at 10 C.
Figure C.5. (From Alderdice et al. 1958). Variation in hatching
rate of chum salmon eggs reared at IOC, results of 7-day exposure
to prescribed oxygen levels at intervals through incubation.
Critical level depended on stage of embryonic development in
degree days.
For the purposes of our review, and in view of the great
importance of dissolved oxygen in incubation of salmonid embryos,
we provide some detail on the calculation of critical oxygen
levels by Alderdice et al. (1958):
Oxygen respired by the embryo diffuses through a thin
enclosing spherical capsule of species-specific diameter and
thickness. If a homogenous spherical body uses oxygen at a
constant rate, and if the oxygen tension may be assumed to be
maintained at zero in the center of the body, then
C0 =
Ar2/6D, where
concentration of oxygen at surface of the sphere in
atmospheres,
58 /SECTION II
-------
A „,
..-' TSl'
,-r.v/--"!"-'--1
Figure C.6. (From Alderdice et al. 1958). Lethal and calculated
critical levels of dissolved oxygen for chum salmon ova incubated
at 10 C. •
A = oxygen consumption of the sphere in ml/g/min.
r = total radius of sphere in cm.
D =» diffusion coefficient of oxygen through the capsule
in ml/cm per cm2 of capsule area/min.
This formula may be applied to the egg before a functional
circulatory system develops to estimate ambient oxygen required
to maintain respiration at a rate independent of the environ-
mental supply.
When an egg reaches the stage of possessing a functional
circulatory system (about 200 degree-days, calculated in
centigrade), oxygen is transported to the embryo tissue with
greater efficiency. The tension difference needed for dif-
fusion of oxygen in this phase is:
C0 = ArT/3D, where
T = thickness of the capsule in cm
A = oxygen consumption of the sphere, in ml/g tissue/min.
Alderdice et al. (1958) assumed D = 0.0000123 ml/cm/cm2/min,
T = 0.006 cm for chum salmon, weight of the chum egg as 0.29 g,
egg radius as 0.37 cm, and calculated critical oxygen level for
various stages of development as in Figure C.6. Critical oxygen
level is that concentration at which respiratory demand is just
satisfied. The key finding is that oxygen need rises with
59
/SECTION II
-------
development, and by the stage of development at 250 degree-days
has reached 5 ppm at IOC. As noted later, the 5 ppm figure is too
low.
Alderdice et al. (1958) recommended that critical levels of
dissolved oxygen be regarded as a measure of oxygen requirements
for successful incubation, and that studies be undertaken to
determine if theoretically-estimated critical levels of oxygen
are similar to empirical limiting levels.
Silver et al. (1963) showed that growth of Chinook salmon
embryos was restricted before the 24th day at all tested oxygen
concentrations below 11.7 mg/1. For steelhead, growth is res-
tricted before the 30th day at all tested oxygen concentrations
below 11.2 mg/1. Growth of coho salmon at 11 C is restricted
before the 7th day after fertilization at concentrations of
dissolved oxygen at least as high as 6 mg/1, and before the 23th
day at concentrations slightly below 11.9 mg/1. Other data show
restricted growth of embryonic steelhead at 12.5 C before the
llth day at oxygen levels slightly below 10.4 mg/1.
Silver et al. (1963) concluded that the theoretical critical
oxygen levels calculated for embryos by Alderdice et al (1958)
are far below actual limiting oxygen levels for salmonid embryos
throughout most of development at temperatures of 10 to 12.5 C.
These authors ascribe the difference to the impropriety of the
models (summarized above) used by Alderdice et al. The error may
be associated with assumption that the limiting respiratory
surface before establishment of blood circulation is the entire
periphery of the chorion, when in likelihood it is the surface of
the embryo itself. Another possible error is associated with the
post-hatch period. Alderdice et al. assumed that the perivit-
elline fluid after establishment of blood circulation has zero
oxygen. The embryo could not survive with no oxygen in the
fluid. This means that realistic tensions are less than assumed
60 /SECTION II
-------
by Alderdice et al. Higher oxygen concentrations in the water
would be needed to achieve the same embryo development. Silver
et al. (1963) suggested that a realistically determined critical
oxygen concentration in the surrounding water cannot be assumed
to be very much lower than the critical concentration in the
water for newly hatched sac fry.
Davis (1975) extensively reviewed the oxygen requirements of
salmonids, including anadromous forms. He shows a mean threshold
of incipient oxygen response for hatching eggs and larval
salmonids as 8.09 mg/1 (SD = 1.65, SE = 0.58, n = 3) and 76%
saturation (SD = 22.9, SE = 13.2, n = 3).
Development of embryos often proceeds at temperatures less
than 10 C. In the interior plateau and Rocky Mountains, winter
water temperatures during incubation by fall spawners often reach
5 C (Columbia River mainstem), or even less than 1 C (tributaries
at high elevation). Critical levels for dissolved oxygen at low
temperatures are not established, but may be somewhat less than
those reported by Silver et al. (1963). Wickett (1954) tabulates
data from Atlantic salmon (Lindroth 1942) obtained at about 5 C.
These indicate a critical dissolved oxygen level of 5.8 ppm just
before hatching.
Wickett's own data for chum salmon suggest a critical level
of less than 5 ppm for 85-day eggs (faintly eyed) at 3.6-4.9 C.
The limited data indicate that an assumption that critical oxygen
levels are lower at low temperature is reasonable. Hobbs (1937)
noted that the oxygen requirement per unit of tissue and unit of
time at 3 C was about one-third, and at 7 C about one-half what
it is at 12 C.
The fairly convincing results, summarized below, of Silver
et al. (1963) and Shumway et al. (1964), who showed that any
reduction of dissolved oxygen could be shown to reduce length of
61 /SECTION II
-------
fry at hatching, suggest that reduction of dissolved oxygen in
natural environments below natural levels should be assiduously
avoided. Silver et al. (1963) demonstrated that water velocities
must be high enough within the redd to transport enough oxygen to
support all embryos in the redd, but also to deliver sufficient
oxygen to the surface of the chorion around each embryo.
Steelhead and Chinook salmon embryos held at 9.5 and 11 c,
respectively, all died at an oxygen concentration of 1.6 mg/1.
Alderdice et al. (1958) found that the incipient median lethal
level for dissolved oxygen rose with development from about 0.4
ppm early in embryo development to 1.0-1.4 ppm before hatching.
Silver et al. (1963), in an important study, showed that sac
fry from embryos reared at low and intermediate oxygen concen-
trations were smaller and weaker than sac fry from embryos reared
at high concentrations. Figure C.7 clearly shows that length of
steelhead fry is a function of water velocity at given oxygen
concentrations, and that at higher oxygen concentrations, water
velocity increase has less effect on size of fry than at low
oxygen levels.
i*
0
5 ir
I '*
.„
man VELOCITY
O MOCM/HH
A ISO CM/Hit
Q 14 CM/Hft
V * CM/Hit
MMI.VKB errtt* ewtecHiiurKiN m M.LMMMI»II uttit
Figure C.7. (From Silver et al 1963). Mean lengths of steelhead
sac fry when hatched in relation to dissolved oxygen during
incubation, at different water velocities and 9.5 C.
62
/SECTION II
-------
Figure C.8, compared with Figure C.7, allows one to see that
dissolved oxygen is relatively more important as an influence on
steelhead fry size than water velocity. The velocities measured
by Silver et al. approach true pore velocity (the actual velocity
past the surface of the embryo). Apparent velocities as measured
in field studies are lower than actual velocities in the sub-
strate, and cannot be compared absolutely with these laboratory
studies. However, the results of Silver et al. (1963) demon-
strate clearly the combined effects of dissolved oxygen and velo-
cities, regardless of absolute values for velocity.
s
*»
>•
II.I ««/ L
A r.t M«/ L
a s.r M«/ L
* « t «•/ L
0 I.* MO/ L
W SO KW 300 NIOO
VtLOWTT « eCNTIMITCM PCM HOW
Figure C.8. (From Silver et al. 1963). Relationship between
length of steelhead sac fry when hatched and water velocities, at
different dissolved oxygen concentrations and 9.5 C.
Shumway et al. (1964) demonstrated that when embryos were
incubated in a laboratory environment among glass beads, rather
than on a porous plate such as that used by Silver et al. (1963),
a given water volume per unit of time in experimental cells led
to larger embryos at a given oxygen concentration, an effect
especially pronounced when a mix of large and small beads was
used. The clear inference is that true pore velocities were
higher in the substrate mix, and that it is correct to conclude
63
/SECTION II
-------
that velocities in pores in redds are higher than is reflected by
apparent velocity determinations. Figures C.9 and C.10 for
Chinook salmon fry (Silver et al. 1963) duplicate the pattern of
fry size and dissolved oxygen and water velocity found for
steelhead.
As Mason (1969) later showed, any reduction in size at
emergence has important effects on subsequent social success,
hence survival, of post-emergent fry.
Figure C.9. (From Silver et al. 1963) . Mean length of chinook
salmon sac fry at hatching as a function of oxygen level at 11 C.
Effects on emergence itself will be discussed later in this
review.
t*
•'•
I"
s
I"
_ a lit W/L
O If M*/L
* I H«/l.
IM/L
HI
«»'
meet" m ctmintim n« MOUK
Figure C.10. (From Silver et al. 1963). Mean length of chinook
salmon sac fry as a function of water velocity at 11 C.
64
/SECTION II
-------
Shumway et al. (1964) provided a three-dimensional model of
the effects of dissolved oxygen and water velocity on size of
fish at hatching (Figure C.ll). They also modeled the effect of
the two variables on hatching delay.
—t-T T-rrrw.'
prWtB M««l «)W^"tM>K««
Figure C.ll. (From Shumway et al. 1964). Model of influence of
both dissolved oxygen and water velocity on size of fry at
hatching.
McNeil (1969) discussed compensatory growth in alevins by
noting that Brannon's (1965) studies of newly-hatched embryos
incubated at 3, 6, and 12 mg O2/l reported average wet weight of
newly hatched alevins as 22, 30, and 42 mg for the respective
dissolved oxygen levels. At absorption of yolk, fry from eggs
exposed to 3 mg O2/l were over 5 times their weight at hatch, fry
from the 6 mg O2/l history were 4 times weight at hatch, and fry
from eggs provided with 12 mg 02/1 were 3 times their weight at
hatching.
Coble (1961) demonstrated that embryo survival in steelhead
was related to dissolved oxygen concentration in artificial redds
constructed in a field environment, but found that apparent velo-
city and dissolved oxygen effects could not be separated.
65
/SECTION II
-------
Phillips and Campbell (1962) buried newly-fertilized steel-
head and coho salmon ova in stainless steel perforated boxes in a
glass bead medium; the boxes surrounded short standpipes that
were buried in shovel-dug redds in tributaries of Drift Creek, an
Alsea River drainage. Percent survival of steelhead to a time
about 3 weeks after hatching was negligible where mean dissolved
oxygen levels were below 7 mg/1 (Figure C.12).
For coho, high survivals occurred above 8 mg/1 (Figure C.13)
and alevin size correlated with amount of dissolved oxygen. For
steelhead, no relationship between alevin size and oxygen level
was apparent. These authors concluded that oxygen requirements
of incubating embryos were higher than had previously been
suspected.
i i . i .-i . i . t . i . i . i . i
Figure C.12. (From Phillips and Campbell 1962). Survival of
steelhead embryos in relation to dissolved oxygen levels.
Wells and McNeil (1970) noted that the largest and fastest
developing embryos and alevins of pink salmon were found in
Sashin Creek, Alaska, in spawning gravels with high levels of
dissolved oxygen in intragravel water.
Koski (1975) showed that survivals of chum salmon to emer-
gence were about one-third as high when embryos had been sub-
jected to dissolved oxygen levels of less than 3 mg/1 as compared
66
/SECTION II
-------
I—
4 I » r I
wut nnotvn mra* m«t»t««no" t-»Tt
Figure C.13. (From Phillips and Campbell 1962). Survival of coho
salmon embryos in relation to dissolved oxygen level.
to levels over 3 mg/1. Emergence was delayed in groups exposed
to oxygen levels lower than 3 mg/1.
McNeil (1966) stated that oxygen requirements of embryos
rises to a maximum just before hatching, and that larvae are more
tolerant of low dissolved oxygen levels than are embryos. Hays
et al. (1951) reported that the oxygen concentration that would
limit metabolism of Atlantic salmon decreased after hatching.
McNeil attributed this change to availability of vastly increased
respiratory areas (gills) after hatching.
Sowden and Power (1985) found that survival of rainbow trout
embryos in a groundwater-fed streambed depended upon mean dissol-
ved oxygen content and velocity of groundwater in redds. Figures
C.14 and C.15 demonstrate these relationships. Survival increased
directly as oxygen content rose above 6 mg/1.
Mason (1969) exposed coho salmon embryos and alevins to
dissolved oxygen levels of 11, 5, and 3 mg/1. Mortalities to
time of yolk absorption amounted to 17, 23, and about 41% for
these respective dissolved oxygen levels. Embryos subjected to
67 /SECTION II
-------
90
40
<
10
10
0 * * « 8 10
MEAN DISSOLVED OXYGEN CONC ma/t
90,
0 «O 90 120 ISO 200
CALCULATED VELOCITY cm / h
Figure C.14. (From Sowden and Power 1985). Relationship between
percent survival of preemergent rainbow trout embryos and mean
dissolved oxygen content (top graph) and apparent velocity
(bottom graph) of groundwater in redds.
the lower oxygen levels were significantly smaller (p < 0.01) at
hatching (22.9, 25.4, and 28. 1 mm for the 3 , 5, and 11 mg/1
respective oxygen concentrations) . Alevins that were subjected
to 3 mg/1 were about 33 mm long at yolk absorption, while alevins
subjected to the higher oxygen levels were over 37 mm long.
Fry that had been exposed to the most severe hypoxial
conditions were most prone to emigrate after emergence because of
competition in the post-emergence period. Mason compensated for
low dissolved oxygen levels by increasing temperatures so that
fish would emerge from all 3 groups at the same time. Had he
maintained all groups at the same temperature, the social
disadvantage faced by fry with hypoxial histories would have been
68
/SECTION II
-------
exacerbated because they would emerge later and smaller.
The key point in regard to Mason's study and the work of
Silver et al. (1963) and Shumway et al. (1964) is that depri-
vation of dissolved oxygen leads to subtle problems not detect-
able in tests of survival in various oxygen levels. It appears
incorrect to set critical oxygen levels at any arbitrary point,
or to assume that survival to time of emergence is sufficient
evidence of ecological success. In fact, any reduction in
dissolved oxygen levels from saturation appears to reduce
likelihood of survival to emergence or post-emergent survival for
embryos.
Davis (1975) suggested three levels of protection against
effects of low dissolved oxygen concentrations:
Level A: This level is 1 SD above the mean average incipient
oxygen response level (incipient response is defined as a
dissolved oxygen level at which sublethal response to hypoxia
first becomes apparent) for the group. The rationale is that few
members of a fish population, or fish community, will likely
exhibit effects of low oxygen at or above this level. Level A
represents more or less ideal conditions and permits little
depression of oxygen from full saturation. It represents a level
that assures a high degree of safety for very important fish
stocks in prime areas.
Level B: This level represents the oxygen value where the
average member of a species of a fish community starts to exhibit
symptoms of oxygen distress. These values are derived from the
class mean averages of incipient response. Some degree of risk
to a portion of fish populations exists at this level if the
oxygen minimum period is prolonged beyond a few hours.
Level C: At this level a large portion of a given fish
population or fish community may be affected by low oxygen. This
deleterious effect may be severe, especially if the oxygen
minimum is prolonged beyond a very few hours. The values are 1
SD below the B Level, or class average, for the group. This
level should be applied only if the fish populations in an area
are judged hardy or of marginal significance, or of marginal
economic importance and, as such, are dispensable (i.e. in the
socioeconomic sense) .
69 /SECTION II
-------
Davis notes that the use of standard deviations is based on
the statistical concept that in normally-distributed data, about
68% of the values lie within plus or minus 1 SD of the mean.
Thus, the recommended levels span the range of responses that
include both sensitive and insensitive individuals, both within
and between species. Davis prepared a table of oxygen criteria
for various fish communities, using the foregoing protection
levels at various temperatures (Table C.I).
C.4. Fines
One difficulty in relating percentages of fines to survival
of embryos and alevins is absence of a common sieve standard for
definition of "fines". Various workers have used sieve criteria
of 0.83, 0.85, 1.0, 2.0, 3.0, 6.0 and 9.5 mm as limits of
"fines" categories. Section A contains a lengthy description of
efforts to unify gravel characterizations. In general, the
information suggests that the greater is the proportion of fine
sediments in redds, the lower will be survival.
Table C.I. (From Davis 1975).
Oxy*m aittrif tun! OH perwHite nlMMlon vilun totted wilt, it™ teveb of emtoctton u outlined
to Ih* (at P0,'j «
-------
Tagart (1984) measured survival from egg deposition to fry
emergence in 19 redds of naturally-spawning coho salmon in the
Clearwater River in northwestern Washington. The range in
survivals extended from 0.9 to 77.3%. Survival was directly
related to intragravel permeability, and percentage of "good
gravel" (defined as the fractional volume between 3.35 and 26.9
mm) (Figure C.15). Survival was inversely related to the
percentage of "poor gravel" (particle volume under 0.85 mm) . For
9 redds for which minimum and mean dissolved oxygens were
available, Tagart (1984) showed that dissolved oxygen was
inversely related to the percentage of fines under 0.85 mm in
size. The reason for the relationship is not clear, but
biochemical oxygen demand in the substrate may have reduced
oxygen levels where permeability was low (high percentage of
fines), or low permeability prevented interchange of oxygenated
surface waters with intragravel water. It is important to
remember in considering this point that fines were assessed by
Tagart in redds rather than in egg pockets.
Tagart's relationship between survival and permeability in
Figure C.15 is not a strong one because it depends so strongly on
an extreme point for regression development. In a later section
on predictors of survival, we provide a more robust development
of the relationship between permeability and survival.
Tagart's data show survivals to emergence from redds with
less than 20% fines as 31.9%, while that for groups with greater
than 20% fines was 17.7%, a difference significant at p = 0.05.
Trapping results have particular value because no management of
gravel mix or stratification is involved when natural redds are
sampled. However, as we note elsewhere, Tagart did not obtain
his independent variables with reference to the egg pocket.
71 /SECTION II
-------
Ill V8 MERN PERMEP81UTY TO 50% EMERGENCE
Survival to
emergence
ten am ma mm
PERHEflBlLlTY IN CM PER HOUR
Sit V8 FINES
• • *
i • -.m
• n » » <
rtRCEHT FItCS tORRVEL < 0.850 mi
Figure C.15. (From Tagart 1984). Survival of coho salmon embryos
to emergence in natural redds as a function of gravel permeabil-
ity (top graph) and percent fines
-------
Koski (1966) trapped fry emerging from 21 natural redds of
coho salmon in three Oregon streams. Survival was generally
related to a permeability index and loosely to minimum dissolved
oxygen concentration. It was also inversely related to percent-
age of fines, the percentage of fines smaller than 3.3 mm having
the highest correlation (Figure C.16.).
• D**r Creak
» N«*dt* Branch
• Klfnn Creek
IS 10 32 34 3« 38 40 41 44 46 4* SO
PERCENT GRAVEL SMALLER THAN
J.3Z7 MILLIMETERS
Figure C.16. (From Koski 1966). Survival to emergence in
natural coho redds in relation to percentage of fines < 3.3 mm.
He found a similar relationship for chum salmon survival to
emergence in gravels placed in a spawning channel in Washington
(Figure C.17). However, Koski's data on chum salmon were taken
from the channel cell in which fish spawned, not in the egg
pocket. They should not be used for quantitative prediction of
effect of fines on survival in the wild.
73
/SECTION II
-------
H»
•O
W
TO
»
10
10
- ll7.1f-l.7Wx
(0 30
90
Figure C.17. (From Koski 1981). Percent survival of chum salmon
embryos to emergence in relation to fines smaller than 3.3 mm.
Coho salmon survival from green egg to emergence was tested
in artificial stream troughs by Cederholm and Salo (1979). The
inverse relationship between percentage of fines smaller than
0.85 mm and survival (Figure C.18) is fairly strong. The mix and
stratification of gravels in these experiments was chosen to
provide an analog of actual conditions in various streams of the
Clearwater basin, but "actual conditions" refers to spawning
gravels; areas used by fish, and not to the egg pocket. Thus,
the data can provide no quantitative predictor of survival
except for the specific laboratory conditions that were studied.
Some readers may consider the foregoing concerns irrelevant,
for they may not readily see how a gravel mixture in a trough, if
based on data on gravel composition from spawning areas, would
differ from a gravel mixture in a natural redd. The general
maxim that "nothing is as simple as it seems" comes into play,
74
/SECTION II
-------
KX3
U 75
5
i
Rao
<
1975
r i-
75
f m M.11-4.M
5O
25
1977
j - lOT.f«-).!>!
O 5 10 15 20 25 30 O 5 10 15 2O 25 3O
MEAN »f. FINES < QBSOmm dla
Figure C.18. (From Cederholm and Salo 1979). Coho salmon
survival from green egg to emergence in gravel troughs in
relation to percent fines smaller than 0.85 mm.
for the way in which the gravel matrix lies in the redd affects
emergence success. This problem is addressed earlier in the
review with regard to the work of Tappel and Bjornn (1983) , and
later in reference to Irving and Bjornn (1984), and in numerous
references to failure of tools to accurately sample the makeup of
the egg pocket.
Peterson and Metcalfe (1981) measured emergence of Atlantic
salmon eggs that had incubated in various gravel and sand mix-
tures and two directions of water current. Fine sand (0.06-0.5
mm) was more effective than coarse sand (0.5-2.2 mm) in reducing
emergence success. Strong upwelling water current in the gravel
bed mitigated effects of sand (reduced porosity, hence permea-
bility) to some degree, and induced earlier fry emergence. These
workers showed that where the percentage of fine sand rose above
about 12%, survival declined sharply. Where the percent of
coarse sand rose above about 22%, emergence dropped sharply. The
75 /SECTION II
-------
gravel mix used by Peterson and Metcalfe had a high ratio of
small:medium gravels (particles 22-62 mm had a ratio to medium
and fine gravels of 2.3-22 mm of about 5:3), a mix similar to
that found in natural spawning areas for Atlantic salmon, but not
normal for western streams used by Pacific salmon and steelhead.
MacCrimmon and Gots (1986) investigated the effects of fines
<4 mm on survival of rainbow trout from eyed egg to emergence.
Survivals were 51-74% in gravels with 40-100% fines, although
fines led to earlier emergence of smaller alevins. It appears
that alevins responded to a high percentage of fines by exiting
the substrate early, independent of dissolved oxygen levels.
Survivals equaled 87-92% in 0-20% fines. This work was completed
in incubation cells that had vertical water movement adjusted to
130 ml/min regardless of substrate composition. This means that
actual pore velocities were increased in fines.
McCuddin (1977) tested ability of Chinook salmon and steel-
head to survive and emerge in troughs of various gravel-sand
mixtures that were designed to simulate natural spawning areas.
Survival decreased as the proportion of sand in the substrate
increased (Figure C.19) above 10-20%. For tests with newly-
fertilized eggs placed in the substrate, any percentage of 6-12
mm particles above about 10-15% appeared to reduce survival. Any
percentage of fines (< 6 mm) above about 20-25% reduced
survivals.
Fines appeared in McCuddin' s work to interfere more with
emergence than with incubation, as dissolved oxygen levels in the
channel remained above 9 ppm and no relationship was found
between fines and length, weight, or time of emergence of fry.
Typicality of the gravel beds in artificial channels is always an
issue, and in McCuddin's work the egg pocket structure did not
appear typical of a natural redd. The effect of this on his
results is uncertain.
76 /SECTION II
-------
(OOf
s "
5
I"
CHtNOOK
SALMON
1975
PERCENTAGE SEDIMENT LESS THAN ».* MM
00
75
50
15
n
k
\
^
WO
STEELHEAD
TROUT
(976 75
50
«
N . .
MM
\
\
CHINOOK
SALMON
1976
\
,
STEELHEAO
TROUT
I9J*
0 W 40 . 60 0 » *0 60
PERCENTAGE SEWHENT LESS THAN tZ'HM
100
75
50
K
CHINOOK
SALMON
1976
M
tO *0 60
Figure C.19. (From McCuddin 1977). Percent emergence of Chinook
salmon and steelhead in relation to percentage of fines < 6.4 mm
(top 3 graphs) and < 12 mm (bottom 3 graphs).
The percentage of survival of bull trout from fertilized
eggs to emergence was measured in fiberglass screen bags in
artificial redds (Shepard et al. 1984). Open bags were used in
part of the work and emergence traps placed over the redds.
Survival to emergence was negatively correlated with percent
fines (< 6.4 mm) (Figure C.20).
NCASI (1984b) studied the survival of rainbow trout embryos
to emergence (Figure C.21). This work showed that survival was
inversely related to percentage of fines smaller than 0.8 mm.
For each percent increase in fines over the range of 10-30%,
77
/SECTION II
-------
*
I
^
•
z
f-M»l* III. I
1*
MKIMf
Figure C.20. (From Shepard et al. 1984). Survival of bull trout
to emergence in relation to percent fines < 6.4 mm.
survival declined 1.3%. In a second study, each percent increase
over the range 10-40% decreased survival 1.1%. The work also
found a significant negative relationship between survival and
percentage of fines smaller than 6.4 mm. The authors stated that
failure to emerge was probably associated with physical
entrapment, as the dissolved oxygen content of the intragravel
water at any gravel mix was similar. No information on apparent
velocities was obtained.
The high survival (near 90%) at 20% fines (< 6.4 mm) is of
interest. The authors felt that these particles prevented smaller
fines and organic debris from entering the incubation environ-
ments. This bridging effect would also occur in the egg pocket.
We infer, from this and other information, that some fines aid
survival, and that the particular mix and stratification in the
egg pocket governs emergence success.
Where the substrate is supplied with groundwater instead of
surface waters, the relationship between fines and survival would
be an unsatisfactory predictor of survival. Sowden and Power
(1985) reported that survival of rainbow trout was not signifi-
cantly related to the percentage of sediments smaller than 2.0
mm, to dg, or to the fredle index of substrate quality. Rather,
it was strongly related to dissolved oxygen level and water
78 /SECTION II
-------
i
i
•a
8
M
o
DO
B
u
a.
100
90
80
70
60
50
40
30
20
10
Fin* SediMDt* <0.8 ma d
(1) r2 - 0.70
(2) r2 • 0.5S
'N>
x »
S
"^1*
I ^ • i/ 1
X^x]
10 20 30
FIHBS CONCBNTRATIOW
40
100
90
80
70
60
SO
40
30
20
10
Fine Sediments < 6.4 mn d
_2
0.36
—*-=J!.7 . 0
10 20 30
FIHES CONCENTRATION
40
Figure C.21. (From NCASI 1984b). Survival of rainbow trout
embryos and alevins to emergence in various mixes of gravel in
relation to fines of < 0.8 mm (left graph) and < 6.4 mm (right
graph.
velocity, with oxygen content determined by groundwater condi-
tions rather than by factors causing biological oxygen demand
within the redd. Sowden and Power measured survival to the sac-
fry stage, not to emergence. They note that further studies that
take survival to emergence would be desirable, but urge cautious
application of survival models based on substrate particle sizes.
Their omission of the period from sac-fry to emergence makes it
impossible to draw conclusions about effects of fines on
survival.
The extensive studies by Tappel and Bjornn (1983) as
described earlier are pertinent in review of effects of fines,
but will not be re-summarized here. Irving and Bjornn (1984)
extended the laboratory techniques that Tappel and Bjornn used
for study of Chinook salmon and steelhead survival in various
gravel mixes to investigation of survival of kokanee salmon and
cutthroat and rainbow trout. Figure C.22 depicts their data on
survival in relation to the percentage of fines smaller than 6.35
mm, together with those of Tappel and Bjornn (1983). These
workers also prepared isoline graphs for 0-80% survival in
79
/SECTION II
-------
too
Cutthroat Trout
o 10 • ao so 40 go so
Kokin«« Salmon
IO 2O 3O 4O SO SO
0 19 2O 30 40 80 80
0 10 20 30 40 SO 90
tO IO 30 40 SO HO
PERCENTAGE FfMES
Figure C.22. (From Irving and Bjornn 1984). Embryo survival as a
function of percentage of fines smaller than 6.35 mm in
laboratory troughs and gravel mixes,
relation to percentage of fines smaller than 9.5 and 0.85 mm
(Figure C.23) .
The data of Irving and Bjornn, taken at face value, tend to
demonstrate that tolerance of higher percentages of fines (< 0.85
and < 6.35 mm) as depicted in figures C.22 and C.23 is lower for
cutthroat and rainbow trout and kokanee salmon than for chinook
salmon. Irving and Bjornn state that "rainbow and cutthroat
trout and kokanee salmon, tolerated gravels with more fine
particles than Chinook salmon studied by Tappel and Bjornn
(1983)." In making this statement, they refer to a table of
coefficients of determination (r^J of survival as a function of
80
/SECTION II
-------
Cutthroat Trout
10 2O 30 4O 50 60
UJ
o
ce
ui
Q.
PERCENT SMALLER THAN f.Bmm
M
to
d
Z
X
X
tn
ui
Ul
O
(E
UI
o.
30-
20-
10-
Rainbow Trout
P«-118.8-l6.«<80.»»)-0.00rl88.il1
20%
40-|
30-
20.
10-
Kokanea Salmon
4a«
20*
0%
80%
10 20 30 40 80 60
Slaalhaad Trout
40-
30-
20-
10-
40-
30-
20-
10-
10 20 30 40 SO 60
Chinook Salmon
Pt • tS.« - OJ' Isolines of pre-
to mergence in relation to percentage
°'85% in laborat°^ Boughs and
various particle sizes for support of their statement. In fact,
r2 values describe the fit of data to regressions, not the slope'
and thus cannot be used in support.
81
/SECTION II
-------
More importantly, why should chinook salmon have a higher
tolerance to fines than cutthroat trout, kokanee salmon, rainbow
trout, or steelhead (figures C. 22 and C.23) in laboratory mixes
of gravels? The answer probably lies partly with embryo size and
alevin strength. Laboratory mixes of gravels are not packed
tightly by substrate shifts or intrusion of additional sediments
during the incubation process. The relatively large chinook
salmon alevins may butt their way to the surface more success-
fully in these mixes.
The answer more clearly lies with depth of burial of embryos
in the laboratory channels. The channel diagram provided by
Irving and Bjornn indicates a burial depth of about 20 cm (8 in)
for embryos, rather deeper than burial depths in natural redds
for small trout and kokanee salmon, but less than the depth of
chinook salmon egg pockets. In fact, the data used by Irving and
Bjornn (1984) for chinook salmon were obtained from Tappel and
Bjornn (1983), who stated that chinook salmon in Vibert boxes
were actually placed at a depth of about 15-20 cm. But Tappel
and Bjornn provided a diagram of the experimental apparatus that
they used, which showed that the total depth of gravels in the
trough ranged from 15 to 20 cm and the center of the Vibert boxes
lay at about 12-15 cm below the gravel surface. The depth at
which alevins first encountered the experimental gravel mixes lay
approximately at the top of the Vibert boxes. Whether embryos
were actually at 15-20 cm or 12-15 cm, these depths are consid-
erably less than the depths from which chinook salmon and
steelhead alevins must emerge in natural egg pockets. The
relationships shown in figures C.22 and C.23 are consistent with
this explanation.
The point of the foregoing comments is that the conditions
imposed in laboratory research will determine results. It is
inappropriate to extrapolate such information to field condi-
tions. One cannot quantitatively estimate survival of chinook
82 /SECTION II
-------
salmon and steelhead in the field based on gravel composition
through use of laboratory data, for example, as did Stowell et
al. (1983) and Talbert (1985).
The graphs provided by Tappel and Bjornn (1983) and Irving
and Bjornn (1984) may tempt the reader to extrapolate laboratory
data to the field, yet these authors noted " predictions of
embryo survival generated by the equations may be inaccurate when
applied to field conditions" (Tappel and Bjornn 1983), and
"Embryo survival in natural stream (sic) may nor may not match
the rates presented in these papers" (Irving and Bjornn 1984).
Tappel and Bjornn (1983) and Irving and Bjornn (1984) sug-
gest that the greatest applicability of their model functions is
in predicting the relative change in embryo survival rates that
may occur if changes occur in the spawning and incubation sub-
strate. We submit that the greatest applicability of their
laboratory data, however elegant the laboratory approach and
resultant data, is to conditions in the laboratory. It is
incorrect to assume, for example, that a 10% incremental increase
in particles smaller than 0.85 mm will result in a predictable
decline in embryo survival of a given salmonid in a field
environment, as is implied by these authors and by others who
assume that laboratory data can be quantitatatively applied to
the field. We explain why in several places in our report, but
draw the attention of the reader particularly to the section on
predictive tools for assessing the effects of fine sediments in
the field.
Everest et al. (1981) also criticized laboratory studies of
effects of intragravel conditions on survival to emergence
because they are not useful in predicting survival to emergence
in the field. Gravel mixes used in the laboratory bear little
resemblance to particle size composition within the vertical
zones of egg pockets found in nature, and planting of eyed eggs
83 /SECTION II
-------
at uniform depth is not representative of nature, according to
these workers. They concluded that only vertical subsampling of
gravel cores from natural environments will show the actual
conditions that fry must face during emergence.
We point out, however, that embryos are usually found at the
lower portion of the egg pocket in nature in a very limited
vertical zone, making the criticism of uniform burial depth moot.
More critical is the fact that laboratory studies fail to dupli-
cate egg pocket structure, and merely place a thoroughly-mixed
substrate matrix in a study "cell". This point is discussed
further in the next section.
It is appropriate to treat laboratory studies of embryo
survival as models useful in assessing mechanistic responses
rather than as exact analogs of nature that permit accurate
assessment of quantitative biological response.
84 /SECTION II
-------
D. EMERGENCE FROM GRAVELS
D.I. Entrapment by fines
White (1942) reported that where Atlantic salmon incubated
in areas with high quantities of sand, 80% of the eggs were dead
and 20% of the complement could not emerge through the compact
sand layer. White found entombed fry even in redds where
emergence had occurred.
Koski (1966) excavated in known redds of coho salmon that
had been surrounded with netting to assess emergence success. He
found one redd in which 260 emaciated dead fry were present sev-
eral inches below the surface of the gravel. No data on gravel
composition are available for that redd.
Phillips et al. (1975) followed Koski's observation of
entombed fry by examining effects of fines on emergence success.
They prepared 8 mixtures of sand and gravel, then inserted coho
and steelhead fry into the substrate through a vertical standpipe
arrangement. Emergence success declined (from near 100%) above
about 10% fines (1-3 mm) for steelhead and coho. Presence of 20%
fines reduced emergence success about 60-70% (Figure D.I).
J 1WVC
ao
80
40
zo
o
c
*•» JN,^
1 \
> 10 2
;
K
i\
0 J
k
•1
!\^
1
0 4
"*^.
j
*K-
1
0 !
- ^ — f
• MEAN'
—BANGt,
i^s-L
i- j"
,' " v~~
SO 6O 7
COHO-.-
J
4
>
e
3
«* *
£ X
§ 5
M 4
*:
1.0
0.9
^ STEEUHEflQ^
Y- I.466-O.OI6X
COHO
Y« ).300-0.OI7X
—' ' '
_i_
0 fO 20 30 40 90 60 70
PERCENTAGE OF I-3 MM SAND
Figure D.I. (From Phillips et al. 1975). Survival to emergence
of steelhead and coho salmon from gravels with various
percentages of sand 1-3 mm.
85
/SECTION II
-------
The amount of fines (< 3.3 mm) in spawning gravels used by
coho salmon in unlogged Oregon watersheds varied from 27 to 55%
(Koski 1966 and Moring and Lantz 1974). The implication that one
might draw, using laboratory studies of survival to emergence in
various gravel mixes (Phillips et al. 1975), could be that
survivals of coho salmon and steelhead from pre-emergent state to
emergence would be 25-50% in undisturbed environments. The
incremental effects of incubation from deposition to pre-emergent
state would be subtracted from these percentages. However, the
data of Phillips et al. (1975) only illustrate that emergence
success declined in the laboratory mixes of gravel that contained
high percentages of fines. They do not permit quantitative
predictions in field situations, partly because no information is
available for the egg-to-alevin stage and partly because fines in
"spawning gravels" cannot safely be used as an indicator of
conditions in egg pockets, or even in "redds".
Data from natural redds as reported by Koski (1966) and
Tagart (1984) support a mean survival from deposition to
emergence of about 27 and 30%, respectively, in undisturbed
(Alsea watershed) and partially-logged (Clearwater River,
Washington) drainages. We use these data extensively as examples
in the report section on predictive tools.
Bjornn (1969) studied survival and emergence of steelhead
trout and Chinook salmon in gravels with various amounts of
granitic sands in troughs. Steelhead emerged with undiminished
success (about 50%, calculated from green egg placement) in
gravels with percentages of sand (< 6 mm) as high as 20% (Figure
D.2). Chinook emergence success was undiminished (about 70-75%,
from green eggs to emergence) until sands exceeded about 10-
15%.(Figure D.2). "Control" survivals were adjusted upward by
Bjornn to about 90% for both species because of variables un-
related to fines, but without effect on the trends in survival
86 /SECTION II
-------
10
20 30 40 90
PtRCCNtAOC SAND
10
TO
Figure D.2. (From Bjornn 1969). Percent survival of steelhead
and Chinook salmon to emergence from various mixes of sand.
after given percentages of sand were reached in tests.
NCASI (1984b) reported significant negative relationships
between survival and percentages of fines smaller than 0.8 mm in
test mixes of gravels, probably because of interference of sands
with emergence, inasmuch as dissolved oxygen content did not
differ in various mixes of sand.
The remarks in the previous section concerning application
of laboratory data to field situations pertain in reference to
the laboratory data of Bjornn (1969) and NCASI (I984b). The data
demonstrate that emergence success declines in gravels with a
high percentage of fines. They do not permit prediction of emer-
gence in natural redds from knowledge of percentage of fines in
"spawning gravels" or "redds" and survivals in laboratory
situations.
87 /SECTION II
-------
As a further commentary on the criticisms of Everest et al.
(1981) of laboratory studies, it should be noted that use of an
"emergence box" such as a Vibert perforated container that
permits fish to leave through holes, may slightly vitiate
criticism of unrealistic gravel mixes by providing a simulated
egg pocket. The word "may" is required because even a Vibert box
cannot provide a correct analog of the egg pocket structure and
attendant bridging of fines in the pocket centrum and higher in
the pocket.
The reason the egg pocket is important in simulation exer-
cises is that it tends to be structured toward large particles
while upper redd portions tend to have more fines. This means
that emerging alevins begin moving vertically from an area that
in theory should have more pore space. Dr. W. Platts reported
that a chinook salmon redd component, removed intact with
multiple freeze probes, had what appeared to be "tunnels" through
the egg pocket (W. Platts, personal communication).
Movement upward by alevins, as reported by Bams (1969),
promotes dropping of fines into the deeper pores. It is very
likely that this gradation is an important component of
intragravel ecology. Clearly, exceptions to the gradation will
occur, especially where very fine particles are added to a
cleansed substrate (Adams and Beschta 1982). In geographic areas
where sands make up the bulk of .the fines, gradation and bridging
are very likely to be important.
D.2. Effects of fines on size of emergent fry
The effect of fine sediments on size of emergent fry has
been reported by several workers. Tappel and Bjornn (1983) found
that size of steelhead fry that emerged from gravels with low
percentages of fines slightly exceeded that of fry from gravels
with high percentages of fines, but size of chinook salmon fry
88 /SECTION II
-------
varied little through the range of experimental gravel mixtures.
The effect of different incubation histories (steelhead were
placed in gravels as newly-fertilized ova while Chinook salmon
were placed in gravels as eyed embryos) is unknown. The obser-
vations of Tappel and Bjornn on effect of fines on emergent fry
size should be considered in light of the shallow depth of burial
of Chinook salmon embryos in the laboratory mixes of gravel. Had
the Chinook salmon been forced to emerge through 25-30 cm, the
results might have been quite different.
Inasmuch as gravels with high percentages of fines and low
permeability tend to have low. dissolved oxygen levels, embryo
development is slower. Hence size of emergent fry would be
reduced, with the potential subsequent ecological disadvantages
noted by Mason (1969).
Phillips et al. (1975) reported that coho salmon fry that
emerged from high percentages of sand were smaller than those
from gravels with low percentages, but that steelhead fry were
similar in size after emergence. Koski (1966) found that coho
salmon size at emergence directly related to permeability of the
substrate (Figure D.3). In a study of intragravel ecology of
chum salmon, Koski (1981) showed that fish emerging from gravels
with high proportions of sand were smaller. He attributed this
to restriction by sand of size of fish that could physically
emerge.
Hausle and Coble (1976) were unable to find a relationship
between percentages of sand and size of emerging brook trout, but
the sand mixes ranged from 0 to 25% and overall survivals from
hatching to emergence averaged 70% Absence of high percentages
of sand probably prevented definition of a size-related effect.
NCASI (1984b) also found no effect of fines on size of emerging
rainbow trout. We cannot ascertain the reason for this in their
report,
89 /SECTION II
-------
u
S
:
k>
Q9.40
* «g no mg 24« iflo us
WI1AH BMMeABIUTT INDEX (MILUUTCRS/i SECOHOSt
Figure D.3. (From Koski 1966). Mean weight of emergent coho fry
from natural redds in relation to the mean gravel permeability
index in the redd.
McCuddin (1977) stated that he found no relationship between
fines and size of emerging Chinook salmon or steelhead, although
survival was related to percentage of fines. McCuddin's labor-
atory procedures involved placement of fry in a horizontal pipe
section 25 cm long and 10 cm in diameter. Numerous 1 cm holes
were drilled in the pipe section. His diagrams indicate that the
horizontal pipe lay on the bottom of his experimental troughs,
and he states that his troughs contained 25 cm of gravel. A
vertical standpipe, connected to the horizontal pipe, permitted
gentle introduction of alevins to the horizontal emergence
chamber, to which the fish were then confined by a foam rubber
stopper. Introduced fry could emerge by passing through the 1 cm
holes or out the open ends of the horizontal pipe. The top of
the horizontal pipe cylinder lay only 15 cm below the gravel
surface. McCuddin placed newly-fertilized eggs 21 cm beneath the
gravel surface (in gravel without the horizontal pipe), and later
90 /SECTION II
-------
introduced swim-up alevins to horizontal pipe chambers. He then
compared survival to emergence of swim-up fry and newly fertil-
ized eggs, and drew several conclusions regarding effects of
fines on emergence. One of these was that emergence success of
swim-up fry was greater than had been reported by other workers.
He compared emergence success of fish that passed upward through
21 cm of gravel matrix with that of fish that emerged through as
little as 15 cm of matrix in his systems. The latter group
initiated upward movement from an open cylinder with a volume of
1.96 liters (probably reduced somewhat by spill-in of gravels
into the ends of the pipes).
Tagart (1984) reported a strong inverse correlation between
percentage of fines and mean length of fry emerging from natural
redds. He also found a positive correlation between dissolved
oxygen concentration in redds and fry length at emergence, and a
negative relationship between percentage of fines and dissolved
oxygen. The relative effects of these variables are unknown. In
the sense of empirical effect, it is unimportant whether low
dissolved oxygen or impeded emergence reduced size and survival.
In terms of improved understanding of intragravel ecology, how-
ever, such distinctions are of obvious importance.
Smaller fry, bearing a substantial yolk sac, may move
through small substrate particles more easily than larger fry.
The egg sac is malleable, and may almost "ooze" through pore
constrictions. Dill and Northcote (1970) investigated intra-
gravel movements of coho salmon, showing both vertical and
lateral movement, but the gravel sizes used in their study were
too large (1.9-6.3 cm) to permit inference about normal sub-
strata with fines. Bams (1969) suggested that presence of large
quantities of fines may stress embryos and lead to premature
emergence.
Inasmuch as embryo and alevin size tends to be related to
91 /SECTION II
-------
adult female size and to species, some of the conflicting results
on effects of fines on emergence size is probably a function of
female size. In other words, given the same mix of fines, one
would expect that rainbow trout fry would better be able to exit
the gravel than would Chinook salmon fry. Corrections for
species differences in embryo size were incorporated by Shirazi
et al. (1981) (Figure II.A.3).
MacCrimmon and Gots (1986) investigated effects of various
sediment mixes on emergence success and size of rainbow trout
alevins. They found a strong positive relationship between
alevin size (weight and length) at emergence and the geometric
mean particle size. The mixes with high proportions of fines led
to much earlier emergence of fry that did not have fully absorbed
egg sacs. We are unable to determine why NCASI (1984b) and
MacCrimmon and Gots (1986) obtained contradictory results, but
the reason may involve gravel types and percentages of various
particle sizes, which interacted with percentages of fines.
D.3. Effects of fines on emergence timing
Bams (1969) described the emergence-oriented behavior of
sockeye salmon fry as swimming motions upward, presumably
oriented to gravitational force. Normal emergence movements are
slow and appear restrained, with long periods of rest between
movements. However, in favorable substrata, movements of 5 cm
per minute were frequently recorded. Bams described the field of
movement of fish released from a given point as an inverted cone
with a vertical axis. Fry can drop backward or pull themselves
backward by flexion and a pulling action provided by tail
purchase.
Bams reported that when fish confronted a sand barrier near
the surface of an experimental gravel bed, they "butted" upward
with repeated short thrusts. This action loosened the sand
grains, which fell down and past the butting fish, forming an
92 /SECTION II
-------
open passage as the fish worked upward. This behavior can be
related to descriptions in Section B.I of redd structure and
bridging of fines. Bridging within the egg pocket may be
breached by butting behavior.
Koski (1966) reported mean length of the emergence period in
coho redds was 30-39 days, and that 90% of fry emerged from redds
in 15-20 days. The number of days to first emergence was
inversely but loosely related to the amount of fines smaller than
3.3 mm in redds of coho salmon (Figure D.4). The total period of
emergence was greatest for redds with highest percentages of
fines.
u
o
z
u
13
04
u
a
o
t*
VI
Q
b,
O
A
U
a
2
§
I1S-
i 1 4
A 1 V
113
112-
III
110-
I0<»
108
107-
106
IOS-
104
103-
102-
• 9 Deer Creek
• A - Ftynn Creek
m- Needle Br.
A Ik
A
• •
A
• A
m
m
B
*
A A
• • A
1 — i i t 1 1 1 1 1 1 1
18 30 3Z 34 36 3d 40 42 44 46 48
PERCENT GRAVEL SMALLER THAN
3.3Z7 MILLIMETERS
Figure D.4. (From Koski 1966). First emergence of coho fry from
natural redds in relation to gravel composition.
For chum salmon, Koski (1975) noted that the number of
93
/SECTION II
-------
temperature units required for the first 5% of emerging fry to
reach the surface decreased with increasing percentages of fines
(Figure D.5). This may show that high sand compositions cause
some stress in alevins in the substrate, leading to rapid
emergence.
1900
1900
1700
I«OO
ISOO
t MOO
^
*
I IBOO
• I7OO
N
^ ISOO
I isoo
I MOO
f 1300
IZOO-
1987 -«»
i
»oo
I9OO
ISOO
1700
HOO
1500
MOO
M!
«7
871
BIS
7 SO
1968 -«9
98!
927
an
8I«
7SO
709
•MS
rt»
»,itf*>i
W93
•038
BTI
em
» *O M SO
Figure D.5. (From Koski 1975). Relation between the percent of
fines <3.3 mm in the gravel and temperature units needed for
first 5% of chum salmon fry to emerge.
Hausle and Coble (1976) recorded increases, rather than the
decreases reported by most workers, in the time required for
emergence of brook trout in gravels that contained higher
percentages of fines (< 2.0 mm).
McCuddin (1977) stated that he found no relationship between
94 /SECTION II
-------
timing of fry emergence in chinook salmon and steelhead and
percentages of sand in the substrate. However, his analysis may
have been incomplete. Figure D. 6, although not sufficiently
labeled for the reader to identify the species and year, shows
considerable differences. In the top graph, emergence through
0-22% sand peaked earlier than at 41-52% sand. On the contrary,
emergence peaked earlier at 52% sand in the bottom graph. Based
on other information in his thesis, we believe the top graph is
for steelhead; the bottom for chinook salmon. Although
McCuddin's data do not shed light on the cause for the diff-
erence, it may relate to stress level in the two species. The
behavior of chinook salmon seemed to parallel that recorded for
chum salmon by Koski (1975), a study in which higher percentages
of fines correlated inversely with number of temperature units
required for emergence. Steelhead observed by McCuddin seemed to
behave in a manner opposite to Koski's chum salmon. The degree
to which laboratory circumstances affected the results is
unknown.
Figure D.6. (From McCuddin. 1977). Number of fry emerging daily
from trough mixes of sand of various percentages.
MacCrimmon and Gots (1986) found that most rainbow trout
alevins incubated in 60-100% fines loadings began moving toward
95
/SECTION II
-------
the surface of the substrate immediately after hatching, while
those in 0-20% mixes tended to move deeper in the incubator
columns. MacCrimmon and Gots (1986) also found in uniform
substrata that alevins emerged earlier in finer homogeneous
gravels than in coarser gravels.
The weight of evidence shows that alevins emerge earlier
from gravels with high percentages of fines. We interpret this
as an adaptive mechanism that increases survival. Head and body
size increase as the yolk sac is absorbed, which should make
passage through fines more difficult. Early emergence would
trade mortality in the substrate against mortality caused by
early emergence into surface waters.
Bams (1969) reported that pre-emergent sockeye salmon could
be induced to move out of the gravel by a reduction in flow in
the redd environment, while for alevins at an earlier stage of
development a flow reduction led to burrowing. Bams explained
these adaptations by noting that escape into the stream might be
appropriate for survival of fish nearly ready to emerge, but that
burrowing would be appropriate for alevins with much yolk
unabsorbed.
96 /SECTION II
-------
III. SUBSTRATE CHARACTER AND ECOLOGY OF REARING SALMONIDS
A. SUBSTRATE CHARACTERIZATION
Substrate characterization in rearing and wintering habitat
for salmonids and for macroinvertebrates requires techniques
suited for quantification of large particles that are usually not
found in areas selected by salmonids for spawning, as well as
assessment of fines. Visual assessment, coring, embeddedness,
and free matrix particles each may have a role for particular
purposes.
A.I. Visual assessment
An example of visual assessment utility is provided by
substrate surface evaluations used by the Instream Flow
Incremental Methodology (IFIM). IFIM practitioners place several
transects across study sites, and obtain various measurements
along a tape, or tag line, stetched over the stream on each
transect. Conditions in each segment are defined fay measurements
or assessments at the ends of the segment, called "verticals".
Thus, substrate condition in each segment usually consists of the
average of two measurements. This average condition is assumed
to obtain in the "cell" bounded by the tape, the segment ends,
and a cross-stream line usually halfway upstream toward the next
transect and another cross-stream line usually halfway downstream
to the next transect (Figure A.l).
Quantification of substrate at each measurement point is
visual, with a coding system of complexity that depends on
investigator decision. The first digit, for example, could
reflect the dominant particle size, the second digit the second
most dominant size, the third a percentage of fines smaller than
6 mm. Coding could include embeddedness level if required.
97 /SECTION III
-------
TRANSECTS ESTABLISHED
A. A study site Is selected to represent a homogeneous
stream reach. Starting at a hydraulic control, transects
are placed across the stream at Intervals determined
ay tfw channel configuration.
O POINT MEASUREMENTS OF DEPTH,
VELOCITY. AND SUBSTATE
T^
8. Point measurements of depth, velocity, and substrate
are spaced along the transects. The measurements
am taken initially wherever there Is a change In the
bottom profile of the stream channel. Subsequent
measurements during changed /low conditions
an taken at the same location as the original
measurements.
C. Each cross-section Is divided into subsections, and the
po** measurements are used to determine average
values of depth, velocity, and substrate for each sub-
section. The subsection values along the transects
are assumed to extend halfway to the adfacent
cfoss-sectfons.
O. A matrix ot rectangular cells Is created with each cell
having an average depth, velocity, end substrate. The
original transect placement Is critical and illustrates
the need for knowledge ot the computer procedures
before attempting Held work
Figure A.I. (From Hilgert 1982).
into cells.
Subdivision of IFIM study site
Table A.I offers an example of codings.
Substrate scoring has been used by workers to describe
habitat suitability for aquatic insects (Sandine 1974, Bjornn et
al. 1977, Brusven and Prather (1974), and for fish (Grouse et al.
1981). Scoring categories differ somewhat among workers,
primarily because of sieve size differences. An example of such
substrate scoring is offered in Table A.2 (Grouse et al. 1981).
The predominant particle size is assigned a rank number, as is
the second most dominant substrate. The third rank is the size
of material surrounding the predominant substrate particles, and
98
/SECTION III
-------
Table A.I. Examples of substrate codings used in IFIM studies,
(From WDF 1983).
Code
Description
Diameter
mm in
0
1
2
3
4
5
6
7
3
9
Organic detritus
Silt, clay
Sand
Small gravel
Medium gravel
Large gravel
Small cobble
Large cobble
Boulder
Bedrock
<2
<2
2-12
12-38
38-76
76-152
152-305
>305
<0.1
<0.1
0.1-0.5
0.5-1.5
1.5-3.0
3,0-6.0
6.0-12.0
>12. 0
Table A.2. (From Grouse et al. 1981). Substrate characteristics
and associated ranks for calculation of Substrate Scores, as
modified from Sandine (1974).
Rank
1
2
3
4
5
6
7
8
1
2
3
4
5
Characteristic
particle size or type
Organic cover over 50% of bottom surface
< 1-2 mm
2-5 mm
5-25 mm
25-50 mm
50-100 mm
100-250 mm
> 250 mm
Embeddedn e s s a
Completely embedded, or nearly so
3/4 embedded
1/2 embedded
1/4 embedded
Unembedded
a - Extent to which predominant-sized particles are covered by
finer sediments.
the fourth rank is the level of embeddedness of predominant
substrate by material ranked in the third evaluation. The sum of
the ranks constitutes a single Substrate Score.
Substrate score was related by Grouse et al. (1981) to dg,
as shown in Figure A. 2. The visual scoring system appears to
99
/SECTION III
-------
correlate reasonably well with
E
E
50
40
UJ
2
o
tE
30
Ul 20
O
UJ
10
r-.96
JL
_U
_L
10 II 12 13 14 19 16 17
SUBSTRATE SCORE
18
Figure A. 2. (From Grouse et al. 1981). Geometric mean particle
size of laboratory sediments and substrate score obtained by
evaluating four sample areas in each channel. Points are means
of two replications.
Shirazi and Seim (1981) assessed the efficacy and accuracy
of visual assessment of substrate in connection with evaluation
of spawning gravels. Section II.A of the current review
discusses this assessment. Criticism of the visual assessment
method (Everest et al. 1981) as failing to determine conditions
in the egg pocket would not pertain as appropriately to
descriptions of the substrate relevant to juvenile rearing.
However, examination of surface conditions will not adequately
measure crevice availability for macroinvertebrate hiding or for
salmonid refugia in winter, especially in armored substrate
surface zones. At best, visual assessments would be indicators
of microhabitat conditions in the surface zone in areas not
armored.
Konopacky (1984) compared substrate scoring with mean par-
ticle size and percentage of fines assessed in substrate coring
100
/SECTION III
-------
in several streams in central Idaho. He believed he could
determine any two of these statistics from the third. Figures
A.3 and A.4 show his data in two forms.
I"
Is
•V «t
imwv- LOWCT
it.« tut IKH ILK
CIK-
If DIME Ml ID
Figure A.3. (From Konopacfcy 1984) . Substrate score, percent
fines, and mean particle size of riffles in several streams of
central Idaho.
Ocular assessment of surface fines as an indicator of gravel
composition has been abandoned on the Payette National Forest
because of failure of the surface system to accurately (in
comparison with core samples) detect the percentage of fines
smaller than 6.3 mm (D. Burns, Payette National Forest, personal
communication). It cannot discriminate between heavily-
sedimented and pristine habitats in the range of values found on
the Forest.
101
/SECTION III
-------
"
V«»* »l**Wlt tlfl. «"t* f
Figure A.4. (From Konopacky 1984). Relationships between percent
sediment and mean riffle particle size, percent sediment, and
substrate score for several streams in central Idaho.
A.2. Photographic measurement
Chapman et al. (1986) determined percentages of 3 groups of
particle diameters on photos of the substrate surface in exposed
gravels and underwater in the Columbia River to depths of 10 feet
in high water velocities. Figures A.5 and A.6 offer examples of
the photos and data that can be developed from them. The
percentage of materials smaller than 7.6 cm was a suitable
indicator of surface gravels at various points across the river
channel, but detection of fines in the <6 mm category was not
attempted. Surface armoring made the attempt pointless.
Burns (1978) used photographs to measure surface substrate
composition (section II.A). The Payette National Forest has
since abandoned use of photos for substrate characterization.
Photographic assessment could not distinguish between areas where
embeddedness was 19% and those where it equalled 30%. Comparison
102 /SECTION III
-------
Figure A.5. (From Chapman et al. 1985). Photographs of substrate
in the Columbia River at fixed distance from the camera. Scale
marks are 2.54 cm apart.
of core samples with photo evaluations showed that where the
percentage of fines smaller than 6.3 mm was equal to or over 30%,
photo-assessed surface fines were estimated as only 10%. This
seems reasonable, for armoring would probably make the gravel
surface unlike gravels at two or more inches below the surface.
A.3. Core samples
Core samples, whether obtained by freeze probes or in McNeil
cylinders, offer the most complete assessment of substrate com-
ponents. Study objectives would determine whether core strata
should be analyzed as a lumped sample to the depth normally
examined in assessment of spawning gravels to evaluate potential
103
/SECTION III
-------
100-
H Lwgw ttwn 112cm
TOkcta
LEGEND
o Lcnmr transact
a Mkfefl* transact
SDkcta
36 kefs -00m
STATION
-6On
-90m
Figure A.6. (From Chapman et al. 1985). Mean percentages of
gravels smaller than 7.6 cm and larger than 13.2 cm as measured
from photographs along a transect normal to streamflow, both
above the stream edge and at depth.
for winter habitat for salmonids or confined to the top few cm of
the core, the zone more .important to macroinvertebrates and
salmonid fry.
Sieve examination of substrate cores appears an awkward way
to approach evaluation of suitability of substrate for summer
104
/SECTION III
-------
rearing or winter hiding habitat. In fact, visual methods,
photographic techniques, or embeddedness measurements seem more
direct. However, measures of coarseness may have utility. Much
attention has been devoted to fines because of their effects upon
reproductive success, but the other end of the particle scale has
been virtually ignored.
Part of the problem with core samples as a descriptor of
winter habitat is that they are difficult to obtain in the stream
zones where they are most needed, that is, in areas with the
higher proportions of rubble and boulders. In theory, a descrip-
tor such as "% finer than 100 mm" should offer utility for
analysis of winter refugia. However, if this descriptor were
adopted and coring used to define winter habitat, variance would
be high (Adams and Beschta 1980) and bias would increase. Large
particles are not sampled satisfactorily by McNeil cylinders (the
sampler is stymied when two large particles lie partly in and
partly out of the cylinder and the cylinder cannot be forced deep
enough into the substrate) or by freeze probes (large particles
are frequently frozen to the core only at one end and are
extracted without a surrounding matrix of small particles).
In stream zones dominated by medium gravels and smaller
particles, freeze-cores or samples from McNeil cylinders have
less bias associated with them. Thus for some streams, coring
may be useful as a substrate descriptor. (Carried to extreme,
this would mean that these tools would work best as descriptors
of rearing conditions in streams badly damaged by sedimentation.)
Geometric mean particle size and geometric variance or a
fredle index offer descriptors of the substrate, but suffer from
bias associated with the sampling tools. Thus, failure of the
McNeil or freeze-core systems to accurately sample large
particles would bias dg as well as percentage of particles in
large size categories. Bias would be reduced in these situations
105 /SECTION III
-------
by increased core sampler diameters, but with an obvious cost in
practicality.
Cores can serve to evaluate such engineering features as
road crossings. Figure A.7 illustrates temporal changes in fines
as a result of road construction measures over 3 years in one
stream.
[Pj Upitr»»mSIU
QB Oownjtr»»m Silt
Fin*
1983
1984
1993
Figure A.7. (From Munther and Frank 1986b). Percent fine
sediment <6.35 mm and > 0.21 mm from core samples in Randolph
Creek (Lolo National Forest).
A.4. Embeddedness
Embeddedness is generically defined as the amount of fine
sediment that is deposited in the interstices between larger
stream substrate particles (Burns 1984, Burns and Edwards 1985).
Embeddedness ratings.have been developed and applied in rearing
and overwintering habitat rather than in spawning gravels.
Kelley and Dettman (1980) measured embeddedness of particles
larger than 4.5 cm in a California stream, and Klamt (1976)
estimated the degree to which key rocks or dominant rocks in
streams were embedded (using 25, 50, and 75% as embeddedness
levels).
106
/SECTION III
-------
Specifically, Burns used embeddedness level to refer to the
proportion of an individual matrix particle (4.5 to 30.0 cm in
greatest diameter) surrounded by fine sediment (< 6.3 mm in
diameter). The proportion is calculated by Burns and Edwards
(1985) as:
E = d2 (100)/dx , where
E = % embeddedness,
di = longest diameter of a matrix particle of 4.5-30 cm
greatest diameter at right angles to the plane of depos-
ition of fine particles (< 6.3 mm diameter), and
d2 = distance along d± covered by fine sediment (< 6.3 mm
diameter) or "embedded" in the stream bottom.
Figure A.8 demonstrates the relationships among variables used to
calculate embeddedness.
The population of single matrix particles must be sampled to
characterize substrate conditions (Burns 1984) . Burns treated an
embeddedness measurement made for one rock as one observation.
Burns used a 60 cm steel hoop to define particles in the
substrate to be measured, a 30 cm transparent ruler to measure
lineal dimensions and water depth, and a float and stopwatch to
measure water velocity. These simple tools offer practical means
F.ttiel*
(Perpendicular to PE)
d,
E- ^(100)
Figure A.8. (From Burns and Edwards 1985)
criteria. See text above.
Embeddedness
107
/SECTION III
-------
of assessing embeddedness, provided the measure itself is deemed
useful.
Burns examined embeddedness in a specific stratum within
various streams, making no effort to obtain a simple random
sample or even a stratified random sample of conditions
reflecting average stream character. His approach should
substantially reduce variance. He chose a sample stratum in each
stream that had laminar surface flow over a cobble bottom suited
for winter cover selection by overwintering juvenile salmonids.
Kelley and Dettman (1980) used a random-toss method to quantify
embeddedness for assessment of general stream condition.
Burns sampled embeddedness in 19 tributaries to the South
Fork of the Salmon River, chosen to represent the full range of
past development ("undeveloped", "partially developed" with low
road mileage constructed, and "developed" with heavily roaded
mileage). Burns found that streams with more development had
more-embedded substrate than undeveloped or partially developed
streams. The embeddedness in developed, partially developed, and
undeveloped streams averaged a respective 44, 24, and 24%. The
three categories had significantly different mean embeddedness (p
= 0.00009). Developed streams had a significantly higher mean
embeddedness than the other two categories (p = 0.05). Burns was
unable to distinguish partially-developed and undeveloped streams
by embeddedness information.
Burns (1984) regressed percentages of fine sediment from
core samples (Lund 1982 and Corley and Newberry 1982) against
mean embeddedness for 11 sites for which both measures were
available. His regression, significant at p = 0.01, had an r2 =
0.64 (Figure A.9).
108 /SECTION III
-------
, r » a.m. 41 -
If*** pvfrtu* f tfw
Figure A.9. (From Burns 1984). A regression between mean % fines
from core samples (Lund 1982, Corley and Newberry 1982)
collected in 1981 and mean % embeddedness assessed in 1983 from
11 locations in the South Fork Salmon River.
Each data point in Figure A.9 represented a mean derived
from 40 core samples and at least 100 individual matrix component
(individual rock) embeddedness measurements. There appears to be
a real relationship between embeddedness and percentage of fines.
Burns did not address temporal changes or variation of embedded-
ness measures within a sample site. He also regressed relative
frequency of free matrix particles (loose rocks) against mean
embeddedness. This regression (r2 = 0.82), significant at p =
0.01 (Figure A.10), suggested that at the regression intercept of
45% embeddedness, no rocks were free, and that at 0% embedded-
ness, only about 85% of rocks would be free. A "free" particle
should probably be defined as one not wedged by either fines <6
mm in diameter or very fine gravel. The latter should explain
why 85% of rocks are free at 0% embeddedness. Burns suggested
that free matrix particles may offer a measure more sensitive
than embeddedness percentages in conditions between 0 and 50%.
109
/SECTION III
-------
60
Ul
n
at
0)
a
-o
-------
distance below the surface where armoring has occurred,
Kelley and Dettman (1980) found embeddedness a useful
measure of substrate character in Lagunitas Creek, California.
They did not indicate that they stratified habitats to eliminate
zones outside the criteria of Burns and Edwards (1985), and
apparently estimated embeddedness percentage on particles larger
than 45 mm in diameter visually rather than with the system of
Burns and Edwards.
Munther and Frank (1986 a,b,c) used embeddedness measure-
ments to quantify conditions in Montana streams. They noted that
excavation below the surface layer of substrate is commonly
needed to reach the substrate level where all further particles
are completely embedded. They removed all free matrix particles
from the area of the sample hoop, then proceeded to remove and
measure all embedded matrix particles. The resulting statistics
apparently are the same as those obtained by the embeddedness
techniques described by Burns (1984).
Munther and Frank (1986 a,b,c) regressed free matrix
particles on embeddedness (eg. Figure A.11). They reported
coefficients of determination of 0.73 to 0.92.
Embeddedness offers a useful "before and after" or "above
and below" measure of changes over time or space. Burns (1983
and 1984) and Burns and Edwards (1985) demonstrated that Mule
Creek, a Monumental Creek tributary, contributed to downstream
degradation of Monumental Creek, by using upstream-downstream
embeddedness measures. They also reported that Boulder Creek, a
tributary of the Little Salmon River, had high embeddedness (42%)
immediately downstream from logging and road construction
relative to an upstream control area (20% embeddedness).
Ill /SECTION III
-------
Embrddtdness
Embeddedness vs Free Matrix
Lolo N.F.
Figure A.11. (From Munther and Frank 1986a). Relation between
embeddedness and free matrix particles on the Lolo National
Forest.
0.90.
The final sentence in Burns and Edwards (1985) states that
data acquisition needs to be extensive before the extent and
degree of impact to fish habitat from man-caused sedimentation
can be properly evaluated. D. Burns (personal communication)
estimated that for samples of 100 particles in a carefully-
selected stratum, inter-mean differences of about 12-18% could be
detected, but that about 400 samples would be needed to assess an
inter-mean difference of about 5%.
Munther and Frank (1986c) compared similar morphologic sites
on developed and undeveloped streams. In 4 of 8 pairings of
riffles, tailouts, and runs, they found significant differences
in embeddedness between developed and undeveloped streams
(Table A.3).
112
/SECTION III
-------
Table A. 3. (From Munther and Frank 1986c). Results of t-tests
for comparison of embeddedness means of paired streams and
stations. Martin and Meadow creeks are developed drainages
(Meadow Creek reading 3.63 mi/section, 18.9 mi2, Martin Creek
roading 4.24 mi/section, 6.8 mi2) and Tolan and Moose creeks are
undeveloped (Tolan Creek roading 1.15 mi/section, 18.5 mi2, Moose
Creek roading 0.28 mi/section, 15.5 mi2).
Coopared Stations
Meadow riffle (Site 1, Sta A)
Tolan riffle (Site 1, Sta A)
Meadow riffle (Site 2, Sta A)
Tolan riffle (Site 1, Sta A)
Meadow pool tallout(Slte 2 Sta B)
Tolan pool tallout (Site 2 Sta A)
Moose pool tallout (Site 1 Sta A)
Martin pool tallout(Slte 1 Sta C)
Moose pool tallout (Site 1 Sta C)
Martin pool tallout(Slte 1 Sta C)
Moose riffle (Site 1, Sta B)
Martin riffle (Site 1, Sta A)
Moose riffle (Site 1, Sta E)
Mirtln rlffle(Slt« 1, Sta A)
Moose run (Site 1, Sta D)
Martin run (Sit* 1, Sta B)
Results Two tailed probability
Significant Difference
at .01 level
No Significant Difference
at .05 level
Significant Difference
at .01 level
Ho Significant Difference
at .05 level
Ho Significant Difference
at .05 level
No Significant Difference
at .05 level
Significant Difference
at .01 level
Significant Difference
at .05 level
.000
.974
.003
.064
.542
.145
.000
.030
It is important to sample embeddedness in a carefully-
defined stratum. Clearly, one should expect embeddedness to
differ among various stream gradients and in various positions
within a stream section, even within a pool. Burns and Edwards
(1985) described a series of criteria for sampling, which we
summarize below:
1. Float speed across a randomly-thrown steel hoop should
indicate a surface velocity of 24 cm/s to 66.7 cm/s.
2. Water depth should not.be less than 15 cm or the hoop or part
of the hoop lie in an eddy caused by a pool or large boulder.
3. Particles in the hoop should not all be less than 4.5 cm or
greater than 30 cm. (This restriction obviously implies that
some particles can be larger or smaller than these limits.)
113 /SECTION III
-------
When Burns (1984) measured embeddedness in spawning areas,
samples were not taken where water depth was less than 30 cm,
core sampling had disturbed the site, spawning had occurred, or
where particles in the hoop were all less than 4.5 cm or greater
than 30 cm. Samples were avoided where the hoop partly or
completely lay in an eddy caused by a pool or large boulder.
Klamt (1976), Kelley and Dettman (1980) and Burns (1984)
intended embeddedness measures to pertain to habitat suited for
rearing or winter hiding rather than for spawning gravels. In
severely armored surfaces, percentages of free particles may
offer useful ancillary measures of substrate condition for winter
hiding. In boulders, it may be impossible to obtain suitable
measures of embeddedness or percentage of free particles.
A.5. Sediment traps
A variant of analysis of gravel composition that may serve
in "above and below" and "before and after" comparisons is the
placement of buckets of washed gravels in pool tailouts above and
below such features as road crossings. An example of the results
of such comparisons, where the crossing caused sedimentation, is
found in Figure A.12.
Bucket samples still require sieving for determination of
percentages of fines. Theoretically, it should be possible to
withdraw the bucket from the substrate after covering it with a
plastic cap, then to evaporate all moisture from the bucket.
Weight of contents before placement and after evaporation should
measure increment of particles. This would shed no light on
particle sizes of incoming material. Munther and Frank covered
the buckets with "chicken wire" to minimize scouring of particles
on the surface of the bucket.
Buckets of cleaned gravels embedded in the substrate are
114 /SECTION III
-------
subject to effects of gravel shifting and vandalism. Neverthe-
less, they may serve in particular circumstances. An example
might be to evaluate increment of fine particles below a road
crossing during the period between late June (after spring
freshets) and October.
Gilt Edge
UJest Fk. Big Creek
Fin*
SHiment
M ••
12
10
8
6
Upstream SH*
Dovnstr*am Sit*
Eaaatii
1985
Figure A.12. (From Munther and Frank 1986a). Percent fine
sediment <6.35 mm and >0.212 mm in buckets in Gilt Edge Creek and
in West Fork of Big Creek above and below road crossings. Three
buckets of washed gravel were placed above and below each
crossing.
115
/SECTION III
-------
B. SUBSTRATE UTILIZATION BY REARING SALMONIDS
B.I. Newly-emerged frv
All salmonid fry utilize shallow water areas with low velo-
city after they emerge from the redd. In many streams these
areas also have a substrate composed of fine sediments and
organic debris, although fry also utilize quiet shallows in
rubble. Recent work by T. Hillman in the Wenatchee River (T.
Hillman, personal communication) indicates that substrate type is
not important for age 0 Chinook salmon and steelhead, but rather
that these fish key on other habitat features, such as velocity.
Everest and Chapman (1972) and Lister and Genoe (1970)
reported that fish use faster, deeper water as they become
larger, but that newly-emerged fry remain close to the stream
margin in quiet water. Availability of these marginal areas in
spring and summer is important, but would be considered unlikely
to control overall output of anadromous smolts or recruitment of
resident adult salmonids to the fishable phase (see discussion
later in this report on limiting factors).
B.2. Ea_rr_and fincrerlings
Effects of fines on juveniles in the size groupings classed
as age 0, I, and II are of particular concern because abundance
of these fish is generally conceded to affect recruitment of fish
to the fishable phase or to adulthood.
After the first few weeks of stream life, juvenile salmonids
can be called fingerlings or parr. They begin using deeper,
faster water for feeding (Everest and Chapman 1972, Lister and
Genoe 1970, Campbell and Neuner 1985). They associate with
velocity shear lines and occupy habitat much like that used by
adults.
116 /SECTION III
-------
These sites would usually have less sedimented substrata
than inshore areas, simply because of stream competency.
Saunders and Smith (1965) observed in a silted stream that brook
trout tended to associate with small patches of clean bottom, and
suggested that turbulence kept those areas clear, or that clear
areas provided better feeding places. Another explanation is
that clean patches may lie close to favorable velocity strata
that fish prefer.
Grouse et al. (1981) demonstrated in laboratory stream
channels that coho salmon production (g/m2/112 days) related
directly to substrate score in both spring and summer (Figure
B.I). Grouse et al. considered substrate scoring (see Table A.2)
as a reasonable and meaningful way to evaluate quality of rearing
habitat for salmonids.
ao
no
,.
it a M B
SUBSTRATE SCORE
V
§
3
cc
u
0.0
r-93
i ii i< a
SUBSTRATE SCORE
Figure B.I. (From Grouse et al. 1981). Cumulative production of
coho salmon (g/m2/112 days) in spring (left graph) and summer
(right graph) as a function of substrate score.
Figure B.I contains elements of trophic relationships
confounded with fish behavior because production equals growth in
biomass, whether the tissue survives or dies, hence is an amalgam
of instantaneous growth rate and fish density.
Alexander and Hansen (1983) experimentally reduced sandy
117 /SECTION III
-------
bedload sediments in a Michigan stream by means of a sediment
settling basin, and observed the control (upstream from sediment
basin) and treatment (downstream from sediment basin) reaches for
6 years. They used ratios of treatment to control populations,
growth, and production by size-grouped fish to evaluate effects
of reductions in bedload fines. The basin reduced sand bedload
by 86%.
Small brown and rainbow trout increased by 40% in the
treated area. Trout production increased 28%, but growth rate
changed little, hence most of the increase was associated with
increased numbers of fish (survival), and, apparently, with
improved habitat and production of macroinvertebrates. The
useful experimental approach of Alexander and Hansen (1983)
provides excellent and conclusive data on the negative effects of
sediment on population density and growth in the test stream. We
suggest that a similar experimental technique could profitably be
applied to one or more streams in Idaho.
Bjornn et al. (1977) studied effects of granitic sediments
on juvenile Chinook salmon, steelhead and cutthroat in central
Idaho streams. In laboratory stream studies they evaluated three
levels of sediment embeddedness, one-third, two-thirds and full
in 1974 and two levels, one-half and full in 1975. Juvenile
anadromous salmonids were placed in the channels in excess of
estimated carrying capacity and traps captured fish that
departed. The experiments used both wild and hatchery trout and
chinook salmon.
Figure B.2 depicts the density of fish 5 days after summer
experiments were begun. Test areas (with sediment embedded)
tended to hold less fish than did control environments that did
not have embedded sediments. Tests continued for 35 days with
age 0 steelhead of hatchery origin also showed that embedded
environments held less fish than control channels (Figure B.3).
118 /SECTION III
-------
Q C»Mi«l (wHfc.nl wJImwl)
to
•MCI11 •*••
• ••IMlll I/I
I
1
IH,
u—
i
M
f9
I
m
(i
|
"
•M
k*
•JhMM)
' 1
114 » * T 1
IN, IN] IM, IN, 5H. C«, CK,
WHO 1
•»-^HATCHtBT U W11B
Figure B.2. (From Bjornn et al. 1977). Densities of fish in
artificial stream channels after 5 days during summers of 1974
and 1975. SH^ = age I steelhead: CK0 = age 0 chinook; F = fully
embedded.
E3-...
i - »»-*,.,
Figure B.3. (From Bjornn et al. 1977). Density of age 0
hatchery steelhead 35 days after introduction to the test
channels. Fish could leave at will.
Konopacky (1984) placed steelhead trout and chinook salmon
(each in allopatry) in pools with rubble cover in laboratory
stream channels that were supplied by in-channel drift food from
either sand/pebble or gravel riffles. A downstream trap
permitted volitional residence. Densities of fish at the end of
119 /SECTION III
-------
the studies equalled 9-10 fish/m2, and fish did not grow in test
or control channels with either sand or gravel riffles. The
Konopacky (1984) results are difficult to relate to natural
streams because beginning densities of fish were extremely high
(steelhead densities of 10 fish/m2 and chinook salmon densities
of 12.5/m2)), the tests were conducted with large hatchery fish
(steelhead of 61 mm and chinook salmon 102 mm in length) that
were accustomed to hatchery fish loadings and naive to natural
drift at the beginning of the study, and all work was conducted
after mid-September.
One would expect that fish densities for naturally-produced
fish in Konopacky's channels would adjust, as a result of social
interaction, to under 1 fish/m2, but ending densities were
usually 9-10/m2. Given the limited food-producing area of
riffles in the artificial channels, it is not surprising that
fish could not grow, whether supplied from sand or gravel
riffles.
Other work by Konopacky in natural environments suggests
that each pool in his artificial channels should have held less
than 8 g of chinook salmon bioraass, or the equivalent of less
than 1 juvenile chinook salmon at 9 g, for the 3.36 m2 riffle
areas for production of drift above the pools. Density problems
obscured any effect of substrate type on fish production. How-
ever, the study concepts used by Konopacky merit pursuit, and
have promise for better defining relationships between riffle
condition and downstream fish production.
Alexander and Hansen (1986) experimentally added sand sedi-
ment to Hunt Creek, a Michigan stream. The addition increased
bedload 4-5 fold and significantly reduced brook trout numbers
and habitat. The brook trout population declined to less than
half its normal abundance, although fish growth was not affected.
Population adjustment occurred via a decrease in brook trout
120 /SECTION III
-------
survival rates, particularly in the egg to fry and/or fry to fall
fingerling stages of the life cycle. However, the authors
ascribed the deterioration in survival largely to loss of cover,
pool volume, depth, and channel widening. Figure B. 4 indicates
test:control ratios of numbers of brook trout before and after
bedload addition. Interestingly, although macroinvertebrate
density declined sharply after treatment, the decline in food
base was apparently offset by reduced fish abundance, hence no
effect on growth was noted.
2.0-
g
<
tT.
O
£ .-OH
• Spring Ralto
Foll Ratio
1988 1970 1972 874 1979 197 B 880 1902
YEAR
Figure B.4. From Alexander and Hansen (1986). Ratios of the
total number of brook trout present in treated (T) and control
(C) areas each spring and fall. Dashed line is for pretreatment
years and solid line for post-treatment.
Bj ornn et al. (1977) added sediment to a section of Knapp
Creek, a tributary of the Middle Fork of the Salmon River in
Idaho. Careful evaluation of the subsequent changes over a
period of several days, two weeks, and a year showed that total
fish density and density of Chinook salmon in summer decreased as
fine sediment decreased the amount of pool area (Figure B.5).
Decreases in the area of pool deeper than 0.30 m caused a linear
decrease in fish numbers in the pool. The authors cautioned that
because fish utilize only a small proportion of large pools, the
decrease in fish numbers with lost pool volume may not occur in
large pools.
121
/SECTION III
-------
l*t mrm tntmnt
mt IK I ton
fit
I 1 * Nil '»"
Itttlllt lltlllttl
Figure B.5. (From Bjornn et al. 1977). Fish densities in control
(unsedimented) and test (sedimented) sections of Knapp Creek
prior to sedimentation and at 1 day (1), 4 days (2) after first
addition of sediment and 3 days (3) and 13 days (post) after
second addition, and one year later (1975).
Correlational work in two additional streams revealed no
relationship between fish density and percentage of fine
sediments in riffles, as measured by core sampling. The riffles
examined were partially or fully embedded. Even in stream areas
with 66% fines smaller than 6.35 mm, no significant decrease in
fish density could be shown (Figure B.6). Because the relation-
ships shown between fish density and percentage of fines are
within-stream, seeding should not be a major concern. The best
correlations were with percentage of pool area having cover and
with drifting insect abundance. Lack of a clear effect of
addition of fines in Bjornn et al. (1977) may have been a result
of the small area involved in the experiment (291 m^ in 165 m of
stream).
122
/SECTION III
-------
1.0
9,9
1O 20 39 40 SO 6O 70
% SEDIMENT < 1.35 MM IN RIFFLES
10
1O 2O 30 40 SO 60 7O
% SIOIMf NT < 4.1SWM IN ilFFLIS
Figure B.6. (From Bjornn et al. 1977). Fish density in pools
versus percentage of sediment <6.35 mm in riffles in Bearskin
Cree)c (left graph) and Elk Creek (right graph), August 1975.
Saunders and Smith (1965) reported that low standing crops
of brook trout were associated with silting, and that the chief
effect on rearing habitat was destruction of hiding places. The
streams studied by these workers had higher standing crops of
trout after scouring removed the silt from the surface of the
substrate.
Shepard et al. (1984) found a significant positive relation-
ship (p<0.001) between bull trout abundance and substrate score
(the higher the score, the less fines were present and/or the
lower the degree of embeddedness). Score consisted of the sum of
the number (Table B.I) of the dominant particle size, subdominant
particle size, and embeddedness. These authors did not provide
information on stratifications that might have influenced their
results. Figure B.7 (Shepard et al, 1984) depicts the relation-
ship between substrate score and density of bull trout longer
than 75 mm.
123
/SECTION III
-------
Table B.I. (From Shepard et al. 1984). Substrate
characteristics and associated ranks for calculating substrate
score.
Rank
Characteristic
Particle atie clasa (range)
•» >
1 Silt and/or detritua
2 Sand (<2.0 BO)
3 Small gravel (2.0-6.4 mm)
4 Large gravel (6.5-64.0 mm)
5 Cobble (64.1-256.0 ma)
6 Boulder and bedrock (>296.0 mm)
Eabeddedneaa—
1 Completely enbedded (or nearly ao)
2 3/4 enbedded
3 1/2 eabedded
4 1/4 eabedded
5 Unembedded
I/ Extent to vhich doninant cited particle* arc
burled In land and ailt (aee Bjornn et al. 1977
for an illuetration).
tUMIIUTt ICOMt
Figure B.7. (From Shepard et al. 1984). Relationship between
juvenile bull trout density (fish at least 75 mm in length per
100 m2 of stream area) and substrate score for 26 stream reaches
in the Swan River drainage, Montana.
124 /SECTION III
-------
Shepard et al. provided no information on other habitat variables
or on seeding effects.
Konopacky (1984) examined a spectrum of natural pool/riffle
environments in central Idaho, obtaining data on riffle and pool
areas, pool volume, fish abundance, riffle sediment, discharge,
and invertebrate drift. He found that upstream portions of study
streams contained resident, relatively large salmonids, while
downstream reaches contained primarily age 0 chinook salmon. The
salmon were either products of spawning within the study areas or
had immigrated from downstream areas. More-upstream areas had
less fine sediments, hence substrate particles of greater mean
size.
Konopacky noted that fines, in riffles did not consistently
correlate with density of chinook salmon in the pools below the
riffles. Pool area and invertebrate drift at dusk accounted for
more variation. Riffle area, mean particle size in riffles, and
pool area accounted for more variation in biomass than did other
variables. Fine sediments in riffles influenced invertebrate
drift, which in turn influenced fish diet. Resident trout in
upstream reaches of study sites used cobble, boulders and vege-
tative cover within those areas during cold periods.
Edie (1975) reported a negative relationship between fines
smaller than 0.85 mm and the density of age 0 trout but not older
trout in the Clearwater River basin of Washington. His correl-
ations for age 0 trout also were significant and positive for
stream gradient, discharge, riffle area, percentage of the basin
cut, percentage of steelhead in the population, and abundance of
Cgttus rhotheus. He stated that the negative correlation for
percentage of fines was of importance to the land manager, who
presumably would want to avoid increases in fines so that
juvenile steelhead would be unaffected. However, it appears that
125 /SECTION III
-------
small juvenile steelhead tend to prefer somewhat higher gradients
where sediments are less likely to accumulate, and the negative
correlation with fines, while real, may in fact mislead. Edie
was unable to find a correlation of coho abundance with any
habitat variable.
Kelley and Dettman (1980) related density of age 0 steelhead
in late summer to embeddedness (Figure B.8). We examined the
report of Kelley and Dettman, and found that seven data points
had not been recorded in Figure B.8. Hence we plotted all 22
points in Figure B.9. Although the model in the latter figure
is not as "clean" as that in Figure B.8, we see a strong negative
relationship between embeddedness and juvenile steelhead
densities.
e
!„
e t.i
(t)
Figure B.7. (From Kelley and Dettman 1980). Age 0 steelhead
density in short sections of Lagunitas Creek, California. Dots
indicate riffles, squares indicate glides.
126
/SECTION III
-------
1.1-
1.0-
0.9-
_ a8"
H
i
| 0.5-
9 0.4-
a 0.3-
az-
ai-
01 0.2 0.3 0.4 os ae a?
EMBEOOEONESS IE)
Figure B.8. (Adapted from Kelley and Dettman 1980). Density of
steelhead in Lagunitas Creek, California, in relation to
embeddedness, all points plotted.
The Kelley and Dettman data are strengthened because they
were obtained in one stream that had been fully seeded in late
spring with hatchery steelhead to fill any carrying capacity not
occupied by naturally-spawned progeny. Kelley and Dettman did not
provide other data that might permit evaluation of relative
effects of embeddedness and other variables on fish density, but
the model in Figure B.8 is qualitatively convincing. However, we
do not consider it useful for quantitative prediction of effects
of embeddedness on steelhead in Idaho.
D. Burns and R. Thurow (unpublished data, South Fork Salmon
River) used sampling data from a range of stream gradients and
orders to develop a general relationship between embeddedness and
density of age 0 and age 1 Chinook salmon, steelhead and
cutthroat trout (Figure B.9). Highly variable seeding contrib-
uted to data scatter. Other physical and biological habitat
variables are confounded within the relationship. High fish
densities did not occur where embeddedness exceeded 35-40%, but
127
/SECTION III
-------
Figure B.9. (From unpublished data of R. Thurow and D. Burns).
Density of age 0 and 1 Chinook salmon, steelhead, and cutthroat
trout in granitic watersheds of the Idaho batholith in relation
to percent embeddedness.
did at embeddedness levels under 30-35%. Bull trout appeared to
flourish at low embeddedness as well (Figure B.10). Many streams
of different order, species suitability, and seeding levels were
Figure B.10. (from unpublished data of R. Thurow and D. Burns).
Density of bull trout in streams in granitic drainages in central
Idaho.
128
/SECTION III
-------
included in the relationships in the previous figures. Although
higher densities occurred in sites with lower embeddedness, we
cannot state with any confidence that embeddedness affected
density of fish.
When Thurow and Burns examined maximum combined . fish
densities (steelhead, cutthroat, all chinook salmon) found in any
single site at each embeddedness interval (10-19, 20-29, 30-39,
40-49, 50-59, and 60-69%), they calculated an inverse
relationship between embeddedness and fish density (Figure B.ll).
Their aim in this procedure was to reduce any effects of
inadequate seeding and other confounding variables. Lumping by
embeddedness strata does not eliminate problems of grouping
different habitats and stream orders. High embeddedness may co-
vary with other habitat features. Multi-variate regression
analysis within strata (selected to reduce variance caused by
such factors as gradient, riparian condition, etc.) may profit-
ably be used to extract the correlational effect of embeddedness
alone.
'"'"u T 'T!TrT''
±J±hmtHi:K
Figure B.ll. (From unpublished data of R. Thurow and D. Burns).
Relation between maximum fish density at any single site to
various embeddedness levels.
129
/SECTION III
-------
In order to avoid reliance on a single snorkel site, we re-
grouped data of Thurow and Burns in tabular form by embeddedness
intervals (Table B.2). The information for steelhead shows
Table B.2. (adapted from unpublished data of Thurow and Burns).
Mean and range in density of juvenile steelhead and age 0 and 1
Chinook salmon within embeddedness intervals. Numbers are in
fish/100 m2.
% No.
Embeddedness sites
10-19 25
20-29 19
30-39 9
40-49 10
50-59 1
60-69 2
No. juv. sthd.
mean (range)
3.25(0-14.46)
4.46(0-24.83)
2.89(0- 7.74)
4.45(0- 9.41)
4.46( )
0.37( )
No. juv. chin,
mean (range)
5.57(0-44.2)
5.05(0-40.0)
3.39(0-11.4)
0.44(0- 3.3)
0 ( )
little apparent effect of embeddedness level on fish density.
Densities of Chinook salmon tended to decline at higher
embeddedness levels.
Another difficulty with the "maximum density" approach in
Figure B.ll is that it confounds habitat features at different
sites (see locations noted beside each point in the figure) with
the effect of embeddedness. That is, the single data point for
low embeddedness was derived from Chamberlain Creek, and the data
points for the higher embeddedness levels were obtained from the
South Fork Salmon River.
Edwards and Burns (1986) correlated stream gradient with
embeddedness, but obtained an r2 of only 0.19 (not significant, p
= 0.07). We believe the stratification procedure used by Burns
and Edwards (1985) reduces the possibility of securing a signifi-
cant negative slope for the regression of embeddedness on grad-
ient by selecting a limited range of velocities, depths, and
particle sizes for embeddedness determinations. We believe that
the probability of p = 0.07 may reflect a real relationship
130 /SECTION III
-------
between gradient and "embeddedness" in the wider sense of the
term; that is, in a stream at large rather than in the strata
used by Edwards and Burns.
Certainly fines accumulation is a function of stream
gradient. Figure B.12, from Platts (1974b) illustrates this
point. The percentage of fines in Figure B.12 declined sharply
until a stream gradient of about 4-5% was reached, then declined
more gradually, according to Platts1 figure.
Figure B.12. (From Platts 1974b).
boulders to stream gradient.
Relationship of fines and
Munther and Frank (1986 a,b,c) reported r2 values for linear
relationships between embeddedness and fish populations in
several streams in Montana (Deerlodge, Lolo, and Bitterroot
national forests). In 53 such regressions, coefficients of
determination were usually termed "extremely low" by the authors.
One r2 exceeded 0.70, but only 3 observations were included and
that test was considered by the authors to be invalid.
Konopacky et al. (1985) provided data on snorkel counts of
abundance of age 0 Chinook salmon and pool embeddedness. The
methods section of the report does not indicate methods used to
131 /SECTION III
-------
measure embeddedness. Personal communication with Konopacky and
Shoshone-Bannock tribal biologists indicates embeddedness was
estimated visually in pools rather than with the accurate and
detailed measurement techniques of Burns and Edwards (1985).
Their data do not support a relationship between embeddedness and
chinook salmon density in Yankee Fork and Bear Valley creeks,
both Salmon River tributaries, but must be evaluated in light of
the visual estimation procedure that they used.
C. Johnson (Cottonwood District, Bureau of Land Management)
provided unpublished data on fish densities in his district in
relation to embeddedness. His embeddedness values were obtained
with the methods of Burns and Edwards (1985). Density of age 0
and 1 rainbow and cutthroat trout was negatively related to
embeddedness level (Figure B.13), but with r2 of only 0.07.
Summer standing crop, in kg/ha, correlated negatively with
embeddedness level at r2 = 0.27 (Figure B.14). Another data set,
as
^0.4
|
>as
CO
go-2
a
0.1
01 0.2 0.3 0.4 0.5 0.6
% MEAN EMBEDDEDNESS
Figure B.13. (From data of C. Johnson, BLM). Density of age 1+
rainbow and cutthroat trout in stream sections in the Cottonwood
District, BLM.
132
/SECTION III
-------
90
80
70
3
r
940
§
I30
*10
n
, r- u.£t
*
*
\
\
\ »
*
*
\ *
m \
\
\t
* \
*
» \^
/
•
" »\
10 20 30 40 50 60
X MEAN EMBEDDEONESS
Figure B.14. (From data of C. Johnson, Cottonwood District, ELM).
Summer standing crop in kg/ha of rainbow and cutthroat trout of
age 0 and 1-*-.
presumably for different study sites, showed a weak positive
correlation (r2 = 0.06) between age 0 rainbow and cutthroat trout
and embeddedness level (Figure B.15). Another data set, for
streams with an overall embeddedness > 35% and/or substrate with
more than 30% fines < 6.3 mm, had a negative correlation of fish
density with mean embeddedness level (r2 = 0.20)(Figure B.16).
0.9
0.8
0.7
0.8
O.S
0.4
0.3
0.2
0.1
a
•
r*. 0.064
-
.•-•''.'
",.*''
' • • : • '
O 0.1 0.2 O3 04 05
X MEAN EMBEDOEDNESS
Figure B.15. (From data of C. Johnson, Cottonwood District, BLM).
Density of age 0 rainbow and cutthroat in relation to
embeddedness in stream sites in the Cottonwood District.
133
/SECTION III
-------
c*
\
1
I
to
u.
as
0.4
0.3
0.5
0.1
0
•
r2-0.20
•
""""•..
***** •
"*••-, •
• *•»
* " "*"**•
•
o.i 0.2 as 0.4 0.5
% MEAN EMBEDDEDNESS
Figure B.16. (From data of c. Johnson, Cottonwood District, BLM) .
Density of age 1+ rainbow and cutthroat trout in relation to
embeddedness in streams with mean embeddedness in excess of 35%
and/or substrate with > 30% fines in the < 6.3 mm category.
Gamblin (1986) provided information on cutthroat trout
density in 21 streams in the Coeur d'Alene River and Hayden Creek
drainages in relation to embeddedness. He used the embeddedness
measurement system of Burns and Edwards (1985). He found no
significant relationship (p=0.05) between embeddedness and trout
densities. The plot of the data (r2 = 0.12) indicated a positive
relationship to about 40% embeddedness (Figure B.17).
Although the data on fish density and embeddedness provide
only weak coefficients of determination in all cases, the weight
of the evidence tends to indicate that areas with high embed-
dedness (50% embeddedness may define "high") tend to have lower
134
/SECTION III
-------
i
£
rt
10
as 30 3*
Figure B.17. (From Gamblin 1986). Relationship between age 1+
cutthroat trout density and embeddedness level in streams in the
Coeur d'Alene River and Hayden Creek drainages.
densities of salmonids. However, the low r2 values suggest to us
that other factors may be more important than embeddedness.
Embeddedness should co-vary with morphology of the stream
channel, and the latter may have greater influence on fish
rearing densities than embeddedness level.
Konopacky et al. (1985) assessed fish density and substrate
fines as part of a feasibility study for habitat improvement.
Figure B.18 shows the relationship between percentage of fines
smaller than 4 mm and age 0 Chinook salmon per 100 m2 for two
streams in August, 1984. Fines were evaluated in riffles at 25
equidistant points in each of three riffle cross-sections by eye.
Fish counts were by snorkel observations. The data, for Bear
Valley Creek and Yankee Fork, tributaries within the Salmon River
drainage, do not support a negative relationship between riffle
fines and fish density in pools. Again, seeding, cover, and
stream order are confounded in the data, although interstream
variability was reduced because the authors obtained their data
in several strata within the same stream. Figure B.19 shows the
135 /SECTION III
-------
35-
30-
25
1=
2 **?o
§1
X 15
81
If to
X 91
o
V
e
1O 20 3O 40 SO 6O
% RNES S 4 mm
Figure B.18. (Extracted from Konopacky et al. 1985).
Relationship between visually-assessed percentage of fines
smaller than 4 mm and density of 0-age chinook salmon.
relationship between percent riffle fines and total density in
pools of all chinook salmon, steelhead, and cutthroat trout.
This figure also does not support a negative relationship between
fines and fish density, up to 40% fines.
ra^-
* E
35-
30-
25
15
10 20 30 40
% FINES S 4 mm
50
Figure B.19. (Adapted from Konopacky et al. 1985). Relationship
between fines smaller than 4 mm and total density of chinook
salmon, steelhead, and cutthroat trout.
136
/SECTION III
-------
Hillman et al. (1986) investigated microhabitat selected by
juvenile Chinook salmon and steelhead in Red River, a Clearwater
River tributary. They found most Chinook salmon of age 0 over
sand-gravel substrate in summer. In this fines-embedded stream,
they found chinook salmon densities greater than 60 fish per 100
m2 in habitat with water velocities less than 20 cm/s, depths
from 20-80 cm, and with associated cover in the form of undercut
banks.
Petrosky and Holubetz (unpublished report, 1986) plotted
snorkel-observed densities of age 0 chinook salmon in many Idaho
streams in relation to visually-assessed percentage of fines.
Their figures (figures B.20 and B.21) seem to support a negative
K - :
19
100
Figure B.20. (From Petrosky and Holubetz 1986). Density of age 0
chinook salmon in relation to visually-assessed percentage of
fines <5.0 mm in Marsh Creek, Bear Valley/Elk creeks, upper
Salmon River, and Valley Creek, all central Idaho streams.
relationship between fines* and chinook salmon density. [In
examining the figures, the reader should ignore the medians.]
However, the data in figures B.20-B.21 include data points from a
wide spectrum of habitat conditions. Because we would expect
fines to vary inversely with increasing gradient, we stratified
137
/SECTION III
-------
40
O
o
99-
cro-w AWJ KtiArr
• tin cnetK Aim tnmw/mms
BO
00
100
Figure B.21. (From Petrosky and Holubetz 1986). Density of age 0
Chinook salmon in relation to visually-assessed percentage of
fines <5.0 nun in Marsh, Knapp, and Elk creeks in central Idaho.
source tabulations (obtained from C. Petrosky, personal communi-
cation) by gradients of 0-1%, 1-2%, and over 2%, plotted densi-
ties against fines in these strata, then calculated regressions
and correlations. The resulting relationships in Figures B.22-
B.24 provide little support for a generalization about effects
of fines on rearing densities of Chinook salmon. Coefficients of
determination (r2) were only 0.21, 0.084, and 0.002 for the
respective gradient strata, although they decreased with gradient
as one might expect. Only the regression for gradients of 1-2%
was statistically significant (F = 8.45, p =0.05). The fact that
r2 values were extremely low leads us to the tentative conclusion
that other habitat features were much more important than fines.
The combined r2 value for the data for all gradients combined is
only 0.058. Another important source of data scatter in the
foregoing three figures may be seeding, but if we take the data
138
/SECTION III
-------
> . in
1.1.i to
m urn Hi •«••• • rmhu
Figure B.22. (plotted from data of Petrosky and Holubetz 1986).
Relationship between snorkel-assessed densities of age 0 chinook
salmon and visually-determined percentage of fines in stream
gradients of 0-1%.
90-1
49-
40-
30"
3S-
w-
B-
r*« 0.084
r' MS
* I« 1 t 99
10 10 30 40 SO «0 70 80 »O WO
% LARGE 10.8-S.O MM1 + FINE t-O.S MMI SEDIMENT
Figure B.23. (plotted from data of Petrosky and Holubetz 1986).
Relationship between snorkel-assessed densities of age 0 chinook
salmon and visually-determined percentage of fines in stream
gradients of 1-2%.
139
/SECTION III
-------
t' i o.90»
r « «.1
««• t
"* 1>~*~' n —w
i IMKK »••»«"• • "« '•«••
w M no
Figure B.24. (plotted from data of Petrosky and Holubetz 1986).
Relationship between snorkel-assessed densities of age 0 chinook
salmon and visually-determined percentage of fines in stream
gradents of over 2%.
at face value, fines "explained" less than 6% of the variability
in fish density.
For Valley Creek and tributaries, Petrosky and Holubetz
(1986) collected information on brook trout density as well as
chinook salmon abundance, together with data on percentage of
fines. We prepared a correlation of brook trout density in 32
study sites in Valley Creek with fines smaller than 5 mm (Figure
B.25). The regression, significant at p = 0.01 (F = 8.57, r2 =
0.22) indicates that brook trout density in Valley Creek was
higher in the presence of more fines. A similar treatment for
the upper Salmon River (Figure B.26) shows the same significant
relationship (F = 41.62, r2 = 0.64).
140
/SECTION III
-------
13-
_ a
M
I
K 4
Z
F -S-S9B
1 130 dt.
10 2O 30 40 SO 60 70
\ LARGE (O.8-50MMI » FIME t«O.8 MM) SEDIMENT
9O
10O
Figure B.25. (From Petrosky and Holubetz 1986). Density of age
1-f brook trout in relation to fines < 5.0 mm in Valley Creek and
tributaries.
i
^.0-64
F =41.62
1 8.23 d.f.
M
•
O>
3
I- 4
3
en
o
O
cc
CD
10 20 ' 30 40 50 60 70
% LARGE (0.8-5.0 MM> + FINE(
-------
The available field data for the northern Rockies have
provided information that indicates that sediment contributes to
the variability in densities of chinook salmon juveniles. The
data do not permit quantitative assessment of effects. Field
correlations have sometimes shown a relationship between
abundance of other salmonid species and embeddedness and fines.
In the case of brook trout, fish densities appear higher
where more fines are present. This finding is at apparent odds
with the information of Alexander and Hansen (1986), but may be
explainable partly on the basis of stream morphology. Sand in
Hunt Creek in Michigan increased bedload by 4-5 times and
dramatically altered stream morphology for the entire one-mile
test section. Sand measurements reported by Petrosky and
Holubetz (1986) pertain to short stream habitat units.
Laboratory experiments and confounded field data offer some
indication that embeddedness and' fines affect fish rearing
densities in the northern Rockies. Experiments in field sites in
which full seeding is assured and a number of habitat variables
are examined would be needed before more can be said about the
relative and quantitative role of sediments in affecting rearing
densities of juvenile chinook salmon and other species. As noted
earlier, stream morphology may outweigh embeddedness and levels
of fines as legislators of fish rearing density.
Any generalizations about utility of sedimented substrata
for fish rearing habitat require bounds. Depositional streams
such as Idaho's Silver Creek often lack gravel or rubble in many
productive rearing areas, and would rank low in substrate
scoring. It is important to note that stream stability and
nutrient availability in spring-fed streams lead to abundant
growth of macrophytes, with associated populations of amphipods
and other macroinvertebrates. The thrust of this review of
142 /SECTION III
-------
rearing conditions is not directed at these relatively scarce (in
the northern Rockies) environments, although winter hiding areas
may remain important in spring-fed streams when macrophytes die
off and hence do not offer cover. Hunt (1969) and White (1972)
showed clearly that provision of winter cover could substantially
increase carryover of juveniles and adults and increase average
density of salmonids in spring-fed streams.
It is important to state that habitat utilization by parr
and fingerlings differs on a diel basis. Daytime habitat
preference places fish near the path taken by drift food and
often places fish over less-sedimented substrata. At night these
fish tend to move inshore and settle in the shallows on the
bottom (Hartman 1963, Edmundson et al. 1968, Campbell and Neuner
1985, Hillman et al. 1986). Species differ in their ability to
utilize low light intensities for feeding, but in general,
although feeding may become intense at dusk, quiescence inshore
appears to be the norm for salmonids of the northern Rockies at
night in streams.
B.3. Adults
Adult resident salmonids utilize microhabitat features that
tend to have less-sedimented substrata because they use focal
points in areas subject to more stream energy. They utilize
deeper, faster areas for feeding stations, although attempting to
obtain food with maximum gain in relation to energy expended
(Puckett and Dill 1985, Bachman 1984) by using focal points in
low velocities adjacent to faster water (Campbell and Neuner
1985). Areas close to incoming food streams would tend to have
less sediment than more-inshore zones.
Adult salmonids have been noted at night in relatively
shallow and quiescent water, resting on the substrate (Campbell
and Neuner 1985). These locations are different from daytime
focal points. In high-gradient streams they tend to lie in
143 /SECTION III
-------
pockets of low velocity that are sand-silt depositional areas.
Fish that remain in cover such as rubble during the day may leave
it even at low temperatures at night and move to the stream edges
(Campbell and Neuner 1985 and unpublished data of, J. Griffith,
Idaho State University). The proportion of the population that
so behaves is unknown. Movement out of the rubble and to the
stream edges at night during very low temperatures in winter is
mystifying as to adaptive significance. Micro-differences in
water temperature may be involved, as the fish seek warmer water
at a time when predation should not occur.
Adults require deeper water of suitable velocity with
adequate cover than do juveniles. Hence loss of pool volume to
sediment deposition (eg. Bjornn et al. 1977) (or to cross-
sectional breakdown because of livestock grazing or lost large
woody debris, although this topic is beyond the current review)
should logically reduce suitability of a stream for adult
salmonids. Examination of Hunt (1969) may be instructive. His
worJc showed that increases in average depth and pool volumes,
with concomitant reduction in stream surface area and quantity of
fines in the substrate, substantially benefited adult salmonids
in Lawrence Creek, Wisconsin.
The excellent research of Alexander and Hansen (1983, 1986)
demonstrates that fine sediments can limit fish populations. But
their results cannot serve as quantitative models for the north-
ern Rockies. The generally low coefficients of determination
found in field correlations of fish densities and fines or
embeddedness levels in the northern Rockies suggest to us that
other factors have strong influences on fish densities. These
factors will not be ascertained by single-factor evaluations of
sediment effects, nor will the quantitative influence of
sediments be determined without holistic ecological evaluations.
144 /SECTION III
-------
C. INSECT ABUNDANCE
C.I. insect density
McClelland (1972} added sediment to 0'Kara Creek, a Selway
River tributary in the Idaho Batholith, to examine effects of the
addition on aquatic macroinvertebrates. Insect abundance and
diversity generally declined as a result of sediment injection.
Pteronarcyjs califomica and Arctopsvche err and is were highly
sensitive to bottom sediment. Isogenus sp., Rhithrogena
robusta. Arcvnoptervx sp. , Acroneuria pacifica, Ephemere11a
grandis. and Rhyacophila acropedes were moderately sensitive to
small and medium amounts of sediment, and highly sensitive to
large amounts. McClelland found that the microhabitat area
beneath cobble was very important for most of the species
studied. Under-cobble areas sealed by fines could not be used by
insects.
Brusven and Prather (1974) investigated laboratory and field
preferences of aquatic insects for various substrate types.
Substrates with cobble were generally preferred over those
without cobble by 5 species of stream insects that represented
Ephemeroptera (Ephemeralla grandis), Plecoptera fPteronarcvs
californica), Trichoptera (Brachycentrus sp. and Arctopsyche
grandis), and Diptera (Atherix variegata). Substrata with
unembedded cobble were slightly preferred over half-embedded
cobble, and completely embedded cobble in fine sand proved
unacceptable to most of the species. Embeddedness affected
Brachycentrus sp. and Atherix variegata less than was the case
for other species, because of autecological requirements.
Reiser and Bjornn (1979) summarized literature that reported
the highest production of aquatic macroinvertebrates in streams
with gravel and rubble-size substrate. Pennak and Van Gerpen
(1947) found, as have other workers, that numbers of benthic
145 /SECTION III
-------
invertebrates decreased in the progression rubble to gravel to
sand. Nuttall (1972) found that sand deposition tended to lead
to increased abundance of only a few forms, to the detriment of
many species. Tubificids increased in sandsr as did two types of
mayflies. Table C.I lists reactions of various species to sand,
according to Nuttall.
Table C.I. (From Nuttall 1972). The reaction of various species
of invertebrates to sand deposition in the River Camel in Great
Britain.
1. Species immediately eliminated by sand deposition.
Leuctra niger Baetis pumilus
L. hippopus Polycentropus kincri
L.._ geniculata Gammarus pulex
Amphinemura sulcicollis Polycelis felina
Ephemera danica Sericostoma personatum
2. Species showing an increase in numbers with sand deposition.
Tubificidae Rhithrogena semicolorata
Baetis rhodani
3. Species not immediately affected by sand deposition.
Leuctra fusea Gyrinus sp.
Caenis rivulorum Naididae
Protonemoura meyeri Ancylastrum fluviat.ile
Hvdropsyche instabilis
Nuttall attributed most of the decline in those species
negatively affected by sand to the shifting substrate rather than
to abrasion. He also noted that rubble supports more animals
than sand does, a phenomenon correlated with the amount of
available living space and with the greater lodgement of organic
materials among stones, which provides food for
macroinvertebrates.
Cummins and Lauff (1969) found that substrate particle size
was the main factor involved in raicrohabitat selection for four
species of macroinvertebrates but of lesser importance for 6
others, among 10 species studied.
Alexander and Hansen (1986) reported reductions in benthic
macroinvertebrates as a result of increases in sandy bedload in
146 /SECTION III
-------
treatment areas of Hunt Creek in Michigan. They compared
treatment populations with those in control areas to which no
sand had been added.
While sand is not good insect habitat, clay is better and
muddy sand still better (Behning 1924), the latter holding 20 to
40 times more animals per m2 than clean sand. It seems likely
that muddy sand would not persist in a substrate that shifts in
response to currents, and that Nuttall's comments about
instability being responsible for reduced insect abundance are
reasonable. Hynes (1970) recalculated data for stream samples
collected by another worker, and showed that shifting sand
yielded the lowest number of organisms of any substrate sampled
(Table C.2),
Table C.2 demonstrates extremely high densities of
invertebrates in macrophytes. The plants colonize sediment-laden
areas, trap more sediment, and often support higher densities
than nearby gravels. These "islands" of macrophytes and
entrapped sediments cause a form of channel "braiding" that can
increase local velocities and cleanse gravels.
Table C.2. (From Hynes 1970). Mean number of animals per m2 on
various substrata at 3 stations in Doe Run, Kentucky, based on 12
months of sampling.
Station number I II III
Miles from source 0.1 1.9 3.1
Fissidens beds 87,600 102,000
Nasturtium beds 9,000
Myriophyllum beds 34,700
Myosotis beds 9.960
Bare riffles 2,700 7,550
Travertine riffles 1,530
Rubble riffles 1,540
Silty to sandy pools 4,640 5,830
Shifting sand 1,860
147 /SECTION III
-------
Hynes showed that rubble riffles had a fauna composed mostly
of Ephemeroptera (70%), while mayflies were absent from shifting
sand. Mayflies are known to drift more than any other macro-
invertebrate order, with the occasional exception of chironomids.
Thus, the effect of fijnes on mayflies may be expected to have a
consequent effect on abundance of drift. The shifting sand
substrate had a preponderance of isopods, amphipods, caddis, and
snails.
Number of organisms is only one component of the ecological
story, of course. "Turnover" of animals is important in
contributing to the drift food source used by most salmonids in
streams. Turnover time (permanence of station along the
upstream-downstream axis) and contribution to drift could provide
a different pattern than standing crop information. One suspects
that turnover is more rapid in some forms of mayflies and
dipterans than in isopods, amphipods, and gastropods. Dipterans
in sand tend to consist of burrowing forms not available to fish.
Where fines reduce suitability of the benthic zone beneath
the surface (insect habitat in interstices and beneath rocks),
indirect effects on fish production may occur. High embeddedness
levels, for example, may potentially limit insect production to
the upper surfaces of gravel or rubble particles.
The point of these comments is that the low standing crop of
organisms on rubble riffles as reported by Hynes (1970) is
probably not indicative of the relative importance of such
riffles in relation to a shifting sand substrate as sources of
drift food for fish.
Tebo (1955) found that logging-related siltation sharply
reduced the abundance of macroinvertebrates in a North Carolina
stream. The incremental silt consisted of a layer of "sterile
sand" and micaceous material that accumulated in some areas to a
148 /SECTION III
-------
depth of 25 cm. When the silted substrate was cleansed by flood
flows, it again supported insects at the same rate as control
sites not affected by siltation from logging.
Martin (1976) noted that silt in the Clearwater River basin
in Washington had a short residence time, and was moved by
freshets, which can occur even in summer. He suggested that when
sediment was present, "poor substrate habitat" reduced bottom
fauna populations, but his data did not reveal a consistent
significant effect of sedimentation.
Crevice shelter in the substrate is a first requirement for
invertebrates in flowing water (Hynes 1961}(note: the siinuliids,
net-spinners, and some mayflies that rely on camouflage for
protection are exceptions) . Hynes noted that as an insect grows
it needs a larger crevice, hence moves through a progression of
successively larger crevices as it grows (this is a pattern
similar to the movement into faster, deeper water by fish as they
grow, although related more obviously to cover in the case of
insects), Cederholm and Lestelle (1974) suggested that loss of
intragravel living space causes the inverse relationship between
percentage of fines (< 0.84 mm) and insect density in substrate
samples.
Diversity also decreases in fines (Williams and Mundie 1978,
Chutter 1969). Ephemeropterans and plecopterans do not prosper
where fines predominate. Erman and Mahoney (1983) studied
streams 6-10 years after logging in northern California with and
without narrow buffers. Unbuffered streams showed considerable
but incomplete recovery of diversity of macroinvertebrates, The
mean diversity of unbuffered logged streams was about 9% lower in
1980-81, compared to 25% lower after an initial post-logging
study in 1975. On the other hand, narrow-buffered streams
changed little between post-logging and about 6 years later,
relative to controls. Mean diversity was about 12% lower than in
149 /SECTION III
-------
controls in 1980 as compared to 12% in 1975. The logged streams
had significantly more fine seiment in the top substrate layers
than did comparable control streams.
Erman and Erraan (1984) tested effects of median particle
size an heterogeneity on rates at which Ephemeroptera and,
specifically, Paraleptophlebia memorialis. colonized cleaned and
sorted gravels in trays. More mayflies colonized median particle
sizes of 32 mm and 8 mm than 2 mm (Figure C.I).
I to
I 11
»H tmml
Figure C.I. (From Erman and Erman 1984). Colonization of
cleaned, graded substrate trays in a stream by mayflies in
relation to partial size. Letters in graph refer to hetero-
geneity tests not relative to the particle size:colonization
point made in the present text.
NCASI (1984c) studied insect density and biomass in
artificial channels to which fines were added. Biomasses
declined as fines were added. The study suggested that up to 20%
fines would not negatively influence biomass in communities
already heavily colonized by chironomids, amphipods, and
gastropods. Where these sediment-tolerant groups are not well-
represented, NCASI indicated that biomass should be expected to
decline with higher proportions of fines.
150
/SECTION III
-------
Konopacky (1984) sampled benthic macroinvertebrates over
three months in stream troughs supplied with sand-pebble or
coarse gravel substrata. The number of organisms was greater in
the sand-pebble mix, but biomasses in the two substrata were
similar (Figure C.2). Mean weight of individual insects was
greater in the coarse gravel mix.
1Kb ®
iflfc
Figure C.2. (From Konopacky 1984). Numbers (top graph) and
biomass (bottom graph) of macroinvertebrates over 3 months in
sand-pebble {white boxes) and coarse gravel (black boxes)
substrata in stream channels supplied with water from a nearby
stream.
Bjornn et al. (1977) permitted insects to colonize
experimental channels at Hayden Creek for 15 days, then examined
standing crops of insects in locations with 4 levels of
embeddedness with and without adding sediment to the channels
in addition to that represented in the embeddedness indices).
Table C.3 (from Bjornn et al. 1977) summarizes the results.
151
/SECTION III
-------
4
2
1
0.8
o
>
s
2
O
5
30
10
13
9
a
4
• HI llm
SAND- COARSE
PEBBLEQRAVEt
31
JUL
IS
AUQ
31
AUQ
13
SEP
27
3EP
11
OCT
ALL INVERTEBRATES
CH1RONOMIOAE
EPHEMEROPTERA
TRICHOPTERA
SIMULItDAE
OTHER INVERTEBRATES
Figure C.3. (From Konopacky 1984). Biomass/individual in two
substrate mixes over a 3- month period in stream channels
supplied with water from a nearby stream.
Table C.3. (From Bjornn et al. 1977). Number of benthic insects
per 0.093 m2 sample in riffles without and with sediment, at 4
embeddedness conditions.
Level of cobble
embeddedness
0
1/3
2/3
Full
Mean insect density
Without sediment With sediment
21.2 28.9
53.8 62.9
72.1 84.8
64.9 25.1
Only at full embeddedness did insect density decline from the
highest density (which occurred at 2/3 embeddedness). The
152 /SECTION III
-------
authors monitored several species for specific reactions to
embeddedness (Epeorus albertae. Cinygroula sp., Ephemerella
tibialis. Baetis bicaudatus. Simulium sp.). Only insects of the
family Chironomidae did not decline significantly (p = 0.05) at
high embeddedness. Riffles with sediment added supported
slightly larger insect populations than did the control riffles,
a finding attributed by the authors to the slight to moderate
layer of sediment around cobbles, a condition they considered
more natural and often more favorable than the unnaturally clean
cobbles in riffles without sediment. E. albertae was partic-
ularly intolerant to increased sedimentation, especially at full
embeddedness, as were B. bicaudatus and Simulium sp. Table C.4
(Bjornn et al. (1977) summarized reactions of taxonomic groups to
embeddedness and sediment addition.
Table C.4. (From Bjornn et al. 1977). Mean insect densities
(number/0.093 m2) in test (sediment added) and control (no added
sediment) riffles at experimental channels at Hayden Creek, 1974.
tn
ToMhMctl Cowral 211 !3I 71.1 «4 »
tat !*.» «J M.I JJ I'
Tnfi* al ipccfc) *<«• Mow CoHM 100 13,1 130 It.O
Tot 90 11.1 111 17-0
OmiAffM* Central 11.5 4JO JJJ Ml
T«t «.l ».** 14.0* !.!'
Contml 1,1 7.( J.t • »
TMI 1J t.l tO.4 0.)*
CoMral 01 M 49 <»
Trt O.I t.l' l.t J.I
trail Kk-rfmt Cowfot II 110 33-J 101
tat IT.O- ».7 T».I H.3*
StnmH*m *. (tew) Cortmt 40.4 »4 «.0 *4 I
Tnl 44.7 11*0 1M.V 3 »t
m tp. (p>r«l Conln>» l.t 3J4 233 114
T«t 1.4 W.7 11.4 1.7*
Control IK 43.» J)4
TcK )».!• 4».l )1J
• > i%«Ukin< dVrtTCM* it JX Itrrl
t * ilf ntflanl rfYfertne* it S% ttrrtt Tot nttiril l«f t
Bjornn et al. (1977)' extended insect density sampling to
natural stream areas in tributaries of the Middle Fork of the
Salmon River. In Elk Creek, which contained large amounts of
fines in riffles, the authors used adjacent manually cleaned and
uncleaned plots in riffles to check laboratory results. The
153 /SECTION III
-------
total number of insects per 0.093 m2 on cleaned plots was 1.5
times the number collected from uncleaned plots after 45 days of
recolonization. There were about 4 times more mayflies and 8
times more Alloperla sp. (the dominant stonefly) on cleaned plots
than on uncleaned plots. The effects of sediment cleaning can be
seen in Figures C.4 and C.5, from Bjornn et al. (1977).
411
lit
M XI
•
m
** IM
0
X. It
*•
I! I
m «•
_ II
A
s i
* n rt'::
i_l_J:
C "I
II-
X ^ II-
1
ft 1 -
1 1 ttl
1
fl fl
I 1
D
1 1
1
!
1
I
Kits
Figure C.4. (From Bjornn et al. 1977). Density of benthic
insects in cleaned (test) and uncleaned (control) sections in Elk
Creek riffles. A = total insects? B = number of species; C =
Amel.et.us sparsatus; D = Paraleptophlebia heteronea.
_cti
Figure C.5. (From Bjornn et al. 1977). Density of benthic
insects in cleaned (test) and uncleaned (control) sections in Elk
Creek riffles. A = Ephemeroptera; B = Rhithroqena robusta; C =
Alloperla sp.; D = Ephemerella inermis-infreauens complex.
154
/SECTION III
-------
Bjornn et al. (1977) summarized their results by stating
that several substrate parameters, namely (1) predominant
substrate, (2) level of cobble embeddedness, and (3) size of
sediment surrounding cobble can be compared with criteria for
benthic insect habitat in Idaho batholith streams. Less insects
lived where cobble embeddedness by fine sand (< 6.35 mm) exceeded
2/3, and where large amounts of sand were present. On the other
hand, embeddedness caused by a heterogeneous mix of sediment
around pebbles and small cobbles provided good habitat for
insects. Completely clean cobbles apparently do not provide as
suitable an environment for benthic insects as partially embedded
cobbles.
Bachman (1958) reported a significant reduction in standing
crop of aquatic insects in a northern Idaho stream after
sedimentation had increased as a result of logging, and Tebo
(1955) reported similar results for a small stream in North
Carolina. Cederholm and Lestelle (1974) found highly significant
negative correlations between the mean percentage of fines
smaller than 0.84 mm and the mean number of benthic insects per
0.10 m2 sample (Figure C.6.)
1
150
100
50
y«-33.1* * 394.5
r--.85
»»-22.5x+280.5
r»-.95
I
1
60 7.0 8.0 90 10.0 11.0 12.0
June 12-24
6.0 70 8.0 9.0 10.0 11.0 (2.0
July 26-Aug6
Meon Percent Fines <0.84I mm
Figure C.6. (From Cederholm and Lestelle 1974). Mean numbers of
bottom organisms correlated with mean percentages of fines
smaller than 0.841 mm in diameter.
155
/SECTION III
-------
Munther and Frank (1986 a,b,c) reported extremely poor
correlations between embeddedness or free-matrix particles and
various macroinvertebrate habitat utilization indices. Table C.5
lists invertebrate niche characteristics and r2 values, none of
which exceeded 0.32, and all of which were non-significant.
Table C.5. (From Munther and Frank (1986 a,b,c). Linear
relationships between embeddedness, free matrix particles and
invertebrate habitat utilization indices for 20 sample locations
in the Deerlodge, Lolo, and Bitterroot national forests.
particle H^asurgnqpt Invertebrate Measurement
Embeddedness J
Embeddedness 1
Embeddedness 1
Embeddedness I
Embeddedness 1
Efflbeddedness 1
Embeddedness J
Embeddedness 1
Embeddedness %
Free Hatrlx >
Free Hatrlx J
Free Matrix %
Free Katrlx 1
Free Matrix J
Free Hatrlx S
Free Hatrlx %
Free Matrix %
Free Matrix t
Taxa are Sedlrent Tolerant .05
Ccmunity Numbers are S«d. Tolerant .21
Biomass Is Sedlrent Tolerant .01
Taxa require Interstlcial Spaces .00
Coon. Numbers require Inters tic. Space.07
aiomaas retire Interatlclal Spaces .08
Taxa use Top of Substrate .05
Community NUnbers use Top of Substrate.It
Blomass use Top of Substrate .32
Taxa are Sediment Tolerant .06
Community Winters are Sed. Tolerant .22
Blomasa Is Sediment Tolerant .04
Taxa Requiring Interstical Spaces .00
Com. Nbnfcers Req. Interstle. Space .02
Blomsa Requiring Interstlcial Space .07
Taxa Using Top of Substrate .17
Commit? Hbmbers use Top of Substrate.21
Biomass Using Top of Substrate .22
Munther and Frank (1986 a,b,c) attempted to relate embed-
dedness and free matrix particles to macroinvertebrate habitat
groupings (forms that use the substrate surface, interstitial
users, and burrowers, sediment-tolerant groups). Munther and
Frank felt that one possible explanation for failure of abundance
of species groupings to correlate with embeddedness could have
been that some forms are neither discrete nor obligate residents
of the habitat zones to which they were assigned. The approach
of these authors has a logical basis in that one would expect
more sediment-tolerant forms to utilize embedded habitat
T
differently from less tolerant animals, and that residents that
depend on the areas beneath the gravel surface should suffer
declines in embedded zones.
156 /SECTION III
-------
In light of the results of Spaulding (1986), in which a
strongly significant negative effect was shown for embeddedness
on insect density and yet r2 equaled only 0.13, some of the
coefficients of determination found by Munther and Frank may
require more attention. Three dependent variables had r2 values
greater than 0.13; percent of community numbers that are sediment
tolerant (r2 = 0.21), percent of community numbers that use the
top of the substrate (r2 = 0.14), and percent of the biomass that
uses the top of the substrate (r2 = 0.32).
Munther and Frank had four dependent variables that had r2
values greater than 0.13 when related to free matrix particles;
percent of community numbers that are sediment tolerant (r2 =
0.22), percent of taxa using top of the substrate (r2 = 0.17,
percent of community numbers and biomass using top of the
substrate (r2 =* 0.24 and r2 = 0.22, respectively). Of course,
values for r2 do not demonstrate regression slope. Munther would
have to reanalyze data so that the regressions could be
evaluated.
Spaulding (1986) extensively evaluated two sections of Big
Creek, Utah. One section in a rest-rotation grazing allotment
was compared with another that had been rested from grazing for
four years. He extracted 32 Surber samples on transects in the
treatment (grazed) and control (rested) sections. Mean abundance
of macro invertebrates was 99.6 on the ungrazed and 73.7 on the
grazed sections (significant at p = 0.05). The greatest
differences occurred among the collector and shredder groups,
with the former more abundant in the grazed site and shredders
more plentiful in the ungrazed section.
Spaulding (1986) regressed insect abundance against percent
surface embeddedness (measured with U.S. Forest Service methods)
and found a significant negative effect of embeddedness on
density (F = 11.6, 1 and 62 d.f., p = 0.05, r2 = 0.13). Although
157 /SECTION III
-------
the degree of variability explained by the data was not high,
insect abundance was reduced by an absolute 21% where
embeddedness was 3/4 to full in comparison to where embeddedness
was 1/2 to 3/4.
C.2. Insect drift
Bjornn et al. (1977) examined effects of sediment on
quantity of drifting insects in artificial stream channels.
Drift at sunset was not related to embeddedness in the channels
(Figure C.7), or to sediment addition in given levels of
embeddedness, according to the authors. However, they did not
try to quantify the immediate effect of sediment introduction on
insect drift. . Q
1 ,
1 '
CHUfMCIl
m— » 7 r i
*« H TH I T
Figure C.7. (From Bjornn et al. 1977). Insect drift at sunset
in two experimental channels. 1/2 = half cobble embeddedness; F
= full embeddedness; Y = fish in channels; N - no fish.
Horizontal scale is temporal.
In correlational sampling in Bearskin Creek, Bjornn et al .
(1977) found that drift density in pools was significantly (p =
0.05) related to riffle area, but not to sedimentation in riffles
(Figure C.8) .
158
/SECTION III
-------
I*
= 1.0
a)
10 40 •»
nmi MIA IM'I
10
5"
SI
» mtumn < t.» nm m »rmn
Figure C. 8. (From Bjornn et al. 1977). Relationship between
riffle area and insect drift (upper graph) and between riffle
fines and insect drift (lower graph).
Drift density did not'correlate with drift density in Bearskin
and Elk creeks, both tributaries of the Middle Fork Salmon River
(Figure C.9).
The results of Bjornn et al.(1977) are somewhat surprising.
Sedimented substrata should offer fewer hiding places and livable
substrate (Hynes 1970 and Cederholm and Lestelle 1974) for
aquatic insects than would clean substrate. The literature on
predation suggests that prey vulnerability decreases as
environmental complexity increases (Huffaker 1958, Stein and
Magnuson 1976, Saiki and Tash 1979).
Wilzbach et al. (1986) investigated effects of reducing
habitat complexity in trout-macroinvertebrate predation by
covering the bottoms of test pools with fiberglass screening to
deny insects access to crevices. They showed that cutthroat
trout captured a higher percentage of drifting prey in crevice-
159 /SECTION III
-------
1.0
0.5
rsO.141
BEARSKIN CREEK
O 0.5 1.0 1-3
INSECT DRIFT I NOS./M*]
0.4
ELK CREEK
nO.539
"1334
INSECT DRIFT [NOS.'M3J
Figure C.9. (From Bjornn et al. 1977). Relationship between
fish density and insect drift in Bearskin and Elk creeks, Middle
Fork Salmon River, Idaho.
covered pools. The authors felt that the increased efficiency
was associated with increased visibility of drifting prey, rather
than with inability of prey to escape to crevices.
Embedded substrata should be somewhat similar to a substrate
covered with fiberglass screening. That is, visibility of
drifting prey should increase and the work required to find prey
should decrease as the substrate becomes smoother. Net energetic
efficiency of salmonids should increase in this circumstance.
Bjornn et al. (1977) (Figure C.10) showed that density of fish in
the fully sedimented channel was much lower than density in the
unsedimented channel, that fish in both environments grew in
length and weight, but that fish in the unsedimented channel were
longer and weighed more than in the unsedimented one. Perhaps
improved foraging efficiency, in the sense of net gain for units
160 /SECTION III
-------
of work, was responsible. This is of scientific interest, but
the negative influence of sedimentation on density and growth
rate is the main concern.
11'
MI
*
— 1
I
^3 |*° (*^°i •••*!•••••>
1 - *r»-«»tt
4
1
j,
•
1
j"
1 - *Mt-l«l«
• •10
1
^
1
1 k
1
l>10
1 rl
Figure C.10. (From Bjornn et al. 1977). Fish density, length,
weight, and condition factor in control and embedded channels,
for age 0 hatchery steelhead, 1975. Condition factor =
lengthvweight.
Konopacky (1984) felt that work reported by Bjornn et al.
(1977) and other University of Idaho researchers probably did not
take adequate account of in-channel vs. inflow drift, and area of
substrate that produced food upstream from the drift-sampling
points. Konopacky blocked incoming drift by placement of nets
0.25 mm in mesh opening across the entire inflow. Thus, drift
out of the stream channels originated within the channels. He
compared drift from a sand-pebble mixture (2.1 mm median particle
size) and a gravel (36.9 mm median size) (Figure C.ll). Drift
from the sand-pebble mix contained more organisms but similar
biomass of invertebrates in comparison with the gravel mix
(Figure C.ll). Individual weight of organisms tended to be
larger from the gravel-bottomed channel than from the sand-pebble
mix.
161
/SECTION III
-------
«*.
C C C C €> ft *>
I' UJLv-TWEWr^-^
rtrt.
cm • o
o • o
ikL
Figure C.ll. (From Konopacky 1984). Abundance (top graph) and
biomass (bottom graph) of invertebrate drift from channels with a
sand-pebble (white boxes) and gravel (black boxes) substrate.
The gravel used by Konopacky had 5.9% less surface area and
17 times more interstitial area than the sand-pebble mix. The
gravel did not offer a heterogenous habitat and probably did not
i
trap as much detritus as a mixture of gravel, small gravel, and
sand. Bjornn et al. (1977) showed that a substrate that was 1/3
to 2/3 embedded contained more insects than a substrate free of
embeddedness. Kennedy (1967) and Rabeni and Minshall (1977)
offered data that showed that a substrate with a mix of particle
162 /SECTION III
-------
sizes offers better habitat for macroinvertebrates,
163 /SECTION III
-------
D. RUBBLE COVER FOR SALMONID WINTER HIDING.
D.I. Effects of temperature on substrate use
Riinmer et al. (1983) observed that juvenile Atlantic salmon
moved into the substrate in rearing areas at water temperatures
below 10 C (Figure D.I).
160
MO
120
100
80
6O
40
5 K
OT
b 140
i ®°
too
80
SO
40
20
o
*-"*\
• * "\ -j
^_\ / -4
. 1
.-H
r^ \
197V
'
. Ah
197V
1379 \
Auq. 1 Sep
^.
1
1
I
*
I
Silt A I6O
140
120
KX3
80
60
i 40
=*->*-pT*fn ! 1
•5
" i IMO
*co
IOC
00
eo
40
zo
7777* tyrfrtr o
ifrrrcijir rr V
Del 1 New. 1
Sbft A | — ••
/""V^1 /ww
Lx-^
J^r
: fcXL
[ i
•
^J
2 4 • a BBt4 WIBZ02Z2426
Wottr Tifnptfohr* ('CJ
Figure D.I. (From Rimmer et al. 1983). Temporal changes in
number of visible Atlantic salmon juveniles in two sites (left
graphs), and counts in relation to water temperature in 3 years
(right graphs).
Hartman (1963) reported that brown trout tended to associate
with the stream bottom in winter at temperatures of 0.5 C, but to
be well above the substrate in spring at 12.5 C. Juvenile
chinook salmon and steelhead trout move into the substrate at
temperatures below about 5 C (Chapman and Bjornn 1969), at least
164
/SECTION III
-------
during the day. Figures D.2 and D.3 graphically demonstrate
cover-seeking behavior of steelhead and chinook salmon in an
experimental environment in laboratory tanks.
O
Figure D.2. (From Chapman and Bjornn 1969). Numbers of juvenile
steelhead visible above a rubble substrate during the day in
laboratory tanks in relation to water temperature.
3,
COHTHOLt
/ v
w I*
Figure D.3. (From Chapman and Bjornn 1969). Chinook salmon and
steelhead juveniles above rubble substrate during the day as a
function of temperature.
165
/SECTION III
-------
The experiments first noted above (Figure D.2) were
conducted with age 1+ steelhead, and those in Figure D.3 with age
0 steelhead and Chinook salmon. Everest et al. (1986) reported
that age 0 steelhead remained active above the substrate at lower
temperatures than yearling steelhead (Figure D.4). The ecolog-
ical or physiological significance of this difference is unknown.
•
100
78
= 80
* 1+«t««lh«ad
Figure D.4. (From Everest et al. 1986). Percentage of juvenile
steelhead observed active in a laboratory stream as temperatures
declined.
The laboratory observations stimulated tests in artificial
streams that more closely simulated natural environments of
juveniles. Fish that were actively moving downstream in the
Lemhi River in Idaho were placed in stream troughs at 12.2 C and
at 0-10 C, with substrate consisting of relatively smooth gravel
and clean rubble. A higher percentage of fish remained in the
lower-temperature environments with rubble. Downstream movement
was reduced in the tests in water of higher temperature (Chapman
and Bjornn 1969)(Figure D.5).
166
/SECTION III
-------
too
"
s
t»
h-
X
> IOO
w
_i
W
w
SUBSTRATE
STEELHCAO ^HUBBtE
R
^
M
„
K
yj>
QWUVK.
io ra-.
I
"•
CHI
P
i
NC
ran
n r
_J 6
TEST 1 I
TIHP.- o- 10 e. IMC.
Figure D.5. {From Chapman and Bjornn 1969), Percentage of
active downstream migrants that left experimental troughs in 10
days. Data include 2 replicates, 2 water temperatures, and 2
substrate types.
Morrill (1972) assessed downstream migration of age 0
chinook salmon from two stock sources (Lemhi River and Stanley
area of upper Salmon River) in stream channels with rock and
rubble {termed good substrate) and from gravel or shale (poor
substrate) and constant or declining water temperatures (fall
months, 1970 and 1971). As temperature declined below 10 C. ,
fish began emigrating. No residual fish were seen above the
substrate in daytime at temperatures below 5-7 C. No difference
in behavior was noted for the two different stocks of Chinook
salmon.
Morrill also found that more of the larger chinook salmon
moved out of channels with unsatisfactory substrate than did
smaller salmon, and that as chinook salmon approached 80 mm in
size, they showed increasing preference for suitable substrate in
which to hide. Morrill contended that suitable substrate for
winter hiding was an .important part of winter habitat
requirements.
Stuehrenberg (1975) found that age 0 steelhead trout and
chinook salmon did not remain in riffle sections of artificial
167
/SECTION III
-------
streams in winter when sediment had filled the interstitial
spaces of the gravel substrate. Age 1 steelhead used the depth
of pools for winter cover, hence tended to remain in the
artificial streams.
Observations of densities of juvenile chinook salmon that
remain in artificial channels at reduced temperatures as a
function of substrate type should be viewed with some caution.
Miller (1971) found that retention of fish in channels was a
function of water temperature, but that population density, race
of fish, and moonlight all influenced movement out of experi-
mental channels. In 4 of 5 experimental densities of juvenile
chinook salmon, more fish held in troughs with 13-15 cm gravels
than in 5-8 cm gravels. However, at the most-replicated density
(75 fish per trough, or 8.6 fish/m2}, more fish remained in the
channels with smaller gravels.
Several times in this section we have referred to daytime
observations of winter behavior. J. Griffith (unpublished) has
observed that some juvenile rainbow trout move out of daytime
winter cover and into shallower water during the night, even at
low temperatures, T. Hillman (unpublished) observed that pre-
smolt chinook salmon in winter remain in the substrate during the
day, but often appear on the bottom at night near the crevices or
interstices that they used during the day. The percentage of fish
that so move has not been determined, and the adaptive signifi-
cance of such movement is obscure.
D. 2 . Winter carrying., capacity
The next step in experimentation on effects of rubble on
downstream movements of juveniles, reported by Bjornn (1971),
involved placement of rubble piles in late summer on otherwise
relatively smooth substrate in Big Springs Creek, a Lemhi River
tributary. Subsequent electrofishing in winter demonstrated that
more fish inhabited the areas with rubble piles than control
168 /SECTION III
-------
areas without rubble (Figure D.6). Fish continued to use the
rubble areas for summer habitat and were found in greater numbers
in rubble late in the succeeding winter.
its
no
to.
O
29
mT
J
I
AUQ JAN
1997 19 *•
JUL
19*8
MAR AUO
DM O«9
Figure D.6. (From Bjornn 1971). Number of juvenile rainbow-
steelhead yearlings electro-sampled from control and test (rubble
added) sections of Big Springs Creek.
Hillman et al. (1986) demonstrated in Red River, a
Clearwater River tributary with high percentages of fines in the
substrate, that density of juvenile Chinook salmon at declining
temperatures could be increased several-fold if rubble piles were
provided for hiding cover (Table D.I), especially when rubble was
provided under sheltering streambanks. By March of the
succeeding year, cobble piles in the open stream were partially
displaced 0.5-1.0 m downstream and severely embedded with fines.
Cobble placed under the banks was not as severely embedded. No
sampling occurred after early November until March, hence it is
unclear exactly when embed,dedness increased in rubble piles, or
how long juvenile chinook salmon remained in them.
Bjornn et al. (1977) found that embedded channels that
contained boulders held fewer chinook salmon, steelhead, and
169
/SECTION III
-------
cutthroat trout than channels with boulders and no embeddedness.
Everest et al. (1986) reported that large rubble associated with
large boulders offered better steelhead overwintering habitat
than did rubble or boulders surrounded by small rubble, or
boulders alone (Figure D.7) .
Table D.I. (From Hillman et al. 1986). Numbers and densities
per m2 (in parentheses) of age 0 chinook salmon in control (no
rubble piles added) and test (rubble piles added) areas of Red
River, South Fork Clearwater River, Idaho in early winter 1985
and in March, 1986.
Habitat Date
Glide
Riffle
10/10/85
11/08/85
3/17/86
10/10/85
11/08/85
3/17/86
Exposed substrate
Altered Control
12(2.4)
16(3.1)
0(0)
10(2.4)
8(1.9)
0(0)
0(0)
0(0)
0(0)
0(0)
0(0)
0(0)
Underbank substrate
Altered Control
33(10.3)
20(6.3)
5(1.6)
18(8.2)
16(7.3)
1(0.5)
8(1.3)
4(0.6)
1(1.6)
0(0)
1(0.2)
0(0)
GOOD
POOR
Figure D.7. (From Everest et al. 1986) Examples of winter habitat
in which 1+ steelhead were observed in Fish Creek, a Clackamas
River tributary.
It is logical to extend the observations of Everest et al.
(1986) with the data of Bjornn et al. (1977), which showed that
juvenile steelhead, chinook salmon, and cutthroat trout tended to
remain in unsedimented channels with boulders and to leave
170 /SECTION III
-------
embedded channels (Figure D.8). We conclude that embedded
habitat reduces winter carrying capacity.
coiinoiQ
1— J
i 1 n
1
iRi
i
1
u
X.
X.
f;i;
|
1
n
Tt»T M. 14 It II I* 11 II II I* II
*Ctt*-Mi ft, ««. IM, IM, («. IM, M.JH Cl, tl,,
rWH M WMWHMW
Figure D.8. (From Bjornn et al. 1977) . Densities of fish
remaining in artificial stream channels after 5 days in winter
tests, 1975. W = wild; H = hatchery; CK0 = age-0 chinook salmon,
SH0 = age 0 steelhead; CT0 = age 0 cutthroat trout; CT1(2 = a
-------
1969, Hillman et al. 1986). Coho salmon tend to aggregate on the
bottom of pools in winter (Hartman 1965, Mason 1976, Bustard and
Narver 1975a). Cunjak and Power (1986) reported that brook and
brown trout in the Credit River did not enter the substrate in
winter, instead they aggregated beneath cover in pools and close
to point sources of groundwater discharge.
Hartman (1963) reported that brown trout hid within the
substrate in cold water, and Stuart (1957) recorded a downstream
movement of brown trout in October as water temperatures dropped
or in advance of declines to winter temperatures less than 4 C.
Allen (1951) reported for the Horokiwi Stream in New Zealand that
brown trout fed all winter, generally had no winter check on
scales, and did not move seasonally. The Horokiwi Stream
generally had water temperatures in excess of 7 C.
Rimmer et al. (1983) found a dramatic shift by Atlantic
salmon from summer feeding stations to winter hiding locations in
the rubble substrate as water temperatures decreased below about
10 C (Figure D.9). These workers established that this shift
140
BO
JlOO
vt
•S 80
I sc
40
SC
2 4 6 B O IZ 14 16 18 ZO Z2 24 Z6 ZS
Dull h *tH ember
Figure D.9. (From Rimmer et al. 1983). Chronological changes in
numbers of visible juvenile salmon counted, and relationship
between number counted and water temperature in the same years,
and salmon trapped as they left the site.
172
/SECTION III
-------
occurred within the habitat areas used in summer, and that it was
not accompanied by downstream movement out of the area.
Cutthroat trout of larger size tend to leave summer habitat
in the Middle Fork of the Salmon River (Mallet 1963) and the
upper Clearwater River (Ball 1971, Hogander et al. 1974) and to
return in spring. Chapman (1962), Salo and Bayliff (1958), and
Shapovalov and Taft (1954) found little or no downstream
movements of juvenile coho salmon during fall and winter.
Petersen (1980) found that juvenile coho salmon in the
Clearwater River on Washington's Olympic Peninsula often moved
considerable distances downstream in fall, then entered wall-base
channels (small streams that issue forth from the base of incised
terrain along main river channels) or spring ponds, where they
over-winter. Everest et al. (1986) stated that small ponds
provided important overwintering habitat for juvenile coho salmon
in Fish Creek in the Clackamas River, and suggested that
provision of this type of habitat would be a productive
enhancement tool. Steelhead used large boulders and rubble for
overwintering in the main channel of the stream.
Murphy et al. (1986) examined effects of various habitat
features on winter fish densities in Alaskan coastal streams.
Fine sediment, defined as percentage of cross-stream transect
length covered by particles smaller than 2 mm, was not
significantly related to density of coho salmon fry, parr, Dolly
Varden parr, or trout parr (steelhead and cutthroat combined).
Coho salmon parr density was significantly related to undercut
banks, pool volume, and instream debris (large woody debris).
Debris would be a correlate of pool volume. Dolly Varden parr
density was significantly related only to woody debris; trout
parr density to undercut banks and canopy density. Sediment
percentage on transects averaged less than 11.2% in the most
sedimented stratum, and no single sediment percentage exceeded
173 /SECTION III
-------
19.2%. Average sediment ranged from 7.8 to 11.2% on old-growth,
buffered, and clear-cut strata. These surface sediment
proportions are not particularly high when compared with, say,
the South Fork Salmon River in Idaho (Platts and Megahan 1975},
although fines in the latter were assessed as smaller than 4.7 mm
rather than 2 mm. Particle size distributions in samples from
the South Fork Salmon River indicated that the percentage of
particles smaller than 5 mm (about 35%) was roughly equivalent to
20% particles smaller than 2 mm.
Apparent differences in winter behavior in various popu-
lations may be associated with fish size, severity of winter
temperatures to which the population has adapted, availability of
refugia from freshets, and time-series behavior. One would
expect, for example, that streams with a snow-melt hydrograph
would have salmonid populations adapted to in-substrate
overwintering of small fish (juvenile chinook salmon, steelhead,
and probably one-and two-year-old cutthroat and bull trout).
That populations of salmonids can adapt genetically to local
temperature patterns was shown by Miller (1971), who demonstrated
that three races of spring chinook salmon, al 1 from the upper
Salmon River drainage in Idaho, differed significantly in temper-
ature-related behavior in artificial channels. Cunjak and Power
(1986), comparing substrate hiding in brook trout in northern
latitudes of Canada with above-substrate aggregations beneath
cover in the southern latitudes, hypothesized that these
different behaviors reflected differences in severity of winter.
Severe freezing in the stream would favor overwintering in the
substrate, while alternating freezing and thawing and different
types of ice formation would make a more active wintering
behavior adaptive.
Where suitable rubble substrate is not available, pre-smolt
salmon and steelhead appear to move downstream until they find
174 /SECTION III
-------
such habitat. These fish would not have to return in the
following spring. Absence of nearby clean rubble should be
particularly limiting for juvenile cutthroat and bull trout, or
for young resident rainbow trout, which should have difficulty in
returning -against freshet flows from the same downstream
migration as made by larger, older fish of the same species.
Juvenile salmonids (eg. age-0 steelhead) that must remain in the
stream for two or more years would also suffer from absence of
winter hiding habitat in the rearing areas. Ice and ice scour
would cause a difficult situation for juveniles unable to find
refugia within summer habitat areas.
Larger, older fish capable of reascending streams in spring
should have difficulty in burrowing down into interstices in
summer habitat, simply on grounds of relative size of fish and
crevices. Movement downstream is probably adaptive for these
fish.
In coastal watersheds with a rainfall hydrograph and where
ice formation in rivers is unusual, as well as in areas with
mixed hydrologic pattern, steelhead use rubble refugia in winter
(Bustard and Narver 1975a, Everest et al. 1986) while coho salmon
aggregate over bottoms of pools, under log jams, or utilize
spring ponds and wall-base channels (Peterson 1980, Tschaplinski
and Hartman 1983).
Mason (1976) and Ruggles (1966) offered experimental refugia
to coho, in the form of flat rocks on pool bottoms, but were
unable to increase habitat capacity during winter. Bustard and
Narver (1975b) reported that coho do not enter rubble substrate
during winter. It appears that quiet, relatively deep pool
bottoms, undercut banks, and log jams offer the best winter
habitat. Tschaplinski and Hartman (1983) showed that water
velocities in the areas used by juvenile coho during winter were
always less than 0.3 m/s.
175 /SECTION III
-------
Bustard and Narver (1975a) reported that focal point
velocities used by young steelhead in winter were directly
related to water temperature (Figure D.10). Both steelhead and
coho salmon yearlings tended to use deeper water than age-0 fish
(coho salmon take one or two years and steelhead 2-4 years to
reach smolt size) (Figure D.ll) in the area studied by these
authors.
FOCAL POINT VELOCITIES
0---0 AGE 0 STEEIHEAO N = 36. , .0. 34*
•—• AGE 1*SIKIHEAD N *66. r .0.32**
*
u
MAXIMUM AREA VELOCITIES
o—oAGE 0 STEELHEAO t4*36.,.Q49f*
•—•ACE 1*STEEI.HEAO N«« •
r -O.J3"
Figure D.10. (From Bustard and Narver 1975a). Mean focal point
and maximum area water velocities selected by age-0 and age 1+
steelhead in relation to temperature.
Bustard and Narver found little winter use by age-0
steelhead of substrata with diameters smaller than 10 cm, at
least in the main portion of Carnation Creek, but it is unclear
whether this was because of scarcity of fines. The report tends
to indicate that fine particles were available but not used.
Steelhead did not winter in side channels as did coho salmon.
Bustard and Narver (1975b) showed that both coho salmon and
cutthroat preferred sidepools that offered overhanging bank cover
176 /SECTION III
-------
AOt 0 COHO
N' 117*
I» 7J.7 CM
tD
ACt 0 StCClHCAD
H'Tt
i« ]t.7CM
ACt I* JIHIMMD
N»III
2*11.1 CM
n n i n s
OtPTH OF WATtl ICMl
Figure D.ll. (From Bustard and Narver 1975a). Depth of water
selected by coho salmon and steelhead at water temperatures of 7
C or less.
and clean rubble substrate as opposed to silted rubble. In tests
that offered equal amounts of side-bay rubble with and without
silt, cutthroat were at least 10 times more abundant in clean
rubble areas (Figure D.12).
n
jfz
im
COHO SALMON
CUTTHROAT TROUT
3.5-5.8C
Figure D.12. (From Bustard and Narver 1975b) Percentage of coho
salmon and cutthroat trout that chose clean and silted rubble
when offered equal amounts of clean and silted rubble in
sidepools. Left circles pertain to periods of obligatory
residence, right circles to volitional access to main stream.
177
/SECTION III
-------
D.4. Conditions in the Northern Rockies
Throughout the Northern Rockies, snow-melt hydrographs
dominate flow patterns except in relatively uncommon spring-fed
streams. Addition of sediments to the substrate would tend to
block juvenile salmonids from essential winter refugia. Most
species of juvenile salmonids in the region either are known to,
or thought to, seek refuge in the substrate in winter, although
they may leave these sites during the night. Inasmuch as such
night shifts, to whatever extent they occur, are followed by
disappearance of fish during the following day, winter refugia
are apparently required during the day.
Embedded substrata would potentially limit carryover of
resident juveniles to the next growing season, and force
downstream movement of anadromous fish. Whether the latter
movement would reduce survival over that occurring, on average,
in rearing areas has not been determined. The fact that
juveniles stay in rearing areas if offered suitable overwintering
conditions suggests that fall migrants seek the most-upstream
available substrate near rearing areas.
We do not know if winter habitat is in short supply. The
work of Mason (1976) showed that winter cover controlled
overwintering abundance of coho salmon. Chapman (1966)
hypothesized that population regulation in winter was solely
space related, without a food component. Cunjak and Power (1986)
suggested that space limitations in winter mean that many fish
simultaneously seek a common goal (suitable winter space). This
results in a clumping or squeezing effect where fish are
restricted in their occupation of a limited spatial commodity
(Cunjak and Power 1986) .
We suggest that at the higher densities of salmonids
observed by Thurow and Burns (unpublished) and by Petrosky and
Holubetz (1986), about 50-60 fish/100 m2, winter densities in
178 /SECTION III
-------
stream reaches where fish winter where they reared would also
equal about one fish per two square meters. Given that
interstices do not occur everywhere in any stream reach of
several hundred meters, in fact lie in "clumped" or contagious
distributions, densities of fish per unit area must be much
higher in those limited areas during the winter if all fish in
the rearing areas were to find refuge in the substrate. The
possibility of competition for winter spaces certainly exists.
Why should lower, larger stream areas in the northern
Rockies region not provide better habitat than smaller tribu-
taries? The answer .may lie with ice scour, which occurs in
winters with extended extreme cold when rapid warming follows the
cold period. Small streams at high elevation tend to freeze over
quickly, and are insulated from cold air by bridging of deep snow
in most of the northern Rockies. They may have less tendency to
lose ice cover during temporary warm spells, and to offer more
moderate water temperatures beneath the snow. Larger streams do
not freeze over as readily, but may often reach colder tempera-
tures than snow-insulated tributaries and have more extensive
substrate disturbance during ice scour. Ice block buildup can
develop massive proportions, and breakup may bring catastrophic
scouring.
The solid cover of ice and snow at high elevation has an
additional benefit for salmonids that overwinter there. Warm-
blooded predators do not have access beneath the snow in most
such areas. In the larger streams at lower elevation, birds and
mammals may have easier access to prey.
Another advantage accrues to fish that overwinter in the
same areas in which they reared. They do not have to move
downstream in the low, clear flows of fall to reach wintering
habitat. They continue to have access to familiar cover or move
relatively short distances to winter habitat. The longer the
179 /SECTION III
-------
journey to winter habitat in low, clear water, the higher will be
the likelihood that the transient is taken by a predator, either
warm— or cold-blooded. It might be argued that anadromous smolts
must traverse the same distance (equivalent to the distance from
rearing area to downstream wintering area) sooner or later. We
contend that later is better, at least for fish of smolt size,
largely because spring flows are higher and water more turbid,
factors that should reduce predation*
In light of the foregoing discussion, we suggest a scenario
for the worst of all worlds for salmonids of the northern
Rockies. This would be a winter of light snow and cold air
temperatures (no snow bridging for protection), followed by a
late winter "Chinook" or thaw with accompanying ice scour, in
turn followed by low flows in the following summer because of low
snow pack. Anadromous fish that move toward the sea in the
spring could avoid the summer low flows but face truly difficult
passage problems in reservoirs because of low discharge in
spring.
Hillman et al. (1986) found that rubble piles that they
placed in Red River became embedded with fines between November
and March. Although no data exist to establish the fate of fish
that were using the piles for wintering in November, the role of
fines in filling interstices of winter habitat may be important.
If the displaced fish were less likely to survive the displace-
ment and subsequent search for replacement winter habitat than
other members of the cohort that departed in early fall, addition
of rubble piles conceivably increased mortality for fish induced
to winter in them.
Whether the foregoing, or other hypotheses, represent
reality, it seems prudent to follow the lead offered by existing
information. That is, to assume that fish seek winter habitat in
upstream locations, and addition of sediment that reduces
180 /SECTION III
-------
availability of this environmental requisite probably increases
mortality.
181 /SECTION III
-------
IV. FINE SEDIMENTS AND CHANNEL MORPHOLOGY
Although we have not specifically discussed channel
morphology as a function of sediment recruitment, we believe we
cannot ignore this topic. Our work assignment dealt with
sediment-related criteria for evaluation of best management
practices. We concentrated on the intragravel environment and
fine sediments as if the larger sediment components were fixed.
Although an appropriate approach for the assignment, it ignores
the role of unusual storm events and landslides in controlling
channel shape. This report section, while certainly not
exhaustive, may serve to caution readers about the importance of
channel structure in legislating conditions for salmonid spawning
and rearing.
A. AGGRADATION AND DEGRADATION
Lisle {1982) recorded changes in channel structure after a
major storm event in northwestern California. Gaging station
cross-sections widened by up to 100% and aggraded as much as 4 m,
then degraded to stable levels over 5 years or more. Pools and
bar relief diminished. Figure A.I plots width, mean depth and
mean velocity against discharge for preaggradation, peak
aggradation, and post-aggradation periods. Aggradation reduced
friction, increased width, decreased depth, and increased average
velocity for given discharges. Increased aggradation tended to
increase the effectiveness of moderate discharges in bed load
transport, hence in shaping the streambed. Slight increases in
sediments from the watershed would more easily be transported
under these aggraded circumstances.
The net effect of aggradation would be to reduce channel
diversity. More fine sediments would move at low and moderate
discharges, and channel roughness would decrease, suggesting
182 /SECTION IV
-------
probable increases in embeddedness levels.
No'trt Fork Trimly River
o
u 0
I
woo;
002
E
£
1
£
o 1 »«•••»•
19
Oitehtrf*
100
Figure A.I. (From Lisle 1982). Relationships of stream width,
mean depth, and mean velocity to discharge in three periods;
preaggradation (I) , peak aggradation (II) , and postaggradation
(III) . Qc = discharge at initiation of gravel transport, Q^ =
discharge at intersection of hydraulic geometries for pre- and
peak aggradation periods.
Nolan and Marron (1985) recorded major changes in bed
profiles as a result of floods in Redwood Creek and in the San
Lorenzo River in California. Figures A. 2 and A. 3 indicate that
the changes were quite dramatic. The authors noted that changes
in channel geometry in northwestern California tend to be long-
lasting, and that human activity may have reduced the storm
magnitude required to cause channel changes from landslides.
Pearce and Watson (1983) recorded channel morphology after
183 /SECTION IV
-------
10 20 30
HORIZONTAL DISTANCE.
IN METERS FROM LEFT MONUMENT
Figure A. 2. (From Nolan and Marron 1985). Pre-and post-flood
cross profiles of Redwood Creek at a gaging station. Profile of
1958 shows pre-aggradation state, that of 1973 shows the post-
flood profile. Profile of 1981 was still elevated.
two landslide episodes in six deeply incised streams in New
Zealand. Debris torrents carried about 4,700 m3 of logs and led
to formation of major log jams in five streams. This material
led to multi-stepped stream profiles, aggradation of channel
reaches up to 150 m long to mean depths of 1.2-4.1 m of deposits,
reductions in average gradient, and a reduction in average
particle size in the substrate. The authors estimated that
sediment stored behind log jams equaled normal supply of sediment
for 50-220 years from hillslopes to stream channels. Log jam
failure would lead to major morphological changes in higher-order
streams.
. 3-Oi
!i.
m tt
It"
iio.s
-t-
10 IS JO
HORIZONTAL DISTANCE.
IN METtRS FROM LEFT MONUMENT
Figure A. 3. (From Nolan and Marron 1985). Cross profile of the
San Lorenzo River before and after January 1982 flood. Signifi-
cant quantities of sediment had been removed by March 1982.
184 /SECTION IV
-------
Beschta (1983a) studied aggradation in the upper Kowai basin
in New Zealand. A 150-year storm there in 1951 stored sediments
in the stream channel that remain to date. As distance down-
stream from sediment input points increased, widths and lengths
of the deposits increased and depths of deposition decreased. He
noted that the major storm caused hillside stability thresholds
to be exceeded, and that thresholds can vary with rock type,
faulting, weathering rates, and land use. In another paper,
Beschta (1983b) reported that the channel of the Kowai River
widened in the vicinity of landslides, and that the widening from
aggradation proceeded downstream at a rate of less than 1- kra/yr.
185 /SECTION IV
-------
B. ROLE OF BED MATERIAL IN GOVERNING MORPHOLOGY OF STREAMS
Beschta and Platts (1986) discussed the important role of
bed material in influencing channel characteristics. They noted
that a change in the median particle size can influence the
frequency and magnitude of bedload transport and may affect
channel dimensions.
Increased deposition of fines between gravel particles may
affect subsequent bedload transport. Fines in interstices of the
bed may delay the onset of bed movement during large flows
Beschta and Platts (1986); a movement critical to removal of
accumulated fines in spawning gravels. These authors also
recognized that salmonids have evolved and adapted to the natural
sediments of the stream channel, and that a complex mix of
sediments, in combination with certain hydraulic conditions, is
needed. They note, for example, that the optimum spawning
substrate appears to consist of gravel with small amounts of
fines and small rubble to support egg pockets and stabilize the
bed in high flows. This recognition of the importance of rubble
to the spawning environment constitutes one of the few in the
literature. One can infer from Bj ornn et al. (1977) that a mix
of various particle sizes also is needed by macroinvertebrates.
Whether fines enter a stream from logging, road-building,
natural erosion processes, or exposure of streambanks by
livestock overgrazing and trampling would seem at first glance to
be irrelevant. The micro-effect of a given increment of
particles of a given size should be the same. However, different
sediment accelerators and the resulting income to the stream may
have profoundly different effects on substrate particle
compositions (Sullivan et al. 1986). A landslide as a debris
torrent, whether natural or man-caused, brings large and small
sediments and woody debris with it, forming a source of materials
186 /SECTION IV
-------
for major or localized aggradation and altered morphology.
Recruitment of sand from roads or livestock use may well provide
a source of materials that more insidiously alter not only micro-
environments for fish and insects but stream width, depth, and
friction factors. Table B.I, from Sullivan et al. (1986) notes
the differing effects on channel morphology of debris torrents,
landslides, and volcanic fines in several example streams.
Table B.I. (From Sullivan et al. (1986). Case histories of
channel response and recovery from disturbances.
Area
Northern
California
Middle Fork
Willamette,
Oregon
South Fork
Salmon River,
Idaho
Mt. St. Helens
Blast Area
Type of
Disturbance
Logging-induced
landslides, surface
erosion; large floods
Logging-induced
landslides, debris
flows, large flood
Logging-induced
landslides, surface
erosion; large floods
Large increases in
fine sediment; LOD
added, then salvaged
Channel
Response
Widening,
aggradation,
loss of riparian
vegetation
Widening,
aggradation,
loss of riparian
vegetation
Aggradation,
deposition of
fines, infilling
of pools
Aggradation,
fining of the
bed, filling of
pools
Recovery Period (Yrs.)
Channel Riparian
5-60 100-200
10-20
10
100 • 200
Little impact
8 - 25? 100 - 200
187
/SECTION IV
-------
C. WOODY DEBRIS AND CHANNEL MORPHOLOGY
Woody debris, in conjunction with sediments, has profound
effects on salmonid habitat in streams. An extensive literature
has developed around this topic and will not be reviewed here,
except in summary. We suggest that the reader obtain Bisson et
al. (1987), Sedell and Swanson (1984), Sedell et al. (1985), and
Everest et al. (1986) for entry into the literature on woody
debris. We will only use this topic to further note the
importance of stream structure.
Unpublished data of Dr. P. Bisson, as included in Sullivan
et al. (1986) show that removal of woody debris reduces the
stream area in pools, increasing that in riffles (Figure C.I).
Strewn Am (%)
80
70
60
SO
40
30
20
10
0
Str»»m Clasn Up
Figure C.I. (From Sullivan et al. 1986). Stream surface area in
pools and riffles in zones from which woody debris was cleaned
out and in uncleaned zones.
Bisson et al. (1986) reported that hydraulic characteristics
of streams influenced habitat utilization patterns of coho
salmon, steelhead, and cutthroat trout. Riffles favored
steelhead, pools favored coho, and cutthroat used both habitats,
although less effectively than the favored species. The authors
188 /SECTION IV
-------
related body and fin shape of the three species to habitat
utilization. Coho salmon have a deep, laterally compressed body
with large median and paired fins, a morphology that promotes
rapid turns and quick but transient burst swimming. Steelhead
have a cylindrical shape with short median fins and large paired
fins; attributes well adapted to holding position in swift water.
Cutthroat trout are not morphologically specialized, a finding
that may explain why coho and steelhead dominate them in
sympatry.
Dolloff (1986) removed debris smaller than 60 mm and large
debris that was not embedded in the stream channel from sections
of two streams. Subsequent population densities and production
by coho salmon and dolly varden were higher in the uncleaned
sections (Table C.I).
Table C.I. (From Dolloff 1986). Average production and
difference in production of coho salmon and dolly varden in
cleaned and uncleaned sections of Tye and Toad creeks during
three summers, 1979-1981.
Species and age
class
Coho salmon
AgeO-t-
Age 1 +
Dolly Varden
Age 14
Age 2 +
Coho salmon
Age 04
Age 14
Dolly Varden
Age 1 4
Age 24
' Icnncd 1 ii
fg/m-)
Tye Creek
0.80
0.25
0.13
0.12
Tond Creek
0.43
0,25
0 13
0.24
','^T'
0.76
0.39
0.16
0.24
0.48
0.35
0 24
0.46
I5i (Terence
4-5
-36
-19
-50
-10
-29
-46
-48
Dolloff attributed the decline in density and production in
the cleaned areas to reduced visual isolation in summer and to
loss of winter cover. Reductions in production appeared somewhat
189 /SECTION IV
-------
more severe for dolly varden than for coho salmon.
The foregoing studies illustrate how stream structure may
favor or detrimentally affect particular salmonids.
190 /SECTION IV
-------
D. RELATIVE ROLE OF STREAM MORPHOLOGY AND FINES
In the northern Rockies, stream morphology, as controlled by
sediment and geologic structure and, in some cases, by woody
debris, together with bank integrity and riparian vegetation, may
be more important as density legislators than is the quantity of
fines in and on the substrate. Woody debris may be especially
important in headwater streams and in smaller, low gradient
stream reaches in Idaho. We do not mean to imply that excessive
fines have no negative impact on reproductive success, or that
our opinion should be taken to mean that we believe control of
recruitment of fines is unnecessary.
It is crucial to our understanding of stream systems in the
northern Rockies that more holistic approaches be developed for
studies of factors influencing salmonid survival and densities.
Much more attention must be paid to effects of stream morphology.
We need investigations of the role of woody debris in affecting
use of habitat specifically by steelhead, Chinook salmon,
cutthroat trout, bull trout, and rainbow trout. We would benefit
from investigations of riparian vegetation as cover and source of
terrestrial insects. Winter habitat as a population legislator in
relation to summer habitat requires study.
Without these types of research, without recognition of
multivariate response-surface functions, questions about
quantitative effects of fines or embeddedness on fish are fated
to elicit vague and equivocal responses.
Studies of effects of various habitat features on fish
abundance require imaginative approaches to assure full seeding
so that the effect of recruitment is minimized. Transfer of fish
from one site to another to assure full recruitment may solve
this problem. Stocking of hatchery fry or fingerlings may serve.
191 /SECTION IV
-------
V. TOOLS FOR PREDICTING EFFECTS OF FINE SEDIMENTS ON FISH
AND AQUATIC MACROINVERTEBRATES
We use the term "tools" in this report segment in the sense
of "models". Specific techniques for securing data on fines in
and on the substrate can be found in the summaries by Levinski
(1986), and in the papers of Brusven and Prather (1974), Grouse
et al. (1981), Burns and Edwards (1985), Lotspeich and Everest
(1981), Shirazi et al. (1981), Walkotten (1976), and Platts et
al. (1979) . Terhune (1958) provides details on the Mark VI
standpipe, used to assess permeability and apparent velocity in
gravels.
A. INTRAGRAVEL SURVIVAL OF EMBRYOS TO EMERGENCE OF ALEVINS
In light of section II, we now discuss several tools that
have been proposed for prediction of effects of fine sediments in
field conditions. The models include the fredle index, geometric
mean particle size, percentage of fines, and gravel permeability.
Platts et al. (1983), Burns and Edwards (1985), Shirazi and Seim
(1979), Terhune (1958), and Levinski (1986) provide procedural
details for various types of physical sampling tools. We confine
our comments primarily to questions of sampling validity, bias,
and precision.
Any ' and all of these independent variables, to offer
utility, reality, and permit quantitative prediction of sur-
vivals, must be shown to reflect conditions in the egg pocket of
the salmonid redd. It is not sufficiently accurate to assume
that data from the salmonid redd applies to the redd pocket, for
the redd consists of substantial areas outside egg pockets
(Figure A.I). Much misleading information and unnecessarily high
variances may have been reported in the literature because
workers accepted "redd" and "egg pocket" as synonymous.
192 /SECTION V
-------
i-XA;?'-Sfiv; ' * r £•••-*•'" «•••* * t^vijfjg r. rf-5H
COMPLETED EGO POCKET
Figure A.I. Schematic conception of egg pockets within a
salmonid redd, in this case a chinook salmon. This diagram is
based on numerous nightly measurements of developing redds and
excavation of redds (Chapman et al. 1986). Note cobbles that
form the centrum in the bottom of the egg pocket, multiple egg
pockets spaced sequentially upstream, and depths of egg placement
in pockets.
We concede that it is possible that conditions "in the redd"
may serve as indicators of conditions in the egg pocket.
Inasmuch as data from only one egg pocket have been examined
(Platts et al. 1979), substantial testing of the relationship
between average conditions within the redd periphery and those in
the egg pocket is required .before possibility progresses to fact.
Currently-available data that relate survival to gravel
composition and permeability in "redds" may be viewed as broadly
instructive but specifically speculative as quantitatively
193
/SECTION V
-------
predictive analogs. Some of the field sampling information may
be accurate, in that coring or permeability samples actually
included information from an egg pocket or pockets, but we have
no means of determining if that was the case.
In order to demonstrate the generic problem of sampling in
egg pockets, we have prepared a diagram to scale that shows three
McNeil cores inserted into a Chinook salmon redd in longitudinal
section (Figure A.2). The diagram of the redd section was
developed with the experience of sequential nightly measurements
by T. E. Welsh of redd construction in 274 Chinook salmon redds
in the Columbia River, and excavations in 369 redds (Chapman et
al. 1983 and 1986) .
/
TAIL SPILL
Figure A. 2. Scale schematic diagram of egg pockets in a chinook
salmon redd in the Columbia River, with three hypothetically-
placed McNeil core samplers to scale. This information is based
on observations of Chapman et al. (1983, 1986). Note failure of B
to reach centrum of egg pocket and that A and C lie outside
pockets.
The ridges between eggtpockets in Figure A.2 may be slightly
wider in upstream-downstream direction than those one might find
in smaller streams or redds of smaller adults, but this diff-
erence is immaterial in the following discussion. We have
slightly exaggerated the number of cobbles in the egg pocket
194
/SECTION V
-------
centrum in order to illustrate the sampling problem.
Cores A and C would sample materials entirely out of the egg
pocket and partly in consolidated materials not cleansed by the
female. Core B would sample above the egg pocket portion holding
most of the eggs (see Figure A.I), thus missing the large
particles in the centrum.
The tools shown as A, B, and C in Figure A. 2 have a core
diameter of 15 cm. Doubling the diameter would reduce bias
associated with whether one treats large particles at the edge of
the cylinder as in or out of the sample. The larger cylinder
would enclose a larger sample, but would not solve the problem of
sampling outside of the egg pocket components that harbor the
bulk of the incubating embryos.
While a core of depth similar to that in Figure A. 2 would
more probably reach the bottom of the egg pocket in core B in,
say, the shallower egg pockets of a coho salmon or steelhead
redd, problems represented by cores A and C would remain. In the
case of redds of smaller fish, such as resident rainbow, risk
would increase of extending the core through the egg pocket into
the consolidated material below it. Shortening the core for
smaller fish would reduce this likelihood while leaving problems
of missed egg pockets unsolved. Reducing core diameter for small
fish and predominantly small substrate particles, while
advisable, would not solve problems represented by A and C in
Figure A.2.
The section in Figure A.2 does not address the problem of
laterally missing the egg pocket. Figure II.B.2, from Hawke
(1978) shows longitudinal placement of egg pocket in plan view,
and demonstrates that the unwary redd sampler is much more likely
to miss than hit the egg pocket.
195 /SECTION V
-------
While a freeze-core, say a triple-probe sampler, would not
eliminate problems represented by A and C in Figure A. 2, or
errors in plan-view context, extraction of an intact core permits
the field worker to visually examine the core as a vertical
representation of the substrate. Consolidated materials below
the egg pocket can be identified and eliminated, and absence of
large particles at the bottom of the core would be grounds for
sample discard.
Shirazi et al. (1981) compared McNeil and freeze-core
systems, and found (see section II.A.I) that triple-probe freeze
cores and McNeil samples yielded similar results in spawning
habitat. The reader should not interpret these results as per-
taining to the egg pocket or to the problems elucidated above.
In the following subsections, we suggest that careful field
researchers might find it profitable to observe redd development
for a number of redds, identify egg pockets during the construc-
tion process, then survey the location of the pockets. Triangu-
lation or surveying instruments would permit later location of
the exact egg pocket for coring or permeability measurements. If
level-rod elevations were taken in the egg pockets, and diameter
of cobble or large gravel particles at the bottom of the pocket
were estimated, it would be possible to take core samples, even
with a McNeil sampler, exactly within the egg pocket and to
assure that the core extends to the bottom of the pocket and no
further. The McNeil cores could not be used for information on
vertical stratification, for only freeze cores provide these
data. For large fish, such as Chinook salmon, longer tubes might
be required for the McNeil sampler to extend to the bottom of the
pocket.
We believe that data scatter in, for example, the relation-
ship between fines and survival or between dg and survival would
decrease for information obtained solely within the egg pocket.
196 /SECTION V
-------
With data from egg pockets and redd capping, quantitative models
and predictive tools could then be developed for field data.
Even if redd capping were never successful. knowledge of
structure and composition of the egg pocket could be applied to
laboratory experimental conditions, making them better analogs of
egg pockets in the field.
In the material to follow, we suggest some predictive tools
that might be developed, using available information on gravels
and survivals as examples. The models shown are not intended for
use directly as quantitative predictors of survivals in the
northern Rockies or elsewhere.
A.I. Fredle index
In this section, we re-analyze survival data from natural
redds, as reported by Tagart (1976, 1984) and Koski (1966). We
calculated the fredle index for gravel samples as if the latter
were obtained in egg pockets, although they were obtained "in the
redd". Figures A.l and A. 2 show why samples taken with McNeil
cores probably included materials not pertinent for the egg
pocket.
We have not used the fredle index with laboratory data.
Laboratory data of Phillips et al. (1975), as tabulated with the
fredle index by Lotspeich and Everest (1981), demonstrate that
emergence (data for period from button-up fry to emergence only)
positively relates to the fredle index. However, these data do
not include the full intragravel period and should not serve for
quantitative prediction in field conditions. They merely demon-
strate that a relationship exists between fredle index and
emergence success.
The fredle index incorporates the advantages of dg, as eluc-
idated by Platts et al. (1979) and by Shirazi et al. (1981), with
a mixing index. We deem the mixing index desirable because
197 /SECTION V
-------
gravels with quite different percentages of fines and of cobble
can have the same dg (Lotspeich and Everest 1981) . The mixing
index must incorporate assessment of large as well as small
particles because of the important role of large gravel and
cobble in egg pockets in natural redds. We do not agree with
elimination of large particles (Adams and Beschta 1981, Tappel
and Bjornn 1983) from model functions.
For all works in which natural redds were capped with
netting and survival to emergence was measured in concert with
sieving of redd particles, we plotted particle size compositions
by weight for gravel samples from each redd, determined dg and
appropriate percentile particle sizes (^25, d75), then calculated
the fredle index. We next plotted survivals from these redds
against the fredle index. (Figure A.3). Figure A.3 represents an
80"
7D-
150-
I 40-
2CH
to
-All redds
r3' O.M
F • «.ta
1 S 37 df.
t966 only
r2 - 0.54
F-IJ09
1 + 19 df,
' r i v
5 8 7 8 910
OJ 03 04 OS 08 08 I 33
FREDLE INDEX (l|)
Figure A. 3. (From Tagart, 1984 and file data at Oregon State
University, from work of Koski 1966). Survival to emergence of
coho salmon from natural redds in relation to the fredle index.
Note low r2. Regression is significant at p = 0.05.
example of an analysis that should be done with survivals and
198 /SECTION V
-------
gravels from egg pockets. The figure does not purport to offer a
quantitative prediction tool. The fredle index should not drop
below 1.0, but does in a few cases from Koski' s work because we
had to extrapolate to estimate some small size components in the
data retrieved from files at Oregon State University. We
requested these because Koski's thesis did not contain data for
individual redds.
The fredle index for a given survival as plotted in Figure
A. 3 may be lower or higher than was the case in actual egg
pockets. The fredle index from corings of Tagart (1984) and
Koski (1966) may include more coarse fines and less small fines
than would be present in an egg pocket, because of such problems
as corings on the "ridges" between egg pockets, failure of cores
to penetrate to the bottom of unconsolidated redd materials, and
intrusion of fines in egg packets in a manner different from that
outside the pocket.
The plots in Figure A. 3 would lead one to infer that as the
fredle index drops below about 1.5, one should expect survival to
decline below 25%. Again, these data should not be used for
quantitative predictions in the field. They are from coho redds
and have a high scatter about the semi-log regression; a scatter
possibly caused in part by gravel sampling outside of the egg
pocket.
We suggest that to develop useful and accurate quantitative
predictions of survival in relation to the fredle index, it would
be necessary to freeze-core repeatedly in the redd being sampled
until an egg pocket is identified by presence of several embryos
or alevins frozen in the core. Any undisturbed substrate that
"floors" the egg pocket should be eliminated from the sample, and
the remaining core components should be sieved. Ideally, several
egg pockets would be so-sampled, thus providing data that would
permit assessment of variance and, incidentally, better simul-
199 /SECTION V
-------
ation in the laboratory of egg pocket structure and composition.
Unfortunately, the work suggested in the preceding para-
graph, however useful in defining egg pocket structure and
composition, would disturb the natural redd and kill embryos,
preventing unbiased assessment of survival to emergence by redd
capping. Survival to the time of freeze-coring is not suffi-
cient, as it does not include the emergence phase. It may be
necessary to accept a lesser objective. That is, one could cap
the redd to assess survival to emergence, then freeze-core two or
three times, progressing in an upstream centerline of the redd
from just forward of the tailspill, and taking care that the
subsequent sample is not disturbed by earlier samples. Any
sample that does not incorporate one or more large gravel or
cobble particles near the bottom of the core (egg pocket compo-
nents) would be thrown away.
One may ask: "What is the improvement of the suggested
system over McNeil coring after emergence is complete?". The
answer is that one is assured, with the freeze-core, that any
undisturbed egg pocket "floor" is removed from the core, and it
is possible to visually determine that the core has a pocket base
component (large gravel or cobble) present. Without prior
mapping of the pocket, there is still no absolute assurance that
only the egg pocket is incorporated in a freeze-core, but this
system should be much superior to McNeil coring.
Another, better system involves observation of redd con-
struction progress, triangulation or bearings and distances from
a survey stake to the egg pocket, and sampling only in surveyed
egg pockets for as many redds as time and money permit. Survey^
ing could be combined with level-rod work so that elevation of
the bottom of the egg pocket is assessed in relation to a
temporary benchmark. Thus, it would later be possible to
determine whether a freeze-core extracts the full vertical
200 /SECTION V
-------
spectrum of the egg pocket.
Freeze-cores obtained after emergence is complete would
provide representation of gravel condition at emergence time, not
average conditions faced by incubating embryos. We expect that
the percentage of fines in a redd would tend to increase as
intrusion of suspended bedload continues through the period from
redd construction (cleanest condition) to emergence (highest
percentage of small fines). Ringler (1970) provided data that
permit us to estimate in the tributaries that he sampled, redds
(not necessarily egg pockets) constructed in one year will
acquire fines smaller than 0.833 mm at a rate of a little over 1%
per month for the following year. The rate of accretion should be
much greater in the incubation period than in summer in the Alsea
River tributaries studied by Ringler, so a linear calculation
does not reflect reality. We have no data on the rate of fines
intrusion in redds in the northern Rockies.
It is also likely that the microdistribution of fines will
deepen during emergence, as alevins butt toward the surface and
fines drop down into the egg pocket. The combined effects of
upward butting and gravity as several hundred embryos leave an
egg pocket may well move fines lower, but analysis probably would
not be able to detect this change in stratified samples.
The procedure suggested above for assessment of conditions
in the egg pocket in relation to emergence is labor-intensive,
hence expensive. We suggest that the alternative is to continue
to use flawed data of low accuracy and precision and to extra-
polate inappropriate laboratory data to field environments.
A. 2 . Geomejtric mean particle size
At the cost of information on gravel grading provided by the
fredle index, the dg offers another model of survival against
particle size. Again, we do not agree with inclusion of
201 /SECTION V
-------
laboratory data in the model. Even where laboratory data may fit
reasonably well with the function for data from natural redds, we
would not include them, for they would merely contribute to the
"fallacy of misplaced concreteness". In the present context,
this would simply mean that sample error would decrease as
degrees of freedom increase, without logical physical and
biological support.
In Figure A. 4, we have plotted survival to emergence for
natural redds (Koski 1966, 1981, Tagart 1984) against dg. This
relationship was developed for all 39 redds and for data from
Koski only. The latter provided a significant regression and r^
of 0.47. Data of Koski (1966) indicate that when dg drops below
about 5.0, survival declines below 25%.
A procedure for sampling that is the same as that suggested
for the fredle index is needed for assessing
freeze-coring after redd capping.
involving
90
70
BO
50
40
30
20
TO
8 10 12 14 1« 18 20
(tg
Figure A. 4. (Adapted from Tagart 1984 and data obtained from
Oregon State University as acquired by Koski 1966). Survival to
emergence in individual coho salmon redds as a function of d,-..
202 /SECTION V
-------
We also plotted the survival data for 39 individual redds of
Koski (1966) and Tagart (1984) as a function of dg in the mode of
Shirazi et al. (1981). Then we placed the curve of the latter
authors through the data (Figure A.5). The scatter and placement
of individual redd data is so great that the curve of Shirazi et
al. (1981) would not be an appropriate descriptor.
90
70
SO
*O
30
2O
1O
w
IS
20
Figure A.5.
survival to
(From Figure 4 and Shirazi et al. 1981). Curve of
emergence as a function of
q
from Shirazi et al.
(1981) placed through data on survival ana dg for 39 individual
redds.
A.3. Percentage of fines
Use of only percentages of fines in models of survival to
emergence explicitly neglects large particles in the redd.
However, if we use data exclusively from natural redds, we might
assume that fish behavior implicitly incorporate large particles
in the analysis because redds constructed by adults normally lie
203
/SECTION V
-------
in areas that have suitable egg pocket materials. In any event,
fines should be assessed with freeze-coring procedures described
above, making concerns about large particles moot, as large
gravel and cobble of the egg pocket would be a part of acceptable
cores. McNeil cores would also serve for assessment of percent-
ages of fines if they originated in egg pockets.
We plotted percentage of fines smaller than 0.85, 2.0, and
6.0 mm in figures A.6, A.7, and A.8, respectively. These plots,
Figure A. 6. (Adapted from Tagart 1984 and Koski 1966). survival
of coho salmon to emergence in natural redds in relation to fines
smaller than 0.85 mm in redd corings.
Figure A.7. (Adapted from Tagart 1984 and Koski 1966). Survival
to emergence of coho salmon in natural redds as a function of
fines < 2.0 mm.
204
/SECTION V
-------
20 30
% S 8.0 MM
50 60
70 80
Figure A.3. {Adapted from Tagart 1984 and Koski 1966). Survival
of coho salmon to emergence in natural redds as a function of
fines < 6.0 mm.
various forms of which have been used by many investigators,
would lead one to infer that survival of embryos to emergence
declines as the percentage of fines smaller than 0.85 mm
increases. The relationship was significant (p = 0.05) for all
redds combined and for redds trapped by Koski.
Survival was significantly related to percentage of
particles smaller than 2.0 mm and smaller than 6.0 mm for all
redds combined, although the "F" value was somewhat lower for the
larger category of fines.
We also attempted to plot survivals at percentages of
particles smaller than 0.85 and 9.5 mm (Tappel and Bjornn 1983),
again using only data from natural redds. Isolines of survival
could not rationally be placed through the data.
NCASI (1984b) prepared regressions of survival to emergence
for all available works, and we plotted NCASI data on the same
figure (Figure A.9). The regressions have major differences in
205 /SECTION V
-------
placement and slope, only partly because of different categoriz-
ation of fines.
10 2O 3O
PERCENT FINE SEGMENT
Figure A.9. (Adapted from NCASI 1984b). Regressions of survival
against percentages of fines for all available works.
We can conclude that fines decrease survival, but cannot
predict the amount of the decrease in field conditions from a
given increase in fines from available information. The
relationships between fry emergence and percentage of fines that
Stowell et al. (1983) developed (Figures A. 10 and A.11) include
much data that should not, with present knowledge , be used for
field environments. Even the data for field environments are
flawed, as we have abundantly discussed elsewhere. A procedure
like that suggested for assessing survival in relation to the
fredle index in egg pockets would be needed to evaluate effects
of fines, and to better focus laboratory studies so that the
resulting data can serve for field predictions with some
*
confidence.
206
/SECTION V
-------
100
PERCENT FINE SEDIMENT
Figure A.10. (From Stowell et al. 1983). Fine sediment by depth
versus alevin emergence for Chinook salmon.
100
20
PERCENT FINE SEDIMENT
Figure A.11. (From Stowell et al. 1983). Fine sediment by depth
versus alevin emergence for steelhead.
207
/SECTION V
-------
A.4. Permea b i1itv
Platts et al. (1979) showed that permeability is a function
of dg, which itself is related to porosity. Permeability, the
ability of gravels to pass fluid under given hydraulic head, is
related to survival in natural redds. In our opinion, permea-
bility as measured in standpipes quite often has not been
obtained in egg pockets, but rather elsewhere in the redd, which
increased variability in the data. Only the condition within the
natural egg pocket has direct import for embryos incubating
there. This does not mean that conditions in the redd outside
the egg pocket have no influence on conditions in the pocket.
One need only examine Figure II.B.I and figures II.B.10-11 to
support the latter statement.
In Figure A. 12, we plot the available data on survival and
permeability from natural coho salmon redds. The regression is
significant (F = 11.69, 1 and 28 d.f., p = 0.005, r2 = 0.30), The
model does not permit quantitative prediction in egg pockets, but
Figure A.12. (Plotted from data of Tagart 1984 and Koski 1966).
Survival to emergence in natural coho redds as a function of
permeability.
208
/SECTION V
-------
shows that where permeability, as measured by Tagart (1984) and
Koski (1966), declined below about 2,000 cm/h, survival declined
below 25%.
Moring (1982) stated, after showing that permeability
decreased in a stream draining a clearcut drainage in comparison
to unlogged control streams, that this variate may be a more
important indicator of intragravel conditions than are changes in
fine sediments. We believe that if permeability in the egg
pocket could be ascertained, it would provide a measure of gravel
quality useful in predicting survival. Permeability assessment is
easier by far than gravel coring, work on permeability is needed
in natural redds in the northern Rockies, and too little
attention has been directed to assuring that permeability data
came from egg pockets.
Examination of Figure A.2, above, and Figure II.B.2, from
Hawke (1978), illustrates how permeabilities would often derive
from locations outside the egg pocket. In fact, the probability
of driving a standpipe well-point into the egg pocket centrum on
any single sample without knowledge of precise pocket location is
likely to lie much below 50% because of longitudinal, lateral,
and even vertical error.
We suggest, as one sampling alternative, that multiple
measurements, perhaps 10 to 20, should be obtained from each
sampled redd, and that the highest measurement, or perhaps the
top tenth of measurements, might be accepted as egg-pocket
permeability. One would not be absolutely certain that the well
point of the standpipe does not penetrate the egg pocket "floor",
or that the well-point has reached the most important zone in the
egg pocket, but use of the top 10% of measurements should reduce
such errors. Unlike the situation for gravel coring, multiple
permeabilities could be obtained before the redd is capped, but
we see unacceptable risk that driven pipes would loosen gravels
209 /SECTION V
-------
and alter the emergence environment.
Another possibility for the careful field worker would be to
observe redd construction for as many adults as possible, with
tools for triangulating or obtaining bearings and distances to
egg pockets for many redds. Subsequent permeability information
could be obtained from re-located egg pockets only. Egg pockets
can be identified during redd development where water over the
redd is clear enough for the observer to see to the substrate.
Either a semi-permanent truncated standpipe could be driven and
left in the egg pocket, or a standpipe could be driven into the
pocket near the end of the incubation period. Other sampling
possibilities could be developed.
Another advantage of the surveying system for locating egg
pockets is that a level rod could be placed in the completed egg
pocket before the female deposits eggs there, and elevation
recorded in relation to a temporary benchmark. This procedure
would later permit the researcher to use a leveling instrument to
determine when the standpipe well-point lies exactly at the
desired elevation in the egg pocket.
A.5. Electronic probe measurement of pore velocity
A similar set of comments could be made about instantaneous
measurement of pore velocity with electronic probes that measure
water movement independent of vector. Such a measure should cor-
relate with permeability and survival, but must come from within
the egg pocket to reduce data variability and increase relevance.
Multiple measurements within a redd might be easier than coring,
and perhaps one should take the top decile or the top measurement
as indicative of conditions in the egg pocket. Unfortunately, we
can find no survivals to emergence for correlation with probe
data. It is our understanding that the electronic probe has not
worked satisfactorily in Idaho because of difficulty in inserting
the probe in substrata, in part due to relatively large
210 /SECTION V
-------
particles. The tool should be more satisfactory in gravels used
by small adult trout or kokanee, or in spawning areas of pink
salmon.
A.6. Dissolved oxygen assessment
Dissolved oxygen could be used to assess conditions in the
substrate, but measurements should be taken in the egg pocket
centrum, just as for other independent variables used to relate
survival to physical conditions. We base this caveat on Cooper's
(1965) work on water currents within the redd as opposed to
currents in the surrounding undisturbed gravels.
Davis (1975) has suggested that valuable salmonid stocks
should enjoy minimum dissolved oxygen levels of 9.75 mg/1.
Stocks of moderate value would have a minimum of 8.00 mg/1, and
other stocks could have a minimum of 6.50 mg/1. The level for
valued stocks would be 98% of saturation at 0-15 C; stocks of
intermediate and lower value would enjoy 76-79% and 54-64%
saturation, respectively, at 0-15 C (see Table II.C.I).
We suggest that an example of application of the criteria of
Davis to a valued Chinook salmon stock would be to determine if
dissolved oxygen in egg pockets exceeds 98% saturation. If not,
no addition of fines should occur, whether a result of best
management practices or not. As long as oxygen level exceeds 98%
saturation, land management practices could be considered compat-
ible with production of fish. We do not suggest this example as
the only possible application of criteria; merely as one
alternative. Fines lying over the egg pocket centrum could affect
emergence success even where dissolved oxygen remained at 98%
saturation. Dissolved oxygen might provide one of several
criteria.
A.7. Best available information
As noted in earlier report sections, we do not favor extra-
211 /SECTION V
-------
polation of the available laboratory data on survival to field
situations/ for in this we quite agree with Everest et al.
(1986). We have abundantly explained our reservations about the
laboratory data that we reviewed.
Tappel and Bjornn (1983) stated:
"Although embryo survival instreams may not exactly parallel
"equation" predictions (from laboratory data), the equations
should provide a good index of relative changes in survival. As
an example, suppose the steelhead equation predicted that embryo
survival in a stream would decrease from 30 to 60% as a result of
increases in fine sediment in the spawning substrate. Even if
survival in the stream was not 80% for a given substrate before
increases in fine sediment occurred, the 20% reduction in embryo
survival predicted by the equation should be close to the
decrease in embryo survival in the stream. If embryo survival
was actually 50% in the stream instead of 80%, survival of
steelhead embryos should still decrease by 20% as a result of a
given increase in fine sediment."
They further wrote:
"Predicting the consequences of an increased deposition of
fine sediment in spawning areas is an important application of
our research." "Such predictions could be used in the
equations presented here to forecast the effects of human
activities on steelhead and Chinook salmon embryo survival."
Headers who have patiently followed the logic trail through
the thickets of the literature analyses in Section II of our
report will not be surprised that we disagree with the foregoing
assertions. We do not consider it appropriate to predict the
percentage change in embryo survival in field situations from
laboratory tests that did not duplicate egg pocket structure and
gravel size distributions. The only natural egg pocket for which
dg has been reported for chinook salmon (Platts et al. 1979) had
a d— of 32 mm, 49% higher than the highest dg (21.5 mm) used by
Tappel and Bjornn (1983), about three times larger than the
average dg (11.1 mm), and eight times the lowest dg (4.0 mm).
The data of Tappel and Bjornn (1983) reflect good laboratory
research. They can be used to infer that the mix of fines, as
212 /SECTION V
-------
indicated by weight percentages of fines of two diameters, has an
effect on survivals of steelhead and Chinook salmon embryos in
natural egg pockets.
In spite of the foregoing arguments, we realize that a
substantial body of information in laboratory and field
conditions shows that increases in percentages of fines in
gravels negatively affects survival. We consider it appropriate
to summarize it here.
From Tappel and Bj ornn (1983) and other sources, we provide
Tables A.1-A.3, which summarize what we consider to be the best
available information on intragravel survival to emergence in
salmonids of the northern Rockies in relation to various
potentially useful tools. We consider the tables as a "state of
the art" information summary.
Table A.I. Best available information on survival to emergence
of salmonids in certain intragravel conditions. Sources include
Tappel and Bjomn (1983), Davis (1976), Irving and Bjornn (1984),
Lotspeich and Everest (1981), and Shirazi et al. (1981), McCuddin
(1977), and materials that we developed in our report.
P.O.
Fines indicesa
<6.3mm <0.85and <9.5mm
Fredle
80% survival
Chinook salmon 98 10,000
Steelhead
Rainbow trout
Kokanee
Cutthroat trout
Bull trout
20
98
98
98
98
98
10,000
10,000
10,000
10,000
10,000
20
20
10
10
25
Use Fig. II.C.23 15
at 80% survival
ii ii 15
i< ii 15
n 11 15
" " 15
Use cutthroat data 15
in Fig. II.C.23 at
80% survival
6
6
6
6
6
a - D.O. in % saturation, K in cra/h, fines in % by weight, dg in
mm, Fredle with no dimension.
213
/SECTION V
-------
Table E.2. Estimated dose:response relationships from best
available information on intragravel survivals to emergence
between 80% and 25%. Each indicated increment or decrement in an
independent variable would correspond with a 10% decrease in
survival to emergence. Sources include Tappel and Bjornn (1983),
Davis (1976), Irving and Bjornn (1984), Lotspeich and Everest
(1981), and Shirazi et al. (1981), McCuddin (1977), and materials
that we developed in our report.
Increment (+) or decrement (-) causing 10%
reduction in survival between 80% and 25% survival3
Species
P.O.
K
<6.3mm <0.85 and <9.5mm
Chinook -5% -1,600 +4%
Steelhead -5% -1,600 +4%
Rainbow trout -5% -1,600 -t-5%
Cutthroat trout -5% -1,600 +3%
Kokanee -5% -1,600 +4%
Bull trout -5% -1,600 +3%
See Fig. II.C.23
tt
I?
it
11
-1.6
-1.6
-1.6
-1.6
-1.6
-1. 6
Fredle
-1.0
-1.0
-1.0
-1.0
-1.0
-1.0
a - D.O. % saturation, K in cm/h, dg in mm, Fredle w/o dimension.
Table E.3. Best available information on intragravel conditions
below which survivals to emergence would be equal to or lower
than 25%. Sources include Tappel and Bjornn (1983), Davis
(1976), Silver et al. (1963), Phillips and Campbell (1962),
Sowden and Power (1985) , Shumway et al. (1964), Irving and
Bjornn (1984), Lotspeich and Everest (1981), and Shirazi et al.
(1981), McCuddin (1977), and materials that we developed in our
report.
Fines indices3
Species
Chinook
Steelhead
Rainbow trout
Cutthroat trout
Kokanee
Bull trout
D.O.
65%
65%
65%
65%
65%
65%
K
2,000
2,000
2,000
2, 000
2,000
2, 000
<6.3mm <0.85 and <9.6mm
45%
45%
35%
30%
45%
30%
a - dimensions as in Table E.I.
See Fig II.C.23
tr n
f*
7
7
7
7
7.0
F_redle
1.2
1.2
1.2
1.2
1.2
1.2
It would be inappropriate to place confidence limits about
the estimates in the foregoing tables, even if it were possible.
The data come from a mix of sources, each of which has an unknown
214
/SECTION V
-------
statistical distribution, and from situations with different
experimental and sampling biases.
215 /SECTION V
-------
B. SUBSTRATE FINES AND EFFECT ON REARING DENSITIES
We found little information that indicates the quantitative
effects of sediments on actual rearing densities in the field for
most species of the northern Rockies. We will discuss predictors
in no preference order in the material that follows.
B.I. Embeddedness
Evidence that embeddedness affects densities of juvenile or
adult salmonids in field environments in summer is only
moderately convincing. Munther and Frank (1986 a,b,c) reported
very low r^ values for 53 relationships between embeddedness and
fish densities. Konopacky et al. (1985) provided data on
embeddedness and fish densities in various reaches of the same
stream (thus reducing, although probably not eliminating, data
scatter caused by highly variable seeding, different limnological
features, etc.). It appears that his data were obtained by a
visual estimate rather than with the careful techniques of Burns
and Edwards (1985) , hence should be considered less useful. His
results lend no support for a negative relationship between
embeddedness and Chinook salmon density.
Thurow and Burns (unpublished), provided fish densities and
embeddedness data that can be used to develop a negative rela-
tionship between embeddedness and fish densities, and particu-
larly maximum densities of age 0 Chinook salmon, but we cannot
accept their models as quantitative predictors of fish density
because streams of different order, gradient, size, and seeding
levels were included without stratification. Their model for
maximum density of age 0 chinook salmon included only stream
habitats in Chamberlain Creek and the South Fork Salmon River,
hence may reduce some of the effects of extraneous environmental
variables.
216 /SECTION V
-------
Data obtained from C. Johnson (BLM) and from Gamblin (1986)
are only moderately convincing and of mixed message in regard to
effects of embeddedness on salmonid densities.
Stowell et al. (1983) used experimental data of Bjornn et
al. (1977) to prepare models of effects of embeddedness on
densities of age 0 steelhead, age 0 Chinook salmon, and age 1
steelhead (Figure B.I). The relationships for age 0 chinook and
SUMMER REARING CAPACtTY -RUN"
AgeO Chinook
C
Y • 2.4B — 0.044M +
£
LLJ
-------
density of 90 fish per 100 m2 at the same embeddedness.
In fact, Bjornn et al. should not be used for prediction of
field densities of Chinook salmon at any embeddedness level. At
0% embeddedness, the densities of Chinook reported by Bjornn et
al. (1977) and used by Stowell et al. are about 250 fish/100 m2,
four times higher than the maximum density found anywhere by
Thurow and Burns (unpublished) or by Petrosky and Holubetz
(1986). The effect of these excessive densities on the responses
of juveniles to embeddedness is unknown. Stowell et al. (1983)
attempted to normalize for the high densities of fish in the
laboratory environments of the artificial streams by treating
maximum density (250 fish/100 m2) as a fish density of 100%
(Figure B.I), and predicting fish densities in percentages of
maximum density.
Stowell et al. (1983), based on experimental data of Bjornn
et al. (1977), would predict age 0 steelhead densities of 800
fish per 100 m2 at 50% embeddedness and 1000 fish/100 m2 at 0%
embeddedness, many times higher than densities found in natural
environments. Densities of age 1 steelhead are predicted as 150
fish/100 m2 at 50% embeddedness and over 300 fish/100 m2 at 0%
embeddedness. In fact, the only density for steelhead that
appears close to that to be expected in the wild was at 100%
embeddedness (about 30 age 1 fish/lOOm2). Stowell et al. (1983)
attempted to normalize fish densities as percentages, with
maximum densities at an embeddedness of 0%. However, we have no
way of knowing the effect of the inordinately high densities
reported by Bjornn et al. (1977) on reaction of fish to
embeddedness in the wild.
In summary, we reject the predictions of fish density from
embeddedness in the artificial environments of Bjornn et al.
(1977), even where densities were normalized as percentages. Had
densities been more in concert with those in field environments,
218 /SECTION V
-------
and had other variables been handled differently, (for example if
juveniles emerged into the channels, or had the investigators
seined wild fish of smaller size and placed them in the channels
at appropriate densities earlier in the summer) the effects of
embeddedness could have differed markedly. Had other aspects of
the trough environments offered compensatory features probably
available in the wild, such as more normal pool/riffle structure
and more natural cover, other results, hence different models in
Figure B.I, might have been developed.
While we emphasize that we see much scientific merit in
mechanistic work in artificial channels, we point out, as for
laboratory study of embryo survival to emergence, that experi-
mental circumstances govern outcome, certainly in quantitative,
and perhaps in qualitative senses. We feel it is not appropriate
to apply the available information in quantitative ways to
natural stream systems.
We suggest that field evaluations of embeddedness effects on
fish density should be undertaken in carefully-documented stream
strata and fully-seeded fish populations. This should entail
collection of data on habitat variables and addition of approp-
riate numbers of juveniles to test areas early in the summer to
assure full seeding. Juveniles from nearby areas may be seined
or electro-sampled for addition to test strata. Stepwise
multiple regressions should help to assess the significance and
incremental effect of embeddedness on r2, but the main point is
that field experiments at full seeding in several environmental
strata are needed to assess effects of embeddedness on density of
fish.
Bjornn et al. (1977) recognized the importance of field
investigations, examining habitat variables and fish densities in
conjunction with experimental additions of fines in natural
streams of the Middle Fork Salmon River. The authors were unable
219 /SECTION V
-------
to show an adverse effect of sediment on abundance of fish except
where sediment so reduced pool volume that living space
decreased. Their work should be extended to a much longer stream
segment.
B.2. Effect of fines jgn rearing densities of fish
Konopacky et al. (1985) visually estimated percentages of
fines smaller than 4 mm in diameter and total density of Chinook
salmon, steelhead, and cutthroat trout. At least to percentages
of fines in excess of 40%, we could find no negative relation-
ship between fines and densities (to the contrary, the
relationship appeared positive, although we did not calculate
regressions).
We used the data of Petrosky and Holubetz (1986) to
calculate a weakly negative relationship between percentages of
fines smaller than 5 mm and densities of age 0 Chinook salmon.
The overall r2 was only 0.058. When we stratified the data by
stream gradient, only the 1-2% gradient stratum had a signifi-
cantly negative relationship between fines and fish density (r2 =
0.084) .
Bjornn et al. (1977) added granitic fines to field environ-
ments in two tributaries of the Middle Fork Salmon River, but
could not show an effect of fines until pool volume was greatly
reduced. No relationship could be shown between percentage of
fines (<6.35 mm) in riffles and density of fish in pools just
downstream. Konopacky (1984) found no correlation of Chinook
salmon density in pools and fines in the riffle upstream.
We draw the very obvious inference that a superabundance of
fines will reduce density of rearing fish. "Superabundance"
means enough fines to reduce living space and overcome ability of
the stream to process sediment recruitment so that food
production on riffles is smothered and sharply declines.
220 /SECTION V
-------
Regretfully, the model for a relationship of lesser amounts of
sediment and fish density has not been developed, nor does
embeddedness sampling to date offer a satisfactory surrogate.
The model would require data on pertinent habitat variables as
well as substrate condition, and assurance of full seeding.
Stratification by stream gradient and geology would be desirable.
B.3. Substrate score and fish density
Grouse et al. (1981) related coho salmon production to
substrate score in both spring (r2 = 0.75) and summer (r2 = 0.90)
in artificial stream channels. The scoring function contains
unknown elements of cover and energetics. Shepard et al. (1984)
related numbers of bull trout per 100 m2 to substrate score (r2 =
0.40) in 26 stream reaches in the Swan River system in Montana.
No data on other habitat features or seeding were offered.
Substrate score, an amalgam of substrate particle size and
embeddedness, should offer more information than embeddedness
alone or fines alone, and is worth pursuing in future studies of
fish density in relation to habitat variables. This variate
currently has insufficient documentation for application to the
species of the northern Rockies.
B.4. Best available information
Available information on effects of fines on salmonid
rearing densities does not permit a broad statement on effects of
embeddedness level or various percentages of surface fines. A
very conservative view of the data would be to state that rearing
densities are often lower at embeddedness levels greater than
50%. A conservative view in habitat protection, on the other
hand, might state that any embeddedness level greater than 25% in
rearing areas risks loss of winter habitat in interstices and
should be avoided until better information becomes available.
221 /SECTION V
-------
C. MACROINVERTEBRATE RESPONSES TO FINES
Response of aquatic macroinvertebrates to fine sediments has
received considerable research attention. We will discuss embed-
dedness and percentage of fines as predictors.
C.I. Insect density as a function, Q_f_embeddedness and fines
Mean insect density was shown by Bjornn et al. (1977) to
maximize at an embeddedness level of two-thirds in artificial
stream channels. In natural stream areas, these authors cleaned
plots and compared insect abundance there to uncleaned controls.
Final abundance of insects (45 days later) was greater on
undisturbed sites than on cleansed plots. The work by Bjornn et
al. appears to show that a substrate without any fines offers
less habitat niches and diversity than one with some fines around
the bases of larger particles.
Munther and Frank (1985 a,b, c) were unable to show that
embeddedness explained aquatic insect biomasses in 20 sample
locations in Montana. Even when various stratifications by
insect species assemblages were attempted, statistical analysis
could "explain" little variation caused by embeddedness.
Although Bjornn et al. (1977) offered data on insect density
in artificial stream channels, their data should have more appli-
cability to field situations than is the case for fish. Coloni-
zation of artificial stream substrata occurred naturally, reduc-
ing or eliminating the problem of inappropriate starting dens-
ities, and the animals involved were very small relative to the
channels, thus reducing problems with cover, and suitability of
sympatric animals. However, the mix of gravels in the channels
may not apply to natural streams. The substrate in riffle
sections consisted of a layer of cobbles 6.35 to 12.60 cm in
diameter over a 0.4 m layer of gravel (size not noted), and that
222 /SECTION V
-------
in pools consisted of 0.3 m boulders placed on the bottom. Thus
we cannot conclude that an optimum substrate in the wild should
have an embeddedness level of two-thirds.
Bjornn et al. (1977) correlated benthic insect densities
with the level of cobble embeddedness at Knapp Creek in 1974-75.
Embeddedness categories included unembedded, one-fourth, one-
half, three-fourths, or fully embedded. The authors found a low
correlation between insect densities and embeddedness for
mayflies and for the caddis, Brachycentrus sp., but not
consistently positive or negative. For example, Brachycentrus
abundance correlated positively with embeddedness in riffles to
which sediments were added in 1974, but significantly and
negatively in both test and control riffle areas in 1975. For
all species combined, embeddedness correlated significantly and
positively with abundance of insects in 1974 on test riffles, but
negatively and non-significantly on test riffles in 1975.
Cleansing of test plots in Elk Creek, another natural
system, led to increased insect abundance after several weeks
(when embeddedness had reached about one-fourth), in comparison
to control areas that were about three-fourths embedded. No
means are available to compare effects of two-thirds embeddedness
with these data.
Fines have been shown to correlate negatively with abundance
of aquatic insects (Cederholm and Lestelle 1974) . An increase
from 7% up to about 9% percent fines < 0.84 mm appeared to cause
a 50% reduction in abundance of benthic insects, although hidden
correlates, such as larger fines, may have been involved. The
fines were evaluated in core sampling, hence surface fines may
have increased relatively much more than subsurface fines,
Bjornn et al. (1977) state that a two-thirds embeddedness
level would correspond with a core sample containing 30% or more
223 /SECTION V
-------
sediment, a situation in which all of the interstitial spaces in
the gravel are filled with fines. The data sets of Cederholm and
Lestelle and of Bjornn et al. appear to conflict, probably
because the two works used different classifications of fines.
Saunders (1986) reported that insect standing crops declined
where embeddedness was high, and that insects were less abundant
where embeddedness exceeded 3/4 than where it was between 1/2 and
3/4.
C.2. Insect drift as a function of embeddedness and fines
Embeddedness did not affect drift of insects in artificial
channels studied by Bjornn et al. (1977). Konopacky (1984)
provided data that indicated that a substrate of gravel produced
more drift of sizes appropriate for use by fish than did a sand-
pebble substrate, although more absolute numbers of insects
drifted from the latter.
No useful relationship between substrate type and
invertebrate drift is available for predicting effects of fines
on the insect drift resource in streams of the northern Rockies.
C.3. Best available information
The thrust of the available information is reasonably
consistent in showing that at high embeddedness, insect abundance
declines. It appears that the embeddedness level at which
insects decline in abundance is about 2/3 to 3/4. Although no
relationship between drift and fines is clear, logic leads us to
infer that where embeddedness increases sufficiently to reduce
insect densities on and in the substrate, drift density/m3 of
water should also decrease.
224 /SECTION V
-------
D. EFFECTS OF FINES ON WINTER HABITAT OF SALMONIDS
D.I. Embeddedness and winter habitat
Stowell et al. (1983) used data from Bjornn et al. (1977) to
develop models of the relationship between embeddedness level and
winter carrying capacity (Figure D.I). Intuitively, we conclude
WINTER CAAflYIKa CAPACITY TOOL"
• l
A««>CMnoo*
v
O
•
tMBEOOTONtSS lfVfl.{PCTl
Figure D.I. (From Stowell et al. 1983). Winter carrying
capacity of pools and substrate embeddedness, for age 0 chinook
salmon, steelhead, cutthroat trout, and age 1 and 2 cutthroat
trout.
that more winter habitat in the substrate exists at 0%
embeddedness level than at 50% or, certainly, at 100%
225
/SECTION V
-------
embeddedness levels. However, we do not believe that data from
artificial streams at Hayden Creek (Bjornn et al. 1977) can be
applied to natural stream areas.
The first obstacle to acceptance of Figure D.I, from Stowell
et al. (1983), is that the density of fish in artificial stream
channels at Hayden Creek was excessive at test initiation.
Figure D.I would predict densities at 0% embeddedness of 250
chinook/100 m2, over 700 age 0 steelhead/100 m2, 1,600 age 0
cutthroat trout, and over 100 age 1 and 2 cutthroat trout. All
fish were added to the channels when temperatures led to immedi-
ate entry into the substrate. Had initial densities been moder-
ate, densities after 5 days might have been different, even at
full embeddedness.
The second problem is that 5 days may or may not bring fish
numbers into equilibrium with the available habitat, and an
interaction between time and embeddedness level would destroy the
models in Figure D.I. The third major difficulty is that a
different substrate mix or presence of other habitat features
different from those in the artificial streams, might well affect
the functions in Figure D.I.
The Aforegoing three problems lead us to reject Figure D.I as
a predictor of the height of the 100% point on the right axis of
each model in the figure, and as an indication of the shape of
the functions.
We have no doubt that functional relationships exist between
embeddedness and winter holding capacity of the substrate for
salmonids, and that those relationships differ by fish size and
perhaps by species. However, quantitative prediction is cur-
rently not possible. We assume that winter capacity for fish in
the substrate is greatest at 0% embeddedness level, and that it
must decline at some embeddedness betweeen 0% and 100%.
226 /SECTION V
-------
Research effort should be devoted to examination of the
effects of various substrate types, starting and ending fish
densities, and fish size and species on the shape of the
relationship between winter holding capacity and embeddedness. It
may be desirable to undertake this work in natural stream
channels in which seeding upstream and in the study area is known
to be adequate (perhaps based on late summer snorkel index monit-
oring) , then to modify substrate condition and rely on extraction
of fish with electrofishing gear within a few weeks after temper-
atures decline to levels that cause entry of fish into the sub-
strate. Artificial stream channels may play a role, but multiple
treatments and semi-natural conditions and surety of appropriate
fish behavior will mandate few replicates and treatments because
of costs. On the other hand, one season of field work could
define fish density and embeddedness functions sufficiently to
permit preliminary field use.
Concomitant research should establish the relative survival
of fish that remain in the substrate in rearing areas and that of
fall "migrants" that move out of rearing areas into larger
streams. Recent development of technologically-advanced tags
that can be detected in induction coils without handling of fish
by humans offer promise for use on anadromous fish in the Snake
River basin, where all juveniles pass Lower Granite and Little
Goose dams. Work of this type has been proposed and will be
implemented on 1987 (R. Thurow, personal communication).
D.2. Best available information
Bjornn et al. (1977) reported that at an embeddedness level
of about 2/3, corresponding to about 30% fines in the substrate,
all of the interstices in the substrate were filled with fines.
This would completely remove rearing areas with such high
embeddedness from use by overwintering fish. Reduction in winter
habitat must occur at embeddedness levels somewhere between 0%
227 /SECTION V
-------
and 66%.
Best available information indicates that substrate
interstices are needed by overwintering salmonids in the northern
Rockies, and that fines decrease availability of interstices.
Risk, uncertainty, and prudence suggest that any incremental
fines above current conditions would be likely to reduce
overwinter survival of salmonids.
228 /SECTION V
-------
E. MONITORING
Where the environment already contains large amounts of
intragravel and surficial fines, and where redds and spawning
gravels cannot be monitored, management agencies may have to
adopt other approaches. To evaluate long-term changes in
substrate conditions, the manager may have to employ a
combination of McNeil coring, embeddedness, substrate scoring,
and ocular estimates of surface fines.
Earlier in this section and elsewhere in our report we
extensively criticized existing data on intragravel composition
as they apply to embryo survival. These and other criticisms do
not imply that McNeil corings, embeddedness, ocular estimates of
fines, or substrate scoring will not serve to monitor long- or
short-term changes in substrate composition in actual or
potential spawning and rearing areas. On the contrary, these
techniques may work very well as indices of habitat quality. Our
argument is with quantification of biological effects of fine
sediments, not with habitat monitoring.
Shortage of funds and trained manpower for habitat research
and evaluation makes it imperative that data collection in
monitoring programs have a clearly-defined purpose and that it
receive regular scrutiny of both biologists and statisticians so
that waste may be avoided. In very heavily sedimented streams,
where spawning gravels are buried, or seeding levels so low that
redds do not occur, the advice of a hydro-geomorphologist may be
required before a monitoring program is developed.
229 /SECTION V
-------
VI. GENETIC RISK, BIOLOGICAL COMPENSATION AND LIMITING FACTORS
A. GENETIC STRESS ON POPULATIONS
Use of tools for evaluation of sedimentation need not
require a presumption of fish population limitation in a
numerical sense by fines. Although erosion and sedimentation are
natural processes, and all streams in the northern Rockies a
product of millenia of process operation, acceleration of erosion
processes and concomitant stream processing of fines should be
viewed as a genetic risk for fish stocks.
The genetic material represented by indigenous fish species
and stocks of the northern Rockies has adapted to evolving and
instantaneous climatic, edaphic, and biological constraints and
potentials. Risk accompanies acceleration of change and addition
of stress in this system, particularly in the case of resident
and anadromous stocks already subjected to heavy fishing pressure
(usually a discriminant stress), engineered obstacles {often
causing indiscriminant stress), and manipulation by resource
managers (discriminant and indiscriminant stresses).
In the case of anadromous fish stocks of the Columbia River
system, the Northwest Power Act has brought into being a new
bureaucracy, the Northwest Power Planning Council, with a mandate
to protect and enhance salmon and steelhead stocks of the basin,
and to bring fish into parity with power generation. The Council
has stated that anadromous fish losses associated with hydropower
dams amount to 5.6 - 10.1 million adults, and set tentative near-
term goals of doubling mid-1970s runs of 2.5 million fish to 5
million.
It has been reasonably estimated that the natural environ-
ments of the Columbia River system can produce a total of one
million adult salmon and steelhead. Thus, hatcheries will have
230 /SECTION VI
-------
to produce the complement of the Council-ordered increase, or
about 4 million fish. Genetic change is a concomitant of intense
management by hatchery supplementation with accompanying fish-
eries . Thus, natural stocks of the upper basin, often those in
the steeper areas most subjected to accelerated erosion from road
building and logging, will face increasing selective pressures
downstream. No one knows if selective pressures from stresses
caused by sedimentation would be additive, multiplicative, or co-
vary with other selective forces in the life cycle.
B. ROLE OF BIOLOGICAL COMPENSATION
Biological compensation is the adjustment of birth, death,
and growth in reaction to fluctuations in recruitment and
mortality (Nicholson 1954). Ricker (1954), Ricker (1958), and
many others have shown by means of such models as curves of stock
and recruitment that fish stocks have high survival when
population numbers are reduced by fishing or by density-
independent factors, and to have low survival when numbers become
excessive when agents of mortality are reduced in intensity, or
when recruitment is excessive (Fraser 1969). A similar pheno-
menon can be demonstrated for growth; that is, fish in popul-
ations at low density tend to grow faster than those at high
density (Fraser 1969).
It is useful to offer an example of the role of compensation
in the present context. If fine sediments increase in the
substrate as a result of accelerated erosion in the watershed, in
turn resulting in reduced intragravel survival of salmonid
embryos, we should expect to see emerging survivors of the
incubation period grow more rapidly and survive at higher rates,
thus "compensating" for higher embryonic mortality. (Although
compensation operates at the individual fish level, biologists
often think of compensation as a population phenomenon, as if the
population were a super-organism.) If intragravel mortality is
231 /SECTION VI
-------
high enough, even compensatory mechanisms will not serve to
adjust for it and the population will be depressed (limited).
On the other hand, if fine sediments not only reduce
intragravel survival and emergence, but also reduce abundance of
macroinvertebrates and fill interstices used by salmonids for in-
stream cover in the rearing period and for overwinter hiding
places, compensation must be less effective, and carrying capa-
city would be reduced.
C. LIMITING FACTORS
Limiting factor theory has roots in the classic ecological
literature (Pearl 1926, Solomon 1949, Volterra 1931, Lack 1954,
Nicholson 1954), These works, as well as basic ecology texts,
describe relationships among birth rates, death rates, population
distribution, innate capacity for increase, empirical rate of
increase, and discuss population legislation by limiting factors.
Figure A.I illustrates birth and death rates in a regulated
population, population growth in relation to density, and
approach of population size to equilibrium (often designated as
"K") with time. At population densities that exceed K, death
rate will, in the long run, exceed birth rate. In the short run,
loss of animals to mortality will bring the population toward K.
In populations smaller than K, birth rates will exceed death
rates in the long run. In the short run, survival rate will
increase until the population reaches K.
Populations may oscillate about K as a result of various
time lags (lag between ovulation and emergence, lag in predator
reaction to prey abundance, etc.), and K itself may vary from
year to year in response to climatic or biologic factors. Hence
K may be considered a long-term average rather than an instant-
aneous absolute. The limiting factors that determine K may vary
232 /SECTION VI
-------
can
•; o •
X 0.13
A. _
c
•
•o
c
o
3
a
o
a.
Population density
Population density
TlfTlB
Figure A.I. (From Ricklefs 1973). Birth and death rates in a
regulated population (top), population growth in relation to
density (middle), and approach of population size to equilibrium
(K) with time.
with the seasons.
Nicholson (1954) offered elegant examples of limiting
factors in operation. He studied sheep blowflies (Lucilia
cuprina) in laboratory cultures. He provided everything in excess
except one requisite; water, sugar, space in excess for example,
and larval food in limited quantity, or all factors in excess
except food for adults. Under these circumstances, provision of
double the limiting requisite allowed the population to double.
In natural ecosystems, all factors but one are rarely in
perfect excess, and doubling supply of a limiting factor would
not normally double the population size. Another factor
generally would limit in this circumstance before the population
doubled.
233
/SECTION VI
-------
Mason (1976) provided an excellent example of "lifting the
lid" of food limitation by feeding marine invertebrates to coho
salmon in a stream on Vancouver Island. In comparison to unfed
stream sections, fed sections produced larger coho salmon with
high lipid reserves, and test populations 6- to 7-fold more
numerous in terms of fish per 100 m2 by late September. By
February, test (fed) populations of coho salmon had declined to
about 25 fish per 100 m2, a density close to that of nearby
natural populations not subjected to testing. Mason concluded
that winter hiding spaces limited carryover populations of pre-
smolt coho. This study illustrated that the limit on smolt
output may be set by a suitable space mechanism in winter.
Although it is generally conceded that coho salmon abundance
is determined by suitable space in the summer-fall rearing period
(Chapman 1962, 1965), the possibility, remains that behavioral
mechanisms that regulate density in the rearing period may also
have an underlying constraint of over-winter space availability.
Hunt (1969) showed that resident trout density in a Wisconsin
stream could be increased by provision of winter cover.
Large woody debris (LWD) has been shown to exert a strong
effect on abundance of coho salmon (House and Boehne 1986), and
Heifetz et al. (1986) showed that LWD also increased winter
habitat for coho salmon, Dolly Varden char, and steelhead. Lid-
lifting sometimes has unforeseen effects. Sih et al. (1985)
examined studies of lid-lifting through predator control, noting
two unanticipated effects: (1) removal of keystone predators may
increase density of a dominant prey, decreasing diversity, hence
stability, and (2) removal of top predators may release a middle
level predator and thus reduce numbers of primary consumers; a 3-
trophic-level effect. Forty percent of all predator control
studies resulted in unforeseen effects in the system; some
negative, some positive.
234 /SECTION VI
-------
McFadden (1969) defined population regulation as operation
of density-dependent factors, as distinct from population
limitation, which refers to operation of density-independent
factors. In other words, density-independent factors legislate
K, and regulatory factors govern populations within the K
constraint. A population so-governed would be termed "K-
selected" in the evolutionary sense (Pianka 1970). The salmonid
populations in streams of the northern Rockies fit this termin-
ology. McFadden (1969) hoped to de-emphasize the guest for
single limiting processes or factors and to increase interest in
a flexible scheme that would synthesize information about a
variety of population processes and environmental factors.
As an example of a flexible viewpoint, McFadden (1969) noted
that where a fish stock is very abundant (potential overseeding),
behavioral interference leads to retention of part of the egg
complement or to destruction of previously-completed redds by
late spawners, and to an increased proportion of adults spawning
in less-suitable areas where redds may be exposed to subsequent
freezing, dessication, scouring, or siltation. Where great
numbers of eggs are deposited, oxygen depletion may occur in
redds.
Where suitable spawning gravel is available in limited
quantity (or where large numbers of fish use the gravels),
mortality in the spawning and incubation phases will thus be
density dependent (and will act as a population regulator).
Where extensive areas of gravel of poor quality are present
relative to fish abundance, mortality during the spawning and
incubation stages will be density independent (Allen 1962).
Hence in the one case, spawning gravel may limit population
recruitment (as often appears to be the case for sockeye salmon
and pink and chum salmon in Canada and Alaska) . In the second
case another life history phase may limit (as often appears to be
the case for species with extended stream rearing phases, such as
235 /SECTION VI
-------
spring and summer Chinook salmon, steelhead, resident rainbow
trout, bull trout and cutthroat trout).
Most areas of the northern Rockies do not appear to suffer
from excessive spawner densities at this time, hence are unlikely
to support populations of fish that are regulated by density
dependent factors in the reproductive phase. Kokanee may be an
exception, where limited spawning gravels support large lake
populations. More usually, one should expect density-independent
survivals in the reproductive phase, as a result of flooding, ice
scour, or widespread siltation.
P. Bisson (unpublished draft report 1986) warned of the many
pitfalls in the process of identifying limiting factors. He
specifically listed (1) excessive reliance on professional
judgement, (2) extrapolation in space and time, (3) over-
simplification of complex ecological situations, (4) exclusive
focus on one aspect of life history, (5) failure to consider
critically important factors. Examples of these problems were
offered by Bisson. In the context of the present review and
synthesis, item (2) is exemplified by erroneous extrapolation of
laboratory studies to field environments by workers seeking tools
for management and regulation. Item (3) is typified by evalu-
ations of short stream segments without knowing whether the
carrying capacity of a segment is limited by some factor in the
study section or by factors removed in time and space. Fishery
workers have generally confined field studies to daytime in
favorable weather periods, thus risking pitfall (5) . Instances
of pitfall (1) are so numerous as to require no example.
Laboratory studies of emergence success embody elements of
pitfall (3). Bisson's paper should be widely read.
Assessment of limiting factors for salmonid stocks in the
northern Rockies remains an art form rather than a scientific
system. Peters (1986) forthrightly discussed work in Montana on
236 /SECTION VI
-------
effects of forest practices on stream ecology by stating:
"The sampling of fish populations in this study has raised more
questions than the effort has answered so far. Recent literature
searches and on-going fisheries work further clouded our results
with uncertainty as to what exactly we are measuring in the
limited fish population sections. Are we measureing the carrying
capacity of that section or the result of some limiting factor
either spatially or temporally removed from this site? The
ramifications of this difficulty is that our sampling design may
only show impacts if the summer habitat contains the limiting
factor(s) on the particular population. Studies of cutthroat
trout in the Flathead River system indicate that some fish enter
the interstitial areas in the substrate while some adult fish
move to pools in the larger river system (personal communication
with Pat Graham)."
Professional judgement, empirical experience, and the
knowledge base for salmonid ecology must be applied with tailor-
ing to each situation, whether the biologist wishes to manipulate
nature for improved fisheries or merely wishes to understand and
maintain the status quo.
In the face of uncertainty and a limited and fragmentary
knowledge base, we contend that resource managers should attempt
to maintain maximum ecological diversity as a hedge. In the
context of the present report, this concept means that the
manager would seek to increase natural diversity in habitats
damaged by sediments. This concept means, in the case of the
intragravel environment, minimal (or no) introduction of fines to
the existing gravel matrix. It means making every reasonable
effort to reduce sediment recruitment from basin development.
The rearing phase of salmonid life history appears to be
somewhat plastic in regard to effects of surficial fines on fish
abundance and growth, but it seems prudent not to increase fine
sediments. This prudence would conform to the concept of
maintaining diversity in stream systems.
The overwintering phase of salmonid life history requires
237 /SECTION VI
-------
interstices for hiding. Any increment of fines that decreases
abundance and volume of interstices constitutes a risk. It would
be imprudent of land managers to decrease habitat diversity by
permitting more fines to enter winter cover. The information
base is simply too scant to justify any risk, especially where
fisheries of high value are involved.
238 /SECTION VI
-------
VII. PREDICTIVE TOOLS, MANAGEMENT, AND REGULATORY UTILITY
In the following material we define predictors as
independent variables that can serve to quantify effects of fines
on fish and macroinvertebrates. We define threshold values as
levels of independent variables beyond which serious damage to
fish or macroinvertebrate populations would occur.
A. PREDICTORS FOR FISH
We found no functional predictors that would serve
environmental regulators in evaluating quantitative effects of
sediment on the natural incubation, rearing, or wintering phases
of salmonid life history in the northern Rockies. Some
predictors have promise that can be met by additional research.
A considerable body of laboratory and field research has
been devoted to effects of fines on the intragravel environment
in the northern Rockies, particularly in the Idaho batholith, and
in the Pacific Northwest. It is clear from the available infor-
mation that increased amounts of fines that reduce geometric mean
particle size, fredle index, permeability, or dissolved oxygen in
intragravel water also tend to reduce survival of embryos to
emergence. Inasmuch as we cannot support quantitative predictors
for use in field environments, either in the northern Rockies or
elsewhere, we offer no threshold values. Uncertainty and risk
factors lead us to suggest that any incremental increase in
intragravel fines in egg pockets should be avoided until
functional relationships can be developed.
The only threshold values that have been developed for
intragravel parameters in relation to fishery values are those of
Davis (1975) for dissolved oxygen. These relate to conditions in
the egg pocket because they were developed directly from
investigations of embryo survival in various oxygen concen-
trations, rather than indirectly (through indexing fines in the
239 /SECTION VII
-------
redd or permeability, for example) . We see no reason why these
cannot be adopted. We suggest that threshold levels of minimum
dissolved oxygen content in intragravel water in the egg pocket
could be set at 98% of saturation at temperatures of 0-10 C for
high value fisheries, 76% of saturation at 0-10 C for fisheries
of moderate value, and 54-57% of saturation for fisheries of
lower value. Judgement and negotiation would have to set fishery
value categories. Chinook salmon, steelhead, and westslope
cutthroat trout would nearly always fall in the high-value group.
Abundance of salmonids, with the exception of brook trout,
appears loosely and negatively correlated with embeddedness level
and percentage of surficial fines. Embeddedness appears prom-
ising as a tool for evaluation of effects of sedimentation on
rearing phases of salmonids. No threshold value can be set. We
can make no recommendation regarding surficial fines as they may
affect salmonid rearing densities.
The relationship of surface fines and embeddedness to
overwintering success in salmonid populations of the northern
Rockies has not been guantified. The body of available labor-
atory and field information leads us to infer that maximum
crevice availability (minimum embeddedness) should provide the
greatest overwintering capacity. No threshold value can be set
for embeddedness as it affects overwintering. In view of uncer-
tainty and risk factors, we believe it prudent and conservative
on the side of salmonid gene pools to permit no man-caused
incremental embeddedness until functional relationships can be
established, at least in areas that support fisheries of moderate
to high value. How these fishery values may be established lies
beyond the scope of our report, but we suggest that judgement of
biologists and negotiation with resource managers must determine
them.
240 /SECTION VII
-------
B. PREDICTORS FOR MACROINVERTEBRATES
The relationship between aquatic insect density and
embeddedness appears to us to be sufficiently strong that we feel
reasonably secure in stating that at embeddedness levels above
about 2/3 or 3/4, aquatic insect density declines. We see no
evidence that embeddedness levels below 2/3 reduce density; in
fact, density tends to be higher at embeddedness of 2/3 than at
very low embeddedness levels. This is probably a phenomenon
caused by environmental complexity and space availability in
mixed particle sizes. In view of uncertainty regarding other
aspects of effect of fines on stream ecology in the northern
Rockies, it would be prudent to consider an embeddedness level of
2/3 as a tentative threshold level. Embeddedness this high would
probably violate needs of sediment-free interstices for
overwintering space for fish.
We see no relationship between embeddedness and insect
drift, although drift should logically decline at very high
embeddedness levels.
241 /SECTION VII
-------
C. MEASURES OF LAND USE
Real and detectable relationships appear to exist between
disturbances such as stream crossings and fines (Munther and
Frank 1986 a,b,c). Where above/below and before/after studies can
be conducted with adequate sample sizes, it should be possible to
determine whether given land management activies cause fines to
increase in corings or in cleaned gravels in buckets inserted in
the stream bottom. Cleansed gravels of moderate to high geometric
mean particle size offer the most sensitive "trap" for sediments
in these evaluations.
Edwards and Burns (1986) prepared correlations of embed-
dedness and various drainage characteristics in 19 tributaries of
the Payette National Forest. Their results, for 23 independent
variables, provide several significant (p < 0.05) relationships.
(Table C.I).
Table C.I. (From Edwards and Burns 1986) .
embeddedness with 23 independent variables,
probability that correlations equaled zero.
Correlations of
with
and
Variable r
1. Drainage size -0.19
2. Stream mileage -0.30
3. Stream gradient -0.44
4. Water yield -0.31
5. Percent glaciation -0.51
6. Drainage density -0.09
7. Road density 0.74
8. Road acres/decade 0.79
9. Acres sediment slope group 6+ -0.27
10. Ac. sed. slope group 11-12 -0.19
11. Acres road 0.72
12. Acres road on sed. slope grps.
w/mass waste hazard during original
2 decades post-construction 0.45
13. Acres road on sed. slope grps.
w/mass waste hazard current 0.65
14. Acres riparian road 0.25
242
0.04
0.09
0.19
0.09
0.26
0.01
0.55
0. 62
0.07
0.04
0.52
0.20
0.42
0.06
p of r = 0
0.46
0.23
0.07
0.22
0. 03
0.71
0. 00
0.00
0.28
0.46
0. 00
0.06
0.00
0. 32
/SECTION VII
-------
Table C.I, continued.
Variable r r*. p of r -_ 0
15. Number stream crossings 0.61 0.37 0.01
16. Natural sediment yield -0.22 0.05 0.38
17. Accelerated sediment 1955-65 0.44 0.20 0.07
18. % accel. sed. over natural
1955-65 0.60 0.36 0.01
19. Total sediment 1955-65 0.11 0.01 0.67
20. Accelerated sed. 1975-85 0.29 0.08 0.25
21. % accel. sed over natural
1975-85 0.52 0.27 0.03
22. Total sediment 1975-85 -0.16 0.02 0.54
23. Natural sediment/acre -0.15 0.02 0.54
The authors eliminated variables 9, 10, 16, 22 and 23
because the correlations were low and did not make physical
sense. After transformation with logarithms for non-linear
relationships and tests of interactions, the authors used
stepwise regressions to examine a best combination of independent
variables. A significant relationship was:
Embeddedness = 34.59 + 12.90 Iog10(acres road with current
mass waste hazard + 1) - 0.16(% glaciation).
This regression accounted for 83% of observed variability in
embeddedness and was highly significant (p = 0.0009, F = 37.01, 2
and 15 d.f.).
The next best regression was for embeddedness against road
density and percent glaciation. After examining interactions,
Edwards and Burns calculated a stepwise regression of the
following form:
Embeddedness = 32.17 + 10.95 Iog10(road density + 1)(acres
road with current mass waste hazard + 1) -
0.13(% glaciation).
This relationship was highly significant (r2 = 0.85, p = 0.0009,
F = 42.8, 2 and 15 d.f.).
Equations of the foregoing types, based on careful field
work, offer utility for assessing arbitrarily-selected threshold
243 /SECTION VII
-------
levels in embeddedness in fish habitat. However, a prerequisite
is that functional quantitative relationships between fish
densities and embeddedness, or winter habitat capacity and
embeddedness, must be developed.
Moring (1982) showed that stream gravel permeability
decreased in a clear-cut watershed in the Oregon Coast Range
after logging. Permeabilities dropped from a prelogging average
of about 4,900 cm/h to an average of 1,100 cm/h in the first year
after logging, then remained at an average of about 2,400 cm/h
for the next 6 years. Both the first-year and later changes were
significant (p < 0.01). No similar decline occurred in a
drainage in which 25% of the watershed was clear-cut, and per-
meabilities in a control watershed did not differ significantly
before and after logging in the test drainages (Figure C.I).
_ 5
I 3
S-,
_v It I
a 3
1
Illmlllll
FLTNN CR.
(control)
•llllllllll
DEIR CB.
[partial cut)
NEEDIE BR.
(dear-p.-«
SAMPUNG YEAR
Figure C.I. (From Moring 1982). Annual average permeability in
three tributaries of the Alsea River, Oregon, 1962-1973. In Deer
Creek and Needle Branch watersheds, road construction was com-
pleted after the 1964-65 season, and logging after the 1965-66
season.
Clear-cut logging by Georgia-Pacific Corporation in the
Needle Branch watershed was in accord with what would have been
244 /SECTION VII
-------
termed best management practices at the time. Permeabilities in
the streams were recorded at two-week intervals in three
permanent Mark VI standpipes installed in each stream. Had they
been obtained in egg pockets, the results would probably have
differed in absolute means and extent of change. We would expect
egg pockets to have higher mean permeabilities and to show
greater change over the incubation period than the temporal
changes in permanent pipes.
245 /SECTION VII
-------
D. ROLE OF JUDGEMENT IN CRITERIA FOR MANAGEMENT PRACTICES
In the absence of quantitative models of fish density in
relation to embeddedness, fishery managers and water quality
regulators may have to use professional judgement and negoti-
ations with development-oriented interests to establish interim
estimated threshold levels for embeddedness. Instances may
develop in which best management practices would lead to sediment
recruitment above judgement-based and negotiated limitations. In
this case, uncertainty and risk factors may be be deemed to
mandate that no incremental development occur in the watershed.
Fishery values and estimated stream recovery rates may be
important considerations in the negotiated interim settlement of
threshold levels. Certain stocks and species may justify more
stringent protective limitations than others.
We acknowledge that our inability to find or develop quan-
titative functional relationships that can be used as regulatory
tools will disappoint some readers and please others. Among the
latter may be some who have a vested interest in logging and road
construction. Our failure may also please those who have con-
tended that best management practices should be sufficient pro-
tection for aquatic communities, or that protective criteria must
have quantitative tools before they can be justified. We direct
the following comments to advocates of best management practices.
The interface between water and air is, to a very real
degree, the interface between relatively easy and very difficult
ecological assessments. Too often, it is also the interface
below which research funding has been insufficient. Even after 35
years of intensive work on the effects of fine sediments on the
intragravel environment, we are only now focusing on that which
directly influences survival of embryos to emergence. Only now
246 /SECTION VII
-------
have researchers begun to realize the importance of habitat for
overwintering by salmonids in the northern Rockies.
Quantitative tools for evaluation of effects of fines will
not supplant scientific judgement. The complexity and variety of
stream systems in the northern Rockies will make it impossible to
broadcast fixed criteria. We suggest that every system and sub-
system must be evaluated as a discrete unit; that conservatism in
favor of the fishery resource is prudent, especially in the case
of high-value fisheries, and should be quite amply justified
simply on the grounds that unforeseen externalities are easy to
cause and hard to rectify; and, finally, that informed judgement
should drive the system, rather than any arbitrary set of best
management practices. We feel that best management practices can
too easily be interpreted to mean economical practices. While
"economical" in this sense serves the timber or mining interests
well, it totally neglects consideration of the state of knowledge
in stream ecology and hidden costs passed on to the general
public in the form of deteriorated fishery resources and future
rehabilitation requirements.
Criteria for evaluation of best management practices,
however desirable, will not relieve the resource manager of his
responsibility in watershed husbandry. One may easily lose sight
of the fact that accelerated stream sedimentation reflects poor
land husbandry. But after all, what good land manager would
voluntarily sluice away the soil resource? Thus the two most
important criteria for critical evaluation of land management
practices in areas of valuable fisheries or slow recovery, be
those practices best or otherwise, should be: (1} Are we
conservative enough to avoid unforeseen externalities, and (2)
Will these measures result in zero acceleration in soil wasting?
We believe these criteria would meet the prudent and conservative
caveats noted above in sections A and B.
247 /SECTION VII
-------
The foregoing two criteria place the responsibility for good
stream protection through good land management where it belongs;
on the land manager. The aquatic resource professional must use
best judgement to establish habitat requirements and protective
measures in site-specific circumstances, thus providing guidance
for the land manager. Negotiation will have to precede final
decisions on fishery values, and protection measures.
248 /SECTION VII
-------
VIII. SUMMARY OF RESEARCH NEEDS
In this section we summarize our principal suggestions
regarding research that we believe will be required for
development of quantitative tools for assessment of effects of
fines on salmonids in the northern Rockies.
A. INTRAGRAVEL ENVIRONMENT
We recommend that assessment of survival to emergence be
pursued for salmonid redds by means of careful mapping of egg
pocket locations and depths in relation to temporary benchmarks,
followed by redd capping shortly before onset of emergence.
Timing of emergence can be determined from knowledge of spawning
time and accrued temperature units in the substrate.
Depths and precise locations of egg pockets should be
established when redds are mapped during spawning. A program for
periodic sampling of permeabilities and dissolved oxygen percent
saturation in the egg pocket centrum should be designed. Three-
probe freeze cores of egg pockets should be obtained after
alevins have emerged from the redd.
Concurrent programs should be conducted in redds outside the
egg pocket and in "spawning gravels" around the intensively-
studied redds so that relationships between survival (as assessed
from redd capping) and various physical statistics in the egg
pocket, in the redd area outside the pocket, and in the nearby
unused gravels can be compared. This should permit researchers
to learn whether conditions outside the egg pocket can serve as
surrogates for parameters in the pocket.
In addition to permeability and percent dissolved oxygen
saturation, researchers should assess geometric mean particle
249 /SECTION VIII
-------
size in and outside the egg pocket in up to three vertical strata
and for the composited vertical strata. Other statistics to be
calculated include the fredle index, and percentage of fines by
weight in various sieve categories. These data will permit
workers to assess conformity, if any, of survival and conditions
in the field to information obtained in various laboratory
studies.
Even if redd capping proves impossible in field situations
in Idaho, the work described above will establish physical
conditions in egg pockets, permitting laboratory workers to
develop better, more accurate, physical analogs for field
conditions. Even live:dead ratios obtained by hydraulically
sampling egg pockets before emergence begins (after appropriate
physical statistics have been obtained) would bring laboratory
work to date into better focus, offering a substitute for
survivals to emergence-
As soon as data become available on egg pockets and the egg
pocket centrum, we suggest that laboratory studies be designed to
duplicate, as closely as possible, conditions in the pocket.
The type of work typified by Tappel and Bjornn (1983) should be
pursued, but with better surrogates for natural egg pockets. The
independent variates that need to parallel natural conditions
this work include:
1. Geometric mean particle size.
2. Fredle index.
3. Gravel mix components.
4. Permeability.
5. Organic matter in the gravel mix.
6. Physical structure of the surrogate pocket.
250 /SECTION VIII
-------
B. REARING HABITAT
Embeddedness, substrate score, surface percentage of fines,
and other habitat variables such as gradient, riparian cover, in-
stream cover, stream orientation (north-south, east-west) need to
be evaluated in a program of randomly selected sites in several
habitat strata in the northern Rockies. It may be necessary to
adjust densities of fish (at least in anadromous fish areas in
which seeding by adults may not be sufficient to lead to full
recuitment of juveniles) in study sites by means of adding fish
seined, minnow-trapped, or electrofished from areas not
intensively studied. Fish abundance and biomass would then be
examined in relation to independent habitat variables.
Stream areas in which resident species not subjected to
heavy fishing pressure are the target animals may not require
population adjustments before studies begin. But whether
densities are enhanced or not, habitat and species variability
may well dictate that large numbers of study sites be evaluated.
251 /SECTION VIII
-------
C. WINTERING HABITAT
Intensive study should be allocated to the question of over-
wintering habitat used by anadromous and resident salmonids in
the northern Rockies. The first priority is to establish
relative survivals of fish that overwinter in areas near the
summer rearing zones and those that move downstream in fall to
overwinter in larger streams. "PIT" tags can be used for these
studies on steelhead and Chinook salmon in Idaho streams because
detection devices for the tags will provide for reasonably
efficient sampling at Lower Granite and Little Goose dams on the
Snake River.
The next or concurrent step is to better define where and
how densely salmonids are found in winter. Condition of habitat
in which fish are found should be defined in terms of gradient,
substrate particle sizes, embeddedness, cover, width, flow, and
other habitat variables. Areas not used, lightly occupied, and
heavily occupied should be described. We need to learn if low
wintering use by salmonids is associated with high incidence of
fines in the substrate.
Study of overwintering habitat and survivals may be the most
important work that we suggest in section VIII.
252 /SECTION VIII
-------
D. EXPERIMENTAL MANIPULATION OF FINES
We believe that test and control areas should be set up in
several drainages to evaluate the effect of trapping and removing
fine sediments on salmonid density and growth. This may require
careful selection of upstream control zones and downstream test
zones separated by large-volume sediment traps. Alexander and
Hansen (1983 and 1986) offer useful reading for those interested
in such manipulations. The size of the experimental environments
should be large enough to lead to detectable changes in habitat
quality and fish populations. Alexander and Hansen (1986)
separated a two-mile stream reach into a one-mile upstream
control and a one-mile downstream test area by removing fines at
the midpoint in a sediment trap.
253 /SECTION VIII
-------
IX. SUMMARY
1. Several measures of streamtaed character offer indices of
habitat quality. These include percentages of fines (most useful
categories in the northern Rockies are <6.35 mm, <0.85 mm, <9.5
mm) , geometric mean particle size (dg) , fredle index, and
permeability.
2. Percentage of fines in the intragravel environment varies
spatially and temporally. To reduce this sampling variability
where survival is to be related to substrate fines, samples
should be taken in redd egg pockets during the period when
embryos are incubating within the pockets.
3. Geometric mean particle size (dg) offers a workable measure-
ment as a companion to percent fines, for a more complete index
of habitat quality. High survivals tend to occur in association
with high dg, but failure of researchers to sample in egg pockets
and to relate dg to survival from deposition to emergence reduces
utility of the relationship for quantitative predictions of
survival,
4. Tappel and Bjornn (1983) tested embryo survival in relation
to gravel mixtures on the basis of two substrate variables, fines
percentages <9.5 mm and <0.85 mm. They eliminated particles
larger than 25.4 mm and used relatively small dg. This neglects
the fact that for many salmonids, the egg pocket centrum in
natural redds is formed of particles much larger than 25.4 mm.
We did not support use of these laboratory results to quantit-
atively predict conditions in natural egg pockets.
5. Field research has yet to verify the utility of the fredle
index (dg/sg) in a variety of field situations. Because it
embodies 9g and a mixing index, it offers promise.
6. Visual assessment offers a reasonable measure of major change
in surficial fines where used as a time-trend indicator. It does
not necessarily reflect conditions deeper in the substrate or
differences in depths of intrusions of fines.
7. Permeability offers a useful tool for correlations with
survival and for assessment of fines intrusions.
8. Assessment of conditions in the substrate as they control
salmonid incubation and emergence requires sampling in natural
egg pockets. To determine overall condition of "spawning
gravels" without reference to embryo survival, a somewhat less-
restrictive stratification could suffice.
9. The structure of salmonid redds is critical to the role of
fine sediments in affecting survival of incubating embryos and
emergence success. The lack of understanding of redd internal
254 /SECTION IX
-------
structure by researchers has hampered quantification of the role
of sediment in affecting incubation and emergence.
10. Survival of salmonid embryos is positively related to
apparent velocity and permeability. Dissolved oxygen affects
both emergence success and timing.
11. Survival to emergence tends to decrease as the proportion of
fine sediments increases in the incubation environment.
12. Studies on the effects of fines on emergence size have
shown conflicting results for the same species and among species.
Female size may have caused some variance.
13. Alevins tend to emerge earlier from gravels with high
percentages of fines. This does not appear to have any direct
deleterious effects on the alevin, and may be an adaptive
mechanism that has the effect of trading-off mortality in surface
waters against mortality in the substrate.
14. Various forms of substrate scoring have been used to
describe habitat suitability for aquatic insects and fish.
Substrate score correlates reasonably well with dg. Visual
assessments at best offer indicators of microhabitat conditions
in the surface zone in areas not armored.
15. Photographic assessment of fines has been pursued by a few
workers but appears to underestimate fines smaller than 6.3 mm
when compared with core samples.
16. Core samples seem to offer the most complete assessment of
substrate components, but do not serve well for evaluation of
summer rearing or winter hiding habitat. Measures of coarseness
may have utility concert with visual methods, photographic
techniques or embeddedness measurements.
17. Embeddedness, free matrix particles, and percent of fines
are related. Embeddedness works best where sand is an important
component of the substrate. These measures offer useful "before
and after" or "above and below" measure of changes over time and
space. Sediment traps may provide useful tools as well.
18. All newly emerged salmonid fry utilize shallow water areas
with low velocities, where fines tend to accumulate. Newly-
emerged fry remain close to the stream margin in quiet water,
making availability of these areas in spring and summer
important, but not likely to be limiting.
19. Effects of fines on juveniles in the size groupings classed
as age 0, I and II are of particular concern. As juveniles grow
they begin using deeper faster water. The weight of evidence
tends to indicate that areas with high embeddedness (50%
255 /SECTION IX
-------
embeddedness may define "high") tend to have lower densities of
salmonids. The data do not consistently or convincingly support a
negative relationship between fines and fish rearing density. In
some cases density correlates positively with fines.
20. Loss of pool volume due to sediment deposition reduces
suitability of a stream for adult salmonids.
21. Macroinvertebrate biomass and diversity decrease where fines
predominate. Also aquatic insect density declines at embed-
dedness levels above about 2/3 or 3/4. We saw no relationship
between embeddedness and insect drift.
22. At water temperatures below 10 C, salmonids often begin to
emigrate or seek cover in the substrate. This behavior varies by
species/race.
23. From the available data we infer that as embeddedness
increases, winter carrying capacity declines. However, fish
response to embedded habitat may vary between species and
populations. Fish appear to seek winter habitat in upstream
locations. Addition of sediment that reduces availability of
this environmental requisite probably increases mortality.
24. We believe the role of channel structure in legislating
conditions for salmonid spawning and rearing is of great
importance and deserves greater research emphasis.
25. Aggradation reduces channel diversity, leading to probable
increases in embeddedness levels.
26. Fines in interstices of the bed may delay the onset of bed
movement during large storm flows, a movement critical to the
removal of accumulated fines in spawning gravels.
27. Woody debris is an important component of stream structure,
especially in smaller streams. Stream structure may favor or
detrimentally affect particular salmonids.
28. For any of the independent variables that have been
previously studied (fredle index, dg, percent fines, and
permeability) to offer utility, reality, and permit quantitative
prediction of survivals, they must be shown to reflect conditions
in the egg pocket of the salmonid redd. Given the existing
information, we do not favor extrapolation of laboratory data on
survival to field situations.
29. In order to develop useful and quantitative predictions of
survival in relation to the fines, fredle index and geometric
mean particle size, we suggest several alternative sampling
scenarios, all of which incorporate redd capping to assess
survival to emergence and freeze-core sampling of the egg pocket.
256 /SECTION IX
-------
30. As with the other procedures, care should be taken when
assessing permeability that the measures come from within egg
pockets. We suggest that permeability assessment is easier by
far than gravel coring and we offer some possibilities for
sampling.
31. Dissolved oxygen measurements from within the egg pocket
could be used to assess incubation conditions.
32. Field evaluations of embeddedness effects on fish density
should be undertaken in carefully documented stream strata and
fully-seeded fish populations. This may require the addition of
juveniles from other nearby areas to the test strata. Such field
experiments at full seeding in several environmental strata are
needed to assess effects of embeddedness on density of fish.
33. The relationship between density of fish and percentage of
fines in rearing areas is fairly weak until there is a
superabundance of fines. "Superabundance" means enough fines to
reduce living space and overcome ability of the stream to process
sediment recruitment so that food production on riffles is
smothered and sharply declines.
34. Substrate score currently has insufficient documentation for
application to the species of the northern Rockies but does merit
pursuit in future studies for it should offer more information
than embeddedness alone or visually-estimated fines alone.
35. The embeddedness level above which aquatic insects decline in
abundance is about 2/3 to 3/4.
36. No useful relationship between substrate type and insect
drift is available for predicting effects of fines on the insect
drift resource in streams of the northern Rockies.
37. We believe that functional relationships exist between
embeddedness and winter holding capacity of the substrate for
salmonids and that those relationships differ by fish size and
perhaps by species. It is currently impossible to form
quantitative predictors. Research should examine the effects of
various substrate types, starting and ending fish densities and
fish size, and species on the shape of the relationship between
winter holding capacity and embeddedness.
38. We contend that resource managers should attempt to maintain
maximum ecological diversity as a hedge. This concept means, in
the case of the intragravel environment, for example, minimal (or
no) introduction of fines to the existing gravel matrix and every
reasonable effort to reduce sediment recruitment from basin
development. In overwintering habitat, the diversity hedge would
prevent any increment of fines until predictive tools are
257 /SECTION IX
-------
developed.
39. We found no functional predictors that would serve
environmental regulators in evaluating quantitative effects of
sediment in the natural incubation, rearing, or wintering phases
of salmonid life history in the northern Rockies. In the absence
of criteria, we suggest that due to uncertainty and risk factors,
any incremental increase in intragravel fines in egg pockets
should be avoided until functional realtionships can be
developed.
40. We suggest that threshold levels of minimum dissolves oxygen
content in intragravel water in the egg pocket could be set at
98% of saturation at temperatures of 0-10 C for high value
fisheries, 76% of saturation at 0-10 C for fisheries of moderate
value, and 54-57% of saturation for fisheries of lower value.
Judgement and negotiation would have to set fishery value
categories.
41. We make no recommendation regarding surficial fines as they
may affect salmonid rearing densities.
42. No threshold value can be set for embeddedness as it affects
overwintering. We believe it prudent and conservative on the
side of salmomid gene pools to permit no man-caused incremental
embeddedness until functional relationships can be established,
at least in areas that support fisheries of moderate to high
value.
43. As mentioned before, insect density declines at embeddedness
levels above about 2/3 or 3/4. It would be prudent to consider an
embeddedness level of 2/3 as a tentative threshold level.
Embeddedness this high would probably violate needs of sediment-
free interstices for overwintering space for fish.
44. Real and detectable relationships exist between land-
disturbing activities and increased fines in the aquatic
environment.
45. In view of uncertainty and environmental variability,
professional judgement must play an important role in evaluating
effects of fine sediments in salmonid habitat in the northern
Rockies.
46. Regulatory agencies may have to provide interim criteria for
non-point source sediment delivery to salmonid habitat. We
summarize best available information from the body of laboratory
and field information. We offer examples of the variables that
might serve for this purpose. We urge that research proceed apace
to develop criteria that will permit quantitative predictions of
survival or density of fish and to evaluate efficacy of any
interim criteria that are adopted.
258 /SECTION IX
-------
47. Assessment of survival to emergence on salmonid redds should
be conducted by mapping of egg pocket locations and depths, redd
capping shortly before emergence, sampling of permeability and DO
in the egg pocket, and three-probe freeze-coring in the egg
pocket after emergence. Similar sampling should be conducted in
substrate areas outside the egg pocket to compare physical
statistics that may serve as surrogates for parameters in the
pocket. Even where redd capping fails to provide good inform-
ation on survival, data on egg pocket structure will permit
laboratory studies to better model field conditions.
48. Assessment of rearing habitat requires an evaluation of
random sites in several habitat strata in the northern Rockies.
Embeddedness, substrate score, surface percentage of fines, and
other habitat conditions such as gradient, riparian cover,
instream cover, and stream orientation should be defined.
Densities of, fish may need to be adjusted by addition of fish
from nearby areas.
49. Much work is needed to define overwintering habitat.
Survival of fish overwintering in summer rearing areas versus
downstream areas and where and at what density salmonids spend
the winter are the major evaluations of concern.
50. Experimental manipulation of fines should help evaluate the
effect of fine sediment on salmonid density and growth.
259 /SECTION IX
-------
Acknowledqments
We thank several members of the Technical Advisory Committee
for their thoughtful comments and criticisms of an earlier draft
of this report. We thank Dale McCullough specifically for his
editorial criticisms, and J. S. Griffith for his counsel. Review
by these colleagues does not imply responsibility for, or
approval of, the contents of our report.
Researchers in the program supervised by T. C. Bjornn, at
the University of Idaho, deserve a special thanks. Their efforts
through the years have provided high-quality research in pro-
fusion, and numerous working hypotheses pertinent to effects of
fines on salmonids of the northern Rockies. The criticisms that
we have offered of that work do not detract in the slightest from
our admiration of the intelligence and diligence represented in
in the many graduate theses and publications that have evolved
from the program.
We have come to better appreciate the human ingenuity
represented in the many research works that we reviewed and
criticized. If no risks had been taken, criticism and progress
would wither. We dedicate our report to those who took the risks.
260 /SECTION IX
-------
X. LITERATURE CITED
Adams, J. N. and R. L. Beschta. 1980. Gravel bed composition
in Oregon coastal streams. Can. J. Fish. Aquat. Sci.
37:1514-1521.
Alderdice, D. W., W. P. Wickett, and J. R. Brett. 1958. Some
effects of exposure to low dissolved oxygen levels on
Pacific salmon eggs. J. Fish. Res. Board Can. 15:229-250.
Alexander, G. R. and E. A. Hansen. 1983. Sand sediment in a
Michigan trout stream Part II. Effects of reducing sand
bedload on a trout population. N. Am. J. Fish. Mgmt. 3:365-
372.
Alexander, G. R. and E. A. Hansen. 1986. Sand bed load in a
brook trout stream. N. Am. J. Fish. Mgt. 6:9-23.
Allen, K. R. 1951. The Horokiwi stream. A study of a trout
population. New Zealand Marine Dept., Fish. Bull. No. 10.
1962. The natural regulation of population in the
salmonidae. New Zealand Sci. Rev., 20:58-62.
Bachman, R. W. 1958. The ecology of four North Idaho trout
streams with reference to the influence of forest road
construction. M.S. thesis, Univ. of Idaho. 97 p.
Bachman, R. A. 1984. Foraging behavior of free-ranging wild and
hatchery brown trout in a stream. Trans. Am. Fish. Soc.
113(1):l-32.
Ball, K. W. 1971. Initial effects of catch and release on
cutthroat trout in an Idaho stream. M.S. thesis, Univ.
of Idaho, Moscow.
Bams, R. A. 1969. Adaptations of sockeye salmon associated with
incubation in stream gravels, pp. 71-81 In T.G. Northcote,
ed., Symposium on Salmon and Trout in Streams, H.R.
MacMillan Lectures in Fisheries, Univ. of British Columbia,
Vancouver,B.C.
Behning, A. 1924. Einige Ergebnisse qualitativer und
guantitativer Untersuchungen der Bodenfauna der Wolga.
Verh. int. Verein. theor. angew. Limnol. 2:71-94.
Beschta, R. L. 1983a. Channel changes following storm-induced
hillslope erosion in the upper Kowai Basin, Torlesse
Range, New Zealand. J. Hydrology (N.Z.) 22:93-111.
. 1983b. Long-term changes in channel widths of the
261 /SECTION X
-------
Kowai River, Torlesse Range, New Zealand. J. Hydrology
(N.Z.) 22:112-122.
Beschta, R. L. and W. L. Jackson. 1979. The intrusion of fine
sediments into a stable gravel bed. J. Fish. Res. Bd. Canada
36:204-210.
Beschta, R. L. and W. S. Platts. 1986. Morphological features of
small streams: significance and function. Water Res. Bull.
22(3}:369-379.
Bianchi, D. R. 1963. The effects of sedimentation on egg survival
of rainbow trout and cutthroat trout. M.S. thesis, Montana
State College, Bozeman. 28 p.
Bisson, P. A. 1986. Importance of identification of limiting
factors in an evaluation program. Unpublished draft report,
Weyerhaeuser Company, Tacoma, WA. 22 p.
Bisson, P. A., R. E. Bilby, M. D. Bryant, C. A. Dolloff,
G. B. Grette, R. A. House, M. L. Murphy, K. V. Koski,
and J. R. Sedell. 1987. In: T. Cundy and E. Salo, (Eds.)
Large woody debris in forested streams in the Pacific
Northwest: Past, present, and future. Proc. Sympos. on
Forestry and Fishery Interactions, Feb. 12-14, 1986.
Univ. Washington, Seattle.
Bisson, P. A. and J. R. Sedell. 1984. Salmonid populations in
streams in clearcut vs. old-growth forests of Western
Washington. In W.R. Meehan, T.R. Merrel, Jr. and T.A. Hanley
eds. Wildlife Relationships in Old-Growth Forests:
Proceedings of a symposium held in Juneau, AK 12-15 April
1982. Am. Inst. Fish. Res. Biol. pp. 121-129.
Bisson, P. A., K. Sullivan, and J. L. Nielsen. 1986. Channel
hydraulics, habitat utilization, and body form of coho
salmon, steelhead trout, and cutthroat trout in streams.
Manuscript submitted to: Can. J. Fish and Aguat. Sci.
Bjornn, T. C. 1969. Embryo survival and emergence studies. Job -
No.5, Salmon and Steelhead Invest., Project F-49-R-7, Annual
Completion Rep. Idaho Dept. of Fish and Game. 11 p.
. 1971. Trout and salmon movements in two Idaho
streams as related to temperature, food, stream flow,
cover, and population density. Trans. Am. Fish. Soc. 100(3):
423-438.
Bjornn, T. C., M. A. Brusven, M. P. Molnau, J. H. Milligan, R. A.
Klamt, E. Chacho, and C. Schaye. 1977. Transport of granitic
sediment in streams and its effects on insects and fish.
Forest, Wildlife and Range Station. Technical Report,
262 /SECTION X
-------
Project B-036-IDA.
Brannon, E. L. 1965. The influence of physical factors on the
development and weight of sockeye salmon embryos and
alevins. Int. Pac. Salmon Fish. Comm. Progress Rep. No. 12.
Brusven, M. A. and K. Prather. 1974. Influence of stream sedi-
ments on distribution of macrobenthos. J. Ent. Soc. B.C.
71:25-32.
Burner, C. J. 1951. Characteristics of spawning nests of Columbia
River salmon. U.S. Fish and Wildlife Ser., Fishery Bull. 61,
Vol.52.
Burns, D. C. 1978. Photographic and core sample analysis of fine
sediment on the Secesh River - 1978. USDA Forest Service,
McCall, Idaho.
. 1983. 2630 Habitat. Informal letter of December
19, 1983. Subject: CD'A Mines Thunder Mountain Mine
Project. To: E. Dodds, Team Leader. Payette National For-
est Files, McCall, Idaho.
. 1984. An inventory of embeddedness of salmonid
habitat in the South Fork Salmon River drainage, Idaho. For
the Payette and Boise National Forest.
Burns, D. C. and R. E. Edwards. 1985. Embeddedness of salmonid
habitat of selected streams on the Payette National Forest.
For the Payette National Forest.
Bums, J. W. 1972. Some effects of logging and associated road
construction on Northern California streams. Trans. Amer.
Fish. Soc. 101(1}:1-17.
Bustard, D. R. and D. W. Narver. 1975a. Aspects of winter ecology
of juvenile coho salmon (Oncorhynchus kisutch) and steelhead
trout fSalmo gairdneri). J. Fish. Res. Bd. Can. 32(5):667-
680.
Bustard, D. R. and D. W. Narver. 1975b. Preferences of juvenile
coho salmon fOncorhynchus kisutch) and steelhead trout
(Salmo gairdneri) relative to simulated alteration of winter
habitat J. Fish. Res. Bd. Can. 32 (5) :681-687.
Campbell, H. J. 1954. The effect of siltation from gold dredging
on the survival of rainbow trout and eyed eggs in Powder
River, Oregon. Oregon Game Commission 3 mimeo pp.
Campbell, R. F. and J. H. Neuner. 1985. Seasonal and diurnal
shifts in habitat utilized by resident rainbow trout in
Western Washington Cascade Mountain streams. Pages 39-48 in
263 /SECTION X
-------
F. W. Olson, R. G. White, and R. H. Hamre, eds., Symposium
on Small Hydropower and Fisheries, Am. Fish. Soc., Aurora,
CO, 1-3 May 1985.
Cederholm, C. J. and L. C. Lestelle. 1974, Observations on the
effects of landslide siltation on salmon and trout resources
of the Clearwater River, Jefferson County, Washington 1972-
73. Univ. Washington, Fish. Res. Inst. Final Rep. FRI-UW-
7404. 89 p.
Cederholm, C.J., L.M. Reid, B.C. Edie, and E.O. Salo. 1981.
Effects of forest road erosion on salmonid spawning gravel
composition and populations of the Clearwater River,
Washington. Contribution No. 568, College of Fish., Univ.
of Washington, Seattle, WA.
Cederholm, C.J. and L. M. Reid. 1986. The impacts of
sedimentation on coho salmon populations of the Clearwater
River. Symposium on Streamside Management: Fishery and
Forestry Interactions. Univ. of Wash., Seattle.
Cederholm, C.J. and E.O. Salo. 1979. The effects of logging road
landslide siltation on the salmon and trout spawning gravels
of Stequaliho Creek and the Clearwater River basin,
Jefferson County, Washington, 1972-1978. Univ. Washington,
Fish. Res. Inst., Final Report Part III, FRI-uw-7915. 99 p.
Cederholm, C. J., W. J. Scarlett, and E. O. Salo. 1977. Salmonid
spawning gravel composition data summary from the Clearwater
River and its tributaries, Jefferson County, Washington,
1972-76. Univ. Washington Fish. Res. Inst. Circ. 77-1. 135
pp.
Chapman, D.W. 1962. Aggressive behavior in juvenile coho salmon
as a cause of emigration. J. Fish. Res. Bd. Canada 19(6):
1047-1080,
1965. Net production of juvenile coho salmon in
three Oregon streams. Trans. Am. Fish. Soc. 94(l):40-52.
. 1966. Food and space as regulators of salmonid
populations in streams. Am. Nat. 100(913):345-357.
Chapman, D. W. and T. C. Bjornn. 1969. Distribution of salmonids
in streams with special reference to food and feeding. Pages
153-176 in H.R. MacMillan Lectures in Fisheries, Symposium
on Salmon and Trout in Streams.
Chapman, D. W., D. E. Weitkamp, T. L. Welsh, and T. H. Schadt.
1983. Effects of minimum flow regimes on fall chinook
spawning at Vernita Bar 1978-82. Report to Grant County
Public Utility District, Ephrata, WA. by Don Chapman Con-
264 /SECTION X
-------
sultants, Inc., and Parametrix, Inc. 123 p.
Chapman, D. w., D. E. Weitkamp, T. L. Welsh, M. B. Dell, and T.
H. Schadt. 1986. Effects of river flow on the distribution
of chinook salmon redds. Trans. Am. Fish. Soc. 115(4):537-
547.
Chutter, E.M. 1969. The effects of silt and sand on the
invertebrate fauna of streams and rivers. Hydrobiologia
34:57-76.
Coble, D. W. 1961. Influence of water exchange and dissolved
oxygen in redds on survival of steelhead trout embryos.
Trans. Amer. Fish. Soc. 90: 469-474.
Cooper, A. C. 1956. A study of Horsefly River and the effect of
placer mining operations on sockeye spawning grounds. Int.
Pac. Sal. Fish. Comm. Bull. 18. 58 p.
1965. The effect of transported stream sediments on
survival of sockeye and pink salmon eggs and alevin. Int.
Pac. Sal. Fish. Comm. Bull. 18. 71 pp.
Cordone, A. J. and D. E. Kelley. 1961. The influence of inorganic
sediments on the aquatic life of streams. Calif. Fish and
Game 47:189-228.
Corley, D. R. and D. D. Newberry. 1982. Fishery habitat survey of
the South Fork Salmon River - 1981. USFS, Boise and Payette
National Forests.
Grouse, M. R., C. A. Callahan, K. w. Maleug,and S. E. Dominguez.
1981. Effects of fine sediments on growth of juvenile coho
salmon in laboratory streams. Trans. Am. Fish. Soc.
110:281-286.
Cummins, K.W. and G.H. Lauff. 1969. The influence of substrate
particle size on the microdistribution of stream
macrobenthos. Hydrobiologia 34:145-181.
Cunjak, R. A. and G. Power. 1986. Winter habitat utilization
by stream resident brook trout (Salvelinus fontinalis) and
brown trout fSalmo trutta). Can. J. Fish. Aquat. Sci. 43:
1970-1981.
Davis, J. C. 1975. Minimal dissolved oxygen requirements of
aquatic life with emphasis on Canadian species: a review.
J. Fish. Res. Bd. Can. 32:2295-2332.
Dill, L. M. and T. G. Worthcote. 1970. Effects of gravel size,
egg depth, and egg density on intragravel movement and
emergence of coho salmon (Oneorhynchus kisutch) alevins.
265 /SECTION X
-------
J. Fish. Res. Bd. Canada 27:1191-1199.
Dolloff, C. A. 1986. Effects of stream cleaning on juvenile
coho salmon and dolly varden in southeast Alaska. Trans.
Amer. Fish. Soc. 115:743-755.
Edie, B. G. 1975. A census of the juvenile salmonids of the
Clearwater River Basin, Jefferson County, Washington, in
relation to logging. M.S. thesis, Univ. of Washington,
Seattle, WA.
Edmundson, E., F. E. Everest, and D. W. Chapman. 1968. Permanence
of station in juvenile chinook salmon and steelhead trout.
J. Fish. Res. Bd. Can. 25 (7) :1453-1464.
Edwards, R. and D. Burns. 1986. Relationships among fish habi-
tat embeddedness, geomorphology, land disturbing activities
and the Payette National Forest sediment model. Payette
National Forest, U.S. Forest Service, McCall, Idaho.
Envirocon, Ltd. 1984. Kemano completion hydroelectric develop-
ment. Enviromental studies: Vol. 5, Fish Resources of the
Nechako River system: Baseline information.
Erman, D. C. and N. A. Erman. 1984. The response of stream
macroinvertebrates to substrate size and heterogeneity.
Hydrobiologia 108:75-82.
Erman, C. D. and D. Mahoney. 1983. Recovery after logging in
streams with and without bufferstrips in Northern
California. Cal. Water Res. Center, Univ. of California,
Contribution No. 186.
Everest, F. H., R. L. Beschta, J. C. Scrivener, K. V. Koski,
J.R. Sedell and C. J. Cederholm. 1986. Fine sediment and
salmonid production - a paradox. Draft report.
Everest, F. H. and D. W. Chapman. 1972. Habitat selection and
spacial interaction by juvenile chinook salmon and steelhead
trout in two Idaho streams. J. Fish. Res. Bd. Can. 29(1):91-
100.
Everest, F. H., F. B. Lotspeich, and W. R. Meehan. 1981. New
perspectives on sampling, analysis,and interpretation of
spawning gravel quality. In N.B. Armantrout. ed.,
Proceedings of a Symposium on Aquisition and Utilization of
Aquatic Habitat Inventory Information, Western Div. Am.
Fish. Soc., Portland, OR 99:325-333.
Everest, F. H. and W. R. Meehan. 1981. Forest management and
anadromous fish habitat productivity. Pages 521-530 in
46th North Am. Wildl. and Nat. Res. Conf. Trans. Wildl.
266 /SECTION X
-------
Mgmt. Inst., Washington D.C.
Fraser, F.J. 1969. Population density effects on survival and
growth of juvenile coho salmon and steelhead trout in
experimental stream channels. Pages 253-266 in H.R.
MacMillan Lectures in Fisheries, Symposium on Salmon and
Trout in Streams, T.G. Northcote, ed. Inst. of Fisheries,
Univ. of British Columbia, Vancouver, B.C.
Gamblin, M.S. 1986. Taft/Bell sediment and fishery monitoring
project. Idaho Dept. Fish and Game draft progress report
for BPA Intergovernmental Agreement DE-AI 79-85 BP 23203.
Gangmark, H. A. and R. G. Bakkala. 1958. Plastic standpipe for
sampling streambed environment of salmon spawn. USDI Fish
and Wildlife Ser. Spec. Sci. Rep: Fish. No. 261.
Hall, J. D., and R. L. Lantz. 1969. Effects of logging on the
habitat of coho salmon and cutthroat in coastal streams. pp
353-375. In T. G. Northcote, ed., a Symposium on Salmon
and Trout in Streams. Univ. Brit. Columbia, Vancouver.
Hansen, E. A., G. R. Alexander and W. H. Dunn. 1983. Sand
sediment in a Michigan trout stream Part I. A technique for
removing sand bedload from streams. N. Am. J. Fish. Mgmt.
3:355-364.
Hardy, C. J. 1963. An examination of eleven stranded redds of
brown trout (Salmo trutta), excavated in the Selwyn River
during July and August, 1960. N.Z. J. Sci. 6:107-119.
Harrison, C. W. 1923. Planting eyed salmon and trout eggs.
Trans. Am. Fish. Soc. 52:191-200.
Hartman, G. F. 1963. Observations on behavior of juvenile brown
trout in a stream aquarium during winter and spring. J.
Fish. Res. Bd. Can. 20(3):769-787.
1965. The role of behavior in the ecology and
interaction of underyearling coho salmon fOjicorhynchus
kisutch) and steelhead trout (Salmo gairdneri). J. Fish.
Res. Bd. Can. 22:1035-1081.
Hausle, D. A. and D. W. Coble. 1976. Influence of sand in redds
on survival and emergence of brook trout (Salvalinus
fontinalis). Trans. Am. Fish. Soc. 105(1):57-63.
Hawke, S. P. 1978. Stranded redds of quinnat salmon in the
Mathias River, South Island, New Zealand. N.Z. J. Marine
and Freshwater Res. 12(2):167-171.
Hays, F.R., I.R. Wilmot, and D.A. Livingstone. 1951. The oxygen
267 /SECTION X
-------
consumption of the salmon egg in relation to development and
activity. J. Exper. Zool. 116(3):377-395.
Heifetz, J., M.W. Murphy, and K.V. Koski. 1986. Effects of
logging on winter habitat of juvenile salmonids in Alaskan
streams. N. Amer. J. Fish. Mgmt. 6:52-58.
Hilgert, P. 1982. Evaluation of instream flow methodologies for
fisheries in Nebraska. Nebraska Tech. Series No. 10. 50 p.
Hillman, T. W., J. S. Griffith, and W. s. Platts. 1986. The
effects of sediment on summer and winter habitat selection
by juvenile chinook salmon in an Idaho stream. Unpublished
Report. Idaho State Univ., Dept. of Biological
Sciences.
Hobbs, D. F. 1937. Natural production of quinnat salmon, brown
and rainbow trout in certain New Zealand waters. N. Z.
Mar. Dep. Fish., Bull. 6. 104 p.
Hogander, G., T. C. Bjornn, and S. Pettit. 1974. Evaluation of
catch and release regulations on cutthroat trout in the
North Fork of the Clearwater River. Idaho Dept. Fish and
Game, Project No. F-59-R-5. Job No. 1.
House, R.A. and P.L. Boehne. 1986. Effects of instream
structures on salmonid habitat and populations in Tobe
Creek, Oregon. N. Amer. J. Fish. Mgt. 6:38-46.
Huffaker, C. B. 1958. Experimental studies on predation:disper-
sion factors and predator-prey oscillations. Hilgardia
27:343-383.
Hunt, R. L. 1969. Effects of habitat alteration on production,
standing crops and yield of brook trout in Lawrence Creek,
Wisconsin.
Hynes, H.B.N. 1961. The invertebrate fauna of a Welsh mountain
stream. Arch. Hydrobiol. 57(3):344-388.
1970. The ecology of running waters. Univ. of
Toronto Press, Toronto. 555 p.
Irving, J.S. and T.C. Bjornn. 1984. Effects of substrate size
composition on survival of kokanee salmon and cutthroat and
rainbow trout. Idaho Coop. Fish. Res. Unit Tech. Rep. 84-6.
Kelley, D. W. and D. H. Dettman. 1980. Relationships between
streamflow, rearing habitat, substrate conditions, and
juvenile steelhead populations in Lagunitas Creek, Marin
County, 1979. Unpub. Rept. Marin County Water District. 23p.
268 /SECTION X
-------
Kennedy, H. D. 1967. Seasonal abundance of aquatic invertebrates
and their utilization by hatchery-reared rainbow trout.
U.S. Fish and Wildlife Service, Bur. Sport Fish, and Wildl.,
Tech. Pap. 12, Washington, D.C.
Klamt, R. R. 1976. The effects of coarse granitic sand on the
distribution and abundance of salmonids in the central Idaho
batholith. M.S. thesis, Univ. of Idaho, Moscow. 85 p.
Konopacky, R.C. 1984. Sedimentation and productivity in salmonid
streams. Ph.D. Dissertation, Univ. of Idaho, Moscow.
Konopacky, R.C., E.C. Bowles, and P.J. Cernera. 1985. Natural
propagation and habitat improvement: Salmon River habitat
enhancement. Bonneville Power Adm., Div. Fish and wildlife.
Project No. 83-359.
Koski, K. V. 1966. The survival of coho salmon (Oncorhyj^chus
kisutch) from egg deposition to emergence in three Oregon
streams. M.S. thesis, Oregon St. Univ., Corvallis. 84 p.
. 1975. The survival and fitness of two stocks of
chum salmon (Qnchorhynchus keta) from egg deposition to
emergence in a controlled stream environment at Big Beef
Creek. Ph.D. dissertation, Univ. Washington, Seattle. 212
pp.
1981. The survival and quality of two stocks of chum
salmon COncorhynchus ketaj from egg deposition to emergence.
Rapp. P.-V. Rev. Cons. Int. Explor. Mer. 178:330-333.
Lack, D. 1954. The natural regulation of animal numbers.
Oxford U. Press, London.
Lestelle, L. C, 1978. The effects of forest debris removal on a
population of resident cutthroat trout in a small headwater
stream. M. S. thesis, Univ. Washington, Seattle, 86 pp.
Levinski, C.L. 1986. Technical assessment of alternative
techniques for evaluating the impacts of stream sediment of
fish habitat. Idaho Dept. Health & Welfare, Div. of
Environment, Water Quality Bureau Report.
Lindroth, A. 1942. Sauerstoffverbrauch der Fische. I.
Verschiendene Entwicklungs-und Altersstadein vom Lachs und
Hecht. Z. vergl. Fhysiol., 29 (4) :583-594.
Lisle, T. E. 1982. Effects of aggradation and degradation on
riffle-pool morphology in natural gravel channels, north-
western California. Water Resources Research 18:1643-1651.
Lister, D. B. and H. S. Genoe. 1970. Stream habitat utilization
269 /SECTION X
-------
by cohabitating underyearlings of Chinook (Qncorhvnchus
tshawytscha) and coho (O. kisutchJ salmon in the Big
Qualicum River, British Columbia. J. Fish. Res. Bd. Can.
27:1215-1224.
Lotspeich, F. D, and F. H. Everest. 1981. A new method for
reporting and interpreting textural composition of spawning
gravel. Res. Note PNW - 369, Pac. Northwest Forest and
Range Exp. Sta., USDA.
Lund, J. A. 1985. Fine sediment analysis of chinook spawning
sites in the Secesh River, 1984. USFS, Payette National
Forest.
Mahoney, D. and D, C. Erman. 1984. An index of stored fine
sediment in gravel bedded streams. Water Res. Bull. 20(3):
343-348.
Mallet, J. L. 1963. The life history and seasonal movements of
cutthroat trout in the Salmon River, Idaho. M.S. thesis,
Univ. of Idaho, Moscow. 62 p.
Martin, D. J. 1976. The effects of sediment and organic detritus
on the production of benthic macro-invertebrates in four
tributary streams of the Clearwater River, Washington. M.
S. thesis. Univ. Washington, Seattle.
Mason, J. C. 1969. Hypoxial stress prior to emergence and
competition among coho salmon fry. J. Fish. Res. Bd. Can.
26:63-91.
1976. Response of underyearling coho to'
supplemental feeding in a natural stream. J. Wildl. Mgt.
40:775-778.
Mason, J. C. and D. W. Chapman. 1965. Significance of early
emergence, environmental rearing capacity, and behavioral
ecology of juvenile coho salmon in stream channels, J. Fish.
Res. Bd. Canada, 22(1):173-190.
McClelland, W, T. 1972. The effects of introduced sediment on the
ecology and behavior of stream insects. Ph.D thesis, Univ.
of Idaho, Moscow.
MacCrimmon, H. R. 1954. Stream studies on planted Atlantic
salmon. J. Fish. Res. Bd. Can. 11:362-403.
MacCrimmon, H. R. and B. L. Gots. 1986. Laboratory observations
on emergent patterns of juvenile rainbow trout, Salmo
gairdneri, relative to test substrate composition. Fifth
trout stream habitat improvement workshop, sponsored by
American Fisheries Society, Penn. State Fish Comm., Sport
270 /SECTION X
-------
Fishery Res. Found,, Trout Unlimited, and U. S. Forest
Service. Lock Haven University, Lock. Haven, PA. pp. 63-76.
McCuddin, M. E. 1977. Survival of salmon and trout embryos and
fry in gravel-sand mixtures. M.S. thesis, Univ. of Idaho.
Moscow, Id.
McDonald, J. G. and M. P. Shepard. 1955. Stream conditions and
sockeye fry production at Williams Creek. Fish. Res. Bd.
Can. Prog. Rep. Pac. Coast Stn. 104:34-37.
McFadden, J.T. 1969. Dynamics and regulation of salmonid
populations in streams. Pages 313-329 in H.R. MacMillan
Lectures in Fisheries, Symposium on Salmon and Trout in
Streams, T. G. Northcote, ed. Inst. of Fisheries, Univ. of
British Columbia, Vancouver, B.C.
McNeil, W. J. 1962. Variations in the dissolved oxygen content
of intragravel water in four spawning streams of
southeastern Alaska. U.S. Fish and Wildlife Ser. Spec. Sci.
Rep. Fish. No. 402. 15 p.
. 1966. Effects of the spawning bed environment on
reproduction of pink and chum salmon. U.S. Fish and Wildlife
Ser., Fish. Bull. 65 (2) :495-523.
. 1969. Survival of pink and chum salmon eggs and
alevins. Pages 101-116 in T.G. Northcote, ed., Symposium on
Salmon and Trout in Streams, H.R. MacMillan Lectures in
Fisheries, Univ. of British Columbia, Vancouver, B.C.
McNeil, W. J., and W. H. Ahnell. 1964. Success of pink salmon
spawning relative to size of spawning bed materials. U. S.
Fish and Wildlife Serv. Spec. Sci. Rep.—Fish. 469. 15 pp.
Meehan, W. R. and D. N. Swanston. 1977. Effects of gravel
morphology on fine sediment accumulation and survival of
incubating salmon eggs. USDA For. Ser. Res. Paper PNW-220.
Megahan, W. F., W. S. Platts, and B. Kulesza. 1980. Riverbed
improves over time: South Fork Salmon. Symposium on
Watershed Management, Vol 1. Am. Soc. Civ. Eng., New York.
Milhous, R. T. and P. C. Klingeman. 1973. Sediment transport
system in a gravel-bottomed stream. Pages 293-303 in
Proceedings of the 21st Annual Hydraulics Div. Specialty
Conf. at Montana State Univ., Bozeman. 15-17 August 1973.
Miller, W. H. 1971. Factors influencing migration of chinook
salmon fry (Oncorhvnchus tshawvtscha) in the Salmon
River, Idaho. Doctoral dissertation, Univ. of Idaho.
271 /SECTION X
-------
Moring, J. R. 1982. Decrease in stream gravel permeability after
clear-cut logging: an indication of intragravel conditions
for developing salmonid eggs and alevins. Hydrobiologia
88:295-298.
Moring, J. R., and R. L. Lantz. 1974. Immediate effects of
logging on freshwater environment of salmonids. Oregon
Wildl. Comm. Res. Div. Proj. AFS-58, Final Report, 101 p.
Morrill, C. F. 1972. Migration response of juvenile chinook
salmon to substrates and temperatures. M.S. thesis, Univ. of
Idaho, Moscow.
Munther, G. and G. Frank. 1986a. 1985 fisheries habitat and
aquatic environment monitoring report - Rock Creek drainage
of the Deerlodge National Forest. Deerlodge National
Forest, U.S. Forest Service, in cooperation with Montana
Department of Fish, wildlife and Parks.
1986b. 1985 fisheries habitat and
aquatic environment monitoring report - Lolo National
Forest. "Validation of aquatic habitat quality and fish
population assumptions used to predict effects of manage-
ment activities and evaluation of actual effects". U.S.
Forest Service, Missoula.
. 1986c. 1985 fisheries habitat and
aquatic environment monitoring report - Bitterroot Nation-
al Forest. Bitterroot National Forest, U.S. Forest Service
in cooperation with Montana Department of Fish, Wildlife
and Parks.
Murphy/ M. L., J. Heifetz, S. W. Johnson, K.V. Koski, and J. F.
Thedinga. 1986. Effects of clear-cut logging with and
without buffer strips on juvenile salmonids in Alaskan
streams. Can. J. Aquat. Sci. 43:1521-1533.
NCASI (National Council of the Paper Industry for Air and Stream
Improvement, Inc). 1984a. The effects of fine sediment on
salmonid spawning gravel and juvenile rearing habitat -
A literature review. NCASI Tech. Bull. No. 428.
1984b. A laboratory study of the effects
of sediments of two different size characteristics on
survival of rainbow trout fSalmo gairdneri) embryos to fry
emergence. NCASI Tech. Bull. No. 429.
. 1984c. The relationship between fine
sediment and microinvertebrate community characteristics-
A literature review and results from NCASI five sediment
studies. NCASI Tech. Bull. No. 418.
272 /SECTION X
-------
Nicholson, A.J. 1954. An outline of the dynamics of animal
populations. Aust. J. Zool. 2:9-65.
Nolan, K. M. and D. C. Marron. 1985. Contrast in stream-
channel response to major storms in two mountainous areas
of California. Geology 13:135-138.
Nuttall, P. M. 1972. The effects of sand deposition upon the
macroinvertebrate fauna of the River Camel, Cornwall.
Freshwater Biol. 2:181-186.
Pearce, A. J. and A. Watson. 1983. Medium-term effects of two
landsliding episodes on channel storage of sediment.
Earth Surface Processes and Landforms 8:29-39.
Pearl, R. 1926. The biology of population growth. Williams and
Norgate, London.
Pennak, R. W. and E. D. Van Gerpen. 1947. Bottom fauna production
and physical nature of the substrate in northern Colorado
trout streams. Ecology 28:42-48.
Peters, D. 1986. First year results of the evaluation of the
effect of forest land management on tributaries in Rock
Creek and the Bitterroot River drainages. Rept. Montana
Dept. Fish, Wildlife and Parks. 15 p.
Peterson, N. P. 1980. The role of spring ponds in the winter
ecology and natural production of coho salmon COncorhynchus
klsutch) on the Olympic Peninsula, Washington. M.S. thesis,
Univ. of Washington, Seattle.
Peterson, R. H. and J. L. Metcalfe. 1981. Emergence of Atlantic
salmon fry from gravels of varying composition: A laboratory
study. Can. Tech. Rep. Fish. Aquat. Sci. No. 1020.
Petrosky, C.E. and T.B. Holubetz. 1986. Idaho habitat evaluation
for off-site mitigation record. Idaho Dept. of Fish and Game
Annual Report, Fiscal year 1985. Prepared for US Dept. of
Energy, Bonneville Power Administration, Div. of Fish and
Wildlife, Contract No. DE-AI79-84BP13381, Project No. 83-7.
Phillips, R. W. 1971. Effects of sediment on the gravel
environment and fish production. In Forest Land Uses and
Stream Environment. Oregon State Univ. 1970.
Phillips, R. W., and H. J. Campbell. 1962. The embryonic
survival of coho salmon and steelhead trout as influenced by
some environmental conditions in gravel beds. 14th Annual
Rept. Pac. Mar. Fish. Comm., Portland, OR p. 60-73.
Phillips, R. W., R. L. Lantz, E. W. Claire, and J. R. Moring.
273 /SECTION X
-------
1975. Some effects of gravel mixtures on emergence of coho
salmon and steelhead trout fry. Trans. Am. Fish. Soc.
104(3):461-466.
Pianka, E.R. 1970. On r- and K- selection. Amer. Naturalist
104.
Platts, W.S. 1974a. Stream channel sediment conditions in the
South Fork Salmon River, Idaho. U.S. Forest Service Progress
Rep. IV.
1974b. Georaorphic and aquatic conditions
influencing salmonids and stream classification, with
appliction to ecosystem classification. Surface Environ-
ment and Mining Program, U.S. Dept. Agriculture.
Platts, W. s., and W. F. Megahan. 1975. Time trends in riverbed
sediment composition in salmon and steelhead spawning areas:
South Fork Salmon River, Idaho. Trans. 40th N. Am. wildl.
Conf. Wildl. Mgmt. Instiute. 229-239.
Platts, W.S., R.L. Nelson, 0. Casey, and V. Crispin. 1933.
Riparian-stream habitat conditions on Tabor Creek, Nevada,
under grazed and ungrazed conditions. Pages 162-174 in
Western Proceedings 63rd annual conference of the Western
Association of Fish and Wildlife Agencies; Boise, ID.
Platts, W. S., M. A. Shirazi,and D. H. Lewis. 1979. Sediment
particle sizes used by salmon for spawning with methods for
evaluation. US Environmental Protection Agency. EPA-600/3-
79-043.
Pollard, R. A. 1955. Measuring seepage through salmon spawning
gravel. J. Fish. Res. Bd. Canada 12(5):706-741.
Puckett, K. J. and L. M. Dill. 1985. The energetics of feeding
territoriality in juvenile coho salmon (Oncorhynchus
kisutch). Behavior 92:97-111.
Rabeni, C. F., and G. W. Minshall. 1977. Factors affecting
microdistribution of stream benthic insects. Oikos 29:
33-43.
Reiser, D. W. and T. C. Bjornn. 1979. Influence of forest and
rangeland management on anadromous fish habitat in Western
North America: Habitat requirements of anadromous salmonids.
Pac NW For. and Range Exp. Stn. Gen. Tech. Rep. PNW-96.
Ricker, W.E. 1954. Stock and recruitment. J. Fish. Res. Bd.
Canada 11(5):559-623.
. 1958. Maximum sustained yields from fluctuating
274 /SECTION X
-------
environments and mixed stocks. J. Fish. Res. Bd. Canada
15(5):991-1006.
Ricklefs, R. E. 1973. The economy of nature. Chiron Press, New
York.
Rimmer, D. M., U. Paim, and R. L. Saunders. 1983. Autumnal
habitat shift of juvenile Atlantic salmon (Salmo salar) in a
small river. Can. J. Fish, and Aquat. Sci. 40:671-680.
Ringler, N. H. 1970. Effects of logging on the spawning bed
environment in two Oregon coastal streams. M.S. thesis,
Oregon State University, Corvallis. 96 p.
Ruggles, C. P. 1966. Depth and velocity as a factor in stream
rearing and production of juvenile coho salmon. Can. Fish.
Cult. 38:37-53.
Saiki, M. K. and J. C. Tash. 1979. Use of cover and dispersal by
crayfish to reduce predation by largemouth bass. Pages 44-
48 in D.L. Johnson and R.A. Stein, eds. Response of fish to
habitat structure in standing water. North Central Div., Am.
Fish. Soc., Spec. Publication 6, Bethesda, MD.
Salo, E. 0. and W. H. Bayliff. 1958. Artificial and natural
production of silver salmon, Oncorhynchus kisutch. at
Minter Creek, Washington. Washington Dept. Fish. Res.
Bull. 4. 75 p.
Sandine, M. F. 1974. Natural and simulated insect-substrate
relationships in Idaho Batholith streams. M.S. thesis,
University of Idaho, Moscow.
Saunders, J. W., and M. W. Smith. 1965. Changes in a stream
population of trout associated with increased silt. J. Fish.
Res. Bd. Canada 22 (2):395-404.
Scrivener, J. C. and M. J. Brownlee. 1981. A preliminary analysis
of Carnation Creek gravel quality data 1973-1980. In
Proceedings of a Conference on Salmon Spawning Gravel: a
renewable resource in the Pacific Northwest. Univ. of
Washington, Seattle.
Sedell, J.R. and F.J. Swanson. 1984, Ecological characteristics
of streams in old-growth forests of the Pacific Northwest.
Pages 9-16 in W.R. Meehan, T.R. Merrel, Jr., and T.A.
Hanley, eds. Fish and wildlife relationships in old-growth
forests. Am. Inst. Fish. Res. Biol., Juneau, Alaska.
Sedell, J.R., F.J. Swanson, and S.V. Gregory. 1985. Evaluating
fish response to woody debris. In T.J. Hassler, ed.
Proceedings of the Pacific Northwest Stream Habitat
275 /SECTION X
-------
Management Workshop, October 10-12f 1984. Western Div. Am.
Fish. Soc. and Coop. Fish, Unit, Humboldt State Univ.,
Arcata, CA.
Shapovalov, L, and W. Berrian. 1940. An experiment in hatching
silver salmon (Qn_cp_rhy_nchus kisutch) eggs in gravel. Trans.
Am. Fish. SOC. 69:135-140.
Shapovalov, L. and A. C. Taft. 1954. The life histories of the
steelhead rainbow trout and silver salmon. Calif. Dept. Fish
and Game, Fish Bull. 98, 375 p.
Shaw, P. A. and J. A. Maga. 1943. The effect of mining silt on
yield of fry from salmon spawning beds. Calif. Fish and
Game 29:29-41.
Shelton, J. M. 1955. The hatching of Chinook salmon eggs under
simulated stream conditions. Prog. Fish Cult. 17:20-35.
Shepard, B. B., S. A. Leathe, T. M. Weaver, and M. D. Enk. 1984
Monitoring levels of fine sediment within tributaries to
Flathead Lake and impacts of fine sediment on bull trout
recruitment. Paper presented at the Wild Trout III
Symposium. Yellowstone Nat'1 Park, WY 24-25 Sep. 1984.
Sheridan, W. L. and W. J. McNeil. 1968. Some effects of logging
on two streams in Alaska. J. Forestry 66:128-133.
Shirazi, M.A. and W.K. Seim. 1979. A stream systems evaluation -
An emphasis on spawning habitat for salmonids. EPA-600/3-
79-109, Corvallis Environ. Res, Lab., Corvallis, OR. 32 p.
Shirazi, M. A. and W. K. Seim. 1981. Stream system evaluation
with emphasis on spawning habitat for salmonids. Water
Resources Research 17 (3) :592-594.
Shirazi, M. A., W. K. Seim, and D. H. Lewis. 1981.
Characterization of spawning gravel and stream system
evaluation. Proceedings of the Conf. on Salmon-Spawning
Gravel, Rep.# 39. Wash. Water Res. Research Center, Pullman
Shumway, D. L.. C. E, Warren and P. Doudoroff. 1964. Influence of
oxygen concentration and water movement on the growth of
steelhead trout and coho salmon embryos. Trans. Am. Fish.
Soc. 93:342-356.
Sin, A., P. Crowley, M. McPeek, J. Petranka, and K. Strohmeier.
1985. Predation, competition, and prey communities: A
review of field experiments. Ann. Rev. Ecol. Syst. 16:269-
311.
Silver, S. J., C. E. Warren and P. Doudoroff. 1963. Dissolved
276 /SECTION X
-------
oxygen requirements of developing steelhead trout and
Chinook salmon embryos at different water velocities.
Trans. Am. Fish. Soc. 92(4):327-343.
Solomon, M.E. 1949, The natural control of animal populations,
J. Anim. Ecol. 18:1-35.
Sowden, T. K. and G. Power. 1985. Prediction of rainbow trout
embryo survival in relation to groundwater seepage and
particle size of spawning substrates. Trans. Am. Fish.
Soc. 114:804-812.
Spaulding, S. 1986. A comparison of macroinvertebrate com-
munities from a grazed and rested section of stream in
a livestock/fisheries interaction study site, Big Creek,
Utah. Unpublished report, senior thesis for Fish. 195,
Humboldt State Univ., Arcata, CA.
Stein, R. A. and J. J. Magnuson. 1976. Behavioral response of
crayfish to a fish predator. Ecology 57:751-761.
Stowell, R. A., A. Espinosa, T, C. Bjornn, W. S, Platts, D. C.
Burns, and J.S. Irving. 1983. A guide for predicting
salmonid response to sediment yields in Idaho batholith
watersheds. Unpub. Rept. R1/R4, USDA - FS.
Stuart, T. A. 1953. Spawning migration, reproduction, and young
stages of lock trout fSalmo trutta L.J. Freshwater and
Salmon Fisheries Research, No.5. Scottish Home Dept.,
Edinburgh. 39 p.
1957. The migation and homing behavior of brown
trout (SaImp trutta L.)- Freshw. Salmon Fish. Res.,
Scotland. 18: 27 p.
Stuehrenberg, L. C. 1975. The effects of granite sand on the
distribution and abundance of salmonids in Idaho streams.
M.S. thesis, Univ. of Idaho/ Moscow. 490 p.
Sullivan, K., T. E. Lisle, C, A. Dolloff, G. E. Grant, and
L. M. Reid. 1986. Stream channels — the link between
forests and streams. In: T. Cundy and E. Salo (Eds.).
Proc. Symposium on Forestry and Fishery Interactions,
Feb. 12-14, 1986. Univ. Washington, Seattle.
Tagart, J. V. 1976. The survival from egg deposition to emergence
of coho salmon in the Clear-water River, Jefferson County,
Washington. M. S. thesis, Univ. Washington, Seattle.
. 1984. Coho salmon survival from egg deposition to
emergence. In Proceedings of the Olympic Wild Fish
Conference, J.M. Walton and D.B. Houston, eds. Port Angeles,
277 /SECTION X
-------
Wa., March 1983.
Talbert, D. E. 1983. Anadromous fish habitat monitoring in Pete
King and Deadman Creeks - Annual report. USFS - Clearwater
National Forest.
. 1985a. Anadromous fish habitat monitoring in Pete
King and Deadman Creeks - Progress report, FY '82-84. USFS -
Clearwater National Forest.
. 1985b. Anadromous fish habitat monitoring in Pete
King and Deadman Creeks - Progress report, CY85. USFS -
Clearwater National Forest.
Tappel, P. D. and T. C. Bjornn. 1983. A new method of relating
size of spawning gravel to salmonid embryo survival. North
Am. J. Fish. Mgmt. 3:123-135.
Tebo, L. B., Jr. 1955. Effects of siltation, resulting from
improper logging, on the bottom fauna of a small trout
stream in the Southern Appalachians. Prog. Fish-Cult.
17(2):66-70.
Terhune, L. D. B. 1958. The MARK VI groundwater standpipe for
measuring seepage through salmon spawning gravel. J. Fish.
Res. Bd. Canada 15:1027-1063.
Thurow, R. and D. Burns. Unpublished. Fish density and
embeddedness data. USFS - Payette National Forest.
Tschaplinski, P. J. and G. F. Hartman. 1983. Winter distribution
of juvenile coho salmon fOncorhynchus kisutch^ before and
after logging in Carnation Creek, British Columbia, and some
implcations for overwinter survival. Can. J. Fish. Aquat.
Sci. 40(4):452-461
Turnpenny, A.W.H and R. Williams. 1980. Effects of sedimentation
on the gravels of an industrial river system. J. Fish. Biol.
17:681-693.
Vaux, W. G. 1962. Interchange of stream and intragravel water in
a salmon spawning riffle. U.S. Fish and Wildl. Ser., Spec.
Sci. Rep. - Fish. No. 405.
Volterra, V. 1931. Lecons sur la Theorie Mathematigue de la
Lutte pour la Vie. Gauthier-Villars, Paris.
Vronskiy, B.B. 1972. Reproductive biology of the Kamchatka River
Chinook salmon [Qncorhynchus tschawtscha (Walbaum)]. J.
Ichthyology 12:259-273.
Walkotten, W. J. 1976. An improved technique for freeze sampling
273 /SECTION X
-------
streambed sediments. USDA For. Ser. Res. Note PNW-281.
Washington Department of Fisheries. 1983. WDF letter to
prospective IFIM investigators on recommended substrate
particle size codes for IFIM studies. 12 July, 1983.
Wells, R. A. and W. J. McNeil. 1970. Effect of quality of the
spawning bed on the growth and development of pink salmon
embryos and alevins. US Fish and Wildl. Ser. Spec. Sci. Rep.
Fish. No. 616:1-6.
Wendling, F. L. 1978. Final report on gravel porosity studies
along the trans-Alaska pipeline. U.S. Fish & Wildlife Serv.
Spec. Rept. 18. 77 p.
White, H. C. 1939. Bird control to increase the Margaree River
salmon. Bull. Fish. Res. Bd. Canada 58:30 p.
. 1942. Atlantic salmon redds and artificial spawning
beds. J. Fish. Res. Bd. Can. 6(l):37-44.
White, R.J. 1972. Trout population responses to habitat changes
in Big Roche-a-Cri Creek, Wisconsin. Doctoral Dissertation.
Univ. of Wisconsin, Madison, WI.
Wickett, W. P. 1954. The oxygen supply to salmon eggs in spawning
beds. J. Fish. Res. Bd. Ca. 11:933-953.
. 1958. Review of certain environmental factors
affecting the production of pink and chum salmon. J. Fish.
Res. Bd. Can. 15:1103-1126
Williams, D.D. and J.H. Mundie. 1978. Substrate size selection
by stream invertebrates and the influence of sand. Limnology
and Oceanography 23:1030-1033.
Wilzbach, M. A., K. W. Cummins, and J. D. Hall. 1986. Influence
of habitat manipulations on interactions between cutthroat
trout and invertebrate drift. Ecology 67(4):898-911.
279 /SECTION X
------- |