AN
ANALYSIS
OF
THE
EFFECTS
OF
TEMPERATURE
ON
SALMONIDS
OF
THE
PACIFIC
NORTHWEST
WITH
IMPLICATIONS
FOR
SELECTING
TEMPERATURE
CRITERIA
December
2000
Kathleen
Sullivan
Sustainable
Ecosystems
Institute
Portland,
Oregon
Douglas
J.
Martin
Martin
Environmental
Seattle,
Washington
Richard
D.
Cardwell
Parametrix
Inc.
Kirkland,
Washington
John
E.
Toll
Parametrix
Inc.
Kirkland,
Washington
Steven
Duke
Weyerhaeuser
Company
Tacoma,
Washington
6
X
V
W
D
L
Q
D
E
O
H
(
F
R
V
\
V
W
H
P
V
,
Q
V
W
L
W
X
W
H
3
9
3
8
6
:
7
D
\
O
R
U
V
)
H
U
U
\
5
R
D
G
3
R
U
W
O
D
Q
G
/
2
U
H
J
R
Q
<
:
5
4
<
7
H
O
=
8
3
6
0
5
7
9
0
8
3
3
;
H
0
P
D
L
O
=
V
H
L
#
V
H
L
1
R
U
J
)
D
[
=
8
3
6
0
5
7
9
0
9
<
3
8
:
H
E
V
L
W
H
=
Z
Z
Z
1
V
H
L
1
R
U
J
ii
iii
TABLE
OF
CONTENTS
Table
of
Contents.......................................................................................................................................................
II
Abstract
....................................................................................................................................................................
IV
Acknowledgments
....................................................................................................................................................
V
Section
1
Introduction
and
Objectives.............................................................................................................
1­
1
Section
2
Review
of
The
Physiologic
Response
of
Fish
to
Environmental
Temperature
............................
2­
1
ACUTE
TEMPERATURE
EFFECTS
...........................................................................................................................
2­
3
SUBLETHAL
TEMPERATURE
EFFECTS
...............................................................................................................................
2­
4
ASSOCIATING
BIOLOGICAL
EFFECTS
AND
STREAM
TEMPERATURE
..........................................................................
2­
8
ESTABLISHING
WATER
QUALITY
CRITERIA
...........................................................................................................
2­
10
CONCLUSIONS....................................................................................................................................................
2­
10
Section
3
Temperature
Characteristics
of
Streams
Used
in
analysis.............................................................
3­
1
TEMPERATURE
DATA............................................................................................................................................
3­
1
TEMPERATUE
CHARACTERISTICS
..........................................................................................................................
3­
2
COMPARISON
OF
STUDY
SITES
TO
OTHER
TEMPERATURE
STUDIES
IN
WASHINGTON,
OREGON,
AND
IDAHO.............
3­
5
TEMPORAL
VARIATION
AT
A
SITE
...........................................................................................................................
3­
9
TEMPERATURE
INDICES
........................................................................................................................................
3­
9
CONCLUSIONS....................................................................................................................................................
3­
11
Section
4
Assessment
of
Risk
of
Salmon
Species
to
Acute
Temperatures
in
Streams
and
Rivers..............
4­
1
INTRODUCTION.....................................................................................................................................................
4­
1
ACUTE
THERMAL
EFFECTS
CURVES
ASSOCIATED
WITH
50%
MORTALITY................................................................
4­
2
COMPARISON
OF
LT50
AND
LT10
MORTALITY
RELATIONSHIPS
..............................................................................
4­
4
ACUTE
THERMAL
EFFECTS
CURVES
AT
10%
MORTALITY
.......................................................................................
4­
6
ACUTE
EXPOSURE
CHARACTERIZATION.................................................................................................................
4­
7
ACUTE
RISK
CHARACTERIZATION
..........................................................................................................................
4­
9
DISCUSSION.......................................................................................................................................................
4­
12
CONCLUSIONS....................................................................................................................................................
4­
13
Section
5
Development
and
Corroboration
of
a
Bioenergetics­
Based
approach
To
Evaluating
Salmon
Growth
in
Relation
to
Environmental
Temperatue..............................................................................................
5­
1
INTRODUCTION.....................................................................................................................................................
5­
2
GROWTH
MODEL
..................................................................................................................................................
5­
4
Consumption                                           ..
5­
5
Specific
Growth
Rate                                       
.5­
11
Growth
Model                                          ..
5­
14
APPLICATION
OF
THE
METHOD
FOR
PREDICTING
GROWTH
IN
NATURAL
STREAMS..................................................
5­
14
COMPARISON
OF
MODEL
PREDICTIONS
TO
OBSERVED
FISH
GROWTH...................................................................
5­
22
DISCUSSION.......................................................................................................................................................
5­
30
CONCLUSIONS....................................................................................................................................................
5­
33
Section
6
Quantifying
Growth
Effects
in
Relation
to
Temperature
Thresholds
...........................................
6­
1
INTRODUCTION.....................................................................................................................................................
6­
2
GROWTH
SIMULATION
METHOD.............................................................................................................................
6­
3
TEMPERATURE
EFFECTS
ON
WEIGHT
GAIN............................................................................................................
6­
5
RELATIVE
TEMPERATURE
EFFECTS
ON
GROWTH
...................................................................................................
6­
5
RISK
ASSOCIATED
WITH
GROWTH
LIMITATION
.....................................................................................................
6­
10
GROWTH
LOSS
AND
TEMPERATURE
CRITERIA......................................................................................................
6­
12
DISCUSSION.......................................................................................................................................................
6­
12
CONCLUSIONS....................................................................................................................................................
6­
14
iv
Section
7
Relationship
between
Existing
and
Proposed
Temperature
Criteria
and
Risk
Assessment
Findings..................................................................................................................................................................
7­
1
CRITERIA
AND
METHODS
TO
DERIVE
THEM............................................................................................................
7­
2
Experimental
Information­
based
Method
(
EPA)                            ...
7­
2
Field
Observation
Methods                                     ..
7­
5
Risk
assessment
Approach                                    ..
.7­
6
Review
Approach                                         ..
7­
8
TEMPERATURE
CRITERIA
.....................................................................................................................................
7­
9
DISCUSSION.......................................................................................................................................................
7­
13
CONCLUSIONS....................................................................................................................................................
7­
14
Section
8
A
Discussion
of
the
Scientific
and
Management
Implications
of
Findings
THE
REGULATORY
CONTEXT
OF
TEMPERATURE
CRITERIA......................................................................................
8­
1
THE
CHARACTERISTICS
OF
TEMPERATURE
CRITERIA
.............................................................................................
8­
2
THE
BASIS
FOR
DERIVATION
OF
TEMPERATURE
CRITERIA
......................................................................................
8­
3
QUANTITATIVE
ANALYSES
TO
ASSESS
THE
EFFECTS
OF
TEMPERATURE
ON
FISH
IN
NATURAL
ENVIRONMENTS..........
8­
4
THE
SCIENTIFIC
BASIS
FOR
TRANSLATING
QUANTITATIVE
BIOLOGICAL
ANALYSES
TO
TEMPERATURE
CRITERIA
........
8­
8
TEMPERATURE
THRESHOLDS
BASED
ON
RISK
ASSESSMENT
..................................................................................
8­
9
TEMPERATURE
CRITERIA
DERIVED
FROM
RISK
ASSESSMENT...............................................................................
8­
11
UNCERTAINTIES
IN
APPLYING
CRITERIA
IN
NATURAL
ENVIRONMENTS....................................................................
8­
13
CONCLUSIONS....................................................................................................................................................
8­
14
Section
9
Literature
Cited..................................................................................................................................
9­
1
Section
10
Appendices.....................................................................................................................................
10­
1
APPENDIX
A..
DATA
USED
TO
DEVELOP
RELATIONSHIPS
BETWEEN
GROWTH
RATE,
TEMPERATURE,
AND
CONSUMPTION
DETERMINED
FROM
LABORATORY
STUDIES.
...............................................................................
A­
1
APPENDIX
B.....
TEMPERATURE
DATA
FOR
THE
PACIFIC
NORTHWEST
REGION,
FROM
U.
S.
GEOLOGICAL
SURVEY
WATER
RESOURCES
DATA.
................................................................................................................
B­
1
APPENDIX
C
ACUTE
EFFECTS
OF
TEMPERATURE
ON
SALMON
AND
TROUT:
DATA
USED,
ANALYSES
AND
ASSUMPTIONS...................................................................................
UNDER
SEPARATE
COVER
v
ABSTRACT
To
administer
the
Clean
Water
Act,
the
U.
S.
Environmental
Protection
Agency
and
state
water
quality
agencies
throughout
the
nation
have
adopted
numeric
and
qualitative
criteria
that
establish
environmental
conditions
known
to
protect
aquatic
life
from
adverse
effects.
Pacific
Northwest
states
have
adopted
temperature
criteria
designed
specifically
to
protect
fish
with
emphasis
on
salmonid
species
because
water
temperature
plays
a
role
in
virtually
every
aspect
of
salmon
life.
Adverse
levels
of
temperature
can
affect
growth,
behavior,
disease
resistance,
and
mortality.
In
recent
years,
the
EPA
and
National
Academies
of
Science
and
Engineering
have
promoted
risk
assessment
techniques
to
develop
water
quality
criteria,
including
formal
protocols
that
have
been
peer
reviewed
nationally.
Risk
assessment
is
designed
to
combine
the
information
from
biological
studies
with
an
analysis
of
each
population's
exposure
to
quantified
effects.
Risk
occurs
when
the
stress'
magnitude,
frequency
and
duration
exceed
the
species'
ability
to
deal
with
that
stress.
A
risk­
based
approach
seems
ideally
suited
to
developing
criteria
for
and
assessing
temperature
risk
to
fish
because
exposure
has
been
well
documented
through
temperature
monitoring
and
extensive
research
on
the
lethal
and
sublethal
effects
on
salmon
physiology
has
been
conducted
over
the
past
40
years.
Nevertheless,
risk­
based
approaches
have
not
yet
been
used
to
establish
temperature
criteria
in
recent
state
agency
reviews
of
water
quality
standards.

In
this
paper
we
develop
a
risk­
based
approach
to
analyze
summertime
temperature
effects
on
juvenile
salmon
species.
We
use
available
research
findings
to
quantitatively
evaluate
the
biological
effects
of
temperature
in
combination
with
measured
stream
temperature
ranging
from
very
cold
to
very
warm.
Many
currently
exceed
Washington's
temperature
standard.
Acute
risk
to
high
temperatures
was
assessed
using
laboratory­
derived
values
of
mortality
in
relation
to
duration
of
exposure.
Despite
warm
temperatures,
the
risk
analysis
found
that
direct
mortality
from
temperature
is
unlikely
in
the
range
of
temperature
in
study
streams
because
temperatures
high
enough
to
cause
mortality
are
either
never
observed,
or
occur
over
too
short
of
periods
of
time
to
cause
death.
The
analysis
suggested
that
there
is
little
or
no
risk
of
mortality
if
annual
maximum
temperature
is
less
than
26oC,
although
site­
specific
analyses
are
suggested
when
annual
maximum
temperature
exceeds
24oC
to
affirm
this
result
in
local
river
conditions.
Short­
term
occurrence
of
temperatures
sufficient
in
duration
and
magnitude
to
cause
mortality
is
feasible,
within
parts
of
the
Pacific
Northwest
region,
and
therefore
streams
in
other
geographic
areas
or
streams
with
known
temperature
extremes
should
be
individually
evaluated
with
the
method.
Chronic
exposure
to
temperature
was
based
on
the
growth
potential
of
fish
as
assessed
using
a
simplified
bioenergetics
approach
developed
in
the
report.
This
analysis
found
that
growth
predicted
from
ambient
temperatures
is
somewhat
less
than
the
maximum
potential
growth
in
all
streams
regardless
of
temperature
regime,
because
no
stream
experienced
temperatures
that
fully
optimized
growth
all
of
the
time
during
the
summer
rearing
period.
Generally
the
effect
of
temperature
regime
on
growth
was
small
in
the
range
of
streams
studied,
but
growth
effects
were
evident
at
higher
temperatures.
The
results
suggest
that
quantitative
analysis
of
growth
effects
can
be
determined
with
reasonably
simple
methods
that
can
be
applied
at
specific
sites
or
at
a
region
scale
to
identify
appropriate
temperature
thresholds.
Assuming
a
10%
growth
loss
represents
an
appropriate
risk
level,
an
upper
threshold
for
the
7­
day
maximum
temperature
of
16.5oC
is
appropriate
for
coho
and
20.5oC
is
appropriate
for
steelhead.
Criteria
derived
in
this
manner
are
somewhat
lower
than
those
developed
in
a
U.
S.
E.
P.
A.
paper
in
1977
and
close
to,
but
not
identical,
to
those
currently
specified
in
Washington
and
Oregon
criteria.
vi
ACKNOWLEDGMENTS
The
Oregon
Forest
Industries
Council,
Washington
Forest
Protection
Association
and
Weyerhaeuser
Company
provided
funding
for
this
project.
We
thank
John
Heffner,
Brian
Fransen,
and
Jason
Walters
who
collected
the
temperature
and
fish
data
that
made
the
risk
assessment
analysis
presented
in
this
report
possible.
We
acknowledge
the
significant
contributions
of
our
many
colleagues
cited
in
this
report
who
conducted
the
laboratory
experiments
over
the
past
four
decades
on
which
the
lethal
and
growth
effects
analyses
are
based.
We
remain
especially
inspired
by
the
lifelong
work
of
J.
R.
Brett
who
provided
so
much
understanding
of
the
underlying
physiology
of
salmon
that
translates
into
their
ecologic
behavior.
We
thank
Margaret
Spence
and
Lucinda
Tear
for
their
assistance
with
the
re­
analysis
of
laboratory
growth/
temperature
data
borrowed
from
our
co­
workers'
previously
published
work.
Peter
Farnum
contributed
significantly
with
his
careful
review
of
our
work.
Finally,
we
would
like
to
acknowledge
the
contributions
of
the
scores
of
volunteers
who
participated
in
the
Timber/
Fish/
Wildlife
statewide
temperature
study
in
1990,
and
the
hundreds
of
volunteers
who
participated
in
the
study
of
stream
temperatures
in
the
Puget
Sound
region
under
the
sponsorship
of
the
University
of
Washington
Center
for
Urban
Studies
and
Derek
Booth.
Their
data
help
to
establish
the
extent
to
which
results
from
the
analyses
presented
in
this
paper,
based
on
a
relatively
few
streams,
may
be
applicable
to
the
rest
of
the
streams
of
Washington.

Preferred
Citation:

Sullivan,
K.,
D.
J.
Martin,
R.
D.
Cardwell,
J.
E.
Toll,
and
S.
Duke.
2000.
An
analysis
of
the
effects
of
temperature
on
salmonids
of
the
Pacific
Northwest
with
implications
for
selecting
temperature
criteria.
Sustainable
Ecosystems
Institute,
Portland
Oregon.

Copies
of
this
report
and
Appendix
C
(
under
separate
cover)
may
be
obtained
from:

Librarian,
Sustainable
Ecosystems
Institute,
0605
SW
Taylors
Ferry
Rd,
Portland,
Oregon
97219
(
503.246.5008).
The
report
can
also
be
downloaded
in
PDF
format
from
SEI's
website:
www.
sei.
org.
1­
1
SECTION
1
INTRODUCTION
AND
OBJECTIVES
Maintaining
the
quality
of
aquatic
environments
that
allow
fish
and
other
organisms
to
grow
and
prosper
is
a
primary
objective
of
the
Clean
Water
Act
adopted
by
Congress
in
1972.
To
administer
the
Act,
the
U.
S.
Environmental
Protection
Agency
(
EPA)
and
state
water
quality
authorities
throughout
the
nation
have
adopted
numeric
and
qualitative
criteria
that
establish
environmental
conditions
known
to
protect
aquatic
life
from
adverse
effects.
Historically,
physical
environmental
characteristics
have
been
used
to
indicate
the
minimum
requirements
for
biological
health.
The
criteria
address
naturally
occurring
conditions
that
may
be
affected
by
human
activities
(
e.
g.,
temperature,
sediment,
dissolved
oxygen,
pH,
and
nutrients)
and
numerous
exogenous
pollutants
introduced
by
manufacturing
and
agricultural
activities.
Water
quality
criteria
are
often
a
single
value
defining
thresholds
of
favorable
or
adverse
conditions
(
Suter
et
al.
1993).
The
public
and
the
regulatory
system
have
accepted
simple
physical
criteria
as
indicators
of
biological
health,
although
natural
systems
are
dynamic
and
often
exhibit
a
range
of
water
quality
conditions
over
time
in
response
to
many
non­
anthropogenic
factors.
Thus,
even
though
criteria
are
often
an
over­
simplification
of
real
biological
response,
they
have
generally
been
accepted
as
necessary
to
effectively
administer
the
regulatory
system.

The
Environmental
Protection
Agency
and
other
agencies
have
conducted
water
quality
research
over
the
years
to
accomplish
two
major
objectives:
1)
develop
sound
cause­
andeffect
relationships
between
water
quality
conditions
and
biological
response,
and
2)
develop
repeatable
methodologies
that
use
research
findings
to
craft
regulatory
water
quality
criteria
grounded
in
sound
science.
A
primary
technique
used
by
researchers
is
to
subject
fish
and
other
aquatic
organisms
to
pollutants
in
a
controlled
laboratory
setting
to
determine
the
relationships
between
dosage,
length
of
exposure
and
biological
responses
such
as
growth
loss,
stress,
altered
behavior,
disease,
or
death.
Such
laboratory­
based
research
has
been
a
cornerstone
of
fisheries
science
during
this
century
and
its
validity
has
been
confirmed
in
field­
based
studies
(
Brett
1971,
Shuter
et
al.
1980,
Baker
et.
al.
1995,
Filbert
and
Hawkins
1995).
Conversely,
field
observations
alone
are
often
not
reliable
for
deriving
water
quality
criteria
because
of
variability
in
the
natural
environment
and
the
complexity
of
factors
controlling
natural
systems
and
habitat
response.
Brett
(
1971)
observed
that
"
it
is
inherently
difficult
to
examine
existing
conditions
and
deduce
the
important
biological
factors
which
have
occurred
in
the
past
to
explain
the
present."
Laboratory
studies
were
the
basis
for
EPA
recommended
temperature
criteria
(
U.
S.
EPA
1977),
and
field
studies
have
been
used
mainly
for
validating
the
appropriateness
of
water
quality
criteria
(
Hansen
1989,
Mount
et
al.
1984).

Most
water
quality
criteria
were
originally
adopted
in
the
1970s
(
e.
g.,
U.
S.
EPA
1980)
with
relatively
little
revision
since
implementation
(
Hansen
1989).
In
recent
years,
water
quality
agencies
in
the
Pacific
Northwest
have
conducted
scheduled
reviews
of
criteria
to
reassure
their
effectiveness
or
change
them
if
necessary
(
ODEQ
1995,
WDOE
1999).
Interest
in
the
validity
of
temperature
criteria
has
been
particularly
keen
because
of
concern
that
temperature
is
one
of
the
habitat
elements
that
has
contributed
to
the
decline
in
certain
runs
of
salmon
and
trout
in
the
region
(
NAS
Committee
on
Protection
and
Management
of
Pacific
Northwest
Anadromous
Salmonids
1996).
Within
the
home
range
of
salmon
in
Washington,
Oregon,
and
Idaho,
over
2500
streams
are
currently
listed
on
Clean
Water
Act
section
303(
d)
lists,
many
for
exceeding
summer
temperature
criteria.
High
summertime
temperatures
in
these
streams
are
due
in
part
to
a
variety
of
land
use
1­
2
and
manufacturing
activities
that
have
historically
impacted
temperature
regimes
(
Sullivan
et
al.
1990),
as
well
as
natural
phenomenon
that
affect
stream
temperature.

The
risk
to
salmon
and
trout
populations
associated
with
temperature
is
perceived
to
be
high
because:
1)
the
potential
for
biological
effects
exists
according
to
laboratory­
derived
results;
and,
2)
many
populations
are
already
exposed
to
temperatures
exceeding
those
believed
to
induce
negative
biological
consequences.
Water
temperature
plays
a
role
in
virtually
every
aspect
of
salmon
life
(
Brett
1995;
Weatherly
and
Gill
1995),
and
adverse
levels
of
temperature
can
affect
behavior
(
e.
g.
migration
delays
and
timing),
disease
resistance,
growth,
and
mortality
(
Brett
1956).
Such
concerns
have
led
agencies
to
reconsider
temperature
requirements
and
tolerances
of
these
species,
with
emphasis
on
those
listed
as
threatened
or
endangered:
chinook
and
coho
salmon,
steelhead
trout,
and
bull
trout.
Recent
reviews
have
called
for
lowering
of
temperature
criteria
to
levels
thought
to
be
more
desirable
(
less
stressful)
for
these
species
(
e.
g.
ODEQ
1995;
WDOE
1999,
U.
S.
EPA
in
preparation).

The
scientific
justification
for
these
recommendations
relies
largely
on
review
of
the
scientific
literature
and
application
of
a
number
of
implicit
assumptions
concerning
the
temperatures
that
occur
and
those
that
cause
adverse
effects.
They
also
include
safety
factors
to
ensure
that
adverse
effects
and
exposures
are
not
underestimated.
These
assumptions
and
safety
factors
are
usually
developed
using
best
professional
judgment.
A
more
objective
risk
assessment
technique,
where
adverse
effects
are
placed
in
an
exposure
context
to
identify
population
risk,
has
not
yet
been
applied
in
these
temperature
criteria
reviews,
despite
its
accepted
value
for
establishing
criteria
(
Suter
and
Mabrey
1994)
and
risk
for
other
pollutants
(
U.
S.
EPA
1995).

Risk
assessment
involves
comparing
effects
and
exposure
periods
to
achieve
probability
of
adverse
effects
for
the
defined
exposure.
This
process
includes:
1)
biological
effects
characterization,
2)
exposure
characterization,
and
3)
a
risk
characterization
that
combines
the
two.
The
effects
characterization
requires
a
quantitative
measure
of
the
biological
effects
of
temperature,
and
the
exposure
characterization
requires
a
quantitative
measurement
of
the
temperatures
occurring
in
the
fish's
environment.
These
quantitative
measures
are
expressed
as
probabilities
for
the
risk
characterization,
where
the
in
situ
temperature
regime
is
related
to
the
temperature
biological
effects
relationships
to
estimate
the
likelihood
of
adverse
biological
impacts.

A
risk­
based
approach
seems
ideally
suited
to
developing
criteria
for
and
assessing
temperature
risk
to
aquatic
life.
Fish
are
constantly
exposed
to
temperatures
that
vary
by
minutes,
hours,
days,
weeks,
and
months
depending
on
celestial
forces
that
guide
the
earth
around
the
sun.
Fish
are
thermoconformers;
that
is,
they
cannot
maintain
body
temperatures
much
different
from
the
water
in
which
they
occur.
Thus
their
exposure
is
variable,
ranging
over
the
full
array
of
optimal
and
suboptimal
temperatures.
Considerable
laboratory
study
has
been
conducted
on
a
variety
of
salmon
and
trout
species
to
characterize
their
responses
to
temperature;
these
data
may
be
sufficient
to
characterize
the
responses
of
some
species
and
life
stages.
Finally,
temperature
is
easy
to
measure
and
there
is
an
abundance
of
data
available
in
Washington
and
elsewhere
to
characterize
temperature
regimes
and
to
evaluate
exposure
with
considerable
precision.

The
objectives
of
this
paper
are
(
1)
to
review
relevant
temperature
research
and
(
2)
to
evaluate
the
biological
risks
associated
with
ambient
temperature
regimens
on
populations
of
two
species
of
juvenile
salmonids
using
a
probabilistic
risk
assessment.
This
1­
3
assessment
is
based
on
laboratory
data
concerning
the
effects
of
temperature
on
growth
and
mortality.
The
analyses
concentrate
on
the
summer
rearing
life
history
phase
of
species
within
the
Salmonidae
family
that
dwell
in
stream
environments,
namely
juvenile
coho
salmon
and
steelhead
trout.
There
has
also
been
considerable
study
of
the
thermal
requirements
of
chinook,
sockeye,
pink,
and
chum
salmon,
but
since
these
species
are
not
typically
found
in
western
Washington
streams
during
the
summer
months,
they
will
not
be
directly
considered
here.
This
analysis
illustrates
the
use
of
risk
analysis
for
objectively
deriving
temperature
criteria;
similar
techniques
could
be
applied
to
other
species,
stocks,
and
life
history
phases
(
Hokanson
1977).
We
also
use
results
to
evaluate
the
biological
effects
of
existing
and
proposed
temperature
criteria
in
the
context
of
risk
assessment
techniques.

This
report
contains:

·
 
A
review
of
the
scientific
literature
regarding
the
effects
of
water
temperature
on
direct
acute
mortality
and
growth
of
fish,
with
emphasis
on
salmonids
during
fresh
water
rearing
(
Section
2)

·
 
A
description
of
the
temperature
data
collected
from
a
variety
of
stream
conditions
in
Washington
used
in
the
quantitative
analyses,
with
a
discussion
of
temperature
indices
(
Section
3).

·
 
A
synthesis
of
available
scientific
information
into
a
quantitative,
risk­
based
approach
to
evaluating
biological
response
to
acute
or
lethal
temperatures
in
natural
streams
(
Section
4),

·
 
A
comprehensive
development
of
a
new,
quantitative
approach
to
assessing
fish
growth
in
response
to
long­
term
exposure
to
temperatures
in
natural
streams
(
Section
5),

·
 
A
synthesis
of
the
risk­
based
approaches
to
suggest
temperature
criteria
for
coho
and
steelhead
(
Section
6),

·
 
A
discussion
of
the
use
of
scientific
information,
including
methods
developed
in
this
report,
to
identify
temperature
standards
in
federal
and
state
regulatory
approaches
(
Section
7).

·
 
A
brief
summary
of
key
findings
and
a
synthesis
of
information
for
policy­
makers
and
scientists
(
Section
8).
1­
4
2­
1
SECTION
2
REVIEW
OF
THE
PHYSIOLOGIC
RESPONSE
OF
FISH
TO
ENVIRONMENTAL
TEMPERATURE
ABSTRACT
In
this
section,
the
biological
effects
of
temperature
on
fish
are
briefly
reviewed,
with
emphasis
on
the
fresh
water
rearing
phases
of
salmonids.
Lethal
and
non­
lethal
effects
are
discussed.

Key
findings:

·
 
Many
of
the
lethal
and
non­
lethal
effects
of
temperature
on
salmonids
are
well
understood
and
in
many
cases
have
been
quantitatively
established.

·
 
Both
lethal
and
sub­
lethal
effects
of
temperature
depend
on
its
magnitude
in
relation
to
duration
of
exposure.

·
 
Fish
have
behavioral
and
physiological
mechanisms
to
tolerate
temporary
excursions
into
stressful
temperature
levels.
If
exposure
and
magnitude
limits
are
exceeded,
mortality
can
occur.

·
 
Growth
has
been
widely
used
to
evaluate
the
sub­
lethal
response
of
fish
to
temperature.

INTRODUCTION
Temperature
is
a
dominant
factor
affecting
aquatic
life
within
the
stream
environment
(
Hynes
1970).
Temperature
influences
all
aspects
of
fish
life,
as
well
as
those
of
the
macroinvertebrates
(
Sweeney
and
Vannote
1986)
and
primary
producers
(
algae,
bacteria
etc.)
that
dwell
within
the
stream
and
serve
as
food
for
fish
(
Hynes
1970).
As
summarized
by
Brett
(
1956
pg.
76):

"
Because
of
the
all­
pervading
nature
of
environmental
temperature,
the
fundamental
thermal
requirement
of
fishes
is
an
external
environmental
temperature
most
suitable
to
their
internal
tissues 
Temperature
sets
lethal
limits
to
life;
it
conditions
the
animal
through
acclimation
to
meet
levels
of
temperature
that
would
otherwise
be
intolerable;
it
governs
the
rate
of
development;
it
sets
the
limits
of
metabolic
rate
within
which
the
animal
is
free
to
perform;
and
it
acts
as
a
directive
factor
resulting
in
the
congregation
of
fish
within
given
thermal
ranges,
or
movements
to
new
environmental
conditions."

Quantitatively
defining
the
effects
of
temperature
on
key
biological
functions
is
essential
for
understanding
how
temperature
contributes
to
fish
success
as
well
as
how
it
places
species
at
risk.
Temperature
effects
have
been
extensively
studies
for
all
aspects
of
fish
life.
Although
review
of
all
temperature
effects
are
beyond
the
scope
of
this
report,
we
note
that
there
are
excellent
references
where
temperature
effects
are
discussed
more
fully
(
e.
g.,
Groot
et
al.
1995),
or
where
specific
species
are
reviewed
in
detail
(
e.
g.
McCullough
2­
2
1999
on
chinook
salmon).
We
narrow
our
review
to
aspects
of
temperature
affecting
the
rearing
of
salmonid
species
in
the
fresh
water
environment.

Two
important
elements
of
temperature
affect
the
growth
and
survival
of
fish:
1)
the
relationship
between
temperature,
metabolism,
and
food
conversion
efficiency
over
long
periods,
and
2)
the
thermal
tolerance
of
fish
to
lethal
temperatures
over
relatively
short
periods.
Both
aspects
are
important
because
ambient
stream
temperature
may
vary
from
very
low
levels
in
winter
to
occasionally
high
peaks
in
summer
(
e.
g.,
Beschta
et
al.
1987;
Sullivan
et
al.
1990).

The
thermal
tolerance
to
temperature
has
been
the
focus
of
considerable
laboratory
testing
for
many
fish
species,
including
salmonids,
beginning
early
in
this
century
and
continuing
today
(
see
reviews
by
Fry
1967,
NAS/
NAE
1973,
Coutant
1977).
Much
of
the
available
laboratory
research
on
temperature
tolerances
was
performed
prior
to
1980
and
was
stimulated
principally
by
the
need
to
assess
the
impact
of
heated
effluent
from
power
plants,
dams
and
other
facilities
(
Hokanson
1977).
Since
that
time,
temperature
research
has
focused
on
studying
additional
species
and
refining
the
understanding
of
contributing
factors
such
as
the
effect
of
acclimation
temperatures,
daily
diurnal
temperature
fluctuations,
food
rations,
and
the
interaction
of
temperature
with
other
pollutants
(
e.
g.,
Elliott
1976,
Wurtsbaugh
and
Davis
1977,
Brett
et
al.
1982,
Thomas
et
al
1986,
Coutant
and
Talmage
1977).

These
and
other
studies
show
that
fish
respond
to
temperature
through
physiological
and
behavioral
adjustments
that
depend
on
the
magnitude
and
duration
of
temperature
exposure.
Upper
and
lower
temperature
extremes
that
cause
death
after
exposures
ranging
from
minutes
to
96
hours
are
termed
acute
temperature
effects.
Temperatures
causing
thermal
stress
after
longer
exposures,
ranging
from
weeks
to
months,
are
termed
chronic
temperature
effects.
Endpoints
of
exposure
to
temperature
over
longer
periods
(
chronic
Effects
of
Temperature
on
Salmonids
10
15
20
25
30
35
Minutes
Hours
Days
Weeks
Duration
Temperature
(
oC)
Upper
Critical
Lethal
Limit
Zone
of
Resistance
Mortality
can
occur
in
proportion
to
length
of
exposure
Behavioral
adjustment
(
no
grow
th,
no
mortality)
Zone
of
Tolerance
Zone
of
Preference
Grow
th
response
depends
entirely
on
food
availability
Optimal
grow
th
at
all
but
starvation
ration
Reduced
grow
th
Rapid
death
Figure
2.1
General
biological
effects
of
temperature
on
salmonids
in
relation
to
duration
and
magnitude
of
temperature.
2­
3
effects)
are
sublethal
and
may
include
growth,
competitive
interactions,
change
in
behavior,
or
disease.
Temperature
ranges
defined
by
acute
and
chronic
temperature
effects
are
referred
to
as
the
zones
of
thermal
resistance
and
tolerance
(
Elliott
1981,
Jobling
1981).
The
range
of
physiological
response
relative
to
temperature
is
summarized
in
Figure
2.1.
The
range
of
temperature
over
which
feeding
occurs
without
signs
of
abnormal
behavior
is
referred
to
as
the
optimum
temperature
range
(
Elliott
1981).

ACUTE
TEMPERATURE
EFFECTS
The
acute
effects
of
temperature
are
frequently
expressed
as
effects
on
survival
that
result
from
exposure
to
elevated
temperatures
for
specified
time
periods.
Mortality,
expressed
as
the
median
lethal
time
(
LT50),
and
the
ultimate
upper
incipient
lethal
threshold
(
Brett
1952)
have
been
the
most
common
endpoints
measured.
The
median
lethal
time
is
the
duration
eliciting
50%
mortality
at
a
specific
temperature.
The
ultimate
upper
incipient
lethal
limit
is
the
temperature
at
which
acute
mortality
does
not
increase
with
any
further
increase
in
the
temperature.

Laboratory
studies
repeatedly
show
that
salmon
have
the
ability
to
extend
their
temperature
tolerance
through
acclimation.
Brett
(
1956)
reports
that
the
rate
of
increase
in
ability
to
tolerate
higher
temperatures
among
fish
is
relatively
rapid,
requiring
less
than
24
hours
at
temperatures
above
20oC
(
e.
g.,
Figure
2.2).
Acclimation
to
low
temperatures
(
less
than
5oC)
is
considerably
slower
(
Brett
1956).
Studies
of
the
acute
temperature
effects
on
salmonids
have
yielded
remarkably
consistent
results
between
studies
and
among
salmon
species
(
Brett
1956;
Lee
and
Rinne
1980),
indicating
temperature's
influence
on
fish
with
similar
biochemistry
and
physiology.
The
upper
lethal
limit,
that
is
the
temperature
at
which
death
occurs
within
minutes,
ranges
from
27o
to
30oC
for
salmonids
(
Jobling
1981).
Fish
acclimated
at
cold
temperatures
can
have
upper
lethal
limits
3o
to
4oC
lower.
Many
species
of
fish
have
considerably
higher
upper
thermal
levels
than
members
of
the
Salmonidae
family,
which
are
classified
in
cold
water
temperature
guilds
(
Magnuson
et
al.
1979).
At
temperatures
below
the
upper
lethal
limit,
fish
can
tolerate
each
successively
lower
temperature
for
exponentially
increasing
intervals
of
time.

Behavioral
mechanisms
may
allow
fish
in
situ
to
resist
short­
term
extreme
temperature,
and
acclimation
will
promote
resistance
to
high
temperature,
although
there
is
an
upper
Figure
2.2
Example
of
a
relationship
between
the
time
(
min)
for
50%
mortality
of
brown
trout
(
Salmo
trutta,
and
the
lethal
temperature
(
oC)
at
different
acclimation
temperatures.
(
From
Elliott
1981).
2­
4
limit
to
the
temperature
to
which
fish
can
acclimate
(
Jobling
1981).
This
resistance
to
the
lethal
effects
of
thermal
stress
enables
fish
to
make
excursions
for
limited
times
into
temperatures
that
would
eventually
be
lethal
(
Brett
1956;
Elliott
1981).
The
period
of
tolerance
prior
to
death
is
known
as
the
"
resistance
time"
and
the
duration­
temperature
is
termed
the
"
zone
of
resistance"
(
Figure
2.1)
(
Hokanson
1977,
Jobling
1981).
Laboratory
studies
have
repeatedly
found
that
salmon
can
spend
very
lengthy
periods
in
streams
of
24oC
or
less
without
suffering
mortality.
Thus
temperature
as
a
direct
cause
of
death
generally
ceases
at
temperatures
less
than
24
°
C
(
Brett
1956).
Acute
effects
are
not
generally
considered
below
this
level,
because
continuous
long
duration
exposure
to
temperature
of
this
magnitude
is
not
likely
to
occur
in
natural
environments
within
the
species'
normal
geographical
range.
Laboratory
studies
testing
daily
fluctuations
in
temperature
as
large
as
13.5
°
C
did
not
shown
effects
on
growth
or
mortality
of
salmonids,
although
lethal
levels
were
never
exceeded
(
Thomas
et.
al.
1986).

SUBLETHAL
TEMPERATURE
EFFECTS
Chronic
exposure
to
sublethal
temperatures
can
have
a
broad
range
of
effects
on
the
various
functions
of
fish.
Brett
(
1971)
described
25
physiological
responses
for
sockeye
salmon
and,
similarly,
Elliott
(
1981)
identified
19
similar
characteristics
for
brown
trout.
The
relationship
between
these
responses
and
temperature
follow
two
general
patterns:
either
the
response
(
e.
g.,
standard
metabolic
rate,
active
heart
rate,
gastric
evacuation)
increases
continuously
with
rise
in
temperature,
or
the
response
(
e.
g.,
growth
rate,
swimming
speed,
feeding
rate)
increases
with
temperature
to
maximum
values
at
optimum
temperatures
and
then
decreases
as
temperature
rises
(
Brett
1971,
Elliott
1981).
In
the
latter
case,
the
form
of
the
responses
to
temperature
and
the
optimum
temperatures
are
not
always
the
same
for
different
functions,
and
the
optimum
temperature
for
a
response
may
change
if
there
is
an
alteration
in
another
factor
such
as
energy
intake
(
Elliott
1981).

Based
on
this
theory,
fish
rarely
occur
within
a
temperature
regime
that
is
optimal
for
all
functions
given
the
natural
diel
and
seasonal
variability
in
water
temperature.
Consequently,
fish
have
developed
mechanisms
to
survive
various
levels
of
thermal
stress
both
above
and
below
optimal
ranges
to
maintain
the
health
and
survival
of
a
population.
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
0
5
10
15
20
25
30
35
40
Temperature
(
oC)
Specific
Growth
Rate
(%
weight/
day)
Yellow
Walleye
Lake
Trout
Sockeye
Salmon
Northern
Pike
Figure
2.3
Example
of
a
multi­
family
specific
growth
rate
curves
(
after
Christie
and
Regier
1988)
2­
5
Several
studies
have
indicated
that
growth
under
fluctuating
temperatures
is
essentially
the
same
as
that
under
constant
temperature
if
the
fluctuating
temperature
is
expressed
as
the
time­
weighted
mean1
(
Thomas
et
al.
1986;
Brett
1971;
Everson
1973;
Iverson
1972;
Wurtsbaugh
and
Davis
1977).

Fish
are
poikilothermic
and
temperature
plays
a
key
role
in
regulating
their
metabolic
functions.
Fish
tolerate
suboptimal
temperatures
by
metabolic
adjustment
and
behavioral
thermoregulation
(
Elliott
1981).
For
example,
as
temperatures
increase
above
the
optimum
for
feeding,
the
feeding
rate
declines
and
is
completely
inhibited
at
temperatures
several
degrees
below
the
incipient
lethal
level
(
e.
g,
at
22oC
for
brown
trout,
Elliott
1981;
and
24oC
for
sockeye
and
chinook
salmon,
Brett
1971).
Similarly,
the
metabolic
rate
of
and
scope
for
activity
declines,
reducing
the
overall
energy
expenditure,
which
helps
to
conserve
energy
and
reduce
thermal
stress.
Behavioral
adjustments,
such
as
movements
to
cooler
refuge
sites,
also
enable
fish
to
avoid
thermal
stress.
Numerous
observers
have
reported
significant
changes
in
salmonid
activity
at
or
near
22
°
C
(
Donaldson
and
Foster
1941;
Griffiths
and
Alderdice
1972;
Wurtsbaugh
and
Davis
1977;
Lee
and
Rinne
1980;
Bisson
et.
al.
1988;
Nielsen
et
al.
1994,
Tang
and
Boisclair
1995;
Linton
et
al.
1997;
Biro
1998).
This
temperature
is
consistent
with
a
sharp
drop
in
food
consumption
and
conversion
efficiency
observed
in
laboratory
studies
(
Brett
et
al.
1982).
At
very
low
temperatures,
salmonids
have
been
observed
to
cease
feeding
and
seek
cover
under
banks
or
within
stream
gravels
(
Everest
and
Chapman
1972).

How
large
fish
grow
is
fundamentally
determined
by
environmental
and
population
factors
that
determine
the
availability
of
food.
Temperature,
however,
regulates
how
much
growth
can
occur
with
the
food
that
is
available.
Growth
is
dependent
on
the
energy
consumed
by
the
fish
balanced
by
its
energy
expenditures
to
meet
basic
demands
such
as
metabolism
and
swimming.
What
is
left
over
can
be
used
to
grow
body
mass
and
reproductive
capability.
The
long­
term
exposure
of
salmonids
to
environmental
temperature
during
their
freshwater
rearing
phase
has
an
important
influence
on
the
size
fish
achieve
and
potentially
the
timing
at
which
they
reach
readiness
for
smolting
(
Weatherly
and
Gill
1995).

The
size
of
salmonids
during
juvenile
and
adult
life
stages
influences
survival
and
reproductive
success.
Although
the
large
majority
of
anadromous
salmonid
growth
occurs
in
the
ocean
environment,
growth
of
juveniles
in
natal
streams
is
especially
important
for
anadromous
salmonids
that
must
reach
minimum
sizes
before
they
can
smolt
(
Weatherly
and
Gill
1995).
Holtby
and
Scrivener
(
1989)
and
Quinn
and
Peterson
(
1996)
demonstrated
that
the
size
achieved
by
juvenile
coho
at
the
end
of
their
first
summer
growing
period
was
a
strong
determinant
of
their
later
success
in
overwintering
and
smolting.
Larger
size
generally
conveys
competitive
advantage
for
feeding
in
the
freshwater
environment
(
Puckett
and
Dill
1985,
Nielsen
1994)
for
both
resident
and
anadromous
species.
Mason
(
1976)
and
Keith
et
al.
(
1998)
found
that
the
smaller
fish
tend
to
be
those
that
are
lost
from
rearing
populations.
Brett
et
al.
(
1971)
described
the
freshwater
rearing
phase
of
juvenile
sockeye
as
one
of
restricted
environmental
conditions
and
generally
retarded
growth.
This
synopsis
is
also
generally
true
for
salmonid
species
that
dwell
in
stream
and
river
environments
for
lengthy
periods
of
time.

To
explore
the
effects
of
prolonged
exposure
to
temperature,
numerous
investigators
have
found
growth
to
be
a
reliable
and
measurable
integrator
of
a
variety
of
physiological
1
The
maximum
temperature
studied
cannot
be
high
enough
to
elicit
mortality.
2­
6
responses
(
Brett
1971,
1995;
Iverson
1972;
Brungs
and
Jones
1977;
Wurtsbaugh
1973).
Growth
rate
is
the
most
frequently
reported
measure
of
fish
health
from
laboratory
studies
and
occasionally
from
field
studies.
Growth
can
be
viewed
as
the
net
effect
of
the
environment
on
the
relation
between
food
consumption,
metabolism,
and
activities
of
an
organism
(
Warren
1971).
Growth
integrates
a
host
of
specific
physiological
responses
to
temperature,
including
metabolic
rate
(
basal
and
active),
feeding
and
digestion,
and
swimming
performance
or
the
ability
to
hold
position
with
the
current
(
Brett
1995;
Weatherly
and
Gill
1995).

Laboratory
studies
demonstrate
that
virtually
all
fish,
including
salmonids,
grow
best
within
a
range
of
temperatures.
Optimal
growth
generally
occurs
at
a
midpoint
of
temperatures
where
the
fish
live,
and
it
declines
in
waters
that
are
warmer
or
cooler.
The
range
of
temperature
at
which
growth
occurs
is
generally
wide,
and
usually
reflects
the
ambient
temperatures
likely
to
be
found
within
the
natural
range
of
the
specie's
habitats
(
Hokanson
1977).
Significant
differences
in
growth
curves
exist
among
fish
families
(
Figure
2.3),
from
Christie
and
Regier
1988),
but
growth
curves
are
often
similar
for
species
within
the
same
genera
and
family.
Because
all
salmonids
have
a
similar
biokinetic
range
of
tolerance,
performance,
and
activity,
they
are
classified
as
temperate
stenotherms
(
Hokanson
1977)
and
are
grouped
in
the
cold
water
guild
(
Magnuson
et
al.
1979).

The
effect
of
temperature
on
growth
varies
significantly
with
the
ration
of
available
food
(
Figure
2.4).
For
example,
in
Figure
2.4,
sockeye
salmon
held
at
optimum
temperature
and
fed
satiation
rations
achieved
600%
more
growth
than
fish
held
at
optimum
temperature
with
starvation
rations.
As
ration
increases
from
maintenance
level
(
no
net
growth)
to
satiation
or
excess
level
(
more
than
is
needed
for
growth,
metabolism,
and
all
physiological
functions),
the
optimum
temperature
for
growth
shifts
progressively
to
higher
temperatures.
This
response
is
consistent
for
all
salmonids
where
laboratory
studies
are
available
(
Brett
1971;
Everson
1973;
Iverson
1972;
Wurtsbaugh
and
Davis
1977).
Figure
2.4
Basic
relationship
between
temperature,
ration
and
growth
of
7­
12
month­
old
sockeye
salmon
(
from
Brett
et.
al.
1969).
2­
7
The
relationship
between
food
and
temperature
must
be
taken
into
account
when
considering
the
productivity
of
fish
populations
(
Filbert
and
Hawkins
1995).
Many
studies
have
observed
an
increase
in
the
growth
and
productivity
of
fish
populations
in
streams
when
temperature
(
and
correspondingly)
food
is
increased.
This
tends
to
occur
even
in
the
cases
where
temperatures
exceed
preferred
and
sometimes
lethal
levels
(
Murphy
et
al.
1981,
Hawkins
et.
al.,
1983,
Martin
1985,
Wilzbach
1985).
This
situation
indicates
that
starved
fish
require
somewhat
lower
temperature,
although
the
low
environmental
temperature
tends
to
create
conditions
of
low
food
supply
(
Weatherly
and
Ormerod
1990).

The
forgoing
discussion
indicates
that
the
optimum
temperature
for
fish
extends
over
a
broad
range
depending
on
the
function
and
the
presence
of
other
interacting
factors.
This
optimum
(
preferred)
range
is
defined
by
Elliott
(
1981)
as
the
range
over
which
feeding
occurs
and
there
are
no
external
signs
of
abnormal
behavior,
i.
e.,
thermal
stress
is
not
obvious.
This
delineates
a
wider
range
than
the
peak
optimal
temperatures
where
growth
is
maximized.

Within
the
optimum
temperature
range,
research
has
identified
a
preferred
temperature
range,
which
is
defined
as
the
temperature
around
which
all
individuals
will
ultimately
congregate
regardless
of
their
prior
temperature
exposure
history
(
Fry
1947).
Some
investigators
specifically
define
the
optimal
temperature
as
the
temperature
at
which
maximum
growth
occurs,
and
refer
to
the
range
of
temperature
where
growth
occurs
as
"
preferred"
temperatures.
Determining
this
range,
however
has
resulted
in
considerable
variability
within
the
same
species
due
to
different
experimental
test
procedures
and
the
multiplicity
of
environmental
factors
that
affect
fish
preference
(
Elliott
1981,
Jobling
1981).
This
uncertainty
has
led
Elliott
(
1981)
to
conclude
that
the
optimum
temperature
range
defined
based
on
physiologic
response
is
a
more
realistic
concept
for
studies
on
thermal
stress
then
definitions
based
on
concepts
of
"
preference".
Different
uses
of
the
terminology
can
create
confusion.

Elliott's
optimum
temperature
definition
fits
well
with
the
behavioral
response
of
fish
to
natural
temperature
regimes.
For
example,
Brett
(
1971)
showed
how
behavioral
thermoregulation
by
juvenile
sockeye
resulted
in
energy
conservation.
Vertical
movements
in
a
thermally
stratified
lake
over
the
course
of
a
day
enabled
the
juveniles
to
maximize
the
efficiency
of
food
conversion
into
growth
by
controlling
energy
intake
and
metabolism
as
temperature
followed
the
solar
cycle.
The
sockeye
salmon
exhibited
varied
behavior
in
selecting
temperatures
that
did
not
solely
reflect
the
preferred
temperature
available
to
them
within
the
lake.
Other
field
studies
have
also
documented
salmonid
utilization
of
temperatures
outside
of
the
preferred
range
when
those
within
or
near
the
preferred
level
were
readily
accessible
(
e.
g.
Matthews
et
al.
1994,
Brett
1971,
Biro
1998).

Metabolic
characteristics
are
not
the
only
response,
but
they
are
the
most
important
and
most
easily
quantifiable.
Less
quantifiable
in
a
dose­
response
context
are
relationships
involving
temperature
and
disease
resistance,
and
temperature
effects
on
sensitivity
to
toxic
chemicals
and
other
stressors.
It
is
well
recognized
that
temperature
can
decrease
disease
resistance
in
the
most
sensitive
individuals
within
each
species'
population
and
influence
their
sensitivity
to
certain
toxic
chemicals
(
e.
g.,
Cairns
et
al.
1978).
However,
the
study
of
Cairns
et
al.
(
1978)
concluded
that
"
temperature­
toxicity
interactions
are
far
more
complex
than
earlier
literature
has
indicated,"
and
increased
temperature
does
not
necessarily
lead
to
increased
sensitivity
to
toxic
chemicals.
For
example,
Linton
et
al.
(
1997)
found
that
sublethal
levels
of
ammonia
enhanced
growth
at
higher
temperatures
and
Dockray
et
al.
(
1996)
found
better
performance
at
high
temperature
when
pH
was
2­
8
low.
For
temperature
to
affect
the
occurrence
of
disease,
disease­
causing
organisms
must
be
present,
and
either
those
organisms
must
be
affected
by
temperature
or
fish
must
be
in
a
weakened
state
due
to
the
effect
of
temperature.
In
addition,
some
diseases
may
be
more
prevalent
at
high
temperature,
others
are
more
prevalent
at
low
temperature,
and
some
are
not
apparently
related
to
temperature.
Therefore,
for
disease
and
pollutants,
the
specific
nature
and
local
presence
of
the
disease­
causing
organism
or
pollutant
influences
its
interaction
with
temperature.

The
response
of
fish
to
temperature
in
natural
streams
is
not
only
based
on
physiological
functions
but
also
on
the
overall
interaction
with
other
ecological
factors
(
e.
g.,
predators,
prey
abundance,
and
competitors).
Differences
among
species
can
confer
competitive
advantages
in
relation
to
environmental
variables
that
are
reflected
by
the
species'
distribution
(
Brett
1971,
Baltz
et.
al.
1982,
Reeves
et
al.
1987,
DeStaso
and
Rahel
1994).
Natural
stream
environments
nearly
universally
have
increasing
temperature
from
headwaters
to
their
mouths
(
Hynes
1970),
largely
reflecting
systematic
changes
in
a
variety
of
critical
environmental
factors
that
control
heat
transfer
processes
(
Sullivan
et
al.
1990).
Systematic
changes
in
the
occurrence
or
dominance
of
species
within
river
systems
in
part
reflects
these
temperature
patterns.

ASSOCIATING
BIOLOGICAL
EFFECTS
AND
STREAM
TEMPERATURE
Identification
of
appropriate
temperature
criteria
to
protect
fish
is
complicated
by
the
highly
variable
nature
of
temperature
at
stream
sites,
coupled
with
the
differing
temperature
requirements
of
fish
species.
Water
temperature
at
individual
sites
varies
0
5
10
15
20
25
00
02
04
06
08
10
12
14
16
18
20
22
Temperature
(
oC)

Deschutes
Hard
Cr.
Incubation
Summer
Rearing
Overwintering,

Smolting
(
1
yr)
Em
ergence
(
fry)
Migration
(
adults
)

Jan
Feb
Mar
Apr
May
June
July
Aug
Sept
Oct
N
ov
D
ec
Figure
2.5
Annual
temperature
regime
of
the
Deschutes
River
(
148
km2)
and
Hard
Creek
(
2.3
km2),
a
headwater
tributary,
near
Vail,
Washington.
Data
are
hourly
measurements.
2­
9
significantly
with
time,
ranging
from
lows
in
winter
to
highs
in
summer,
with
daily
fluctuations
depending
on
stream
and
climatic
characteristics
(
Figure
2.5).
The
life
history
phases
of
salmonids
are
generally
adapted
to
the
prevalent
temperature
as
illustrated
for
two
stream
sites
in
Figure
2.5,
although
species
and
stocks
have
specific
life
history
timing
(
Weatherly
and
Gill
1995).
Growth
of
alevins
within
the
gravel
bed
and
as
fry
and
resident
adults
within
the
stream
is
a
function
of
temperature;
the
timing
of
movement
of
alevins,
fry
and
adults
also
depends,
in
part,
on
temperature.

The
intent
of
temperature
criteria
is
to
index
or
describe
key
characteristics
of
the
temperature
regimen
that
have
important
measurable
impacts
on
individuals
or
populations.
Many
authors
favor
identification
of
criteria
appropriate
for
each
life
history
phase
(
Bell
1973,
Reiser
and
Bjornn
1979,
Armour
1991)
that
reflect
the
temporal
variability
of
temperature
through
the
year.
Hokannson
(
1977)
described
a
quantitative
means
of
establishing
criteria
for
each
life
history
phase
of
percids
(
perch,
walleye)
in
a
procedure
he
termed
the
"
Envelope
Method."
Quantitative
estimates
of
fish
response
to
temperature
for
maturation
cycles,
spawning
times,
migrations,
activity
and
spatial
distribution
are
compared
to
seasonal
changes
in
temperature.
In
Hokannson's
example,
natural
history
observations
supplement
experimental
data
where
available.
Temperature
ranges
that
indicate
optimal,
sub­
optimal,
and
lethal
temperatures
are
plotted
or
tabulated
for
each
life
stage
period
(
e.
g.,
migration,
spawning,
incubation,
rearing)
to
show
the
range
and
temporal
distribution
of
temperature
preference/
tolerance
regimes
(
temperature
envelopes)
during
the
hydrologic
year.

Recent
development
of
temperature
standards
in
Oregon
and
Washington
have
also
endorsed
life
history­
based
criteria
for
salmonid
and
char
species,
and
promoted
reach­
and
watershed­
based
approaches
for
determining
criteria
(
ODEQ
1995;
WDOE
1999).
The
intent
is
to
identify
a
series
of
criteria
that
can
be
applied
to
limit
impact
on
all
species
and
life
stages
that
may
exist
in
a
stream
reach
at
that
time
of
year.
There
has
been
some
interest
in
"
tailoring"
criteria
to
specific
time
of
the
year,
species,
and
even
individual
stream
reaches.
However,
reach­,
species­,
or
temporally­
specific
criteria
can
create
enormous
data
collection
and
management
issues.

Criteria
that
can
be
applied
on
a
regional
basis
require
indices
of
the
key
characteristics
of
temperature
regimes
that
are
biologically
meaningful,
measurable
without
extraordinary
means,
and
sensitive
to
human­
caused
effects.
Factors
to
consider
when
reducing
the
variable
summer
temperature
regime
to
simpler
indices
include:
1)
the
temperature
threshold
that
reflects
biological
effects
(
e.
g.,
usually
a
maximum
but
can
be
a
minimum);
Table
2.1
Characteristics
of
temperature
regimes
relevant
to
temperature
criteria.

Regime
Characteristic
Variables
Temperature
Threshold
‰
Acute
‰
Sub­
lethal
Temperature
Fluctuation
Characteristic
‰
Maximum
‰
Mean
‰
Minimum
‰
Fluctuation
(
maximum­
minimum)

Averaging
Period
‰
Instantaneous
maximum
‰
7­
Day
average
‰
Monthly
or
seasonal
average
2­
10
2)
the
temperature
statistic
within
the
amplitude
of
fluctuation
(
e.
g.,
maximum,
mean
or
minimum);
and
3)
the
averaging
period
that
characterizes
temperature
exposure
(
e.
g.,
hourly,
daily,
or
weekly)
(
Table
2.1).

ESTABLISHING
WATER
QUALITY
CRITERIA
In
recent
years,
the
EPA
and
National
Academies
of
Science
and
Engineering
have
promoted
risk
assessment
techniques
to
develop
water
quality
criteria,
including
formal
protocols
that
have
been
peer­
reviewed
nationally
(
Parkhurst
et
al.
1996,
U.
S.
EPA
1995).
Risk
assessment
is
designed
to
enhance
understanding
of
the
potential
adverse
effects
of
a
pollutant
on
species
by
combining
the
information
from
biological
studies
with
an
analysis
of
each
population's
potential
exposure
to
those
effects.
Risk
occurs
when
the
stress'
magnitude,
frequency,
and
duration
exceed
the
species'
ability
to
deal
with
that
stress.
Risk
has
little
to
do
with
the
organisms'
or
species'
sensitivity
to
a
stress
or
to
the
concentration
or
level
of
environmental
stress;
risk
depends
entirely
on
whether
the
combination
of
exposure
and
sensitivity
exceeds
the
organism's
ability
to
withstand
or
adapt
to
the
stress
(
Suter
et
al.
1993;
U.
S.
EPA
1992).

Recent
risk
assessment
techniques
use
more
available
data
and
disclose
more
uncertainties
than
assessments
based
on
comparing
a
number
denoting
an
effect,
criterion
or
an
exposure
(
Parkhurst
et
al.
1996;
Solomon
et
al.
1996).
Exposures
and
potential
effects
may
be
represented
as
probabilities
of
occurrence.
Uncertainties
about
exposures
and
effects
can
also
be
expressed
(
e.
g.,
as
95%
confidence
limits)
and
used
in
decisionmaking
Risk
assessment
techniques
have
been
used
to
derive
water
quality
criteria
for
aquatic
life
(
U.
S.
EPA
1993),
wildlife,
and
human
health,
and
are
being
considered
as
one
of
the
site­
specific
water
quality
criteria
tools
(
Spehar
and
Adams,
1998).
They
can
be
used
to
evaluate
broad
effects
if
general
patterns
of
exposure
are
known
or
to
develop
stream­
specific
criteria.

To
derive
a
meaningful
biological
measure
for
specific
life
phase
requirements,
careful
consideration
must
be
given
to
both
magnitude
and
duration
of
temperature,
since
these
factors
together
have
great
effect
on
the
risk
that
temperature
poses
to
fish.
In
the
remainder
of
this
report,
we
will
use
a
risk
assessment
approach
to
quantitatively
estimate
acute
and
chronic
effects
of
temperature
on
salmonids.
Risk
assessment
requires
a
quantitative
analysis
of
fish
response
to
temperature,
and
a
quantitative
assessment
of
the
exposure
to
temperature
that
a
fish
may
experience
during
the
period
of
interest.
The
overlap
between
effects
and
exposure
determines
the
risks
associated
with
temperatures
experienced
in
the
aquatic
environment.

CONCLUSIONS
The
implications
of
this
research
to
the
question
of
establishing
temperature
criteria
are:

·
 
The
effects
of
temperature
on
physiologic
functions
during
the
freshwater
phase
of
salmonid
life
history
are
reasonably
well
understood,
and
in
many
cases
have
been
quantitatively
established
in
a
laboratory
setting.

·
 
Salmon
and
trout
have
physiological
and
behavioral
mechanisms
that
resist
death
at
high
and
low
temperatures
unless
extreme
maximums
are
achieved.
2­
11
Establishing
temperature
criteria
for
water
quality
standards
is
benefited
by
consideration
of
both
duration
and
magnitude
of
temperature
within
these
extremes.

·
 
The
effects
of
temperature
on
other
factors,
such
as
resistance
to
disease
or
pollutants,
are
more
variable,
depending
on
site
conditions,
and
are
less
well
characterized.
2­
12
3­
1
SECTION
3
TEMPERATURE
CHARACTERISTICS
OF
STREAMS
USED
IN
ANALYSIS
Abstract
Temperature
data
collected
during
the
summer
months
from
a
number
of
streams
and
rivers
in
Washington
are
used
for
biological
analysis
in
following
sections
of
this
report.
In
this
section,
we
introduce
these
data
and
summarize
the
site
and
temperature
statistics
for
each
of
the
monitoring
stations,
and
compare
them
among
sites
and
with
previous
temperature
studies.
Temperature
data
span
a
range
of
temperatures
from
cold
to
warm,
and
many
stream
sites
exceed
current
temperature
criteria.
Sites
are
shown
to
be
broadly
representative
of
temperatures
observed
in
many
Washington
streams.
Temperature
indices,
including
annual
maximum,
7­
day
maximum,
and
7­
day
mean
temperature
are
closely
related
at
each
site,
and
any
can
be
used
to
index
stream
temperature
measured
over
longer
periods.

KEY
FINDINGS
INCLUDE:

·
 
The
data
used
in
this
report
are
broadly
representative
of
stream
temperatures
in
fishbearing
streams
found
in
forested,
rural,
and
urban
streams
in
Washington.
Temperature
patterns
are
also
probably
representative
of
many
streams
throughout
the
Pacific
Northwest.

·
 
Temperatures
span
a
range
of
temperatures,
from
12o
to
26oC
in
the
annual
maximum
water
temperature.
This
temperature
range
is
within
the
range
that
salmonids
may
experience
growth
and
lethal
effects
from
short
and
long­
term
exposure.

·
 
There
is
year­
to­
year
variation
in
temperatures
at
sites,
which
affects
short­
duration
temperature
indices.

·
 
Various
temperature
indices
such
as
annual
maximum,
7­
day
maximum,
and
7­
day
mean
are
closely
related
to
one
another.

TEMPERATURE
DATA
Analysis
of
the
biological
effects
of
temperature
that
follows
in
Sections
4,
5
and
6
is
based
on
temperature
recorded
at
19
stream
sites
in
the
Chehalis,
Deschutes,
and
Toutle
river
watersheds,
located
in
the
Coast
Range
and
the
west
slopes
of
the
Cascade
Mountains
in
Washington.
Temperature
has
been
monitored
over
the
years
for
various
monitoring
and
research
projects.
All
sites
are
located
on
portions
of
the
river
systems
where
forestry
is
the
dominant
land
use.
Sites
with
hourly
temperature
records
varying
from
very
cool
headwater
streams
that
support
cutthroat
steelhead,
and
coho
populations
to
warm
river
mainstems
with
more
diverse
fish
communities
were
selected.

Three
sites
are
located
in
the
mainstem
of
the
headwaters
of
the
Chehalis
River
near
the
town
of
PeEll
and
represent
the
largest
river
in
our
analysis.
Bankfull
stream
widths
3­
2
average
30
to
60
meters.
The
amount
of
shade
varies
with
stream
width,
ranging
from
low
to
high.
The
river
and
its
tributaries
flow
through
second
growth
forests,
and
riparian
areas
are
in
various
stages
of
regrowth
following
past
logging­
related
disturbance.
Eight
tributaries
to
the
headwaters
of
the
Chehalis
River
were
monitored.
Each
is
approximately
5­
10
meters
in
width
at
their
confluence
with
the
mainstem.
Tributary
streams
were
logged
to
the
stream
banks,
and
in
some
cases
cleaned
of
woody
debris,
in
the
1970'
s.
Most
are
now
well
shaded
with
second­
growth
alder
and
Douglas­
fir
plantations.
The
mainsteam
of
the
Chehalis
River
supports
fall
chinook
nearly
as
far
upstream
as
site
3.
The
lower
portions
of
the
tributaries
support
steelhead
and
coho
spawning,
incubation
and
rearing.
Several
of
these
streams
are
the
location
of
marine
nutrient
and
fish
carcass
supplementation
research
previously
reported
in
the
literature
(
Bilby
et
al.,
1996,
1998).
Porter
Creek
is
a
tributary
to
the
lower
Chehalis
River
flowing
from
the
Capitol
Forest
near
Olympia.
It
is
well
shaded
with
a
predominantly
alder
overstory.
This
stream
was
the
site
of
a
woody
debris
addition
study
(
Cederholm
et
al.
1997).

Four
sites
in
the
Deschutes
River
basin
were
monitored,
including
the
mainstem,
near
the
town
of
Vail
and
at
the
downstream
end
of
the
forest
land
use
zone.
A
2000­
ha
tributary,
Thurston
Creek,
and
two
small
streams
(<
300
ha)
in
the
headwaters
(
Hard
and
Ware
Creeks)
have
been
monitored
since
1974.
Previous
monitoring
information
is
available
in
Sullivan
et
al.
(
1987).
The
smallest
tributaries
support
cutthroat
trout
populations,
while
coho
use
the
lower
tributaries
and
mainstem.
Anadromous
fish
are
excluded
from
the
upper
tributaries
by
a
barrier
falls.

Two
sites
are
located
in
the
Mt.
St.
Helens
blast
zone.
These
streams
have
experienced
vegetative
recovery
since
the
eruption
in
1980,
and
currently
support
populations
of
steelhead
and
coho,
that
at
times
are
supplemented
by
hatchery
fish.
Previous
research
on
the
interaction
of
temperature
and
fish
production
has
been
reported
by
Bisson
et
al.
(
1988).

Sites
represent
a
range
of
small
to
large
streams
with
shade
varying
from
0
to
100%.
Maximum
potential
shade
naturally
varies
among
the
sites
with
stream
width.
However,
current
shade
is
lower
than
potential
at
many
sites
due
to
past
forest
practices
or
natural
disturbance.

Water
temperature
was
sampled
to
the
nearest
0
º
C
each
hour
by
an
electronic
temperature
recording
device
(
HoboTemp
®
or
Omnidata
®
)
calibrated
at
the
time
of
deployment
and
field­
checked
at
least
once
each
month.
The
temperature
recorded
by
the
instrument
was
the
average
temperature
for
the
hour.
Water
temperature
probes
were
placed
in
the
stream
near
the
bank
and
out
of
direct
exposure
to
sunlight.

Temperature
Characteristics
Temperatures
span
a
range
from
predominantly
cold
to
predominantly
warm
as
indexed
by
the
annual
maximum
temperature
(
the
single
highest
hourly
temperature
during
the
year)
(
Table
3.1,
Figure
3.1).
Multiple
years
of
data
were
available
at
some
sites.
The
selection
of
years
to
include
in
the
analysis
was
arbitrary,
largely
reflecting
the
ready
access
to
data
in
the
archives.
Although
additional
years
or
sites
could
have
been
included,
data
would
fall
entirely
within
the
range
of
data
observed
at
the
example
sites.
We
did
not
feel
that
is
was
as
important
to
include
a
large
number
of
sites
in
this
temperature
analysis,
as
it
was
to
select
sites
that
span
the
range
of
temperatures
likely
to
occur
within
Washington
to
the
extent
possible
with
the
data
available
to
us.
3­
3
Table
3.1
Basin
and
temperature
characteristics
of
18
stream
sites
used
in
acute
(
Section
4)
and
growth
risk
analysis
(
Section
5).
These
sites
are
referenced
as
temperature
study
sites
in
the
text.

Site
Watershed
Basin
Area
(
km2)
7­
Day
Maximuma
oC
7­
Day
Meanb
oC
Annual
Maximumc
oC
Season
Mediand
oC
Year
Measured
Deschutes
River
mainstem
Deschutes
145.0
21.0
18.4
22.5
15.0
1994
Thurston
Creek
Deschutes
9.1
14.9
14.1
15.5
12
1994
Hard
Creek
Deschutes
3.0
14.0
13.0
14.0
11.0
1994
Ware
Creek
Deschutes
2.8
17.5
16.1
18.3
12.9
1994
Huckleberry
Creek
Deschutes
5.3
18.4
17.6
18.5
15.5
1991
Chehalis
River
mainstem
(
Site
1)
Chehalis
181.8
21.1
18.9
22.1
15.6
1997
Chehalis
River
mainstem
(
Site
2)
Chehalis
57.5
22.1
18.2
23.2
14.5
1997
Chehalis
River
mainstem
(
Site
3)
Chehalis
29.5
20.6
18.6
21.4
14.3
1997
Crim
Creek
Chehalis
22.0
18.8
16.9
19.4
14.3
1997
Lester
Creek
Chehalis
10.4
18.4
16.3
19.0
14.2
1997
Thrash
Creek
Chehalis
16.7
15.3
14.3
15.8
12.3
1997
Rogers
Creek
Chehalis
13.1
15.7
14.1
16.1
12.6
1997
Big
Creek
Chehalis
9.0
16.5
14.6
16.9
12.5
1997
Sage
Creek
Chehalis
5.3
16.5
14.6
16.9
12.5
1997
Salmon
Creek
Chehalis
8.9
15.8
14.2
16.2
12.3
1997
Mack
Creek
Chehalis
2.8
12.9
12.5
13.1
11.7
1997
Porter
Creek
Chehalis
25
17.5
16.3
18.6
14.4
1990
Hoffstadt
Creek
Toutle
25.6
24.5
18.4
26.0
14.0
1988
Harrington
Creek
Toutle
8
19.1
16.7
20.5
13.3
1988
a
maximum
value
of
the
7­
day
moving
average
of
the
daily
maximum
temperature
b
maximum
value
of
the
7­
day
moving
average
of
the
daily
mean
temperature
c
instantaneous
maximum
d
median
of
daily
mean
temperature
from
June
1
to
September
15
The
coolest
measured
stream
was
Mack
Creek.
The
temperature
never
exceeded
13oC
at
any
time
during
the
summer
(
Figure
3.1).
The
warmest
temperatures
recorded
in
the
temperature
study
occurred
in
Hoffstadt
Creek
located
within
the
Mt.
St.
Helens
blast
zone,
although
this
stream
continues
to
cool
with
vegetation
regrowth
since
previous
studies
(
Bisson
et
al.
1988).
The
mainstem
of
the
Chehalis
River
(
sites
1,
2
and
3)
experienced
the
longest
duration
of
high
temperature
at
or
above
20oC.
The
Chehalis
River
is
among
the
warmest
rivers
in
Washington
and
well
exceeds
existing
state
temperature
standards.
The
contrast
in
seasonal
temperature
regime
between
a
consistently
warm
and
a
consistently
cool
site
within
the
same
time
period
and
watershed
is
shown
in
Figure
3.2.
Except
for
Hoffstadt
Creek,
the
temperatures
of
the
other
sixteen
sites
fell
somewhere
between
these
two.
All
streams
that
exceed
16oC
annual
maximum
temperature
exceed
current
Washington
water
quality
temperature
standards.

The
minimum
temperatures
observed
during
the
period
between
June
1
and
September
15
were
between
7
and
9
oC
in
all
streams
(
Figure
3.3).
This
temperature
is
close
to
groundwater
temperature
and
was
typically
experienced
early
in
June.
The
maximum
temperature
observed
reflects
site
characteristics
such
as
openness
to
the
sky,
stream
depth,
and
the
extent
of
groundwater
inflow
(
Sullivan
et
al.
1990).
Despite
large
differences
in
the
annual
maximum
temperatures
among
sites
(
Figure
3.1),
most
streams
also
spent
a
considerable
amount
of
time
at
the
same
temperatures,
most
notably
in
the
range
between
12
and
17
oC.
This
temperature
is
coincident
with
the
optimal
temperature
range
of
many
salmonids.
3­
4
6
8
10
12
14
16
18
20
22
24
1­
Jun­
97
1­
Jul­
97
31­
Jul­
97
30­
Aug­
97
29­
Sep­
9
Chehal
is
R.
(
2)
Mack
Cr
.

Figure
3.2
Daily
maximum
temperature
at
one
of
the
warmest
sites
(
Chehalis
River
site
2),
and
one
of
the
coolest
sites
(
Mack
Creek).
Temperature
Study
Sites
10
12
14
16
18
20
22
24
26
28
M
ack
C
r.
H
ard
C
r
T
hursto
n
C
r
T
hrash
C
r.
R
ogers
C
r.
Salm
o
n
C
r.
B
ig
C
r.
Sage
C
r.
Ware
C
r
H
uc
k
lebe
rry
C
r.
P
o
rter
C
r.
Le
s
te
r
C
r.
C
rim
C
r.
H
arringto
n
C
r
C
hehalis
R
.
(
3)
D
eschutes
R
.
C
hehalis
R
.
(
1)
C
hehalis
R
.
(
2)
H
o
f
f
s
tadt
C
r
Annual
Maximum
Temperature
(
oC)

Figure
3.1
Annual
warmest
temperature
at
the
19
temperature
study
sites
included
in
the
temperature
risk
assessment.
3­
5
Comparison
of
Study
Sites
to
Other
Temperature
Studies
in
Washington,

Oregon
and
Idaho
Stream
temperatures
at
the
nineteen
sites
are
representative
of
most
other
streams
and
rivers
in
Washington,
and
probably
elsewhere
in
the
Pacific
Northwest
as
well.
As
an
indication
of
how
well
the
example
data
span
the
range
of
temperatures,
we
compare
the
relative
occurrence
of
annual
maximum
temperature
at
sites
included
in
other
large­
scale
temperature
studies
that
have
been
conducted
in
Washington.
The
19
risk
analysis
sites
represent
temperature
patterns
in
the
same
range
as
the
89
streams
included
in
the
1990
Timber/
Fish/
Wildlife
statewide
temperature
study;
the
earlier
statewide,
multi­
agency
study
represented
a
broad
range
of
stream
conditions
found
primarily
in
forests
located
throughout
Washington
(
Sullivan
et
al.
1990).
Using
the
annual
maximum
temperature
to
index
long­
term
temperature,
the
distribution
of
temperature
at
sites
were
similar
between
the
two
studies,
although
this
analysis
includes
a
few
more
warm
sites
and
a
few
less
cold
sites
(
Figure
3.4).
Temperatures
from
seven
sites
included
in
both
studies
tended
to
be
warmer
than
expected
with
mature
riparian
vegetation,
due
in
part
to
past
land
use
or
recent
natural
disturbance.
The
1990
study
contained
a
number
of
undisturbed
sites.
Hourly
Temperature
Distribution
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
0.45
5
10
15
20
25
Temperature
(
oC)
Proportion
of
Hours
Deschutes
R.

Hard
Creek
Thurston
Cr
Ware
Cr
Chehalis
Site
1
Chehalis
Site
2
Chehalis
Site
3
Crim
Cr
Lester
Cr
Mack
Cr
Rogers
Cr
Sage
Cr
Thrash
Cr
Salmon
Cr
Big
Cr
Hoffstadt
Cr
Huckleberry
Cr
Harrington
Cr
Porter
Cr
Figure
3.3
Proportion
of
2567
hourly
temperature
observations
by
temperature
at
19
temperature
study
sites
measured
between
June
1
and
September
15.
3­
6
Figure
3.4
also
includes
the
cumulative
distribution
of
maximum
temperatures
measured
in
a
synoptic
study
of
570
urban/
rural
streams
throughout
the
Puget
Sound
region
coordinated
by
the
University
of
Washington
Center
for
Urban
Studies.
Temperature
data
were
collected
by
volunteers
within
a
2­
hour
interval
in
the
afternoon
of
August
19,
1998.
Data
from
this
study
do
not
necessarily
represent
the
hourly
maximum
temperature
since
temperatures
did
not
necessarily
coincide
with
the
hottest
hour
of
the
year.
However,
the
data
are
representative
of
the
long­
term
average
daily
maximum
for
August
(
Derek
Booth,
pers.
com.)
and
are
likely
to
be
within
a
few
degrees
of
the
annual
maximum
(
Sullivan
et
al.,
1990).
Streams
tended
to
be
small
tributary
streams.
Stream
temperatures
at
these
sites
tended
to
be
slightly
cooler
than
risk
analysis
sites,
and
no
streams
were
as
warm
as
those
included
in
this
analysis.

In
further
consideration
of
how
the
risk
sites
represent
the
range
of
high
temperatures
observed
throughout
the
Pacific
Northwest
region,
we
examined
published
U.
S.
Geological
Survey
temperature
records
from
Washington,
Oregon
and
Idaho.
We
selected
the
year
1978­
79
for
several
reasons:
there
were
more
sites
recording
temperature
in
the
1970'
s
and
1980'
s
than
are
operative
today,
and
this
was
a
period
of
rather
high
temperature
throughout
the
region
due
to
the
1977­
78
drought.
The
annual
maximum
temperatures
are
shown
by
state
in
Figure
3.5.
(
Note
differences
in
the
number
of
sites
in
each
state.)
The
U.
S.
G.
S.
sites
are
primarily
on
larger
rivers,
although
some
smaller
streams
are
also
included.
For
example,
the
Columbia
River
mainstem
is
represented
6
times
in
the
Washington
data
and
3
times
in
the
Oregon
data.
Rivers
such
as
the
Columbia,
Skagit,
Yakima,
Snake,
Deschutes
(
OR),
Willamette,
Rogue,
and
Umpqua,
to
name
a
few,
are
included
in
this
data
set.
(
See
Appendix
A
for
a
listing
of
U.
S.
G.
S.
sites.)
None
of
the
sites
in
the
Timber/
Fish/
Wildlife
study,
University
of
Washington
survey,
or
the
risk
sites
(
Table
3.1)
were
located
within
the
zone
of
influence
of
dams.
Dams
often
cause
local
heating
or
cooling
depending
on
the
release
depth
from
the
upstream
reservoir.
A
few
of
the
U.
S.
Geological
Survey
sites
were
located
below
dams.
These
sites
were
generally
colder
than
expected
given
the
size
of
the
river
at
these
locations.
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
10
15
20
25
30
Annual
MaximumTemperature
(
oC)
Cumulative
Proportion
of
Sites
Risk
Analysis
(
n=
19)

TFW
1990
(
n=
89)

U.
Wash.
Center
Urban
Studies
(
n=
570)

Figure
3.4
Cumulative
distribution
of
the
annual
maximum
temperature
at
temperature
study
sites
included
in
this
study
in
comparison
to
the
89
sites
included
in
the
statewide
study
of
temperature
conducted
by
Sullivan
et
al.
(
1990)
and
the
570
sites
included
in
the
University
of
Washington
Center
for
Urban
Studies
synoptic
survey
of
urban
and
rural
streams
in
the
Puget
Sound
area.
The
1990
TFW
study
included
sites
located
throughout
Washington.
3­
7
Of
the
129
U.
S.
G.
S.
sites,
11
(
8.5%)
had
an
annual
maximum
temperature
greater
than
26oC,
the
highest
temperature
observed
at
the
risk
analysis
sites.
One
of
those
was
in
Washington
(
Yakima
River),
1
was
observed
in
Idaho
(
Snake
River)
and
9
were
observed
in
Oregon.
The
John
Day
River
in
eastern
Oregon
reached
temperatures
as
high
as
31oC.
The
remainder
of
the
Oregon
sites
exceeding
26oC
were
concentrated
in
the
southwestern
corner
of
the
state,
including
the
Applegate,
Siuslaw,
N.
and
S.
Fork
Umpqua,
Calapooia
River,
and
Elk
Creek.

The
four
sources
of
data
cited
in
this
report
provide
perspective
on
the
temperatures
of
small
forested
and
urban
streams
and
moderate
to
large
size
rivers
in
both
the
dry
interior
and
wet
coastal
zones.
The
sites
included
in
the
risk
analysis
are
broadly
representative
of
temperatures
of
moderate
to
small
size
streams
(
all
sites
had
basin
area
less
than
200
km2).
Larger
rivers
tend
to
fall
within
the
temperature
ranges
observed
in
the
smaller
rivers.
However,
it
is
appropriate
to
recognize
that
the
largest
rivers,
and
those
in
some
geographic
areas,
have
different
temperature
regimes
than
most
Pacific
Northwest
streams,
and
if
temperature
is
of
concern,
these
should
be
specifically
evaluated
to
determine
whether
the
duration
of
specific
temperatures
exceeds
adverse
levels.

The
data
from
the
University
of
Washington
website,
the
statewide
Timber/
Fish/
Wildlife
study,
and
the
U.
S.
Geological
Survey
are
presented
merely
to
establish
how
well
the
19
risk
analysis
sites
used
in
this
report
(
Table
3)
represent
the
streams
found
in
a
variety
of
geographic
and
land
use
settings
that
occur
in
the
Pacific
Northwest
region.
Only
the
records
from
the
19
sites
are
used
for
the
analyses
of
acute
and
chronic
temperature
effects
that
follow
in
Sections
4,5,
and
6
of
this
report.
Regional
Temperature,
U.
S.
G.
S.
1979
0
5
10
15
20
25
30
35
1
10
100
1000
10000
100000
1000000
Basin
Area
(
km2)
Annual
Maximum
Temperature
(
oC)
Washington
(
n=
34)
Oregon
(
n=
78)
Idaho
(
n=
17)

Range
of
temperature
at
risk
analysis
sites
Figure
3.5.
Annual
maximum
temperature
for
all
stream
and
river
sites
listed
in
the
U.
S.
Geological
Survey
Water
resources
data
for
Washington,
Oregon,
and
Idaho
for
the
year
1978­
79.
3­
8
Deschutes
River
(
Km
60.2)

15
16
17
18
19
20
21
22
23
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
Year
7­
Day
Maximum
Temperature
(
oC)

Figure
3.6
Annual
maximum
temperature
of
the
Deschutes
River
near
Vail,
Washington
from
1975­
1995.

Deschutes
River,
WA
(
Km
60.2)

0.00
0.05
0.10
0.15
0.20
0.25
5
10
15
20
25
Temperature
(
oC)
Proportion
of
Hours
1988
1989
1990
1991
1992
1993
1994
1995
Figure
3.7
Frequency
distribution
of
hourly
temperature
for
8
years
in
the
Deschutes
River
near
Vail,
WA.
We
had
no
hourly
temperature
records
for
comparable
time
intervals
from
the
coldest
streams
(
e.
g.,
<
12oC).
These
are
most
likely
to
occur
in
well­
shaded,
small
headwaters
streams
(
Black,
2000).
When
well­
shaded,
these
streams
tend
to
hover
near
groundwater
temperature
(
typically
6o­
10oC
during
the
summer,
depending
on
geographic
location)
with
little,
if
any,
daily
fluctuation.
Such
patterns
were
evident
in
Norwegian
Creek
(
western
Washington)
and
Cee
Cee
Ah
Creek
(
eastern
Washington)
as
examples
taken
from
the
TFW
statewide
study
in
1990
(
Sullivan
et
al.
1990).
To
represent
this
type
of
stream,
we
assigned
constant
temperatures
through
the
summer.
All
temperature
indices,
including
daily
mean
and
daily
maximum
were
set
to
10oC
for
"
Ten
Site"
and
8oC
for
"
Eight
Site".
3­
9
Temporal
Variation
at
a
Site
Temperature
regimes
at
sites
often
vary
somewhat
from
year
to
year
due
to
climatic
factors.
Annual
variation
is
illustrated
in
the
7­
day
maximum
temperature
at
the
Deschutes
River
mainstem
(
Figure
3.6).
The
long­
term
mean
at
this
site
was
20.3oC
averaged
over
20
years
but
the
upper
temperature
ranged
±
1.25oC
the
average.
The
shape
of
the
frequency
distribution,
available
in
hourly
increments
for
eight
years
since
1988,
was
similar
from
year
to
year,
but
tended
to
shift
up
or
down
the
temperature
scale
(
Figure
3.7).
Therefore,
the
time­
averaged
characteristics
of
temperature
are
likely
to
vary
on
an
annual
basis
at
a
site.
However,
the
relationship
between
the
indexing
characteristics
such
as
maximum
7­
day
temperature
and
the
overall
distribution
of
temperature
during
the
same
summer
period
should
remain
consistent.

Temperature
Indices
Later
in
this
report
(
Sections
6
and
7),
we
will
discuss
the
temperature
indices
that
have
been
used
to
characterize
the
complex
long­
term
temperature
regime
experienced
by
the
biological
community
inhabiting
streams
(
e.
g.,
Figures
2.5
and
3.2)
in
ways
that
are
meaningful
ecologically.
These
indices
or
metrics
are
considered
necessary
as
per
Clean
Water
Act
requirements
to
establish
temperature
criteria
that
are
protective
of
salmonids
or
other
designated
uses.
When
water
quality
criteria
are
exceeded
in
water
bodies,
activities
that
contribute
to
or
cause
the
exceedances
to
those
water
bodies
may
be
restricted.
Many
streams
and
rivers
are
currently
identified
as
exceeding
water
quality
criteria
according
to
the
305b
reports
from
the
states
in
the
Pacific
Northwest
region,
with
a
large
number
of
them
listed
for
temperature
impairment.
Therefore,
the
temperature
criteria
take
on
significant
legal
and
economic
meaning,
and
their
appropriateness
is
of
great
concern
to
the
public,
scientists,
and
regulators.

Temperature
criteria
generally
specify
a
temperature
threshold
calculated
over
an
averaging
period
(
Table
2.1).
For
example,
Washington's
current
criteria,
sometimes
also
referred
to
as
standards,
specifies
the
annual
maximum
temperature,
expressed
as
the
maximum
hourly
temperature
that
occurs
each
year.
Oregon
specifies
the
average
of
the
daily
maximum
temperature
of
the
7
warmest
consecutive
days
(
ODEQ
1995).
The
U.
S.
EPA
(
1977)
recommends
the
average
of
the
daily
mean
temperature
of
the
7
warmest
consecutive
days
(
MWAT).
Each
of
these
measures
for
each
temperature
site
is
listed
in
Table
3.1.

There
is
an
implicit
assumption
with
these
indices
that
they
are
representative
of
temperatures
that
is
biologically
meaningful
in
some
way.
The
relationship
between
short­
term
indices
and
acute
temperatures,
which
are
typically
expressed
for
short
intervals,
may
be
direct
as
discussed
in
Section
4.
However,
the
threshold
values
associated
with
state
water
quality
criteria
appear
to
be
selected
to
prevent
long­
term
chronic
effects
as
discussed
in
Sections
5
and
6.
Therefore,
there
is
an
implicit
assumption
that
short­
term
indices,
based
on
temperature
measured
for
only
a
few
hours
each
year,
represent
the
effects
of
long­
term
exposure.
This
assumption
is
worthy
of
evaluation.
Furthermore,
there
is
no
consensus
on
how
to
report
stream
temperature
with
meaningful
but
simplified
measures:
laboratory
and
field
studies
use
a
wide
variety
of
methods,
and
seemingly,
no
two
are
alike.
Lack
of
standardized
methods
for
reporting
temperature
among
both
the
physical
and
biological
sciences
makes
comparison
among
3­
10
studies
difficult,
and
the
selection
of
temperature
criteria
based
on
field
ecological
and
laboratory
studies
tenuous.

Our
analyses
of
acute
and
chronic
biological
effects
associated
with
natural
stream
temperatures
that
follow
in
Sections
5
and
6
of
this
report
rely
on
hourly
records
summarized
only
to
average
daily
temperature.
These
analyses
are
therefore
not
limited
by
the
lack
of
consensus
on
methodology
to
compress
long­
term
temperature
regimes
into
very
short
duration
indices.
We
will,
however,
use
results
based
on
detailed
temperature
records
in
Section
6
to
evaluate
whether
short­
term
indices
are
reliable
indictors
of
at
least
some
long­
term
biological
responses.

Nevertheless,
although
there
is
some
debate
as
to
whether
short­
term
indices
are
appropriate
to
represent
long­
term
exposure,
it
appears
that
all
of
the
short­
term
indices
are
closely
related
to
one
another
(
Figure
3.8).
This
makes
selection
among
them
a
matter
of
procedural
and
logistical
questions,
rather
than
a
biological
question,
since
all
similarly
index
the
characteristics
of
the
upper
tail
of
the
distribution
of
the
temperatures
sampled.

Perhaps
a
more
important
question
is
how
well
the
short­
term
measures
correlate
with
temperature
characteristics
occurring
over
longer
periods.
The
median
temperature
for
the
period
from
June
1
to
September
15,
a
long­
term
measure,
is
shown
in
relation
to
the
three
short­
duration
indices
in
Figure
3.9.
Although
more
variable,
the
short­
term
indices
are
well
correlated
with
the
season
median,
indicating
that
short­
duration
measures
can
meaningfully
characterize
seasonal
temperature
patterns,
albeit
with
some
loss
of
precision.
Not
surprisingly,
the
7­
day
mean
temperature
(
MWAT)
is
best
correlated
with
the
season
median,
probably
because
each
is
respectively
characterizing
the
central
tendency
of
the
temperature
within
the
daily
and
seasonal
period.

y
=
1.117x
­
1.427
R2
=
0.996
10
12
14
16
18
20
22
24
26
1
0
1
2
1
4
1
6
1
8
2
0
2
2
2
4
2
6
7­
Day
M
aximum
Temperature
(
oC)
Line
o
f
1:
1
Correspond
ence
y
=
1.5524x
­
6.2644
R2
=
0.87
1
0
1
2
1
4
1
6
1
8
2
0
2
2
2
4
2
6
1
0
1
2
1
4
1
6
1
8
2
0
2
2
2
4
2
6
7­
Day
M
ean
Temperature
(
oC
)
Line
of
1:
1
Cor
respondence
y
=
0.6353x
+
4.5983
R2
=
0.89
1
0
1
2
1
4
1
6
1
8
2
0
2
2
1
0
1
2
1
4
1
6
1
8
2
0
2
2
2
4
2
6
7­
Day
M
aximum
T
emperature
(
oC
)
Line
o
f
1
:
1
Correspondence
Figure
3.8
Relationships
between
temperature
indices
including
annual
maximum,
7­
day
mean
(
MWAT),
and
7­
day
maximum.
3­
11
CONCLUSIONS
·
 
The
data
used
in
this
report
are
broadly
representative
of
stream
temperatures
in
fishbearing
streams
found
in
forested,
rural,
and
urban
streams
in
Washington.
Their
temperatures
also
appear
to
be
representative
of
many
streams
throughout
the
Pacific
Northwest,
based
on
comparisons
of
data
from
other
sources
in
the
region.

·
 
Observed
temperatures
at
study
sites
span
a
range
of
temperatures,
from
13o
to
26oC
in
the
annual
maximum
water
temperature.
This
range
encompasses
most
of
the
temperatures
where
salmonids
may
experience
acute
and
chronic
effects
from
short
and
long­
term
exposure.

·
 
Various
temperature
indices
such
as
annual
maximum,
7­
day
maximum,
and
7­
day
mean
that
are
often
used
in
temperature
criteria
are
closely
related
to
one
another
and
can
be
compared
or
used
interchangeably
with
the
appropriate
correlation
relationships.

·
 
Measures
representing
long
duration
exposure,
such
as
the
median
temperature
observed
over
the
summer
period
are
related
to
short­
term
measures.

·
 
There
is
year­
to­
year
variation
in
temperatures
at
stream
sites,
which
is
reflected
in
short­
duration
temperature
indices.
7­
Day
Mean
=
1.409x
­
3.01
R2
=
0.83
7­
Day
Maximum
=
1.80x
­
6.39
R2
=
0.62
Annual
Max
=
1.976x
­
8.11
R2
=
0.59
10
12
14
16
18
20
22
24
26
28
10
12
14
16
18
20
Season
Median
(
oC)
Temperature
(
oC)
7­
Day
Maximum
7­
Day
Mean
Annual
Maximum
Figure
3.9
Relationship
between
season
median
temperature
(
June
1­
Sept
15)
with
short
duration
indices.
3­
12
4­
1
SECTION
4
ASSESSMENT
OF
RISK
OF
SALMON
SPECIES
TO
ACUTE
TEMPERATURE
IN
STREAMS
AND
RIVERS
Abstract
In
this
section,
we
examine
temperature
records
from
streams
and
rivers
in
Washington
spanning
a
range
of
summertime
maximum
temperatures
to
determine
whether
acute
lethal
temperature
conditions
exist,
and
if
they
could
be
associated
with
water
quality
temperature
criteria.
A
relationship
between
temperature
and
duration
of
exposure
sufficient
to
cause
mortality
was
established
based
on
previously
published
research.
Hourly
temperature
records
were
scanned
for
occurrences
of
sufficient
continuous
duration,
defined
for
each
level
of
temperature
to
cause
mortality
within
salmonid
populations.
Although
at
least
one
stream
had
temperature
as
high
as
26oC,
a
temperature
that
can
be
lethal
to
salmonids,
the
length
of
exposure
was
not
sufficient
to
cause
mortality.
We
found
no
occurrence
of
acute
lethal
temperature
conditions
in
any
of
the
stream
sites,
which
are
broadly
representative
of
streams
and
rivers
in
the
Pacific
Northwest.

Key
findings
of
this
chapter:

‰
There
is
sufficient
information
to
quantitatively
define
the
lethal
effects
of
temperature
on
salmonids.

‰
No
occurrences
of
acute
lethal
temperatures
were
observed
at
stream
sites
with
a
wide
range
of
temperatures
including
many
with
annual
maximum
temperatures
that
well
exceed
current
water
quality
standards.

‰
Nevertheless,
lethal
level
temperatures
of
sufficient
duration
to
cause
mortality
have
been
reported
in
the
Pacific
Northwest.
Therefore,
although
not
a
common
occurrence,
attention
should
be
paid
to
local
site
conditions
that
can
lead
to
acute
mortality.

‰
A
temperature
threshold
of
26oC
is
suggested
to
prevent
mortality
of
salmon
and
trout
species
in
natural
rivers
and
streams.
Further
analysis
of
temperature
to
determine
exposure
is
suggested
for
streams
where
annual
maximum
temperature
is
between
24o
and
26oC.

INTRODUCTION
Temperature
duration
and
lethality
relationships
have
been
established
through
laboratory
study
for
most
salmon
species.
Acute
effects
of
temperature
typically
have
been
assessed
as
effects
on
survival
that
result
from
continuous
exposure
to
elevated
temperatures
for
specified
periods
of
time
(
usually
from
1
to
96
hours).
Mortality
has
been
commonly
expressed
as
the
duration
eliciting
mortality
of
some
specified
portion
of
the
population
at
a
specific
temperature
(
Brett
1952).
This
is
a
measure
of
mortality
from
temperatures
occurring
within
the
zone
of
resistance
(
Figure
2.1)
(
Fagerlund
et
al.
1995);
that
is,
where
the
temperature
must
be
experienced
for
some
duration
greater
than
1
hour
before
4­
2
mortality
occurs.
These
temperatures
are
more
likely
to
occur
than
the
ultimate
lethal
thresholds
where
mortality
of
most
or
all
of
the
population
occurs
within
a
very
brief
time.

Risk
analysis
is
performed
by
quantitatively
relating
key
temperature
characteristics
with
specific
measures
of
probable
population
response
to
those
temperatures.
We
examined
the
likelihood
that
exposure
to
temperature
is
of
sufficient
magnitude
and
duration
that
it
causes
direct
mortality
within
the
fish
population
using
conventional
probabilistic
risk
assessment
procedures
(
Parkhurst
et
al.
1996).
Laboratory
mortality
data
available
from
the
literature
were
used
to
develop
temperature
effects
relationships.
Temperature
data
from
streams
monitored
continuously
during
the
summer
months
were
used
to
assess
exposure.
Acute
effects
would
most
likely
be
associated
with
the
occasional
spikes
of
warm
temperature
that
may
induce
mortality.

ACUTE
THERMAL
EFFECTS
CURVES
ASSOCIATED
WITH
50%
MORTALITY
Past
research
has
emphasized
the
exposure
duration
causing
50%
mortality
in
the
population
at
a
given
temperature,
as
the
most
common
lethality
measure.
Data
from
several
sources
were
used
to
generate
curves
showing
the
relationship
between
temperature
and
duration
to
50%
mortality
(
EPA
1977,
Brett
1952,
and
Golden
1978).
Each
curve
estimates
the
length
of
time
that
50%
of
a
population
can
survive
at
some
temperature
above
its
upper
incipient
lethal
temperature.
This
temperature
is
referred
to
as
the
LT50.
At
each
successively
lower
temperature,
the
duration
of
exposure
must
be
longer
to
achieve
the
same
amount
of
mortality
(
Figure
2.2).

EPA
(
1977,
page
11
of
text
and
page
38
of
Appendix
C)
provides
a
regression
equation
relating
exposure
time
(
in
minutes)
to
the
LT50
(
in
°
C):

50
log
10
LT
b
a
t
×
+
=
(
4.1)

where
t
is
the
exposure
time,
and
a
and
b
are
coefficients
of
the
relationships.
Equation
4.1
can
be
arranged
to
b
a
t
LT
/
)
(
log
50
10
-
=
(
4.2)

The
regression
coefficients,
a
and
b,
are
provided
in
EPA
(
1977)
for
many
fish
species,
including
all
those
identified
above,
except
cutthroat
trout
(
pages
55­
58
of
Appendix
B).
From
the
coefficients
provided,
curves
can
be
generated
for
selected
species
of
salmon
and
trout:
pink
salmon,
chum
salmon,
coho
salmon,
sockeye
salmon,
and
chinook
salmon.
The
coefficients
in
EPA
(
1977)
were
gathered
from
many
different
sources,
including
Brett's
1952
paper
summarizing
his
study
of
lethal
temperatures
for
the
five
salmon
species.
It
is
a
necessary
assumption
of
this
analysis
that
the
data
from
these
laboratory
studies
conducted
on
a
small
number
of
fish
and
a
few
stocks
are
representative
of
the
species,
and
that
these
relationships
correctly
characterize
the
mortality/
temperature
relationships.
Golden
(
1978,
Figure
4
on
page
14)
provides
regression
coefficients
for
cutthroat
trout.
Steelhead
LT50
curves
were
generated
using
data
from
Alabaster
and
4­
3
Table
4.1
Regression
coefficients
for
the
relationship
between
duration
and
percent
mortality
of
the
sample
population
for
laboratory
studies
on
salmon
and
trout
species.
Most
of
the
data
are
taken
from
a
summary
U.
S.
E.
P.
A
document
(
1977).
Studies
report
LT50,
unless
otherwise
noted.

Species
Acclimation
Temperature
(
oC)
Source
Age/
Size
a
b
N
R
Notes
Coho
salmon
15
Brett
1952
Juvenile
20.4066
­
0.6858
6
­
0.9681
20
"
"
20.4022
­
0.6713
4
­
0.9985
23
"
"
18.9736
­
0.6013
5
­
0.9956
17
Coutant
1970
Adult
5.9068
­
0.1630
5
­
0.9767
Reported
acclimation
temp.
was
the
Columbia
River
temp
(
at
Priest
Rapids)
during
fall
migration.

Rainbow
trout
15
Alabaster
and
Downing
1966
Juvenile
15.650
­
0.5000
2
­

(
Steelhead)
18
Alabaster
and
Welcomme
1962
"
18.4654
­
0.5801
5
­
0.9787
D.
O.
at
7.4
mg/
l
18
"
"
13.6531
­
0.464
5
­
0.9742
D.
O
at
3.8
mg/
l
20
Alabaster
and
Downing
1966
"
19.6250
­
0.6250
2
­

20
Craigie
1963
Yearling
14.6405
­
0.4470
3
­
0.9787
Raised
in
soft
water,
tested
in
soft
water
20
"
"
15.0392
­
0.4561
3
­
0.9917
Raised
in
soft
water,
tested
in
hard
water
20
"
"
15.1473
­
0.4683
3
­
0.9781
Raised
in
hard
water,
tested
in
soft
water
20
"
"
12.8718
­
0.3837
3
­
0.9841
Raised
in
hard
water,
tested
in
hard
water
Cutthroat
trout
23
Golden
1978
Juvenile
18.092
­
0.56523
?
­
0.996
Hatchery
fish
only
13­
25
(
fluctuating)
"
"
22.543
­
0.71999
?
­
0.999
Hatchery
fish
only
23
"
"
18.3166
­
0.5723
?
­
0.999
Hatchery
and
wild
fish
pooled
13­
25
(
fluctuating)
"
"
18.1515
­
0.5723
?
­
0.992
Hatchery
and
wild
fish
pooled
Chinook
salmon
20
Blahm
and
McConnell
Juv
(
spring
run)
21.3981
­
0.7253
3
­
0.9579
50%
mortality
20
Unpub.
data
"
22.6664
­
0.7797
4
­
0.9747
10%
mortality
20
"
"
20.9294
­
0.7024
3
­
0.9463
90%
mortality
20
"
Juv
(
fall
run)
22.2124
­
0.7526
4
­
0.9738
50%
mortality
20
"
"
21.6756
­
0.7438
4
­
0.9550
10%
mortality
20
"
"
20.5162
­
0.6860
3
­
0.9475
90%
mortality
Sockeye
salmon
15
Brett
1952
Juvenile
15.8799
­
0.5210
7
­
0.9126
20
"
"
19.3821
­
0.6378
5
­
0.9602
23
"
"
20.0020
­
0.6496
4
­
0.9981
20
McConnell
and
Blahn
1970
Juv/
underyearling
16.7328
­
0.5473
6
­
0.9552
20
Unpublished
data
"
17.5227
­
0.5861
6
­
0.9739
10%
mortality
20
"
"
15.7823
­
0.5061
6
­
0.9539
90%
morality
4­
4
Downing
(
1966)
and
Alabaster
and
Welcomme
(
1962),
as
cited
in
the
EPA
document.
Data
were
combined
where
studies
were
reasonably
comparable.
Coefficients
for
each
species
are
provided
in
Table
4.1
for
experiments
that
tested
a
range
of
acclimation
temperatures.
Generally,
the
higher
the
acclimation
temperature,
the
higher
the
LT50
temperature.

The
acute
thermal
effects
curves
were
generated
in
Excel
â
using
regression
coefficients
provided
in
EPA
(
1977)
and
Golden
(
1978)
for
a
range
of
exposure
times.
While
the
equations
provided
in
EPA
(
1977)
were
based
on
exposure
times
measured
in
minutes,
we
converted
them
to
hours
in
order
to
be
consistent
with
temperature
measurements
at
field
sites.
To
generate
the
hourly
curves,
equation
4.2
was
modified
to
b
a
t
LT
/
)
)
60
((
log
50
10
-
×
=
(
4.3)

Although
it
was
assumed
that
the
regression
coefficients
in
Appendix
B
of
EPA
(
1977)
were
correct,
one
appeared
to
be
in
error.
The
value
for
a
was
given
as
16.2444
for
pink
salmon
at
an
acclimation
temperature
of
20
°
C
from
Brett's
study
(
1952).
The
resulting
curve
did
not
match
the
one
presented
in
Figure
5
of
Brett
(
1952).
To
generate
a
curve
more
representative
of
Brett's
(
1952)
figure,
a
value
of
13.2444
was
used
for
a
instead.

A
few
of
the
studies
included
in
EPA
(
1977)
were
excluded
from
the
analysis.
These
were
studies
in
which
the
fish
being
tested
showed
signs
of
gas
bubble
disease
or
other
effects
of
gas
supersaturation.

COMPARISON
OF
LT50
and
LT10
MORTALITY
RELATIONSHIPS
Most
of
the
available
information
on
thermal
effects
is
based
on
50%
survival.
We
felt
that
it
was
appropriate
to
use
a
more
conservative
population
measure
for
risk
assessment.
Therefore,
we
also
expressed
acute
effects
as
the
duration
of
time
needed
to
elicit
10%
mortality
(
LT10)
for
each
temperature
and
species
studied.
LT10
was
selected
because
it
is
the
amount
of
mortality
considered
acceptable
in
the
control
groups
for
acute
toxicity
tests
(
ASTM
1997),
and
90%
(
100%
­
10%)
is
a
recommended
protection
level
for
species
populations
(
SETAC
1994;
Solomon
et
al.
1996).

In
the
EPA
(
1977)
document,
two
unpublished
studies
provided
regression
coefficients
for
both
50%
and
10%
(
LT10)
mortality
curves
at
acclimation
temperatures
of
15oC
or
higher.
McConnell
and
Blahm
(
1970)
calculated
regression
coefficients
for
sockeye
salmon;
and
Blahm
and
McConnell
(
1970)
calculated
regression
coefficients
for
both
spring
and
fall
runs
of
chinook
salmon.
Using
the
regression
coefficients
generated
from
these
studies,
LT50
and
LT10
values
for
sockeye
(
Table
4.2)
and
chinook
(
Table
4.3)
were
calculated
for
a
range
of
time
periods,
along
with
their
ratio.

For
the
range
of
exposure
times,
the
LT10
values
were
98.0
to
99.7%
of
the
LT50
values.
This
is
consistent
with
Brett
(
1958,
page
76
and
Figure
4),
who
indicated
that
differences
between
temperatures
for
50%
mortality
and
those
for
<
50%
mortality
are
relatively
small.
This
"
implies
that
temperatures
of
this
order
have
only
to
increase
slightly
to
cause
a
large
difference
in
mortality."
Based
on
visual
inspection
of
the
LT50
and
LT10
curves,
the
slopes
were
similar.
That
is,
on
the
log­
time
scale,
the
differences
between
the
LT50
and
LT10
curves
were
approximately
constant.
(
There
was
insufficient
information
4­
5
Table
4.2
Relationship
between
LT50
and
LT10
for
sockeye
salmon.
Acclimation
temperature
is
20oC.
Data
from
McConnel
and
Blahm
(
1970),
unpublished
data
(
cited
in
U.
S.
EPA
1977).

Time
(
hrs)
LT50
(
oC)&
LT10
(
oC)#
LT10/
LT50
Ratio
(%)
Delta
T
(
oC)

.1
29.2
28.6
98
0.6
.25
28.4
27.9
98.12
0.5
.5
27.9
27.4
98.21
.0.5
1
27.3
26.9
98.31
0.5
2
26.8
26.3
98.41
0.4
3
26.5
26.0
98.47
0.4
4
26.2
25.8
98.52
0.4
6
25.9
25.5
98.58
0.4
8
25.7
25.3
98.63
0.4
12
25.4
25.0
98.7
0.3
16
25.1
24.8
98.74
0.3
20
24.9
24.6
98.78
0.3
24
24.8
24.5
98.81
0.3
32
24.6
24.3
98.86
0.3
40
24.4
24.1
98.9
0.3
60
24.1
23.8
98.98
0.2
80
23.8
23.6
99.03
0.2
&
regression
coefficients
a=
16.7328,
b=­
0.5473
#
regression
coefficients
a=
17.5227,
b=­
0.5861
Table
4.3
Relationship
between
LT50
and
LT10
for
chinook
salmon.
Acclimation
temperature
is
20oC.
Data
from
McConnel
and
Blahm
(
1970),
unpublished
data
(
cited
in
U.
S.
EPA
1977).

Type
Time
(
hrs)
LT50
(
oC)&
LT10
(
oC)#
LT10/
LT50
Ratio
(%)
Delta
T
(
oC)

Spring
runa
0.1
28.4
28.1
98.74
0.4
0.25
27.9
27.6
98.86
0.3
0.5
27.5
27.2
98.95
0.3
2
26.6
26.4
99.13
0.2
4
26.2
26.0
99.23
0.2
6
26.0
25.8
99.28
0.2
8
25.8
25.6
99.33
0.2
10
25.7
25.5
99.36
0.2
16
25.4
25.2
99.43
0.1
24
25.1
25.0
99.49
0.1
40
24.8
24.7
99.57
0.1
60
24.6
24.5
99.64
0.1
80
24.4
24.3
99.68
0.1
100
24.3
24.2
99.72
0.1
Fall
runb
0.1
28.5
28.1
98.65
0.4
0.25
28
27.6
98.6
0.4
0.5
27.6
27.2
98.56
0.4
2
26.8
26.3
98.49
0.4
4
26.4
25.9
98.44
0.4
6
26.1
25.7
98.42
0.4
8
26
25.5
98.4
0.4
10
25.8
25.4
98.39
0.4
16
25.6
25.1
98.36
0.4
24
25.3
24.9
98.33
0.4
40
25
24.6
98.3
0.4
60
24.8
24.4
98.27
0.4
80
24.6
24.2
98.25
0.4
100
24.5
24.1
98.24
0.4
&
regression
coefficients
LT50:
a=
21.3981,
b=­
0.7253
lT10:
a=
22.6664,
b=­
0.7797
#
regression
coefficients
LT50:
s
a=
22.2121,
b=­
0.7526
LT10
a=
21.6756,
b=­
0.7438
4­
6
presented
in
Appendix
B
of
the
EPA
(
1977)
document
to
statistically
compare
the
slopes.)
Had
the
differences
not
appeared
constant,
the
application
of
a
singe
adjustment
factor
would
not
have
been
appropriate.

ACUTE
THERMAL
EFFECTS
CURVES
AT
10%
MORTALITY
The
adjustment
factor
estimated
from
the
McConnell
and
Blahm
(
1970)
and
Blahm
and
McConnell
(
1970)
data
for
sockeye
and
chinook
salmon
was
assumed
to
be
appropriate
to
use
in
estimating
LT10
curves
for
the
other
salmon
and
trout
species.
For
the
other
studies
from
which
LT50
curves
were
generated,
LT10
curves
were
estimated
by
applying
a
factor
of
0.98
to
each
LT50
curve.
The
equation
used
to
calculate
the
estimated
LT10
values
is
98
.
0
)
)
60
*
(
(
log
10
10
×
-
=
b
a
t
LT
(
4.4)

The
estimated
LT10
curves
for
15oC
acclimation
for
four
salmon
species
are
provided
in
Figure
4.1.
(
Relationships
for
all
species
and
acclimation
temperatures
are
graphically
depicted
in
Appendix
C,
under
separate
cover.)

The
resulting
LT10
lethal
curves
are
very
similar
among
salmon
species
although
cutthroat
trout
have
higher
tolerance
to
high
temperature
than
the
other
species
(
Figure
4.1).
We
note
that
using
the
relationship
based
on
fluctuating
acclimation
temperature
(
Golden
1978)
produced
higher
LT50'
s
than
when
tested
at
constant
temperature.
Continuous
exposures
of
3
to
30
hours
are
necessary
to
cause
mortality
at
temperatures
between
24o
to
26oC,
varying
by
species.
The
duration
of
time
necessary
to
cause
mortality
decreases
sharply
with
small
increments
of
temperature
above
approximately
26oC.
Short
duration
excursions
(
less
than
two
hours)
above
27oC
are
very
likely
to
cause
mortality
of
some
individuals
in
the
population
because
only
one
hour
duration
is
necessary
0.01
0.1
1
10
100
1000
24
25
26
27
28
29
30
Temperature
(
oC)
Time
to
10%
Mortality
(
Hours)

Cutthroat
Rainbow
(
Steelhead)
Chinook
Coho
Figure
4.1
Duration
curve
for
the
LT10
acute
effects
of
temperature
for
coho
and
chinook
salmon
and
cutthroat
and
steelhead
trout,
acclimated
at
15oC.
(
Data
from
Brett
1952,
Alabaster
and
Downing
1966,
Golden
1978;
see
Appendix
C).
4­
7
ACUTE
EXPOSURE
CHARACTERIZATION
The
maximum
temperatures
juvenile
salmonids
experienced
in
situ
were
determined
through
exposure
characterization,
(
U.
S.
EPA
1992).
Summertime
temperatures
for
sites
described
in
Section
3
(
Table
3.1)
were
assembled
for
this
study.
The
occurrence
of
potentially
lethal
temperatures
was
determined
by
examining
the
hourly
temperature
record
at
each
of
the
19
temperature
sites
for
exposure
periods
defined
as
the
number
of
continuous
hours
at
or
above
the
exposure
temperature
(
temperature
was
rounded
to
the
nearest
oC).
We
used
a
lower
level
of
16oC
since
it
approximates
the
optimum
temperature
for
several
salmon
species
(
Weatherly
and
Gill
1995)
although
it
is
important
to
note
that
acute
mortality
does
not
commence
until
24
º
C
under
naturally
fluctuating
conditions.
Exposure
was
based
on
hourly
temperatures
so
that
we
could
capture
relatively
short
duration
effects
and
the
LT10
data
were
expressed
in
hours.

An
exposure
period
is
the
number
of
consecutive
hours
each
temperature
(
measured
to
the
nearest
1.0
°
C)
occurred
within
a
period
when
the
temperature
was
at
or
above
16
°
.
For
example,
if
the
temperature
increased
from
below
16
°
to
16.6
°
,
a
count
of
one
was
added
to
16oC.
If
the
next
hourly
temperature
was
18.5
º
C,
then
a
count
of
one
was
added
to16o,
17o
and
18
º
C.
If
the
following
hourly
temperature
decreased
back
to
17
º
C,
then
a
count
of
one
was
added
to
16o,
and
17o,
and
so
forth.
As
soon
as
the
next
temperature
retreated
below
each
temperature
degree
category,
the
counting
(
i.
e.,
duration
of
exposure)
for
that
temperature
ceased.
When
temperatures
dropped
below
16
º
C,
the
entire
exposure
period
was
concluded.
In
the
example
above,
there
was
one
exposure
period
where
temperature
reached
16
°
for
3
hours,
17
°
for
two
hours,
and
18
°
for
one
hour.

Exposure
to
temperatures
above
16oC
varied
significantly
among
sites.
Table
4.2
lists
the
number
of
exposure
periods
for
temperatures
greater
than
16oC
and
the
number
of
hours
by
temperature
category
for
the
warmest
continuous
exposure
period
at
each
site.
At
some
sites
there
were
no
exposure
periods
at
any
time
during
the
summer
months,
while
at
others
both
the
magnitude
and
duration
of
exposure
were
relatively
large.
Hoffstadt
Creek
in
the
Mt.
St.
Helens
Blast
Zone
experienced
the
warmest
temperature
(
26.0oC)
which
lasted
one
hour
(
Figure
4.3A
and
4.4).
This
stream
is
shallow,
and
it
heats
and
cools
rapidly
over
the
course
of
the
day.
The
deeper
Chehalis
River
sites
(
Figure
4.3B)
did
not
reach
quite
as
high
a
temperature
(
22oC),
but
experienced
higher
temperatures
over
much
longer
continuous
periods.
4­
8
Table
4.
4
Number
of
exposure
sequences
where
temperature
was
continuously
greater
than
or
equal
to
16o
C
for
one
or
more
hour
at
each
of
the
sites
and
the
duration
of
continuous
exposure
(
hours)
at
each
temperature
for
the
warmest
period.

Number
of
Hours
by
Temperature
for
the
Maximum
Sequence
at
the
Site
Site
Total
Number
of
Sequences
_
16_
_
17_
_
18_
_
19_
_
20_
_
21_
_
22_
_
23_
_
24_
_
25_
_
26_

Big
Creek
6
4
0
0
0
0
0
0
0
0
0
0
Chehalis
River
Main
1
49
455
70
21
16
12
8
2
0
0
0
0
Chehalis
River
Main
2
69
44
33
26
21
16
1
6
1
0
0
0
Chehalis
River
Main
3
62
44
32
23
17
11
0
0
0
0
0
0
Crim
Creek
54
22
14
7
3
0
0
0
0
0
0
0
Deschutes
River
Main
70
139
65
17
14
10
7
5
0
0
0
0
Hard
Creek
0
0
0
0
0
0
0
0
0
0
0
0
Harrington
Creek
36
25
17
12
9
5
0
0
0
0
0
0
Hoffstadt
Creek
83
19
16
15
13
12
7
8
7
5
5
1
Huckleberry
Creek
34
405
65
12
0
0
0
0
0
0
0
0
Lester
Creek
47
14
8
5
0
0
0
0
0
0
0
0
Mack
Creek
0
0
0
0
0
0
0
0
0
0
0
0
Porter
Creek
30
21
15
9
0
0
0
0
0
0
0
0
Rogers
Cr.
1
2
0
0
0
0
0
0
0
0
0
0
Sage
Cr.
17
7
0
0
0
0
0
0
0
0
0
0
Salmon
Cr.
2
3
0
0
0
0
0
0
0
0
0
0
Thrash
Cr.
0
0
0
0
0
0
0
0
0
0
0
0
Thurston
Cr.
0
0
0
0
0
0
0
0
0
0
0
0
Ware
Cr.
24
34
23
13
0
0
0
0
0
0
0
0
4­
9
ACUTE
RISK
CHARACTERIZATION
To
assess
acute
risk
to
lethal
temperature,
the
periods
when
duration
of
temperature
equaled
or
exceeded
the
LT10
curve
(
e.
g.
Figure
4.1)
were
examined
to
determine
if
the
duration
of
exceedance
was
of
sufficient
length
to
cause
mortality.
If
the
time
spent
at
any
temperature
equaled
or
exceeded
the
time
necessary
to
elicit
mortality
at
the
LT10
level,
then
an
exceedance
occurred.
If
an
exceedance
occurred,
we
assumed
that
10%
of
the
population
died.
The
number
of
exceedances
that
occurred
during
the
summer
interval
was
counted.
It
follows
that
if
10%
of
the
population
died
at
each
exposure,
then
the
cumulative
risk
to
the
population
(
Lethal
Risk)
could
be
calculated
using
the
number
of
exceedances
(
n)
experienced
by
the
population:

[
]
n
)
10
.
0
0
.
1
(
0
.
1
Risk
Lethal
-
-
=
(
4.5)

where
lethal
risk
is
defined
as
the
proportion
of
the
population
that
dies
due
to
the
temperature
exposure.
For
example,
if
a
stream's
temperature
exceeded
an
LT10
once,
the
cumulative
total
mortality
risk
would
be
[
1.0­(
0.90)
1]
or
10%;
if
it
exceeded
an
LT10
four
times,
the
total
mortality
risk
would
be
[
1.0­(
0.90)
4]
or
34%.

There
were
no
indicated
periods
of
acute
exposure
for
chinook,
coho,
steelhead,
or
cutthroat
at
any
of
the
stream
locations
despite
higher
temperatures
at
some
sites.
The
warmest
period
of
exposure
at
each
of
the
19
temperature
study
sites
in
relation
to
species'
LT10
exposure
curves
is
shown
in
Figure
4.2.
Some
sites
did
not
exceed
optimal
temperature
(
approximately
16oC)
at
any
time
during
the
summer
and
do
not
appear
on
the
figure.
For
those
that
exceeded
acute
lethal
temperatures
(
24oC
or
greater),
the
exposure
was
generally
much
less
than
required
to
cause
mortality.

The
only
site
that
exceeded
24oC
(
the
temperature
zone
where
mortality
depends
on
duration
of
exposure)
was
Hoffstadt
Creek,
which
peaked
at
26.0oC
for
one
hour
(
Figure
4.2
and
4.3A).
Hoffstadt
Creek
had
an
additional
two
exposure
periods
reaching
25oC
(
three­
hour
duration)
and
twelve
reaching
24oC
(
five­
hour
duration).
Although
the
one­
hour
maximum
temperature
approached
the
magnitude
and
duration
that
could
elicit
mortality
at
this
site,
the
duration
of
exposure
at
24
°
,
25
°
,
and
26oC
for
Hoffstadt
Creek
was
too
short
to
directly
cause
mortality
(
Figure
4.3A).
The
Chehalis
0
.0
1
0
.1
1
1
0
1
0
0
1
0
0
0
1
0
0
0
0
1
5
2
0
2
5
3
0
T
e
m
p
e
r
a
tu
r
e
(
oC
)
H
o
u
rs
o
f
C
o
n
tin
u
o
u
s
E
x
p
o
C
u
t
th
r
o
at
S
te
e
lh
e
ad
C
h
in
o
o
k
C
o
h
o
Figure
4.2
Hours
of
continuous
exposure
at
temperature
sites
(
see
Table
4.1)
and
LT10
lethality
relationships
for
fish
species.
4­
10
A)
Hoffs
tadt
Creek
0.1
1
10
100
1000
15
20
25
30
Tempe
rature
(
oC)
Hours
of
Continuous
Exposure
LT10
C
ut
thro
at
LT10
C
hino
o
k
LT10
Steelhead
LT10
C
o
ho
B)
Che
halis
Rive
rM
ains
tem
Site
1
0.1
1
10
100
1000
15
20
25
30
Tem
pe
rature
(
oC)
Hours
of
Continuous
Exposure
Figure
4.3
Example
of
how
acute
exposures
were
summarized:
maximum
number
of
sequential
hours
at
each
temperature
for
A)
Chehalis
River
mainstem
(
site
1)
and
B)
Hoffstadt
Creek
sites.
Diamonds
are
the
exposure
for
each
temperature
for
the
warmest
period
at
each
site.
River
mainstem
experienced
much
longer
duration
at
higher
temperatures
(
Figure
4.3B),
but
never
came
close
to
lethal
temperatures.

No
acute
risks
were
identified
since
none
of
the
water
temperatures
persisted
long
enough
to
exceed
the
lethal
thresholds
(
LT10s)
of
salmonid
species
likely
to
occur
in
these
streams.
Since
n
equaled
0
at
all
sites,
the
probability
of
acute
effects
was
0.
4­
11
Small
to
moderate
size
streams
and
rivers
have
sufficient
fluctuation
in
the
daily
temperature
that
water
does
not
spend
long
continuous
duration
at
lethal
temperature.
The
daily
fluctuation
is
strongly
dependent
on
stream
depth
(
Adams
and
Sullivan,
1990,
Sullivan
and
Adams,
1990).
Within
much
of
the
Pacific
Northwest,
relatively
shallow
streams
and
rivers
may
be
capable
of
achieving
temperatures
that
can
cause
mortality
within
a
few
hours.
Their
shallowness,
however,
will
generally
preclude
them
from
maintaining
high
lethal
temperatures
for
more
than
a
few
hours
during
the
day,
and
temperatures
can
fluctuate
widely
over
the
solar
cycle
(
Brown
1969).
For
example,
Figure
4.4
shows
Hoffstadt
Creek
during
its
warmest
7­
day
period
in
1990.
The
duration
of
lethal
temperature
is
short.

Conversely,
large
deep
rivers
may
maintain
higher
average
temperature,
but
the
daily
fluctuation
is
smaller.
For
example,
for
much
of
the
summer
and
its
length,
the
Columbia
River
maintains
an
average
temperature
of
22oC
with
a
daily
fluctuation
of
0.5
oC.
In
the
U.
S.
G.
S.
water
resource
records
discussed
in
Section
3,
the
S.
Umpqua
River
in
southern
Oregon
reached
30oC
annual
maximum
temperature.
Nevertheless,
the
daily
minimum
temperature
on
the
hottest
days
did
not
exceed
21o
to
23.5oC,
indicating
that
duration
of
extreme
temperature
was
short,
and
probably
within
lethal
thresholds.
Thus,
duration
of
exposure
at
higher
temperatures
can
be
significant,
but
lethal
temperatures
that
cause
mortality
in
a
few
hours
are
not
likely
to
occur
in
most
cases.
The
dependence
of
daily
fluctuation
on
stream
depth
makes
the
case
that
very
large,
deep
rivers
are
different
than
most
of
the
streams
and
rivers
in
the
region
(
Section
3),
and
those
rivers
with
extreme
temperatures
should
probably
be
analyzed
individually
to
fully
understand
their
local
site
conditions.

Table
4.5
shows
the
calculated
temperature
for
1­
hour
and
6­
hour
exposure
that
would
cause
mortality
of
10%
of
the
population
based
on
relationships
provided
in
Table
4.1,
and
the
calculation
methods
described
in
this
section.

Hoffstadt
Creek,
Aug
7­
14,
1990
10
12
14
16
18
20
22
24
26
28
0
24
48
72
96
120
144
168
192
Hours
Temperature
(
oC)

Figure
4.4
Hourly
temperature
during
the
warmest
7­
day
period
at
Hoffstadt
Creek
in
1990.
4­
12
The
temperatures
at
which
mortality
occurs
at
exposures
of
1
hour
and
6
hours
if
very
similar
among
species.
This
temperature
is
between
25.4oC
and
27.4oC,
averaging
26.4oC.
There
is
a
low
probability
that
the
6­
hour
duration
temperature
occurring
in
most
natural
streams
in
Washington,
as
suggested
by
Table
4.1.
Therefore,
an
acute
temperature
for
salmon
of
approximately
26oC
in
the
annual
maximum
temperature
is
suggested
to
prevent
mortality
with
a
margin
of
safety.
We
also
suggest
that
a
site­
specific
analysis
of
temperature
be
performed
when
annual
maximum
temperature
is
between
24o
and
26oC,
to
determine
whether
duration
is
sufficient
to
cause
mortality.
It
would
appear
that
in
most
cases
it
will
not,
but
until
more
analyses
are
completed
over
a
broader
geographic
range
this
will
not
be
known
with
certainty.
Site
specific
analyses
would
also
be
warranted
at
sites
of
thermal
pollutant
sources
that
alter
the
normal
daily
temperature
fluctuation.

Table
4.5.
Temperature
of
1­
hour
and
6­
hour
duration
that
will
cause
mortality
of
10%
of
the
population
for
salmon
and
trout
species.

Species
Age/
size
Acclimation
Temp
(
oC)
1­
Hour
Temperature
(
oC)
6­
Hour
Temperature
(
oC)

Coho
salmon
juvenile
15
26.6
25.5
20
27.2
26.1
23
28.0
26.8
Rainbow
trout
juvenile
15
27.2
25.7
18
27.5
26.9
20
28.0
26.8
Rainbow
trout
yearling
20
28.2
26.5
20
28.5
26.8
20
28.0
26.3
20
28.3
26.3
Cutthroat
trout
juvenile
23
28.3
27.1
13­
25
(
fluctuating)
28.5
27.9
23
28.3
26.3
13­
23
(
fluctuating)
28.0
26.7
Chinook
salmon
juvenile
15
26.8
25.4
20
27.5
26.8
jacks
19
30.1
27.1
juvnile
spring
20
26.5
25.5
20
26.8
25.8
Juvenile
fall
20
26.6
25.6
20
26.8
25.7
Sockeye
salmon
juvenile
15
26.5
25.1
20
26.0
25.9
23
27.5
26.3
underyearling
20
26.8
25.4
Chum
salmon
juvenile
15
26.3
24.9
20
27.8
25.9
23
28.2
26.6
Discussion
We
conclude
that
direct
mortality
from
temperature
is
unlikely
to
affect
populations
living
in
streams
with
temperatures
similar
to
those
evaluated
in
this
paper.

Mortality
associated
with
acute
lethal
temperatures
has
been
previously
reported
from
streams
similar
to
those
in
this
study.
In
both
cases,
natural
and
logging­
related
disturbance
caused
complete
denudation
of
the
watershed,
including
all
riparian
vegetation
and
in­
channel
cover.
In
the
years
immediately
following
the1980
eruption
of
Mt.
St.
4­
13
Helens,
temperatures
of
streams
within
the
blast
zone
were
extreme
relative
to
those
typically
recorded
in
the
Pacific
Northwest.
Bisson
et
al.
(
1988)
and
Martin
et
al.
(
1986)
reported
temperatures
in
Hoffstadt
Creek
and
other
moderate
size
streams
in
the
blast
zone
that
exceeded
29
°
C
for
short
daily
intervals
several
times
during
the
summer.
The
lethality
curve
suggests
that
even
short
duration
exposure
to
such
warm
temperatures
could
cause
mortality
of
individual
fish
and
repeated
exposures
could
have
a
measurable
effect
on
salmon
populations.
In
this
case,
the
acute
effects
analysis
would
have
predicted
mortality,
although
we
do
not
have
temperature
data
from
these
earlier
studies
to
calculate
the
cumulative
mortality.
Populations
of
juvenile
coho
salmon
successfully
survived
the
temperature
episodes
exceeding
lethal
levels
(
Bisson
et
al.
1988).
However,
Martin
et
al.
(
1986)
also
reported
significant
mortality
of
juvenile
coho
within
populations
in
the
same
streams
as
Bisson
et
al.
(
1988)
that
was
proportional
to
the
magnitude
of
daily
temperature
fluctuations
(
maximum
fluctuation
of
17
°
C)
when
temperatures
exceeded
26
°
C.

Temperatures
have
since
sufficiently
declined
with
vegetation
re­
growth
that
mortality
is
not
expected
at
temperatures
existing
eight
years
after
the
eruption
and
beyond.
Hall
and
Lantz
(
1969)
reported
no
reduction
in
numbers
of
coho,
and
a
25%
reduction
in
cutthroat
trout,
when
summertime
temperatures
reached
30oC
in
a
small
Oregon
stream
(
Needle
Branch)
after
clearcut
logging
and
severe
burning.

Elsewhere
in
the
Pacific
Northwest,
water
temperatures
greater
than
24oC
are
measured
relatively
rarely
in
streams
and
rivers
(
see
Section
3).
Based
on
the
foregoing
analysis,
one
can
conclude
there
is
low
risk
of
acute
lethality
to
salmonid
species
juveniles
from
temperatures
observed
at
the
nineteen
sites
in
this
risk
assessment.
The
temperatures
at
these
sites
are
representative
of
the
potentially
lethal
temperature
of
most
of
the
natural
streams
up
to
5th
order
found
in
the
Pacific
Northwest,
including
many
with
disturbed
riparian
forest
cover
(
Sullivan
et
al.
1990).
Streams
in
this
analysis
include
many
sites
affected
by
land
use
and
catastrophic
natural
disturbance.
Nevertheless,
if
temperatures
should
reach
as
high
as
28oC
for
as
little
as
an
hour,
mortality
of
some
portion
of
the
population
can
be
expected.

CONCLUSIONS
·
 
There
is
sufficient
information
to
quantitatively
define
the
lethal
effects
of
temperature
on
salmonids.

·
 
No
occurrences
of
acute
lethal
temperatures
were
observed
at
stream
sites
with
a
wide
range
of
temperatures
including
many
with
annual
maximum
temperatures
that
well
exceed
current
water
quality
standards.

·
 
Nevertheless,
lethal
level
temperatures
of
sufficient
duration
to
cause
mortality
have
been
reported
in
the
Pacific
Northwest.
Therefore,
although
not
a
common
occurrence,
attention
should
be
paid
to
local
site
conditions
that
can
lead
to
acute
mortality.

·
 
A
temperature
threshold
of
26oC
is
suggested
to
prevent
mortality
of
salmon
and
trout
species
in
natural
rivers
and
streams.
Further
analysis
of
temperature
to
determine
exposure
is
suggested
for
streams
where
annual
maximum
temperature
is
between
24o
and
26oC.
4­
14
5­
1
SECTION
5
DEVELOPMENT
AND
CORROBATION
OF
A
BIOENERGETICS
 
BASED
APPROACH
TO
EVALUATING
SALMON
GROWTH
IN
RELATION
TO
ENVIRONMENTAL
TEMPERATURE
Abstract
Growth
is
an
important
biologic
function
for
juvenile
salmonids
rearing
in
streams
and
rivers.
In
this
section,
we
develop
a
quantitative
method
to
evaluate
the
effects
of
the
long­
term
temperature
on
the
growth
of
salmonids.
The
mathematical
model
considers
the
effects
of
temperature
and
food
consumption
on
daily
growth
rate.
When
applied
over
time
to
measured
stream
temperature
regimes,
the
model
simulates
the
weight
gain
of
salmon
species,
and
can
be
used
to
assess
the
importance
of
the
cumulative
or
chronic
effects
of
temperature
on
growth.
At
present,
the
model
is
formulated
to
assess
the
growth
of
juvenile
salmonid
species
during
the
summer
months
with
variable
temperatures.
Previous
researchers
have
used
similar
approaches,
although
this
specific
method
has
not
been
explicitly
described
previously.
We
use
well­
established
bioenergetics
principles
and
models
available
in
the
scientific
literature
to
help
develop
it.

The
relatively
simple
formulation
appears
to
predict
weights
well
according
to
a
number
of
comparisons
of
observed
and
predicted
growth
at
stream
sites
where
fish
populations
had
been
sampled.
The
method
is
sensitive
to
temperature,
making
it
a
useful
tool
for
evaluating
salmon
response
to
temperature
in
natural
streams,
and
it
allows
direct
comparisons
among
sites
and
species
if
desired.
Application
of
bioenergetics
principles
to
field
observations
using
this
tool
suggests
ecological
adaptation
to
environmental
temperature
and
food
supply.
The
model
may
prove
useful
for
helping
field
investigators
sort
out
the
complex
relationships
between
population
dynamics,
environmental
temperature,
and
food
supply
that
control
growth
in
natural
streams.

Because
the
model
is
central
to
determining
the
effects
of
temperature
on
fish
growth
in
relation
to
temperature,
the
mathematical
approach
is
developed
in
detail
in
this
section.
If
the
scientific
basis
for
the
growth
analysis
is
not
of
interest,
the
reader
is
urged
to
move
directly
to
Section
6,
where
the
method
is
applied
to
develop
temperature
criteria.

Key
findings
include:

‰
Methods
of
predicting
growth
based
on
quantitative
bioenergetics
principles
can
be
applied
to
streams
to
assess
the
effects
of
temperature
on
juvenile
salmonid
growth,
with
results
that
are
consistent
with
observed
wild
fish
population
growth
patterns.

‰
The
method
is
sensitive
to
temperature
and
can
be
applied
to
the
daily
temperature
regime
making
it
a
useful
tool
for
assessing
the
biological
impacts
of
temperature
in
natural
streams.
5­
2
INTRODUCTION
Of
primary
interest
in
selecting
temperature
criteria
is
the
prevention
of
adverse
effects
of
longterm
exposure
to
temperatures
detrimental
to
fish.
To
explore
the
effects
of
prolonged
exposure
to
temperature,
numerous
investigators
have
found
growth
to
be
a
reliable
and
measurable
integrator
of
a
variety
of
physiological
responses
(
Brett
1971,
1995;
Iverson
1972;
Brungs
and
Jones
1977;
Wurtsbaugh
1973).
Growth
rate
is
the
most
frequently
reported
measure
of
fish
health
from
laboratory
studies
and
occasionally
from
field
studies.
By
corollary,
the
weight
a
fish
gains
over
a
time
interval
is,
in
part,
determined
by
the
ambient
temperature
of
the
water
in
which
it
lives.

The
size
of
salmonids
during
juvenile
and
adult
life
stages
influences
survival
and
reproductive
success.
Although
the
large
majority
of
anadromous
salmonid
growth
occurs
in
the
ocean
environment,
growth
of
juveniles
in
natal
streams
is
especially
important
for
anadromous
salmonids
that
must
reach
minimum
sizes
before
they
can
smolt
(
Weatherly
and
Gill
1995).
Holtby
and
Scrivener
(
1989)
and
Quinn
and
Peterson
(
1996)
demonstrated
that
the
size
achieved
by
juvenile
coho
at
the
end
of
their
first
summer
growing
period
was
a
strong
determinant
of
their
later
success
in
overwintering
and
smolting.
Larger
size
also
conveys
competitive
advantage
for
feeding
in
the
freshwater
environment
(
Puckett
and
Dill
1985,
Nielsen
1994)
for
both
resident
and
anadromous
species.
Mason
(
1976)
and
Keith
et
al.
(
1998)
found
that
the
smaller
fish
tend
to
be
those
that
are
lost
from
rearing
populations.
Brett
et
al.
(
1971)
described
the
freshwater
rearing
phase
of
juvenile
sockeye
as
one
of
restricted
environmental
conditions
and
generally
retarded
growth.
This
synopsis
is
generally
true
for
salmonid
species
that
dwell
in
stream
and
river
environments
for
lengthy
periods
of
time.

Growth
can
be
viewed
as
the
net
effect
of
the
environment
on
the
relation
between
food
consumption,
metabolism,
and
activities
of
an
organism
(
Warren
1971).
The
bioenergetics
of
many
fish
species,
including
salmonids,
has
been
widely
studied
and
quantified
(
reviewed
by
Adams
and
Breck
1990,
Brett
1995,
and
others).
The
food
fish
consume
is
allocated
to
maintaining
basic
metabolic
functions,
growth
and
waste.
The
most
generalized
representation
of
bioenergetics
is
of
the
form:

Consumption
=
Metabolism
+
Growth
+
Wastes
Researchers
have
developed
mathematical
relationships
for
the
rate
functions
for
various
energetic
processes
(
summarized
in
Brett
1995).
We
propose
that
a
bioenergetics
approach
can
be
used
to
assess
the
effects
of
temperature
exposure
because
many
of
the
energetic
relationships
are
based
on
temperature
and
the
energy
balance
determines
growth.

Bioenergetics
models
vary
in
their
formulations
of
specific
functions
and
in
the
complexity
with
which
they
attempt
to
characterize
the
details
of
each
(
Ney
1993).
The
most
comprehensive
bioenergetics
models
consider
all
modes
of
energy
intake
and
expenditures
including
consumption,
respiration,
active
metabolism,
specific
dynamic
action
(
swimming),
wastes
(
including
egestion
and
excretion),
and
growth
(
including
body
tissue
and
gonads)
(
Kitchell
et
al.
1974,
Hewett
and
Johnson
1992).
Bioenergetics
models
combine
these
relationships
into
an
energy
balance
where
the
energy
consumed
and
the
energy
expended
to
all
functions
must
sum
to
one.
Elliott's
(
1976)
relationships
for
brown
trout
are
also
widely
cited.
Bioenergetics
models
have
been
developed
for
a
number
of
applications
in
fisheries
management
and
their
uses
are
many
and
varied
(
Hansen
et
al.
1993,
Ney
1993).
They
have
been
used
routinely
to
manage
the
rearing
of
fish
in
hatcheries
(
McLean
et
al.
1993)
and
populations
in
natural
environments
(
Hanson
et
al.
1997).
Increasingly,
5­
3
they
have
been
used
to
explore
ecological
responses
to
environmental
conditions
(
Filbert
and
Hawkins
1995,
Preall
and
Ringler
1989,
Railsback
and
Rose
1999).

J.
F.
Kitchell
and
colleagues
at
the
University
of
Wisconsin
have
made
bioenergetics
analysis
available
to
a
wide
variety
of
scientists
and
managers.
They
have
summarized
research
for
many
fish
species,
including
salmonids,
and
packaged
the
energy
functions
in
software
for
easier
use
of
the
multiple
mathematical
statements
required
for
the
energy
balance
(
Kitchell
et
al.
1974,
Hewett
and
Johnson
1992,
Hansen
et.
al.
1997).
In
practice,
model
users
are
urged
to
supply
data
to
calibrate
model
parameters
and
to
validate
population
growth.
There
are
a
variety
of
data
needs
in
conducting
full
bioenergetics
analysis.
The
Wisconsin
models
require
a
relatively
large
number
of
parameters
(
15­
30),
some
of
which
are
measured
from
the
population
and
environment
of
interest
(
Hanson
1997).
Default
values
are
supplied
if
the
user
is
unable
to
develop
local
data,
often
times
borrowed
from
similar
species
that
have
received
greater
laboratory
and
field
study.
The
proliferation
of
parameters,
each
with
its
own
estimation
error,
has
led
some
to
criticize
bioenergetic
models
for
being
insensitive
statistically
and
difficult
to
apply
(
Hansen
et
al.
1993).

For
practical
applications,
it
is
often
desirable
to
construct
the
simplest
model
possible
that
can
capture
the
key
environmental
or
biological
effects
of
interest
(
Kitchell
et
al.
1974).
Moreover,
Ney
(
1993)
has
suggested
that
elaborate
energetic
characterizations
may
not
be
necessary
to
provide
satisfactory
answers
to
some
bioenergetics
questions.
Although
grwoth
is
only
one
of
the
bioenergetic
functions,
many
authors
argue
that
it
integrates
a
host
of
specific
physiological
responses
to
temperature,
including
metabolic
rate
(
basal
and
active),
feeding
and
digestion,
and
swimming
performance
or
the
ability
to
hold
position
with
the
current
(
Brett
1995;
Weatherly
and
Gill
1995).
For
example,
Brett
et
al.
estimated
weight
gain
of
sockeye
(
1971)
and
chinook
(
1982)
in
relation
to
environmental
temperature
assuming
that
the
relationship
between
temperature,
food
consumption
and
growth
rate,
such
as
illustrated
in
Figure
2.4,
adequately
integrates
the
organism's
response
to
temperature.
Mallet
et
al.
(
1999)
applied
a
temperature­
modified
form
of
a
von
Bertalanffy
growth
model,
which
has
no
explicit
of
energetics,
to
estimate
the
growth
of
grayling
in
a
European
River.
Up
to
25
to
30%
of
the
energy
consumed
by
salmonids
is
allocated
to
growth
and
the
remainder
is
allocated
to
the
other
energy
demands
(
Brett
et
al.
1982,
Brett
1995).
With
the
exception
of
respiration,
the
pattern
of
response
of
all
of
the
energetic
functions
to
temperature
is
similar
to
that
of
growth
rate,
with
maximums
and
minimums
of
activity
peaking
and
declining
at
similar
optimal
temperatures
(
e.
g.
Brett
et
al.
1971,
Hansen
et
al.
1997).
Energy
consumed
by
respiration
continually
increases
reaching
maximums
at
temperatures
coincident
with
shut
down
of
other
metabolic
functions,
including
growth
(
e.
g.,
24oC
for
salmonids,
Brett
1995).

Growth,
or
more
precisely,
gain
in
weight
for
our
purpose,
is
one
of
the
few
energetic
functions
that
can
be
readily
measured
in
natural
environments,
and
it
perhaps
is
the
quality
of
greatest
interest
for
juvenile
salmonids.
Because
bioenergetic
functions
respond
similarly
to
growth,
the
approach
of
Brett
et
al.
(
1982)
assumes
that
the
other
components
of
the
energy
equation
can
be
ignored
yielding
a
modeling
approach
that
requires
only
a
few
parameters.

In
this
section,
we
develop
a
mathematical
approach
to
predict
growth
from
temperature
and
food
consumption,
using
realistic
estimates
of
food
supply.
It
follows
the
approach
of
Brett
et
al.
(
1971,
1982)
in
that
only
growth
rate
as
mediated
by
temperature
and
food
consumption
are
accounted
for
in
the
energy
balance.
However,
we
use
the
bioenergetics
approach
to
estimate
the
interaction
of
food
consumption
with
temperature.
We
develop
the
rationale
for
each
of
the
components
of
the
growth
model,
and
parameterize
them
based
on
laboratory
and
field
studies
of
fish
populations.
We
then
compare
model
simulation
results
with
fish
population
weight
gain
at
a
number
of
stream
5­
4
sites
to
evaluate
model
performance.
This
analysis
considers
growth
during
summer
rearing
because
high
water
temperatures
are
of
a
particular
concern
in
how
they
may
restrict
growth.
The
analysis
focuses
primarily
on
coho
and
steelhead
species
because
1)
they
occur
widely
in
Pacific
Northwest
streams,
penetrating
well
into
the
headwaters,
and
2)
necessary
biological
data
were
available
for
constructing
the
model
and
evaluating
its
performance.
The
growth
model
is
then
used
in
Section
6
to
explore
methods
for
quantifying
growth
response
to
identify
biologically­
based
water
quality
criteria.

GROWTH
MODEL
The
basis
for
the
mathematical
formulation
for
daily
weight
gain
in
juvenile
salmonids
is
its
relationship
to
temperature
and
food
consumption.
This
relationship
has
been
graphically
depicted
for
sockeye
and
chinook
salmon
by
Brett
and
colleagues
(
summarized
in
Brett
1995
and
Weatherly
and
Gill
1995).
The
relationship
for
sockeye
salmon
is
shown
in
Figure
2.4
of
this
report.

The
change
in
weight
is
calculated
for
defined
scenarios
of
temperature,
food
availability,
and
species
size
characteristics
on
a
daily
basis,
and
summed
through
the
growth
period.
Weight
gain
is
determined
by
multiplying
the
daily
specific
growth
rate
by
the
body
weight:

i
i
i
w
g
w
×
=
D
(
5.1)

where:

)
(
gram
day
t,
body
weigh
day)
/
gram
(
gram
day
rate,
growth
specific
/
day)
(
gram
day
growth,
s
day'
current
t
body
weigh
t
body
weigh
growth
growth
i
w
i
g
i
w
i
i
i
=
×
=
=
D
Next
day's
weight
is
computed
by
adding
the
daily
growth
to
the
current
day's
weight:

(
)
i
i
i
i
i
w
g
w
w
w
×
+
=
D
+
=
+
1
1
(
5.2)

Weight
at
some
time
t
can
be
computed
as
a
function
of
an
initial
weight
w0
and
daily
growth
rate
coefficients:

(
)
Õ
-
=
+
×
=
1
0
0
1
t
i
i
t
g
w
w
(
5.3)

The
daily
specific
growth
rate
is
a
function
of
the
water
temperature
to
which
the
fish
are
exposed
that
day
and
daily
food
consumption:

)
,
(
f
g
i
i
i
C
T
g
=
(
5.4)

where:
5­
5
day)
/
gram
(
gram
day
n
consumptio
daily
C)
(
day
e,
temperatur
t
body
weigh
food
×
=
°
=
i
C
i
T
i
i
In
turn,
the
consumption
is
a
function
of
the
water
temperature
to
which
the
fish
are
exposed
that
day,
the
weight
of
the
fish,
and
the
food
supply
(
Ri):

)
,
,
f(
i
i
i
i
R
W
T
C
=
(
5.5)

Looking
at
equations
5.1,
5.4
and
5.5,
one
can
see
that
growth
is
being
modeled
as
a
function
of
water
temperature,
consumption,
and
size
of
the
fish.
The
interaction
of
these
terms
is
discussed
in
detail
as
the
mathematical
relationships
are
developed.
The
fact
that
consumption
is
both
a
dependent
and
an
independent
variable
of
temperature
introduces
some
complexity
into
the
growthcalculation
that
must
be
addressed.

The
growth
model
is
developed
in
two
parts.
First,
the
relationship
between
consumption,
temperature
and
weight
(
equation
5.5)
is
estimated
following
the
approach
discussed
by
Stewart
and
Ibarra
(
1991),
and
used,
for
example,
in
the
Wisconsin
bioenergetics
model
(
Hewett
and
Johnson
1992,
Hanson
1997).
Second,
the
relationships
between
specific
growth
rate,
temperature
and
consumption
(
equation
5.4)
are
developed
for
several
salmon
species.
We
draw
from
previously
published
laboratory
experiments
to
establish
these
relationships.
These
studies
exposed
juvenile
salmonids
for
periods
sufficiently
long
to
determine
the
rate
of
growth
(
e.
g.,
change
in
weight
from
one
interval
to
the
next)
of
the
sample
population
at
a
given
temperature
and
food
consumption.
Experimental
data
for
coho
salmon
(
Oncorhynchus
kisutch)
were
obtained
from
Everson
(
1973)
and
for
steelhead
trout
(
Oncorhynchus
mykiss)
from
Wurtsbaugh
and
Davis
(
1977).
We
assumed
that
these
experimental
data
are
representative
of
the
species'
response
to
temperature
and
consumption
in
the
natural
environment.
(
Data
from
these
experiments
is
provided
in
Appendix
A.)

Consumption.
There
is
a
maximum
level
of
food
consumption
for
each
species
that
constitutes
satiation
or
fullness.
There
is
also
a
minimum
consumption
required
to
maintain
standard
metabolism.
Consumption
is
generally
expressed
as
a
proportion
of
food
mass
to
fish
body
mass
consumed
each
day
(
g
gbw
 
1d­
1),
or
alternatively,
as
a
percent
of
body
mass
per
day.
Within
the
range
of
consumption
between
satiation
and
minimum
maintenance,
the
growth
rate
varies
with
consumption
as
illustrated
for
sockeye
salmon
in
Figure
2.4.

The
consumption
rate
at
satiation,
(
Cs
),
and
by
corollary,
at
lesser
amounts
of
food,
varies
with
temperature
and
changes
systematically
with
the
fish's
weight
(
Brett
1995).
Salmonids
respond
physiologically
to
temperature
by
altering
food
consumption
and
the
efficiency
with
which
food
is
converted
to
growth
(
Weatherly
and
Gill
1995).
Cs
is
at
a
maximum
at
optimal
temperature
and
declines
at
colder
and
warmer
temperatures
(
Brett
1971),
yielding
the
characteristic
shape
of
growth/
temperature
response
(
e.
g.,
Figure
2.4).
For
example,
coho
consumption
was
40%
greater
at
21oC
than
at
11oC
in
the
laboratory
experiments
(
Everson
1973),
but
dropped
off
sharply
above
22oC
(
Brett
et
al.
1982).
Consumption
at
satiation
also
declined
with
increasing
fish
weight
(
Brett
1995).
The
rate
of
consumption
of
Everson's
(
1973)
experimental
fish
at
3.6
grams
was
45%
of
that
of
fish
weighing
2.0
grams
at
the
same
temperature.

There
is
a
maximum
consumption
rate
for
each
species,
(
Cmax),
a
key
benchmark
established
in
laboratory
studies
at
optimum
temperature,
low
weight,
and
unlimited
food
supply.
Generally,
the
5­
6
consumption
at
1
gram
of
weight
and
optimum
temperature
is
the
highest
consumption
likely
to
be
observed
for
the
species
(
Cmax),
and
serves
as
an
important
reference
point
for
growth
computations
(
Hanson
1997,
Beauchamp
et
al.
1989,
Hewett
and
Johnson
1987).
The
maximum
consumption
at
satiation
rations
for
other
temperatures
and
weights
(
Cs)
is
less
than
Cmax
,
and
can
be
calculated
as
a
proportion
of
maximum
(
p)
according
to
equation
5.6.

max
C
C
p
s
=
(
5.6)

where:

(
)
)
d
g
g
,
day
t
body
weigh
of
gram
food
of
(
gram
re
temperatu
optimum
and
weight
reference
at
ration
satiation
ess
dimensionl
n
consumptio
normalized
1
­
1
­
bw
1
­
1
­
max
=
=
C
p
Cmax
varies
by
species
(
sockeye:
Brett
1971;
chinook:
Brett
et
al.
1982;
coho:
Everson
1973,
steelhead:
Wurtsbaugh
and
Davis
1977).

To
appropriately
estimate
growth
rate
(
g),
it
is
necessary
to
establish
the
consumption
at
each
weight
(
Cw)
and
temperature
(
CT).
We
follow
the
approach
used
in
the
Wisconsin
bioenergetics
model
(
Hanson
1997)
where
consumption
equations
are
of
the
basic
form:

)
(
max
T
f
p
C
C
T
×
×
=
(
5.7)

CB
w
W
CA
C
×
=
(
5.8)

We
develop
equations
for
each
independently,
and
then
we
will
bring
the
two
effects
together
by
referencing
consumption
relative
to
Cmax
and
calculating
the
reduction
from
that
benchmark
due
to
each
of
the
two
factors.
Hanson
(
1997)
notes
that
developing
a
set
of
parameters
for
these
relationships
may
be
accomplished
by
deriving
them
from
published
reports,
estimating
them
from
specifically
designed
field
or
laboratory
experiments,
or
borrowing
parameters
from
closely
related
species.

The
allometric
relationship
between
consumption
and
weight
(
eq.
5.8)
generally
has
the
form
of
a
negative
power
function
(
Ricker
1975)
whose
terms
are
the
intercept,
CA,
and
the
coefficient,
CB.
The
intercept
value
of
CA
is
the
consumption
of
a
1­
gram
fish
and
has
units
of
grams
per
gram
of
body
mass
per
day
(
g
g­
1
bw
d­
1).
These
terms
must
be
estimated
from
ad
libitum
feeding
experiments
conducted
at
the
optimum
temperature.
There
are
relatively
few
such
laboratory
studies
reporting
values
for
salmonid
species
of
interest
in
this
report,
although
several
other
salmon
species
have
been
extensively
studied.
The
laboratory
studies
of
juvenile
growth
for
coho
(
Everson
1973)
and
steelhead
(
Wurtsbaugh
and
Davis
1977)
were
not
designed
specifically
to
determine
the
allometric
relationship
of
consumption
to
weight.
However,
in
both
studies,
a
series
of
one­
month
long
feeding
trials
were
conducted
on
individuals
drawn
from
a
population
of
fish
that
was
maintained
in
a
natural
stream
between
experiments
over
a
year­
long
period.
Thus,
the
fish
grew
at
natural
rates
between
experiments
and
represented
a
range
of
weights
from
1
to
4
grams,
which
is
within
the
range
of
size
expected
for
salmonids
in
natural
streams
in
the
first
year.
5­
7
We
examined
the
laboratory
data
of
Everson
(
1973)
and
Wurtsbaugh
and
Davis
(
1977)
to
estimate
CA
and
CB
for
coho
and
steelhead
looking
only
at
trials
where:
1)
fish
weights
were
reasonably
near
1
gram;
2)
rations
were
considered
to
be
at
satiation;
and
3)
temperatures
were
optimal,
as
suggested
by
high
growth
rates.
Data
matching
these
criteria
were
limited
and
some
of
our
selected
data
values
only
marginally
fit
these
criteria.
Nevertheless,
a
relationship
between
consumption
and
weight
at
satiation
ration
followed
an
allometric
relationship
with
CB
equal
to
 
0.254
for
coho
(
R2=
0.19)
and
 
0.311
for
steelhead
(
R2=
0.41),
albeit
the
R2
is
relatively
low.
(
The
poor
fit
in
the
relationship
could
be
because
some
of
the
experimental
data
points
did
not
match
the
criteria
as
closely
as
desirable).
Stewart
and
Ibarra
(
1991)
and
the
Wisconsin
user's
manual
recommend
values
for
CB
of
 
0.275
for
coho,
based
on
the
work
of
Beauchamp
et
al.
(
1989),
and
 
0.30
for
steelhead
based
on
Rand
et
al.
(
1993).
Although
these
studies
were
conducted
with
larger
fish,
the
results
are
very
similar.
Calibrating
predictions
of
weight
gain
of
populations
in
natural
streams,
as
described
later,
we
found
that
selecting
values
for
CB
of
 
0.275
for
both
coho
and
steelhead,
as
suggested
by
other
researchers,
produced
satisfactory
modeling
results.
Setting
CB
at
 
0.3
for
steelhead
slightly
but
consistently
underestimated
growth,
suggesting
that
consumption
was
not
declining
at
quite
this
high
of
a
rate
in
natural
populations.

Maximum
consumption
observed
in
a
few
of
the
laboratory
trials
suggested
values
for
the
intercept,
CA,
of
0.11
and
0.16
g
g­
1d­
1
for
coho
and
steelhead,
respectively.
The
allometric
relationship
developed
from
all
trial
data
suggested
CA
equaled
0.083
and
0.16
for
coho
and
steelhead,
respectively.
Calibrating
the
model
predictions
in
natural
streams
suggested
that
CA
was
closer
to
0.10
for
coho
(
10%
body
weight
per
day).
The
User's
Manual
for
the
Wisconsin
model
(
Hanson
1997)
recommends
values
of
CA
between
0.15
and
0.35,
which
are
higher
than
those
we
derived
using
the
laboratory
data.
We
selected
values
for
CA
of
0.10
and
0.16
for
coho
and
steelhead
respectively.
These
are
also
the
values
of
Cmax
for
each
species.

Bartell
et
al.
(
1986),
Beauchamp
et
al.
(
1989)
and
Hanson
et
al.
(
1997)
have
noted
that
the
allometric
parameters
for
the
dependence
of
consumption
on
body
mass
(
CA
and
CB)
are
among
the
parameters
that
have
the
greatest
influence
on
bioenergetics
predictions.
In
subsequent
model
predictions,
described
later
in
this
section,
we
found
this
to
be
true.
Within
the
narrow
range
of
values
separating
adult
from
juvenile
studies,
there
was
relatively
little
effect
of
CB
on
weight
predictions,
and
we
selected
the
widely
cited
values.
However,
the
laboratory
studies
of
juvenile
fish
did
not
support
CA
values
greater
than
we
selected.
Growth
predictions
are
sensitive
to
this
parameter.

Food
consumption
in
relation
to
weight
as
calculated
with
the
allometric
parameters
is
shown
in
Figure
5.1.
There
are
significant
differences
in
consumption
levels
between
these
two
species.

The
proportion
of
consumption
at
each
weight
(
Cw)
relative
to
Cmax
,
is
defined
by
the
slope
of
the
allometric
equation
(
CB)
and
can
be
calculated
as:

CB
w
W
p
=
(
5.8)
5­
8
To
determine
the
effects
of
temperature
on
food
consumption,
a
number
of
authors
have
used
the
Thornton
and
Lessem
algorithm
(
1978)
(
Beauchamp
et
al.
1989,
Stewart
and
Ibarra
1991,
Hanson
et
al.
1997).
The
algorithm
estimates
the
maximum
consumption
at
each
temperature,
expressed
as
the
proportion
of
the
maximum
consumption
at
the
optimal
temperature
(
pt).
The
algorithm
fits
two
sigmoid
curves
to
specified
parameters,
which
include
the
optimal
temperature
and
the
upper
and
lower
temperatures
where
consumption
nears
zero.
Thus,
the
general
shape
of
the
relationship
between
temperature
and
consumption
is
assumed
and
key
temperatures
must
be
known
or
estimated
to
fit
the
proper
shape
of
the
curve
for
each
species.
The
user's
manual
for
the
Wisconsin
bioenergetics
model
(
Hanson
1997)
suggests
values
for
the
key
parameters
of
the
Thornton
and
Lessem
algorithm
drawn
from
Stewart
and
Ibarra
(
1991)
and
Rand
et
al.
(
1993).
We
began
with
those
parameters,
but
then
calibrated
them
to
fit
the
experimental
data
for
coho
(
Everson
1973)
and
steelhead
(
Wurtsbaugh
and
Davis
1977)
(
Figure
5.2).
Small
adjustments
to
the
suggested
parameters
appeared
to
slightly
improve
the
fit
compared
to
the
laboratory
observations,
although
our
final
curve
for
coho
is
very
similar
to
that
presented
by
Stewart
and
Ibarra
(
1991).
The
parameter
values
we
derived
are
provided
in
Table
5.1
and
the
fitted
relationships
are
shown
in
Figure
5.2.
The
Thornton
and
Lessem
equation
is
awkward
to
use
in
an
EXCEL
â
spreadsheet
format.
We
fit
a
polynomial
regression
to
the
Coho
0.0
0.2
0.4
0.6
0.8
1.0
0
5
10
15
20
25
30
Temperature
(
oC)
Polynomia
Line
Fit
Steelhead
0
0.2
0.4
0.6
0.8
1
0
5
10
15
20
25
30
Temperature
(
oC)
Polynomial
Line
Fit
Figure
5.2
Maximum
consumption
in
relation
to
temperature
computed
with
the
Thornton
and
Lessem
algorithm
(
1978).
Consumption
at
Satiation
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0.18
0
1
2
3
4
5
6
7
8
9
10
11
Weight
(
gram)
Consumption
(
g
g
­
1d­
1)

Coho
Steelhead
Cmax
Figure
5.1
Consumption
at
satiation
in
relation
to
weight
as
modeled
by
selected
parameters
for
CA
and
CB
described
in
text.
5­
9
Thornton
and
Lessem
algorithm
results
to
provide
an
equation
for
calculating
pt
at
each
temperature
(
Figure
5.2).

3
3
2
2
1
0
T
T
T
p
t
×
+
×
+
×
+
=
l
l
l
l
(
5.9)

The
fit
was
quite
close
(
Figure
5.2),
especially
in
the
range
of
temperatures
likely
to
be
observed
during
the
summer,
when
the
relationship
was
used
for
growth
modeling.
The
polynomial
terms
describing
the
equations
are
provided
in
Table
5.1.

We
now
have
mathematical
expressions
that
account
for
the
influence
of
weight
(
pw),
and
temperature
(
pt)
on
consumption
at
satiation
where
each
is
expressed
in
proportion
to
Cmax.
The
maximum
consumption
rate,
ps,
at
each
combination
of
weight
and
temperature
expressed
relative
to
Cmax,
(
a
constant
value
for
each
species)
is
calculated
according
to
equation
5.10:

t
w
s
p
p
p
×
=
(
5.10)

and
ranges
between
0
and
1.
The
actual
amount
of
food
consumed
(
Cs),
expressed
in
grams
of
prey
mass
in
relation
to
grams
of
body
mass
per
day,
(
g
gbw
­
1d­
1),
is
equal
to:

max
C
p
C
s
s
×
=
(
5.11)

Table
5.1
Estimated
physiological
parameter
values
used
in
equations
calculating
consumption
for
coho
and
steelhead.

Relationship
Parameter
Coho
Steelhead
CA
0.10
0.16
Allometric
CB
­
0.275
­
0.275
CQ
7
5
CK1
0.4
0.2
CTO
15.6
14.0
CTM
18
17.9
CTL
24
24
Thornton
and
Lessem
function
for
temperature
dependence
(
1978)
f
CK4
0.2
0.1
l
0
­
0.1419
­
0.1229
l
1
0.0544
0.0607
l
2
0.0061
0.0055
Regression
fit
to
Thornton
and
Lessem
algorithm
generated
by
above
parameters
l
3
­
0.0003
­
0.0003
f
Terminology
as
used
by
Hanson
et
al.
(
1995),
see
Appendix
B
The
value
of
Cs
computed
over
a
range
of
temperature
and
weight
is
illustrated
for
coho
in
Figure
5.3.
As
expected
from
the
formulation,
the
highest
levels
of
consumption
occur
near
optimal
temperatures
and
at
lowest
weight.
The
deepest
shades
on
the
contour
map
represent
starvation.
Falling
below
the
minimum
maintenance
consumption
(
approximately
0.03
g
gbw
­
1d­
1
for
coho
juding
from
the
laboratory
trials)
for
extensive
periods
of
time
should
result
in
no
growth,
or
even
5­
10
weight
loss.
The
maximum
potential
consumption
is
relatively
moderate
for
most
of
the
temperatures
and
weights
that
juveniles
are
likely
to
encounter
in
the
freshwater
growth
phase
of
their
life
history
due
to
the
interaction
of
the
temperature
and
weight.
However,
consumption
is
severely
limited
only
at
high
and
low
temperatures
at
all
weights.
Consumption
approaches
the
maximum
potential
(
0.10
g
gbw
­
1d­
1
for
coho)
only
for
a
relatively
few
combinations
of
weight
and
temperature,
and
therefore
for
probably
relatively
little
time
during
the
life
of
a
fish.
Maximum
consumption
at
all
weights
is
achieved
at
optimal
temperature
(
Topt),
where
growth
rate
is
greatest
for
each
level
of
consumption
(
approximately
17oC
for
coho).

The
food
consumption
illustrated
in
Figure
5.3
represents
the
upper
maximum
controlled
by
the
physiology
of
the
fish
at
each
combination
of
weight
and
temperature.
We
refer
to
this
as
the
maximum
potential
consumption.
In
the
natural
stream
environment,
as
well
as
in
the
laboratory,
the
amount
of
food
available
to
consume
may
be
less
than
the
maximum
that
the
fish
can
potentially
consume
and
the
fish
may
experience
some
degree
of
hunger.
We
distinguish
consumption,
which
we
use
to
refer
to
the
physiological
response
controlling
food
intake,
from
ration,
which
we
use
to
refer
to
the
food
supply.
Ration,
which
we
express
as
%
satiation,
is
100%
satiation
at
the
maximum
consumption
(
Cs)
at
each
weight
and
temperature.

Consumption
in
natural
streams
depends
on
food
supply,
competition
for
food,
and
the
size
of
the
fish
based
on
the
past
regime
of
water
temperature.
The
Wisconsin
model
accounts
for
ecological
constraints
on
the
maximum
consumption
rate
(
Cs)
by
imposing
an
additional
proportionality
(
Pvalue
that
can
also
range
from
0
to
1
at
rations
less
than
the
physiological
maximum
(
ps).
Thus,
when
modeling
consumption
at
less
than
satiation
for
each
temperature
and
weight,
Ci
may
be
calculated
as
max
C
p
P
C
s
value
i
×
×
=
(
5.12)
2
4
6
8
10
12
14
16
18
20
22
24
26
Temperature
(
oC)
1
2
3
4
5
6
7
8
9
10
Weight
(
grams)
Coho
Food
Consum
ption,
P
roportion
of
Body
W
eight/
day
Figure
5.3
Illustration
of
the
relationship
of
the
maximum
food
consumption
at
satiation
with
temperature
and
coho
weight.
Contours
are
the
daily
consumption
rate,
Cs,
expressed
in
prey
mass
per
fish
body
mass
per
day
(
g
gbw
­
1
d­
1).
This
figure
illustrates
the
maximum
potential
consumption
where
food
supply
is
not
limited
and
fish
can
eat
to
satiation.
5­
11
It's
important
to
note
that,
on
any
given
day,
field
consumption
may
be
limited
by
either
the
physiological
limits
imposed
by
temperature
or
weight
or
by
the
food
supply.
Consumption
in
streams
must
be
determined
from
in
situ
observations
of
feeding,
or
inferred
from
weight
gain.

Specific
Growth
Rate.
The
specific
growth
rate
(
g)
is
the
daily
growth
rate
in
relation
to
temperature
and
consumption,
expressed
in
proportion
of
body
weight
per
day.
Specific
growth
rate
functions
("
growth
curves")
are
defined
with
data
from
the
laboratory
studies.
This
relationship
was
established
for
sockeye
by
Brett
(
1971),
and
for
chinook
by
Brett
et
al.
(
1982).
Both
Brett
(
1995
pp.
28­
29)
and
Weatherly
and
Gill
(
1995)
recently
reaffirmed
this
relationship,
some
form
of
which
appears
to
apply
to
all
species
of
salmonids.

Growth
rate
curves
for
coho
and
steelhead
have
not
been
previously
published,
although
the
requisite
laboratory
studies
were
available
to
develop
them.
Experiments
on
growth
of
juvenile
coho
reported
by
Everson
(
1973)
were
conducted
at
temperatures
between
11.1o
and
22.5oC
and
rations
between
satiation
and
starvation.
Experiments
on
steelhead
growth
reported
by
Wurtsbaugh
and
Davis
(
1977)
were
conducted
at
temperatures
between
6.9o
and
22.5oC
and
the
full
range
of
rations.
Growth
rate
of
the
population
during
the
experimental
period,
represented
by
the
average
population
weights,
was
calculated
as:

)
)
(
5
.
0
(
)
(

0
0
t
w
w
w
w
g
e
e
×
+
×
-
=
(
5.13)

where
we
and
w0
are
the
weights
at
the
end
and
beginning
of
the
experiment,
respectively,
and
t
is
the
number
of
days
(
generally
25)
in
each
trial.
Dry
weights
were
used
to
calculate
the
growth
rates
because
the
moisture
content
of
fish
is
similar
to
that
of
their
prey
in
natural
streams
(
Winberg,
1971)
and
thus
the
dry
weight
relationships
would
appropriately
match
growth
curves
of
fish
living
in
natural
environments
to
their
natural
food
supply.

We
used
standard
linear
regression
to
build
mathematical
expressions
for
the
growth
rate
g
(
g
gbw
­

1d­
1)
from
food
consumption,
C,
(
g
gbw
­
1d­
1),
temperature,
T,
(
oC),
and
initial
weight,
wo,
for
coho
and
steelhead.
To
reduce
multicollinearity
problems
during
the
model
building
process,
the
independent
variables
were
centered
by
subtracting
each
sample
value
from
the
population
mean.
The
models
were
built
with
the
Reg
procedure
in
SAS
Ò
Version
6.12
(
SAS
Institute
Inc.,
1989).
The
general
form
of
the
model
is:

W
T
C
C
C
T
T
g
×
+
×
×
+
×
+
×
+
×
+
×
+
=
6
5
2
4
3
2
2
1
0
c
c
c
c
c
c
c
(
5.14)

The
relationship
for
coho
includes
squared
and
linear
terms
for
temperature
and
consumption
and
has
high
R2
(
0.93)
and
low
root
mean
square
error
(
0.0023).
Similarly,
the
relationship
for
steelhead
includes
squared
and
linear
terms
for
temperature
and
consumption
as
well
as
a
linear
term
accounting
for
weight
of
the
fish.
This
relationship
also
has
high
R2
(
0.97),
and
low
root
mean
square
error
(
0.0021).
Statistically
determined
coefficients
for
these
relationships
are
provided
in
Table
5.3.
It
is
essential
to
note
that
the
growth
rate
equation
(
eq
5.14)
will
compute
erroneous
growth
rates
if
the
consumption
term
is
not
appropriately
constrained
for
each
weight
and
temperature
as
described
earlier.
Thus,
the
use
of
this
model
always
requires
a
two
step
process
where
consumption
level
is
estimated
and
then
growth
rate
is
selected.
5­
12
The
growth
curves
for
coho
and
steelhead
resulting
from
the
full
growth
model
formulation
calculated
at
a
weight
of
1
gram
are
shown
in
Figure
5.4.
They
show
the
familiar
form
of
sockeye
and
chinook
that
have
been
previously
published
by
Brett
and
others
(
Brett
1969,
Brett
et
al.
1982,
Weatherly
and
Gill
1995),
although
the
curves
in
Figure
5.4
are
derived
mathematically.
Brett
et
al.
(
1982)
discussed
how
optimum
temperature
declines
with
decreasing
ration,
thus
skewing
the
growth
rate
curves
towards
cooler
temperature
with
less
food.
The
mathematical
formulation
of
equation
5.14
also
produces
skew
in
optimal
temperature
with
declining
consumption.
The
crossproduct
(
c
5)
between
temperature
and
consumption
determines
this
shape.

The
growth
curves
for
each
species
are
similar,
but
differ
in
details
such
as
optimal
temperatures,
growth
rate,
and
consumption
levels
at
which
growth
rates
are
achieved.
One
can
see
that
the
growth
rate
varies
widely
with
the
temperatures
and
food
availability
that
salmon
are
likely
to
encounter
in
the
natural
environment
of
Pacific
Northwest
streams
and
rivers,
suggesting
that
fish
size
should
be
strongly
influenced
by
these
two
parameters.

In
natural
streams,
temperature
varies
over
the
course
of
the
day
and
some
temperature
value
must
be
chosen
to
represent
the
daily
temperature.
Laboratory
tests
have
noted
that
the
daily
average
temperature
approximates
the
constant
exposure
test
conditions.
Experiments
where
temperatures
were
fluctuated
to
mimic
daily
temperature
regimes
have
found
either
no
effect
from
variable
temperature
(
Thomas
et
al.
1986,
Dickerson
and
Vinyard
1999)
or
an
improvement
in
growth
(
Spigarelli
et
al.
1982,
Weatherly
and
Gill
1995).
Water
temperature
fluctuated
with
natural
stream
temperatures
in
the
experiments
of
Everson
(
1973)
and
Wurtsbaugh
and
Davis
(
1977).
We
select
the
daily
mean
temperature
as
appropriate
to
represent
the
temperature
related
to
daily
growth
rate.

Table
5.2
Coefficients
for
specific
growth
rate
relationship
to
temperature
(
T)
and
consumption
(
C).
Coho
data
are
from
Everson
(
1973),
Steelhead
data
are
from
Wurtsbaugh
and
Davis
(
1977).

Variable
Coho
Steelhead
c
0
­
0.010649
0.00631
c
1
0.00096624
­
0.0007403
c
2
­
0.00008312
­
0.00003909
c
3
0.450620
0.4302104
c
4
­
3.02056
­
1.43765
c
5
0.01677
0.00735
c
6
NA
­
0.00517
c
7
NA
NA
Regression
R­
square
0.93
0.97
5­
13
Figure
5.4.
Specific
growth
rate
curves
for
coho
salmon,
steelhead
trout
at
1
gram
weight.
Coho
growth
curves
based
on
Everson
(
1973),
steelhead
curves
based
on
Wurtsbaugh
and
Davis
(
1977).
Each
line
is
the
ration
is
expressed
in
%
satiation,
ps.

Coho
Salmon
­
0.010
­
0.005
0.000
0.005
0.010
0.015
0.020
0.025
0.030
0
10
20
30
Tempe
rature
(
oC)
Grow
th
Rate
(
g
g
­
1d­
1
)
100%

80%

60%

40%

30%
Steelhead
­
0.010
­
0.005
0.000
0.005
0.010
0.015
0.020
0.025
0.030
0.035
0
10
20
30
Temperature
(
oC)
Growth
Rate
(
g
g­
1d­
1)
100
80%

60%

40%

30%
5­
14
Growth
Model
Substituting
equations
5.13
and
5.11
back
into
equation
5.3
gives
a
model
for
computing
weight
as
a
function
of
initial
weight
and
the
time
series
of
daily
water
temperatures
and
food
supply
to
which
a
fish
is
exposed:

)
1
(

1
6
5
2
4
1
0
3
2
2
1
0
0
-
-
=
×
+
×
×
+
×
+
×
+
×
+
×
+
+
×
=
Õ
i
i
i
i
t
i
i
i
t
w
T
C
C
C
T
T
w
w
c
c
c
c
c
c
c
(
5.15)

The
model
simulates
the
change
in
mass
of
an
individual
fish
or
a
cohort
of
specified
size.
We
define
a
cohort
as
a
group
of
similar
aged
fish
of
the
same
species
experiencing
identical
environmental
conditions
(
Hanson
et
al.
1997).
Cohort
weights
may
be
represented
by
the
average
population
weight.
It
is
important
to
recognize
that
the
estimates
of
weight
gain
do
not
consider
population
interaction
effects
(
Walters
and
Post
1993).
Thus,
they
do
not
account
for
changes
in
population
density
that
can
also
affect
the
average
weight
and
biomass
of
populations.
Measuring
growth
as
the
difference
in
a
cohort's
weight
between
two
dates
is
subject
to
biases
from
any
size­
dependent
movement
and
mortality
(
Railsback
and
Rose
1999).
Our
estimates
only
address
temperature
and
food
effects
with
the
assumption
that
changes
in
number
or
weight
are
unbiased
by
size.

The
growth
formulations
(
equations
5.7­
5.12)
were
easily
programmed
into
an
EXCEL
Ò
spreadsheet
to
perform
the
calculations
of
growth
over
time.
To
estimate
weight
gain
of
populations
in
natural
streams,
three
parameters
must
be
determined:
initial
weight
(
w0),
temperature
(
T)
and
the
food
consumption
(
Ci)
that
must
reflect
the
food
supply.

APPLICATION
OF
THE
METHOD
FOR
PREDICTING
GROWTH
IN
NATURAL
STREAMS
In
the
remainder
of
this
section,
we
apply
the
growth
model
to
fish
populations
observed
in
natural
streams
to
demonstrate
model
behavior
and
corroborate
its
predictions.
We
show
a
number
of
simulations
where
the
three
input
parameters
were
known
with
varying
degrees
of
certainty.
No
new
experiments
to
determine
food
consumption
or
to
validate
the
growth
model's
ability
to
accurately
predict
weight
gain
were
conducted.
Instead,
model
performance
is
examined
using
data
from
a
number
of
previously
reported
studies.

The
parameters
required
by
the
model
are
rather
modest.
Daily
mean
temperature
is
known
with
certainty
at
any
site
where
temperature
is
continuously
recorded,
and
we
restrict
growth
simulations
to
sites
where
this
condition
was
met.
Initial
weight
and
food
consumption
must
be
determined
from
fish
population
characteristics
observed
in
streams
or
from
known
food
consumption
amounts.
Of
these
two
variables,
the
consumption
characteristics
of
juvenile
salmonids
in
natural
streams
are
by
far
the
least
well
quantified
(
Filbert
and
Hawkins
1995,
Railsback
and
Rose
1999)
and
difficult
to
obtain
(
e.
g.,
Martin
1985).
For
this
reason,
we
begin
simulations
with
data
from
a
field
experiment
where
food
was
well
known
and
the
assumptions
about
the
3
input
parameters
were
limited.

Mason's
Feeding
Experiment.
Mason
(
1976)
reported
an
experiment
where
juvenile
coho
in
a
natural
stream
on
Vancouver
Island,
British
Columbia
were
fed
to
satiation
for
a
2­
month
summer
period.
Data
from
this
experiment
were
ideal
for
comparing
observed
with
predicted
growth
since
environmental
limitations
on
food
supply
were
reduced,
if
not
5
-
15
eliminated,
for
most
of
the
individuals
in
the
population.
Thus
errors
associated
with
estimating
consumption
were
minimized
and
the
other
two
input
parameters
were
well
known.

Mason
(
1976)
reported
that
temperatures
were
between
12o
and
13oC
during
the
study.
We
used
a
constant
12.5oC
temperature
for
the
simulation.
Mason
(
1976)
provided
information
on
individual
fish
size
within
the
population.
We
modeled
three
cases
where
the
initial
weight
was
set
to
the
weight
of
the
largest,
average,
and
smallest
fish
at
the
start
of
the
experiment.
(
Figure
5.5).
Fish
were
feed
daily
at
what
Mason
calculated
was
a
satiation
ration
for
the
population.
For
this
simulation,
growth
was
predicted
at
the
maximum
potential
consumption
determined
by
the
weight
and
temperature
(
eq.
5.11),

assuming
that
the
available
rations
supplied
100%
of
the
potential
consumption
and
there
was
no
food
limitation.
We
also
modeled
the
full
range
of
rations
(
ps)
expressed
as
percentage
of
satiation
of
shown
in
increments
of
80,
60,
40,
and
30%
in
Figure
5.5.

After
simulating
growth
for
the
two­
month
period,
the
estimated
weights
of
the
largest
and
average
fish,
computed
at
high
food
ration,
were
very
similar
to
their
observed
weights
(
Figure
5.5).
The
largest
fish
in
the
population
increased
its
weight
by
286%
(
2
g
to
5.9
g)
during
this
period.
The
predicted
weight
for
that
fish,
assuming
100%
satiation,
was
within
2%
of
its
observed
weight.
These
results
appear
to
corroborate
growth
predictions
and
confirm
Mason's
conclusion
that
some
of
the
fish
obtained
satiation
rations.
Progressively
smaller
individuals
within
the
population
apparently
ate
at
lesser
rations
than
the
largest
individual.
(
It
should
be
noted
that
it
is
not
known
whether
the
fish
that
held
these
ranks
at
the
end
of
the
study
were
the
same
individuals
as
those
at
the
start.)
The
weight
of
the
average
size
fish
was
consistent
with
growth
predictions
at
80%
ration.
The
smallest
fish
did
not
fare
nearly
as
well,
achieving
little
weight
gain
consistent
with
Coho
Growth
Mason,
1976
0
1
2
3
4
5
6
7
Initial
Observed
Observed
Smallest
Predicted
Smallest
Observed
Average
Predicted
Average
Observed
Largest
Predicted
Largest
Weight
(
grams)

Ration
100%

Ration
80%

Ration
60%

Ration
40%

Ration
30%
July
28
Sept
28
Largest
Coho,
Mason
(
1976)

0
1
2
3
4
5
6
7
Figure
5.5.
Observed
and
predicted
growth
of
coho
in
an
experimental
feeding
study
in
a
natural
channel
in
British
Columbia
(
Mason
1976).
The
growth
of
the
smallest,
average
and
largest
fish
at
the
start
of
the
feeding
trial
were
modeled
for
the
full
range
of
satiation
(
shown
in
bars).
Weight
of
the
smallest,
average
and
largest
fish
at
the
end
of
the
experiment
are
shown
relative
to
growth
simulations.
The
predicted
daily
weight
of
the
largest
size
fish
is
shown
at
right.
5
-
16
subsistence
at
minimum
maintenance
rations.
Mason
(
1976)
observed
that
the
smaller
fish
tended
to
be
those
dying
or
migrating
downstream,
emphasizing
the
importance
of
larger
size
providing
survival
benefit
through
more
effective
feeding
strategies
(
Puckett
and
Dill
1985,
Nielsen
1994).

Overall,
Mason
(
1976)
observed
many
large
fish
remaining
after
the
feeding
experiment;
moreover,
the
average
population
weight
increased
significantly.
This
has
been
taken
as
evidence
that
coho
growth
in
streams
tends
to
be
limited
by
food
supply.
Other
field
experiments
have
showed
improved
growth
resulting
from
activities
that
increased
light
(
and
temperature)
in
streams,
presumably
improving
the
food
resources
(
Hawkins
et
al.
1983,
Martin
1985,
Bisson
et
al.
1988).

Fish
Populations
Living
in
Natural
Stream
Conditions.
As
they
grow
during
the
summer,
juvenile
fish
living
in
streams
in
the
Pacific
Northwest
experience
temperatures
that
are
relatively
low
in
spring
after
emergence
from
the
gravel,
warm
during
the
summer,
and
cool
again
in
the
fall
after
mid­
September.
Food
resources
may
be
limited
and
the
manner
in
which
natural
populations
regulate
their
numbers
to
match
individual
and
population
growth
needs
with
available
food
supply
is
complex
and
largely
unknown
(
Walters
and
Post
1993).

In
the
next
set
of
simulations
we
estimate
the
average
growth
of
a
cohort
of
age
0
coho
and
steelhead
in
natural
streams
having
natural
and
unknown
food
supplies.
Fish
population
data
were
available
from
sites
representing
a
number
of
treatments
and
controls
in
experiments
previously
reported
in
the
literature
(
Table
5.3).
Physical
habitat
characteristics
varied
among
sites.
Like
many
streams
within
the
region,
the
stream
segments
where
fish
were
sampled
generally
contained
low
amounts
of
large
woody
debris
(
LWD)
and
varying
amounts
of
shade.
The
examples
include
sites
with
water
temperature
spanning
very
warm
to
cool
(
but
not
cold,
e.
g.,
less
than
12oC).
Porter
Creek
was
the
site
of
a
habitat
improvement
experiment
where
LWD
was
added
to
increase
pool
habitat
(
Cederholm
et
al.
1997).
Several
sites
were
severely
impacted
by
the
Mount
St.
Helens
eruption
in
1980,
but
have
since
experienced
recovery
of
vegetation
(
Bisson
et.
al
1988).
A
dam­
break
flood
scoured
portions
of
Huckleberry
Creek
and
its
floodplain
in
1990
after
several
years
of
population
monitoring
during
a
feeding
experiment
(
Fransen
et
al.
1993).
Bear
Creek
is
a
tributary
to
the
Bogachiel
River,
above
a
barrier
falls,
that
flowed
through
undisturbed
old
growth
forest
(
Martin
1985).

Although
the
experimental
objectives
and
study
designs
differed,
fish
populations
in
these
stream
segments
were
sampled
routinely
in
a
similar
way
providing
comparable
data
sets
to
draw
from.
Populations
were
sampled
by
electrofishing
using
the
three­
pass
removal
method
(
Young
and
Robson
1978,
Bisson
et
al.
1988).
The
lengths
of
all
captured
fish
were
measured,
while
weights
were
sub­
sampled.
We
selected
a
number
of
cases
from
the
available
data
where
fish
populations
were
sampled
at
the
beginning
and
end
of
the
summer.
A
few
sites
were
sampled
mid­
season.
Early
season
sampling
was
conducted
between
March
and
July
while
end
of
season
sampling
was
conducted
between
September
and
November.
Coho
occurred
at
all
six
sites
and
steelhead
were
found
at
four.
The
time
periods
encompassed
by
the
temperature
and
fish
population
data
did
not
always
overlap
at
all
sites,
and,
in
a
few
cases,
daily
temperatures
were
not
available.
5
-
17
Table
5.3.
Characteristics
of
fish
populations
as
determined
by
field
surveys.

Data
Sources
Temperature
Characteristics
Fish
Population
Surveys
Model
Predictions
Species
Site
Experiment
Reference
Year
Regime
Season
median
(
oC)
Annual
maximum
(
oC)
Dates
sampled
Days
(
t)
Initial
weight
(
g)
End
weight
(
g)
Increase
in
weight
(%)
Ave.
Daily
Growth
Rate
(
g
g­
1d­
1)
Estimated
weight
(
g)
Difference
from
observed
(%)

Coho
Huckleberry
Artificial
feeding
Fransen
et
al.
1993
1987
Cool
12.5
15.5
7/
13­
10/
1
80
3.0
3.6
16
.0064
5.2
+
44
1988
Cool
11.6
15.0
5/
12­
10/
24
163
0.6
4.2
600
.0114
3.8
­
10
1989
Cool
12.0
14.5
7/
16­
7/
28
101
2.0
4.1
105
.0063
3.8
­
7
Dam­
break
flood
1991
Warm
15.5
18.5
5/
15­
10/
9
147
0.9
3.2
240
.0082
3.1
­
3
altered
valley
1991
Warm
15.5
18.5
6/
20­
10/
9
113
4.1
6.3
54
.0082
5.8
­
8
1998
Warm
14.0
18.0
6/
23­
9/
15
84
2.7
4.0
57
.0072
5.0
+
16
Porter
Cederholm
et
al.
1997
1988
Warm
13.5
18.0
6/
28­
9/
28
92
3.0
6.1
103
.0069
5.7
­
7
Placement
of
large
woody
debris
1989
Warm
14.0
17.0
6/
19­
9/
18
91
3.4
6.0
76
.0069
6.4
+
7
1990
Warm
14.6
18.6
6/
18­
8/
27
70
3.7
6.5
76
.0085
6.6
+
2
1991
Warm
12.8
17.3
6/
4­
9/
30
118
2.5
5.4
116
.0080
6.4
+
19
Harrington
Volcanic
eruption
Bisson
et
al.
1988
1984
f
Very
warm
16.5
d
29.0
6/
24­
10/
9
108
1.3
6.3
380
NA
NA
NA
Vegetation
regrowth
1990
Very
warm
13.3
20.5
6/
22­
9/
20
88
3.3
6.0
82
.0066
5.8
­
2
Hoffstadt
Volcanic
eruption
Bisson
et
al.
1988
1984
f
Very
warm
16.7
d
29.5
6/
26­
10/
3
100
2.4
5.6
133
NA
NA
NA
Vegetation
regrowth
1990
Very
warm
15.0
26.0
6/
21­
9/
20
88
3.6
6.1
69
.0054
5.9
­
5
Big
Salmon
carcass
enhancement
Bilby
et
al.
1996,
1998
1994
Cool
12.9
16.1
7/
13­
9/
8
57
4.6
6.0
30
.0093
6.9
+
15
Salmon
Salmon
carcass
enhancement
Bilby
et
al.
1996,
1998
1994
Cool
12.6
16.2
7/
20­
9/
8
50
3.6
5.5
53
.0069
5
­
9
Steelhead
Porter
Cederholm
et
al.
1997
1988
Warm
13.5
18.0
6/
28­
9/
28
92
0.8
3.5
338
.0173
3.9
+
11
Placement
of
large
woody
debris
1989
Warm
14.0
17.0
6/
19­
9/
18
91
0.6
3.3
450
.0210
3.6
+
9
1990
Warm
14.6
18.6
6/
18­
8/
27
70
0.6
3.2
433
.0235
3.1
­
3
1991
Warm
12.8
17.3
6/
4­
9/
30
118
0.4
3.4
750
.0191
3.8
+
12
Harrington
Volcanic
eruption
Bisson
et
al.
1988
1984
Very
warm
16.5
d
29.0
6/
24­
10/
3
102
0.7
2.9
.0228
NA
Vegetation
regrowth
1990
Very
warm
13.3
20.5
6/
22­
9/
14
84
0.5
3.7
640
.0076
4.2
+
14
Hoffstadt
Vegetation
regrowth
Bisson
et
al.
1988
1990
Very
warm
15.0
26.0
6/
21­
9/
20
91
2.1
5.5
162
.0180
4.2
­
24
Salmon
Salmon
carcass
enhancement
Bilby
et
al.
1996,
1998
1994
Cool
12.6
16.2
7/
20­
9/
8
50
1.4
3.2
129
3.4
+
6
Cutthroat
Bear
Control
for
riparian
buffer
experiment
Martin,
1985
1978
Cool
12.2
q
14
d
7/
6­
11/
11
128
2.2
3.4
54
NA
NA
f
Continuously
recorded
temperature
data
not
available;
population
weights
not
simulated
with
model.

d
Estimated
from
annual
maximum
temperature
by
relationship
found
in
Section
3.
q
Through
September
5­
18
Consumption
Rates
in
Natural
Streams.
Ideally,
consumption
rates
in
natural
populations
are
determined
by
observation
of
food
intake
or
availability
(
e.
g.,
Martin
1985,
Filbert
and
Hawkins,
1995).
Such
data
are
rare.
In
lieu
of
direct
measurement,
consumption
can
be
approximated
by
examining
the
pattern
of
growth
over
the
growing
period
(
Beauchamp
et
al.
1989,
Hanson
et
al.
1997).
Consumption
may
be
inferred
by
estimating
how
much
food
must
have
been
consumed,
given
the
prevailing
temperature
regime,
to
have
maintained
growth
at
the
observed
rate
(
Hanson
et
al.
1997).
To
estimate
consumption
indirectly,
populations
should
be
sampled
several
times
over
the
summer
duration,
because
consumption
rates
should
decline
through
time
as
fish
grow,
based
on
the
expected
allometric
effect.
Such
data
are
also
rare.

A
naturally­
spawned
coho
population
was
sampled
multiple
times
during
the
summer
of
1991
at
Huckleberry
Creek.
We
selected
1991
data
because
fish
were
sampled
a
number
of
times
over
a
long
period
from
March
to
October,
and
we
focused
on
the
lower
of
two
sites
on
the
same
stream
because
the
population
was
naturally
spawned
fish
with
no
augmentation.
Coastal
cutthroat
trout
were
sampled
multiple
times
at
Bear
Creek
in
1978
(
Martin
1985).
We
assume
that
the
cutthroat
trout
sampled
at
this
site
are
representative
of
steelhead
trout,
for
the
purposes
of
establishing
feeding
patterns
only.
We
use
these
two
sites
to
evaluate
food
consumption
patterns
and
to
establish
the
relationships
used
to
estimate
ration
for
growth
simulation
of
the
other
populations
in
our
study.

The
consumption
of
food
by
cutthroat
trout
at
Bear
Creek
was
measured
using
stomach
evacuation
techniques
(
Martin
1985).
In
situ
measurements
of
food
consumption
were
computed
by
the
modified
Bajkov
method
(
1935).

At
Huckleberry
Creek,
neither
food
consumption
nor
food
supply
was
measured.
Apparent
food
consumption
by
coho
between
each
sampling
was
derived
using
the
growth
rate
curve.
First,
growth
rate
was
computed
for
the
full
range
of
consumption
from
maintenance
to
satiation
at
the
mean
temperature
and
population
weight
for
each
interval.
Consumption
was
then
inferred
by
selecting
the
value
from
the
computed
Weight,
Growth,
and
Food
Consumption
for
Age
0
Coho
0
0.02
0.04
0.06
0.08
3/
15
5/
4
6/
23
8/
12
10/
1
Date
(
1991)
Rate
(
g
g­
1d­
1)

0
1
2
3
4
Mean
Weight
(
g)
Growth
rate
Consump
rate
Mean
Wt
Weight,
Growth
and
Food
Consumption
for
Age
0
Cutthroat
0
0.04
0.08
0.12
0.16
5/
20
7/
9
8/
28
10/
17
12/
6
Date
(
1978)
Rate
(
g
g
­
1d­
1)

0
0.5
1
1.5
2
2.5
3
3.5
4
Mean
Weight
(
g)
Growth
rate
Consump
rate
Mean
Wt
Figure
5.6.
Mean
weight
and
estimated
consumption
and
growth
rates
for
age
0
coho
at
Huckleberry
Cr.
,
lower
site,
and
age
0
cutthroat
trout
at
Bear
Creek.
5
-
19
range
that
corresponded
most
closely
to
the
observed
growth
rate.
Important
assumptions
accompany
this
analysis;
only
food
and
temperature
influence
the
observed
data,
our
growth
equations
capture
these
factors
correctly,
there
are
no
effects
from
population
dynamics
and
there
are
no
sampling
errors.
Therefore,
this
approach
is
only
an
approximation
and
it
is
preferable
to
establish
consumption
estimates
by
direct
measurement
of
food
supply
or
intake,
such
as
was
done
for
cutthroat.

Average
population
weight
and
estimated
(
coho)
and
measured
(
cutthroat)
growth
and
consumption
rates
for
the
two
populations
are
shown
in
Figure
5.6.
Growth
rate
was
calculated
using
equation
5.13
assuming
a
linear
growth
rate
between
each
sampling
interval.
Weight
gain
slowed
late
in
the
season
and
growth
rate
and
consumption
declined
from
initially
higher
rates
for
both
species
(
Figure
5.6).
This
pattern
is
consistent
with
what
would
be
expected
for
an
allometric
consumption
relationship,
where
consumption
declines
with
increasing
weight.
However,
it
can
also
suggest
diminishing
food
supply
that
may
accompany
increasing
size
in
an
environment
where
food
supply
is
fixed.

The
pattern
of
consumption
observed
at
the
sites
is
of
particular
interest
with
regard
to
its
implications,
vis­
à­
vis
the
physiological
limits
of
temperature
and
weight
versus
food
supply.
Maximum
potential
consumption
(
Cs)
was
computed
using
the
mean
temperature
and
weight
for
each
interval.
It
is
plotted
with
apparent
consumption
in
Figure
5.7.
The
apparent
consumption
was
greatest
during
the
early
half
of
the
summer
and
declined
through
the
later
half
for
both
species.
(
Note
that
the
time
intervals
differ
for
the
two
sites).

An
important
implication
of
the
observations
in
Figure
5.7
is
that
the
observed
consumption
of
both
species
was
remarkably
similar
to
each
species's
physiologicallydetermined
potential
consumption.
This
was
true
for
coho
from
soon
after
emergence
from
the
gravels
until
the
end
of
June
or
early
July
and
into
September
for
cutthroat
trout.
Cutthroat
trout
actually
showed
greater
consumption
than
estimated
during
this
period,
but
values
were
probably
within
the
error
range
due
to
estimates
of
Consumption
for
Age
0
Coho
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
15­
Mar
14­
Apr
14­
May
13­
Jun
13­
Jul
12­
Aug
11­
Sep
11­
Oct
Consumption
(
g
g­
1d­
1)
Observed
Maximum
Potential
Consumption
for
Age
0
Cutthroat
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
13­
Jun
13­
Jul
12­
Aug
11­
Sep
11­
Oct
10­
Nov
10­
Dec
Consumption
(
g
g­
1d­
1)
Observed
Maximum
Potential
Figure
5.7.
Maximum
potential
consumption
based
on
temperature
and
weight
for
the
period
in
relation
to
observed
consumption.
5
-
20
average
temperature
and
because
the
maximum
potential
is
based
on
steelhead
specific
growth
curves.
During
these
periods,
the
data
suggest
that
the
physiological
factors
were
controlling
consumption
and
the
average
size
fish
was
eating
at
satiation.
That
is,
even
if
there
were
more
food,
the
individual
fish
represented
by
the
average
population
weight
would
not
eat
it.
Later
in
summer,
the
apparent
consumption
was
less
than
maximum
potential
for
both
species.
The
difference
between
the
apparent
and
maximum
potential
consumption
estimates
suggests
the
degree
of
food
limitation
in
the
streams.
The
closer
the
observed
lines
are
to
potential,
the
closer
available
ration
is
to
satiation.

It
is
interesting
to
note
that
it
is
during
the
early
summper
period
that
the
number
of
individuals
in
a
population
following
emergence
is
determined,
suggesting
that
population
density
adjusts
to
match
the
food
supply
(
Chapman
1966).
Carl
(
1983)
observed
that
the
number
of
coho
in
a
population
adjusted
in
response
to
population
density,
but
that
their
daily
growth
rates
were
not
dependent
on
density
during
the
period
when
population
adjustment
after
emergence
occurred
(
May
to
July).
Similarly,
rainbow
trout
growth
rate
was
not
correlated
with
their
population
density.
Carl's
(
1983)
observations
of
population
dynamics
and
growth
are
consistent
with
results
from
our
analysis
of
consumption.

For
coho,
the
difference
between
apparent
and
potential
consumption
was
much
larger
than
for
cutthroat
in
the
later
half
of
the
summer.
Apparent
consumption
declined
to
near
starvation
ration
by
the
end
of
the
summer,
while
estimated
potential
consumption
remained
relatively
high
because
both
temperature
and
weight
were
moderate.
Potential
consumption
actually
increased
during
part
of
this
period
reflecting
favorable
(
closer
to
optimum)
temperatures.
Declining
food
supply
for
coho
during
this
period
could
be
explained
by
increasing
size
of
individuals
in
the
population
and
limited
food
supply.
Although
the
effects
of
size
on
consumption
are
not
great,
the
absolute
volume
consumed
by
each
fish
must
increase
since
the
consumption
is
expressed
as
a
proportion
of
body
weight
each
day
(
Hanson
1997).
The
same
food
supplies
that
may
have
been
adequate
for
the
population
at
small
size
may
represent
a
much
smaller
proportion
as
the
fish
grow.
The
decline
could
also
reflect
a
change
in
the
food
supply.
Studies
generally
show
fairly
steady
or
increased
food
availability
during
the
summer,
depending
on
habitat
conditions
(
e.
g.,
under
forested
vs.
open
riparian
canopies,
Hetrick
et
al.
1998).

Observed
and
maximum
potential
consumption
were
similar
during
the
entire
period
for
cutthroat
trout
(
Figure
5.7).
During
the
time
period
coincident
with
coho,
cutthroat
were
apparently
consuming
at
or
near
maximum
potential
(
satiation)
and
food
supply
did
not
appear
to
be
a
major
factor
limiting
the
food
consumption
by
this
species.
The
decline
in
potential
consumption
in
the
fall
can
largely
be
attributed
to
lower
temperature.
One
reason
that
consumption
is
nearer
the
maximum
for
much
of
the
summer
period
may
be
that
the
cutthroat
fry
emerge
later
and
are
smaller
than
the
coho
in
late
summer.
Thus,
satisfying
juvenile
food
demands
is
less
challenging.

The
apparent
consumption
patterns
observed
in
populations
inhabiting
natural
streams
have
important
implications
for
the
growth
simulations.
When
consumption
can
be
assumed
to
equal
maximum
potential
(
eq
5.11)
there
is
no
need
to
calibrate
the
consumption
estimates
to
account
for
the
ecological
constraints
on
food
supply
(
eq
5.12).
This
appears
to
be
a
reasonable
assumption
for
age
0
cutthroat
throughout
the
summer.
We
assume
that
age
0
steelhead
have
similar
feeding
patterns
as
cutthroat,
and
therefore,
we
assume
that
we
can
use
the
maximum
potential
consumption
to
represent
age
0
steelhead.
This
also
appears
to
be
a
reasonable
assumption
for
coho
until
the
end
of
June.
5
-
21
After
June,
apparent
consumption
by
coho
declined
significantly
relative
to
maximum
potential
and
the
consumption
estimate
must
be
modified
by
a
factor
reflecting
food
limitation
(
P­
value
in
the
Wisconsin
model
terminology).
To
do
so
for
coho,
we
fit
a
linear
regression
to
the
apparent
coho
consumption
from
July
to
October
(
Figure
5.7)
Maximum
Potential
Consumption,
Cs
0.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16
23­
May
2­
Jun
12­
Jun
22­
Jun
2­
Jul
12­
Jul
22­
Jul
1­
Aug
11­
Aug
21­
Aug
31­
Aug
10­
Sep
20­
Sep
30­
Sep
Consumption
(
g
g­
1d­
1)
Steelhead
Coho
P­
value
adjustment
to
coho
A)

Relative
Consumption,
ps
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0.90
1.00
23­
May
2­
Jun
12­
Jun
22­
Jun
2­
Jul
12­
Jul
22­
Jul
1­
Aug
11­
Aug
21­
Aug
31­
Aug
10­
Sep
20­
Sep
30­
Sep
Relative
Consumption,
P
Coho
Steelhead
B)

Figure
5.8
Estimates
of
consumption
at
Huckleberry
Creek.
Data
points
are
estimates
of
maximum
potential
consumption
based
on
temperature
and
weight:
A)
Consumption
as
a
proportion
of
body
mass;
B)
Relative
consumption,
calculated
as
daily
consumption
divided
by
Cmax
generated
by
the
consumption
equations.
The
line
is
the
P­
value
for
coho
that
adjusts
consumption
based
on
field
observations
suggesting
ecological
constraints
on
food
supply.
5
-
22
using
a
dated
time
step.
Maximum
consumption
values
and
the
ecological
adjusted
relationship
are
shown
in
Figure
5.8A.
For
readers
familiar
with
the
Wisconsin
model,
the
relative
consumption
for
each
day,
(
pi),
is
also
shown
in
Figure
5.8B.
During
growth
simulation,
each
consumption
value
is
calculated
in
the
defined
time
frame
(
site
adjusted
P­
Value
and
maximum
potential)
and
the
value
that
is
lower
is
selected.
(
During
the
spring
and
early
part
of
the
summer,
the
P­
value
is
not
calculated
since
the
continuation
of
the
calculation
outside
the
time
interval
on
which
it
is
based
may
estimate
unrealistic
values,
although
it
does
appear
to
fit
the
data
during
this
interval
as
well.)
Ideally,
population
weight
data
would
be
available
at
each
site
where
the
growth
simulations
are
conducted,
at
least
for
coho.
However,
this
information
was
not
available
at
other
sites
(
Table
5.4)
since
it
is
common
to
measure
populations
only
at
the
beginning
and
end
of
the
summer.
Therefore,
all
population
growth
simulations
for
coho
and
steelhead
that
follow
use
the
consumption
estimates
developed
at
Lower
Huckleberry
Creek,
depicted
in
Figure
5.8A,
to
estimate
consumption.

Comparison
of
Model
Predictions
to
Observed
Fish
Growth
The
apparent
consumption
estimates
were
used
to
simulate
growth
of
the
fish
populations
listed
in
Table
5.3,
using
initial
weights
and
temperature
recorded
over
the
period
between
the
initial
and
final
sampling
of
populations.
We
first
simulated
coho
population
growth
at
the
two
sites
at
Huckleberry
Creek
where
the
coho
consumption
estimates
were
developed
(
Figure
5.8).
Presumably,
errors
would
be
less
here
than
at
sites
where
consumption
was
not
calibrated
site­
specifically.
The
simulation
results
and
observed
population
weights
at
the
two
sites
are
shown
in
Figure
5.9
Between
the
first
and
last
sampling
time,
the
naturallyspawned
coho
at
the
lower
site
increased
in
weight
more
than
threefold
from
0.9
to
3.2
grams.
The
predicted
weight
was
within
3%
(
0.1
g)
of
the
observed
weight
at
the
end
of
the
summer.
The
predicted
and
observed
values
are
so
close
that
the
simulation
line
appears
to
connect
the
observations.
The
model
correctly
predicted
the
ending
population
weight
and
those
observed
at
each
mid­
season
sampling.
This
is
perhaps
not
a
surprising
result
given
that
the
food
consumption
estimates
(
Figure
5.8)
were
developed
at
this
site.
However,
the
simulation
shows
that,
when
food
intake
is
calibrated
at
a
given
location,
the
model
can
accurately
predict
weight
gain.
Huckleberry
Creek,
Coho
0
1
2
3
4
5
6
7
28­
May
11­
Jun
25­
Jun
9­
Jul
23­
Jul
6­
Aug
20­
Aug
3­
Sep
17­
Sep
1­
Oct
Weight
(
g)
Low
er
Predicted
Low
er
Observed
Upper
Predicted
Upper
Observed
Figure
5.9
Observed
and
simulated
weight
gain
at
the
lower
and
upper
sites
at
Huckleberry
Creek.
5
-
23
Importantly,
the
model
simulations
are
sensitive
to
the
consumption
estimates.
For
example,
increasing
daily
consumption
by
10,
20
and
30%
above
the
values
shown
in
Figure
5.8
through
the
simulation
period
yields
final
weights
of
coho
at
the
lower
Huckleberry
site
of
4.0,
4.9,
and
6.0
grams
respectively
(+
25,
+
53
and
+
88%
greater
than
observed).
Thus,
errors
have
a
cumulative
effect
since
they
are
applied
over
time.
Interestingly,
this
relatively
moderate
increase
in
consumption
results
in
increased
population
weights
that
are
proportionately
consistent
with
those
observed
in
field
studies
that
found
improved
production
in
streams
where
shade
had
been
removed,
and
presumably,
food
had
increased.
(
Hawkins
et
al.
1983,
Hetrick
et
al.
1998).
These
results
suggest
that
food
supply
doesn't
have
to
increase
by
a
large
amount
to
significantly
affect
the
weight
that
fish
achieve
over
the
course
of
the
summer.

Predicted
weight
at
the
upper
Huckleberry
site
was
also
close
to
that
observed,
but
predictions
were
not
as
accurate
(­
8%)
(
Figure
5.9).
The
upper
site
had
a
large
influx
of
larger
fish
between
the
May
and
June
sampling
dates,
so
the
simulation
was
initiated
in
June
after
the
population
was
determined.
The
upper
site
had
a
slightly
different
P­
value
at
the
end
of
the
summer,
and
the
value
was
used
to
model
this
site
only
(
not
shown).
Growth
of
this
population
was
low
during
the
latter
part
of
the
summer
and
the
rate
of
decline
of
the
P­
value
was
greater
at
this
site
than
the
lower
site,
possibly
because
of
the
larger
size
of
fish
in
the
population.
Nevertheless,
despite
nearing
minimum
maintenance
ration
at
the
end
of
the
season,
the
large
size
of
the
population
was
maintained
into
the
fall
and
no
weight
loss
was
observed.
The
model
prediction
was
actually
more
accurate
at
the
end
of
the
season
than
at
some
of
the
mid­
points,
where
it
tended
to
underestimate
the
weight.

In
the
next
set
of
simulations,
we
applied
the
model
to
the
rest
of
the
populations
listed
in
Table
5.3,
except
those
where
continuously
recorded
temperature
data
were
not
available.
The
consumption
estimates
established
for
the
population
at
the
lower
site
at
Huckleberry
Creek
in
1991
were
used
for
all
of
the
coho
simulations.
We
would
expect
that
applying
these
assumptions
to
other
sites
could
be
a
significant
source
of
error,
since
presumably
food
supply
differs
among
sites
and
the
growth
estimates
are
sensitive
to
this
parameter.
Furthermore,
errors
accumulate
due
to
the
effect
of
weight
on
consumption.
We
show
the
simulation
for
coho
and
steelhead
at
Porter
Creek
(
1988)
in
Figure
5.10.
Porter
Creek
is
exemplified
because
the
prediction
accuracy
was
average
for
both
coho
and
steelhead
(
Table
5.3),
the
site
was
far
removed
from
Huckleberry
Creek,
the
simulation
period
was
long,
and
both
species
were
present.
Predicted
weights
closely
approximated
observed
weights
for
both
species
(
Coho:
­
7%,
Steelhead:+
11%).
Porter
Creek,
1988
0
1
2
3
4
5
6
7
24­
Jun
2­
Jul
10­
Jul
18­
Jul
26­
Jul
3­
Aug
11­
Aug
19­
Aug
27­
Aug
4­
Sep
12­
Sep
20­
Sep
28­
Sep
Weight
(
g)
Coho
P
redicted
Coho
Observed
Steelhead
Predicted
Steelhead
Observed
Figure
5.10
Weight
prediction
and
observed
fish
weights
for
coho
and
steelhead
populations
at
Porter
Creek,
1988.
5
-
24
Predicted
relative
to
observed
weights
at
the
end
of
the
simulation
period
are
shown
for
all
of
the
site
by
year
data
sets
in
Figure
5.11
and
listed
in
Table
5.3.
The
accuracy
of
the
growth
simulations
was
estimated
by
calculating
the
sum
of
unexplained
residual
variations
between
observed
and
simulated
length
for
each
cohort
of
coho
(
n=
14)
and
steelhead
(
n=
7)
through
the
following
equation:

2
1
)
(
1
i
i
i
predicted
observed
predicted
n
i
W
W
W
n
SCE
-
=
å
=
(
eq.
5.16)

SCE
for
both
coho
and
steelhead
was
0.08
of
the
observed
for
both,
while
the
average
gain
Predicted
Relative
to
Observed
Weights:
Coho
Salmon
­
100%
­
80%
­
60%
­
40%
­
20%
0%
20%
40%
60%
80%
100%

Huckleberry
87
Huckleberry
88
Huckleberry
89
Huckleberry
91
Huckleberry
98
Huckleberry
(
Up)

Big
94
Salmon
94
Hoffstadt
90
Harrington
90
Porter
88
Porter
89
Porter
90
Porter
91
Predicted
Relative
to
Observed
(%)
A.)

Predicted
Relative
to
Observed
Weights:
Steelhead
Trout
­
100%
­
80%
­
60%
­
40%
­
20%
0%
20%
40%
60%
80%
100%

Salmon
94
Hoffstadt
90
Harrington
90
Porter
88
Porter
89
Porter
90
Porter
91
Predicted
Relative
to
Observed
(%)
B.)

Figure
5.11.
Differences
between
predicted
and
observed
average
weights
of
fish
populations.
5
-
25
in
weight
during
the
period
was
1.25
and
4.14
times
initial
weight,
respectively.
Some
predictions
were
above
and
some
were
below
the
observed
weights,
with
no
apparent
patterns.
Model
predictions
averaged
within
4%
of
observed
for
both
species
considering
the
sign
of
the
difference
between
observed
and
predicted,
and
11%
when
averaging
its
absolute
value.
The
largest
prediction
error
occurred
at
Huckleberry
Creek
in
1987
(+
44%).
There
was
very
little
weight
gain
at
this
site
during
this
year
and
our
consumption
estimates
were
clearly
too
high.
As
another
example,
the
steelhead
population
at
Hoffstadt
Creek
was
under­
predicted
by
24%,
perhaps
because
the
population
was
augmented
by
hatchery
fish.
Since
there
are
no
assumptions
about
food
supply
used
to
model
steelhead,
this
error
suggests
that
one
or
more
of
the
model
parameters
were
probably
incorrect
for
this
population
and
year.
Our
unexplained
error
is
consistent
with,
but
somewhat
less
than
achieved
by
Mallet
et
al.
(
1999)
for
grayling
in
a
European
river
where
they
used
a
similar
approach
for
accounting
for
temperature
but
did
not
consider
consumption.

The
close
correspondence
of
observed
and
predicted
weights
was
welcome,
if
not
somewhat
surprising,
given
potential
errors
associated
with
using
consumption
estimates
extrapolated
from
other
streams
in
the
case
of
coho,
or
from
other
species
(
cutthroat),
albeit
closely
related,
in
the
case
of
steelhead.
While
the
physiological­
based
estimates
of
consumption
should
be
broadly
applicable
to
the
species
as
formulated
in
this
paper,
varying
perhaps
by
genetics
among
races,
the
ecological
constraints
on
food
supply
can
vary
significantly
among
sites.
Many
factors
potentially
influence
food
supply;
population
dynamics
and
competition
for
food,
(
Walters
and
Post
1993),
riparian,
inchannel
habitat
characteristics
that
control
primary
and
secondary
production
(
Hawkins
et
al.
1983,
Bilby
and
Bisson
1988,
Hetrick,
Murphy
1998,
Railsback
and
Rose
1999)
and
even
the
energy
content
of
prey
(
Stewart
and
Ibarra
1991).
Information
about
these
riparian,
in­
channel
habitat
and
population
characteristics
are
not
directly
measured,
but
are
integrated
by
growth
rate
in
the
time
series
of
daily
temperatures,
consumption
and
body
weights.
The
average
growth
rate
as
a
function
of
annual
maximum
temperature
is
shown
in
Figure
5.12.
Railsback
and
Rose
(
1999)
observed
that
measured
growth
rates
of
trout
in
the
Sierra
Nevada,
and
those
predicted
by
the
Wisconsin
model,
varied
from
year
Growth
Rate
of
Fish
Populations
Coho
=
­
0.0002T
+
0.0117
R2
=
0.20
Sthd=
­
0.0011T
+
0.0399
R2
=
0.50
0.000
0.005
0.010
0.015
0.020
0.025
10
15
20
25
30
Annual
Maximum
Temperature
(
oC)
Daily
Growth
Rate
(
g
g
­
1d­
1)
Coho
Steelhead
Figure
5.12
Growth
rate
calculated
by
the
model
averaged
for
the
sampling
period
in
relation
to
annual
maximum
temperature
at
each
site.
5
-
26
to
year
and
with
site
characteristics
such
as
flow.
We
observed
similar
variation
in
the
sites
sampled
in
multiple
years
(
Table
5.3),
although
we
did
not
attempt
to
account
for
differences
other
than
those
caused
by
differences
in
temperature.

For
the
final
two
simulations,
we
examined
how
modeling
weight
over
the
full
range
of
rations
and
weights
of
individuals
within
the
population
would
relate
to
estimates
based
on
averages.
Presumably,
individuals
are
feeding
over
the
range
of
possible
rations,
from
minimum
maintenance
to
satiation,
with
consumption
limited
by
their
hierarchical
position
within
the
population
(
Puckett
and
Dill
1985,
Nielsen
1994).
In
theory,
some
fish
Coho
Huckleberry
C
reek
0
1
2
3
4
5
6
7
8
9
10
30
40
50
60
70
80
90
Length
(
mm)
Weight
(
grams)
Coho
Huckleberry
Creek
0
1
2
3
4
5
6
7
8
9
10
30
40
60
80
100
%
of
Satiation
Weight
(
grams)

Steelhead
Salmon
C
reek
0
1
2
3
4
5
6
7
8
9
10
40
50
60
70
80
90
Length
(
mm)
Steelhead
Salmon
C
reek
0
1
2
3
4
5
6
7
8
9
10
30
40
60
80
100
%
of
Satiation
Weight
(
grams)

Figure
5.13.
Examples
of
length­
weight
relationships
for
coho
salmon
population
at
the
end
of
summer
growing
season
at
Huckleberry
Creek
and
the
steelhead
trout
population
at
Salmon
Creek.
Also
shown
are
the
predicted
weight
for
each
ration
for
Huckleberry
Creek
and
Salmon
Creek,
calculated
with
the
site's
temperature
from
the
period
between
population
sampling,
and
a
starting
weight
as
observed
at
the
site
at
the
initial
sampling.
5
-
27
may
be
feeding
at
satiation
ration,
although
population
density
controls
could
prevent
fish
from
achieving
this
level
of
ration
(
Brett
1995).
Individuals
feeding
at
less
than
minimum
maintenance
ration
over
extended
periods
of
time
are
likely
to
be
lost
from
the
population
due
to
starvation.

Weight
gain
was
simulated
for
the
coho
population
at
lower
Huckleberry
Creek
(
1991)
and
the
steelhead
population
at
Salmon
Creek
(
1994)
for
the
intervals
between
fish
population
samplings,
with
weight
initialized
at
the
observed
average
population
weight.
The
maximum
potential
consumption
was
computed
each
day
based
on
temperature
and
fish
weight,
then
categorized
into
5
rations
between
100%
satiation
and
minimum
maintenance
(
30,
40,
60,
80
and
100%).
A
minimum
of
30%
satiation
was
chosen
because
this
is
very
near
starvation
but
still
allows
some
growth
at
some
temperatures.

Predicted
ending
weight
at
each
ration
level
is
shown
in
comparison
to
the
range
of
individual
weights
observed
in
the
population
on
the
final
sampling
date
in
Figure
5.13.
The
range
of
weights
observed
within
the
two
populations
near
the
end
of
the
summer
was
wide,
but
typical
of
what
is
usually
observed:
the
largest
fish
were
4
to
6
times
larger
than
the
smallest
fish
for
both
species.
There
was
a
range
of
weight
among
individuals
within
the
population
early
in
the
season,
although
it
was
considerably
narrower
than
it
was
at
the
end
of
the
summer
(
Figure
5.14).
The
growth
model
predicted
a
range
of
weights
that
was
similar
to
observed.
Model
predictions
suggested
that
coho
as
large
as
9.5
grams
could
occur
at
100%
satiation
ration
and
the
largest
fish
observed
was
7.4
grams,
a
weight
consistent
with
high
consumption,
but
less
than
satiation.
For
steelhead,
the
largest
predicted
individual
was
about
3.5
grams,
considerably
smaller
than
the
largest
individual
in
the
population
assigned
to
the
0
age
class
by
the
field
biologist.
Seventy­
five
percent
of
the
steelhead
individuals
were
within
the
predicted
range.
The
steelhead
that
are
larger
than
the
100%
ration
prediction
could
actually
be
yearling
fish
since
scale
analysis
was
not
performed,
and
it
can
be
difficult
to
distinguish
age
of
the
fish
from
observation
alone.

Exact
correspondence
in
individual
weights
was
not
expected
since
the
average
population
weight
was
used
to
initialize
the
simulation.
If
we
apply
the
same
analysis
of
the
growth
of
the
largest,
average
and
smallest
fish
in
the
coho
population
at
Huckleberry
Creek
as
we
did
earlier
for
Mason's
population,
we
estimate
the
average
degree
of
satiation
for
each
ranked
fish
(
Figure
5.14).
During
this
period
there
was
only
a
small
change
in
the
number
of
fish
in
the
population,
although
we
do
not
know
if
the
fish
that
held
these
ranks
at
the
end
of
the
summer
were
the
same
one's
that
held
them
at
the
beginning.
As
in
Mason's
experiment,
the
largest
fish
ate
at
a
higher
satiation
level
than
smaller
fish
in
the
population.
The
largest
fish
is
estimated
to
have
eaten
at
73%
satiation,
the
average
fish
at
62%,
and
the
smallest
fish
at
45%
satiation,
averaged
over
the
growth
period.
These
results
are
consistent
with
Fausch
(
1983),
who
found
that
dominant
fish
obtained
more
energy
during
natural
feeding.
5
-
28
Modeling
individuals
within
the
range
of
population
sizes
produced
qualitatively
similar
results
as
modeling
the
average
fish
for
the
range
of
rations,
in
terms
of
the
range
of
weights
predicted
and
the
average
consumption.
Perhaps
this
is
because
the
average
population
ration
was
in
the
mid­
range
of
rations.
Nevertheless,
the
growth
model
suggested
a
plausible
range
of
weights
at
the
individual
scale
using
a
range
of
rations,
and
the
model
closely
approximated
the
average
population
weight
and
at
the
population
scale
using
average
weights
and
field­
calibrated
ration
(
Figure
5.11)

Finally,
we
conducted
a
brief
sensitivity
analysis
to
demonstrate
model
behavior
over
a
range
of
temperatures
and
food
consumption
levels.
Weight
gain
was
simulated
using
the
same
time
interval,
starting
weight,
and
range
of
rations
from
satiation
to
minimum
maintenance,
allowing
only
the
temperature
to
vary
for
each
site.
June
1
to
Sept
15
was
selected
to
represent
the
primary
growth
period
for
coho
and
steelhead
since
both
species
should
have
emerged
from
the
stream
bed
by
this
date,
and
our
data
set
from
temperature
sites
was
complete
during
this
interval.
A
representative
initial
weight
was
estimated
by
regressing
weight
of
all
the
populations
sampled
early
in
the
summer
season
with
time.
Steelhead
tended
to
be
very
similar
in
weight
early
in
the
season,
despite
differences
in
sample
dates,
averaging
0.5
grams
on
June
1
(
julian
day
152).
Coho
weight
varied
more,
and
in
some
cases,
included
some
sites
with
hatchery­
raised
fish.
Coho
averaged
1.4
grams
on
June
1.
Temperature
simulations
included
site
temperature,
which
varied
Coho
Growth
Huckleberrry
1991
0
2
4
6
8
10
12
14
16
Initial
Observed
Observed
Smallest
Predicted
Smallest
Observed
Average
Predicted
Average
Observed
Largest
Predicted
Largest
Weight
(
grams)

Ration
10
Ration
80
Ration
60
Ration
40
Ration
30
May
15
Oct
9
Figure
5.14
Observed
and
predicted
weight
of
coho
in
Huckleberry
Creek
(
1991).
The
growth
of
the
smallest,
average,
and
largest
fish
at
the
start
of
the
summer
(
May
15)
were
modeled
for
the
full
range
of
satiation
(
shown
in
bars).
Weight
of
each
class
at
the
end
of
the
period
(
Oct
9)
are
shown
at
the
end
of
the
summer
relative
to
growth
simulations.
5
-
29
through
the
season,
as
well
as
constant
temperature
for
the
period
(
Figure
5.15).
Simulated
weights
based
on
each
site's
temperature
regime
are
plotted
as
circles;
constant
temperature
predictions
are
the
solid
parabolic
lines
at
each
ration,
and
the
observed
fish
population
weights
are
plotted
as
squares.
The
solid
lines
computed
at
constant
temperature
follow
the
growth
rate
curves
on
which
they
are
based,
but
they
reflect
the
effect
of
long­
term
exposure
while
compensating
for
effects
of
increasing
weight
on
Coho
Salmon
0
1
2
3
4
5
6
7
8
9
10
11
0
5
10
15
20
25
30
Temperature
(
oC)
Weight
(
grams)
Predicted
30%
Predicted
40%

Predicted
60%
Predicted
80%

Predicted
100%
Average
Field
Population
Constant
Tempature
100%
Satiation
80%

60%

40%

Steelhead
0.0
1.0
2.0
3.0
4.0
5.0
0
5
10
15
20
25
30
Tempera
ture
(
oC)
Weight
(
grams)
Predic
ted
20%
Predic
ted
40%

Predic
ted
60%
Predic
ted
80%
Predic
ted
100%
A
v
erage
Field
Population
C
o
ns
t
a
n
t
T
emp
e
r
a
t
ur
e
1
0
0
%
S
at
ia
t
io
n
8
0
%

6
0
%
4
0
%
3
0
%

Figure
5.15.
Summary
of
observed
and
predicted
weights
for
coho
salmon
and
steelhead
trout.
Solid
lines
are
the
weight
predicted
at
constant
temperature
for
106­
day
simulation
(
June
1­
Sept
15).
Circles
are
the
predicted
weights
at
each
site
and
ration
based
on
the
daily
temperature
regime
for
the
period.
Squares
are
the
average
weight
of
observed
populations.
Initial
weight
was
1.4
grams
and
0.5
grams
for
coho
and
steelhead,
respectively.
5
-
30
consumption.

Weight
can
be
strongly
affected
by
both
the
prevailing
temperature
and
the
amount
of
food
available
to
fish.
Either
factor
can
have
the
same
level
of
effect,
although
maximum
weight
can
only
occur
when
prevailing
temperatures
are
near
the
species'
physiological
optimum
temperature
and
food
supply
is
high.
Low
to
moderate
weight
can
result
from
many
combinations
of
temperature
and
food
supply,
perhaps
helping
to
explain
why
it
is
difficult
to
determine
whether
food
or
temperature
is
limiting
growth
in
natural
streams
from
empirical
field
observations
alone.

Weight
is
maximized
when
temperature
over
the
course
of
the
summer
is
closer
to
optimal
temperature
(
Figure
5.15).
Washington
streams
and
rivers
tended
to
be
near
optimal,
despite
significant
differences
in
the
temperature
patterns
at
the
sites
when
indexed
by
the
annual
extremes
such
as
annual
temperature
(
Figure
3.1).
The
effects
of
temperature
are
more
pronounced
at
higher
levels
of
food.
Interestingly,
observed
weights
of
coho
and
steelhead
populations
show
patterns
consistent
with
those
simulated
across
the
range
of
temperatures,
although
food
supply
at
sites
was
not
well
known.
The
biomass
of
fish
populations
measured
at
the
study
streams
is
moderate
to
high
relative
to
those
reported
from
the
coastal
areas
of
the
Pacific
Northwest
and
Alaska
(
Bisson
and
Bilby
1998).
At
very
low
levels
of
food,
growth
is
very
low,
but
is
somewhat
better
at
cooler
temperatures.

The
weights
predicted
using
observed
temperatures
do
not
deviate
far
from
those
calculated
with
a
constant
temperature
(
Figure
5.15).
Thus,
since
there
is
little
loss
of
information
using
the
constant
temperature,
these
simulations
suggest
that
the
weight
gain
at
sites
much
warmer
and
colder
than
those
available
in
our
data
sets
are
realistic.
The
constant
temperature
can
be
represented
by
the
median
temperature
of
the
period.
Figure
5.13
suggests
that
the
model
is
sensitive
to
temperature,
and
would
estimate
significantly
different
weights
for
significantly
different
temperature
profiles.
The
similarity
in
the
weights
predicted
with
the
growth
simulation
methods
at
many
of
the
sites
(
Figure
5.15)
does
not
appear
to
reflect
lack
of
sensitivity
of
the
model,
but
rather
the
relatively
narrow
range
of
average
temperature
at
streams
within
the
sample.
Population
weights
should
be
less
in
streams
with
predominantly
warmer
or
colder
than
occur
in
our
data
sets.

DISCUSSION
We
have
developed
a
bioenergetics­
based
approach
to
assessing
the
effects
of
temperature
on
growth
of
salmonids.
At
this
time,
it
has
only
been
applied
to
age
0
steelhead
trout
and
coho
salmon
living
in
natural
streams.
Growth
of
other
salmonids
could
easily
be
modeled
with
the
same
approach
if
the
basic
consumption
and
growth
functions
can
be
established.

There
are
many
examples
of
growth
modeling
based
on
bioenergetics
principles
available
in
the
literature.
Most
use
similar
methods
for
estimating
consumption
terms
but
provide
more
comprehensive
evaluation
of
energetic
functions
in
addition
to
growth
(
Beauchamp
et
al.
1989,
Kitchell
et
al.
1974,
Hewlett
and
Johnson
1987,
Beauchamp
et
al.
1989,
Hanson
et
al.
1997.).
Our
formulation
differs
from
these
in
that
energy
consumed
by
growth,
as
evidenced
by
observed
growth
rate,
is
the
only
energy
function
considered,
and
we
make
no
attempt
to
close
the
energy
balance
between
intake
and
expenditure.
Thus,
we
view
our
approach
as
bioenergetics­
based
but
not
a
true
bioenergetics
model.
However,
the
growth
model
developed
here
shares
many
key
elements
with
bioenergetic
5
-
31
models,
and
therefore
its
application
enjoys
many
of
the
same
challenges
and
criticisms
(
Ney
1993,
Hansen
et
al.
1993).

In
a
review
of
bioenergetics
models,
Ney
(
1993)
noted
that
corroborative
studies
of
bioenergetics
models
showed
a
number
of
deficiencies
that
appear
to
compromise
their
ability
to
estimate
consumption
and
growth
of
non­
captive
fish
accurately.
Some
critics
feel
the
models
are
overly
complex
(
Ney
1993)
and
prone
to
errors
in
parameter
estimation
(
Boisclair
and
Leggett
1989).
A
comprehensive
model
of
energy
functions
results
in
the
proliferation
of
parameters,
which
may
create
difficulties
in
adequately
informing
a
number
of
the
input
variables,
and
defining
some
of
the
energetic
relationships.
Hansen
et
al.
(
1993)
comment
that
bioenergetics
models
having
20
or
more
input
parameters,
each
with
its
own
estimation
error,
can
lead
to
estimates
of
consumption
or
growth
that,
in
some
circumstances
may
have
to
differ
by
100%
or
more
to
be
judged
statistically
different
(
Boiscalir
and
Leggett
1989.)
This
poses
particular
problems
for
trying
to
use
these
models
to
sort
out
complex
ecological
interactions.
Beauchamp
et
al.
(
1989)
noted
that
bioenergetics
models
are
used
more
frequently
to
predict
consumption
than
growth
because
of
the
additional
errors
associated
with
bringing
in
other
equations.
To
the
contrary,
Hanson
(
1997)
argues
that
forcing
a
balance
of
the
energy
budget
acts
to
limit
error
propagation
(
Bartell
et
al.
1986).

The
growth
model
developed
in
this
paper
is
relatively
simple,
especially
in
field
application,
but
the
errors
are
unconstrained.
Our
model
shares
some
of
the
same
relationships
that
are
most
sensitive
to
the
errors
that
are
found
in
full
bioenergetics
models.
Specifically,
the
consumption
terms
that
apply
to
the
physiologic
controls
of
temperature
and
allometry,
as
well
as
the
food
supply,
are
both
important
in
estimating
growth
(
Stewart
et
al.
1983,
Beauchamp
et
al.
1989,
Bartell
et
al.
1986,
Hanson
et
al.
1993).
The
physiologic
relationships
appear
to
be
reasonably
well
established
for
salmonid
species,
given
the
similarity
of
parameter
values
developed
from
different
studies
(
Hanson
et
al.
1997).
We
achieved
good
modelling
results
using
these
values.

There
have
been
questions
as
to
whether
such
laboratory
study
results
can
be
used
for
predicting
response
of
fish
living
in
natural
environments.
Laboratory
studies
have
unique
conditions
of
food,
environment,
and
population
pressures
that
themselves
may
create
stresses
that
may
not
be
observed
in
natural
settings.
In
our
analyses,
the
gain
in
weight
of
fish
living
in
stream
environments
was
closely
approximated
by
relationships
derived
from
laboratory
studies.
This
suggests
that
laboratory
studies
can
be
used
with
some
confidence
to
predict
responses
in
more
natural
ecological
settings.

Achieving
a
quantitative
understanding
of
the
cause
and
effect
linkage
between
food
consumption
and
temperature
in
natural
environments
is
a
significant
challenge,
as
suggested
by
Figure
5.14,
where
different
conditions
can
produce
similar
population
weight.
Field
studies
are
labor
intensive
and
field
estimates
are
laden
with
their
own
assumptions
and
subject
to
their
own
errors
(
Ney
1990).
Trying
to
discern
such
relationships
by
empirical
observation
alone
is
problematic,
given
the
multivariate
and
dynamic
nature
of
the
interaction
and
the
difficulty
of
measuring
some
of
the
key
fundamental
relationships
in
natural
environments
(
Brett
1971,
Boisclair
and
Leggett
1991,
Railsback
1997).
While
the
growth
model
can
not
solve
these
problems,
it
can
help
field
investigators
to
develop
physiologically­
based
hypotheses
that
may
help
them
understand
the
responses
they
observe.
5
-
32
We
found
that
our
simplified
application
of
bioenergetics
principles
provided
some
useful
structure
to
the
analysis
of
field
observations
of
environmental
temperature
and
food
supply
that
appeared
to
address
some
of
the
complexity
of
the
interactions.
This
could
lead
to
greater
insight
into
their
inter­
relationships
when
compared
to
empirical
observation
alone.
For
example,
our
results
suggest
that
food
limitation
existed
for
some
species
and
not
for
others
(
at
least
at
this
age),
and
that
food
limitation
varies
in
time
in
ways
that
have
important
implications
for
population
dynamics.
Additional
study
of
food
consumption
and
supply
in
natural
streams,
where
data
are
extremely
limited,
would
be
very
useful,
given
the
importance
of
these
factors
in
determining
fish
productivity
and
response
to
temperature.
It
seems
clear
from
this
analysis
that
there
is
more
opportunity
to
affect
productivity
(
population
weight
and
density)
by
changing
food
supply
than
by
changing
temperature
within
the
range
of
temperature
observed
in
the
study
streams,
at
least
during
the
summer
months.
This
interpretation
is
consistent
with
field
studies
that
have
shown
increased
productivity
with
canopy
removal,
despite
increases
in
temperature
that
at
times
have
appeared
to
be
adverse
(
Hawkins
et
al.
1983,
Bisson
et
al.
1988).

Importantly,
the
hypotheses
presented
by
the
model
can
be
explicitly
tested
in
field
experimentation,
and
rejected
if
model
results
fail
to
predict
observed
responses.
In
this
first
application,
the
model
hypotheses
were
confirmed,
but
more
carefully
implemented
field
experiments
would
be
beneficial.
We
caution
that
these
results
corroborate
the
utility
of
the
model,
but
do
not
constitute
a
rigorous
test
of
the
model
or
its
underlying
assumptions
(
Hansen
et
al.
1993).
Validation
would
best
be
achieved
by
field
and
laboratory
experiments
to
confirm
the
growth
curves
and
allometric
functions
and
to
independently
determine
food
availability,
rather
than
estimate
consumption
from
observed
change
in
average
population
weight
through
time
(
e.
g.
Filbert
and
Hawkins
1995,
Martin
1985).
Such
tests
would
help
refine
the
input
parameters,
as
well
as
reveal
whether
our
assumption
that
energetic
functions,
other
than
those
captured
in
the
allometric
and
growth
rate
functions,
can
be
ignored
in
estimating
growth
effects
of
environmental
temperature
at
a
necessary
level
of
precision.

It
was
somewhat
surprising
that
we
were
able
to
achieve
such
good
predictions
of
weight
gain
compared
to
observations
of
fish
populations
given
the
sparse
amount
of
data
used
to
develop
consumption
estimates.
Weight
estimates
for
populations
of
juvenile
salmonids
were
generally
within
11%
of
observed
following
growth
over
several
months
during
a
rapid
growth
phase.
However,
the
model
is
always
likely
to
perform
best
using
sitespecific
estimates
for
food
consumption.
Nevertheless,
extrapolated
estimates
proved
satisfactory
for
our
purposes.

We
conclude
that
the
methods
developed
in
this
section
perform
well
for
the
purpose
of
assessing
the
effects
of
environmental
temperature
on
juvenile
salmonid
growth.
Ney
(
1993)
concluded
that,
in
their
present
state
of
development,
bioenergetics
models
are
best
suited
for
making
relative
rather
than
absolute
predictions
such
as
comparing
outcomes
of
different
habitat
and
food
availability
scenarios.
This
is
how
we
emphasize
use
of
this
method.
The
method
is
used
in
Section
6
to
evaluate
the
effects
of
temperature
regime
on
salmonid
growth
relative
to
temperature
thresholds
that
could
be
used
as
water
quality
criteria.
5
-
33
Conclusions
‰
Methods
of
predicting
growth
based
on
quantitative
bioenergetics
principles
can
be
applied
to
streams
to
assess
the
effects
of
temperature
on
juvenile
salmonid
growth,
with
results
that
are
consistent
with
observed
fish
population
growth
patterns.

‰
The
method
is
sensitive
to
temperature
and
can
be
applied
to
the
daily
temperature
regime
making
it
a
useful
tool
for
assessing
the
biological
impacts
of
temperature
in
natural
streams.
5
-
34
6­
1
SECTION
6
QUANTIFYING
GROWTH
EFFECTS
IN
RELATION
TO
TEMPERATURE
THRESHOLDS
Abstract
In
Section
5,
we
developed
a
method
to
estimate
the
effects
of
temperature
and
food
consumption
on
the
gain
or
loss
of
weight
of
coho
salmon
and
steelhead
trout
during
the
summer
months.
In
this
section,
we
use
the
methodology
specifically
to
identify
temperature
indices,
including
the
duration
and
magnitude
of
threshold
temperatures
that
minimize
the
negative
growth
effects
of
the
temperatures
that
occur
over
long
periods
of
time.
The
analysis
provides
an
objective
method
to
establish
temperature
criteria
based
on
protecting
the
opportunity
for
growth,
rather
than
avoiding
catastrophic
impacts.

Analyses
demonstrated
that
some
effects
from
the
ambient
temperature
occurred
at
all
sites,
because
no
stream
spends
all
of
its
time
at
a
fish's
optimal
temperature
for
growth.
Sites
with
both
chronically
high
and
chronically
low
temperatures
were
estimated
to
experience
significant
growth
loss.
Using
the
7­
day
maximum
temperature,
and
allowing
10%
growth
loss,
the
upper
threshold
for
coho
salmon
was
found
to
be
16.5oC
and
the
lower
limit
was
13.0oC.
The
range
for
steelhead
was
wider
at
14.5
to
21oC.
Many
sites
included
in
the
analyses
exceed
current
temperature
criteria
for
Washington
(
annual
maximum
of
16oC).
Streams
that
far
exceed
the
criteria
also
were
predicted
to
have
high
growth
loss.
Those
near
the
threshold
criteria
(
±
1oC)
appeared
to
experience
the
best
temperatures
for
growth.
Streams
with
lower
temperature
(<
13oC)
also
had
higher
growth
loss
and
temperature
less
than
10oC
were
adverse
to
growth
of
both
coho
and
steelhead.
The
method
could
be
used
to
identify
thresholds
for
other
temperature
indices
such
as
annual
maximum
and
7­
day
mean,
or
other
levels
of
growth
loss.

Key
findings
include:

‰
It
is
feasible
to
apply
a
risk­
based
approach
that
uses
data
without
safety
factors
and
produces
an
estimate
of
cumulative
risks.
This
technique
is
a
true
quantitative
benchmark
that
is
measurable
and
testable.

‰
The
majority
of
temperatures
experienced
by
salmonids
are
generally
suboptimal
for
growth,
and
these
exert
some
cost
on
the
maximum
potential
growth.

‰
Colder
water
temperatures
are
not
necessarily
better
for
rearing
salmonids,
and
warmer
water
temperatures
are
not
necessarily
worse.
Concepts
of
safety
factors
in
selecting
thresholds
need
to
be
exercised
with
some
caution.
Thresholds
that
are
too
low
can
also
negatively
effect
growth.

‰
Despite
what
appear
to
be
large
differences
in
temperature
among
sites,
especially
with
regard
to
the
warmest
temperatures
that
occur
each
summer,
there
was
less
of
a
difference
in
the
predominance
of
temperatures
that
are
important
to
growth.
Sites
with
significantly
different
temperature
regimes
often
have
similar
predicted
growth
risk.
6­
2
‰
An
upper
threshold
for
the
7­
day
maximum
temperature
of
16.5oC
minimizes
growth
losses
for
coho.
A
7­
day
maximum
temperature
or
20.5oC
minimizes
growth
losses
for
steelhead,
based
on
an
analysis
that
does
not
consider
population
dynamics.

Introduction
One
of
the
most
important
aspects
of
temperature
in
natural
environments
is
its
effect
on
growth.
Growth
is
regulated
by
a
complex
interrelationship
between
food
supply,
population
dynamics,
and
the
physiologic
responses
of
the
fish
to
temperature
(
Weatherly
1972).
Trying
to
discern
such
relationships
by
empirical
observation
alone
is
problematic,
given
the
multivariate
and
dynamic
nature
of
the
interaction
and
the
difficulty
of
measuring
some
of
the
key
fundamental
relationships
in
natural
environments
(
Brett
1971,
Boisclair
and
Leggett
1991,
Railsback
1997).
The
physiological
responses,
including
energy
consumption
and
expenditures
are
generally
studied
under
laboratory
conditions,
and
have
been
quantified
for
many
species
of
salmonids
(
Weatherly
and
Gill
1995,
Brett
1995).
Field
studies
are
labor
intensive
and
they
are
laden
with
their
own
assumptions
and
subject
to
their
own
errors
(
Ney
1990).
Consequently,
there
are
relatively
few
field
studies
that
have
successfully
established
the
linkage
(
Martin
1985,
Filbert
and
Hawkins
1995).
A
combination
of
field
and
laboratory
study
offers
the
best
hope
for
establishing
a
quantitative
understanding
of
the
cause
and
effect
linkage
between
growth
and
temperature
in
natural
environments
(
Hansen
et
al.
1993).

Laboratory
studies
have
produced
quantitative
relationships
between
energy
consumption
and
expenditure
mechanisms.
The
rate
at
which
most
energetic
functions
proceed
is
mediated
by
ambient
environmental
temperature.
Bioenergetics
models
have
been
developed
to
help
manage
understanding
of
the
multitude
of
physiologic
responses
to
temperature
(
Kitchell
et
al.
1974),
accounting
for
energy
consumption
and
expenditures.
Some
have
been
packaged
into
software
programs
(
Hewlett
and
Johnson
1992,
Hanson
et
al.
1997),
and
have
been
proven
to
be
useful
tools
for
a
number
of
applications
in
fisheries
management
(
Hansen
et
al.
1993).
These
include
the
rearing
of
fish
in
hatcheries
(
McLean
et
al.
1993),
and
populations
in
natural
environments
(
Hanson
et
al.
1997).
The
downside
to
such
models
is
that
they
often
require
many
parameters
to
inform
a
number
of
energetic
functions,
most
of
which
are
difficult
to
quantify
in
natural
environments.
Thus
their
application
to
explore
ecological
responses
to
environmental
temperature
are
more
limited,
although
researchers
have
recently
found
them
promising
for
this
purpose
(
Filbert
and
Hawkins
1995,
Preall
and
Ringler
1989,
Railsback
and
Rose
1999).

In
Section
5,
we
developed
and
corroborated
a
bioenergetics­
based
approach
that
can
be
used
to
evaluate
the
effects
on
growth
of
the
variable
temperatures
that
occur
over
the
juvenile
rearing
period
in
natural
stream
environments.
The
model
treats
the
population
as
a
cohort,
and
does
not
account
for
population
density
effects.
Thus,
the
method
allows
us
to
focus
on
temperature
differences
among
streams
while
minimizing
biological
data
requirements.
The
mathematical
model
simulates
weight
gain
over
a
specified
duration,
and
requires
only
three
input
parameters
(
temperature,
initial
weight,
and
daily
food
consumption).
The
method
is
sufficiently
simple
that
it
can
be
applied
in
field
experimental
studies.
Only
the
food
consumption
term
is
difficult
to
assess
in
streams
(
Filbert
and
Hawkins
1995).
To
our
knowledge,
no
simple
methodology
for
assessing
food
availability
has
been
developed.
6­
3
The
ecological
constraints
on
food
supply
can
vary
significantly
among
sites,
and
they
are
potentially
influenced
by
many
factors:
population
dynamics
and
competition
for
food,
(
Walters
and
Post
1993),
riparian,
in­
channel
habitat
characteristics
as
they
control
primary
and
secondary
production
(
Hawkins
et
al.
1983,
Bilby
and
Bisson
1988,
Hetrick,
Murphy
1998,
Railsback
and
Rose
1999)
and
the
energy
content
of
food
prey
(
Stewart
and
Ibarra
1991).
Information
about
these
riparian,
in­
channel
habitat
and
population
characteristics
is
embodied
in
the
time
series
of
daily
temperatures,
consumption
and
body
weights
of
fish
as
they
grow
in
natural
environments
over
time.

Despite
difficulties
in
establishing
in
situ
food
consumption,
we
were
able
to
generate
estimates
of
consumption
for
juvenile
coho
and
steelhead
from
observations
of
fish
growth
over
time
that
produced
close
correspondence
between
simulated
and
observed
weight
gain,
even
when
parameters
were
extrapolated
among
streams.
Weight
gain
predicted
for
16
populations
of
coho
and
8
populations
of
steelhead
were
generally
within
11%
of
the
observed,
during
a
rapid
growth
period
where
weight
gain
ranged
from
67
to
415%.
Based
on
corroboration
with
observed
population
growth,
we
concluded
that
the
method
is
a
useful
tool
for
quantifying
the
effect
of
temperature
regime
on
growth,
though
not
biomass
until
population
effects
are
accounted
for.

In
this
section,
we
use
the
model
to
perform
a
series
of
relative
comparisons
of
growth
effects
from
observed
stream
temperatures
from
a
number
of
stream
sites
with
widely
varying
temperature
profiles.
We
pay
particular
attention
to
interpreting
results
relative
to
temperature
thresholds
that
are
often
used
for
water
quality
criteria.

Growth
Simulation
Method
The
growth
model
was
used
to
simulate
weight
gain
for
coho
and
steelhead
using
standard
timelines,
initial
weights,
and
consumption
estimates
for
each
species
(
Table
6.1).
Only
the
temperature
varied,
according
to
measured
daily
temperature
at
the
19
sites.
Thus,
this
assessment
isolated
the
effects
of
the
long­
term
temperature
on
growth.
Temperature
profiles
at
the
sites
varied
from
very
warm
to
cold
(
Table
6.2,
and
further
described
in
Section
3),
and
many
exceed
the
current
Washington
Department
of
Ecology
temperature
criteria.
The
temperature
data
set
did
not
include
very
cold
streams
(
12oC
or
less).
These
typically
occur
in
the
extreme
headwaters,
and
are
small
and
non­
fish­
bearing
in
this
region
(
Black
2000).
To
represent
these
streams,
we
also
conducted
two
simulations
using
a
constant
temperature
of
8o
and
10oC.

Growth
of
coho
and
steelhead
populations
was
characterized
by
the
average
population
weight
according
to
the
parameters
identified
in
Table
6.1.
Simulations
were
run
from
June
1
through
September
15
Table
6.1
Input
variables
for
growth
simulations.
See
Section
5
for
a
full
description
of
the
growth
model
and
the
determination
of
parameters.

Input
Variables
Coho
Steelhead
Simulation
Period
June
1­
Sept
15
June
1­
Sept
15
Initial
Weight
(
g)
1.4
0.5
Food
Consumption
(
g
g­
1d­
1)
Declines
through
time
with
ecological
constraint
(
see
Section
5,
Figure
5.8)
Assumed
at
maximum
potential
calculated
by
temperature
and
weight
(
see
Section
5,
Figure
5.8)

Temperature
(
oC)
Daily
mean
calculated
from
hourly
temperature
measurements
Daily
mean
calculated
from
hourly
temperature
measurements
6­
4
because
we
had
a
complete
temperature
record
for
all
of
the
sites
and
this
period
encompasses
most,
if
not
all,
of
the
growth
occurring
during
the
summer
rearing
period.

The
consumption
was
varied
according
to
observed
or
estimated
rates
inferred
from
growth
of
fish
between
time
periods
as
described
in
Section
5.
The
analysis
found
that
there
are
important
differences
between
fish
species
in
their
consumption
patterns.
Age
0
steelhead
appear
to
have
no
environmental
constraint
on
food
supply,
subject
only
to
physiological
constraints
imposed
by
temperature
and
weight.
Coho
show
ecological
constraint
on
food
supply
beginning
in
late
June
that
lasts
through
the
summer
season.
Fausch
(
1984)
observed
similar
patterns,
concluding
that
these
differences
are
likely
to
influence
the
species'
response
to
temperature.

Coho
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
Weight
(
g)

Steelhead
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
Weight
(
g)

Figure
6.1
Estimated
weight
after
106­
day
growth
period
using
the
same
initial
weight
and
food
consumption
assumptions
at
each
site.
6­
5
Temperature
Effects
on
Weight
Gain
The
predicted
weights
after
107
days
of
simulated
growth
are
shown
in
Figure
6.1.
Because
the
initiating
assumptions
are
the
same,
the
predicted
weights
are
similar.
Any
differences
in
observed
weight
are
due
to
the
temperature
at
each
of
the
sites.
Coho
vary
up
to
1.2
grams
(
31%
of
the
mean),
and
cooler
sites
generally
had
greater
weights.
The
predicted
weights
of
steelhead
vary
by
only
0.5
grams
among
sites
(
12%
of
the
mean),
despite
large
differences
in
the
temperatures.

Relative
Temperature
Effects
on
Growth
Our
approach
to
evaluating
the
effects
of
temperature
on
salmonid
growth
is
to
estimate
the
weight
gain
achieved
during
the
summer
according
to
the
observed
temperature
regimes
and
estimates
of
food
availability,
then
compare
that
growth
to
a
reference
point.
Our
reference
point
is
the
growth
the
fish
would
have
achieved
if
temperatures
had
been
at
optimal
for
growth,
and
at
the
same
level
of
food
consumption.
Brett
(
1971),
Brett
et
al.
(
1982),
Railsback
and
Rose
(
1997),
Preall
and
Ringler
(
1989)
used
a
similar
reference
approach,
and
growth
models
as
a
basis
for
prediction.

Establishing
Optimal
Temperature
and
Growth.
Previous
studies
have
established
a
maximum
potential
growth
as
the
point
of
reference.
Many
have
used
the
size
calculated
at
constant
optimal
temperature
and
maximum
consumption
(
a
reference
only
observed
at
optimal
temperature
and
low
weight).
(
See
Section
5
for
a
full
discussion
of
this
point).
This
represents
the
maximum
possible
growth
that
fish
could
achieve,
and
it
is
probably
only
observed
in
the
laboratory
setting
where
both
of
those
parameters
can
be
maintained
at
required
levels
and
for
a
short
period
in
a
fish's
life.

It's
important
to
note
that
optimal
temperature
is
not
static
but
varies
with
weight
and
ration
(
Stewart
and
Ibarra
1991,
Brett
et
al.
1971,
Brett
1995).
The
change
in
optimal
temperature
with
rations
between
the
minimum
required
for
basic
metabolic
functions
and
the
maximum
at
satiation
is
shown
for
coho
and
steelhead
in
Figure
6.2.
Note
that
consumption
rates
differ
for
the
two
species,
as
does
the
range
of
optimal
temperatures.
There
are
also
changes
in
consumption
rates
as
the
fish
gain
weight,
referred
to
as
allometric
relationships
(
Brett
1995,
Stewart
and
Ibarra
1991).
In
the
optimal
growth
simulations,
we
reduce
consumption
according
to
fish
weight
and
any
ecological
constraints
on
the
food
supply.
Optimal
temperature
is
then
reduced
accordingly.
This
yields
realistic
estimates
of
optimal
growth
reflecting
well­
documented
allometric
effects.
By
adjusting
optimal
temperature
with
consumption,
we
believe
the
reference
optimal
growth
simulation
realistically
Optimal
Temperature
in
R
elation
to
F
o
o
d
Consumption
10
11
12
13
14
15
16
17
18
0
0.05
0.1
0.15
0.2
Consumption
(
g
g­
1d­
1)
Optimal
Temperature
(
oC)
Coho
Steelhead
Figure
6.2
Optimal
temperature
in
relation
to
food
consumption.
Consumption
is
expressed
as
grams
of
prey
per
gram
of
body
mass
per
day.
The
range
displayed
is
from
satiation
to
minimum
maintenance
for
each
species.
6­
6
represents
the
potential
effects
of
temperature.
It
is
not
clear
if
previous
researchers
who
have
used
optimal
growth
benchmarks
have
performed
this
adjustment.

In
the
second
scenario,
weight
is
calculated
allowing
the
daily
growth
rate
(
gi)
to
vary
with
temperature
according
to
the
specific
growth
curves
in
Figure
5.4,
and
food
consumption
assumptions
(
as
illustrated
in
Figure
5.8).
We
express
growth
in
terms
of
the
Reduction
In
Maximum
Growth
(
RMG).
RMG
is
defined
as
the
percentage
of
reduction
in
growth
for
the
site
specific
temperature
compared
to
the
maximum
growth
achieved
at
the
optimum
temperature
for
each
ration.

RMG
(%)
is
calculated
according
to:

100
)
1
(
x
W
W
RMG
optimal
t
t
=
-
=
(
6.1)

A
sketch
of
the
calculation
is
provided
in
Figure
6.3.
Since
no
stream
had
optimal
temperature
all
of
the
time,
the
gain
in
weight
with
varible
temperature
should
be
less
than
that
for
the
optimal
case.
The
RMG
is
expressed
as
a
percentage
of
the
maximum
weight.
The
lower
the
value
of
RMG,
the
less
the
deviation
from
the
optimal
growth
rate.
A
RMG
value
of
0%
suggests
there
is
no
growth
loss
due
to
the
temperature
at
the
site.

The
reduction
in
maximum
growth
due
to
temperature
varied
among
species
(
Table
6.2).
It
is
clear
that
fish
spent
only
a
portion
of
their
time
in
the
optimal
temperature
range
during
their
growth
period,
since
RMG
was
greater
than
zero
at
all
sites
for
both
species.
Consequently,
there
was
some
cost
to
growth
for
salmonids
living
in
Pacific
Northwest
streams
due
to
their
temperatures
during
the
rearing
growth
phase
(
Table
6.2).
However,
the
RMG
were
generally
within
20%
for
both
species
at
all
sites.

0
0.5
1
1.5
2
2.5
3
3.5
4
26­
Mar­
97
6­
Jun­
97
16­
Aug­
97
27­
Oct­
97
Weight
(
grams)

RMG
Grow
th
at
optim
al
tem
p
Grow
th
at
obs
e
rve
d
tem
p
Figure
6.3
Illustration
of
the
reduction
in
maximum
growth
based
on
two
temperature
scenarios;
growth
at
constant
optimal
temperature
and
with
the
variable
growth
rate
based
on
the
site's
temperature
regime.
6­
7
Table
6.2
Location
and
temperature
characteristics
of
temperature
sites
used
in
reduction
in
growth
analysis.

Site
characteristics
Temperature
characteristics
Reduction
from
maximum
growth
(%)

Site
Watershed
Basin
Area
(
km2)
Year
7­
Day
Maximuma
oC
7­
Day
Meanb
oC
Annual
Maximumc
oC
Season
Mediand
oC
Coho
Steelhead
Deschutes
River
mainstem
Deschutes
145.0
1994
21.0
18.4
22.5
15.0
13.5
11.4
Thurston
Creek
Deschutes
9.1
1994
14.9
14.1
15.5
12
10.0
9.5
Hard
Creek
Deschutes
3.0
1994
14.0
13.0
14.0
11.0
19.2
13.7
Ware
Creek
Deschutes
2.8
1994
17.5
16.1
18.3
12.9
14.7
8.7
Huckleberry
Creek
Deschutes
5.3
1991
18.4
17.6
18.5
15.5
18.1
12.7
Chehalis
River
mainstem
(
Site
1)
Chehalis
181.8
1997
21.1
18.9
22.1
15.6
16.8
16.0
Chehalis
River
mainstem
(
Site
2)
Chehalis
57.5
1997
22.1
18.2
23.2
14.5
15.3
13.1
Chehalis
River
mainstem
(
Site
3)
Chehalis
29.5
1997
20.6
18.6
21.4
14.3
13.8
11.2
Crim
Creek
Chehalis
22.0
1997
18.8
16.9
19.4
14.3
11.6
9.9
Lester
Creek
Chehalis
10.4
1997
18.4
16.3
19.0
14.2
9.3
8.5
Thrash
Creek
Chehalis
16.7
1997
15.3
14.3
15.8
12.3
8.8
8.1
Rogers
Creek
Chehalis
13.1
1997
15.7
14.1
16.1
12.6
6.4
7.0
Big
Creek
Chehalis
9.0
1997
16.5
14.6
16.9
12.5
6.2
6.9
Sage
Creek
Chehalis
5.3
1997
16.5
14.6
16.9
12.5
9.1
7.7
Salmon
Creek
Chehalis
8.9
1997
15.8
14.2
16.2
12.3
8.1
7.7
Mack
Creek
Chehalis
2.8
1997
12.9
12.5
13.1
11.7
6.2
9.0
Porter
Creek
Chehalis
25
1990
17.5
16.3
18.6
14.4
17.1
10.5
Hoffstadt
Creek
Toutle
25.6
1990
24.5
18.4
26.0
14.0
24.6
15.1
Harrington
Creek
Toutle
8
1990
19.1
16.7
20.5
13.3
16.5
9.5
Eight
(
Constant)
NA
NA
NA
8
8
8
8
28.1
35.4
Ten
(
Constant)
NA
NA
NA
10
10
10
10
7.1
18.3
a
maximum
value
of
the
7­
day
moving
average
of
the
daily
maximum
temperature
b
maximum
value
of
the
7­
day
moving
average
of
the
daily
mean
temperature
c
instantaneous
maximum
d
median
of
daily
mean
temperature
from
June
1
to
September
1
6­
8
Temperature
criteria
for
water
quality
standards
are
generally
applied
for
time­
averaged
characteristics
of
temperature
such
as
the
warmest
7­
day
average
of
daily
maximum
and
mean
temperature,
or
the
annual
maximum
temperature
(
instantaneous
measure)
(
see
Section
3).
Growth
reduction
(
RMG)
is
shown
relative
to
various
time­
averaged
temperature
indices
in
Figure
6.4.
RMG
was
at
a
minimum
when
the
stream's
temperature
index
most
closed
approximated
the
species'optimal
temperature.
All
streams
had
some
growth
loss
due
to
its
long­
term
temperature,
since
no
stream
had
RMG
equal
to
0.
RMG
tended
to
increase
for
streams
significantly
warmer
or
colder
than
the
optimum
or
lowest
point
of
the
growth
curves.
Generally,
the
sites
with
lowest
growth
loss
had
indexing
mean
temperatures
within
the
range
of
optimal
temperatures
(
e.
g.,
Figure
6.2).
These
are
14o
to
17oC
for
coho
and
11.5
o
to
14oC
for
steelhead
using
the
7­
day
maximum
measure.
Both
species
showed
steep
response
in
growth
loss
at
higher
and
lower
temperatures
than
their
optimal
range.
There
was
significantly
growth
loss
for
both
coho
0
10
20
30
5
10
15
20
25
7­
Day
Maximum
Temperature
(
oC)
Coho
Steelhead
A)

0
10
20
30
5
10
15
20
25
30
Annual
Maximum
Temperature
(
oC)
Coho
Steelhead
B)

0
10
20
30
5
10
15
20
7­
Day
Mean
Temperature
(
oC)
Coho
Steelhead
C)

0
10
20
30
5
10
15
20
Season
Median
Temperature
(
oC)
Coho
Steelhead
D)

Figure
6.4
Estimated
growth
risk
at
temperature
study
sites
in
relation
to
several
time­
averaged
temperature
metrics:
A)
7­
day
maximum,
B)
annual
maximum,
C)
7­
day
mean,
and
D)
median
temperature
for
the
simulation
period.
6­
9
and
steelhead
when
the
7­
day
maximum
temperature
was
less
than
10oC
and
greater
than
24oC.
The
patterns
are
similar
for
both
species,
although
coho
predictions
are
more
variable
with
site
temperature
than
steelhead.
These
two
species
often
cohabit
the
same
streams,
and
with
the
growth
simulation
method,
they
are
predicted
to
have
similar
level
of
response
at
approximately
the
same
temperatures,
despite
fundamental
differences
in
their
specific
growth
rate/
temperature
curves
(
Section
5).

The
general
patterns
in
RMG
described
for
the
7­
day
maximum
temperature
also
held
true
for
all
of
the
time­
averaging
periods
(
Figure
6.4).
Of
course,
the
range
of
temperature
where
growth
was
optimized
varies
with
each
of
the
temperature
indices.
The
relationship
with
the
least
scatter
was
the
7­
day
mean
temperature,
suggesting
it
may
be
the
best
for
temperature
criteria.

While
the
species
have
similar
patterns
of
response,
there
were
also
important,
nonintuitive
differences
in
simulation
results.
Although
the
optimal
temperatures
for
coho
are
higher
than
steelhead
(
Figure
6.2),
their
growth
was
maximized
within
a
narrower
and
lower
range
of
temperatures.
Steelhead
maximized
growth
at
a
wider
and
somewhat
higher
range
of
temperatures.
We
believe
that
these
differences
reflect
the
food
supply
with
which
each
species
is
modeled.
Because
no
food
limitation
was
assumed
for
steelhead,
their
optimal
temperatures
were
on
the
high
end
of
the
optimal
range.
This
gave
steelhead
a
broader
temperature
range
where
growth
was
not
compromised
by
temperature.
Coho
had
a
narrower
range
of
temperature
where
growth
was
optimized,
and
this
range
was
consistent
with
the
optimal
growth
range
indicated
by
their
specific
growth
curves.
Assumptions
of
food
supply
restrictions
helped
ensure
that
the
optimal
growth
for
coho
would
include
those
temperatures
associated
with
low
consumption
(
13
to16.5
oC).

The
short­
duration
indices
appear
to
be
useful
for
characterizing
the
long­
term
temperature
pattern
in
a
way
that
is
meaningful
to
fish
growth.
In
Figure
6.5,
the
daily
mean
temperature
for
the
entire
simulation
period
is
shown
for
three
sites.
Some
sites
Coho
0
5
10
15
20
25
6­
Jun
12­
Jun
18­
Jun
24­
Jun
30­
Jun
6­
Jul
12­
Jul
18­
Jul
24­
Jul
30­
Jul
5­
Aug
11­
Aug
17­
Aug
23­
Aug
29­
Aug
4­
Sep
10­
Sep
Mean
Daily
Temperature
(
oC)
Hoffstadt
Cr
Big
Cr
Hard
Cr
Optimal
Figure
6.5
Temperature
regime
at
selected
sites
with
estimated
optimal
temperature.
6­
10
spend
a
lot
of
time
above
optimal
temperature
(
e.
g.,
Hoffstadt
Creek),
and
some
spend
all
or
most
of
the
time
below
optimal
(
e.
g.
Hard
Creek).
Growth
is
impacted
in
both
these
cases
by
approximately
20%.
The
sites
where
temperature
was
close
to
optimal
for
the
longest
time
(
e.
g.,
Big
Creek),
had
the
least
effects
on
growth.
Most
site
temperatures
tended
to
be
well
below
optimal
early
in
the
summer
growing
season
when
growth
rates
are
maximized,
partly
because
fish
were
smaller.
Better
growth
early
in
the
season
appeared
to
compensate
somewhat
for
very
warm
temperatures
later
in
the
season.
Figure
6.5
also
illustrates
how
optimal
temperature
declines
through
time
with
decreasing
food
consumption
due
to
weight
gain.

Since
the
sites
span
a
range
of
temperature
regimes,
the
site
with
the
lowest
growth
loss
represents
the
temperature
profile
that
best
encourages
growth
for
that
species.
In
a
sense,
this
site
fits
the
concept
of
an
"
index"
stream,
although
in
this
case,
the
reference
is
defined
by
fish
growth
rather
than
its
naturalness
or
lack
of
disturbance.
Interestingly,
Big
Creek
had
the
most
optimal
temperature
for
both
coho
and
steelhead.
This
site
has
an
annual
maximum
temperature
of
16.9oC,
nearly
1
degree
over
the
Washington
water
temperature
criteria.

Risk
Associated
with
Growth
Limitation
Holtby
and
Scrivener
(
1989)
and
Quinn
and
Peterson
(
1996)
found
that
coho
size
at
the
end
of
summer
was
a
primary
factor
influencing
overwintering
survival
and
smolting.
Holtby
and
Scrivener
(
1989)
provided
an
equation
relating
probabilities
of
overwintering
success
to
coho
length.
We
translated
this
relationship
to
an
equation
based
on
weight
using
a
population
length/
weight
relationship
(
Ricker
1975).
The
relationship
between
weight
and
the
probability
of
overwintering
success
from
these
two
studies
are
shown
in
Figure
6.6.
Holtby
and
Scrivener
(
1989)
found
a
large
increase
in
overwintering
success
with
increased
size,
possibly
because
the
coho
were
so
small
at
Carnation
Creek
(
generally
less
than
2
grams).
According
to
their
relationship,
weight
of
about
6
grams
or
more
yields
an
80%
or
better
probability
of
successfully
overwintering.
Note
that
we
have
extended
Holtby
and
Scrivener's
relationship
beyond
the
limits
of
their
data
to
cover
the
larger
fish
sizes
at
our
sites.
Quinn
and
Peterson
(
1996)
found
more
modest
improvement
in
overwintering
success
at
Big
Beef
Creek
with
fish
size,
although
the
fish
in
this
stream
were
significantly
larger
than
Carnation
Creek.
These
authors
found
that
the
probability
of
successfully
overwintering
was
about
50%
for
fish
>
89
mm
(
approximately
8
grams)
and
only
17%
for
fish
<
60
mm
(
approximately
2.5
grams).
These
values
agree
more
closely
with
fish
sizes
in
our
study
streams.
We
connect
the
two
lines
to
determine
a
relationship
of
overwintering
success
with
weight.
Pr
=
5.8511xW+
3.1915
0
20
40
60
80
100
0
2
4
6
8
10
12
Weight
(
grams)
Holtby
and
Scrivener
1989
Quinn
and
Peterson
1996
Figure
6.6
Probability
of
success
for
coho
overwintering
survival
(
from
Holtby
and
Scrivener
1989)
based
on
end
of
summer
fish
size.
Data
from
Quinn
and
Peterson
(
1996)
are
also
shown.
The
two
points
were
connected
with
a
linear
regression.
6­
11
We
use
both
of
the
relationships
to
illustrate
the
potential
effect
of
growth
reduction
on
overwintering
success.
The
probability
of
success
for
each
individual
in
the
population
at
Salmon
Creek
was
calculated
based
on
its
weight
(
shown
as
actual
weight
in
Figure
6.7).
Quinn
and
Peterson's
data
are
used
in
Figure
6.7A
and
Holtby
and
Scrivener's
relationship
is
used
to
calculate
Figure
6.7B.
Despite
the
large
differences
in
estimated
probabilities
with
the
two
relationships,
both
produce
similar
relative
results.
There
is
a
reduction
in
overwintering
success
with
lower
weight;
the
magnitude
of
change
is
approximately
equal
to
the
percentage
change
in
weight.
That
is,
a
10%
reduction
in
growth
calculated
for
the
population
results
in
an
average
reduction
in
overwintering
success
of
9%.

The
important
outcome
of
this
analysis
is
the
suggestion
that
the
relatively
small
changes
in
weight
that
we
calculate
due
to
temperature
(
e.
g.
Figure
6.1)
are
sufficient
to
affect
individual
and
population
overwintering
success
to
some
extent.
A
10%
reduction
in
growth
would
be
difficult
to
statistically
detect
given
the
typical
range
of
sizes
in
natural
populations.
However,
a
20%
reduction
should
be
detectable,
especially
when
field
experiments
are
guided
by
hypotheses
generated
from
the
growth
model.
Brett
et
al.
(
1982)
suggested
a
20%
upper
limit
for
change
in
weight
due
to
temperature
for
chinook
populations
living
in
the
Nechako
River.
This
appears
to
be
somewhat
high
for
coho,
based
on
implications
for
loss
of
overwintering
success.
However,
it
should
be
noted
that
many
factors
affect
the
survival
of
salmon
in
the
marine
environment.

Use
of
either
relationship
extends
them
beyond
the
original
data
or
application
developed
by
the
authors.
Therefore,
even
though
we
use
both
relationships
to
estimate
the
effect
of
growth
reduction
from
temperature
on
overwintering
success,
we
acknowledge
uncertainty
in
this
analysis.
Additional
research
quantifying
the
effect
of
size
on
success
at
later
life
history
stages
would
increase
confidence
in
the
analysis
of
risk
to
growth
loss
Probability
of
Overwintering
Success
Quinn
and
Peterson
0
5
10
15
20
25
30
35
40
0­
10
11­
20
21­
30
31­
40
41­
50
51­
60
61­
70
71­
80
81­
90
91­
100
Probability
Category
Number
Actual
Weight
­
10%
­
20%
A)

Probability
of
Overwintering
Success
Holtby
and
Scrivener
0
5
10
15
20
25
0­
10
11­
20
21­
30
31­
40
41­
50
51­
60
61­
70
71­
80
81­
90
91­
100
Probability
Category
Number
Actual
Weight
­
10%
­
20%
B)

Figure
6.7
Histogram
of
probability
of
overwintering
survival
based
on
weight
of
individuals
within
the
population
and
simulated
effects
with
redcution
in
weight
due
to
temperature.
Probability
calculated
based
on
Quinn
and
Peterson
results
from
Big
Beef
Creek,
WA
(
A)
and
from
Holtby
and
Scriverner's
from
Carnation
Creek
(
B).
6­
12
due
to
temperature.
This,
in
turn,
creates
uncertainty
regarding
the
choice
of
10%
as
the
growth
reduction
limit.
It
appears
clear
that
at
least
5%
growth
loss
can
be
expected
at
all
sites
due
to
long
duration
exposure,
even
when
the
bulk
of
temperatures
are
near
optimal.
The
difference
in
temperature
thresholds
selected
at
10
and
20%
RMG
is
quite
significant.
Increased
understanding
of
the
role
of
juvenile
size
in
determining
success
at
later
life
history
stages
would
improve
confidence
in
selecting
an
appropriate
limit
to
growth
loss.

In
the
Carnation
Creek
study,
improved
growth
of
steelhead
fry
with
increased
temperature
after
logging
did
not
translate
to
larger
smolts
after
two
to
three
years
of
rearing
(
Hartman
and
Scrivener
1990).
Steelhead
will
usually
spend
at
least
one
additional
year
in
the
stream
regardless
of
size
achieved
in
the
first
year.
Thus
the
impacts
on
growth
from
temperature
are
shown
to
be
small
in
this
analysis,
which
is
consistent
with
observations
at
Carnation
Creek.
The
negative
or
positive
effects
on
growth
are
not
great
enough
to
change
weight
sufficiently
to
change
migratory
patterns,
that
is,
to
speed
up
or
delay
them
by
one
year.

This
is
in
contrast
to
coho
at
Carnation
Creek,
where
temperature
increased
growth
sufficiently
to
bring
some
fish
to
smolting
size
in
one
year.
It
should
be
noted
that
Southern
British
Columbia
represents
the
most
northerly
locale
where
coho
are
typically
able
to
reach
smolting
size
in
one
year
(
Sandercock
1991).
Coho
at
Carnation
Creek
typically
migrate
at
2­
years
rather
than
one,
presumably
due
to
lower
temperatures.
In
this
case,
increasing
the
temperature
accelerated
growth
to
the
point
where
the
coho
outmigrated
after
the
first
season.
This
was
interpreted
as
negative
for
the
species,
because
the
timing
of
their
migration
made
them
more
susceptible
to
predation
in
the
estuary
and
ocean
environment.
Where
fish
typically
migrate
in
one
year
(
e.
g.,
Washington
and
Oregon),
growth
improvement
would
probably
benefit
their
success
by
producing
fish
of
larger
size,
according
to
Quinn
and
Peterson
(
1995).

Growth
Loss
and
Temperature
Criteria
We
translate
the
RMG
data
(
Figure
6.4)
to
zones
of
reduction
in
maximum
growth
to
facilitate
identifying
thresholds
of
growth
response
in
Figure
6.8.
The
range
of
temperatures
was
determined
by
ordering
the
site
temperature
data
and
estimating
the
temperature
where
the
10
and
20%
boundaries
occurred.
The
range
of
20%
reduction
encompasses
most
of
the
stream
temperatures
typically
observed
in
the
region.
The
range
of
temperature
where
there
was
relatively
little
effect
less
than
10%)
was
fairly
narrow
for
both
species.
It
should
be
noted
that
at
temperatures
above
and
below
the
ranges
illustrated,
there
is
high
growth
loss
due
to
temperature
(
30%
or
more).

Discussion
The
temperature
assessment
approach
provides
a
method
for
indexing
the
relative
effects
of
stream
temperature
regimes
on
salmonid
growth.
A
value
of
the
approach
is
that
the
relative
effects
of
temperature
and
food
consumption
can
be
evaluated
independently
of
other
habitat
or
population
characteristics,
and
each
other.
Results
can
be
also
be
used
to
directly
compare
growth
effects
among
species.
Because
these
factors
are
assumed
constant
in
this
analysis,
the
growth
estimates
only
account
for
the
direct
effects
of
temperature,
and
do
not
account
for
population
dynamics.
6­
13
Species
were
similar
in
their
range
of
response.
However,
coho
were
more
temperature
sensitive,
and
steelhead
tended
to
grow
somewhat
better
at
warmer
temperatures
than
coho.
If
bioenergetics
relationships
are
correct,
this
can
be
explained
by
differences
in
food
supply.
This,
in
turn,
may
reflect
the
different
foraging
strategies
that
each
species
utilizes
when
coexisting
in
the
same
streams
(
Bisson
et
al.
1988b).
Selection
of
temperature
criteria
in
management
situations
may
be
most
useful
if
they
reflect
the
most
sensitive
species
(
coho)
when
both
species
are
present.

Fish
were
predicted
to
be
growing
near
optimal
within
many
streams,
including
a
number
that
exceed
current
water
quality
standards.
The
temperature
ranges
observed
at
the
sites
included
in
this
analysis
are
representative
of
current
conditions
in
Washington
streams
and
rivers.
The
vegetative
overstory
of
many
has
been
disturbed
within
the
last
50
years,
and
therefore
streams
may
be
currently
warmer
than
they
have
been
at
other
times
in
their
history.
However,
the
range
of
temperature
represented
at
these
sites
is
likely
to
be
representative
of
the
range
that
has
occurred
historically,
given
the
history
and
frequency
of
fire
disturbance
in
the
region
(
Agee
1993).

The
patterns
of
biological
growth
response
in
relation
to
all
of
the
time­
averaged
temperature
regime
metrics
demonstrates
that
they
can
be
used
to
index
the
temperature
regimes
of
sites
in
biologically
meaningful
ways.
Short
averaging
periods
such
as
7­
day
and
even
the
annual
instantaneous
maximum
temperature
are
strongly
indicative
of
the
long­
term
temperature
regime
that
partially
controls
fish
growth
during
the
summer.
7­
Day
Maximum
0
5
10
15
20
25
30
Coho
Steelhead
Temperature
(
oC)

10%
20%

20%
Annual
Maximum
0
5
10
15
20
25
30
Coho
Stealhead
Temperature
(
oC)
20%

20%
10%
7­
Day
Mean
0
5
10
15
20
25
30
Coho
Steelhead
Temperature
(
oC)

20%

10%

20%

Figure
6.8
Ranges
of
temperature
where
reduction
from
maximum
growth
is
0­
10%
and
11­
20%.
At
temperatures
above
and
below
these
ranges
RMG
exceeds
20%.
6­
14
CONCLUSIONS
‰
It
is
feasible
to
apply
a
risk­
based
approach
that
uses
data
without
undefined
safety
factors
and
produces
an
estimate
of
cumulative
risks.
This
technique
is
a
true
quantitative
benchmark
that
is
measurable
and
testable.

‰
Modeled
growth
using
measured
temperature
suggest
that
the
majority
of
temperatures
experienced
by
salmonids
are
generally
suboptimal
for
growth,
and
these
exert
some
cost
on
the
maximum
potential
growth.

‰
Despite
what
appear
to
be
large
differences
in
temperature
among
sites,
especially
with
regard
to
the
warmest
temperatures
that
occur
each
summer,
there
is
considerable
similarity
in
the
predominance
of
temperatures
that
are
important
to
growth.
Sites
with
significantly
different
temperature
regimes
can
have
similar
predicted
effects
on
growth.

‰
An
upper
threshold
for
the
7­
day
maximum
temperature
of
16.5oC
minimizes
growth
losses
for
coho.
A
7­
day
maximum
temperature
or
20.5oC
minimizes
growth
losses
for
steelhead.

‰
Concepts
of
safety
factors
in
selecting
temperature
thresholds
defined
for
salmonids
need
to
be
exercised
with
some
caution.
Thresholds
that
are
both
too
low
and
too
high
can
negatively
affect
growth.

‰
The
criteria
above
assume
10%
growth
loss
as
the
acceptable
level
of
risk.
There
is
uncertainty
associated
with
this
number,
since
there
are
relatively
few
quantitative
data
to
base
it
on.
Further
research
could
help
confirm
acceptable
risk
levels.
7­
1
SECTION
7
RELATIONSHIP
BETWEEN
EXISTING
AND
PROPOSED
TEMPERATURE
CRITERIA
AND
RISK
ASSESSMENT
FINDINGS
Abstract
Understanding
the
biological
effects
of
temperature
on
fish
is
essential
for
effectively
managing
stream
temperature
under
the
Clean
Water
Act.
One
of
the
key
elements
of
water
quality
management
is
to
establish
temperature
criteria
(
e.
g.,
water
quality
standards)
that
will
limit
human­
caused
impacts
to
the
beneficial
uses
of
the
stream
(
e.
g.,
fish).
Ideally,
criteria
are
based
on
an
understanding
of
the
interaction
of
fish
physiology
and
ecology
(
biological
effects)
and
the
physical
watershed
and
climatic
processes
that
control
the
temperature
of
streams
(
exposure).
Scientific
understanding
of
these
factors
should
then
lead
to
criteria
that
are
realistic
and
appropriate
in
assigning
a
temperature
threshold
that
appropriately
reflects
temperature.
In
this
section,
we
review
and
compare
methods
of
determining
temperature
criteria
used
to
select
criteria,
including
those
developed
in
this
report.

Key
findings
of
this
chapter
are:

‰
Risk
assessment
allows
the
effects
of
magnitude,
duration
and
frequency
of
temperature
on
fish
growth
and
survival
to
be
quantified
in
an
objective
and
repeatable
manner.

‰
The
U.
S.
EPA
(
1977)
temperature
criteria
were
found
to
be
the
most
objectively
defined
and
consistent
with
risk
analysis
results.
They
generally
appear
to
allow
up
to
20%
reduction
in
growth
due
to
temperature.

‰
Criteria
derived
from
review
of
scientific
literature
without
quantitative
synthesis
are
generally
consistent
with
risk
assessment
and
U.
S.
EPA
methods,
although
they
tend
to
overestimate
the
benefits
of
cold
temperatures
and
slightly
underestimate
the
positive
growth
effects
at
temperatures
somewhat
higher
than
optimum.

Introduction
From
a
scientific
basis,
methods
for
deriving
temperature
criteria
should
be
explicitly
defined
and
based
on
sound
scientific
data
that
pass
data
quality
screening
criteria
(
ASTM
1997).
A
criteria
derivation
protocol
needs
to
have:

‰
clearly
defined,
transparent
and
repeatable
methodology;

‰
data
quality
objectives,
attributed
data
sources,
and
quality
control
screens;

‰
defined
levels
of
protection
for
species
populations,
communities
or
ecosystems;
7­
2
‰
stated
assumptions,
safety
factors,
and
data
extrapolation
factors;
and
‰
temperature
criteria
that
incorporate
magnitude,
frequency,
and
duration
as
decision
variables.

In
previous
sections
of
this
report
we:
reviewed
the
scientific
literature
elucidating
the
growth
and
acute
lethal
response
of
salmon
to
temperature
during
the
rearing
life
history
phase
(
Section
2);
explored
relationships
between
long­
term
and
short­
term
temperature
indices
(
Section
3);
performed
risk
analysis
on
the
effects
of
temperature
on
mortality
of
salmon
(
Section
4);
developed
a
quantitative
method
to
asses
effects
of
long­
term
exposure
to
temperature
on
growth
(
Section
5);
and
established
temperature
threshold
criteria
(
Section
6).
In
this
section
we
evaluate
water
quality
temperature
criteria
derived
from
several
methods
that
have
been
adopted
or
proposed
by
various
authors
and
agencies
relative
to
their
biological
effects,
including
the
analyses
developed
in
this
report.

Criteria
and
Methods
To
Derive
Them
Review
of
scientific
literature
and
agency
policy
documents
identified
a
number
of
different
approaches
to
derive
water
temperature
criteria
for
the
summer
maximum
temperature.
Methods
fall
into
three
general
categories:

‰
criteria
derived
from
experimental
temperature
tolerance
studies;

‰
criteria
derived
from
field
observations
of
fish
occurrence
under
different
temperature
regimes;
and
‰
criteria
derived
from
professional
review
of
temperature
information.

The
degree
of
objectivity
or
subjectivity
by
which
the
information
is
synthesized
into
recommended
criteria,
the
degree
to
which
data
forms
the
basis
for
the
criteria,
and
the
extent
to
which
population
effects
can
be
probabilistically
determined
varies
between
methods.

Experimental
Information­
based
Method
(
EPA)

The
EPA
has
published
temperature
criteria
for
a
number
of
fish
species
based
on
a
review
of
laboratory­
based
research
on
the
thermal
tolerance
of
fish
(
Brungs
and
Jones,
1977,
also
cited
as
U.
S.
E.
P.
A.
1977).
Brungs
and
Jone's
method
includes
identification
of
acute
and
chronic
threshold
values,
definition
of
averaging
time
of
specific
daily
temperature
characteristics,
and
explicit
treatment
of
safety
factors
to
ensure
the
recommended
criteria
control
population
level
effects.
Temperature
criteria
are
based
on
temperature
tolerance
studies
that
generally
follow
the
protocols
developed
by
the
NAS/
NAE
(
1973).
These
protocols
include
procedures
to
derive
specific
temperature
criteria
for
both
chronic
and
acute
exposure.
Criteria
for
chronic
exposure
are
derived
from
incipient
lethal
temperature
and
physiological
(
bioenergetic)
performance
(
e.
g.,
growth
optima)
data.
The
temperature
assessment
methods
described
in
Sections
4,5
and
7­
3
6
of
this
report
are
based
on
similar
data
such
as
the
acute
lethal
and
growth
temperature/
ration
relationships.

From
the
perspective
of
selecting
temperature
criteria,
some
very
simple
principles
can
be
derived
from
the
growth
curves
that
are
meaningful
(
Figure
5.4).
Beginning
with
the
coolest
temperatures
(
0o
C),
growth
increases
with
temperature
up
to
the
temperature
up
to
optimal
where
the
maximum
growth
rate
is
achieved
without
any
increase
in
ration
(
approximately
16o
C).
This
improved
growth
is
due
to
increased
food
conversion
efficiency
and
consumption.
At
temperatures
above
the
maximum
growth
rate,
growth
rates
cannot
be
maintained
because
consumption
declines
and
metabolic
energy
costs
increase.
Further
increases
or
maintenance
of
growth
rate
must
come
from
increased
food,
if
possible
within
satiation
limits.
Because
the
shape
of
growth
curves
is
broad
at
the
maximum,
there
is
little
or
no
negative
effect
of
temperature
several
degrees
above
optimum.

Brungs
and
Jones
(
1977)
describe
an
objective
method
for
developing
threshold
criteria
based
on
optimal
temperature
and
the
range
of
preferred
temperatures
from
laboratory
derived
growth
curves
available
at
the
time.
The
criterion
for
chronic
exposure
is
expressed
as
the
maximum
seasonal
7­
day
moving
average
of
the
daily
mean
temperature.
Brungs
and
Jones
(
1977)
refer
to
this
metric
as
the
"
maximum
weekly
average
temperature"
(
MWAT).
This
value
may
be
derived
for
different
seasons
and
life
stages
(
e.
g.,
summer
rearing
or
fall/
winter
incubation).
The
scientific
rationale
for
using
the
MWAT
as
a
temperature
limit
is
based
on
data
showing
that
moderate
temperature
fluctuations
can
be
tolerated
as
long
as
the
incipient
lethal
temperature
is
not
exceeded
for
long
periods.
The
method
also
assumes
that
optimum
temperatures
are
neither
necessary
nor
realistic
at
all
times
to
maintain
viable
fish
populations
(
NAS/
NAE
1973).

Criteria
for
protection
from
exposure
to
extreme
temperatures
are
based
on
thermal
tolerance
data.
Regression
equations
of
median
survival
times
(
LT50)
(
e.
g.,
Figure
2.2)
are
used
to
predict
the
upper
incipient
lethal
temperature
for
a
24­
hour
exposure,
and
a
2
°
C
safety
factor
is
subtracted
to
derive
a
short­
term
criterion
that
will
provide
100%
survival
(
Brungs
and
Jones
1977).
Since
LT10'
s
and
LT50'
s
are
very
close,
the
safety
factor
is
sufficient
to
preclude
effects
on
any
fraction
of
the
population.

Protocols
for
deriving
criteria
based
on
experimental
temperature
tolerance
studies
have
not
changed
since
being
proposed
by
the
Environmental
Protection
Agency
(
Brungs
and
Jones
1977).
Various
authors
continue
to
recommend
this
method
as
an
option
for
protecting
fish
habitat
(
Armour
1991).

The
acute
and
chronic
temperatures
for
the
EPA
protocol
were
computed
for
the
rearing
stage
of
seven
native
salmonids
in
Washington
using
the
experimental
temperature
tolerance
data
that
could
be
found
in
the
literature
(
Table
6.1).
This
includes
relevant
data
reported
by
the
NAS/
NAE
(
1973)
and
any
newer
data
that
could
be
found.
Original
criteria
reported
by
Brungs
and
Jones
(
1977)
are
shown
for
comparison
with
the
updated
numbers.
Note,
the
latter
results
are
slightly
different
than
criteria
reported
by
Brungs
and
Jones
(
op
cit)
for
the
same
species
because
our
values
are
based
on
data
from
individual
studies
and
not
on
the
average
of
several
studies.
The
results
from
multiple
evaluations
of
the
same
species
produced
similar
results
(
usually
within
1oC),
and
differences
among
salmonid
species
were
relatively
small.
7­
4
Table
7.1
Water
temperature
criteria
(
maximum
weekly
average
temperature
and
acute
exposure
maximums
during
growth
season)
for
salmonids
based
on
the
EPA
methodology
(
Brungs
and
Jones
1977).
Values
reported
by
Brungs
and
Jones
are
shown
for
comparison.

Species
temp.
(
°
C)
lethal
level
(
°
C)
a
intercept
(
a)
slope
(
b)
optimum
Source
(
°
C)
maximum
(
°
C)
data
Oncorhynchus
c
lark
i
13­
23
fluctuating
25.5
18.1515
­
0.5723
­­
(
a)
Golden
1978
18.50
a,
d
cutth
roat
trout
23
25.7
­­
­­
­­
(
b)
G
olden
1978
18.57
b,
d
16
22.6
­­
­­
­­
(
c)
V
igg
and
Koch
1980
24.20
a
constant
­­
­­
­­
15
(
d)
Dwyer
and
K
ram
er
1975
Oncorhynchus
gorbuscha
24
23.9
14.7111
­
0.4459
­­
(
e)
B
rett
1952
17.97
e,
t
pink
salm
on
23.91
e
Oncorhynchus
keta
23
23.8
15.3825
­
0.4721
­­
(
f)
B
rett
1952
17.93
f,
t
chum
salm
on
9
23.2
15.9272
­
0.5575
­­
(
g)
Blaham
and
Parente,
unpublished
in:
NAS/
NAE
1973
17.73
g,
t
23.89
f
20.90
g
Oncorhynchus
k
isutch
23
25.0
18.9736
­
0.6013
­­
(
h)
B
rett
1952
19.67
h,
k
coho
salm
on
23
25.0
(
hh)
Dehart
1975
10
23.5
18.4136
­
0.641
­­
(
i)
B
laham
and
M
cConnel,
unpublished
in:
NAS/
NAE
1973
19.17
I,
k
10­
13
fluctuating
26.0
­­
­­
­­
(
j)
Thom
as
et
al.
1986
20.00
j,
k
constant
­­
­­
­­
17
(
k)
A
verett
1968
24.30
h
21.80
I
Brungs
a
nd
Jones
(
1977)
estimate
18.00
24
.00
Oncorhynchus
m
ykiss
18
26.5
18.4654
­
0.5801
­­
(
l)
A
labaster
and
W
elcom
m
e
1962
20.30
l,
o
rainbow/
steelhead
20
­­
19.625
­
0.625
­­
(
m
)
Alabaster
and
Downing
1966
19.17
l,
p
16
25.6
­­
­­
­­
(
n)
Hokanson
et
a
l.
1977
19.83
l,
q
constant
­­
­­
­­
17.2
(
o)
Hokanson
et
a
l.
1977
20.30
n,
o
fluctuating
­­
­­
­­
15.5
(
p)
Hokanson
et
a
l.
1977
18.87
n,
p
fluctuating
­­
­­
­­
16.5
(
q)
W
urtsbaugh
and
D
avis
1977
19.53
n,
q
24.39
l
24.35
m
Brungs
a
nd
Jones
(
1977)
estimate
19.00
24
.00
Oncorhynchus
nerka
23
24.8
20.002
­
0.6496
­­
(
r)
Brett
1952
18.27
sockeye
salm
on
20
23.5
16.7328
­
0.5473
­­
(
s)
M
cConnel
a
nd
B
lahm
,
u
npublished
in:
NAS/
NAE
1973
17.83
constant
15
(
t)
B
rett
et
a
l.
1969
23.93
15
(
u)
Shelbourn
et
a
l.
1973
22.80
Brungs
a
nd
Jones
(
1977)
estimate
18.00
22
.00
Oncorhynchus
tshawytscha
20
25.1
22.9065
­
0.7611
­­
(
v)
B
rett
1952
21.03
v,
x
chinook
salm
on
20
24.7
21.3981
­
0.7253
­­
(
w)
Blaham
and
M
cConnel,
unpublished
in:
NAS/
NAE
1973
20.90
w,
x
­­
­­
­­
­­
19
(
x)
B
rett
et
a
l.
1982
23.95
v
23.15
w
a
D
ata
represent
the
ultim
ate
upper
incipient
lethal
tem
perature
w
here
available
o
r
the
upper
incipient
lethal
tem
perature
when
an
estim
ate
o
f
the
u
ltim
ate
level
was
not
available.

Median
Survival
Time
Temperature
criteria
7­
5
Differences
between
maximum
and
minimum
MWAT
and
acute
criteria
are
3.3
°
and
3.5
°
C,
respectively
(
Table
6.1).
This
exercise
demonstrates
that
the
EPA
method
is
highly
objective
and
reproducible;
there
were
no
difficulties
in
updating
the
analysis
with
results
of
more
recent
studies.
Brungs
and
Jones
also
provide
criteria
for
other
life
history
stages.

Field
Observation
Methods
Observation
of
temperature
at
which
fish
occurrence
is
verified
is
another
method
used
for
estimating
fish
temperature
requirements.
Bovee
(
1978)
recommended
the
use
of
fish
observations
where
temperature
is
simultaneously
collected
to
determine
a
"
probability
of
use"
curve
based
on
the
distribution
of
observations.
This
technique
of
characterizing
physical
environmental
conditions
in
conjunction
with
fish
observations
forms
the
basis
for
the
U.
S.
Fish
and
Wildlife
Service
Instream
Flow
Incremental
Methodologies
(
IFIM)
Habitat
Suitability
Index
model
(
HSI)
(
McMahon
1983).
This
method
has
been
used
to
evaluate
population
effects
from
physical
habitat
alterations.
Presumably,
this
method
would
reflect
preference
temperatures
of
the
fish,
but
their
quantitative
relationship
to
optimal
temperature,
growth,
or
lethal
temperatures
cannot
be
known
since
results
would
also
strongly
reflect
the
ambient
temperature
occurring
when
observations
were
made
and
may
not
indicate
true
preference
or
long­
term
exposure.

Eaton
et
al.
(
1995)
proposed
a
technique
for
deriving
the
maximum
thermal
tolerance
of
fish
matching
stream
temperature
records
with
fish
presence
data.
Their
fish
and
temperature
data
matching
system
(
FTDMS)
provides
a
direct
measure
of
the
temperatures
that
are
utilized
by
fish
populations
in
nature.
In
this
protocol,
fish
presence
data
are
matched
with
weekly
mean
temperatures
taken
from
the
same
location
and
time
period
to
derive
a
fish
presence
by
temperature
frequency
distribution.
An
estimate
of
the
maximum
temperature
tolerance
for
a
species
is
assumed
to
be
equivalent
to
the
temperature
at
which
95%
of
the
fish
observations
occur
for
a
large
(
n
=
1000
matches)
database
that
represents
the
geographic
range
limits
of
the
species.
The
95th
percentile
observation
is
proposed
as
a
safe
estimate
of
thermal
tolerance
to
protect
against
inaccuracies
in
temperature
records
and
biased
observations
of
fish
presence
that
may
be
contained
in
the
database.

The
FTDMS
is
recommended
as
an
approach
for
determining
the
maximum
temperatures
that
limit
the
distribution
of
salmonids.
With
regard
to
temperature
criteria,
this
method
seems
to
relate
best
to
the
acute
temperatures
rather
than
the
chronic
temperatures.
Table
7.2
lists
the
mean
weekly
temperature
derived
from
over
1000
field
observations
(
Eaton
et.
al
1995).
The
field
observation
method
of
establishing
upper
lethal
limits
suggested
by
Eaton
et
al.
(
1995)
produced
very
similar
results
to
those
of
Brungs
and
Jones
Table
7.2
Water
temperature
criteria
for
salmonids
based
on
the
fish
and
temperature
data
matching
system
(
FTDMS).
Taken
from
Table
1
in
Eaton
et
al.
(
1995).

Species
Mean
weekly
temperature
(
oC)
for
the
95th
percentile
observation
Oncorhynchus
clarki
(
Cutthroat
trout)
23.2
O.
gorbushcha
(
pink
salmon)
21.0
O.
keta
(
chum
salmon)
16.8
O.
kisutch
(
coho
salmon)
23.4
O.
mykiss
(
steelhead/
rainbow
trout)
24.0
O.
tschawytscha
(
chinook
salmon)
24.0
7­
6
(
1977)
which
were
based
on
laboratory
tests
(
Table
7.1):
field
observation
estimates
were
within
1
°
C
of
EPA's
acute
temperature
criteria
for
all
salmonids
except
chum
salmon.
These
observations
suggest
that
laboratory
derived
criteria
can
predict
the
thermal
tolerance
limits
in
nature
reasonably
well.

RISK
ASSESSMENT
APPROACH
The
temperature
assessment
methods
described
in
Sections
5
and
6
followed
a
risk
assessment
approach
and
provided
a
time­
integrated
and
quantitative
estimate
of
the
influence
of
the
temperature
regime
on
the
potential
growth
of
salmon
within
selected
stream
reaches
where
temperature
was
known.
This
approach
could
be
used
to
develop
site­
specific
or
regional
temperature
criteria
for
both
acute
and
chronic
effects
on
summer
rearing.
Here
we
use
the
results
of
the
acute
and
growth
analyses
to
evaluate
generally
applied
temperature
indices
used
as
temperature
criteria.
The
assessment
of
acute
temperatures
suggested
that
26oC
is
threshold
temperature
for
salmonid
species
(
Section
4).

The
reduction
from
maximum
potential
growth
due
to
temperature
regime
was
calculated
for
a
number
of
stream
segments
with
widely
varying
temperature
regimes
in
Section
6,
based
on
a
growth
model
developed
in
Section
5.
The
range
of
temperatures
where
growth
was
within
10%
and
20%
of
optimum
based
on
those
analyses
is
shown
in
Figure
7.1.
For
each
temperature
indices,
coho
and
steelhead
reduction
from
maximum
growth
(
RMG)
are
7­
Day
Maximum
0
5
10
15
20
25
30
Coho
Steelhead
Temperature
(
oC)

10%
20%

20%
Annual
Maximum
0
5
10
15
20
25
30
Coho
Stealhead
Temperature
(
oC)
20%

20%
10%
7­
Day
Mean
0
5
10
15
20
25
30
Coho
Steelhead
Temperature
(
oC)

20%

10%

20%

Figure
7.1
Range
of
temperature
where
reduction
from
maximum
potential
growth
(
RMG,
%)
was
10
and
20%
during
the
summer
months
for
coho
and
steelhead,
using
three
temperature
indices.
(
See
Section
6
for
methods
and
Table
6.2
for
results
by
site.
RMG
is
greater
than
20%
at
temperatures
outside
of
the
temperature
ranges
indicated.
RMG
is
minimized
near
the
optimal
temperature
for
each
species
and
increases
at
temperatures
warmer
or
cooler
than
the
optimal.
7­
7
plotted
together
to
facilitate
comparison
among
species.
Growth
is
highest
within
the
range
of
optimal
temperatures,
and
declines
at
temperature
higher
and
lower
than
optimal.

There
is
a
narrow
range
of
temperature
where
growth
is
optimized
for
each
species.
Growth
rate
is
highest
at
sites
with
7­
day
maximum
temperatures
between
9oC
to
17oC.
Patterns
are
similar
for
other
temperature
indices
(
7­
day
mean
and
annual
maximum
temperature),
although
the
temperature
range
enveloping
various
levels
of
growth
reduction
vary
with
each
temperature
index.

Steelhead
and
coho
often
occur
together
in
stream
environments,
and
their
growth
responses
are
similar,
although
there
are
important
differences
in
threshold
values.
Selecting
the
criteria
based
on
the
10%
RMG
for
the
more
thermally
sensitive
coho
would
suggest
an
upper
threshold
of
16.5oC
for
the
7­
day
maximum
temperature
and
14.8oC
for
the
7­
day
mean
temperature.
Selecting
the
criteria
based
on
the
10%
RMG
for
the
steelhead
would
suggest
an
upper
threshold
of
20.5oC
for
the
7­
day
maximum
temperature
and
17.0oC
for
the
7­
day
mean
temperature.
The
upper
end
of
the
temperature
range
is
well
below
temperatures
where
behavioral
avoidance
has
been
observed
(
e.
g.
Bisson
et
al.
1988,
Nielsen
et
al.
1994),
and
interspecies
competitive
interactions
have
been
noted
(
e.
g.,
Reeves
et
al.
1987;
Taniguchi
et
al.
1998).

The
growth
effects
predicted
by
the
criteria
will
be
the
same
wherever
the
fish
live.
Even
though
streams
for
resident
trout
may
be
naturally
colder
due
to
proximity
to
headwaters,
the
growth/
food/
and
temperature
effects
should
be
the
same.
It
should
also
be
noted
that
these
criteria
are
only
appropriate
for
streams
with
normal
seasonal
and
daily
temperature
fluctuations.
Streams
or
other
aquatic
environments
(
e.
g.,
thermal
plumes
at
discharge
sites)
with
significantly
different
temperature
patterns
would
require
site
specific
analysis,
i.
e.,
characterization
of
their
temperature
regimes
to
determine
exposure.
Table
7.3
Temperature
criteria
(
oC)
for
growth
of
juvenile
salmonids
derived
from
temperature
analysis
at
values
of
reduction
from
maximum
growth
(
RMG)
of
10%
and
20%.

Coho
Steelhead
Temperature
Index
10%
RMG
20%
RMG
10%
RMG
20%
RMG
MWAT
(
Updated
EPA
method
(
Table
6.1)

7­
day
Maximuma
(
oC)
13.0­
16.5
9.0­
20.5
14.5­
20.5
10.0­
24
7­
day
Meanb
(
oC)
12.8­
14.8
9.0­
19.0
13.0­
17.0
10.0­
19.0
19.7
coho
19.6
steelhead
Annual
Maximumc
(
oC)
13.5­
17.5
9.5­
23.0
15.5­
21.0
10.5­
26.0
amaximum
value
of
the
7­
day
moving
average
of
the
daily
maximum
temperature;

bmaximum
value
of
the
7­
day
moving
average
of
the
daily
mean
temperature;

c
instantaneous
maximum
observed
during
the
summer;
7­
8
Review
Approach
Temperature
criteria
derived
by
the
review
approach
are
based
on
the
professional
interpretation
of
temperature
requirement
information
organized
by
life
stages
and
time
periods
for
each
species
of
interest.
These
temperature
requirements
are
derived
primarily
from
key
review
articles
(
e.
g.,
Bjornn
and
Reiser
1991,
Bell
1973,
1986).
Measurements
of
performance
optima
from
laboratory
studies
and
field
observations
of
temperature
during
different
life
stages
may
also
be
used.

Criteria,
for
example,
are
derived
by
selecting
a
temperature
regime
low
enough
to
protect
the
most
sensitive
life
stage
for
the
summer
juvenile
rearing
period.
Protection
of
this
life
stage
is
assumed
to
protect
all
other
life
stages
that
may
occur
at
the
same
time
(
e.
g.,
adult
holding).
The
review
method
is
not
a
defined
protocol,
but
rather
is
a
general
approach
for
evaluating
temperature
information.
Review­
based
approaches
are
inherently
more
subjective
as
analysts
attempt
to
explicitly
synthesize
a
number
of
factors
and
species
into
one
recommended
criteria.

Table
7.4.
Examples
of
four
water
temperature
envelopes
by
life
stage
summaries
for
spring
chinook
salmon.
(
Temperature
in
oC.)

Life
stage
Bell
(
1973)
Bjornn
and
Reiser
(
1991)
Armour
(
1991)
a
ODEQ
(
1995)

Adult
migration
3.3­
13.3
3.3­
13.3
3.3­
13.3
3.3­
13.3
Spawning
5.6­
13.9
5.6­
13.9
5.6­
13.9
5.6­
12.8
Incubation
5.0­
14.4
5.0­
14.4
5.0­
14.4
4.5­
12.8
Juvenile
rearing
Optimum
7.2­
14.4
Preferred
12­
14
7.9­
13.8
Positive
growth
4.5­
19.1
Optimum
production
10.0­
15.6
aAll
data
are
for
the
recommended
temperature
range
Data
linking
fish
performance
and
temperature
are
evaluated
by
a
professional,
or
group
of
professionals,
who
identify
the
temperature
range
that
provides
some
level
of
protection.
While
the
analysis
may
include
laboratory
or
field
derived
data,
the
manner
in
which
such
data
are
used
is
not
explicitly
defined,
as
it
is
in
the
EPA
and
FTDMS
methods
(
e.
g.
ODEQ
1995,
WDOE
1998a).
Also,
the
level
of
protection1
generally
is
not
explicitly
defined
and
appears
to
vary
depending
on
policy
objectives
and
the
amount
of
available
information.

Bell
(
1973)
conducted
one
of
the
first
reviews
of
temperature
to
establish
criteria,
compiling
most
of
the
information
known
at
the
time,
and
presented
the
data
in
the
form
of
temperature
ranges
or
envelopes
by
species
and
life
stage.
Bell
(
1973)
synthesized
a
temperature
range
from
the
available
information
to
provide
a
recommendation.
However,
he
did
not
describe
the
method
by
which
the
recommendations
were
derived,
including
consideration
of
safety
factors,
and
he
did
not
attribute
the
recommendations
with
specific
citations
(
i.
e.,
only
a
list
of
references
is
given).
Thus
the
scientific
source
for
each
recommendation
cannot
be
verified
and
the
primary
data
sources
that
may
have
been
used
are
not
directly
tied
to
the
final
recommendations.

1
Level
of
protection
refers
to
what
percentage
of
the
individuals
representing
a
race,
subspecies
or
species
are
protected
(
e.
g.,
90,
95
99,
100%).
Similarly,
it
refers
to
the
percentage
of
streams
that
would
be
protected,
and
what
percentage
of
the
time.
For
example,
a
goal
may
be
to
protect
95%
of
the
salmonid
races
and
streams
95%
of
the
time.
7­
9
A
series
of
subsequent
reviews
have
relied
heavily
on
the
work
of
Bell
(
1973)
to
develop
and
revise
temperature
criteria.
For
example
both
Armour
(
1991)
and
ODEQ
(
1995)
cite
Bjornn
and
Reiser
(
1991)
as
the
source
for
some
of
their
temperature
recommendations.
Bjornn
and
Reiser
(
op
cit)
reference
Bell
(
1986),
which
is
the
second
edition
of
Bell
(
1973),
as
source
for
their
temperature
information.
The
Bell
(
1973)
report
is
also
the
basis
for
U.
S.
Fish
and
Wildlife
Service
recommended
temperature
criteria
for
coho
salmon
(
Laufle
et
al.
1986),
chinook
salmon
(
Allen
and
Hassler
1986),
and
steelhead
trout
(
Pauley
et
al.
1986).
These
references
are
also
cited
for
temperature
criteria
in
the
"
ManTech
Report"
sponsored
by
the
National
Marine
Fisheries
Service
(
Spence
et
al.
1996).
The
interdependence
of
these
review
reports
results
in
recommended
temperature
criteria
that
are
remarkably
similar.
For
example,
chinook
criteria
shown
in
Table
7.4
are
the
same
in
each
of
four
different
papers.

This
review
approach
has
formed
the
basis
for
temperature
criteria
in
the
Pacific
Northwest
in
recent
years,
despite
the
lack
of
firmly
documented
primary
data
(
ODEQ
1995,
WDOE
1999).
Criteria
developed
in
Oregon
(
ODEQ
1995)
and
proposed
in
Washington
(
WDOE
1999a)
appear
to
be
based
on
less
quantitative
approaches
than
advocated
by
Hokanson
(
1977),
Brungs
and
Jones
(
1977),
Armour
(
1991)
and
Eaton
et
al.
(
1995),
although
experimental
biological
effects
data
are
available
for
this
purpose.
The
primary
weaknesses
of
review
approaches
are
the
absence
of
a
clearly
defined
decision
process
for
selecting
and
evaluating
temperature
information,
synthesizing
factors
of
safety
and
uncertainty
into
the
criteria,
and
lack
of
clear
linkages
to
field
and
laboratory
data.
In
some
cases,
the
numbers
for
juvenile
rearing
derived
from
literature
reviews
are
consistent
with
those
produced
from
risk
analysis
and
other
quantitative
methods
(
e.
g.,
Bjornn
and
Reiser
1991).
However,
in
most
cases,
the
recommended
ranges
assume
greater
growth
at
lower
temperatures
than
is
likely
to
occur
and
less
growth
at
warmer
temperatures.

Temperature
Criteria
Existing
and
proposed
temperature
criteria,
including
objective
criteria
and
those
that
were
derived
primarily
by
the
review
method
described
above,
are
presented
in
Tables
7.5
and
7.6.
Only
criteria
relevant
to
the
growth
period
of
juvenile
salmon
and
trout,
exclusive
of
bull
trout,
are
shown
for
this
example.
The
Oregon
temperature
criteria
combine
trout
and
salmon
species
together,
with
a
different
standard
for
bull
trout.
The
proposed
Washington
criteria
group
salmon
species
and
steelhead
with
one
criteria,
and
cutthroat
trout
for
another.
Washington's
proposed
criteria
also
vary
the
maximum
temperature
by
specified
periods
during
the
summer
months,
a
detail
we
will
not
address
during
subsequent
analyses.
7­
10
Each
criterion/
standard
in
Tables
7.5
and
7.6
typically
is
assumed
to
represent
the
noeffect
level
for
the
most
sensitive
life
stage
of
the
most
sensitive
species,
plus
a
safety
factor.
The
EPA's
goal
is
to
protect
95%
of
the
species
95%
of
the
time.
This
level
of
protection
is
extended
to
include
economically
important
species,
ecological
keystone
species,
and
threatened
and
endangered
species
(
Stephan
et
al.
1985).
Other
groups
of
scientists
have
suggested
that
protecting
90%
of
the
species
will
protect
aquatic
communities
(
SETAC
1994).
It
is
not
always
necessary
to
protect
95%
of
the
individuals
in
a
population
when
it
is
desired
to
protect
the
species,
based
on
rationales
presented
by
Ricker
(
1975).
However,
in
the
case
of
the
ODEQ
and
WDOE
criteria,
the
actual
level
of
protection
embodied
in
the
criteria
has
not
been
defined.

Temperature
criteria
from
the
various
methods
have
the
following
similarities
and
differences
(
Tables
7.5
and
7.6):

·
 
Most
index
the
juvenile
growth
phase,
which
lasts
several
months,
with
the
warmest
7­
day
period
occurring
during
that
interval.

·
 
No
criteria
use
an
averaging
period
longer
than
a
week.
One
specifies
the
annual
instantaneous
maximum
(
e.
g.,
Washington's
existing
temperature
criteria).

·
 
Criteria
vary
in
whether
the
daily
maximum
or
mean
is
used
to
calculate
temperature
during
the
averaging
period.

·
 
No
criteria
state
an
acceptable
level
of
variation
in
the
threshold
temperature,
indicating
that
are
likely
to
result
from
natural
factors
and
uncertainty.
Table
7.5
Existing
and
proposed
temperature
criteria
for
anadromous
salmon
species
and
steelhead
derived
from
various
methodologies
relevant
to
the
summer
growth
period.
Numbers
are
maximum
allowable
values.

ACUTE
SUB­
LETHAL
Method
Temperature
(
oC)
Metric
Temperature
(
oC)
Metric
EPA
(
Brungs
and
Jones
1977)
24.0
Annual
instantaneous
maximum
a
18.0
Maximum
7­
day
moving
average
of
the
daily
mean
Eaton
et.
al.
(
1995)
23.5
Maximum
7­
day
moving
average
of
the
daily
mean
­­
­­

Risk
Assessment
(
this
report)
25.5
Annual
instantaneous
maximum
a
16.5
Maximum
7­
day
moving
average
of
the
daily
maximum
ODEQ
(
1995)
None
specified
­­
17.8
(
64oF)
Maximum
7­
day
moving
average
of
the
daily
maximum
WDOE
proposed
(
1999)
21.0
(
June­
Sept)
Annual
instantaneous
maximum
16.5
Maximum
7­
day
moving
average
of
the
daily
maximum
WDOE
(
current)
None
specified
­­
16.0
(
AA)
b
18.0
(
A)
21.0
(
B)
Annual
instantaneous
maximum
a
assumed
at
least
a
1­
hour
interval
b
streams
are
classified
as
AA,
B
and
C
according
to
WAC
173­
201­
080
7­
11
·
 
Some
criteria
do
not
specify
upper
acute
temperature
levels,
relying
instead
on
the
temperature
criteria
derived
for
chronic
effects
to
control
maximum
temperature.

·
 
For
most
of
the
criteria,
the
temperature
range
of
16o
to
18oC
is
used
as
the
upper
maximum.

·
 
No
criteria
establish
a
minimum
threshold
temperature.

The
temperature
criteria
derived
from
the
review
approach
are
more
variable,
although
much
of
the
data
used
are
similar
among
the
evaluations
(
ODEQ,
1995
WDOE,
1999).
The
Oregon
and
proposed
Washington
criteria
are
reasonably
similar
to
the
EPA
criteria
for
salmonids
(
Tables
7.5
and
7.6).

The
differences
in
the
temperature
indices
used
by
different
sources
makes
it
difficult
to
compare
them.
In
Table
7.7
we
translate
each
recommended
threshold
to
each
of
the
short­
term
indices,
using
the
relationships
between
indices
shown
in
Figure
3.7
(
Section
3).
When
placed
on
a
common
footing,
it
is
evident
that
there
are
differences
among
the
recommended
criteria.

Table
7.6
Existing
and
proposed
temperature
criteria
cutthroat
trout
derived
from
various
methodologies
relevant
to
the
summer
growth
period.
Numbers
are
maximum
allowable
values.

ACUTE
SUB­
LETHAL
Method
Temperature
(
oC)
Metric
Temperature
(
oC)
Metric
EPA
(
Brungs
and
Jones
1977)
24.2
Annual
instantaneous
maximum
a
18.5
Maximum
7­
day
moving
average
of
the
daily
mean
Eaton
et.
al.
(
1995)
23.2
Maximum
7­
day
moving
average
of
the
daily
Mean
­­
­­

Risk
Assessment
(
this
report)
26.3
Annual
instantaneous
maximum
a
16.5
Maximum
7­
day
moving
average
of
the
daily
maximum
ODEQ
(
1995)
None
specified
­­
17.8
(
64oF)
Maximum
7­
day
moving
average
of
the
daily
maximum
WDOE
proposed
(
1999)
14.5
Annual
instantaneous
maximuma
13.0
Maximum
7­
day
moving
average
of
the
daily
maximum
WDOE
(
current)
None
specified
­­
16.0
(
AA)
b
18.0
(
A)
21.0
(
B)
Annual
instantaneous
maximum
a
assumed
at
least
a
1­
hour
interval
b
streams
are
classified
as
AA,
B
and
C
according
to
WAC
173­
201­
080
7­
12
Table
7.7
Threshold
temperatures
for
short­
term
duration
indices
from
various
sources.
Temperatures
have
been
translated
to
common
values
using
relationships
among
temperature
indices
developed
in
Section
3
(
see
Figure
3.7).
Values
in
bold
type
are
original
reported
numbers.
Authors
report
recommendations
in
a
variety
of
metrics.
Each
recommended
value
is
also
translated
to
the
other
metrics
using
the
regression
relationships
presented
in
Section
3
to
facilitate
their
comparison.

Species
Reference
Sub­
Lethal
Thresholds
Acute
Threshold­­

Annual
Maximum
Temperature
(
oC)
7­
Day
Maximum
(
oC)
7­
Day
Mean
(
MWAT)
(
oC)
Annual
Maximum
Temperature
(
oC)

Coho
salmon
EPA
1977
21.5
21.2
18.0
24.9
Risk
assessment
(
this
report)
17.5
16.5
14.8
25.5
ODEQ
(
1995)
19.0
17.8
16.0
­­

WDOE
(
existing)
16.0
15.5
14.3
­­

WDOE
(
proposed)
17.0
16.5
15.0
21.0
Eaton
(
1995)
­­
­­
23.4
30.5
Steelhead
trout
EPA
1977
24.0
23.0
19.0
26.0
Risk
assessment
(
this
report)
21.0
20.5
17.0
26.0
ODEQ
(
1995)
19.0
17.8
16.0
­­

WDOE
(
existing)
16.0
15.5
14.3
­­

WDOE
(
proposed)
17.0
16.5
15.0
21.0
Eaton
(
1995)
­­
­­
24.0
31.0
The
temperature
analysis
developed
in
this
report
produced
similar
though
not
identical
criteria
to
those
developed
using
a
variety
of
other
methods.
The
thresholds
derived
from
the
risk
assessment
methods
are
somewhat
lower
than
the
EPA
recommendations
(
Brungs
and
Jones
1977),
largely
because
of
the
restriction
to
10%
growth
loss
and
the
realistic
accounting
of
food
consumption.
Brungs
and
Jones
(
1977)
used
the
7­
day
mean
temperature
(
MWAT)
of
18oC
for
coho
and
19oC
for
steelhead.
The
growth
analysis
suggests
that
an
upper
threshold
for
the
7­
day
mean
temperature
of
14.8oC
for
coho
and
17.0oC
for
steelhead
will
maintain
growth
within
10%
of
optimum,
and
19oC
will
maintain
growth
within
20%
of
optimum.

Eaton
et
al.
(
1995)
primarily
studied
the
upper
temperature
limiting
salmonid
distribution
but
not
growth.
Their
numbers
are
several
degrees
higher
than
our
recommendation
for
acute
thresholds.
Indeed,
their
thresholds
are
so
high
that
it
would
suggest
that
coho
and
steelhead
still
live
in
natural
streams
until
maximum
temperatures
reach
upper
critical
lethal
levels
for
a
significant
period
of
time
(
at
least
1
week).
This
appears
to
confirm
that
the
thresholds
we
have
identified
are
conservative
and
not
likely
to
result
in
population
loss.
No
site
included
in
the
risk
assessment
had
7­
day
mean
temperature
anywhere
close
to
the
upper
threshold
that
limits
distribution,
although
this
was
not
true
for
some
rivers
in
the
region
found
in
U.
S.
G.
S.
water
resources
records
(
see
Appendix
B).
Coho,
steelhead
7­
13
and/
or
cutthroat
trout
were
present
in
all
of
these
streams.
We
estimate
that
growth
loss
due
to
temperature
for
Eaton's
(
1995)
upper
limits
for
coho
and
steelhead
would
be
approximately
50%.

Temperature
criteria
derived
with
the
risk­
based
methods
have
only
moderate
agreement
with
criteria
derived
through
various
reviews.
It
appears
that
the
review
approach
tends
to
recommend
similar
temperatures
for
the
lower
end
of
the
range
but
lower
temperatures
for
the
upper
end
of
the
range,
than
was
found
by
risk
analysis
and
other
methods.
While
the
example
ranges
we
cite
are
for
chinook
salmon,
the
temperature
response
curves
for
coho
(
this
report)
are
very
similar
to
those
of
chinook
(
Brett
et
al.
1982).
It
is
difficult
to
directly
compare
with
recommendations
from
these
reviews
since
no
indexing
temperature
measure
is
provided.

ODEQ
(
1995)
criteria
appear
to
match
results
of
the
growth
assessment
reasonably
well,
despite
its
reliance
on
reviews
rather
than
laboratory
data.
ODEQ
(
1995)
specifies
the
maximum
7­
day
temperature
at
17.8oC.
The
7­
day
maximum
criteria
derived
for
coho
are
16.5oC
for
10%
growth
reduction
and
approximately
19oC
for
20%.
However,
if
just
steelhead
were
considered,
the
threshold
would
be
20.5oC
for
10%
growth
reduction.

Current
WDOE
criteria
specifying
an
annual
maximum
of
16oC
for
Class
AA
streams
are
lower
than
that
derived
from
the
risk
assessment
approach.
Risk
assessment
suggests
the
annual
maximum
should
be
between
13.5oC
and
17.5oC
for
coho,
or
between
15.5
oC
and
21.0oC
for
steelhead,
to
maintain
no
more
than
10%
growth
loss.
The
current
criteria
for
Class
A
(<
18oC)
and
Class
B
(
21oC)
streams
is
more
comparable
to
risk
assessment
results.
The
proposed
temperature
criteria
published
as
a
discussion
draft
by
WDOE
(
1999)
are
very
similar
to
those
derived
with
the
growth
assessment
for
10%
growth
loss
for
coho;
16.5oC
7­
day
maximum
for
all
anadromous
salmon
rearing.
This
criteria
is
lower
than
needed
for
steelhead.

Discussion
The
quantitative
analysis
confirmed
that
biologically
meaningful
temperature
thresholds
could
be
identified
with
and
of
the
indices
(
annual
maximum,
7­
day
maximum
or
7­
day
mean).
There
is
no
consensus
on
what
index
to
use
for
temperature
criteria,
introducing
additional
confusion
in
comparing
among
them.
This
study
found
that
all
of
the
most
typical
indices
are
closely
related
to
one
another,
and
that
any
could
be
used
with
satisfactory
results.
The
7­
day
mean
temperature
was
most
closely
correlated
with
growth
loss
estimates
and
therefore
may
be
the
best
indexing
measure
for
this
purpose.
However,
other
measure
are
quite
suitable.
It
is
important
that
the
selected
temperature
match
the
time­
averaging
period
appropriately.
It
should
be
noted
that
the
longer
the
averaging
period,
the
lower
the
threshold
value.

The
growth
analysis
developed
in
this
paper
can
form
a
basis
for
selecting
temperature
criteria,
but
some
other
methods
also
were
reproducible
and
produced
similar
though
not
identical
results.
The
risk
assessment
results
described
in
this
report
rely
on
similar
laboratory
data
as
used
by
Brungs
and
Jones
in
developing
EPA
recommendations
(
1977).
Our
results
suggest
lower
criteria
by
a
few
degrees,
primarily
because
we
use
observed
temperature
regimes
to
estimate
the
growth
of
fish
over
the
long­
term,
and
because
we
account
for
realistic
estimates
of
food
consumption.
Analyzing
temperature
relative
to
duration
did
affect
the
choice
of
thresholds.
In
the
case
of
coho,
temperature
thresholds
7­
14
were
lowered,
while
in
the
case
of
steelhead
the
thresholds
were
similar,
and
could
possibly
be
raised.

CONCLUSIONS
‰
Risk
assessment­
based
approaches
allows
the
effects
of
magnitude,
duration
and
frequency
of
temperature
on
fish
growth
and
survival
to
be
quantified
in
an
objective
and
repeatable
manner.

‰
Moderate
temperatures
are
likely
to
be
more
biologically
productive
for
salmonid
species
than
very
warm
or
cold
temperatures
at
the
level
of
food
availability
that
appears
to
exist
in
streams.

‰
The
U.
S.
EPA
(
1977)
temperature
criteria
were
found
to
be
the
most
objectively
defined
and
consistent
with
risk
assessment
results.
They
generally
appear
to
allow
up
to
18%
reduction
in
coho
growth
due
to
temperature.
8­
1
SECTION
8
A
DISCUSSION
OF
THE
SCIENTIFIC
AND
MANAGEMENT
IMPLICATIONS
OF
FINDINGS
Introduction
In
this
section
we
review
the
information
presented
in
the
report
and
identify
the
scientific
and
management
implications
of
the
results.
This
section
serves
as
both
a
summary
of
key
findings
and
a
synthesis
of
information
for
scientists
and
policy­
makers.

For
ease
of
reading,
we
conduct
this
discussion
without
extensive
referencing
from
within
the
report
or
from
external
documents
or
the
scientific
literature.
Although
these
sources
of
information
are
critical
for
the
context
of
this
discussion,
they
have
been
described
and
referenced
in
detail
in
the
main
body
of
the
report.
We
have
included
several
key
figures
found
in
previous
chapters.

The
Regulatory
Context
of
Temperature
Criteria
The
Clean
Water
Act
requires
states
to
protect
the
public's
values
for
water
bodies.
To
administer
the
CWA,
the
state
water
quality
agencies
must:

·
 
assign
beneficial
uses
to
each
water
body
(
e.
g.,
fishable,
swimmable,
aquatic
life),
·
 
specify
water
quality
criteria
that
are
sufficient
to
protect
the
designated
beneficial
uses,
·
 
assess
and
report
on
the
condition
of
water
bodies
relative
to
those
criteria
(
305b),
·
 
identify
the
sources
of
pollutants,
·
 
develop
various
management
steps
to
protect
or
restore
water
body
conditions
to
meet
criteria,
·
 
monitor
the
water
quality
on
an
ongoing
basis.

The
type
of
regulatory
activities
and
management
restrictions
that
may
be
imposed
depends
on
the
current
and
projected
condition
of
the
water
body
relative
to
the
criteria.
Therefore,
the
water
quality
criteria
have
enormous
legal
and
economic
meaning,
and
their
appropriateness
is
of
great
concern
to
the
public,
scientists,
and
regulators.

States
have
specified
fish
species
in
the
cold
water
guild
(
salmon
and
char)
as
the
designated
beneficial
uses
in
many
streams
and
rivers
of
the
Pacific
Northwest
region.
Water
temperature
plays
a
role
in
virtually
every
aspect
of
fish
life,
and
adverse
levels
of
temperature
can
affect
behavior
(
e.
g.
feeding
patterns
or
the
timing
of
migration),
growth,
and
vitality.
Fish
have
ranges
of
temperature
wherein
all
of
these
functions
operate
normally
contributing
to
their
health
and
reproductive
success.
Outside
of
the
range,
these
functions
may
be
partially
or
fully
impaired,
manifesting
in
a
variety
of
internal
and
externally
visible
symptoms.
Fish
have
a
number
of
physiologic
and
behavioral
mechanisms
that
enable
them
to
resist
adverse
effects
of
temporary
excursions
into
temperatures
that
are
outside
of
their
preferred
or
optimal
range.
However,
high
or
low
temperatures
of
sufficient
magnitude,
if
exceeded
for
sufficient
duration,
can
exceed
their
ability
to
physiologically
adapt
and
can
cause
growth
or
weight
loss,
disease,
competitive
8­
2
Effects
of
Temperature
on
Salmonids
10
15
20
25
30
35
Minutes
Hours
Days
Weeks
Duration
Temperature
(
oC)
Upper
Critical
Lethal
Limit
Zone
of
Resistance
Mortality
can
occur
in
proportion
to
length
of
exposure
Behavioral
adjustment
(
no
grow
th,
no
mortality)
Zone
of
Tolerance
Zone
of
Preference
Grow
th
response
depends
entirely
on
food
availability
Optimal
grow
th
at
all
but
starvation
ration
Reduced
grow
th
Rapid
death
Figure
8.1
The
general
biological
effects
of
temperature
on
salmonids.
Effects
depend
on
the
level
of
temperature
and
its
duration.
Two
important
elements
of
temperature
affect
the
growth
and
survival
of
fish:
1)
the
relationship
between
temperature,
metabolism
and
food
conversion
efficiency
as
it
controls
growth
over
relatively
long
periods,
and
2)
the
thermal
tolerance
of
fish
to
lethal
temperatures
over
relatively
short
periods.
Fish
have
various
physiological
and
behavioral
mechanisms
that
enable
them
to
resist
adverse
effects
of
temporary
excursions
into
temperatures
outside
of
their
preferred
or
optimal
growth
range
but
within
the
upper
critical
lethal
limit.
displacement
by
species
better
adapted
to
the
prevailing
temperature,
or
even
death.
Fish
are
adapted
over
some
evolutionary
time
frame
to
the
prevailing
water
temperatures,
and
climatic
gradient
is
among
the
primary
factors
that
determine
the
extent
of
a
species'
geographic
distribution.

States
in
the
Pacific
Northwest
have
authorized
numeric
temperature
criteria
to
protect
fish
for
nearly
three
decades.
Many
are
currently
reviewing
their
criteria
in
the
triennial
review
process
specified
in
the
Clean
Water
Act,
with
the
help
of
the
U.
S.
EPA,
who
must
also
approve
them.
Many
streams
and
rivers
in
the
Pacific
Northwest
region
have
been
identified
by
the
states
as
exceeding
their
current
water
quality
criteria
according
to
their
305b
reports
to
Congress,
with
a
large
number
of
them
listed
for
temperature
impairment
(
303d
lists).
To
further
add
to
public
concern,
specific
species
or
stocks
within
the
salmon
and
char
genera
have
also
been
listed
as
threatened
or
endangered
under
the
Endangered
Species
Act
in
a
number
of
geographic
locations.
The
ESA
does
not
require
identification
of
a
specific
"
pollutant"
causing
population
decline,
but
high
levels
of
stream
temperature
during
the
summer
months
are
widely
viewed
as
one
of
the
primary
habitat
conditions
that
contributes
to
the
decline
of
cold
water
fish
species
in
freshwater
habitats.

The
Characteristics
of
Temperature
Criteria
Temperature
criteria
typically
have
two
key
elements
 
a
threshold
temperature
that
signals
when
adverse
biological
response
is
likely
to
occur,
and
an
averaging
period
that
indexes
the
duration
of
exposure
likely
to
trigger
that
response.
The
combination
of
the
threshold
temperature
and
the
duration
of
exposure
to
that
temperature
are
an
expression
of
the
risk
imposed
by
the
environmental
temperature
to
the
targeted
fish
species.
Because
of
the
ability
of
fish
to
acclimate
and
adapt
to
temperatures
outside
of
their
optimal
conditions,
both
exposure
duration
and
magnitude
of
temperature
are
necessary
to
determine
the
degree
of
risk
that
this
fluctuation
may
pose
to
each
species.
Environmental
temperature
can
annually
range
over
the
entire
spectrum
from
optimal
to
adverse
in
response
to
the
cyclic
movement
of
the
sun.
Exposure
to
temperature
is
mediated
by
8­
3
stream,
geographic,
and
riparian
forest
characteristics.
A
factor
of
safety
is
typically
added
when
selecting
numeric
criteria
to
account
for
the
uncertainties
in
knowledge
associated
with
each
of
these
elements
and
any
factors
that
are
not
unaccounted
for.

The
averaging
period
has
typically
been
either
the
annual
maximum
temperature
(
observed
for
a
period
as
short
as
an
hour,
but
more
probably
occurring
for
several
hours
on
sequential
days),
or
a
weekly
average
(
generally
focused
on
the
warmest
seven
consecutive
days)
observed
for
the
year.
For
example,
Washington's
current
criteria
specify
the
annual
maximum
temperature,
expressed
as
the
maximum
hourly
temperature
that
occurs
each
year.
Oregon
specifies
the
average
of
the
daily
maximum
temperature
of
the
7
warmest
consecutive
days.
The
U.
S.
EPA
(
1977)
recommends
the
average
of
the
daily
mean
temperature
of
the
7
warmest
consecutive
days
(
MWAT).
Some
have
also
argued
that
the
daily
temperature
fluctuation
should
also
be
accounted
for,
but
this
characteristic
has
not
been
widely
specified
in
states'
criteria.

The
Basis
for
Derivation
of
Temperature
Criteria
A
number
of
different
approaches
have
been
used
to
develop
and
justify
the
temperature
criteria
that
are
currently
widely
used
in
the
Pacific
Northwest
region.
All
draw
upon
a
large
body
of
scientific
research
focused
on
the
thermal
tolerance
of
fish.
There
has
been
considerable
laboratory
testing
for
many
fish
species,
including
salmonids,
beginning
early
in
the
1900'
s
and
continuing
today.
Much
of
the
available
research
on
temperature
tolerances
was
performed
prior
to
1980
and
was
stimulated
principally
by
the
need
to
assess
the
impact
of
heated
effluent
from
power
plants,
dams
and
other
facilities.
Since
that
time,
the
research
focus
has
been
to
add
species
and
refine
the
understanding
of
contributing
factors
such
as
the
effect
of
acclimation
temperatures,
daily
diurnal
temperature
fluctuations,
and
food
rations,
and
to
enhance
understanding
of
the
interaction
of
temperature
with
other
pollutants.
A
considerable
amount
of
the
available
research
has
been
performed
in
the
laboratory
setting.
Ecological
field
studies
have
lagged
behind
laboratory
work,
although
their
application
has
increased
in
recent
years.

Various
methods
have
been
used
to
analyze
temperature
effects
on
fish
to
develop
criteria.
The
methods
vary
in
terms
of
degree
of
objectivity
or
subjectivity
by
which
the
information
is
synthesized
into
recommended
criteria,
the
degree
to
which
data
forms
the
basis
for
the
criteria,
and
the
extent
to
which
population
effects
can
be
probabilistically
determined.
The
temperature
criteria
in
use
in
Pacific
Northwest
states
have
largely
been
drawn
from
professionals'
review
and
interpretation
of
available
scientific
literature
(
e.
g.
ODEQ
1995).
There
also
has
been
some
effort
to
use
the
more
well­
established
scientific
relationships
to
synthesize
objective
analyses
of
threshold
temperatures
and
the
duration
of
exposure
(
e.
g.,
U.
S.
EPA
1977).
This
approach
has
not
been
widely
integrated
into
regulatory
activity.

In
recent
years,
the
EPA
and
the
National
Academies
of
Science
and
Engineering
have
promoted
risk
assessment
techniques
to
develop
water
quality
criteria,
including
protocols
that
have
been
peer­
reviewed
nationally.
Risk
assessment
is
designed
to
enhance
understanding
of
the
potential
adverse
effects
of
a
pollutant
on
a
species
by
combining
the
information
from
biological
studies
with
an
analysis
of
each
population's
potential
exposure
to
those
effects.
These
methods
are
formal,
objective,
and
analytical.
They
have
been
primarily
applied
to
contaminant
pollutants;
guidance
for
other
pollutants
is
still
under
development.
Risk
assessment
can
lead
to
site
or
season­
specific
criteria.
8­
4
To
date,
naturally
occurring
"
pollutants",
such
as
water
temperature,
have
not
been
addressed
with
risk
assessment
techniques
to
determine
criteria.
Instead,
temperature
criteria
generally
are
simple
indices
that
summarize
the
seasonal
and
diurnal
range
of
temperature
observed
in
natural
streams
into
the
averaging
period,
and
that
address
the
complex
array
of
biological
responses
of
all
of
the
life
functions
with
the
temperature
threshold
value.
Although
there
is
a
general
interest
in
tailoring
criteria
to
specific
life
functions
at
specific
times
of
year,
this
has
not
been
widely
accomplished
to
date.
In
most
states,
simple
numeric
indices
are
applied
over
broad
regions
to
primarily
address
the
high
temperatures
that
may
occur
during
the
warm
summer
months,
targeting
the
most
sensitive
species
that
are
likely
to
occur
in
the
water
body.
The
methods
described
earlier
fall
short
of
the
objectively
rigorous
expectations
of
formal
risk
assessment.

Criteria
selection
teams
are
faced
with
a
challenge.
It
is
difficult
to
match
simple
criteria
to
multi­
species
communities
dwelling
in
streams
and
rivers
whose
temperatures
naturally
vary
with
position
in
watershed
and
climate.
Specifying
the
wrong
criteria
could
have
negative,
possibly
catastrophic,
biological
consequences.
At
the
same
time,
the
need
for
management
solutions
that
may
accompany
even
small
changes
in
criteria
can
have
large
economic
and
legal
consequences.
In
addition,
all
approaches
to
developing
biologically
meaningful
temperature
criteria
face
significant
technical
challenges.
Some
of
these
stem
from
the
selection
process
itself.
Subjective
reviews
often
lack
a
clearly
defined
decision
process
for
selecting
and
evaluating
temperature
information,
and
they
fail
to
establish
a
clear
linkage
between
field
and
laboratory
data.
Furthermore,
subjective
evaluations
often
use
unquantified
safety
and
uncertainty
factors.

The
more
that
scientific
research
can
be
used
to
quantitatively
assess
the
extent
that
risk
to
fish
is
minimized,
such
as
those
promoted
as
risk
assessment
techniques,
the
more
confident
the
public
and
regulators
can
be
that
temperature
criteria
are
protective.
Such
confidence
does
not
currently
exist.
Over
the
past
25
years
since
temperature
criteria
were
first
adopted,
there
has
been
considerable
debate
over
them
but
little
scientific
experimentation
to
validate
or
improve
them.
While
the
subjective
analyses
that
form
the
basis
of
current
temperature
criteria
are
apparently
consistent
with
the
scientific
literature,
they
have
failed
to
generate
measurable
hypotheses
that
can
be
scientifically
tested
and
rejected.

The
objective
of
this
report
was
to
synthesize
relevant
temperature
research
and
to
develop
quantitative
risk
assessment
techniques
that
could
be
objectively
applied
to
natural
streams
1)
to
identify
the
risks
posed
by
ambient
temperature
and
to
suggest
temperature
criteria,
and
2)
to
formulate
experimentally
testable
hypotheses.
Analysis
focused
on
the
summer
rearing
phase
of
juvenile
salmonids
because
most
existing
temperature
criteria
target
annual
maximum
temperatures,
salmonids
are
of
primary
interest
in
much
of
the
Pacific
Northwest
region,
and
there
is
a
rich
history
of
laboratory
experimentation
available
to
draw
from.

Quantitative
Analyses
to
Assess
the
Effects
of
Temperature
on
Fish
in
Natural
Environments
The
conceptual
approach
that
frames
this
report
is
that
temperature
is
a
fundamental
component
of
fish
habitat.
Water
temperature
is
the
thermostat
that
controls
energy
intake
and
expenditure.
The
overall
success
of
individual
fish
is
partially
a
result
of
the
cumulative
effect
of
its
environmental
temperature
on
its
ability
to
grow
and
survive
over
8­
5
time.
If
energy
intake
is
adequate
to
fuel
the
physiological
energy
consumption,
mediated
in
large
part
by
the
environmental
temperature,
then
the
organism
can
live
in
a
healthy
state.
The
individual
is
not
likely
to
be
healthy
if
the
water
temperatures
force
energy
consumption
at
a
pace
that
cannot
be
sustained
by
food
intake,
dictated
in
part
by
appetite
and
in
part
by
food
supply,
for
long
periods
of
time.
If
the
duration
of
moderately
negative
temperatures
is
fairly
short,
cessation
of
feeding
or
refuge
seeking
by
the
individual
fish
may
be
sufficient
to
withstand
short­
term
excursions
into
higher
temperatures.
If
this
continues
for
long,
the
fish
loses
growth
opportunity,
and
may
be
displaced
by
competitors
in
the
population.
If
temperatures
reach
a
more
severe
level
of
impairment,
it
creates
physiological
stress,
loss
of
appetite,
and
can
leave
the
fish
open
to
disease
and
competitive
pressures
from
other
species.
Stress
is
exacerbated
at
high
temperature
because
dissolved
oxygen
content
of
the
water
is
inversely
related
to
its
temperature.
If
temperatures
reach
very
high
levels,
it
invokes
significant
stress
that
causes
immediate
death.
For
salmonids,
this
temperature
occurs
at
approximately
30oC.
Low
temperatures
can
also
induce
cessation
of
feeding,
but
unless
water
freezes,
the
fish
also
can
withstand
excursions
into
cold
water
temperature
by
limiting
activity.
Temperature
is
not
the
only
ecological
factor
of
importance
to
biologic
productivity,
but
if
its
central
effect
on
the
individual
can
be
accounted
for,
then
the
influence
of
other
environmental
factors,
such
as
food
supply
and
population
dynamics,
may
become
more
apparent
in
the
complex
ecology
of
natural
environments.

We
work
from
the
assumption
that
there
is
a
continuum
of
biologic
response
to
temperature
that
ranges
from
healthy,
as
indicated
by
maximum
growth,
to
unhealthy,
culminating
in
direct
mortality.
Along
this
spectrum
there
are
a
variety
of
ways
that
temperature
effects
manifest
in
the
organism's
physiologic
condition
or
its
behavior.
Some
of
these
characteristics
can
be
readily
observed
in
natural
environments
and
are
simultaneously
amenable
to
quantification
and
prediction
based
on
measured
temperature.
These
include
the
growth
(
weight
change),
direct
mortality,
and
embryo
development
(
not
addressed
in
this
report).
These
functions
lend
themselves
to
mathematical
expressions,
many
of
which
have
already
been
established
for
many
species,
including
salmonids.
Laboratory
studies
have
shown
that
activity
rates
are
closely
correlated
with
temperature
and
that
they
can
be
predicted
with
some
precision
with
linear
or
non­
linear
equations.

We
were
able
to
develop
models
for
direct
mortality
and
growth
as
a
function
of
temperature
for
several
species
of
salmonids
using
laboratory
data
and
bioenergetic
principles
available
in
the
scientific
literature
(
Figure
8.2).
An
extensive
portion
of
this
report
describes
and
corroborates
these
models.
The
relationships
were
formulated
in
a
way
that
they
can
be
applied
in
natural
ecological
settings.
The
growth
model
simulates
weight
gain
over
time
in
relation
to
daily
temperature
and
food
supply.
Its
formulation
constitutes
a
new
contribution
to
modeling
fish
biological
response
to
habitat
factors.

Importantly,
the
models
predict
qualities
that
allow
them
to
be
corroborated
against
measurable
population
characteristics,
and
therefore
they
produce
hypotheses
that
can
be
rejected
by
direct
observation.
The
methods
are
objective
and
repeatable.
Our
comparisons
of
simulated
growth
(
or
more
specifically
weight
gain)
of
21
populations
living
in
natural
streams
showed
consistent
and
close
agreement
with
observed
weight
characteristics
(
Figure
8.3).

We
were
not
able
to
fully
corroborate
the
acute
temperature
model
because,
when
temperature
records
available
to
us
were
scanned
for
occurrence
of
combinations
of
temperature
and
duration
sufficient
to
cause
mortality,
none
were
found.
The
data
8­
6
Figure
8.2
The
acute
and
chronic
effects
of
temperature
species
have
been
quantified
for
a
number
of
salmonid
species.
Direct
mortality
in
relation
to
exposure
time
for
4
species
is
shown
in
A.
If
the
combination
of
temperature
and
continuous
duration
depicted
by
the
regression
lines
occurs,
10%
of
the
population
is
likely
to
die
in
each
incidence
of
exceedence.
In
natural
stream
environments,
stream
temperatures
must
generally
exceed
the
highest
short
duration
temperatures
(
e.
g.
26oC
or
higher)
for
there
to
be
a
risk
of
direct
mortality
because
temperatures
rarely
remain
at
these
temperatures
continuously
due
to
natural
daily
temperature
fluctuation.

A.

0.01
0.1
1
10
100
1000
24
25
26
27
28
29
30
Temperature
(
oC)
Time
to
10%
Mortality
(
Hours)

Cutthroat
Rainbow
(
Steelhead)
Chinook
Coho
The
relationship
between
daily
temperature,
food
consumption,
and
growth
rate
for
coho
salmon
is
shown
in
B.
Growth
rate
is
strongly
influenced
by
temperature,
with
optimal
growth
occurring
when
fish
feed
at
satiation
ration
and
temperature
is
approximately
17oC
(
optimal
temperature
for
growth).
Growth
rate
declines
with
temperatures
either
warmer
or
colder
than
the
optimal.
Growth
effects
are
significant
at
temperatures
greater
than
22oC
and
less
than
9oC.
Each
line
represents
a
level
of
food
consumption.
This
relationship
and
one
for
steelhead
(
not
shown)
were
used
with
daily
temperature
measured
in
a
number
of
streams
and
rivers
to
assess
the
long­
term
effect
of
temperature
on
weight
gain
during
the
summer
months.
During
this
period,
temperature
in
many
of
the
study
streams
ranged
over
much
of
the
spectrum
of
positive
growth
shown
by
the
curves.

B.

C
o
h
o
S
a
lm
o
n
­
0
.
0
1
0
­
0
.
0
0
5
0
.0
0
0
0
.0
0
5
0
.0
1
0
0
.0
1
5
0
.0
2
0
0
.0
2
5
0
.0
3
0
0
10
20
30
T
e
m
p
e
r
a
t
u
r
e
(
o
C
)
G
r
o
w
t
h
R
a
t
e
(

­
1d­
1)
1
0
0
%

8
0
%

6
0
%

4
0
%

3
0
%
represented
a
wide
variety
of
streams,
including
many
with
high
annual
maximum
temperature.
Therefore,
although
no
mortality
was
reported
in
the
field
studies,
model
predictions
cannot
be
considered
fully
corroborated
until
direct
mortality
is
observed
at
predicted
exposures.
Failure
to
detect
mortality
is
consistent
with
the
general
perception
that
direct
mortality
from
temperature
rarely
occurs
within
the
natural
range
of
a
species
distribution.
Nevertheless,
lethal
temperatures
can
and
have
occurred
in
the
region,
and
there
are
situations
where
further
analysis
for
risk
of
direct
mortality
to
salmonids
is
warranted.

Some
concern
has
been
expressed
that
the
use
of
information
from
laboratory
studies
to
define
temperature
criteria
for
fish
living
in
natural
streams
is
inappropriate.
Such
concerns
confuse
the
synthesis
of
scientific
information
into
temperature
criteria
since
so
much
of
the
most
relevant
information
comes
from
laboratory
experiments.
Laboratory
and
field
studies
each
have
unique
limitations.
Laboratory
studies
are
conducted
in
highly
artificial
environments
that
create
stresses
from
the
experimental
procedures.
Field
studies
are
labor
intensive,
and
discerning
the
effect
of
temperature
by
empirical
observation
in
streams
is
problematic
given
the
multivariate
and
dynamic
nature
of
the
interaction,
and
the
difficulty
of
measuring
some
of
the
key
fundamental
relationships
in
natural
8­
7
Porter
Creek,
1988
0
1
2
3
4
5
6
7
24­
Jun
2­
Jul
10­
Jul
18­
Jul
26­
Jul
3­
Aug
11­
Aug
19­
Aug
27­
Aug
4­
Sep
12­
Sep
20­
Sep
28­
Sep
Weight
(
g)
Coho
Predicted
Coho
Observed
Steelhead
P
redicted
Steelhead
Observed
Figure
8.3
Simulation
of
weight
gain
for
coho
and
steelhead
in
relation
to
temperature
and
food
supply
using
the
growth
model
developed
in
this
report
in
relation
to
observed
weight
of
populations
living
in
natural
streams.
Model
simulations
computed
daily
weight
gain
using
the
measured
daily
temperature
and
estimates
of
food
consumption
derived
from
observing
population
growth
rates
and
back
calculating
how
much
food
had
to
have
been
consumed
to
account
for
the
weight
gain
between
two
sampling
intervals.
Results
shown
for
Porter
Creek
for
both
species
are
similar
to
those
observed
at
most
sites.
Overall,
predicted
weight
deviated
only
8%
from
observed
on
average
for
both
species
in
21
population
simulations.
Such
good
modeling
results
establish
confidence
in
use
of
the
model
to
estimate
the
effects
of
temperature
on
growth
for
determining
temperature
criteria.
It
also
confirms
the
strong
signature
of
prevailing
temperature
on
the
size
of
fish
in
natural
streams.
environments.
Many
scientists
have
argued
that
best
way
to
investigate
ecological
problems
involves
a
combination
of
laboratory
and
field
experiments.

Our
results
suggest
that
laboratory
results
are
of
fundamental
value
in
explaining
observed
fish
growth
in
natural
streams.
We
could
not
reject
the
hypotheses
regarding
growth
in
relation
to
environmental
temperature
using
simulations
based
on
the
laboratory
research.
In
fact,
the
simulations
were
remarkably
representative
of
the
observed
weight
gain
of
naturally
spawned
and
hatchery­
raised
populations
in
streams.
We
also
believe
that
the
direct
mortality
from
the
lethal
temperature
model
would
predict
mortality
consistent
with
the
temperatures
where
death
has
been
observed,
but
we
do
not
know
if
the
proportion
of
population
experiencing
mortality
would
be
as
we
predict.
Confirming
this
may
be
important
for
assessing
the
environmental
factors
controlling
species
distribution,
but
less
important
to
establishing
criteria.
Temperature
criteria
should
primarily
target
sublethal
effects
to
protect
fish
health.

Some
biological
responses
to
temperature
can
be
observed,
but
they
are
not
amenable
to
mathematical
expression
or
prediction.
These
include
behavioral
responses
such
as
cessation
of
feeding
and
seeking
refuge,
and
competitive
interactions.
There
are
internal
physiological
stress
effects
that
stimulate
such
externally
visible
symptoms,
and
additional
study
associating
stress
measures
with
temperature
characterization
would
be
a
useful
augmentation
of
the
analyses
of
growth.
Although
associated
with
environmental
temperature,
the
occurrence
of
some
responses
depends
on
the
presence
of
specific
factors
such
as
cold
water
refuges
or
disease
organisms
that
respond
consistently
with
the
prevailing
temperature.
(
The
role
of
temperature
in
increasing
incidence
of
disease
is
particularly
problematic
since
some
disease
organisms
are
more
virulent
in
cold
temperatures
while
others
are
more
virulent
in
warm
temperatures.)
There
are
also
factors
that
interact
directly,
indirectly
or
independently
of
temperature,
to
affect
the
organism's
condition.
These
include,
but
are
not
limited
to,
ecological
constraints
on
food
supply,
population
interactions,
and
genetic
adaptability.
Increased
understanding
of
these
in
the
context
of
environmental
temperature
would
enhance
understanding
of
the
effects
of
environmental
temperature.
8­
8
The
Scientific
Basis
for
Translating
Quantitative
Biological
Analyses
to
Temperature
Criteria
A
synthesis
of
the
scientific
literature
supports
the
premise
that
the
temperatures
associated
with
the
spectrum
of
biologic
responses
fall
between
low
level
growth
loss
and
direct
mortality,
escalating
as
temperatures
move
towards
the
extremes
of
the
tolerance
range.
The
criteria
suggested
by
our
analyses
for
growth
and
direct
mortality
envelop
these
responses.
For
coho
(
one
of
the
more
comprehensively
studied
species),
the
approximate
temperatures
associated
with
various
biological
effects
that
we
could
broadly
interpret
from
our
modeling
and
the
scientific
literature
are
listed
in
Table
8.1.

On
one
end
of
the
spectrum
is
direct
mortality
from
short­
term
exposure
to
high
temperature.
Clearly,
direct
mortality
is
an
unacceptable
endpoint
condition,
and
would
not
fully
protect
fish.
However,
it
is
important
to
be
on
the
alert
for
these
conditions,
because
there
are
some
geographic,
watershed,
and
climatic
conditions
where
acute
lethal
temperatures
have
been
documented
in
natural
conditions
or
due
to
management
activities.
On
the
other
end
of
the
temperature
spectrum,
positive
growth
for
juveniles,
or
weight
maintenance
for
adults,
is
a
measurable
quality
that
is
very
responsive
to
temperature
(
among
other
factors).
Therefore,
it
can
be
a
sensitive
and
early
indicator
of
the
general
health
of
individual
fish.
While
a
variety
of
ecological
factors
are
known
to
influence
population
characteristics,
the
growth
simulations
showed
that
there
is
a
very
strong
temperature
signature
in
the
size
and
condition
of
fish
observed
in
natural
streams.

Table
8.1
The
spectrum
of
coho
salmon
response
at
temperature
thresholds
synthesized
from
this
report
and
the
scientific
literature.
Threshold
values
are
approximations,
due
to
lack
of
consistency
in
reporting
results
among
studies.
Results
were
standardized
to
7­
day
maximum
temperature
using
regression
relationships
between
various
temperature
indices
described
in
Section
3.
Assumptions
regarding
the
relationship
between
reported
measures
and
7­
day
maximum
temperatures
were
assigned
to
standardize
results.

Biologic
Response
Approximate
Temperature
Threshold
oC
(
Average
of
the
7­
day
daily
maximum
temperature)

Upper
Critical
Lethal
Limit
(
death
within
minutes)
31
Geographic
limit
of
species
(
Eaton
1995)
30
Growth
loss
40%
(
simulated
at
average
food
supply)
30
Acute
threshold
(
this
report)
26
Acute
threshold
selected
by
U.
S.
EPA
1977
25
Complete
cessation
of
feeding
(
laboratory
studies)
24
Growth
loss
of
20%
(
simulated
at
average
food
supply)
22.5
Increased
incidence
of
disease
(
under
specific
situations)
22
Temporary
movements
to
thermal
refuges
22
Growth
loss
of
10%
(
simulated
at
average
food
supply)
16.5
Optimal
growth
at
range
of
food
satiation
(
laboratory)
14­
17
Growth
loss
of
10%
(
simulated
at
average
food
supply)
9.5
Cessation
of
feeding
and
movement
to
refuge
4
8­
9
There
is
no
consensus
among
physical
or
biological
scientists
as
to
how
to
report
temperature
regimes
represented
in
their
studies.
Therefore,
we
had
to
translate
these
reported
temperature
measures
to
a
common
standard
(
7­
day
average
of
the
daily
maximum
temperature)
using
relationships
between
temperature
indices
developed
in
the
report.
The
values
in
Table
8.1
should
therefore
be
viewed
as
approximate1.
One
can
see
that
simulated
growth
loss
values
identify
biologic
effects
closest
to
what
appears
to
be
the
healthiest
conditions.
At
the
temperatures
where
avoidance
behavior
or
competitive
exclusion
can
be
observed,
the
growth
simulation
would
have
predicted
measurable
and
possibly
significant
growth
loss.

Importantly,
the
analysis
leads
to
the
conclusion
that
the
cumulative
effects
on
potential
weight
gain
due
to
the
temperature
regime
for
the
summer
rearing
period
can
be
a
bell
weather
of
more
visible,
and
possibly
more
serious,
effects
observed
at
higher
temperatures.
It
appears
that
temperature
criteria
selected
on
the
basis
of
growth
can
be
protective
without
explicitly
accounting
for
all
biologic
responses
or
causal
mechanisms.
Concentrating
on
those
that
can
be
quantified
and
simulated
(
growth,
direct
mortality,
incubation)
allows
the
interactions
between
biologic
response
to
environmental
characteristics
to
be
quantified.
Thus,
it
may
be
the
most
sensitive
indicator
of
effects
that
can
also
be
measured
in
populations
in
natural
streams
without
sacrificing
fish.
This
also
allows
multiple
species
living
in
a
common
stream
to
be
assessed
and
compared
on
the
same
objective
basis,
and
in
relation
to
observed
and
potential
stream
temperature
regimens.

Temperature
Thresholds
Based
on
Risk
Assessment
The
growth
simulation
method
was
very
sensitive
to
temperature,
predicting
a
range
of
average
population
weights
that
varied
with
temperature
regime.
The
method
is
capable
of
assessing
a
specific
biological
response
on
a
continuous
temperature
scale.
However,
the
results
also
support
the
concept
that
useful
thresholds
can
be
assigned,
experimentally
tested,
and
justified
with
probabilistic
risk
assessment.
When
this
approach
is
applied
at
a
site,
with
interpretation
assisted
by
the
mathematical
model,
it
appears
that
rather
small
changes
in
average
population
weight
could
be
detected.
According
to
typical
size
distributions
in
populations
of
juvenile
salmon
populations,
a
minimum
detectable
weight
loss
or
gain
due
to
any
factor
would
be
approximately
20%.
The
growth
simulation
can
associate
such
small
changes
with
a
temperature
threshold.
Without
an
assist
in
hypothesis
formulation
by
the
growth
simulation
techniques,
it
would
probably
be
difficult
to
have
confidence
in
interpreting
the
influence
of
temperature
on
population
weight
differences
as
small
as
20%
.

Risk
level
for
establishing
thresholds.
A
quantitative
expression
of
the
consequence
of
size
completes
the
formal
appraisal
of
risk.
There
is
ample
evidence
to
suggest
that
larger
size
conveys
many
competitive
and
survival
benefits.
We
attempted
to
associate
risk
with
growth
loss
to
guide
selection
of
threshold
values.
We
did
so
for
coho
salmon
based
on
scientific
literature
that
suggests
that
size
at
the
end
of
the
juvenile
growth
phase
contributes
to
the
individual's
success
at
later
life
history
stages.
We
found
that
weight
loss
as
small
as
20%
of
the
average
population
weight
at
the
end
of
the
juvenile
summer
rearing
phase
may
be
important
in
this
context.
However,
the
research
results
supporting
this
conclusion
are
neither
abundant
nor
sufficiently
consistent
to
have
full
confidence
in
using
them
to
select
risk
criteria.
For
example,
later
success
in
the
marine
environment
1
The
7­
day
maximum
and
annual
maximum
temperature
are
closely
related
and
are
often
within
1o
to
2oC
of
each
other.
8­
10
y
=
1.117x
­
1.427
R2
=
0.996
10
12
14
16
18
20
22
24
26
1
0
1
2
1
4
1
6
1
8
2
0
2
2
2
4
2
6
7­
Day
M
aximum
Temperature
(
oC)
Line
o
f
1
:
1
Correspondence
y
=
0.6353x
+
4.5983
R2
=
0.89
1
0
1
2
1
4
1
6
1
8
2
0
2
2
1
0
1
2
1
4
1
6
1
8
2
0
2
2
2
4
2
6
7­
Day
M
aximum
T
emperature
(
oC
)
Line
o
f
1
:
1
Correspondence
Figure
8.4
The
relationship
between
various
temperature
indices
currently
used
as
the
averaging
period
for
temperature
criteria.
The
averaging
period
has
typically
been
either
the
annual
maximum
temperature
(
observed
for
a
period
as
short
as
an
hour,
but
more
probably
occurring
for
several
hours
on
sequential
days),
or
a
weekly
average
(
generally
focused
on
the
warmest
seven
consecutive
days)
observed
for
the
year.
For
example,
Washington's
current
criteria
specify
the
annual
maximum
temperature,
expressed
as
the
maximum
hourly
temperature
that
occurs
each
year.
Oregon
specifies
the
average
of
the
daily
maximum
temperature
of
the
7
warmest
consecutive
days.
The
U.
S.
EPA
(
1977)
recommends
the
average
of
the
daily
mean
temperature
of
the
7
warmest
consecutive
days
(
MWAT).
All
of
these
indices
represent
the
upper
tail
of
the
distribution
of
temperatures
observed
during
the
summer
months,
and
are
closely
related
to
one
another.
We
conclude
that
any
of
the
indices
can
be
used
for
the
purpose
of
temperature
criteria
because
they
are
closely
related.
Furthermore,
the
short­
term
measures
appear
to
adequately
represent
chronic
exposure
and
long­
term
effects.
also
depends
on
timing
and
location
factors.
Nevertheless,
within
the
juvenile
rearing
phase,
studies
have
consistently
reported
that
larger
size
generally
conveys
a
number
of
advantages
such
as
better
feeding
position
and
lower
mortality.
Such
observations
indicate
that
working
with
growth
or
weight
maintenance
is
a
useful
approach.
The
threshold
level
of
growth
loss
is
an
important
policy
choice
if
it
is
used
to
determine
the
thresholds
in
numeric
temperature
criteria.
Knowledge
of
the
implications
of
growth
and
size
to
organism
success
is
not
as
well
quantified
as
desired
to
guide
that
important
decision.
Establishing
the
effect
of
size
on
organism
success
merits
greater
research
attention.

Nevertheless,
several
lines
of
evidence,
as
well
as
precedent
in
the
scientific
literature,
suggest
growth
loss
values
between
10
and
20%
as
an
appropriate,
risk­
guided
threshold.
We
selected
a
growth
loss
of
10%
as
a
threshold
in
our
discussions
and
to
compare
temperature
criteria
with
other
existing
criteria
derived
from
other
approaches.
This
level
should
prevent
a
statistically
observable
change
in
average
population
weight,
assuming
that
population
numbers
remain
consistent
for
the
period.
It
is
possible
that
somewhat
higher
growth
loss
would
be
acceptable,
although
we
suspect
that
growth
loss
can't
be
much
higher
since
temperatures
associated
with
higher
growth
loss
begin
to
coincide
with
the
outward
manifestation
of
other
adverse
effects
such
as
avoidance
behavior
(
Table
8.1).
We
do
not
attempt
to
quantify
the
response
of
resident
adult
fish
to
growth
loss
as
an
indicator
of
adverse
temperature
effects,
although
the
same
physiologic
mechanisms
manifest
as
weight
loss
in
resident
adults
and
undoubtedly
have
ecological
ramifications.

Averaging
periods
for
criteria.
Temperature
criteria
use
short­
term
averaging
periods
as
indices
of
the
long­
term
response.
Results
provided
in
this
report
confirm
that
these
indices
can
be
used
reliably
to
represent
the
long­
term
temperature
regime.
All
of
the
indices
(
annual
maximum
temperature,
7
day
averages
of
the
daily
maximum
or
8­
11
daily
mean
temperature)
characterize
the
upper
tail
of
the
sampled
temperature
distribution,
and
they
are
closely
correlated
with
each
other.
This
makes
selection
among
them
a
matter
of
procedural
and
logistical
concerns,
rather
than
a
biological
question.
Some
standardization
of
reporting
measures
would
be
most
helpful.
We
urge
scientists
to
provide
at
least
one
of
these
indexing
measures
with
their
study
results,
thus
enabling
comparisons
among
them
as
well
as
their
use
in
supporting
the
development
of
temperature
criteria
for
regulatory
purposes.
We
found
that
the
average
of
the
maximum
7
consecutive
days
of
the
daily
mean
temperature
(
MWAT,
U.
S.
EPA
1977)
was
best
correlated
with
growth
simulations,
but
the
annual
maximum
and
7­
day
maximum
were
also
quite
suitable.
Appropriate
temperature
thresholds
vary
with
each
index.

Temperature
Criteria
Derived
From
Risk
Assessment
Thresholds
generated
from
risk
assessment
are
reasonably
consistent
with
criteria
developed
previously,
including
those
derived
from
subjective
review
methods
and
objective
analysis,
and
those
currently
authorized
by
states.
The
upper
temperature
thresholds
associated
with
10%
weight
reduction
are
16.5o
and
20.5oC
for
coho
and
steelhead
respectively
(
Figure
8.4).
Sub­
lethal
thresholds
suggested
by
the
risk
assessment
method
tend
to
be
slightly
lower
than
those
derived
from
objective
methods
(
e.
g.
EPA
1977),
probably
because
we
directly
accounted
for
realistic
estimates
of
food
7­
Day
Maximum
0
5
10
15
20
25
30
Coho
Steelhead
Temperature
(
oC)

10%
20%

20%
Annual
Maximum
0
5
10
15
20
25
30
Coho
Stealhead
Temperature
(
oC)
20%

20%
10%
7­
Day
Mean
0
5
10
15
20
25
30
Coho
Steelhead
Temperature
(
oC)

20%

10%

20%

Figure
8.4
Temperature
range
for
increments
of
growth
loss
associated
with
long­
term
temperature
regime
as
expressed
for
3
temperature
indices
and
2
species
of
salmonids.
The
inner
range
represents
up
to
10%
growth
loss
and
the
outer
range
represents
up
to
20%
growth
loss.
Above
and
below
the
ranges
shown
the
growth
loss
exceeds
20%
and
was
as
high
as
50%
near
the
extremes
of
the
temperature
range.
Risk
assessment
associated
with
growth
loss
suggests
that
a
10%
limit
would
prevent
any
measurable
effect
on
average
coho
population
weight.
A
loss
of
20%
would
be
detectable
and
the
temperature
associated
with
this
level
of
growth
loss
coincides
with
temperatures
associated
with
avoidance
behavior.
Therefore,
thresholds
selected
at
10%
may
be
most
appropriate
for
establishing
temperature
criteria.
The
threshold
temperature
varies
with
each
index.
8­
12
7­
Day
Maximum
Temperature
0
5
10
15
20
25
Risk
Assess
Bell
1973
EPA
1977
ODEQ
1995
WDOECur
rent
WDOEProposed
(
1999)
Temperature
(
oC)

10%
20%

20%

Figure
8.5
A
comparison
of
temperature
criteria
for
coho
and
recommended
ranges
from
a
variety
of
sources
standardized
to
the
7­
day
maximum
temperature.
The
risk
assessment
is
the
range
of
values
developed
in
this
report.
Bell
(
1973)
is
the
original
source
of
temperature
range
recommendations
that
have
been
widely
used
as
the
basis
for
subjective
analyses.
These
were
used
in
part
to
form
criteria
used
by
Oregon
(
ODEQ
1995)
and
the
current
and
proposed
criteria
for
Washington
(
WDOE).
The
U.
S.
EPA
used
an
objective
approach
based
on
the
growth
curves
to
determine
threshold
criteria.
The
various
methods
vary
in
terms
of
degree
of
objectivity
or
subjectivity
by
which
the
information
is
synthesized
into
recommended
criteria,
the
degree
to
which
data
forms
the
basis
for
the
criteria,
and
the
extent
to
which
population
effects
can
be
probabilistically
determined.
The
temperature
criteria
in
use
in
Pacific
Northwest
states
have
largely
been
drawn
from
professionals'
review
and
interpretation
of
available
scientific
literature
(
e.
g.
ODEQ
1995).
There
has
been
some
previous
effort
to
use
the
more
well­
established
scientific
relationships
to
develop
objective
analyses
of
threshold
temperatures
and
the
duration
of
exposure
(
e.
g.,
U.
S.
EPA
1977).
This
approach
has
not
been
widely
integrated
into
regulatory
activity.
availability
in
the
simulations.
The
risk­
derived
thresholds
tend
to
be
somewhat
higher
than
those
emerging
from
subjective
evaluations
(
e.
g.
WDOE
1999),
possibly
because
we
did
not
add
arbitrary
safety
factors
(
Figure
8.5).
We
believe
that
the
choice
of
low
thresholds
of
growth
loss
(
e.
g.
10%)
provides
an
adequate
margin
of
safety.
Using
the
growth
loss
to
set
the
risk
level
also
prevents
any
unintended
consequences
of
selecting
values
that
are
too
low
for
a
particular
species.
Our
results
demonstrated
that
temperatures
that
are
low
relative
to
a
species'
optimum
have
growth
loss
effects
that
are
comparable
to
those
associated
with
those
that
are
high
relative
to
the
optimum.

When
two
or
more
species
coexist,
as
is
often
the
case,
it
may
be
appropriate
to
select
the
threshold
for
the
more
sensitive
species.
In
the
case
of
coho
and
steelhead,
there
would
be
no
negative
effect
on
steelhead
by
targeting
lower
temperatures
appropriate
for
coho,
the
more
sensitive
species.
If
the
margin
between
species
is
wider,
the
tradeoffs
for
species
could
be
evaluated
in
selecting
the
temperature
threshold
if
growth
models
for
all
species
were
available.

The
fundamental
relationship
quantifying
growth
and
mortality
were
similar,
though
not
identical,
for
the
two
salmon
species
we
modeled.
However
threshold
temperatures
8­
13
generated
through
growth
simulation
varied
between
them,
reflecting
the
differences
in
food
consumption
estimated
by
observing
population
growth
in
natural
streams
that
was
used
in
the
modeling.
This
result
highlights
the
importance
of
food
availability
as
an
important
factor
determining
fish
growth,
a
conclusion
consistent
with
observations
from
field
ecological
studies.
The
simulation
results
suggest
that
coho
populations
ate
at
approximately
50%
of
satiation
rations
and
steelhead,
using
a
different
feeding
strategy,
ate
at
100%
satiation.
Importantly,
food
availability
influences
the
temperature
threshold
for
adverse
effects.
There
is
very
little
documentation
of
how
much
food
is
available
for
fish
dwelling
in
streams
and
rivers,
and
how
management
activities
may
alter
it.
Greater
understanding
of
how
site
and
watershed
conditions
determine
how
much
food
is
available
and
how
it
is
allocated
within
populations
would
allow
understanding
of
how
temperature
affects
total
productivity
in
addition
to
its
effect
on
weight
gain
of
the
individuals
in
the
population.

The
analyses
documented
in
this
report
addressed
the
juvenile
rearing
phase
of
coho
salmon
and
steelhead
trout.
These
species
are
widely
distributed
within
the
region,
they
are
listed
as
threatened
and
endangered
in
a
number
of
locations,
and
there
was
sufficient
laboratory
and
population
data
to
build
models
and
corroborate
them
in
natural
streams.
Nevertheless,
this
is
a
limited
representation
of
the
fish
species
that
occur
in
the
Pacific
Northwest.
Similar
techniques
could
be
applied
to
all
fish
species
if
the
fundamental
laboratory
relationships
used
in
the
growth
simulation
method
were
available.
Currently,
there
are
a
number
of
gaps
in
information
for
key
functions
used
in
modeling
species
of
interest
such
as
cutthroat
trout
and
bull
trout.

Analysis
of
lethal
temperatures
suggested
that
a
threshold
of
26oC
for
annual
maximum
temperature
is
a
signal
of
imminent
risk
of
direct
mortality
to
salmonids.
Although
the
occurrence
of
water
temperature
this
high
is
rare,
it
has
occasionally
been
observed
in
natural
streams
as
well
as
in
those
impacted
by
anthropogenic
activities
in
some
situations.
We
also
recommend
site­
specific
analysis
of
duration
of
exposure
when
annual
maximum
temperature
is
between
24o
and
26oC
in
order
to
assure
that
duration/
magnitude
thresholds
are
not
exceeded.
The
relationship
between
thresholds
for
growth
and
mortality
suggests
that,
if
growth
thresholds
are
met,
lethal
temperatures
will
not
occur.
However,
there
are
situations
where
rivers
and
streams
cannot
be
expected
to
meet
these
criteria,
even
under
natural
conditions.
Acute
criteria
may
be
most
helpful
for
triggering
additional
study
in
certain
situations,
and
for
prioritizing
restoration
activities.

It
may
be
useful
to
vary
temperature
criteria
on
a
seasonal
basis
matching
fish
requirements,
although
differing
criteria
for
too
fine
a
resolution
of
time
may
be
difficult
to
administer
and
may
offer
relatively
little
additional
benefit
if
ambient
temperatures
are
generally
within
exposure
duration
limits.
The
risk­
based
approach
could
be
used
to
investigate
the
need
for
more
finely
tuned
seasonal
criteria.
or
to
develop
site­
specific
criteria.
The
concept
of
selecting
criteria
for
particular
species
appears
valid,
and
the
risk
assessment
method
can
be
employed
to
help
guide
the
selection
of
appropriate
criteria
for
target
species
or
it
can
be
used
to
address
multiple
species
living
in
the
same
location.
There
was
no
indication
in
our
analysis
that
criteria
for
daily
fluctuating
temperature
would
improve
biological
characterization.
Also,
some
states
have
a
maximum
allowable
increase
in
temperature
as
well
as
an
upper
threshold.
The
value
of
this
provision
is
not
immediately
apparent
in
the
context
of
either
acute
or
chronic
effects
analysis
discussed
in
this
report.
8­
14
Uncertainties
in
Applying
Criteria
in
Natural
Environments
There
are
natural
factors
contributing
to
uncertainty
and
variability
when
it
comes
to
administering
temperature
criteria.
There
are
systematic
patterns
in
temperature
dictated
by
watershed
and
geographic
conditions.
It
is
important
to
recognize
that
the
attainment
of
biologically
based
criteria
will
vary
with
watershed
characteristics.
Temperature
regime
also
varies
annually
by
as
much
as
several
degrees
due
to
climatic
factors,
so
it
may
be
appropriate
to
establish
confidence
limits
around
threshold
values
to
determine
whether
water
quality
standards
are
attained.

Conclusion
In
conclusion,
the
analytical
approaches
to
assessing
risk
to
salmon
associated
with
ambient
environmental
temperature
presented
in
this
report
appear
to
be
promising
techniques
for
objectively
defining
temperature
criteria.
They
could
also
assist
ecological
field
studies
to
segregate
the
effects
of
temperature
from
other
habitat
and
population
factors
that
influence
productivity.
The
risk­
based
analyses
support
the
approach
and
general
range
of
numeric
threshold
values
currently
used
as
temperature
criteria
by
Pacific
Northwest
states.
However,
the
specific
numbers
generated
by
quantitative
risk
analysis
techniques
vary
slightly
from
existing
authorized
criteria.
Assuming
the
most
sensitive
salmon
species
is
used
to
select
the
threshold,
and
a
growth
loss
threshold
of
10%,
the
levels
suggested
by
risk
assessment
are
slightly
higher
than
used
by
Washington
and
slightly
lower
than
used
by
Oregon.
Additional
research
is
needed
to
develop
the
biological
basis
for
other
species
of
interest.
SECTION
9
LITERATURE
CITED
Adams,
S.
M.
and
J.
E.
Breck.
1990.
Bioenergetics.
In:
Shreck,
C.
B.
and
P.
B.
Moyle,
eds.
Methods
for
fish
biology.
American
Fisheries
Society.
Bethesda,
MD.

Agee,
J.
K.
1993.
Fire
ecology
of
Pacific
Northwest
forests.
Island
Press,
Washington
D.
C.
493
pp.

Alabaster,
J.
S.
and
A.
L.
Downing.
1966.
A
field
and
laboratory
investigation
of
the
effect
of
heated
effluents
on
fish.
Fish.
Min.
Arg.
Fish
Food
(
Great
Britian)
Ser.
I
Sea
Fish
6(
4):
1­
42.

Alabaster,
J.
S.
and
Welcomme.
1962.
Effect
of
concentration
of
dissolved
oxygen
on
survival
of
trout
and
roach
in
lethal
temperatures.
Nature
194:
107.

Allen,
M.
A.
and
T.
J.
Hassler.
1986.
Species
profiles
and
environmental
requirements
of
coastal
fishes
and
invertebrates
(
Pacific
Northwest)
 
chinook
salmon.
U.
S.
Fish
and
Wildlife
Service
Biological
Report
82
(
11.62).
26
pp.

Anderson,
D.
R.
1986.
Proceedings:
risk
assessment
in
aquatic
ecology.
Electric
Power
Research
Institute,
Palo
Alto,
CA.
Report
No.
EPRI
EA­
4438.

Armour,
C.
L.
1991.
Guidance
for
evaluating
and
recommending
temperature
regimes
to
protect
fish.
U.
S.
Department
of
the
Interior,
Fish
and
Wildlife
Service,
Washington,
D.
C.
Instream
Flow
Information
Paper
28.
Biological
Report
90(
22).
13
pages.

Averett,
C.
L.
1968.
Influence
of
temperature
on
energy
and
material
utilization
by
juvenile
coho
salmon.
PhD
Thesis.
Oregon
State
University,
Corvallis,
OR.

ASTM.
1997.
Annual
Book
of
Standards.
American
Society
for
Testing
and
Materials,
Philadelphia,
Pennsylvania.

Baker,
P.
F.,
T.
P.
Speed
and
F.
K.
Ligon.
1995.
Estimating
the
influence
of
temperature
on
the
survival
of
chinook
salmon
smolts
(
Onchorhyncus
tshawytscha)
migrating
through
the
Sacremento­
San
Joaquin
River
delta
of
California.
Can.
J.
Fish.
Aquat.
Sci.
52:
855­
863.

Baltz,
D.
M.,
P.
B.
Moyle,
and
N.
J.
Knight
1982.
Competitive
interactions
between
benthic
stream
fishes,
riffle
sculpin,
Cottus
gulosus,
and
speckled
dace,
Rhinichys
osculus.
Can.
J.
Fish.
Aquat.
Sci..
39:
1502­
1511.

Bartell,
S.
M.,
J.
E.
Breck,
R.
H.
Gardner,
and
A.
L.
Brenkert.
1986.
Individual
parameter
perturbation
and
error
analysis
of
fish
bioenergetics
models.
Can.
J.
Fish.
Aquat.
Sci.
43:
160­
168.

Beauchamp,
D.
A.,
D.
J.
Stewart,
and
G.
L.
Thomas.
1989.
Corroboration
of
a
bioenergetics
model
for
sockeye
salmon.
Trans.
Am.
Fish.
Soc.
118:
597­
607.

Bell,
M.
C.
1973,
1986,
and
1990.
Fisheries
handbook
of
engineering
requirements
and
biological
criteria.
US
Army
Corps
of
Engineers,
North
Pacific
Division,
Portland,
OR.

Beschta,
R.
L.,
R.
E.
Bilby,
G.
W.
Brown,
L.
B.
Holtby,
and
T.
D.
Hofstra.
1987.
Stream
temperatures
and
aquatic
habitat:
fisheries
and
forestry
interactions.
In:
Salo,
E.
O.
and
T.
W.
Cundy,
eds.,
Proc.
Symposium
Streamside
management:
forestry
and
fishery
interactions.
Held.
Feb
12­
14,
1986.
University
of
Washington,
Seattle,
WA.:
191­
232.

9­
1
Bilby,
R.
E.
and
P.
A.
Bisson.
1992.
Allochthonous
and
autochthonous
organic
matter
contributions
to
the
trophic
support
of
fish
populations
in
clear­
cut
and
old­
growth
forested
streams.
Can.
J.
Fish.
Aq.
Sci.
49:
540­
551.

Bilby,
R.
E.,
B.
R.
Fransen,
and
P.
A.
Bisson.
1996.
Incorporation
of
nitrogen
and
carbon
from
spawning
coho
salmon
into
the
trophic
systems
of
small
streams:
evidence
from
stable
isotopes.
Can.
J.
Fish.
Aquat.
Sci.
53:
164­
173.

Bilby,
R.
E.,
B.
R.
Fransen,
P.
A.
Bisson,
and
J.
K.
Walter.
1998.
Response
of
juvenile
coho
salmon
(
Oncorhynchus
kisutch)
and
steelhead
(
Oncorhynchus
mykiss)
to
the
addition
of
salmon
carcasses
to
two
streams
in
southwestern
Washington,
U.
S.
A.
Can.
J.
Fish.
Aquat.
Sci.
55:
1909­
1929.

Biro,
P.
A.
1998.
Staying
cool:
behavioral
thermoregulation
during
summer
by
young­
of­
the­
year
brook
trout
in
a
lake.
Trans.
Am.
Fish.
Soc.
127:
212­
222.

Bisson,
P.
A.
and
G.
E.
Davis.
1976.
Production
on
juvenile
chinook
salmon,
Oncorhynchus
tshawytscha,
in
a
heated
model
steam.
Fishery
Bulletin
74(
4):
763­
774.

Bisson,
P.
A.,
J.
L.
Nielsen,
and
J.
W.
Ward.
1988a.
Summer
production
of
coho
salmon
stocked
in
Mount
St.
Helens
streams
3­
6
years
after
the
1980
eruption.
Trans.
Am.
Fish.
Soc.
117:
322­
355.

Bisson,
P.
A.
K.
Sullivan,
and
J.
L.
Nielsen.
1988.
Channel
hydraulics,
habitat
use,
and
body
form
of
juvenile
coho
salmon,
steelhead,
and
cutthroat
trout
in
streams.
Trans.
Am.
Fish.
Soc.
117:
262­
273.

Bisson,
P.
A.
and
R.
E.
Bilby.
1998.
Organic
matter
and
trophic
dynamics.
In:
Naiman,
R.
J.
and
R.
E.
Bilby,
eds.
River
ecology
and
management:
lessons
from
the
Pacific
coastal
region.
Springer,
New
York:
373:
398.

Bjornn,
T.
C.
and
D.
W.
Reiser.
1991.
Habitat
requirements
of
salmonids
in
streams.
American
Fisheries
Society
Special
Publication
19:
83­
138.

Black,
J.
2000.
Summer
temperatures
in
small
stream
sources
on
managed
Olympic
Peninsula
timberlands.
M.
S.
Thesis,
U.
Washington,
Seattle,
WA.

Blahm
and
McConnell
1970.
Unpublished
data
cited
in
National
Academy
of
Sciences
and
National
Academy
of
Engineering
(
NAS/
NAE).
1973.

Blahm
and
Parente
1970.
Unpublished
data
cited
in
National
Academy
of
Sciences
and
National
Academy
of
Engineering
(
NAS/
NAE).
1973.

Boisclair,
D.
and
W.
C.
Leggett.
1989.
The
importance
of
activity
in
bioenergetics
models
applied
to
actively
foraging
fishes.
Can.
J.
Fish.
Aquat.
Sci.
46:
1859­
1867.

Boisclair,
P.
and
W.
C.
Leggett.
1991.
If
computers
could
swim
or
fish
could
be
programmed:
a
response
to
comments
by
Hewett
et
al.
(
1991).
Can.
J.
Fish.
Aq.
Sci.
48:
1337­
1344.

Bovee,
K.
D.
1978.
Probability
of
use
criteria
for
the
family
Salmonidae.
U.
S.
Fish
and
Wildlife
Service,
Coop.
Instream
Flow
Service
Group.
Instream
Flow
Information
Paper
No.
4.
FWS/
OBS­
78/
07.

Brett,
J.
R.
1952.
Temperature
tolerance
in
young
Pacific
salmon,
Genus
Oncorhynchus.
Journal
of
the
Fisheries
Research
Board
of
Canada
9(
6):
265­
323.

Brett,
J.
R.
1956.
Some
principles
in
the
thermal
requirements
of
fishes.
1956.
The
Quarterly
Review
of
Biology.
31(
2):
75­
87.

9­
2
Brett,
J.
R.
1958.
Implications
and
assessments
of
environmental
stress.
In
P.
A.
Larkin,
editor.
The
Investigation
of
Fish­
Power
Problems.
H.
R.
MacMillan
Lectures
in
Fisheries,
Univ.
of
British
Columbia,
Vancouver:
60­
83.

Brett,
J.
R.
1971.
Energetic
responses
of
salmon
to
temperature.
A
study
of
some
thermal
relations
in
the
physiology
and
freshwater
ecology
of
sockeye
salmon
(
Oncorhynchus
nerka).
Amer.
Zool.
11(
1):
99­
113.

Brett,
J.
R.
1995.
Energetics.
pp.
3­
68.
In:
C.
Groot,
L.
Margolis,
and
W.
C.
Clarke
[
Editors].
Physiological
Ecology
of
Pacific
Salmon.
UBC
Press,
Vancouver,
B.
C.

Brett,
R.
J.,
J.
E.
Shelbourn,
and
C.
T.
Shoop.
1969.
Growth
rate
and
body
composition
of
fingerling
sockeye
salmon,
Oncorhynchus
nerka,
in
relation
to
temperature
and
ration
size.
J.
Fish.
Res.
Bd.
Canada
26:
2363­
2394.

Brett,
R.
J.,
W.
C.
Clarke,
and
J.
E.
Shelbourn.
1982.
Experiments
on
thermal
requirements
for
growth
and
food
conversion
efficiency
of
juvenile
chinook
salmon
Oncorhynchus
tshawytscha.
Can.
Tech.
Rep.
Fish.
Aquat.
Sci.
No.
1127.

Brown,
G.
W.
1969.
Predicting
temperatures
of
small
streams.
Wat.
Resources
Res.
5:
68­
75.

Brungs
and
Jones.
1977.
Temperature
criteria
for
freshwater
fish:
protocol
and
procedures.
Ecological
Research
Series.,
U.
S.
Environmental
Protection
Agency,
EPA­
600/
3­
77­
061.
130
pp.
[
Also
cited
as
U.
S.
EPA
1977.]

Bustard,
D.
R.
and
D.
W.
Narver.
1975.
Aspects
of
the
winter
ecology
of
juvenile
coho
salmon
(
Onchorynchus
kisutch)
and
steelhead
trout
(
Salmo
gairdneri).
J.
Fish.
Research
Bd.
Can.
32:
667­
680.

Cairns,
J.,
Jr.,
A.
L.
Buikema,
Jr.,
A.
G.
Heath,
and
B.
C.
Parker.
1978.
Effects
of
temperature
on
aquatic
organism
sensitivity
to
selected
chemicals.
Virginia
Water
Resources
Research
Center,
Bulletin
106.
88
pages.

Cederholm,
C.
J.,
R.
E.
Bilby,
P.
A.
Bisson,
T.
W.
Bumstead,
B.
R.
Fransen,
W.
J.
Scarlett,
and
J.
W.
Ward.
1997.
Response
of
juvenile
coho
salmon
and
steelhead
to
placement
of
large
woody
debris
in
a
coastal
Washington
stream.
N.
Am.
J.
Fish.
Manage.
17(
4):
947­
963.

Chapman,
D.
W.
1966.
Food
and
space
as
regulators
of
salmonid
populations
in
streams.
Am.
Midland
Nat.
100:
345­
357.

Chapman,
D.
W.
and
T.
C.
Bjornn
1969.
Distribution
of
salmonids
in
streams
with
special
reference
to
food
and
feeding.
Proc.
Symposium
on
Trout
in
Streams.
In:
T.
G.
Northcote
[
ed].
Symposium
on
salmon
and
trout
in
streams.
H.
R.
MacMillan
Lectures
in
Fisheries.
Univ.
Brit.
Col,
Vancouver,
B.
C.
:
153­
176.

Christie,
G.
C.
and
H.
A.
Regier.
1988.
Measures
of
optimal
thermal
habitat
and
their
relationship
to
yields
for
four
commercial
species.
Can.
J.
Fish.
Aquat.
Sci.
45:
301­
314.

Coutant,
C.
C.
1977.
Compilation
of
temperature
preference
data,
J.
Fish.
Res.
Bd
Can.
34:
739­
745.

Coutant,
C.
C.
and
S.
S.
Talmage.
1977.
Thermal
effects.
J.
Water
Pollution
Control
Federation
49:
1369­
1425.

9­
3
Dehart,
D.
A.
1975.
Resistance
of
three
freshwater
fishes
to
fluctuating
thermal
environments.
M.
S.
Thesis.
Oregon
State
University,
Corvallis,
OR.

De
Staso,
J.
III.,
and
F.
J.
Rahel.
1994.
Influence
of
water
temperature
on
interactions
between
juvenile
Colorado
River
cutthroat
trout
and
brook
trout
in
a
laboratory
stream.
Trans.
Am.
Fish.
Soc.
123:
289­
297.

Dickerson,
B.
R.
and
G.
L.
Vinyard.
1999.
Effects
of
high
chronic
temperatures
and
diel
temperature
cycles
on
the
survival
and
growth
of
Lahontan
cutthroat
trout.
Trans.
Am.
Fish.
Soc.
128:
516­
521.

Dill,
L.,
R.
Ydenberg,
and
A.
Fraser.
1981.
Food
abundance
and
territory
size
in
juvenile
coho
salmon
(
Oncorhynchus
kisutch).
Can.
J.
Zool.
59:
1801­
1809.

Dockray,
J.
J.,
S.
D.
Reid,
and
C.
M.
Wood.
1996.
Effects
of
elevated
summer
temperatures
and
reduced
pH
on
metabolism
and
growth
of
juvenile
rainbow
trout
(
Oncorhynchus
mykiss)
on
unlimited
ration.
Can.
J.
Fish.
Aquat.
Sci.
53:
2752­
2763.

Donaldson,
L.
R.
and
F.
J.
Foster.
1941.
Experimental
study
of
the
effect
of
various
water
temperatures
on
the
growth,
food
utilization,
and
mortality
rate,
of
fingerling
sockeye
salmon.
Trans.
Am.
Fish.
Soc.
70:
339­
346.

Dwyer,
W.
P.
and
R.
H.
Kramer.
1975.
The
influence
of
temperature
on
scope
for
activity
in
cutthroat
trout,
Salmo
clarki.
Trans.
Am.
Fish.
Soc.
104(
3):
552­
554.

Eaton,
J.
G.,
J.
H.
McCormick,
B.
E.
Goodno,
D.
G.
O'Brien,
H.
G.
Stefany,
M.
Hondzo,
and
R.
M.
Scheller.
1995.
A
field
information­
based
system
for
estimating
fish
temperature
tolerances.
Fisheries
20(
4):
10­
18.

Elliott,
J.
M.
1975.
The
growth
rate
of
brown
trout
(
Salmo
trutta
L.)
fed
on
maximum
rations.
J.
Animal
Ecol.
44:
806­
821.

Elliott,
J.
M.
1976.
The
energetics
of
feeding,
metabolism
and
growth
of
brown
trout
(
Salmo
trutta)
in
relation
to
body
weight
,
water
temperature
and
ration
size.
J.
Animal
Ecology
45:
923­
948.

Elliott,
J.
M.
1981.
Some
aspects
of
thermal
stress
on
freshwater
teleosts.
Pages
209­
245
In
A.
D.
Pickering
editor,
Stress
and
Fish.
Academic
Press,
London.

Everson,
L.
B.
1973.
Growth
and
food
consumption
of
juvenile
coho
salmon
exposed
to
natural
and
elevated
fluctuating
temperatures.
M.
S.
Thesis,
Oregon
State
University,
Corvallis,
OR.
68
pp.

Fagerlund,
U.
H.
M,
J.
R.
McBride,
and
I.
V.
Williams.
1995.
Stress
and
tolerance.
pp.
461­
503.
In:
C.
Groot,
L.
Margolis,
and
W.
C.
Clarke
[
Editors].
Physiological
Ecology
of
Pacific
Salmon.
UBC
Press,
Vancouver,
B.
C.

Fausch,
K.
D.
1984.
Profitable
stream
positions
for
salmonids:
relating
specific
growth
rate
to
net
energy
gain.
Can.
J.
Zool.
62:
441­
451.

Filbert,
R.
B.
and
C.
P.
Hawkins.
1995.
Variation
in
condition
of
rainbow
trout
in
relation
to
food,
temperature,
and
individual
length
in
the
Green
River,
Utah.
Trans.
Am.
Fish.
Soc.
124:
824­
835.

Fry,
F.
E.
J.
1947.
Effects
of
the
environment
on
animal
activity.
Univ.
Toronto
Stud.
Biol.
Ser.
55,
Publs.
Ont.
Fish.
Res.
Lab
68:
5­
62.

9­
4
Fry,
F.
E.
J.
1967.
Responses
of
vertebrate
poikilotherms
to
temperature.
Pages
375­
409
in
A.
H.
Rose
editor.
Thermobiology.
Academic
Press,
London.

Fransen,
B.
R.,
P.
A.
Bisson,
J.
W.
Ward,
and
R.
E.
Bilby.
1993.
Physical
and
biological
constraints
on
summer
rearing
of
juvenile
coho
salmon
(
Oncorhynchus
kisutch)
in
small
western
Washington
streams.
In:
Berg,
L.
and
P.
W.
Delaney,
eds.,
Proceedings
of
the
Coho
Workshop.
Nanaimo,
B.
C.
May
26­
28,
1992:
271­
288.

Giannico,
G.
R.
and
M.
C.
Healey.
1998.
Effects
of
flow
and
food
on
winter
movements
of
juvenile
coho
salmon.
Trans.
Am.
Fish.
Soc.
127:
645­
651.

Golden,
J.
T.
1978.
The
effects
of
fluctuating
temperatures
on
the
lethal
tolerance
of
coastal
cutthroat
trout
(
Salmo
clarki
clarki).
MS.
Thesis.
Oregon
State
University,
Corvallis,
OR.

Griffiths,
J.
S.
and
D.
F.
Alderdice
1972.
Effects
of
acclimation
and
acute
temperature
experience
on
the
swimming
speed
of
juvenile
coho
salmon.
J.
Fish.
Res.
Bd.
Canada
29:
251­
264.

Hall,
J.
D.,
and
R.
L.
Lantz.
1969.
Effects
of
logging
on
the
habitat
of
coho
salmon
and
cutthroat
trout
in
coastal
streams.
In:
T.
G.
Northcote
[
ed.].
Proc.
Symposium
on
slamon
and
trout
in
streams.
H.
R.
MacMillan
Lectures
in
Fisheries,
University
of
British
Columbia,
Vancouver:
355­
375.

Hansen,
D.
J.
1989.
Status
of
the
development
of
water
quality
criteria
and
advisories.
Water
Quality
Standards
for
21st
Century,
1989.
Pp.
163­
169.
U.
S.
Environmental
Protection
Agency,
Env.
Res.
Lab.,
Narragansett,
Rhode
Island.

Hansen,
M.
J.,
D.
Boisclair,
S.
B.
Brandt,
S.
W.
Hewett,
J.
F.
Kitchell,
J.
C.
Lucas,
and
J.
J.
Ney.
1993.
Applications
of
bioenergetics
models
to
fish
ecology
and
management:
where
do
we
go
from
here?
Trans.
Am.
Fish.
Soc.
122:
1019­
1030.

Hanson,
P.
C.,
T.
B.
Johnson,
D.
E.
Schindler
and
J.
F.
Kitchell.
1997.
Fish
Bioenergetics
3.0
Users
manual.
University
of
Wisconsin­
Madison
Center
for
Limnology,
U.
Wisconsin,
Madison.

Hartman,
G.
F.
and
J.
C.
Scrivener.
1990.
Impacts
of
forestry
practices
on
a
coastal
stream
ecosystem,
Carnation
Creek,
British
Columbia.
Canadian
Bulletin
of
Fisheries
and
Aquatic
Sciences
223.

Hawkins,
C.
P.,
M.
L.
Murphy,
N.
H.
Anderson,
and
M.
A.
Wilzbach.
1983.
Riparian
canopy
and
substrate
composition
interact
to
influence
the
abundance
of
salmonids,
sculpins,
and
salamanders
in
streams.
Can.
J.
of
Fish.
and
Aquat.
Sci.
40:
1173­
1185.

Hetrick,
N.
J.,
M.
A.
Brusven,
T.
C.
Bjornn,
R.
M.
Keith,
and
W.
R.
Meehan.
1998a.
Effects
of
canopy
removal
on
invertebrates
and
diet
of
juvenile
coho
salmon
in
a
small
stream
in
southeast
Alaska.
Trans.
Am.
Fish.
Soc.
127:
876­
888.

Hetrick,
N.
J.,
M.
A.
Brusven,
W.
R.
Meehan,
and
T.
C.
Bjornn.
1998.
Changes
in
solar
input,
water
temperature,
periphyton
accumulation,
and
allochthonous
input
and
storage
after
canopy
removal
along
two
small
salmon
streams
in
southeast
Alaska.
Trans.
Am.
Fish.
Soc.
127:
859­
875.

Hewett,
S.
W.
and
B.
L.
Johnson.
1992.
Fish
bioenergetics
model
2.
University
of
Wisconsin,
Sea
grant
Institute.
Tech.
Rept
WIS­
SG­
92­
250,
Madison.

Hokanson,
K.
E.
F.
1977.
Temperature
requirements
of
some
percids
and
adaptions
to
the
seasonal
temperature
cycle.
J.
Fish.
Res.
Bd.
Can.
34:
1524­
1550.

9­
5
Hokanson,
K.
E.
F.,
C.
F.
Kleiner
and
T.
W.
Thorslund.
1977.
Effects
of
constant
temperatures
and
diel
temperature
fluctuations
on
specific
growth
and
mortality
rates
and
yield
of
juvenile
rainbow
trout,
Salmo
gairdneri.
J.
Fish.
Res.
Bd.
Can.
34:
639­
648.

Holtby,
L.
B.
and
J.
C.
Scrivener.
1989.
Observed
and
simulated
effects
of
climatic
variability,
clear­
cut
logging,
and
fishing
on
the
numbers
of
chum
salmon
(
Oncorhynchus
keta)
and
coho
salmon
(
O.
kisutch)
returning
to
Carnation
Creek,
British
Columbia.
In:
C.
D.
Levings,
L.
B.
Holtby,
and
M.
A.
Henderson,
eds.
Proceedings
of
the
Natioanl
Workshop
on
effects
of
habitat
alternation
on
slamonid
stocks.
Can.
Spec.
Publ.
Fish.
Aquat.
Sci.
105:
62­
81.

Hynes,
H.
B.
N.
1970.
The
ecology
of
running
waters.
University
of
Toronto
Press,
Waterloo.
555
pp.

Iverson,
R.
A.
1972.
Effects
of
elevated
temperature
on
juvenile
coho
salmon
and
benthic
invertebrates
in
model
stream
communities.
Ph.
D.
thesis,
Oregon
State
University,
Corvallis,
OR.
98
pp.

Jobling,
M.
1981.
Temperature
tolerance
and
the
final
preferendum
 
rapid
methods
for
the
assessment
of
optimum
growth
temperatures.
J.
Fish
Biol.
19:
439­
455.

Keith,
R.
M.,
T.
C.
Bjornn,
W.
R.
Meehan,
N.
J.
Hetrick,
and
M.
A.
Brusven.
1998.
Response
of
juvenile
salmonids
to
riparian
and
insteam
cover
modifications
in
small
streams
flowing
through
secondgrowth
forests
of
S.
E.
Alaska.
Trans.
Am.
Fish.
Soc.
127:
889­
907.

Kitchell,
J.
F.,
J.
F.
Koonce,
R.
V.
O'Neill,
H.
J.
Shugart
Jr.,
J.
J.
Magnuson,
R.
S.
Booth.
1974.
Trans.
Am.
Fish.
Soc.
4:
786­
798.

Laufle,
J.
C.,
G.
B.
Pauley,
and
M.
F.
Shepard.
1986.
Species
profiles
and
environmental
requirements
of
coastal
fishes
and
invertebrates
(
Pacific
Northwest)
 
coho
salmon.
U.
S.
Fish
and
Wildlife
Service
Biological
Report
82
(
11.62).
18
pp.

Lee,
R.
M.
and
J.
N.
Rinne.
1980.
Critical
thermal
maxima
of
five
trout
species
in
the
Southwestern
United
States.
Trans.
Am.
Fish.
Soc.
109:
632­
635.

Linton,
T.
K.,
S.
D.
Reid
and
C.
M.
Wood.
1997.
The
metabolic
costs
and
physiological
consequences
to
juvenile
rainbow
trout
of
a
simulated
summer
warming
scenario
in
the
presence
and
absence
of
sublethal
ammonia.
Trans.
Am.
Fish.
Soc.
126:
259­
272.

Magnuson,
J.
J.,
L.
B.
Crowder
and
P.
A.
Medvick.
1979.
Temperature
as
an
ecological
resource.
Amer.
Zool.
19:
331­
343.

Mallet,
J.
P.,
S.
Charles,
J.
Persat,
and
P.
Auger.
1999.
Growth
modelling
in
accordance
with
daily
water
temperature
in
European
grayling
(
Thymallus
thymallus
L.)
Can.
J.
Fish.
Aquat.
Sci.
56:
994­
1000.

Martel,
G.
1996.
Growth
rate
and
influence
of
predation
risk
on
territoriality
in
juvenile
coho
salmon
(
Oncorhynchus
kisutch).
Can
J.
Fish.
Aq.
Sci.
53:
660­
669.

Martin,
D.
J.
1985.
Production
of
cutthroat
trout
(
Salmo
clarki)
in
relation
to
riparian
vegetation
in
Bear
Creek,
Washington.
Ph.
D.
Thesis,
University
of
Washington,
Seattle,
WA.

Martin,
D.
J.,
L.
J.
Wasserman,
and
V.
H.
Dale.
1986.
Influence
of
riparian
vegetation
on
post­
eruption
survival
of
coho
salmon
fingerlings
on
the
westside
streams
of
Mount
St.
Helens,
Washington.
N.
Am.
J.
Fish.
Manage.
6:
1­
8.

Mason,
J.
C.
1976.
Response
of
underyearling
coho
salmon
to
supplemental
feeding
in
a
natural
stream.
J.
Wildl.
Manage.
40(
4):
775­
788.

9­
6
Matthews,
K.
R.,
N.
H.
Berg,
D.
L.
Azuma
and
T.
R.
Lambert.
1994.
Cool
water
formation
and
trout
habitat
use
in
a
deep
pool
in
the
Sierra
Nevada,
California.
Trans.
Am.
Fish.
Soc.
123:
549­
564.

McCullough,
D.
A.
1999.
A
review
and
synthesis
of
effects
of
alterations
to
the
water
temperature
regime
on
freshwater
life
stages
of
salmonids
with
special
reference
to
chinook
salmon.
U.
S.
EPA
Report
#
EPA
910­
R­
010.
279
pp.

McLean,
W.
E.,
J.
O.
T.
Jensen,
and
D.
F.
Alderdice.
1993.
Oxygen
consumption
rates
and
water
flow
requirements
of
salmon
(
Oncorhynchus
spp.)
in
the
fish
culture
environment.
Aquaculture
109:
281­
313.

McConnell
and
Blahm
1970.
Unpublished
data
cited
in
National
Academy
of
Sciences
and
National
Academy
of
Engineering
(
NAS/
NAE).
1973.

McMahon,
T.
E.
1983.
Habitat
suitability
index
models:
Coho
salmon.
U.
S.
Dept.
Int.,
Fish
Wildl.
Serv.
FWS/
OBS­
82/
10.49.
29
pp.

Mount,
D.
I.,
N.
A.
Thomas,
T.
J.
Norbert,
M.
T.
Barbour,
T.
H.
Roush,
W.
F.
Brades.
1984.
Effluent
and
ambient
toxicity
testing
and
instream
community
and
response
on
the
Ottawa
River,
Lima,
Ohio.
EPA,
Office
of
Water
Enforcement
and
Permits,
Washington,
D.
C.
EPA­
600/
3­
84­
0890.

Murphy,
M.
L.,
C.
P.
Hawkins,
and
N.
H.
Anderson.
1981.
Effects
of
canopy
modification
and
accumulated
sediment
on
stream
communities.
Trans.
of
the
Am.
Fish.
Soc.
110:
469­
478.

Murphy,
M.
L.
1998.
Primary
production.
In:
Naiman,
R.
J.
and
R.
E.
Bilby,
eds.
River
ecology
and
management:
lessons
from
the
Pacific
coastal
region.
Springer,
New
York:
144­
168.

National
Academy
of
Sciences
and
National
Academy
of
Engineering
(
NAS/
NAE).
1973.
Water
quality
criteria
1972.
A
report
fo
the
Committee
on
Water
Quality
Criteria.
US
Environmental
protection
Agency
Pub.
No.
EPA­
R3­
73­
033.
Wasington,
D.
C.
553p.

National
Academy
of
Sciences
Committee
on
Protection
and
Management
of
Pacific
Northwest
Anadromous
Salmonids.
1996.
Upstream­
salmon
and
society
in
the
Pacific
Northwest.
National
Academy
Press,
Washington
D.
C.
452
pp.

Ney,
J.
T.
1993.
Bioenergetics
modeling
today:
growing
pains
on
the
cutting
edge.
Trans.
Am.
Fish.
Soc.
122:
736­
748.

Nielsen,
J.
L.
1992.
Microhabitat­
specific
foraging
behavior,
diet,
and
growth
of
juvenile
salmon.
Trans.
Am.
Fish.
Soc.
121:
617­
634.

Nielsen,
J.
L.,
T.
E.
Lisle,
and
V.
Ozaki.
1994.
Thermally
stratified
pools
and
their
use
by
steelhead
in
northern
California
streams.
Trans.
Am.
Fish.
Soc.
123:
613­
626.

Oregon
Department
of
Environmental
Quality
(
ODEQ).
1995.
Water
quality
standards
review.
Department
of
Environmental
Quality
Standards
&
Assessment
Section.
Portland,
OR.

Oregon
Department
of
Environmental
Quality
(
ODEQ).
1997.
Clean
Version.
Proposed
revisions
to
Division
122:
Hazardous
substance
remedial
action
rules.
Section
340­
122­
084
covers
risk
assessment.

Parkhurst,
B.
R.,
W.
Warren­
Hicks,
R.
D.
Cardwell,
J.
Volosin,
T.
Etchison,
J.
B.
Butcher,
and
S.
M.
Covington.
1996.
Methodology
for
aquatic
ecological
risk
assessment:
A
multi­
tiered
approach.
Project
91­
AER­
1.
Prepared
for
Water
Environment
Research
Foundation,
Alexandria,
Virginia.

9­
7
Pauley,
G.
B.,
B.
M.
Bortz,
and
M.
F.
Shepard.
1986.
Species
profiles
and
environmental
requirements
of
coastal
fishes
and
invertebrates
(
Pacific
Northwest)­­
Steelhead
Trout.
U.
S.
Fish
and
Wildlife
Service
Biological
Report
82
(
11.62).
24
pp.

Preall,
R.
J.
and
N.
H.
Ringler.
1989.
Comparison
of
actual
and
potential
growth
rates
of
brown
trout
(
Salmo
trutta)
in
natural
streams
based
on
bioenergetic
models.
Can.
J.
Fish.
Aquat.
Sci.
46:
1067­
1076.

Puckett,
K.
J.
and
L.
M.
Dill.
1985.
The
energetics
of
feeding
territorality
in
juvenile
coho
samon
(
Oncorhynchus
kisutch).
Behavior
42:
97­
111.

Quinn,
T.
and
N.
Peterson.
1996.
The
influence
of
habitat
complexity
and
fish
size
on
over­
winter
survival
and
growth
of
individually
marked
juvenile
coho
salmon
(
Oncorhynchus
kisutch)
in
Big
Beef
Creek,
Washington.
Can.
J.
Fish.
Aq.
Sci.
53:
1555­
1564.

Railsback,
S.
F.
1997.
Guidance
for
using
a
bioenergetics
model
to
assess
effects
of
stream
temperature.
Tech.
Report
001.8­
97­
8.
Pacific
Gas
and
Electric
Company,
San
Ramon
CA.

Railsback,
S.
F.
and
K.
A.
Rose.
1997.
Bioenergetics
modeling
of
stream
trout
growth:
temperature
and
food
consumption
effects.
Tech.
Report
001.8­
97­
8.
Pacific
Gas
and
Electric
Company,
San
Ramon
CA:
C1­
C25.

Railsback,
S.
F.
and
K.
A.
Rose.
1999.
Bioenergetics
modeling
of
stream
trout
growth:
temperature
and
food
consumption
effects.
Trans.
Am.
Fish.
Soc.
128:
241­
256.

Rand,
P.
S.,
D.
J.
Stewart,
P.
W.
Seelbach,
M.
L.
Jones,
and
L.
R.
Wedge.
1993.
Modeling
steelhead
population
energetics
in
Lakes
Michigan
and
Ontario.
Trans.
Am.
Fish.
Soc.
122:
415­
438.

Reeves,
G.
H.,
F.
H.
Everest,
and
J.
D.
Hall.
1987.
Interactions
between
the
redside
shiner
(
Richardsonius
balteatus)
and
the
steelhead
trout
(
Salmo
gairdneri)
in
western
Oregon:
the
influence
of
water
temperature.
Can.
J.
Fish.
Aquat.
Sci.
44:
1603­
1613.

Reiser,
D.
W.,
and
T.
C.
Bjornn.
1979.
Habitat
requirements
of
anadramous
salmonids.
U.
S.
D.
A.
For.
Serv.
Pac.
NW
For.
And
Range
Exp.
Station.
Gen.
Tech.
Rep
PNW­
96.

Ricker,
W.
E.
1975.
Computation
and
interpretation
of
biological
statistics
of
fish
populations.
Bulletin
of
the
Fisheries
Research
Board
of
Canada.
No.
91.
382
pp.

Sandercock,
F.
K.
1995.
Life
history
of
coho
salmon.
In:
C.
Groot
and
L.
Margolis
(
eds.).
Pacific
Salmon
Life
Histories.
UBC
Press,
Vancouver,
B.
C:
395­
446.

SAS
Institute
Inc.
(
1989)
SAS/
STAT
User's
Guide,
Version
6,
Fourth
Edition.
Cary,
NC:
SAS
Institute
Inc.

SETAC
[
Society
for
Environmental
Chemistry
and
Toxicology].
1994.
Aquatic
risk
assessment
and
mitigation
dialogue
group.
Final
report.
SETAC
Foundation
for
Environmental
Education,
Inc.,
Pensacola,
Florida.
USA.

Shelbourn,
J.
E.
,
J.
R.
Brett,
and
S.
Shirahata.
1973.
Effect
of
temperature
and
feeding
regime
on
the
specific
growth
rate
of
sockeye
salmon
fry
(
Oncorhynchus
nerka)
with
a
consideration
of
size
effect.
J.
Fish.
Res.
Bd.
Canada.
30:
1191­
1194.

Shuter,
B.
J.
,
J.
A.
MacLean,
F.
E.
J.
Fry
and
H.
A.
Regier.
1980.
Stochastic
simulation
of
temperature
effects
on
first­
year
survival
of
smallmouth
bass.
Trans.
Am.
Fish.
Soc.
109:
1­
34.

9­
8
Sokal,
R.
R.
and
F.
J.
Rohlf.
1981.
Biometry.
The
principles
and
practice
of
statistics
in
biological
research.
Second
Edition.
W.
H.
Freeman
and
Company,
New
York.
859
pp.

Solomon,
K.
R.,
D.
B.
Baker,
R.
P.
Richards,
K.
R.
Dixon,
S.
J.
Klaine,
T.
W.
La
Point,
R.
J.
Kendall,
C.
P.
Weisskopf,
J.
M.
Giddings,
J.
P.
Giesy,
L.
W.
Hall,
Jr.,
and
W.
M.
Williams.
1996.
Ecological
risk
assessment
of
atrazine
in
North
American
surface
waters.
Environmental
Toxicology
and
Chemistry.
15:
31­
76.

Spehar,
R.
and
W.
Adams.
1998.
Person
communications
with
Robert
Spehar,
U.
S.
EPA,
Duluth,
Minnesota,
and
D.
William
Adams,
U.
S.
EPA
Science
Advisory
Board,
November
1998
telephone
conversations.

Spence,
B.
C.,
G.
A.
Lomnicky,
R.
M.
Hughes,
and
R.
P.
Novitzki.
1996.
An
ecosystem
approach
to
salmonid
conservation.
TR­
4501­
96­
6057.
ManTech
Environmental
Research
Services
Corp.
Corallis
OR.
(
Available
from
the
National
Marine
Fisheries
Service,
Portland,
Oregon.)

Spigarelli,
S.
A.,
M.
M.
Thommes
and
W.
Prepejchal.
1982.
Feeding,
growth,
and
fat
deposition
by
brown
trout
in
constant
and
fluctuating
temperatures.
Trans.
Am.
Fish.
Soc.
111:
199­
209.

Sprague,
J.
B.
1969.
Measurement
of
pollutant
toxicity
to
fish.
I.
Bioassay
methods
for
acute
toxicity.
Water
Research
3:
793­
821.

Stephan,
C.
E.,
D.
I.
Mount,
D.
J.
Hansen,
J.
H.
Gentile,
G.
A.
Chapman,
and
W.
A.
Brungs.
1985.
Guidelines
for
deriving
numerical
national
water
quality
criteria
for
the
protection
of
aquatic
organisms
and
their
uses.
U.
S.
Environmental
Protection
Agency,
Washington,
D.
C.
NTIS
No.
PB85­
227049.
98
pages.

Stewart,
D.
J.,
and
M.
Ibarra.
1991.
Predation
and
production
of
salmonine
fishes
in
Lake
Michigan,
1978­
88.
Can.
J.
Fish.
Aq.
Sci.
48:
909­
922.

Sullivan,
K.,
S.
H.
Duncan,
P.
A.
Bisson,
J.
T.
Heffner,
J.
W.
Ward,
R.
E.
Bilby,
and
J.
L.
Nielsen.
1987.
A
summary
report
of
the
Deschutes
River
basin:
sediment,
flow,
temperature,
and
fish
habitat.
Technical
Report
044­
5002/
87/
1,
Weyerhaeuser
Company,
Tacoma,
WA
129
pp.

Sullivan,
K.
and
T.
A.
Adams.
1990.
The
physics
of
stream
heating:
An
analysis
of
temperature
patterns
in
streams
environments
based
on
physical
principles
and
stream
data.
Weyerhaeuser
Technical
Report.
044­
5002.90/
2.
Technology
Center,
Tacoma,
WA
98477.

Sullivan
.
K.,
J.
Tooley,
K.
Doughty,
J.
E.
Caldwell,
and
P.
Knudsen.
1990.
Evaluation
of
prediction
models
and
characterization
of
stream
temperature
regimes
in
Washington.
Timber/
Fish/
Wildlife
Report
No.
TFW­
WQ3­
90­
006.
Washington
Dept.
Nat.
Resources,
Olympia,
WA.
224
pp.

Suter,
G.
W.
II,
L.
W.
Barnthouse,
S.
M.
Bartell,
T.
Mill,
D.
Mackay,
and
S.
Paterson.
1993.
Ecological
risk
assessment.
Lewis
Publishers,
Boca
Raton,
Florida,
538
pp.

Suter,
G.
W.
and
J.
B.
Mabrey.
1994.
Toxicological
benchmarks
for
screening
potential
contaminants
of
concern
for
effects
on
aquatic
biota:
1994
revision.
Oak
Ridge
National
Laboratory,
NTIS
No.
DE95oo5996/
WEP.
106
pp.

Sweeney,
Bernard
W.
and
Robin
L.
Vannote.
1986.
Growth
and
production
of
a
stream
stonefly:
influences
of
diet
and
temperature.
Ecology
67(
5):
1396­
1410.

Symons,
P.
1971.
Behavioral
adjustment
of
population
density
to
available
food
by
juvenile
Atlantic
salmon.
J.
Anim.
Ecol.
40:
569­
587.

9­
9
Tang,
M.
and
D.
Boisclair.
1995.
Relationship
between
respiration
rate
of
juvenile
brook
trout
(
Salvilinus
fontinalis),
water
temperature,
and
swimming
characteristics.
Can.
J.
Fish.
Aquat.
Sci.
52:
2138­
2145.

Taniguchi,
Y.,
F.
J.
Rahel,
D.
C.
Novinger,
and
K.
G.
Gerow.
1998.
Temperature
mediation
of
competitive
interactions
among
three
fish
species
that
replace
each
other
along
longitudinal
stream
gradients.
Can.
J.
Fish.
Aquat.
Sci.
55:
1894­
1901.

Theurer,
F.
D.
,
K.
A.
Voos,
and
W.
J.
Miller.
1984.
Instream
water
temperature
model.
Instream
Flow
Info.
Paper
No.
16.
U.
S.
D.
I.
Fish
and
Wildlife
Service
FWS/
OBS­
84/
15.

Thomas,
R.
E.,
J.
A.
Gharrett,
M.
G.
Carls,
S.
D.
Rice,
A.
Moles,
and
S.
Korn.
1986.
Effects
of
fluctuating
temperature
on
mortality,
stress,
and
energy
reserves
of
juvenile
coho
salmon.
Trans.
Am.
Fish.
Soc.
115:
52­
59.

Thornton,
K.
W.
and
A.
S.
Lessem.
1978.
A
temperature
algorithm
for
modifying
biological
rates.
Trans.
Am.
Fish.
Soc..
107:
284­
287.

U.
S.
EPA.
1977.
Temperature
criteria
for
freshwater
fish:
protocol
and
procedures.
Ecological
Research
Series.
U.
S.
Environmental
Protection
Agency
Office
of
Research
and
Development,
Environmental
Research
Laboratory,
Duluth,
Minnesota.
EPA­
600/
3­
77­
061.
130p.
[
Also
cited
as
Brungs
and
Jones,
1977].

U.
S.
EPA.
1980.
Ambient
water
quality
criteria
for
polynuclear
aromatic
hydrocarbons.
EPA­
440/
5­
80/
069.
Available
from
National
Technical
Information
Service,
Springfield,
Virginia.
186
pp.

U.
S.
EPA.
1990.
Guidance
for
data
useability
in
risk
assessment.
Office
of
Emergency
and
Remedial
Response.
EPA/
540/
G­
90/
008.
261
pp.

U.
S.
EPA.
1992.
Framework
for
ecological
risk
assessment.
Risk
Assessment
Forum,
U.
S.
Environmental
Protection
Agency,
Washington,
D.
C.
EPA/
630/
R­
92/
001.
41
pp.

U.
S.
EPA.
1993.
Interim
report
on
data
and
methods
for
assessment
of
2,3,7,8­
tetrachlorodibenzo­
p­
dioxin
risks
to
aquatic
life
and
associated
wildlife.
U.
S.
Environmental
Protection
Agency,
Environmental
Res.
Lab.,
Duluth,
MN.
EPA/
600/
R­
93­
055.

U.
S.
EPA.
1994.
Water
quality
standards
handbook.
Second
edition.
EPA,
Office
of
Water,
Washington,
D.
C.
EPA­
823­
B­
94­
005a.

U.
S.
EPA.
1995.
Draft
proposed
guidelines
for
ecological
risk
assessment.
U.
S.
Environmental
Protection
Agency,
Risk
Assessment
Forum,
Washington,
D.
C.
EPA/
630/
R­
95/
002.
144
pp.

Vigg,
S.
C.
and
D.
L.
Koch.
1980.
Upper
lethal
temperature
range
of
Lahontan
cutthroat
trout
in
waters
of
different
ionic
concentration.
Trans.
Am.
Fish.
Soc.
109:
336­
339.

Walters,
C.
J.,
and
J.
R.
Post.
Density­
dependent
growth
and
competitive
asymmetries
in
size­
structured
fish
populations:
a
theoretical
model
and
recommendations
for
field
experiments.
Trans.
Am.
Fish.
Soc.
122:
34­
35.

Warren,
C.
E.
1971.
Biology
and
water
pollution
control.
W.
B.
Saunders
Company,
Philadelphia.
434
pp.

Washington
Departement
of
Ecology
(
WDOE).
1999a.
Draft:
Evaluating
standards
for
protecting
aquatic
life
in
Washington's
surface
water
quality
standards.
Temperature
criteria.
Preliminary
Review
Draft
Discussion
Paper.
Appendix.
54
pages.,
Washington
Department
of
Ecology,
Olympia,
WA.

9­
10
Washington
Departement
of
Ecology
(
WDOE).
1999b.
Draft:
Evaluating
standards
for
protecting
aquatic
life
in
Washington's
surface
water
quality
standards.
Temperature
criteria.
Preliminary
Review
Draft
Discussion
Paper.
Supplementary
Appendix.
122
pages.
,
Washington
Department
of
Ecology,
Olympia,
WA.

Weatherly,
A.
H.
1972.
Growth
and
ecology
of
fish
populations.
Academic
Press,
London.
285
pp.

Weatherly,
A.
H.
and
H.
S.
Gill.
1995.
Growth.
Pages
101­
158
In:
[
ed]
C.
Groot,
L.
Margolis,
and
W.
C.
Clarke.
Physiological
ecology
of
Pacific
Salmon.
UBC
Press,
Vancouver,
B.
C.
Canada.

Weatherly,
N.
S.
and
S.
J.
Ormerod.
1990.
Forests
and
the
temperature
of
upland
streams
in
Wales:
a
modelling
exploration
of
the
biological
effects.
Freshwater
Biology
24:
109­
122.

Wilzbach,
M.
A.
1985.
Relative
roles
of
food
abundance
and
cover
in
determining
the
habitat
distribution
of
stream­
dwelling
cutthroat
trout
(
Salmo
clarki).
Can.
J.
Fish.
and
Aquat.
Sci.
42:
1668­
1672.

Winberg,
G.
G.
1971.
Symbols,
units,
and
conversion
factors
in
studies
o
freshwater
productivity.
International
Biological
Program,
London.

Wurtsbaugh,
W.
A.
1973.
Effects
of
temperature,
ration,
and
size
on
the
growth
of
juvenile
steelhead
trout,
Salmo
gairdneri.
M.
S.
Thesis,
Oregon
State
University,
Corvallis,
OR.
69.
pp.

Wurtsbaugh,
W.
A.
and
G.
E.
Davis.
1977.
Effects
of
temperature
and
ration
level
on
the
growth
and
food
conversion
efficiency
of
Salmo
gairdneri,
Richardson.
J.
Fish.
Biol.
11:
87­
98.

Young,
W.
D.
and
D.
S.
Robson.
1978.
Estimation
of
population
number
and
mortality
rates.
Pages
137­
164
In:
T.
Bagenal
(
ed)
Methods
for
assessment
of
fish
production
in
fresh
waters.
IBP
Handbook
No
3.
Blackwell
Scientific,
Oxford
London.

9­
11
Section
10­
1
SECTION
10
APPENDICES
Appendix
A.
Data
used
to
develop
relationships
between
growth
rate,
temperature
and
consumption
determined
from
laboratory
studies                 
10­
2
Appendix
B.
Regional
temperature
data
for
the
Pacific
Northewest
Region
from
U.
S.
Geological
Survey
Water
Resources
Data.
                   
10­
11
Appendix
C.
Acute
effects
of
temperature
on
salmon
and
trout:
data
used,
analyses
and
assumptions                    
   .
Under
Separate
Cover
Section
10­
2
APPENDIX
A
DATA
USED
TO
DEVELOP
RELATIONSHIPS
BETWEEN
GROWTH
RATE,
TEMPERATURE
AND
RATION
AND
CONSUMPTION
DETERMINED
FROM
LABORATORY
STUDIES
Coho
salmon
Steelhead
trout
Section
10­
3
DESCRIPTION
OF
METHODOLOGY
The
temperature
risk
assessment
relies
on
growth/
ration/
temperature
relationships
developed
in
laboratory
experiments.
In
these
experiments,
the
growth
of
individual
or
populations
of
fish
are
tracked
over
intervals
of
time
under
known
temperature
and
food
ration.
Previously
published
studies
were
used
to
either
obtain
growth
curves
already
developed
by
the
authors
(
sockeye
and
chinook
salmon),
or
to
develop
curves
from
original
data
provided
by
researchers
(
coho
salmon
and
steelhead
trout).
In
the
studies
used
in
this
analysis,
temperatures
were
maintained
at
constant
levels
for
the
duration
of
each
trial,
and
various
levels
of
food
were
provided
as
treatments.
Trials
were
repeated
at
several
levels
of
temperature.

Data
used
for
growth
of
coho
salmon
was
taken
from
Everson
(
1973).
Growth
of
juvenile
coho
was
studied
in
60
trials
where
temperature
was
varied
from
11.1o
to
22.4oC
and
food
ration
was
varied
from
satiation
to
near
starvation.
Experiments
were
replicated
in
1969
and
1970.
Everson
reported
growth
rates
of
individual
fish.
Table
A.
1
shows
the
trial
averages.
The
average
for
each
trial
at
each
temperature/
treatment
replication
were
used
in
to
develop
specific
growth
curves
in
the
main
body
of
the
report.

Data
used
for
growth
of
steelhead
analysis
taken
from
Wurtsbaugh
and
Davis
(
1977).
Growth
of
juvenile
steelhead
was
studied
in
the
laboratory.
A
total
of
44
trials
were
completed
over
the
course
of
a
year,
varying
temperature
from
6.9o
to
22.5oC
and
ration
over
the
range
from
satiation
to
near
starvation.
Table
A.
2
shows
data
for
each
trial.

Both
studies
were
conducted
at
Oregon
State
University,
and
local
stocks
were
used.
In
both
experiments,
the
fish
were
acclimated
for
approximated
14­
16
days
before
trials,
and
fasted
for
48
hours
before
tests
began.
The
fish
were
fed
for
23
days,
and
their
weights
were
measured
after
25
days.
Section
10­
4
Abbreviations
used
in
the
tables
for
coho:

Variable
Information
Season
Trials
were
run
during
5
intervals
of
time
over
a
1
year
time
span
(
1969­
70)

Temperature
Temperature
was
daily
mean
temperature
averaged
for
the
30
day
interval
of
the
trial
Food
Level
At
each
temperature,
4
food
trials
ranging
from
near
starvation
to
satiation
were
run..
Each
trial
had
3­
4
fish.
Rep
Rep
appears
on
the
Everson
all
sheet.
It
is
the
data
for
each
of
the
3­
4
fish
in
the
trial.

Initial
Wet
Weight
(
g)
Starting
weight
of
the
fish,
measured.

Final
Wet
Weight
(
g)
Ending
weight
of
the
fish,
measured.

Initial
Dry
Weight
(
g)
Starting
dry
weight
of
the
fish,
estimated
based
on
dry
to
weight
weight
factor,
Table
1
Final
Dry
Weight
(
g)
Ending
dry
weight
of
the
fish,
measured
after
sacrificing
the
fish.

Dry
Food
(
mg)
Measured
dry
weight
of
the
food
fed
to
the
fish
during
the
trial.

Wet
Food
(
mg)
Estimated
wet
weight
of
the
food,
based
on
factors
in
Table
1
Init
Weight
Wet
Weight
Consumption
(
g/
g/
day
Estimated
consumption
per
day
for
wet
weight
of
fish
relative
to
the
initial
wet
weight
at
beginning
of
trial.

MidWeight
Wet
Weight
Consumption
(
g/
g/
day)
Estimated
consumption
per
day
for
wet
weight
of
fish
relative
to
the
midpoint
between
start
and
end
wet
weight.

Init
Weight
Dry
Consumption
g/
g/
day
Estimated
consumption
per
day
for
dry
weight
of
fish
relative
to
the
initial
dry
weight
at
beginning
of
trial.

Mid
Weight
Dry
Consumption
g/
g/
day
Estimated
consumption
per
day
for
dry
weight
of
fish
relative
to
the
midpoint
between
start
and
end
dry
weight.

%
Fish
Moisture
Dry/
Wet
Fish
moisture
factor,
from
Table
1
Daily
Growth
Rate
 
Wet
(
g/
g/
d)
Growth
rate
calculated
as
(
end
weight­
start
weight)(.
5*(
start
weight+
endweight)/
30
using
wet
weights
Daily
Growth
Rate
 
Dry
(
g/
g/
d)
Growth
rate
calculated
as
(
end
weight­
start
weight)(.
5*(
start
weight+
endweight)/
30
using
dry
weights
Note
that
rations
and
growth
rates
can
be
expressed
as
grams
of
body
weight
per
day,
referenced
as
g/
g/
d,
g
g­
1
d­
1,
or
g/
g/
day.
These
are
also
sometimes
referenced
as
%
body
Section
10­
5
Table
A.
1
Experimental
data
for
growth
of
coho
salmon
in
relation
to
temperature
and
food
consumption.
Data
is
from
Everson,
1973.

Trial
Fish
Size
Tempera
ture
Food
Level
Initial
Wet
Weight
(
g)
Final
Wet
Weight
(
g)
Initial
Dry
Weight
(
g)
Final
Dry
Weight
(
g)
Dry
Food
(
mg)
Wet
Food
(
mg)
Init
Weight
Wet
Weight
Consumpt
ion
(
g/
g/
day
Mid
Weight
Wet
Weight
Consumption
(
g/
g/
day)
Init
Weight
Dry
Consum
ption
g/
g/
day
Mid
Weight
Dry
Consump
tion
g/
g/
day
%
Fish
Moisture
Dry/
Wet
Wet
Daily
Growth
Rate
(
g/
g/
d)
Dry
Daily
Growth
Rate
(
g/
g/
d)

Trial
1
Med
14.5
1
1.68
1.78
0.30
0.36
344
480
0.0098
0.0095
0.0395
0.035
0.1780
0.0019
0.0066
Trial
2
Med
14.5
2
1.88
2.82
0.34
0.60
924
1289
0.0235
0.0185
0.0946
0.067
0.1780
0.0135
0.0187
Trial
3
Med
14.5
3
1.69
2.91
0.30
0.63
1297
1809
0.0457
0.0310
0.1841
0.111
0.1780
0.0197
0.0251
Trial
4
Med
14.5
4
1.43
2.64
0.25
0.57
1344
1875
0.0443
0.0309
0.1783
0.109
0.1780
0.0200
0.0256
Trial
5
Med
18.6
1
1.58
1.60
0.28
0.30
345
481
0.0103
0.0102
0.0414
0.040
0.1780
0.0003
0.0020
Trial
6
Med
18.6
2
1.82
2.22
0.32
0.49
1023
1426
0.0269
0.0241
0.1083
0.086
0.1780
0.0068
0.0139
Trial
7
Med
18.6
3
3.09
3.91
0.55
0.91
1657
2311
0.0253
0.0222
0.1018
0.077
0.1780
0.0079
0.0165
Trial
8
Med
18.6
4
2.64
3.37
0.47
0.79
1912
2666
0.0350
0.0302
0.1410
0.103
0.1780
0.0084
0.0172
Trial
9
Med
20.8
1
1.33
1.21
0.24
0.23
342
477
0.0122
0.0128
0.0491
0.050
0.1780
­
0.0032
­
0.0016
Trial
10
Med
20.8
2
1.62
2.21
0.29
0.47
957
1334
0.0278
0.0234
0.1121
0.084
0.1780
0.0105
0.0164
Trial
11
Med
20.8
3
1.68
2.66
0.30
0.62
1392
1942
0.0397
0.0307
0.1597
0.106
0.1780
0.0147
0.0223
Trial
12
Med
20.8
4
2.23
3.75
0.40
0.87
2008
2801
0.0433
0.0315
0.1745
0.107
0.1780
0.0174
0.0252
Trial
13
Med
11.1
1
1.77
2.29
0.34
0.45
439
608
0.0116
0.0101
0.0434
0.037
0.1940
0.0086
0.0093
Trial
14
Med
11.1
2
1.92
2.94
0.37
0.62
793
1096
0.0195
0.0153
0.0728
0.054
0.1940
0.0142
0.0169
Trial
15
Med
11.1
3
1.67
2.62
0.32
0.56
949
1312
0.0274
0.0208
0.1022
0.073
0.1940
0.0151
0.0183
Trial
16
Med
11.1
4
1.81
3.38
0.35
0.73
1262
1745
0.0329
0.0227
0.1226
0.078
0.1940
0.0201
0.0233
Trial
17
Med
15.2
1
2.03
2.35
0.39
0.47
445
615
0.0106
0.0098
0.0396
0.075
0.1940
0.0051
0.0063
Trial
18
Med
15.2
2
1.77
2.44
0.34
0.53
764
1057
0.0202
0.0168
0.0751
0.075
0.1940
0.0108
0.0144
Trial
19
Med
15.2
3
1.49
2.86
0.29
0.61
1088
1504
0.0339
0.0232
0.1262
0.075
0.1940
0.0209
0.0237
Trial
20
Med
15.2
4
1.71
3.04
0.33
0.66
1256
1737
0.0345
0.0247
0.1288
0.075
0.1940
0.0186
0.0217
Trial
21
Med
17.8
1
1.59
1.82
0.31
0.36
443
612
0.0132
0.0122
0.0491
0.075
0.1940
0.0048
0.0054
Trial
22
Med
17.8
2
1.85
2.52
0.36
0.53
796
1101
0.0218
0.0181
0.0811
0.075
0.1940
0.0108
0.0133
Trial
23
Med
17.8
3
1.74
2.88
0.34
0.63
1071
1481
0.0286
0.0216
0.1067
0.075
0.1940
0.0160
0.0196
Trial
24
Med
17.8
4
2.42
3.96
0.47
0.89
1413
1954
0.0283
0.0212
0.1054
0.075
0.1940
0.0164
0.0210
Section
10­
6
Table
A.
1
Continued.
Experimental
data
for
growth
of
coho
salmon
in
relation
to
temperature
and
food
consumption.
Data
is
from
Everson,
1973.

Trial
Fish
Size
Tempera
ture
Food
Level
Initial
Wet
Weight
(
g)
Final
Wet
Weight
(
g)
Initial
Dry
Weight
(
g)
Final
Dry
Weight
(
g)
Dry
Food
(
mg)
Wet
Food
(
mg)
Init
Weight
Wet
Weight
Consumpt
ion
(
g/
g/
day
Mid
Weight
Wet
Weight
Consumption
(
g/
g/
day)
Init
Weight
Dry
Consump
tion
g/
g/
day
Mid
Weight
Dry
Consump
tion
g/
g/
day
%
Fish
Moisture
Dry/
Wet
Wet
Daily
Growth
Rate
(
g/
g/
d)
Dry
Daily
Growth
Rate
(
g/
g/
d)

Trial
25
Large
9.4
1
4.11
4.46
0.86
0.95
472
633
0.0051
0.0049
0.0182
0.075
0.2100
0.0027
0.0031
Trial
26
Large
9.4
2
4.08
4.67
0.86
1.00
711
955
0.0078
0.0073
0.0277
0.075
0.2100
0.0045
0.0051
Trial
27
Large
9.4
3
3.70
4.59
0.78
0.99
873
1172
0.0106
0.0094
0.0374
0.075
0.2100
0.0072
0.0078
Trial
28
Large
9.4
4
3.80
4.99
0.80
1.09
1122
1505
0.0132
0.0114
0.0469
0.075
0.2100
0.0091
0.0103
Trial
29
Large
13.1
1
2.85
2.81
0.60
0.56
313
419
0.0049
0.0049
0.0174
0.075
0.2100
­
0.0004
­
0.0024
Trial
30
Large
13.1
2
3.30
3.67
0.69
0.76
569
764
0.0077
0.0073
0.0273
0.075
0.2100
0.0034
0.0031
Trial
31
Large
13.1
3
3.83
4.69
0.80
0.99
902
1211
0.0105
0.0095
0.0374
0.075
0.2100
0.0067
0.0068
Trial
32
Large
13.1
4
3.57
4.55
0.75
0.97
1066
1430
0.0133
0.0117
0.0473
0.075
0.2100
0.0079
0.0085
Trial
33
Large
15.8
1
2.96
2.87
0.62
0.56
330
443
0.0050
0.0051
0.0177
0.075
0.2100
­
0.0012
­
0.0036
Trial
34
Large
15.8
2
3.10
3.20
0.65
0.64
467
627
0.0069
0.0068
0.0246
0.075
0.2100
0.0011
­
0.0003
Trial
35
Large
15.8
3
3.29
3.79
0.69
0.79
761
1022
0.0104
0.0096
0.0368
0.075
0.2100
0.0048
0.0047
Trial
36
Large
15.8
4
3.59
4.65
0.75
1.00
1099
1475
0.0137
0.0119
0.0487
0.075
0.2100
0.0086
0.0094
Trial
37
Small
11.4
1
0.70
0.73
0.12
0.12
142
208
0.0098
0.0098
0.0396
0.075
0.1700
0.0002
­
0.0011
Trial
38
Small
11.4
2
0.76
0.90
0.13
0.16
202
295
0.0130
0.0119
0.0524
0.075
0.1700
0.0057
0.0065
Trial
39
Small
11.4
3
0.72
0.93
0.12
0.16
235
343
0.0158
0.0139
0.0636
0.075
0.1700
0.0081
0.0085
Trial
40
Small
11.4
4
0.69
1.01
0.12
0.18
321
469
0.0228
0.0185
0.0916
0.075
0.1700
0.0126
0.0141
Trial
41
Small
14.8
1
0.77
0.66
0.13
0.11
144
211
0.0091
0.0099
0.0366
0.075
0.1700
­
0.0056
­
0.0083
Trial
42
Small
14.8
2
0.58
0.55
0.10
0.09
145
212
0.0121
0.0125
0.0487
0.075
0.1700
­
0.0024
­
0.0034
Trial
43
Small
14.8
3
0.56
0.60
0.09
0.10
170
249
0.0149
0.0143
0.0598
0.075
0.1700
0.0023
0.0011
Trial
44
Small
14.8
4
0.63
0.75
0.11
0.13
267
390
0.0206
0.0188
0.0827
0.075
0.1700
0.0056
0.0061
Trial
45
Small
17.2
1
0.82
0.75
0.14
0.13
161
235
0.0096
0.0100
0.0385
0.075
0.1700
­
0.0029
­
0.0029
Trial
46
Small
17.2
2
0.65
0.54
0.11
0.08
139
203
0.0104
0.0114
0.0419
0.075
0.1700
­
0.0059
­
0.0094
Trial
47
Small
17.2
3
0.69
0.76
0.12
0.13
210
307
0.0149
0.0142
0.0599
0.075
0.1700
0.0030
0.0040
Trial
48
Small
17.2
4
0.74
0.89
0.13
0.16
277
405
0.0181
0.0165
0.0729
0.075
0.1700
0.0057
0.0068
Trial
49
Med
15.8
1
1.65
1.86
0.31
0.37
352
452
0.0091
0.0086
0.0376
0.075
0.1890
0.0037
0.0049
Trial
50
Med
15.8
2
1.83
2.38
0.35
0.53
643
825
0.0150
0.0131
0.0618
0.075
0.1890
0.0084
0.0131
Section
10­
7
Table
A.
1
Continued.
Experimental
data
for
growth
of
coho
salmon
in
relation
to
temperature
and
food
consumption.
Data
is
from
Everson,
1973.

Trial
Fish
Size
Tempera
ture
Food
Level
Initial
Wet
Weight
(
g)
Final
Wet
Weight
(
g)
Initial
Dry
Weight
(
g)
Final
Dry
Weight
(
g)
Dry
Food
(
mg)
Wet
Food
(
mg)
Init
Weight
Wet
Weight
Consumpt
ion
(
g/
g/
day
Mid
Weight
Wet
Weight
Consumption
(
g/
g/
day)
Init
Weight
Dry
Consump
tion
g/
g/
day
Mid
Weight
Dry
Consump
tion
g/
g/
day
%
Fish
Moisture
Dry/
Wet
Wet
Daily
Growth
Rate
(
g/
g/
d)
Dry
Daily
Growth
Rate
(
g/
g/
d)

Trial
51
Med
15.8
3
1.70
2.49
0.32
0.55
823
1056
0.0207
0.0168
0.0854
0.075
0.1890
0.0126
0.0174
Trial
52
Med
15.8
4
1.93
3.14
0.36
0.73
1202
1543
0.0267
0.0203
0.1101
0.075
0.1890
0.0160
0.0221
Trial
53
Med
19.2
1
1.53
1.56
0.29
0.30
316
406
0.0088
0.0088
0.0364
0.075
0.1890
0.0005
0.0012
Trial
54
Med
19.2
2
1.84
2.22
0.35
0.48
626
804
0.0145
0.0132
0.0600
0.075
0.1890
0.0062
0.0102
Trial
55
Med
19.2
3
1.60
2.22
0.30
0.49
759
975
0.0203
0.0170
0.0838
0.075
0.1890
0.0108
0.0153
Trial
56
Med
19.2
4
1.64
2.62
0.31
0.65
1029
1320
0.0268
0.0207
0.1106
0.075
0.1890
0.0152
0.0226
Trial
57
Med
22.4
1
1.97
1.93
0.37
0.38
404
518
0.0087
0.0088
0.0361
0.075
0.1890
­
0.0007
0.0007
Trial
58
Med
22.4
2
1.54
1.69
0.29
0.33
395
507
0.0109
0.0104
0.0451
0.075
0.1890
0.0032
0.0037
Trial
59
Med
22.4
3
1.62
2.03
0.31
0.45
742
952
0.0196
0.0174
0.0808
0.075
0.1890
0.0075
0.0124
Trial
60
Med
22.4
4
1.84
2.36
0.35
0.51
844
1084
0.0201
0.0173
0.0830
0.075
0.1890
0.0084
0.0126
Section
10­
8
Table
A.
2
Experimental
data
for
growth
of
steelhead
trout
in
relation
to
temperature
and
food
consumption.
Data
is
from
Wurtsbaugh
and
Davis,
1977.

Trial
Temperat
ure
o
C
Season
Initial
Wet
Weight
(
g)
Final
Wet
Weight
(
g)
Initial
Dry
Weight
(
g)
Final
Dry
Weight
(
g)
Dry
Weight
Consumption
(
g/
g/
d)
Mid
Weight
Dry
Consumption
(
g/
g/
d)
%
Fish
Moisture
Dry/
Wet
Final
Dry
Weight
%
Dry
Daily
Growth
Rate
(
g/
g/
d)
Wet
Daily
Growth
Rate
(
g/
g/
d)

Trial
1
6.9
Winter
2.03
1.82
0.443
0.337
0.5
0.04870
0.218
0.185
­
0.011
­
0.436
Trial
2
6.9
Winter
2.01
1.96
0.438
0.380
1.5
0.15345
0.218
0.194
­
0.006
­
0.101
Trial
3
6.9
Winter
1.95
2.1
0.425
0.456
2.5
0.27525
0.218
0.217
0.003
0.296
Trial
4
6.9
Winter
1.91
2.3
0.416
0.534
4.2
0.49874
0.218
0.232
0.010
0.741
Trial
5
9.4
Spring
2.35
2.35
0.470
0.430
2.5
0.28127
0.2
0.183
­
0.004
0.000
Trial
6
9.4
Spring
2.25
2.71
0.450
0.526
4.9
0.59764
0.2
0.194
0.006
0.742
Trial
7
9.4
Spring
2.36
3.19
0.472
0.657
7
0.98800
0.2
0.206
0.013
1.196
Trial
8
9.4
Spring
2.23
3.38
0.446
0.690
10.2
1.44779
0.2
0.204
0.017
1.640
Trial
9
10
Autumn
0.98
0.96
0.190
0.177
2.2
0.10086
0.194
0.184
­
0.003
­
0.082
Trial
10
10
Autumn
1.04
1.31
0.202
0.261
4.3
0.24857
0.194
0.199
0.010
0.919
Trial
11
10
Autumn
1.02
1.64
0.198
0.341
7.9
0.53226
0.194
0.208
0.021
1.865
Trial
12
10
Autumn
1
1.89
0.194
0.412
14
1.06054
0.194
0.218
0.029
2.464
Trial
13
10.1
Winter
1.96
1.66
0.414
0.300
0.6
0.05355
0.211
0.181
­
0.013
­
0.663
Trial
14
10.1
Winter
1.96
1.82
0.414
0.351
1.5
0.14340
0.211
0.193
­
0.007
­
0.296
Trial
15
10.1
Winter
1.97
2.02
0.416
0.396
2.7
0.27391
0.211
0.196
­
0.002
0.100
Trial
16
10.1
Winter
1.94
2.35
0.409
0.496
4.9
0.55443
0.211
0.211
0.008
0.765
Trial
17
12.6
Spring
2.29
2.29
0.463
0.428
3.8
0.42313
0.202
0.187
­
0.003
0.000
Trial
18
12.6
Spring
2.33
2.82
0.471
0.541
6.1
0.77173
0.202
0.192
0.006
0.761
Trial
19
12.6
Spring
2.24
3.17
0.452
0.650
9.1
1.25390
0.202
0.205
0.014
1.375
Trial
20
12.6
Spring
2.28
3.75
0.461
0.810
12.7
2.01701
0.202
0.216
0.022
1.950
Trial
21
13
Winter
1.92
1.7
0.394
0.296
2
0.17235
0.205
0.174
­
0.011
­
0.486
Trial
22
13
Winter
1.84
1.72
0.377
0.316
3
0.26013
0.205
0.184
­
0.007
­
0.270
Trial
23
13
Winter
1.86
2.21
0.381
0.477
5.7
0.61180
0.205
0.216
0.009
0.688
Trial
24
13.3
Autumn
1
0.93
0.199
0.169
2.2
0.10127
0.199
0.182
­
0.006
­
0.290
Section
10­
9
Table
A.
2
Continued.
Experimental
data
for
growth
of
steelhead
trout
in
relation
to
temperature
and
food
consumption.
Data
is
from
Wurtsbaugh
and
Davis,

1977.
Trial
Temperat
ure
o
C
Season
Initial
Wet
Weight
(
g)
Final
Wet
Weight
(
g)
Initial
Dry
Weight
(
g)
Final
Dry
Weight
(
g)
Dry
Weight
Consumpti
on
(
g/
g/
d)
Mid
Weight
Dry
Consumptio
n
(
g/
g/
d)
%
Fish
Moisture
Dry/
Wet
Final
Dry
Weight
%
Dry
Daily
Growth
Rate
(
g/
g/
d)
Wet
Daily
Growth
Rate
(
g/
g/
d)

Trial
25
13.3
Autumn
0.98
1.17
0.195
0.227
4.6
0.24265
0.199
0.194
0.006
0.707
Trial
26
13.3
Autumn
1.02
1.55
0.203
0.315
7.7
0.49822
0.199
0.203
0.017
1.650
Trial
27
13.3
Autumn
0.97
2.11
0.193
0.473
16.5
1.37294
0.199
0.224
0.034
2.961
Trial
28
15.2
Spring
2.28
2.33
0.456
0.419
5.5
0.60184
0.2
0.18
­
0.003
0.087
Trial
29
15.2
Spring
2.25
3.17
0.450
0.669
10.2
1.42656
0.2
0.211
0.016
1.358
Trial
30
15.2
Spring
2.25
3.73
0.450
0.791
15.4
2.38846
0.2
0.212
0.022
1.980
Trial
31
16.2
Summer
1.14
1.21
0.239
0.241
4.9
0.29412
0.21
0.199
0.000
0.238
Trial
32
16.2
Summer
1.23
1.57
0.258
0.345
6.8
0.51315
0.21
0.22
0.012
0.971
Trial
33
16.2
Summer
1.18
1.64
0.248
0.364
9.5
0.72661
0.21
0.222
0.015
1.305
Trial
34
16.2
Summer
1.2
2.13
0.252
0.494
14.3
1.33376
0.21
0.232
0.026
2.234
Trial
35
16.4
Autumn
0.94
1.29
0.185
0.258
8.3
0.45980
0.197
0.2
0.013
1.256
Trial
36
16.4
Autumn
0.92
2.04
0.181
0.469
20.1
1.63423
0.197
0.23
0.035
3.027
Trial
37
19.5
Summer
1.19
1.28
0.258
0.274
6
0.39911
0.217
0.214
0.002
0.291
Trial
38
19.5
Summer
1.21
1.53
0.263
0.341
7.9
0.59621
0.217
0.223
0.010
0.934
Trial
39
19.5
Summer
1.16
1.63
0.252
0.380
10.3
0.81307
0.217
0.233
0.016
1.348
Trial
40
19.5
Summer
1.18
2.02
0.256
0.475
15.7
1.43412
0.217
0.235
0.024
2.100
Trial
41
22.5
Summer
1.1
1.21
0.243
0.258
7.4
0.46327
0.221
0.213
0.002
0.381
Trial
42
22.5
Summer
1.1
1.25
0.243
0.273
9.2
0.59294
0.221
0.218
0.005
0.511
Trial
43
22.5
Summer
1.11
1.41
0.245
0.326
11
0.78515
0.221
0.231
0.011
0.952
Trial
44
22.5
Summer
1.16
1.61
0.256
0.390
13.4
1.08202
0.221
0.242
0.017
1.300
Section
10­
10
LITERATURE
CITED
Brett,
R.
J.,
J.
E.
Shelbourn,
and
C.
T.
Shoop.
1969.
Growth
rate
and
body
composition
of
fingerling
sockeye
salmon,
Oncorhynchus
nerka,
in
relation
to
temperature
and
ration
size.
J.
Fish.
Res.
Bd.
Canada
26:
2363­
2394.

Brett,
R.
J.,
W.
C.
Clarke,
and
J.
E.
Shelbourn.
1982.
Experiments
on
thermal
requirements
for
growth
and
food
conversion
efficiency
of
juvenile
chinook
salmon
Oncorhynchus
tshawytscha.
Can.
Tech.
Rep.
Fish.
Aquat.
Sci.
No.
1127.

Everson,
L.
B.
1973.
Growth
and
food
consumption
of
juvenile
coho
salmon
exposed
to
natural
and
elevated
fluctuating
temperatures.
M.
S.
Thesis,
Oregon
State
University,
Corvallis,
OR.
68
pp.

Wurtsbaugh,
W.
A.
and
G.
E.
Davis.
1977.
Effects
of
temperature
and
ration
level
on
the
growth
and
food
conversion
efficiency
of
Salmo
gairdneri,
Richardson.
J.
Fish.
Biol.
11:
87­
98.
Section
10­
11
APPENDIX
B
REGIONAL
TEMPERATURE
DATA
FOR
THE
PACIFIC
NORTHWEST
REGION,
FROM
U.
S.
GEOLOGICAL
SURVEY
WATER
RESOURCES
DATA
Washington
Oregon
Idaho
From
1978­
79
Water
Resources
Data
U.
S.
Geological
Survey
Section
10­
12
Table
B.
1
Annual
maximum
temperature
for
1978­
79
and
most
extreme
temperature
measured
at
all
stream
and
river
sites
listed
in
the
U.
S.
G.
S.
water
resources
inventory
for
Washington.

State
River
USGS
Station
Number
Basin
Area
(
km2)
1979
Annual
Maximum
Temperature
(
oC)
Extreme
Maximum
(
oC)

Washington
Wynoochee
River
12037400
401
22.5
24
N.
F.
Skokomish
River
12056500
148
14
15
Skokomish
River
12056500
588
16
20.5
Nisqually
River
near
National
12082500
344
18
18.5
Nisqually
River
near
LaGrande
12086500
756
18
18.5
Green
River
12113000
1,033
22.5
24
Cedar
River
near
Landsburg
12117500
313
15.5
19.5
Cedar
River
near
Renton
12119000
477
23
24
M.
F.
Snoqualmie
River
12141300
399
18.5
N.
F.
Snoqualmie
River
1214200
166
21
S.
F.
Snoqualmie
River
108
19
Skagit
River
12179000
3,300
13
Tank
Creek
12197040
6
17.5
Minkler
Creek
12197110
13
21
21
Black
Creek
12197680
1
16.5
17.5
Wiseman
Creek
12197700
8
16
16.5
Skagit
River
near
Sedro
Woolley
12199000
7,809
18
18
Skagit
River
near
Mount
Vernon
12200500
8,011
17.5
17.8
Kalama
River
14223600
325
20.5
21.5
Cowlitz
River
near
Randall
14233400
2,668
18.5
19
Cowlitz
River
below
Mossyrock
Dam
14234810
2,989
11
15
Cowlitz
River
below
Mayfield
Dam
14238000
3,626
12.5
21
Tilton
River
14236200
365
21.5
24.5
Columbia
River
at
Warrendale
14128910
621,600
22
22.5
Columbia
River
at
Bradwood,
OR
14247400
665,900
21.5
22.5
Columbia
River
at
Northport,
WA
12433000
154,600
20
21
Columbia
River
at
Grand
Coulee
Dam
12472900
193,500
19
19
Columbia
River
at
Richland
12473520
251,000
21
21.5
Columbia
River
at
Umatilla
14019250
554,300
22
22.5
Pend
Oreille
River
12398600
65,300
23
24.5
Spokane
River
12433000
15,590
20.5
24.5
Yakima
River
at
Kiona
12510500
14,543
28
29
Snake
River
at
Anatone
13334300
241,000
23.5
25
Snake
River
at
Burbank
13353200
281,800
23.5
24
Section
10­
13
Table
B.
2
Annual
maximum
temperature
for
1978­
79
and
most
extreme
temperature
measured
at
all
stream
and
river
sites
listed
in
the
U.
S.
G.
S.
water
resources
inventory
for
Oregon.

State
River
USGS
Station
Number
Basin
Area
(
km2)
1979
Annual
Maximum
Temperature
(
oC)
Extreme
Maximum
(
oC)

Oregon
Columbia
River
at
Rainier
14245295
664,900
21.5
23.5
Columbia
River
at
Wama
14247295
665,000
21.5
22.5
Columbia
River
at
Vancouver
14144700
624,200
22
23.5
Owyhee
River
13184000
29,300
25
25
Meadow
Creek
near
Starkey
13318050
86
25
25
Meadow
Creek
above
Bear
Cr.
133 
8060
125
26
26.5
Minam
River
13331500
622
24.5
27
Umatilla
River
14020000
339
24.5
25
John
Day
River
14048000
19,600
31
33
Deschutes
River
near
Bend
14064500
4,556
17
17
Deschutes
River
near
Madras
14092500
20,250
14
18
Deschutes
River
near
Moody
14103000
27,200
23
23
DonnerundBlitzen
River
10396000
518
25.5
28.5
Willamette
at
Portland
14211720
28,700
24
27.5
Bull
Run
River
14138850
124
17
17
Fir
Creek
14138870
14
14.5
15
N.
F.
Bull
Run
River
14138900
22
14
S.
F.
Bull
Run
River
14139800
17
M.
F.
Willamette
near
Oak
Ridge
14144800
668
20
23
Hills
Creek
14144900
136
19
22.5
M.
F.
Willamette
above
Salt
Cr.
14145500
1,015
17
25
M.
F.
Willamette
below
N.
Fork
14148000
2,393
19
23.5
M.
F.
Willamette
near
Dexter
14150000
2,593
17.5
18.5
Fall
Creek
14150300
306
25
25
Wineberry
Creek
14150800
114
24
26.5
Fall
Creek
below
Wineberry
Cr.
14151000
482
20.5
26
M.
F.
Willamette
at
Jasper
14152000
3,471
17.5
21
Coast
Fork
Willamette
River
14152500
187
24.5
25.5
McKenzie
River
below
Trail
Cr.
Dam
14158850
477
11
12
McKenzie
River
at
McKenzie
Bridge
14159000
901
13
13.5
S.
Fork
McKenzie
14159200
414
17
17
S.
Fork
McKenzie
near
Rainbow
14159500
539
14
20
Blue
River
14161100
119
23
23
McKenzie
River
near
Vida
14162500
2,409
15
16
Willamette
River
at
Harrisburg
14166000
8,860
20
24
Mary's
River
14171000
412
22.5
23.5
Calapooia
River
at
Holley
14172000
272
29
29.5
Calapooia
River
at
Albany
14173500
963
28
28.5
N.
Santiam
River
below
Boulder
14178000
559
17.5
19.5
Breitenbush
River
14179000
275
16.5
18
N.
Santiam
River
at
Niagara
14181500
1,173
13.5
16.5
S.
Santiam
River
below
Cascadia
14185000
451
24
25
M.
Santiam
River
near
Cascadia
14185800
269
22
22.5
Quartzville
Creek
14185900
257
24
25.5
S.
Santiam
River
near
Foster
14187200
1,443
14.5
15.5
S.
Santiam
River
near
Waterloo
14187500
1,658
18
26
Santiam
River
14189000
4,640
23
23.5
Willamette
River
at
Salem
14191000
18,900
24
25.5
Tualatin
River
near
Gaston
14202500
126
24
Tualatin
River
at
West
Linn
14207500
1,829
23.5
27.5
Section
10­
14
Table
B.
2
Continued
Annual
maximum
temperature
for
1978­
79
and
most
extreme
temperature
measured
at
all
stream
and
river
sites
listed
in
the
U.
S.
G.
S.
water
resources
inventory
for
Oregon.

State
River
USGS
Station
Number
Basin
Area
(
km2)
1979
Annual
Maximum
Temperature
(
oC)
Extreme
Maximum
(
oC)

Oregon
Nehalem
River
14301000
1,728
24.5
24.5
Nestucca
River
14303600
466
23.5
24
Big
Rock
Creek
14304850
18
19.5
Siletz
River
14305500
523
24.5
Siuslaw
River
14307620
1,523
31
31
S.
Umpqua
River
at
Days
Cr.
14308600
1,660
29
30
S.
Umpqua
River
near
Roseburg
14312260
4,657
25
29
N.
Umpqua
River
at
Winchester
14319500
3,481
26
26.5
Umpqua
River
14321000
9,539
27
30
Rogue
River
below
Prospect
14330000
982
20.5
20.5
S.
Fork
Rogue
River
14335075
637
20
20
Rogue
River
at
McLeod
14335075
1.787
14.5
14.5
Big
Bull
Creek
14337500
635
24
24
Rogue
River
near
McLeod
14337600
2,429
14.5
18
Elk
Creek
14337800
204
27.5
28.5
W.
Branch
Elk
Creek
14337870
37
24
25.5
Elk
Creek
near
Trail
14338000
344
31.5
31.5
Rogue
River
at
Dodge
Bridge
14339000
3,147
18.5
20
Rogue
River
at
Raygold
1435900
5,317
19
22
Rogue
River
at
Grants
Pass
14361500
6,369
17
23.5
Rogue
River
near
Merlin
14370400
8,472
22.5
25.5
Rogue
River
at
Marial
14372250
9,873
24.5
27.5
Rogue
River
near
Agnes
14372300
10,202
24.5
26.5
Elliot
Creek
14361600
134
22.5
23
Carberry
Creek
14361700
178
23.5
23.5
Applegate
River
near
Copper
14362000
583
25.5
26.5
Applegate
River
near
Applegate
14366000
1,251
26
28
Applegate
River
near
Wilderville
14369500
1,808
28
28
Section
10­
15
Table
B.
3
Annual
maximum
temperature
for
1978­
79
and
most
extreme
temperature
measured
at
all
stream
and
river
sites
listed
in
the
U.
S.
G.
S.
water
resources
inventory
for
Idaho.

State
River
USGS
Station
Number
Basin
Area
(
km2)
1979
Annual
Maximum
Temperature
(
oC)
Extreme
Maximum
(
oC)

Idaho
Kootenai
River
at
Leonia
12305000
30,407
14
Kootenai
River
near
Copeland
12318500
34,710
19
24
Kootenai
River
at
Porthill
12322000
35
20
23.5
Salmon
River
at
Whitebird
13269000
35,090
25
28
Yankee
Fork
Salmon
River
13296000
505
16
17.5
N.
Fork
Clearwater
near
Canyon
13340600
3,520
21.5
23
Clearwater
River
near
Peck
13341050
6,320
21
25
Clearwater
River
at
Spalding
13342500
24,790
22
28
Bear
River
at
Border,
WY
10039500
6,439
23
23
Salt
River
above
reservoir,
Alpin
WY
13023000
1,160
16.5
21
Snake
River
at
Weiser
13269000
565
28.5
Snake
River
at
Hells
Canyon
Dam
13269000
190,000
19.5
20
Snake
River
near
Irwin
13032500
13,533
15.5
18.5
Snake
River
at
Neeley
13077000
35,200
23.5
23.5
Snake
River
at
Minidoka
13081500
40,700
23
23.5
Snake
River
at
King
Hill
13081500
92,700
20.5
23.5
Willow
Creek
near
Ririe
13058000
1,620
25
Section
10­
16
Calculation
of
Acute
Temperature
Relationships
Appendix
C
Appendix
C
­
Page
1
APPENDIX
C
ACUTE
EFFECTS
OF
TEMPERATURE
ON
SALMON
AND
TROUT:
DATA
USED,
ANALYSES
AND
ASSUMPTIONS
Chinook
salmon
Chum
salmon
Coho
salmon
Cutthroat
trout
Pink
salmon
Rainbow
trout
Sockeye
salmon
DESCRIPTION
OF
METHODOLOGY
This
memorandum
summarizes
how
estimated
acute
thermal
effects
curves
were
generated
for
selected
species
of
salmon
and
trout:
pink
salmon,
chum
salmon,
coho
salmon,
sockeye
salmon,
chinook
salmon,
rainbow
trout
(
steelhead),
and
cutthroat
trout.
Effects
of
elevated
temperatures
on
these
fish
species
were
of
interest;
thus,
curves
were
generated
from
available
data
for
acclimation
temperatures
of
15
°
C
and
higher.

Most
of
the
available
thermal
effects
information
is
based
on
50%
survival;
however,
curves
for
90%
survival
(
10%
mortality)
were
desired.
The
process
by
which
10%
mortality
curves
were
estimated
from
the
50%
mortality
information
is
detailed
here.
Three
attachments
are
provided
to
illustrate
the
data
used
and
analyses:

Attachment
1:
Acute
Effects
of
Temperature
on
Salmonids:
Median
Lethal
Times
(
LT50)
in
Relation
to
Temperature
Attachment
2:
LT50
to
LT10
Conversion
Factors
for
Pacific
Salmon:
Sockeye
and
Chinook.

Attachment
3:
Acute
Effects
of
Temperature
on
Salmonids:
Times
to
10%
Mortality
(
LT10)
In
Relation
to
Temperature.
Calculation
of
Acute
Temperature
Relationships
Appendix
C
Appendix
C
­
Page
2
ACUTE
THERMAL
EFFECTS
CURVES
ASSOCIATED
WITH
50%
MORTALITY
Data
from
several
sources
were
used
to
generate
curves
showing
the
relationship
between
temperature
and
duration
to
50%
mortality
(
EPA
1977,
Brett
1952,
and
Golden
1978).
Each
curve
estimates
the
length
of
time
50%
of
a
species
population
can
survive
at
some
temperature
above
its
upper
incipient
lethal
temperature.
For
the
remainder
of
this
memorandum,
this
temperature
will
be
referred
to
as
the
LT50,
the
temperature
causing
50%
mortality
in
a
population
of
fish
within
a
specified
length
of
time.

EPA
(
1977,
page
11
of
text
and
page
38
of
Appendix
A)
provides
a
regression
equation
relating
exposure
time
(
in
minutes)
to
the
LT50
(
in
EC):

log[
exposure
time]
=
a
+
b*
LT50,

which
can
also
be
written
as
LT50
=
(
log[
exposure
time]
­
a)/
b.

The
regression
coefficients,
a
and
b,
are
provided
in
EPA
(
1977)
for
many
fish
species,
including
all
those
identified
above,
except
cutthroat
trout
(
pages
55­
58
of
Appendix
B).
Golden
(
1978,
Figure
4
on
page
14)
provides
regression
coefficients
for
cutthroat
trout.
The
coefficients
in
EPA
(
1977)
were
gathered
from
many
different
sources,
including
Brett's
1952
paper
summarizing
his
study
of
lethal
temperatures
for
the
five
salmon
species.
In
the
attachments,
the
specific
studies
are
cited
rather
than
the
EPA
(
1977)
document.

The
acute
thermal
effects
curves
provided
in
Attachment
1
were
generated
in
Excel
using
regression
coefficients
provided
in
EPA
(
1977)
and
Golden
(
1978)
and
the
second
form
of
the
regression
equation
presented
above
for
a
range
of
times
(
durations).
For
the
five
salmon
species,
Brett
(
1952)
provided
ultimate
upper
incipient
lethal
temperatures,
and
the
acute
curves
were
discontinued
at
these
values.
For
rainbow
trout
and
cutthroat
trout,
the
curves
were
discontinued
at
25
°
C.

Although
it
was
assumed
that
the
regression
coefficients
in
Appendix
B
of
EPA
(
1977)
were
correct,
one
appeared
to
be
in
error.
The
value
for
a
was
given
as
16.2444
for
pink
salmon
at
an
acclimation
temperature
of
20
°
C
from
Brett's
study
(
1952).
The
resulting
curve
did
not
match
the
one
presented
in
Figure
5
of
Brett
(
1952).
To
generate
a
curve
more
representative
of
Brett's
(
1952)
figure,
a
value
of
13.2444
was
used
for
a
instead.

A
few
of
the
studies
included
in
EPA
(
1977)
were
excluded
from
Attachment
1.
These
were
studies
in
which
the
fish
being
tested
showed
signs
of
gas
bubble
disease
or
other
effects
of
gas
supersaturation.
Calculation
of
Acute
Temperature
Relationships
Appendix
C
Appendix
C
­
Page
3
COMPARISON
OF
LT50
AND
LT10
CURVES
In
the
EPA
(
1977)
document,
two
unpublished
studies
provided
regression
coefficients
for
both
50%
and
10%
(
LT10)
mortality
curves
at
acclimation
temperatures
of
15EC
or
higher.
McConnell
and
Blahm
(
1970)
calculated
regression
coefficients
for
sockeye
salmon;
and
Blahm
and
McConnell
(
1970)
calculated
regression
coefficients
for
both
spring
and
fall
runs
of
chinook
salmon.
Using
the
regression
coefficients
generated
from
these
studies,
LT50
and
LT10
values
were
calculated
for
a
range
of
durations.

Attachment
2
contains
two
tables
(
one
each
for
sockeye
and
chinook
salmon)
of
the
calculated
values
and
their
ratios
(
i.
e,
LT10/
LT50).
For
the
range
of
durations
calculated,
the
LT10
values
were
98.0
to
99.7%
of
the
LT50
values.
This
is
consistent
with
Brett
(
1958,
page
76
and
Figure
4),
who
indicated
that
differences
between
temperatures
for
50%
mortality
and
those
for
<
50%
mortality
are
relatively
small,
"
implying
that
temperatures
of
this
order
have
only
to
increase
slightly
to
cause
a
large
difference
in
mortality."

ACUTE
THERMAL
EFFECTS
CURVES
@
10%
MORTALITY
Based
on
the
comparison
of
LT50
and
LT10
curves
generated
from
the
McConnell
and
Blahm
(
1970)
and
Blahm
and
McConnell
(
1970)
studies,
and
to
be
somewhat
conservative,
LT10
curves
were
estimated
for
the
other
studies
by
applying
a
factor
of
0.98
to
each
curve.
That
is,
each
LT10
value
was
estimated
to
be
98%
of
the
LT50
value
calculated
from
the
regression
equation.
The
estimated
LT10
curves
are
presented
in
Attachment
3.

Based
on
visual
inspection
of
the
LT50
and
LT10
curves
included
in
Attachment
1
from
these
two
studies,
the
slopes
were
similar.
That
is,
on
the
log­
time
scale,
the
differences
between
the
LT50
and
LT10
curves
were
approximately
constant.
(
There
was
insufficient
information
presented
in
Appendix
B
of
the
EPA
(
1977)
document
to
statistically
compare
the
slopes.)
Had
the
differences
not
appeared
constant,
the
application
of
a
singe
adjustment
factor
would
not
have
been
appropriate.

KEY
ASSUMPTIONS/
ISSUES
The
following
text
summarizes
key
assumptions
used
when
evaluating
the
available
data
and
estimating
the
LT10
curves.
Other
issues
relevant
to
the
use
of
these
data
are
also
identified.
The
information
is
presented
in
bullet
form
and
can
be
expanded
upon
at
a
later
date
if
desired.

·
 
The
most
important
assumption
is
that
the
data
provided
in
EPA
(
1977)
were
representative
of
the
same
species
from
different
locations
(
i.
e.,
different
stocks).
There
are
many
factors
to
consider
with
such
an
assumption:
different
environmental
influences
(
water
quality,
temperature
fluctuations),
genetic
differences,
size,
life
stage,
etc.
Calculation
of
Acute
Temperature
Relationships
Appendix
C
Appendix
C
­
Page
4
·
 
As
stated
above,
the
regression
coefficients
reported
in
EPA
(
1977)
were
assumed
to
be
correct.

·
 
The
adjustment
factor
estimated
from
the
McConnell
and
Blahm
(
1970)
and
Blahm
and
McConnell
(
1970)
data
for
sockeye
and
chinook
salmon
was
assumed
to
be
appropriate
to
use
in
estimating
LT10
curves
for
the
other
salmon
and
trout
species.

·
 
Effects
curves
appear
to
differ
for
fish
tested
using
a
constant
acclimation
temperature
versus
a
fluctuating
one
(
see
Attachment
1
for
cutthroat
trout).
Can
results
based
on
constant
acclimation
temperatures
be
applied
to
fish
living
in
streams
with
temperatures
fluctuating
on
a
daily
and
seasonal
basis?

·
 
For
many
of
the
studies,
the
test
fish
were
obtained
from
a
hatchery,
and
sometimes
from
a
limited
number
of
females.
While
this
limited
variability
in
the
biological
responses
to
temperature
because
of
the
genetics,
it
also
may
have
limited
the
representativeness
of
the
results
for
a
wider
population
of
fish
of
same
species.

APPENDIX
C
REFERENCES
Brett,
J.
R.
1952.
Temperature
tolerance
in
young
Pacific
salmon,
genus
Oncorhynchus.
J.
Fish.
Res.
Board
Can.
9(
6):
265­
323.

Brett,
J.
R.
1958.
Implications
and
assessments
of
environmental
stress.
Pages
60­
83
in
P.
A.
Larkin,
editor.
The
Investigation
of
Fish­
Power
Problems.
H.
R.
MacMillan
Lectures
in
Fisheries,
Univ.
of
British
Columbia,
Vancouver.

EPA.
1977.
Temperature
criteria
for
freshwater
fish:
protocol
and
procedures.
U.
S.
Environmental
Protection
Agency
Office
of
Research
and
Development,
Environmental
Research
Laboratory,
Duluth,
Minnesota.
EPA­
600/
3­
77­
061.
130p.

Golden,
J.
T.
1978.
The
effects
of
fluctuating
temperatures
on
the
lethal
tolerance
limits
of
coastal
cutthroat
trout
(
Salmo
clarki
clarki).
Thesis,
Oregon
State
University,
Corvallis,
Oregon.
29p.
Calculation
of
Acute
Temperature
Relationships
Appendix
C
Appendix
C
APPENDIX
C
ATTACHMENT
1
Acute
Effects
of
Temperature
on
Salmonids:
Median
Lethal
Times
(
LT50)
in
Relation
to
Temperature
CHINOOK
SALMON
Acclim.
Source
Age/
Size
a
b
N
r
Time(
hr)
LT50(
C)

15C
Brett(
1952)
juv.
frshwtr
fry
16.4454
­
0.5364
4
­
0.9906
0.1
29.2
0.25
28.5
0.5
27.9
1
27.3
2
26.8
4
26.2
6
25.9
8
25.7
10
25.5
12
25.3
13
25.3
14
25.2
16
25.1
20C
Brett(
1952)
juv.
frshwtr
fry
22.9065
­
0.7611
7
­
0.9850
0.1
29.1
0.25
28.6
0.5
28.2
2
27.4
4
27.0
6
26.7
8
26.6
10
26.4
16
26.2
24
25.9
40
25.7
60
25.4
80
25.3
100
25.1
24C
Brett(
1952)
juv.
frshwtr
fry
18.9940
­
0.5992
9
­
0.9923
0.1
30.4
0.25
29.7
0.5
29.2
1
28.7
2
28.2
4
27.7
6
27.4
8
27.2
10
27.1
16
26.7
24
26.4
40
26.1
56
25.8
72
25.6
96
25.4
120
25.3
140
25.1
17C
Coutant
(
1970)
"
Jacks"
(
1­
2
yrs)
13.2502
­
0.4121
4
­
0.8206
0.1
30.3
0.25
29.3
0.5
28.6
24
25
26
27
28
29
30
31
32
33
34
35
0.1
10
1000
Duration
(
hr)

Temperature
(

C)
LT50
for
15C
Acclim.
(
Brett
1952)

LT50
for
20C
Acclim.
(
Brett
1952)

LT50
for
24C
Acclim.
(
Brett
1952)

UULT50
(
25.1C,
Brett
1952)

LT50
(
17C,
Coutant
1970)

LT50
(
20C,
Coutant
1970)

24
25
26
27
28
29
30
31
32
33
34
35
0
40
80
120
160
Duration
(
hr)

Temperature
(

C)
LT50
for
15C
Acclim.
(
Brett
1952)

LT50
for
20C
Acclim.
(
Brett
1952)

LT50
for
24C
Acclim.
(
Brett
1952)

UULT50
(
25.1C,
Brett
1952)

LT50
(
17C,
Coutant
1970)

LT50
(
20C,
Coutant
1970)
CHINOOK
SALMON
Acclim.
Source
Age/
Size
a
b
N
r
Time(
hr)
LT50(
C)

1
27.8
2
27.1
The
17C
and
19C
temperatures
were
Columbia
River
3
26.7
temperatures
(
at
Grand
Rapids
Dam)
during
fall
4
26.4
migrations
two
different
years.
6
25.9
8
25.6
10
25.4
12
25.2
14
25.1
19C
Coutant
(
1970)
"
Jacks"
(
1­
2
yrs)
9.4683
­
0.2504
4
­
0.9952
0.1
34.7
0.25
33.1
0.5
31.9
1
30.7
2
29.5
3
28.8
4
28.3
6
27.6
8
27.1
10
26.7
20
25.5
26
25.1
CHINOOK
SALMON
Acclim.
Source
Age/
Size
a
b
N
r
Time(
hr)
LT50(
C)

20C
Blahm
&
juv.
(
spring
run)
21.3981
­
0.7253
3
­
0.9579
0.1
28.4
McConnell
(
1970)
0.25
27.9
unpublished
data
0.5
27.5
1
27.1
2
26.6
4
26.2
8
25.8
10
25.7
16
25.4
20
25.3
24
25.1
26
25.1
20C
Blahm
&
juv.
(
spring
run)
22.6664
­
0.7797
4
­
0.9747
0.1
28.1
McConnell
(
1970)
0.25
27.6
unpublished
data
0.5
27.2
10%
mortality
1
26.8
2
26.4
4
26.0
6
25.8
8
25.6
10
25.5
12
25.4
16
25.2
20
25.1
20C
Blahm
&
juv.
(
spring
run)
20.9294
­
0.7024
3
­
0.9463
0.1
28.7
McConnell
(
1970)
0.25
28.1
unpublished
data
0.5
27.7
90%
mortality
2
26.8
4
26.4
6
26.2
8
26.0
10
25.8
16
25.6
20
25.4
24
25.3
32
25.1
24
25
26
27
28
29
30
0.1
1
10
100
1000
Duration
(
hr)

Temperature
(

C)
LT50
for
20C
Acclim.
(
Brett
1952)

UULT50
(
25.1C,
Brett
1952)

LT50
(
20C,
Blahm
&
McConnell
1970)

LT10
(
20C,
Blahm
&
McConnell
1970)

LT90
(
20C,
Blahm
&
McConnell
1970)

24
25
26
27
28
29
30
0
40
80
120
Duration
(
hr)

Temperature
(

C)
LT50
for
20C
Acclim.
(
Brett
1952)

UULT50
(
25.1C,
Brett
1952)

LT50
(
20C,
Blahm
&
McConnell
1970)

LT10
(
20C,
Blahm
&
McConnell
1970)

LT90
(
20C,
Blahm
&
McConnell
1970)
CHINOOK
SALMON
Acclim.
Source
Age/
Size
a
b
N
r
Time(
hr)
LT50(
C)

20C
Blahm
&
juv.
(
fall
run)
22.2124
­
0.7526
4
­
0.9738
0.1
28.5
McConnell
(
1970)
0.25
28.0
unpublished
data
0.5
27.6
1
27.2
2
26.8
4
26.4
8
26.0
10
25.8
16
25.6
20
25.4
24
25.3
34
25.1
20C
Blahm
&
juv.
(
fall
run)
21.6756
­
0.7438
4
­
0.9550
0.1
28.1
McConnell
(
1970)
0.25
27.6
unpublished
data
0.5
27.2
10%
mortality
1
26.8
2
26.3
4
25.9
6
25.7
8
25.5
10
25.4
12
25.3
14
25.2
16
25.1
20C
Blahm
&
juv.
(
fall
run)
20.5162
­
0.6860
3
­
0.9475
0.1
28.8
McConnell
(
1970)
0.25
28.2
unpublished
data
0.5
27.8
90%
mortality
2
26.9
4
26.4
6
26.2
8
26.0
10
25.9
16
25.6
20
25.4
24
25.3
32
25.1
24
25
26
27
28
29
30
0.1
1
10
100
1000
Duration
(
hr)

Temperature
(

C)
LT50
for
20C
Acclim.
(
Brett
1952)

Est.
UULT50
(
25.1C)

LT50
(
20C,
Blahm
&
McConnell
1970)

LT10
(
20C,
Blahm
&
McConnell
1970)

LT90
(
20C,
Blahm
&
McConnell
1970)

24
25
26
27
28
29
30
0
40
80
120
Duration
(
hr)

Temperature
(

C)
LT50
for
20C
Acclim.
(
Brett
1952)

UULT50
(
25.1C,
Brett
1952)

LT50
(
20C,
Blahm
&
McConnell
1970)

LT10
(
20C,
Blahm
&
McConnell
1970)

LT90
(
20C,
Blahm
&
McConnell
1970)
COHO
SALMON
Acclim.
Source
Age/
Size
a
b
N
r
Time(
hr)
LT50(
C)

15C
Brett(
1952)
juv.
frshwtr
fry
20.4066
­
0.6858
6
­
0.9681
0.1
28.6
0.25
28.0
0.5
27.6
1
27.2
2
26.7
4
26.3
6
26.0
8
25.8
10
25.7
12
25.6
16
25.4
20
25.3
32
25.0
20C
Brett(
1952)
juv.
frshwtr
fry
20.4022
­
0.6713
4
­
0.9985
0.1
29.2
0.25
28.6
0.5
28.2
2
27.3
4
26.8
6
26.6
8
26.4
10
26.3
16
25.9
24
25.7
40
25.4
48
25.2
56
25.1
70
25.0
23C
Brett(
1952)
juv.
frshwtr
fry
18.9736
­
0.6013
5
­
0.9956
0.1
30.3
0.25
29.6
0.5
29.1
1
28.6
2
28.1
4
27.6
6
27.3
8
27.1
10
26.9
16
26.6
24
26.3
40
25.9
56
25.7
72
25.5
96
25.3
120
25.1
140
25.0
17C
Coutant
(
1970)
adult
5.9068
­
0.1630
5
­
0.9767
0.1
31.5
Reported
acclimation
temp.
was
the
Columbia
River
0.2
29.6
temp.
(
at
Priest
Rapids
Dam)
during
fall
migration.
0.3
28.5
24
25
26
27
28
29
30
31
32
0.1
10
1000
Duration
(
hr)

Temperature
(

C)
LT50
for
15C
Acclim.
(
Brett
1952)

LT50
for
20C
Acclim.
(
Brett
1952)

LT50
for
23C
Acclim.
(
Brett
1952)

UULT50
(
25.0C,
Brett
1952)

LT50
(
17C,
Coutant
1970)

24
25
26
27
28
29
30
31
32
0
40
80
120
160
Duration
(
hr)

Temperature
(

C)
LT50
for
15C
Acclim.
(
Brett
1952)

LT50
for
20C
Acclim.
(
Brett
1952)

LT50
for
23C
Acclim.
(
Brett
1952)

UULT50
(
25.0C,
Brett
1952)

LT50
(
17C,
Coutant
1970)
COHO
SALMON
Acclim.
Source
Age/
Size
a
b
N
r
Time(
hr)
LT50(
C)

0.4
27.8
0.5
27.2
0.6
26.7
0.7
26.3
0.8
25.9
0.9
25.6
1
25.3
1.1
25.1
1.15
25.0
RAINBOW
TROUT
Acclim.
Source
Age/
Size
a
b
N
r
Time(
hr)
LT50(
C)

15C
Alabaster
&
juvenile
15.6500
­
0.5000
2
­­­
0.1
29.7
Downing
(
1966)
0.25
28.9
0.5
28.3
1
27.7
2
27.1
4
26.5
6
26.2
8
25.9
10
25.7
12
25.6
16
25.3
20
25.1
24
25.0
18C
Alabaster
&
juvenile
18.4654
­
0.5801
5
­
0.9787
0.1
30.5
Welcomme
(
1962)
0.25
29.8
Dissolved
Oxygen
at
7.4
mg/
L
0.5
29.3
2
28.2
4
27.7
6
27.4
8
27.2
10
27.0
16
26.7
24
26.4
40
26.0
80
25.5
120
25.2
160
25.0
18C
Alabaster
&
juvenile
13.6531
­
0.4264
5
­
0.9742
0.1
30.2
Welcomme
(
1962)
0.25
29.3
Dissolved
Oxygen
at
3.8
mg/
L
0.5
28.6
1
27.8
2
27.1
4
26.4
6
26.0
7
25.9
8
25.7
9
25.6
10
25.5
12
25.3
14
25.2
16
25.0
24
25
26
27
28
29
30
31
32
0.1
10
1000
Duration
(
hr)

Temperature
(

C)
LT50
for
15C
Acclim.
(
Alabaster
&
Downing
1966)

LT50
for
18C
Acclim.
(
7.4
DO,
Alabaster
&
Welcomme
1962)

LT50
for
18C
Acclim.
(
3.8
DO,
Alabaster
&
Welcomme
1962)

LT50
for
20C
Acclim.
(
Alabaster
&
Downing
1966)

24
25
26
27
28
29
30
31
32
0
40
80
120
160
Duration
(
hr)

Temperature
(

C)
LT50
for
15C
Acclim.
(
Alabaster
&
Downing
1966)

LT50
for
18C
Acclim.
(
7.4
DO,
Alabaster
&
Welcomme
1962)

LT50
for
18C
Acclim.
(
3.8
DO,
Alabaster
&
Welcomme
1962)

LT50
for
20C
Acclim.
(
Alabaster
&
Downing
1966)
RAINBOW
TROUT
Acclim.
Source
Age/
Size
a
b
N
r
Time(
hr)
LT50(
C)

20C
Alabaster
&
juvenile
19.6250
­
0.6250
2
­­­
0.1
30.2
Downing
(
1966)
0.25
29.5
0.5
29.0
1
28.6
2
28.1
4
27.6
6
27.3
8
27.1
12
26.8
20
26.5
32
26.1
40
26.0
80
25.5
120
25.2
160
25.0
20C
Craigie
(
1963)
yearling
14.6405
­
0.4470
3
­
0.9787
0.1
31.0
Raised
in
soft
water,
tested
in
soft
water
(
SS)
0.25
30.1
0.5
29.4
1
28.8
2
28.1
4
27.4
6
27.0
8
26.8
10
26.5
16
26.1
20
25.9
24
25.7
32
25.4
40
25.2
48
25.0
20C
Craigie
(
1963)
yearling
15.0392
­
0.4561
3
­
0.9917
0.1
31.3
Raised
in
soft
water,
tested
in
hard
water
(
SH)
0.25
30.4
0.5
29.7
1
29.1
2
28.4
24
25
26
27
28
29
30
31
32
0.1
10
1000
Duration
(
hr)

Temperature
(

C)
LT50
for
20C
Acclim.
(
Alabaster
&
Downing
1966)

LT50
for
20C
Acclim.
(
SS,
Craigie
1963)

LT50
for
20C
Acclim.
(
SH,
Craigie
1963))

LT50
for
20C
Acclim.
(
HS,
Craigie
1963)

LT50
for
20C
Acclim.
(
HH,
Craigie
1963)

24
25
26
27
28
29
30
31
32
0
40
80
120
160
Duration
(
hr)

Temperature
(

C)
LT50
for
20C
Acclim.
(
Alabaster
&
Downing
1966)

LT50
for
20C
Acclim.
(
SS,
Craigie
1963)

LT50
for
20C
Acclim.
(
SH,
Craigie
1963)

LT50
for
20C
Acclim.
(
HS,
Craigie
1963)

LT50
for
20C
Acclim.
(
HH,
Craigie
1963)
RAINBOW
TROUT
Acclim.
Source
Age/
Size
a
b
N
r
Time(
hr)
LT50(
C)

4
27.8
6
27.4
8
27.1
12
26.7
16
26.4
20
26.2
32
25.8
40
25.6
60
25.2
72
25.0
20C
Craigie
(
1963)
yearling
15.1473
­
0.4683
3
­
0.9781
0.1
30.7
Raised
in
hard
water,
tested
in
soft
water
(
HS)
0.25
29.8
0.5
29.2
1
28.5
2
27.9
4
27.3
6
26.9
8
26.6
10
26.4
16
26.0
20
25.8
24
25.6
32
25.3
40
25.1
48
25.0
20C
Craigie
(
1963)
yearling
12.8718
­
0.3837
3
­
0.9841
0.1
31.5
Raised
in
hard
water,
tested
in
hard
water
(
HH)
0.25
30.5
0.5
29.7
1
28.9
2
28.1
3
27.7
4
27.3
6
26.9
8
26.6
10
26.3
12
26.1
RAINBOW
TROUT
Acclim.
Source
Age/
Size
a
b
N
r
Time(
hr)
LT50(
C)

16
25.8
20
25.5
24
25.3
32
25.0
CUTTHROAT
TROUT
Acclim.
Source
Age/
Size
a
b
N
r
Time(
hr)
LT50(
C)

23C
Golden
(
1978)
juvenile
18.092
­
0.56523
?
­
0.996
0.1
30.6
1975
tests,
hatchery
only
0.25
29.9
0.5
29.4
1
28.9
8
27.3
16
26.7
24
26.4
32
26.2
40
26.0
60
25.7
80
25.5
120
25.2
160
25.0
13­
25C
Golden
(
1978)
juvenile
22.543
­
0.71999
?
­
0.999
0.1
30.2
1975
tests,
hatchery
only
0.25
29.7
0.5
29.3
2
28.4
4
28.0
8
27.6
16
27.2
24
26.9
40
26.6
80
26.2
160
25.8
240
25.5
300
25.4
560
25.0
23C
Golden
(
1978)
juvenile
18.3166
­
0.5723
?
­
0.990
0.1
30.6
1976
tests,
hatchery
and
wild
trout
data
pooled
0.25
30.0
0.5
29.4
1
28.9
2
28.4
4
27.8
8
27.3
12
27.0
16
26.8
24
26.5
40
26.1
80
25.6
120
25.3
160
25.0
24
25
26
27
28
29
30
31
32
33
0.1
10
1000
Duration
(
hr)

Temperature
(

C)
LT50,
23C
Acclim.
(
Golden
1978,
1975
data)

LT50,
13­
25C
Acclim.
(
Golden
1978,
1975
data)

LT50,
23C
Acclim.
(
Golden
1978,
1976
data)

LT50,
13­
23C
Acclim.
(
Golden
1978,
1976
data)

24
25
26
27
28
29
30
31
32
33
0
120
240
360
480
600
Duration
(
hr)

Temperature
(

C)
LT50,
23C
Acclim.
(
Golden
1978,
1975
data)

LT50,
13­
25C
Acclim.
(
Golden
1978,
1975
data)

LT50,
23C
Acclim.
(
Golden
1978,
1976
data)

LT50,
13­
23C
Acclim.
(
Golden
1978,
1976
data)
CUTTHROAT
TROUT
Acclim.
Source
Age/
Size
a
b
N
r
Time(
hr)
LT50(
C)

13­
23C
Golden
(
1978)
juvenile
18.1515
­
0.5723
?
­
0.992
0.1
30.4
1976
tests,
hatchery
and
wild
trout
data
pooled
0.25
29.7
0.5
29.1
1
28.6
2
28.1
4
27.6
6
27.3
8
27.0
12
26.7
20
26.3
32
26.0
40
25.8
60
25.5
80
25.3
120
25.0
SOCKEYE
SALMON
Acclim.
Source
Age/
Size
a
b
N
r
Time(
hr)
LT50(
C)

15C
Brett(
1952)
juv.
frshwtr
fry
15.8799
­
0.5210
7
­
0.9126
0.1
29.0
0.25
28.2
0.5
27.6
1
27.1
2
26.5
4
25.9
6
25.6
8
25.3
10
25.1
12
25.0
13
24.9
14
24.9
24
24.4
20C
Brett(
1952)
juv.
frshwtr
fry
19.3821
­
0.6378
5
­
0.9602
0.1
29.2
0.25
28.5
0.5
28.1
2
27.1
4
26.7
6
26.4
8
26.2
10
26.0
16
25.7
24
25.4
40
25.1
48
25.0
56
24.9
116
24.4
23C
Brett(
1952)
juv.
frshwtr
fry
20.0020
­
0.6496
4
­
0.9981
0.1
29.6
0.25
29.0
0.5
28.5
1
28.1
2
27.6
4
27.1
6
26.9
8
26.7
10
26.5
16
26.2
24
25.9
40
25.6
56
25.4
72
25.2
96
25.0
120
24.9
220
24.4
23
24
25
26
27
28
29
30
31
32
0.1
10
1000
Duration
(
hr)

Temperature
(

C)
LT50
for
15C
Acclim.
(
Brett
1952)

LT50
for
20C
Acclim.
(
Brett
1952)

LT50
for
23C
Acclim.
(
Brett
1952)

UULT50
(
24.4C,
Brett
1952)

LT50
(
20C,
McConnell
&
Blahm
1970)

23
24
25
26
27
28
29
30
31
32
0
40
80
120
160
200
240
Duration
(
hr)

Temperature
(

C)
LT50
for
15C
Acclim.
(
Brett
1952)

LT50
for
20C
Acclim.
(
Brett
1952)

LT50
for
23C
Acclim.
(
Brett
1952)

UULT50
(
24.4C,
Brett
1952)

LT50
(
20C,
McConnell
&
Blahm
1970)
SOCKEYE
SALMON
Acclim.
Source
Age/
Size
a
b
N
r
Time(
hr)
LT50(
C)

20C
McConnell
&
juvenile
16.7328
­
0.5473
6
­
0.9552
0.1
29.2
Blahm
(
1970)
(
under
yearling)
0.25
28.4
unpublished
data
0.5
27.9
1
27.3
2
26.8
4
26.2
6
25.9
8
25.7
10
25.5
16
25.1
20
24.9
40
24.4
20C
McConnell
&
juvenile
17.5227
­
0.5861
6
­
0.9739
0.1
28.6
Blahm
(
1970)
(
under
yearling)
0.25
27.9
unpublished
data
0.5
27.4
10%
mortality
1
26.9
2
26.3
3
26.0
4
25.8
6
25.5
8
25.3
10
25.2
12
25.0
28
24.4
20C
McConnell
&
juvenile
15.7823
­
0.5061
6
­
0.9539
0.1
29.6
Blahm
(
1970)
(
under
yearling)
0.25
28.9
unpublished
data
0.5
28.3
90%
mortality
2
27.1
4
26.5
6
26.1
8
25.9
10
25.7
16
25.3
20
25.1
24
24.9
44
24.4
24
25
26
27
28
29
30
0.1
1
10
100
1000
Duration
(
hr)

Temperature
(

C)
LT50
for
20C
Acclim.
(
Brett
1952)

UULT50
(
24.4C,
Brett
1952)

LT50
(
20C,
McConnell
&
Blahm
1970)

LT10
(
20C,
McConnell
&
Blahm
1970)

LT90
(
20C,
McConnell
&
Blahm
1970)

24
25
26
27
28
29
30
0
40
80
120
Duration
(
hr)

Temperature
(

C)
LT50
for
20C
Acclim.
(
Brett
1952)

UULT50
(
24.4C,
Brett
1952))

LT50
(
20C,
McConnell
&
Blahm
1970)

LT10
(
20C,
McConnell
&
Blahm
1970)

LT90
(
20C,
McConnell
&
Blahm
1970)
CHUM
SALMON
Acclim.
Source
Age/
Size
a
b
N
r
Time(
hr)
LT50(
C)

15C
Brett(
1952)
juv.
frshwtr
fry
15.8911
­
0.5252
8
­
0.9070
0.1
28.8
0.25
28.0
0.5
27.4
1
26.9
2
26.3
4
25.7
6
25.4
8
25.2
10
25.0
12
24.8
16
24.6
20
24.4
40
23.8
20C
Brett(
1952)
juv.
frshwtr
fry
16.1894
­
0.5168
9
­
0.9750
0.1
29.8
0.25
29.1
0.5
28.5
2
27.3
4
26.7
6
26.4
8
26.1
10
26.0
16
25.6
24
25.2
40
24.8
60
24.4
80
24.2
132
23.8
23C
Brett(
1952)
juv.
frshwtr
fry
15.3825
­
0.4721
4
­
0.9652
0.1
30.9
0.25
30.1
0.5
29.5
1
28.8
2
28.2
4
27.5
6
27.2
8
26.9
10
26.7
16
26.3
24
25.9
40
25.4
56
25.1
72
24.9
96
24.6
120
24.4
240
23.8
23
24
25
26
27
28
29
30
31
32
0.1
10
1000
Duration
(
hr)

Temperature
(

C)
LT50
for
15C
Acclim.
(
Brett
1952)

LT50
for
20C
Acclim.
(
Brett
1952)

LT50
for
23C
Acclim.
(
Brett
1952)

UULT50
(
23.8C,
Brett
1952)

23
24
25
26
27
28
29
30
31
32
0
40
80
120
160
200
240
Duration
(
hr)

Temperature
(

C)
LT50
for
15C
Acclim.
(
Brett
1952)

LT50
for
20C
Acclim.
(
Brett
1952)

LT50
for
23C
Acclim.
(
Brett
1952)

UULT50
(
23.8C,
Brett
1952)
PINK
SALMON
Acclim.
Source
Age/
Size
a
b
N
r
Time(
hr)
LT50(
C)

15C
Brett(
1952)
juv.
frshwtr
fry
12.8937
­
0.4074
8
­
0.9884
0.1
29.7
0.25
28.8
0.5
28.0
1
27.3
2
26.5
4
25.8
6
25.4
8
25.1
10
24.8
12
24.6
16
24.3
20
24.1
24
23.9
20C
Brett(
1952)
juv.
frshwtr
fry
13.2444
­
0.4074
7
­
0.9681
0.1
30.6
16.2444
in
EPA
(
1977),
but
0.25
29.6
the
resulting
curve
does
not
0.5
28.9
match
the
information
1
28.1
presented
in
Brett
(
1952)
2
27.4
<
Fig.
5>
4
26.7
6
26.2
8
25.9
10
25.7
16
25.2
24
24.8
32
24.5
40
24.2
56
23.9
24C
Brett(
1952)
juv.
frshwtr
fry
14.7111
­
0.4459
6
­
0.9690
0.1
31.2
0.25
30.4
0.5
29.7
1
29.0
2
28.3
4
27.7
6
27.3
8
27.0
10
26.8
16
26.3
24
25.9
40
25.4
56
25.1
72
24.8
96
24.6
120
24.3
192
23.9
23
24
25
26
27
28
29
30
31
32
0.1
10
1000
Duration
(
hr)

Temperature
(

C)
LT50
for
15C
Acclim.
(
Brett
1952)

LT50
for
20C
Acclim.
(
Brett
1952)

LT50
for
24C
Acclim.
(
Brett
1952)

UULT50
(
23.9C,
Brett
1952)

23
24
25
26
27
28
29
30
31
32
0
40
80
120
160
200
Duration
(
hr)

Temperature
(

C)
LT50
for
15C
Acclim.
(
Brett
1952)

LT50
for
20C
Acclim.
(
Brett
1952)

LT50
for
24C
Acclim.
(
Brett
1952)

UULT50
(
23.9C,
Brett
1952)
Calculation
of
Acute
Temperature
Relationships
Appendix
C
Appendix
C
APPENDIX
C
ATTACHMENT
2
LT50
to
LT10
Conversion
Factors
for
Pacific
Salmon:
Sockeye
and
Chinook.
Calculation
of
Acute
Temperature
Relationships
Appendix
C
Appendix
C
Attacment
2:
Sockeye
Salmon
(
20C
Acclimation
Temperature)
From
McConnell
&
Blahm
(
1970)
unpublished
data
16.7328
­
0.5473
17.5227
­
0.5861
LT10/
LT50
Delta
T
Time(
hr)
LT50(
C)
Time(
hr)
LT10(
C)
Ratio
(
C)

0.1
29.2
0.1
28.6
98.00%
0.6
0.25
28.4
0.25
27.9
98.12%
0.5
0.5
27.9
0.5
27.4
98.21%
0.5
1
27.3
1
26.9
98.31%
0.5
2
26.8
2
26.3
98.41%
0.4
3
26.5
3
26.0
98.47%
0.4
4
26.2
4
25.8
98.52%
0.4
6
25.9
6
25.5
98.58%
0.4
8
25.7
8
25.3
98.63%
0.4
12
25.4
12
25.0
98.70%
0.3
16
25.1
16
24.8
98.74%
0.3
20
24.9
20
24.6
98.78%
0.3
24
24.8
24
24.5
98.81%
0.3
32
24.6
32
24.3
98.86%
0.3
40
24.4
40
24.1
98.90%
0.3
60
24.1
60
23.8
98.98%
0.2
80
23.8
80
23.6
99.03%
0.2
Calculation
of
Acute
Temperature
Relationships
Appendix
C
___________

Appendix
C
Attachment
2:
Chinook
Salmon
(
20C
Acclimation
Temperature)

From
Blahm
&
McConnell
(
1970)
unpublished
data
Spring
Run
Fall
Run
21.3981
­
0.7253
22.6664
­
0.7797
LT10/
LT50
Delta
T
22.2124
21.6756
­
0.7526
21.6756
­
0.7438
LT10/
LT50
Delta
T
Time(
hr)
LT50(
C)
Time(
hr)
LT10(
C)
Ratio
(
C)
Time(
hr)
Time(
hr)
LT50(
C)
Time(
hr)
LT10(
C)
Ratio
(
C)

0.1
28.4
0.1
28.1
98.74%
0.4
0.1
0.1
28.5
0.1
28.1
98.65%
0.4
0.25
27.9
0.25
27.6
98.86%
0.3
0.25
0.25
28.0
0.25
27.6
98.60%
0.4
0.5
27.5
0.5
27.2
98.95%
0.3
0.5
0.5
27.6
0.5
27.2
98.56%
0.4
2
26.6
2
26.4
99.13%
0.2
2
2
26.8
2
26.3
98.49%
0.4
4
26.2
4
26.0
99.23%
0.2
4
4
26.4
4
25.9
98.44%
0.4
6
26.0
6
25.8
99.28%
0.2
6
6
26.1
6
25.7
98.42%
0.4
8
25.8
8
25.6
99.33%
0.2
8
8
26.0
8
25.5
98.40%
0.4
10
25.7
10
25.5
99.36%
0.2
10
10
25.8
10
25.4
98.39%
0.4
16
25.4
16
25.2
99.43%
0.1
16
16
25.6
16
25.1
98.36%
0.4
24
25.1
24
25.0
99.49%
0.1
24
24
25.3
24
24.9
98.33%
0.4
40
24.8
40
24.7
99.57%
0.1
40
40
25.0
40
24.6
98.30%
0.4
60
24.6
60
24.5
99.64%
0.1
60
60
24.8
60
24.4
98.27%
0.4
80
24.4
80
24.3
99.68%
0.1
80
80
24.6
80
24.2
98.25%
0.4
100
24.3
100
24.2
99.72%
0.1
100
100
24.5
100
24.1
98.24%
0.4
Calculation
of
Acute
Temperature
Relationships
Appendix
C
Appendix
C
APPENDIX
C
ATTACHMENT
3
Acute
Effects
of
Temperature
on
Salmonids:
Times
to
10%
Mortality
(
LT10)
In
Relation
to
Temperature.
CHINOOK
SALMON
­­
Estimated
LT10s
Acclim.
Source
Age/
Size
a
b
N
r
Time(
hr)
LT10*(
C)

15C
Brett(
1952)
juv.
frshwtr
fry
16.4454
­
0.5364
4
­
0.9906
0.1
28.6
0.25
27.9
0.5
27.3
1
26.8
2
26.2
4
25.7
6
25.4
8
25.1
10
25.0
12
24.8
13
24.8
14
24.7
16
24.6
20C
Brett(
1952)
juv.
frshwtr
fry
22.9065
­
0.7611
7
­
0.9850
0.1
28.5
0.25
28.0
0.5
27.6
2
26.8
4
26.4
6
26.2
8
26.0
10
25.9
16
25.7
24
25.4
40
25.1
60
24.9
80
24.8
100
24.6
24C
Brett(
1952)
juv.
frshwtr
fry
18.9940
­
0.5992
9
­
0.9923
0.1
29.8
0.25
29.1
0.5
28.6
1
28.2
2
27.7
4
27.2
6
26.9
8
26.7
10
26.5
16
26.2
24
25.9
40
25.5
56
25.3
72
25.1
96
24.9
120
24.8
140
24.6
24
25
26
27
28
29
30
31
32
33
34
35
0.1
10
1000
Duration
(
hr)

Temperature
(

C)
LT10*
for
15C
Acclim.
(
Brett
1952)

LT10*
for
20C
Acclim.
(
Brett
1952)

LT10*
for
24C
Acclim.
(
Brett
1952)

LT10*
(
17C,
Coutant
1970)

LT10*
(
20C,
Coutant
1970)

24
25
26
27
28
29
30
31
32
33
34
35
0
40
80
120
160
Duration
(
hr)

Temperature
(

C)
LT10*
for
15C
Acclim.
(
Brett
1952)

LT10*
for
20C
Acclim.
(
Brett
1952)

LT10*
for
24C
Acclim.
(
Brett
1952)

LT10*
(
17C,
Coutant
1970)

LT10*
(
20C,
Coutant
1970)
CHINOOK
SALMON
­­
Estimated
LT10s
Acclim.
Source
Age/
Size
a
b
N
r
Time(
hr)
LT10*(
C)

17C
Coutant
(
1970)
"
Jacks"
(
1­
2
yrs)
13.2502
­
0.4121
4
­
0.8206
0.1
29.7
0.25
28.7
0.5
28.0
1
27.3
2
26.6
The
17C
and
19C
temperatures
were
Columbia
River
3
26.1
temperatures
(
at
Grand
Rapids
Dam)
during
fall
4
25.8
migrations
two
different
years.
6
25.4
8
25.1
10
24.9
12
24.7
14
24.6
19C
Coutant
(
1970)
"
Jacks"
(
1­
2
yrs)
9.4683
­
0.2504
4
­
0.9952
0.1
34.0
0.25
32.5
0.5
31.3
1
30.1
2
28.9
3
28.2
4
27.7
6
27.1
8
26.6
10
26.2
20
25.0
26
24.6
20C
Blahm
&
juv.
(
spring
run)
21.3981
­
0.7253
3
­
0.9579
0.1
27.9
McConnell
(
1970)
0.25
27.3
unpublished
data
0.5
26.9
1
26.5
2
26.1
4
25.7
8
25.3
10
25.2
16
24.9
20
24.8
24
24.6
26
24.6
24
25
26
27
28
29
30
0.1
1
10
100
1000
Duration
(
hr)

Temperature
(

C)
LT10*
for
20C
Acclim.
(
Brett
1952)

LT10*
(
20C,
fall
run,
Blahm
&
McConnell
1970)

LT10
(
20C,
fall
run,
Blahm
&
McConnell
1970)

LT10*
(
20C,
spring
run,
Blahm
&
McConnell
1970)

LT10
(
20C,
spring
run,
Blahm
&
McConnell
1970)

24
25
26
27
28
29
30
0
40
80
120
Duration
(
hr)

Temperature
(

C)
LT10*
for
20C
Acclim.
(
Brett
1952)

LT10*
(
20C,
fall
run,
Blahm
&
McConnell
1970)

LT10
(
20C,
fall
run,
Blahm
&
McConnell
1970)

LT10*
(
20C,
spring
run,
Blahm
&
McConnell)

LT10
(
20C,
spring
run,
Blahm
&
McConnell
1970)
CHINOOK
SALMON
­­
Estimated
LT10s
Acclim.
Source
Age/
Size
a
b
N
r
Time(
hr)
LT10*(
C)

20C
Blahm
&
juv.
(
spring
run)
22.6664
­
0.7797
4
­
0.9747
0.1
28.1
McConnell
(
1970)
0.25
27.6
unpublished
data
0.5
27.2
10%
mortality
NOT
ADJUSTED
1
26.8
2
26.4
4
26.0
6
25.8
8
25.6
10
25.5
12
25.4
16
25.2
20
25.1
32
24.9
48
24.6
20C
Blahm
&
juv.
(
fall
run)
22.2124
­
0.7526
4
­
0.9738
0.1
27.9
McConnell
(
1970)
0.25
27.4
unpublished
data
0.5
27.0
1
26.6
2
26.2
4
25.8
8
25.4
10
25.3
16
25.0
20
24.9
24
24.8
34
24.6
20C
Blahm
&
juv.
(
fall
run)
21.6756
­
0.7438
4
­
0.9550
0.1
28.1
McConnell
(
1970)
0.25
27.6
unpublished
data
0.5
27.2
10%
mortality
NOT
ADJUSTED
1
26.8
2
26.3
4
25.9
6
25.7
8
25.5
10
25.4
16
25.1
20
25.0
24
24.9
32
24.7
40
24.6
COHO
SALMON
­­
Estimated
LT10s
Acclim.
Source
Age/
Size
a
b
N
r
Time(
hr)
LT10*(
C)

15C
Brett(
1952)
juv.
frshwtr
fry
20.4066
­
0.6858
6
­
0.9681
0.1
28.0
0.25
27.5
0.5
27.0
1
26.6
2
26.2
4
25.8
6
25.5
8
25.3
10
25.2
12
25.1
16
24.9
20
24.8
32
24.5
20C
Brett(
1952)
juv.
frshwtr
fry
20.4022
­
0.6713
4
­
0.9985
0.1
28.6
0.25
28.1
0.5
27.6
2
26.7
4
26.3
6
26.1
8
25.9
10
25.7
16
25.4
24
25.2
40
24.8
48
24.7
56
24.6
70
24.5
23C
Brett(
1952)
juv.
frshwtr
fry
18.9736
­
0.6013
5
­
0.9956
0.1
29.7
0.25
29.0
0.5
28.5
1
28.0
2
27.5
4
27.0
6
26.8
8
26.6
10
26.4
16
26.1
24
25.8
40
25.4
56
25.2
72
25.0
96
24.8
120
24.6
140
24.5
24
25
26
27
28
29
30
31
32
0.1
10
1000
Duration
(
hr)

Temperature
(

C)
LT10*
for
15C
Acclim.
(
Brett
1952)

LT10*
for
20C
Acclim.
(
Brett
1952)

LT10*
for
23C
Acclim.
(
Brett
1952)

LT10*
(
17C,
Coutant
1970)

24
25
26
27
28
29
30
31
32
0
40
80
120
160
Duration
(
hr)

Temperature
(

C)
LT10*
for
15C
Acclim.
(
Brett
1952)

LT10*
for
20C
Acclim.
(
Brett
1952)

LT10*
for
23C
Acclim.
(
Brett
1952)

LT10*
(
17C,
Coutant
1970)
COHO
SALMON
­­
Estimated
LT10s
Acclim.
Source
Age/
Size
a
b
N
r
Time(
hr)
LT10*(
C)

17C
Coutant
(
1970)
adult
5.9068
­
0.1630
5
­
0.9767
0.1
30.8
Reported
acclimation
temp.
was
the
Columbia
River
0.2
29.0
temp.
(
at
Priest
Rapids
Dam)
during
fall
migration.
0.3
28.0
0.4
27.2
0.5
26.6
0.6
26.2
0.7
25.8
0.8
25.4
0.9
25.1
1
24.8
1.1
24.6
1.15
24.5
RAINBOW
TROUT
­­
Estimated
LT10s
Acclim.
Source
Age/
Size
a
b
N
r
Time(
hr)
LT10*(
C)

15C
Alabaster
&
juvenile
15.6500
­
0.5000
2
­­­
0.1
29.1
Downing
(
1966)
0.25
28.4
0.5
27.8
1
27.2
2
26.6
4
26.0
6
25.7
8
25.4
10
25.2
12
25.1
16
24.8
20
24.6
24
24.5
18C
Alabaster
&
juvenile
18.4654
­
0.5801
5
­
0.9787
0.1
29.9
Welcomme
(
1962)
0.25
29.2
Dissolved
Oxygen
at
7.4
mg/
L
0.5
28.7
2
27.7
4
27.2
6
26.9
8
26.7
10
26.5
16
26.2
24
25.9
40
25.5
80
25.0
120
24.7
160
24.5
18C
Alabaster
&
juvenile
13.6531
­
0.4264
5
­
0.9742
0.1
29.6
Welcomme
(
1962)
0.25
28.7
Dissolved
Oxygen
at
3.8
mg/
L
0.5
28.0
1
27.3
2
26.6
4
25.9
6
25.5
7
25.4
8
25.2
9
25.1
10
25.0
12
24.8
14
24.7
16
24.5
24
25
26
27
28
29
30
31
32
0.1
10
1000
Duration
(
hr)

Temperature
(

C)
LT10*
for
15C
Acclim.
(
Alabaster
&
Downing
1966)

LT10*
for
18C
Acclim.
(
7.4
DO,
Alabaster
&
Welcomme
1962)

LT10*
for
18C
Acclim.
(
3.8
DO,
Alabaster
&
Welcomme
1962)

LT10*
for
20C
Acclim.
(
Alabaster
&
Downing
1966)

24
25
26
27
28
29
30
31
32
0
40
80
120
160
Duration
(
hr)

Temperature
(

C)
LT10*
for
15C
Acclim.
(
Alabaster
&
Downing
1966)

LT10*
for
18C
Acclim.
(
7.4
DO,
Alabaster
&
Welcomme
1962)

LT10*
for
18C
Acclim.
(
3.8
DO,
Alabaster
&
Welcomme
1962)

LT10*
for
20C
Acclim.
(
Alabaster
&
Downing
1966)
RAINBOW
TROUT
­­
Estimated
LT10s
Acclim.
Source
Age/
Size
a
b
N
r
Time(
hr)
LT10*(
C)

20C
Alabaster
&
juvenile
19.6250
­
0.6250
2
­­­
0.1
29.6
Downing
(
1966)
0.25
28.9
0.5
28.5
1
28.0
2
27.5
4
27.0
6
26.8
8
26.6
12
26.3
20
25.9
32
25.6
40
25.5
80
25.0
120
24.7
160
24.5
20C
Craigie
(
1963)
yearling
14.6405
­
0.4470
3
­
0.9787
0.1
30.4
Raised
in
soft
water,
tested
in
soft
water
(
SS)
0.25
29.5
0.5
28.9
1
28.2
2
27.5
4
26.9
6
26.5
8
26.2
10
26.0
16
25.6
20
25.3
24
25.2
32
24.9
40
24.7
48
24.5
24
25
26
27
28
29
30
31
32
0.1
10
1000
Duration
(
hr)

Temperature
(

C)
LT10*
for
20C
Acclim.
(
Alabaster
&
Downing
1966)

LT10*
for
20C
Acclim.
(
SS,
Craigie
1963)

LT10*
for
20C
Acclim.
(
SH,
Craigie
1963)

LT10*
for
20C
Acclim.
(
HS,
Craigie
1963)

LT10*
for
20C
Acclim.
(
HH,
Craigie
1963)

24
25
26
27
28
29
30
31
32
0
40
80
120
160
Duration
(
hr)

Temperature
(

C)
LT10*
for
20C
Acclim.
(
Alabaster
&
Downing
1966)

LT10*
for
20C
Acclim.
(
SS,
Craigie
1963)

LT10*
for
20C
Acclim.
(
SH,
Craigie
1963)

LT10*
for
20C
Acclim.
(
HS,
Craigie
1963)

LT10*
for
20C
Acclim.
(
HH,
Craigie
1963)
RAINBOW
TROUT
­­
Estimated
LT10s
Acclim.
Source
Age/
Size
a
b
N
r
Time(
hr)
LT10*(
C)

20C
Craigie
(
1963)
yearling
15.0392
­
0.4561
3
­
0.9917
0.1
30.6
Raised
in
soft
water,
tested
in
hard
water
(
SH)
0.25
29.8
0.5
29.1
1
28.5
2
27.8
4
27.2
6
26.8
8
26.6
12
26.2
16
25.9
20
25.7
32
25.3
40
25.1
60
24.7
72
24.5
20C
Craigie
(
1963)
yearling
15.1473
­
0.4683
3
­
0.9781
0.1
30.1
Raised
in
hard
water,
tested
in
soft
water
(
HS)
0.25
29.2
0.5
28.6
1
28.0
2
27.3
4
26.7
6
26.3
8
26.1
10
25.9
16
25.5
20
25.3
24
25.1
32
24.8
40
24.6
48
24.5
RAINBOW
TROUT
­­
Estimated
LT10s
Acclim.
Source
Age/
Size
a
b
N
r
Time(
hr)
LT10*(
C)

20C
Craigie
(
1963)
yearling
12.8718
­
0.3837
3
­
0.9841
0.1
30.9
Raised
in
hard
water,
tested
in
hard
water
(
HH)
0.25
29.9
0.5
29.1
1
28.3
2
27.6
3
27.1
4
26.8
6
26.3
8
26.0
10
25.8
12
25.6
16
25.3
20
25.0
24
24.8
32
24.5
CUTTHROAT
TROUT
­­
Estimated
LT10s
Acclim.
Source
Age/
Size
a
b
N
r
Time(
hr)
LT10*(
C)

23C
Golden
(
1978)
juvenile
18.092
­
0.56523
?
­
0.996
0.1
30.0
1975
tests,
hatchery
only
0.25
29.3
0.5
28.8
1
28.3
8
26.7
16
26.2
24
25.9
32
25.7
40
25.5
60
25.2
80
25.0
120
24.7
160
24.5
13­
25C
Golden
(
1978)
juvenile
22.543
­
0.71999
?
­
0.999
0.1
29.6
1975
tests,
hatchery
only
0.25
29.1
0.5
28.7
2
27.9
4
27.4
8
27.0
16
26.6
24
26.4
40
26.1
80
25.7
160
25.3
240
25.0
300
24.9
560
24.5
23C
Golden
(
1978)
juvenile
18.3166
­
0.5723
?
­
0.990
0.1
30.0
1976
tests,
hatchery
and
wild
trout
data
pooled
0.25
29.4
0.5
28.8
1
28.3
2
27.8
4
27.3
8
26.8
12
26.5
16
26.3
24
26.0
40
25.6
80
25.1
120
24.8
160
24.5
24
25
26
27
28
29
30
31
32
33
0.1
10
1000
Duration
(
hr)

Temperature
(

C)
LT10*,
23C
Acclim.
(
Golden
1978,
1975
data)

LT10*,
13­
25C
Acclim.
(
Golden
1978,
1975
data)

LT10*,
23C
Acclim.
(
Golden
1978,
1976
data)

LT10*,
13­
23C
Acclim.
(
Golden
1978,
1976
data)

24
25
26
27
28
29
30
31
32
33
0
120
240
360
480
600
Duration
(
hr)

Temperature
(

C)
LT10*,
23C
Acclim.
(
Golden
1978,
1975
data)

LT10*,
13­
25C
Acclim.
(
Golden
1978,
1975
data)

LT10*,
23C
Acclim.
(
Golden
1978,
1976
data)

LT10*,
13­
23C
Acclim.
(
Golden
1978,
1976
data)
CUTTHROAT
TROUT
­­
Estimated
LT10s
Acclim.
Source
Age/
Size
a
b
N
r
Time(
hr)
LT10*(
C)

13­
23C
Golden
(
1978)
juvenile
18.1515
­
0.5723
?
­
0.992
0.1
29.7
1976
tests,
hatchery
and
wild
trout
data
pooled
0.25
29.1
0.5
28.6
1
28.0
2
27.5
4
27.0
6
26.7
8
26.5
12
26.2
20
25.8
32
25.5
40
25.3
60
25.0
80
24.8
120
24.5
SOCKEYE
SALMON
­­
Estimated
LT10s
Acclim.
Source
Age/
Size
a
b
N
r
Time(
hr)
LT10*(
C)

15C
Brett(
1952)
juv.
frshwtr
fry
15.8799
­
0.5210
7
­
0.9126
0.1
28.4
0.25
27.7
0.5
27.1
1
26.5
2
26.0
4
25.4
6
25.1
8
24.8
10
24.6
12
24.5
13
24.4
14
24.4
24
23.9
20C
Brett(
1952)
juv.
frshwtr
fry
19.3821
­
0.6378
5
­
0.9602
0.1
28.6
0.25
28.0
0.5
27.5
2
26.6
4
26.1
6
25.9
8
25.7
10
25.5
16
25.2
24
24.9
40
24.6
48
24.5
56
24.4
116
23.9
23C
Brett(
1952)
juv.
frshwtr
fry
20.0020
­
0.6496
4
­
0.9981
0.1
29.0
0.25
28.4
0.5
27.9
1
27.5
2
27.0
4
26.6
6
26.3
8
26.1
10
26.0
16
25.7
24
25.4
40
25.1
56
24.9
72
24.7
96
24.5
120
24.4
220
24.0
23
24
25
26
27
28
29
30
31
32
0.1
10
1000
Duration
(
hr)

Temperature
(

C)
LT10*
for
15C
Acclim.
(
Brett
1952)

LT10*
for
20C
Acclim.
(
Brett
1952)

LT10*
for
23C
Acclim.
(
Brett
1952)

LT10*
(
20C,
McConnell
&
Blahm
1970)

LT10
(
20C,
McConnell
&
Blahm
1970)

23
24
25
26
27
28
29
30
31
32
0
40
80
120
160
200
240
Duration
(
hr)

Temperature
(

C)
LT10*
for
15C
Acclim.
(
Brett
1952)

LT10*
for
20C
Acclim.
(
Brett
1952)

LT10*
for
23C
Acclim.
(
Brett
1952)

LT10*
(
20C,
McConnell
&
Blahm
1970)

LT10
(
20C,
McConnell
&
Blahm
1970)
SOCKEYE
SALMON
­­
Estimated
LT10s
Acclim.
Source
Age/
Size
a
b
N
r
Time(
hr)
LT10*(
C)

20C
McConnell
&
juvenile
16.7328
­
0.5473
6
­
0.9552
0.1
28.6
Blahm
(
1970)
(
under
yearling)
0.25
27.9
unpublished
data
0.5
27.3
1
26.8
2
26.2
4
25.7
6
25.4
8
25.2
10
25.0
16
24.6
20
24.4
40
23.9
20C
McConnell
&
juvenile
17.5227
­
0.5861
6
­
0.9739
0.1
28.6
Blahm
(
1970)
(
under
yearling)
0.25
27.9
unpublished
data
0.5
27.4
10%
mortality
NOT
ADJUSTED
1
26.9
2
26.3
3
26.0
4
25.8
6
25.5
8
25.3
10
25.2
12
25.0
20
24.6
32
24.3
52
23.9
CHUM
SALMON
­­
Estimated
LT10s
Acclim.
Source
Age/
Size
a
b
N
r
Time(
hr)
LT10*(
C)

15C
Brett(
1952)
juv.
frshwtr
fry
15.8911
­
0.5252
8
­
0.9070
0.1
28.2
0.25
27.5
0.5
26.9
1
26.3
2
25.8
4
25.2
6
24.9
8
24.6
10
24.5
12
24.3
16
24.1
20
23.9
40
23.3
20C
Brett(
1952)
juv.
frshwtr
fry
16.1894
­
0.5168
9
­
0.9750
0.1
29.2
0.25
28.5
0.5
27.9
2
26.8
4
26.2
6
25.9
8
25.6
10
25.4
16
25.0
24
24.7
40
24.3
60
24.0
80
23.7
132
23.3
23C
Brett(
1952)
juv.
frshwtr
fry
15.3825
­
0.4721
4
­
0.9652
0.1
30.3
0.25
29.5
0.5
28.9
1
28.2
2
27.6
4
27.0
6
26.6
8
26.4
10
26.2
16
25.7
24
25.4
40
24.9
56
24.6
72
24.4
96
24.1
120
23.9
240
23.3
23
24
25
26
27
28
29
30
31
32
0.1
10
1000
Duration
(
hr)

Temperature
(

C)
LT10*
for
15C
Acclim.
(
Brett
1952)

LT10*
for
20C
Acclim.
(
Brett
1952)

LT10*
for
23C
Acclim.
(
Brett
1952)

23
24
25
26
27
28
29
30
31
32
0
40
80
120
160
200
240
Duration
(
hr)

Temperature
(

C)
LT10*
for
15C
Acclim.
(
Brett
1952)

LT10*
for
20C
Acclim.
(
Brett
1952)

LT10*
for
23C
Acclim.
(
Brett
1952)
PINK
SALMON
­­
Estimated
LT10s
Acclim.
Source
Age/
Size
a
b
N
r
Time(
hr)
LT10*(
C)

15C
Brett(
1952)
juv.
frshwtr
fry
12.8937
­
0.4074
8
­
0.9884
0.1
29.1
0.25
28.2
0.5
27.5
1
26.7
2
26.0
4
25.3
6
24.9
8
24.6
10
24.3
12
24.1
16
23.8
20
23.6
24
23.4
20C
Brett(
1952)
juv.
frshwtr
fry
13.2444
­
0.4074
7
­
0.9681
0.1
30.0
16.2444
in
EPA
(
1977),
but
0.25
29.0
the
resulting
curve
does
not
0.5
28.3
match
the
information
1
27.6
presented
in
Brett
(
1952)
2
26.9
<
Fig.
5>
4
26.1
6
25.7
8
25.4
10
25.2
16
24.7
24
24.3
32
24.0
40
23.7
56
23.4
24C
Brett(
1952)
juv.
frshwtr
fry
14.7111
­
0.4459
6
­
0.9690
0.1
30.6
0.25
29.7
0.5
29.1
1
28.4
2
27.8
4
27.1
6
26.7
8
26.4
10
26.2
16
25.8
24
25.4
40
24.9
56
24.6
72
24.3
96
24.1
120
23.9
192
23.4
23
24
25
26
27
28
29
30
31
32
0.1
10
1000
Duration
(
hr)

Temperature
(

C)
LT10*
for
15C
Acclim.
(
Brett
1952)

LT10*
for
20C
Acclim.
(
Brett
1952)

LT10*
for
24C
Acclim.
(
Brett
1952)

23
24
25
26
27
28
29
30
31
32
0
40
80
120
160
200
Duration
(
hr)

Temperature
(

C)
LT10*
for
15C
Acclim.
(
Brett
1952)

LT10*
for
20C
Acclim.
(
Brett
1952)

LT10*
for
24C
Acclim.
(
Brett
1952)
