Steven J. Korel, Wayne A. Hubert, Wyoming Cooperative Fish and Wildlife Research U n P
Unrvers~tyof Wyornmg. Laramie, Wyorn~ng82071
and
Milton G. Parsons, U.S. Foresr Senwe, Wildlife and Fish Ecology Unit. Fon Cdlins, Coloraao 00524
Habitat Features and Trout Abundance Relative Po Gradient
in Some Wyoming Streams
Abstract
Channel gradiant haa been shown to have a negative relation to trout sranding stocks i d a t i n g that aaparation of*\ mn c h e
nch into gndicnt classca may provide a b c m r under~tandingof the rclationship8 bstrrsa habitat and t r o u ~r b u h . Our
major objcctiva w a i to determine if thcra-are significant differencsa in habitat f o n t a m ~ ~atandmg
d
sm& d m u > I 0
mrn between n v o c l ~ r e aof chmnsl gradient; low (0.1-1.4% channel alope) and moderxts(U4.0%). We a h r'
f
- - e
relations between habitat fearurca and trout atanding s t o c h in each class of channel fm uamlterod s t r e a m m z k I
Bow Nntiond For-t. Wyoming. Low-gradiant reacheti were found to have deeper ne-m
water depths, mars nmkmm
AT?
and mare trench pooh than moderate-gradient rapchci, while moderate-gradient re+
had more cobble m b ~ u r rd
mean atandlug stock w u 261-f,$hn in low.gradian1 re.pad: EE
pooh formed by woody debria, and p l u n g s ~ p o o l .
kglha in moderate-gradient reachca. Habitat foaturea corrtlated with trout standings&&
differed between h a two gma-knt
claaacs. Our re~ultndemonnua~athat ~epararionof stream segmcnra into reachen of d a d n i gradient are i m p o m 1 m idenrik
ing featurea of trout hahitat t h a t Ere otherwise obscured by variation over r wider p a d l e n t range.
=
-
.
between h a b i t a t - features and trout stan*
stocks in e a c h otajs of channel.
lntroductlon
Various i n m t i g a t o r s have found the gradient of
stream channels to be negatively correlated with
Study Area
the abundance of brown trout Saimo trutta and
All study strewere in Medicine Bow ?idmad
brook trout Salvelinus fontinalb ( K e m ~ d yand
Forest (MBNJ?l,auuth of Wyoming Hiqhrray 1 3 L
Strange, 1982, Hermansen and Krog 1984,
on the Snowy Range and cast of the Cuatinent~k
Chisholm and Hubert 1986, Fausch in press).-The
Divide on the Sicrra Madre. No native &onids
results suggested that habitat features and strucwere in the st& area, but brook truut and brown
t u r d elements were different among channels o f
trout have b w m introduced. Popufatiom are
differing gradients. In other studies Reeves and
maintained by natural reproduction.
b r d
Everest (1986) used a c h n n n e h p i n g system
trout
inhabit
thc
streams
at
elevations
01 3
(Rosgen 1985) in weJtern Oregon coastal strehrn~
2,450
m.
At
l
o
==
elevations
reaches
zuz Saahand found differences in rnicroha@tnt andrialited by both breok trout and brown m ~ w h k
monid abundance between channels of different
others have o d y brown trout.
p d i c n t s . We felt that the separation of stream
Potential study streams crrendcd 5 -.~a
2233
channeis by gradier-t could provide a better
to 3000 m abmrmean sea lsvcl wid& -' -!e 3a.d
means to understmd the fncrors influencing trout
Platte River draianga in MBNF. Ahm.2 pxecnz
nbundance and to predict brook trout and brown
of all stream dmnnsls in the study r 7 5 u bare a
trout stnndinu- stocks in Rocky Mountain strenms.
low gradient (0J-1.4% channel dapcj;sxd 90 perOur main objective was -.to
- determine if there arc
cent have a ' d e r a t s gmdjam ( l P . a % ) .We;
significant differences in h n b i t = t ~ r ~ - ~ d
selected 15 wmrsraheds thai hnd been .:sinimal+
c
7
1
O
tot&len@h
b-.-7
.
+ -.- affected by human activity with no : r i s ~ g
&
etween two cl~iEi-ofdKO.nncT@ h e ~ : t ; low
clearcut l o g p a g overgrazing, or mrb,?-q. h t n (0.1-1.4%- c h & d
slope) and moderate (1.54.0%). We also d e t e r m i n e d ~ 4 ^ t a l i s i ~ ~ ~ a ~ n tial
n study r e a d i n ware nelected horn ;-!.SFFores~
Service records and U.S. Geologic 5.:top
.. ogrmphic
m
a
p
Low-gradienl
d
:
r
d
e
r
n
tc'The Unit k joindy a&orted by the Uni*enity of Wyoming,
p
d
i
e
n
t
r
e
a
c
h
at
laast
200
m
i
n
k:&
rm
tha Wyoming Cams m d F i h Deparrmenr and tho'U.S. Fuh
marked on the maps. Speciiic sarnp-sirts
.
l a d Wildlife S a m c a .
..
-
-
Northwest Sciwec, Vol. 63 -No. 4. 19t.'?
I
--
175
I
+
selected to represent the range of drainage hasin
areaB (0.7-100 kml)and elevations o c c u r r i n ~in
the stddy area. Selected study reaches and the
upstream watershed area were ground truthed to
confirm thelack of human effects. Channel p a dients were assessed with a surveyor's level prior
to field sampling.
A total of 16 l o w - p d i e n t reaches and 32
moderate-g-radient renches were selected for
study. The 16 low-gradient rcaches were classified
a6 C3 channels using Rosgen's (1985) channel
classification system, while the moderatt-gradient
rcaches were classified as B2 and B3 with 16
reaches in each class. Low-gradient reaches
ranged from 2420 to 2975 m above mean sea
level and tended to be in either headwater or
foothill region^. The wettcd widths during lute
summer ranged from 0.4 to 6.8 m {drainage basin
area range = 0.7-95.0 kma).Moderate-gradient
reaches ranged from 2377 to 2963 m in tlevation and had wettcd widths of 1.1 to 9.3 m
(drainage basin area range = 2.2-96.2 kml).
The underlying geology and the riparian zonc
contributed to differences between the two gradient classes. The low-gadient channels were
almost exclusively in meadow areas with nlluvial
soils where large trees were lacking. but dense
growths of willows, sedges, and grasses allowed
overhanging banks to form as meander processes
occurred. Subsequently, low-gradient channels
tended to have undercut banks and trench pools
with accumulation of fine sediment substrate that
could support aquatic vegetation, but woody
debris was generally lacking. In contrast, mode r a t e - p d i t n t channels were almost exclusively in
coniferous forest, where large trees were present
at or near the water's edge, with geologic formations preventing meander processes from having
a pronounced efiect. So moderate-gradient chnnnels tended t o have more dammed pools and
plunge pools formed by geologic formntions and
woody debris, coarser substrates with less fine
sediment, more riffle habitat, limited amounts of
undercut bank and trench pools, and less aquatic
vegetation than low-gradient channels.
Methods
We used the transect method to measure stream
habitat features over 200-rn reaches (Platts et al.
1983). Riffles marked the upper and lower ends
of study reaches. In 1985 we spaced transects 2 m
176
Kozel, Hubert, and Parsons
apart and made point measurements at 0.3-mintervals; in 1986 the intervals were 4 rn apart and
the point measurements were made at seven
equally spaced locations not including the banks.
Watcr+depths were measured at each point;
ghorc-depth measurements were excluded fram
calculutions of mean depth. At each point interval across a transect, we visually classified cover.
dominant substrate, cmbeddedness, and microhabitat. Cover %.as classified into three categories: aquatic vegetation, woody debris, or
boulder. Substrate was classified into six cntegories (after Platts e l nl. 1983; diameter in parcntheses): s m d fine sediment (dinmeter 5 0.8 mm),
larae fine sediment (0.94.7 mm),g a v e l (4.8-76.0
mm), cobble (77-304mm), smaU boulder (305609
mm), a n d large boulder (1610 mm). Embeddedness (amount of fine sediment surrounding the underlying substrate particles) was rated
from zero to 100% ( P l a t t ~et al. 1983). We
clnssified habitat types according to Bisson el a1
(19e2). Shore depth, undercut bank, and overhanging vegetation were mewured at each bank
with a measuring staff. Bank angle was determined with a clinometer (Platts et al. 4983).
We measured the length of each pool and riffle and noted the physical structure associated
with the formation of the pool-such as geologic
formations (bedrock, boulders) or woody debris
(log check dams, natural log deflectors, and
debris jams). We also computed the percentages
of pool habitat created by geologic controls and
woody debris; most woody debris probably overlaid geologic formations, but where it was the evident feature governing pool formation, the location Wa6 identified.
hut mq
Pearsop
were pt
bctwee
stocks
analyst
differe
gradiel
ad.
197
providl
results
pende~
functic
foundi
variab
were L
variat.
identi
tion i~
ences
'
Res~
Habi;
Twer,
deter
low-g
(Tab1
into
subs1
nel r*
-
W e used the removal method of DeLury
(1951) to estimate trout standing stock over the
study renches during late summer and early
autumn while the streams were at base flow. Each
reach was blocked at the upper and lower end
with small-mesh seines to prevent upstream or
downstream movemcnt of trout. Three depletion
passes were made with a Coffelt Model BP-2
backpack electroshocker. Fish from each pass
were individually weighed to the nearest gram
and measured (total length) to the nearest
millimeter. Only fish longer than 100 mm were -.
used in population estimates. We used program
CAPTURE of White et al (1982)to estimate a o u t
populations, choosing model M(bh) because it
allowed for vnriabdity in behavioral responses of
fish to the first capture attempt.
wert
grad
-.
/
I L-
'
/'
c
(
s
t
);;5,/ .,
L
1,
I
.
I,
,,, ,
\
%
; t
,,,.";,
-
stu6
deej
ban)
to i
-- grar
tion
. , bet1
cha
' and
<
,
\
!
-,
.
-
cla:
, -;
'
'.+
I
.
we1
nd
*
pinG
. from
jn tercover,
nicrocace.
IS, or
: mte?arena
3 mml,
3.76.0
35409
. Em-oundj rated
;). W e
n e l d
i overn bank
deter?83).
.ciatcd
-9logic
debris
3, and
rttages
-is and
:. overhe evie locaeLuy
er the
early
. Each
:r end
am or
~ierion
B P-2
1 pass
,Tam
carest
1 were
'gram
:trout
use it
lses of
We used Scheffe's test for comparison of
means between low-gradient and moderategradient reaches (Zar 1984). An arcsine transformation was performed on all proportional data,
but means of measured values were reported.
Pearson product-moment correlation analyses
were performed to determine univariate relations
between stream habitat features and standing
stocks of trout. Stepwise discriminant function
analyses were performed to identify the major
differences in habitat features between lowgradient and moderate-gradient channels (Nie e t
a1 1975). The univariate comparisons of means
provided insight into habitat differences, but the
resulrs were confounded by the lack of independence among the variables. Discriminant
function analysis was used to eliminate the confounding effect of intercorrelation among habitat
variables. Stepwise multiple-regression analyses
were used to develop models that accounted for
variation in trout standing stock using variables
identified as important in the discriminant function andyses p i e e l nL. 1975). All statistical difierences were accepted as significant at P s 0.05.
Results
Habitat Differences
Twenty-nine habitat variables were analyzed to
determine if they differed in abundance between
low-gradient and moderate-gradient channels
(Tahle 1). The habitat variables were separated
into four group-channel morphology, cover,
substrate and habitat type-based on the channel features they described.
Four of the six channel-morphology variables
were significantly different between the lowgradient and moderate-gradient channels that we
studied (Table I). The low-gradient channels were
deeper and narrower, a larger proportion of the
bank adjoined deep water, and the banks tended
to have more overhang than the moderategrndient channels. The best discriminant function to describe differences in habitat features
between low-gradient and moderate-gradient
channels used two variables, mean shore depth
and the percentage of bank with a bank angle
< 45'. Thie discriminant function properly
clartsified 39 of 48 study reaches.
Of the eight cover features memured, three
were ~ i ~ c n n tdifferent
ly
between the two channel types (Table 1). Both undercut banks and
aquatic vegetation were more abundant in lowgradient reaches than in moderate-gradient
reaches. Woody debris formed an average of only
4 percent of the pool habitat (includes d pool
types) in low-gradient renches, but 24 percent in
moderate-gradient reaches. Discriminant function analysis identified three cover variable8 as
the best to disringuish between the two channel
types-percentage of undercut bank, pool habitat
30 cm deep, and woody debris forming pool
habitat. This discriminant function correctiy
classified 43 of 48 study reaches.
As would be expected, substrate features differed between the two gradient classes (Table 1).
The abundance of large and small sediment was
greater in low-gradient channels, a9 we11 as
embeddedness, while cobble was more abundant
in modernre-gradient channels. Only one substrate variable, cobble, was identifled in the
discriminant analysis and this variable torrectly
classified 41 7 : -18 ~r.:dyreaches.
The a b u n ~ ~ x ~,I c:c:ur habitat types differed
significantly between the two classes of channel
gradients (Table I). Low-gradient reaches tended
to have more trench pools, but fewer riffles.
plunge pools, and dammed pools than in
moderate-gradient reaches. The best discriminant function to describe habitat differences between the two gradient classea included trench
pools, secondary channel pools, and plunge pools.
This discriminant funcrion properly classified 45
of 48 study reaches.
Trout Standing Stocks
The mean standing stock of trout differed
si&~caatly between low-gradient and moderategradient reaches. The menn in low-gradient
reaches was 267 kgiha (SD = 175, range =
50611) nnd in moderate-gradient renches was 102
kgiha (SD = 49, range = 36-204).
Brook trout and brown trout standing stocks
in low-gradient and in moderate-gradient reaches
were evaluated separateiy where allopatric and
sympatric populations occurred (Table 3). Among
dopatric populations, mean brook trout standing
stocks were more than twice as great in lowgradient than in moderate-gradient; however,
mean brown trout standing stockv were similar,
but the sample sizes were small with only 3 and
4 reaches in each category. Among sympatric
populatiom of brook trout and brown trout,
Habitat Features and Trout Abundance Relative to Gradient - - 177
\
TABLE 1. hleans for channel morpholog, ntrenm cover, substratc, and hahitnt variables mennured in
-
low-gradient (0.1-1.4%) and moderatn-jpditnt (154.0%) atraarn renchea on thc Medicine Bow
Fiationnl Forest, Wyoming. Amxisk indica~c*a significant dificrence using Schcfir'e test (P
5 0.05).
>;
,,I,::\
c,,.'
Varidle
,' , i,
Channel gradient (%)
'
:
;
'.'.\ '
1.5-4.0
0.1-1.4
Mtan
Rnngt
Mean
Channcl
padient
Range
Channel morphology
-
Water depth (cm)'
Bank rngle < 90' (70)'
Bank angle
<
45" (%)'
Modcrate
Kidth-LC-depth r a ~ i o
Shore dcpth (cm)
Shore dcpth t 1 5 2 cm ( S b )
Cover
-
--.
Undercut bank (%Im
Overhmging vegetation (70)
Boulder cover (%)
r o o d ! covcr (70)
Aquatic vegetation cover (70).
Pool r 30.5 cm deep (Sb)
Pools with cover (70)
Woodp dchr& control (%)*
Substrate
Large boulder (%)
SmaU boulder (%)
Cobblc ('5)'
Gravel (70)
Large fine sediment (%)'
S m d l fine scdimcnt (%)'
-
Embcddednms (A)*
Habitat type
Riffle (5)'
Rapid (70)
Secondary channel pool (70)
Backwater pool (%)
Trench pool (7'0)'
Plunge pool (To)'
Dammed pool
Glide ( S o )
brook trout standing stocks were about three
times greater in low-gradient channels and standing stocks of both species together averaged almost f o u r times greater in low-gradient reaches.
The habitat features that were correlated with
trout abundance differed between low-gradient
178
Low
Kozel, Hubert, nnd Parsons
and moderate-gradient reaches (Table 3). Total
standing stock increased with increasing amounts
of deep nearshore area. undercut bank, overhang.
ing vegetation, trench pools, and plunge pools,
but decreased with incrwing water dcpth, widthto-depth ratio, and riffle abundance a m 0
I
r
TABLE L Weun standing stocks (kgiha) of d o p a m c and ~ r m .
patric population of brook trout and brown trow
in low-gradient and moderate-gmdicnr stream
reaches on the Medicine Bow National Forcsr,
Wyoming.
Allopatnu
Brook
trout
Brown
trout
Brook
trout
Brown
trout
Both
trout
Mean
283
128
352
103
61
8
3
224
SD
175
34
190
6
6
6
57
16
4
43
42
85
33
23
14
18
14
Channel
g~adient
Low
n
hioderare
Mean
SD
N
7
131
42
14
14
low-gradient reaches. In the moderate-gradient
reaches, standing stock warr positively correlated
with deep nearshore area, overhanging vegetation, aquatic vegetation, fine sediment, embcddedness, and glide habitat, but negatively correlated with width-to-depth ratio, deep pools,
boulder and cobble substrate, and backwater
pools. A few variables werc correlated with trout
abundance in both gradient cIasses; deep nearshore areas and overhnnging vcgetation were
positively correlated, while width-to-depth ratio
was negatively correlated.
Brook trout standing stocks werc found to be
correlated with only one habitat feature among
low-gradi~ntrenches (Table 3); as width-to-depth
ratio increased brook trout abundance declined.
TABLE 3. Statintically significant (P 5 0.05) correlation cocfficienb for relations batwcen trout standing atock .
and measured habitat faaturea in low-gradient and moderate-gradient rcachea on tho Medicine Bow
Nntionnl Foruat, Wyoming.
Channel gradient (%)
*
1.0-1.4
-
Total
trout
(11x16)
Variable
Brook
t-out
(nm13)
1.54.0
Brown
trout
(~39)
Total
UDU
t
(n=32)
Brook
uout
(n=28)
Brown
aour
(n=18)
Channel morphology
Water depth (cm)
Widtheta-depth ratio
Bank angle < 90' ('70)
Shore Depth z 15.2 cm (7%)
Cover
Undercut bank ( 5 )
Overhnnfig vegetation ( A )
Aquatic vegetation cover ( 5 )
Total
)UKltB
1ang'0019,
. idth.
:the
Pool habirar 2 30.5 cm deep (%)
Pooh with cover ( 5 )
Substrnto
Small boulder ( 5 )
Cohblc (%)
Ljrga fine aadirnenr (%)
Small fine sediment (%)
Embeddcdnsas (%)
Riffle (%)
Backwater pool (%)
Trench pool (%)
Plunge pool (%)
Glide (%)
Habitat Features and Trout Abundance Relative to Gradient
179
------
--
-m ,
-
L
Several habitat features were correlated with
brook t r o w abundance in moderate-gradient
reaches. In r n ~ d ~ ~ ~ t e - ~ areaches
d i e n t brook
trout abundance increased with greater proportions of shore depth 2 152 cm deep, overhanging vegetation. aquatic vegetation, aquatic vegctntion cover, large and fine sediment, embeddedntss, and glide habitat. Brook trout abundance
went down a s width-to-depth ratio increased and
the abundance of pool habitat 2 30.5 cm deep
increased.
The habitat features correlated brown trout
standing stocks differed from those related to
brook trout abundance. Among low gradient
reaches, brown trout standing stocks increased
with greater proportions of aquatic vegetation
cover, p o l s with cover, boulder covcr, backwater
pools, ~ e n c pools,
h
and plunge pooh. Within lowgradient reaches, brown trout abundance was
negatively correlated with water depth and the
amount of pool habitat 2 30.5 cm deep. A different set of variables was correlated with brook
trout abundance in moderate-gradient reaches;
the only habitat feature positively correlated with
abundance was water depth. Among moderategradient channels, brown trout standing stock
declined as the abundance of bank with an angle
< 90°, pools with cover and glide habitat
increased.
The discriminant function analyses of habitat
differences between low-gradient and modcrateg-radient reaches identified nine variables thnt
can b e used to separate the two channel types:
mean shore depth, the percentages of bank with
a bank angle < 4S0, undercut bank, pools
formed by woody debris, cobble substrate, trench
pools, seconday channel pools and plunge pools.
These nine variables were used in a stepwise
multipie.regression analysis of all 48 stream
reaches to determine which ones may be of most
value in accounting for variation in trout standing stocks. Two regression models that accounted
for substantial variation in standing stock of trout
(S) were identified. The Fust equation (K2 = 0.64)
included undercut bank (U), trench pools (T) and
pools 2 30 cm deep (P):
The second equation (R1= 0.67) included trench
pools and pools 2 30 cm deep, as well ns secondary channel pools (S) and mean shore depth (D):
180
Kozel, Hubert, and Parsons
These two equntions iIlustrate that habitat
variables included in habitat aasessment modeis
arc to a great degree separating stream reaches
into low-gradient and moderate-gradient classes
and accounting for variation due to. differences
between the two classes. When the nine variables
Was used in stepwise multiple-regression analyses
involving only reaches in individud gradient
classes, there were fewer variables that accounted
for variation in trout standing stocks and the
amount of variation that was accounted for was
much less. The equation which beat accounted
for variation in trout standing stocks among lowgradient channels (R' = 0.36) had only one
variable, undercut bank; the best in moderatedgradient channels (R1= 0.15) had two variables,
pools < 30 cm deep and mean shore d e p t h Lesa
variation in both standing stock estimates and
measured values of the habitat variables within
a gradient class contributed to the reduced ability
of the habitat variables to account for variation
in stnnding stock.
f
c
compo
index
(19871,
variab
of Scru
that th
are to
changr
chann~
Ou
related
gradie:
as bet,
beiievt
into re.
idcntii
abundi
tion o.
model!
define
they a:
for spe
Discussion
Litera
Channel morphology, cover, substrate, and---.. .
habitat types differed between low-gradient
(0.1-1.4%) and moderate-gradient (1.54,0%)
/i
channels sampled in MBNF. We also found dif- /'
ferences in standing stocks of trout in channels
.
- _!!' 6
&
with low and moderate gradients.Wc-belleve the
greatey,.starrdjng stocks-of trout in low-gadient :
reach& wag t to a g e a t e r abundance of pool
habitat $ia&
!' cr
in hew reaches. Total poo
habitat of all types was more abundant in lowgradient reaches (mean = 41% of transect
points) than in moderate-gradient reaches (mea
=I 30%). Also within low-gradient reaches, cov
within p o l habitat-undercut bank and aquatic
vegetation-tended to be more abundant than
in moderate-gradient channels.
\
Measures of the nine habitat variables that
we found to discriminale between low-gradient
and moderate-madient channels are frequently
found in trout-habitat models developed in the
central Rocky Mountains. For example, they are
incorporated in the cover and substrate atmbutes
of the habitat p a l i t y index of Binns and Eiserman (1979), the overhead bnnk cover and instream rubble-boulder-aquatic vegetation variables of the trout cover rating of Wtscha
(1980), the overhead bank covcr and dnap wat2
.
1
Binnu,
P
f
c
Bisrbn. I
I
5
B
(
,
I
1
DeLury
odah
I
1chc5
1
ants
:nccs
ables
Liyses
aient
mted
i the
r
wag
lnted
;lowone
-ated~bles,
-
Less
; and
~ithin
.hility
ation
and
aient
Binna.
;.o%)
d dilnneis
,e the
aient
pool
pooI
1 low'
nsect
mean
20ver
-uatic
than
components of the modified habitat suitnbility
index model for brown trout by Wesche c t aL
(1987), as well as the substrate and undercut bank
variables in the production and biomass models
of Scarnecchia and Bergersen (1987). We believe
that these variables in the various habitat models
are to a great degree accounting for natural
changes in habitat resulting from differences in
channel gradient.
Our data show that habitat features correlated with trout abundance differ between lowgradient and moderate-gradient channels, as well
as between brook trout and brown trout. We
believe that the separation of stream segments
into reaches with similar gradient may enable the
identification of features related to salmonid
abundance that are otherwise obscured by variation over a wider range of gradient. Habitat
models that predict salmonid abundance should
define the range of channel gradient to which
they are applicable, and perhaps be developed
for specific gradient classes or channel types, as
.
that
Gent
.cntly
n the
:y are
-.butes
;Iserd invarfsche
water
i
N. .4., and F. M. Eiserman. 1979. Quantification of
fluvial trout habitar in Wyoming. Tmna. h e r . F i a b
SOC. 108:215.228.
Bisron. P.A., J. L. Nielaea, R. A. Palmason, a d L E. Grove.
1982. A system ot' naming habitat typca in a m d
streams, with c r a m p l e ~ of habitnt utilization by
aaimonidn during low s t r a a d o w . I n N. B. 4.tmmuout,
(cd.). Acquiaition and utilizntion of aquatic habitat
inventory information. hmcrican Fiaherica Sociery,
Western Division. Betheadn. M a ~ i a n d .Pp. 62-73.
Bousru, M. F.'1954. Relationship between trout popula~iom
and cover on s $mall straam. J. WildL Manage.
18229-239.
Chisholm. I. M. 1985. Wintar arrcam conditionn and brook
trout habitat uae on tho Snowy Range, Wyoming.
Univ. of Wyoming, Laramie. M.S. Theaia.
Chi~hoim,I. M., md W. h Hubert. 1986. I d u e n c e of straam
gmaient on standing stock of brook r o u t in the Snowy
Range, Wyoming. Northw. Sci. 60:137-139.
DeLury, D. B. 1951. On tha planning of experiments for
entimation of Gah papulatiom. J. Fish R a Board Con
8681-307.
,
DeVora, P. W.. and R J. White. 1978. Daytime reaponaca
of brown uour (SaLmo trrta) to cover stimuli in atream
channel. Trans. Amcr. Fiab. Soc. 107:763-771.
Dixon. W. J. 1981. BMDP statisricnl software. Univcnity of
California Preaa. Berkalq.
Faulch. K. D. In preaa. Do grndient a n d temperature affect
diatributionr of. and iaterncrionr betwen. brook char
(SdvcLinrr~f o n t i d i s ) and other resident aalmonida
in atrearna? In Proceedings of tho International S y m
posium on Cham and Manu Salmon, October 1988,
Supporo, Jnpan.
well as defining the species of trout fo; which
they are applicable.
Accessibility of streams to the public in
southeastern Wyoming has been related to trout
standing gtocks (Wesche et aL 1987), with standing stocks declining as accessibility increases. We
expect that the trout standing stocks among our
study sites were greater than in many other
streams in MBNF because the sites tended to be
temote without nearby road access, thus fished
less. Comparison to data from Binns and Eiserman (1979), Chkholm (1985) and Lanka (1985)
confirm this belief.
Acknowledgments
We thank the staff of the Medicine Bow National
Forest, especially A. Baucr, L. Frary, N. Schmal,
and M. Wilcox, for assistance and supportj and
I. Chisholm and T.Wcsche for critical review of
the manuscript. Funding was provided by the
U.S. Forcst Service.
Hermamen. H., and C. Krog. 1984. Influcnca of physical factorn on d e n s i v o f stocked brown trout (Sdrno trurta
jario L.) in a Daniah lowland stream. Fish. Manage.
15:107-110.
Kennedy, G. J. A, and C. D. Strnnge. 1982. The distribution
of d m o n i d r in upland s t r e a m in relnrion to depth
and gradient. J. Fiah Biol. 20:579-591.
Lanka, R. P. 1985. Modeling trout atanding stock in s m d
Wyoming s t r e a m b w d upon easily m e w e d habirat
and geomorphologicd characteristics. Univ. of Wyoming. Laramit. MS. Thosir.
Lanka, R. P., W. A. H u b e r t and T.A. Wescha. 1987. Reintionr of geornorphology to atrsam habitat and trout
standing stock in rmall Rocky Mountain atrsam.
Trans. h e r . Fish. Soa. 116:21-28.
Lewia, S. L 1369. Phyaicnl facrorr iaflucncinq fish populations in pools of a trout stream. Tram. Amcr. Fish.
Soc. 98:14-19.
Nie. Y. H., C. H. Hull, J. G. Jcnkana, K. Steinbrcnner, and
D. H. Bear. 1975. SPSS ( S t a t i s t i d package for the
aocial acicnccs), 2nd edition. McCraw Hill, New York.
Platts, W. 5.. W. F. Megahan, and G. W. MinahalL 1983.
Msthodr for evaluating atroam, riparian, and biotic
conditionr. U.S. Forcst Service General Tcchnicnl
Repon INT-138, Ogden. Utnh.
Reeves, G. H., and F. H. Evareat 1986. Stream type as a
predictor of habitat conditions and anadromoua
sdmonid papulatiom in wmtern Oregon stream. US.
Forast S e ~ c a Pncific
,
Nonhwcat Reaearch station.
Corvallia, Oregon.
7.
Habitat Features and Trout Abundance Relative
Gradient
181