&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

-------