HomeMy WebLinkAboutAPA3501February 14,1986
Review Findings
Slat~j9f AlaSKa
~~-.~'#,
DATE:
FILE NO:
SUBJECT:
TELEPHONE NO:
ervisor
heries
and_Game
Ken Florey,e
Oivision of Co
Department of
FROM:
-......~...,-~¥~=~(MORANDUM_"""i
~0~0=~Susitna Program Review Team
~I!)-,....
_M
=M-
--
The attached final draft report "S us itna River Aquatic Studies Review,
Findings and Recommendations of the Susitna Program Review Team"is provided
for your review.I want to thank the team for their excellent corrments on
the draft;we've attempted to incorporate your techni cal comments wherever
appropriate.Please take one more quick look at the report;particularly in
regard to how your comments were treated and let me know by close of business
on March 7 if you have additi ona 1 comments.If I don't hear from you by
then,I'll assume that you are satisfied with the report and we will send it
through headquarters on to the Al aska Power Authority.Thank you for your
help.
cc:K.Tarboxc.Meacham
D.Schmidt
C.Estes
B.Barrett.-
-
ARLIS
Alaska Resources
Library &Information ServICes
Anchorage,Alaska
~O;l·[101MRp.v.10/791
-
-
--
Susitna River Aquatic Studies Review
Findings and Recommendations of the
Susitna Program Review Team
Prepared by
Richard Cannon
Division of Commercial Fisheries
Susitna Aquatic Studies Program
Alaska Department of Fish and Game
February 1986
~,
-
ARLIS
Alaska Resources
Library &InformatlOn ServIces
Anchoritge,Alaska
The following report provides a discussion of the background and
findings of the 1985 Susitna program review which was held on October 2
and 3 in Anchorage by the Alaska Department of Fish and Game.
INTRODUCTION
.-
Commi ss i oner Co 11 i nsworth in his June 18,1985 memorandum
.....
....
.....
-
I'"'"
-!
•L.
_L
(Reorganization of the Susitna Aquatic Studies),assigned the project
coordinator of the Susitna Aquatic Studies program under the direction
of the Commercial Fisheries Division regional supervisor,Central Region
with the responsibility for planning and coordinating all departmental
salmon escapement activities on the Susitna River.In addition,the
coordinator was to administer the contract between the Alaska Power
Authority (APA)and the Division of Commercial Fisheries.The
,coordinator was also instructed to prepare a technical data report which
summarized all salmon escapement data collected to date for the Susitna
River by the combined efforts of the Divisions of Sport Fisheries and
Commercial Fisheries and was to coordinate a cooperative effort among
the fisheries divisions to ensure proper planning for future Susitna
River escapement studies.
In order to accomplish these tasks,the Division of Commercial Fisheries
has initiated the following:1)staff orientation and coordination
activities for staff assigned to Susitna River projects,2)the
preparation of the Susitna salmon escapement summary,and 3)a Susitna
program revi ew.
-1-
-
,....
--
.-
j
I.Coordination Staff orientation and coordination efforts have
included information exchange,joint field inspections of various
projects,and combined Central Region/Susitna Aquatic Studies staff
planning meetings.Although this process is just beginning,significant
improvement in communication and cooperation among the various
departmental projects operating on the Susitna is already evident.
These efforts will continue with an emphasis on including more
involvement by Sport Fisheries Division regional staff in the future.
II.Data Summary A first draft of a "Susitna Salmon Escapement Summary
Report fl has been completed and will be available for departmental review
in January.The report presents escapement data collected to date by
tributary and/or river reaches of the Susitna.The information will be
a available in a bound report format and on computer disk.This summary
should prove very useful during planning for future Susitna salmon stock
assessment programs,developing management plans,and preparing board
reports.
III.Program Review The end of the 1985 open water field season will
complete the final year of aquatic baseline studies which have been
funded to characterize the resources of the Susitna River and to assess
potential impacts resulting from construction and operation of the
proposed hydroelectric project.Fisheries data collected by the Depart-
ment for the Susitna during past years with particular emphasis given to
the last five years of aquatic habitat evaluation will form the basis
for impact assessment and mitigation plans to be prepared by APA .
-2-
-
....
....
j
Beginning in 1986,the objectives of the hydropower development program
will undergo some important changes.Assuming the state will move for-
ward with construction of the dams,a long-term fisheries population
monitoring program will be initiated.This shift in direction of the
program provides the Department,not only the opportunity to assess its
past data collection programs related to the hydropower project,but to
evaluate and to better coordinate all of our fisheries programs for the
Sus itna.
A program review of all its fisheries projects on the Susitna River was
held to assist the Division plan its long-term fisheries monitoring and
stock enumeration programs.A review team was formed which included the
South Central regional supervisors from the Divisions of Habitat,
Commercial Fisheries,and Sport Fisheries,and the chief biometricians
from the Commercial Fisheries and Sport Fisheries Divisions {Table I}.
The team was asked to review the Divisions'fisheries study projects to
evaluate 1}technical merit,2}cost effectiveness,and 3)relevance to
primary management goals.
Table 1.Susitna Aquatic Review Team.
Review Team Staff ~/Position Division Phone Number
Paul Krasnowski Regional Supervisor SPORT FISH Division 267-2168
Michael Mi 11 s Biometrician 111 SPORT FISH Division 267-2369
Kenneth Florey Regional Supervisor COMMERCIAL FISHERIES Div.267-2104
Douglas Eggers Bi ometrici an III COMMERCIAL FISHERIES Div.267-2104
Carl Yanagawa Regional Supervisor HABITAT Division 267-2283
a!Mailing address for all participants -333 Raspberry Road,Anchorage,Alaska 99502
-3-
The Division presently has two primary management goals for the....
Susitna.The first goal is ENSURE THAT OPTIMUM SPAWNING
ESCAPEMENTS ARE MAIIHAINED FOR SUSITNA RIVER STOCKS.The second
primary goal addresses monitoring potential changes resulting from
the construction and operation of the Susitna Hydroelectric Project
on fish populations and their habitats.More specifically the goal
is DESCRIBE THE NATURAL PRE-PROJECT VARIATIONS IN FISH POPULATIONS
AND THEIR HABITATS AT A LEVEL OF RELIABILITY NECESSARY TO DETECT
AND EXPLAIN POSSIBLE FUTURE CHANGES CAUSED BY HYDROELECTRIC
DEVELOPMENT.The review team was asked to use the following list
of questions to structure their comments and recommendations.
l.Do the objectives of the proposed long term monitoring-projects appear to adequately address the Division's primary
goals?How might project objectives be modifi ed to more
clearly address our goals?
,.-
2.If you believe that technical difficulties associated with the
project may exist,what are they and how might they be
remedied?
-
3.Coul d the sampl ing programs be modified to reduce project
costs without unreasonably sacrificing technical quality?
4.What priority should you give these projects in terms of
meeting the Division's primary goals?
-4-
The basic format for the review employed an outline similar to the 1984
project review for the Kenai River Chinook Program.Each project leader
submitted an executive summary and a more detailed operational plan for
their respective projects.These materials in addition to some
"'...
-
,....
additional background information were organized into a briefing book
which was distributed to the review team.
During the first session,six divisional projects on the Susitna were
presented.These included:1)Lower River Salmon Escapement,2)Middle
River Salmon Escapement,3)Middle River Salmon Outmigrant Evaluation,
4)Lower River Spawning Habitat Evaluation,5)Middle River Resident
Fish Study,and 6)Aquatic Habitat Modelling.The first of these proj-
ects represents and ongoing effort by the Division to develop an annual
estimate of total escapement for all salmon species in the Susitna.The
information is primarily used by the Division to set and evaluate post-
fishing season escapement goals for Susitna River salmon.The remaining
five projects are all part of the five year APA baseline studies program
which will be completed during FY 86.
Each project leader presented an overview of their respective project
which included project objectives,experimental design,sample sites,
assumptions,sources of error,and major project results and con-
elusions.
The first set of presentations provided the review team with an oppor-
tunity to examine existing fisheries programs on the Susitna.During
the second session,a proposed long-term program for monitoring hydro-
-5-
"""'
....
....
,.,..
.....
.....
electric caused impacts on fisheries resources was outlined and
discussed.Because the proposed monitoring program consisted primarily
of a continuation of existing projects,the team decided to direct their
comments and recommendations to those projects which may be continued.
In addition,the team wanted to specifically address the problem of
widely varying population estimates for salmon in the lower river.
Various strategies for estimating total Susitna salmon escapements were
al so discussed by the team,however;it was decided to postpone
deve1 opi ng any specifi c recommendati ons on thi s subject until results of
a sonar research program are available and a clearer picture emerges
concerning the future of APA funded studies.
FINDINGS AND RECOMMENDATIONS
The findings and recommendations of the Susitna review team are pre-
sented below.This section of the report is organized into three sub-
sections:1)adult salmon population estimates,2)long term monitoring
of salmon resources,and 3)long term monitoring of resident fish.Each
subsection provides a summary description of the project (i.e.problem
statement,scope,objectives,discussion of assumptions)followed by the
review team1s comments.A summary of the review team's findings and
recommendations is provided on pages 42-46.
-6-
.....
.....
.....
....
.....
I.Adult Salmon Populations Estimates
A.Problem Statement
Salmon stocks utilizing the lower and the middle Susitna River are one facet
of the fisheries resource that may be impacted from operation of the proposed
hydroelectric development at Devil and Watana canyons.Because of this
potential impact,APA has contracted the Department to provide baseline data
on the escapement of anadromous fish to the middle and to a lesser extent,the
lower river reaches •
B.Scope
Adult salmon captured with fishwheels were tagged and released at Flathorn,
1 Sunshine,and Curry stations (Figure 1).Sonar counters and tag/recapture
methods were used concurrently at Sunshine and Talkeetna stations in 1981 and
1982.Sonar counters were used to enumerate salmon populations in the Yentna
River in 1981-1985.Tag recovery and spawning ground surveys were conducted
in all middle river sloughs and streams and to a limited extent in lower river
sloughs and streams.Popul ati on est"imates were deri ved by the Petersen
method.Tag recoveries were also used to evaluate the rate of travel between
fishwheel sites.The adult salmon population parameters (i.e.,age,length,
and sex)were determined from a sUbsample of the fishwheel catch .
-7-
.-
MIDDLE RIVER
LOWER
RIVER
REACH
o 10 20 30
I , I J
MILES
(Approx.Scale)
.....Figure 1.Location of Susitna River study locations .
-8-
.....
-
C.Objectives
1.Determine the abundance of sockeye,pink,chum,and coho salmon at
Flathorn (1984-1985),Yentna (1981-84),Sunshine (1981-85),Talkeetna
(1981-84),and Curry (1981-85)stations.Also,determine the abundance of
chinook salmon at Flathorn (1985),Sunshine (1982-85),Tal keetna
(1982-84),and Curry (1982-85)stations.See Table 2.
2.Evaluate the migrational timing and migrational characteristics (rates of
travel and bank/channel preference)of adult salmon at stations operated
from 1981-85.
""'"!
~,
"...
3.Monitor the age,length,and sex composition of the adult salmon
escapements at the locations and years defined in objective 1.
4.Assess the extent of adult salmon spawning in middle river side channel,
main channel,slough,and tributary habitats (1981-85).
5.Assess the extent of adult salmon spawning in lower river main channel,
side channel,slough (1981-84,except 1983,)and tributary stream mouth
(1984-85)habitats.
-9-
""",
Table 2.Population estimates by location and year for salmon species
.....
2./
ESCAPEMENT.....SAMPLING
LOCATION YEAR CHINOOK SOCKEYE £/PINK CHUM COHO TOTAL
Fl athorn 1984 c/605,800 3.629.900 812,700 190,100 5,238,500r-Station
Yentna 1981 139.400 36,100 19.800 17 ,000 212,300
Station 1982 113,800 447.300 27,800 34,100 623,000
1983 104,400 60,700 10,800 8,900 184,800
1984 149,400 369.300 26,500 18,200 563,400
Sunshine 1981 c/133,500 49,500 262,900 19,800 465,700
Station 1982 52-;900 151,500 443,200 430,400 45,700 1,123,700.....1983 90,100 71,500 40,500 265,800 15,200 483,100
1984 121.700 130,100 1,017,000 765,000 94,700 2,128,500
Talkeetna 1981 c/4,800 2,300 20,800 3,300 31,200
Station 1982 10:900 3,100 73,000 49,100 5,100 141.200
1983 14,400 4,200 9,500 50,400 2,400 80,900
1984 24,800 13,100 177 ,900 98,200 11 ,800 325,800
Curry 1981 c/2,800 1,000 13,100 1,100 18,000.....Station 1982 11-;300 1,300 58,800 29,400 2,400 103,200
1983 9,700 1,900 5,500 21,100 800 39,000
1984 18,000 3,600 116,900 49,300 2,200 190,000
2./Escapement estimates were derived from tag/recapture population estimates except Yentna
Station escapements which were obtained using side scan sonar.
~£/Second-run sockeye salmon escapements only.
E./Chinook salmon were not monitored for escapement.
-10-
.....
......
.....
.....
....
D.Discussion of Methods
The tag/recapture projects used the modified Petersen estimator and the
sonars were 1980 model Bendix side scan units.There were discrepancies
between population estimates from sonar and estimates from the tag/
recapture method.Both estimates have inherent deficiencies.
It should not be assumed that all fish pass over the sonar substrate.
The sector distribution of salmon will vary with site and species with
an undetermined number of salmon passing beyond the counting range.A
major source of error present in sonar counts is related to the methods
of apportionment and the bias inherent in those methods.Although all
fishwheels used to apportion counts were in close proximity to the
counters,it must be recognized that fishwheels are species selective.
,The apportioned sonar counts would then reflect the selected catch-
ability of the fishwheel.In addition,sonar counters are adjusted for
fish velocity and sensitivity,thereby introducing an unknown variance
component into the counts.
Five assumptions were made for estimating population size using tag/
recapture methods and the Petersen estimate in particular.Failure to
meet the following assumptions will bias the population estimate.
1.Ei ther the capture of fi sh for taggi ng or the subsequent capture
and/or observation of fi sh to determi ne the tagged/untagged ratio
was random with respect to the population.
-11-
....
r-
I
2.There was no differential mortality between tagged and untagged
salmon.
3.Tagged salmon mixed randomly within the population.
4.Recovery of tagged salmon was not influenced by the tag.
5.There was no unknown tag loss.
In summary,both methods of enumerating salmon have potential drawbacks
but at this point,they represent the state of the art in estimating
population sizes in glacial river systems.
G.Review Team Comments
The review team focused its attention on unexplained differences in
population estimates generated by sonar counts and tag/recapture methods
for Susitna River salmon.As an example,the estimate for sockeye in
1984 for Flathorn Station was discussed.An escapement of 605,800 fish
at Flathorn was calculated from tag/recapture data;the estimate upriver
at Sunshine Station was 130,100 fish.Most of the 475,700 sockeye
which don't move past Sunshine should be accounted for by the Yentna
escapement and spawni ng areas between Fl athorn and Sunshi ne stations.
The sonar count for Yentna sockeye in 1984 was 149,400 fish;thus an
estimated 326,300 sockeye should be spawning in the river reach between
Flathorn Station and Sunshine Station.However,based on our present
-12-
"..,
-
.....
-
»
knowledge of sockeye production in the lower Susitna,this difference
can not be explained biologically.
Sampl ing error associated with either the Yentna sonar estimate being
negatively biased and/or the Flathorn (tag/recapture)estimate being
positively biased have been offered as explanations at least in part for
these differences.It is generally accepted that both the tag/recapture
estimates and sonar counting techniques have inherent biases.A major
source of error associated with sonar is species apportionment which in
a multi-species,multi-stock system like the Susitna is extremely
difficult to determine.
Use of fishwheels for species apportionment and capturing fish for
tagging was discussed by the review team.Fishwheel catch data strongly
suggests that individual wheels (sites)can be selective for species and
size classes within species and that catch efficiency at a site can vary
within and between seasons.It is acknowledged that fishwheel capture
is not random with respect to the population.
Selection of fishwheel sites is an important consideration for planning
future monitoring programs.The effect of milling on catch data and
tag/recapture estimates for the Sunshine and Curry fishwheel sites has
been proposed as an expl anati on for 1arge differences between
tag/recapture population estimates and estimates derived from spawning
ground counts.Bank migration and cross-channel migration patterns
between wheels is suspected at some sites.Because cross-sectional
-13-
."..
-
.....
-
migration patterns at any site are probably related to velocity
gradients near the bottom which vary with changes in flow and often
shifting channel morphology,comparison of fishwheel data between sites
and even for a given site over time may be suspect.
Preliminary results from the (Biosonics)hydroacoustic study conducted
last summer of the cross-sectional distribution of migrating salmon at
Susitna Station (just upriver from Flathorn)suggest that a high degree
of within season variation in spatial distribution of migrating salmon
occurs at this site.Some mid-channel migrational distribution for all
salmon species was observed,however,approximately 95%of the migration
was identified near the east and west banks.These results emphasize
the importance of developing a monitoring strategy for a site based on
the distributional patterns of the fish.
The review team identified a problem associated with the sample design
employed for tag/recapture population estimates.Because fish are not
captured randomly in fishwheels for tagging,recapture on the spawning
grounds must be random in order to develop a val id Peterson estimate.
Stream surveys used to obtain tagged/untagged ratios on the spawning
grounds have been predominately focused on middle river spawning areas
(above the three rivers confluence).Major spawning areas in the Yentna
River,Chulitna River,and Talkeetna River receive less survey effort.
Because large numbers of fish were tagged and intensive surveys in the
middle river resulted in a relatively high percent of recaptures,the
-14-
-
'"""
precision of population estimates has been reasonably good;however,
due to the violation of assumptions,the accuracy of the estimates are
questionable.The relatively high sockeye salmon estimate for Flathorn
Station,which was discussed above,might be explained by this error in
the sample design.Yentna bound sockeye tagged at F1athorn would be
lost to the surveyed sockeye population,however;this loss was not
considered in the calculation of the population estimate.If the
proportion of tagged sockeye migrating up the Yentna was significantly
larger than those continuing up the Susitna,a large overestimate of the
population would result.
Currently,the stream survey tag recoveries use visual assessment of the
number of tagged and untagged fish present.This data is useless for
the population estimates because the tag number cannot be recorded.If
this is to be continued,a less biased method of recovery should be
used.Samp1 ing a portion of the surveyed streams with beach seines
would provide a better methodology for determining the number of tagged
and untagged fish present.This method would be especially beneficial
in areas where the streams are highly turbid due to glacial influence or
run-off.In addition,individual fish would be handled and the
incidence of tag scars would allow a better estimate of tag retention
rates.
Overall,the tag/recapture program for the Susitna was viewed by the
team as beneficial to the Department and a high priority.Population
estimates for Sunshine Station coupled with Yentna sonar counts
-15-
-
-
.....
-
-
presently provide the only indicator available to the Department for
eva 1uati on and estab 1i shment of post-season escapement goals for the
Susitna.The team concluded that sample design problems for
tag/recapture estimates can be addressed and resolved.
The Program Review Team recommended the following actions:1)Compare
tags recovered at fishwhee1s on the Yentna and Susitna to develop some
indication of how tagged fish are distributed.If Yentna ratios are
higher than Sunshine,it may be possible to account for loss of tagged
fish migrating up the Yentna.2)Numbered tag recovery data from
surveys and fishwheels should be evaluated to assess differences in gear
efficiency among fishwhee1 sites.3)A time stratified population
estimate should be employed in the future.The need for stratification
by size (length of fish)should also be evaluated for salmon species
when different capture probabil i ties are suspected for vari ous size
groups in the population.4)Spawning ground surveys should be
conducted throughout the drainage (more random)to minimize the sample
design problem discussed above.5)Employ capture methods (i.e.beach
seines)which would allow tag recoveries rather than visual counting
methods on the spawning grounds so that size stratification and tag
retention can be more effectively addresed in the future,and 6)The
Chief Fisheries Scientist Office within Commercial Fisheries Division
should consider the species apportionment problems associated with sonar
counts on the Susitna as a high research priority.
-16-
II.Long-Term Monitoring for Salmon
A.Problem Statement
Hydroelectric development may impact the fish resources of the
middle river which has been defined as the reach extending upstream
....
from the three rivers confluence near Talkeetna to Devil Canyon.A
long-term plan to monitor natural variability in the numbers of
salmon which utilize the middle reach needs to be developed.The
design should provide an accurate measure of variability in Susitna
River stocks under natural conditions which could ultimately be
compared to variability under with-project conditions.
This proposed strategy introduces a conceptual approach on how this
task might be accomplished by monitoring the adult salmon and
juvenile salmon outmigration from the middle river reach.Some
additional approaches which could be explQred are also mentioned.
.-
B.Objective To develop a long-term monitoring plan which will
provide the capability to detect hydroelectric development related
impacts to salmon produced in the middle reach of the Susitna
Ri ver.
C.Procedures The basis of this proposal is that natural variation in
pre-project adult escapement and juvenil e outmigrati on wi 11 be
defined and quantified so with-project impacts,if any,can be
-17-
--
--
-
-
detected.With-project variation would have to be outside the
pre-project variation to be considered project impact.This
strategy suggests that pre-project va ri ation can be described by
monitoring over time a Survival Monitoring Index (SMI)based on
annual fry to adult ratios.This plan would require the annual
collection of CPUE data at two points,a IItreatment"and a
II con trol ll
•The objective is to monitor the SMI in the middle reach
of the Susitna River (the treatment)and compare it to the SMI of a
control reach.The ratio of the treatment to the control would be
monitored for several years under natural conditions.A basic
assumption is that any post dam (i.e.,with-project)change in the
ratio outside of the natural variation would indicate a dam-caused
effect.
1.Treatment Adult salmon CPUE data for the treatment population
would be collected with fishwheels at Curry Station (RM 120).
There already exists five years of adult data at this
location.Juvenile CPUE data would be collected at either
Curry or Talkeetna station (RM 130).Juvenile data has been
collected at Talkeetna Station for the last four years.The
rationale for proposing that the outmigrant data collection be
moved to Curry is simplicity of logistics.It is untested if
the da ta from Ta 1 keetna wou 1d be campa rab 1e to Cu rry if pa st
data were to be used.Data should be collected at both points
for several seasons to test the assumption that they are
compa rab 1e.
-18-
.....
-
-
2.Control Adult CPUE data for the control would be taken at
Sunshine Station (RM 80)with fishwheels.Five years of adult
data have been collected at this site.The juvenile data
would be collected near the Parks Highway Bridge (RM 83)just
upstream from Sunshine Station where the river flows through a
single channel.A consideration regarding the future use of
CPUE data is that the sampling design associated with past
studies has emphasized maximizing catch efficiency to either
mark adults or recapture juvenile fish for mark recapture
population estimates rather than maintaining a constant
fishing efficiency.The maintenance of a constant fishing
efficiency or at least an unbiased estimate of it would be
needed to use CPUE data as an index to population size.
3.Calculations A probable way of calculating fry to adult
ratios is to base the ratio on a CPUE index of the number of
adult and juvenile salmon for the years in question.The
ratio is defined:
Fry/Adult Ratio =Outmigrant Index
Inmi grant Index
-19-
Major sample design questions are how to obtain the required
indexes and how to estimate their variances and the variance
of the ratio.For adult salmon,the fishwheels are operating
before the fish arrive and continue operating until virtually
all of the fish have passed by the fishwheels on their way
upstream;so it is fairly easy to convert each day's CPUE to a
24 hour CPUE and then sum over the entire season.The period
when the fishwheels are operating has been and should remain
constant from year to year.
Developing a CPUE index for juveniles poses some additional
problems.Ideally,the period for which the outmigrant traps
is operated would be constant from year to year relative to
the timing of outmigration (Figure 2).The peak outmigration
,~
.-
LLI
::l
a-
u
'-f-TRAPS OPERATlNG------1I .
I Irpeakdelayedf
....."'\2 weeks
" \I,I
1,"\_
'-,J'
Figure 2.Operating Period
-20-
__year X
----year Y
....
..-
I
"""
would have to be detected annually.This may not be possible
however,because climatic differences from year to year change
the timing.There is also evidence that some fry outmigrant
under the ice before the outmigrant traps are in place.To
resolve this problem,the CPUE might be summarized by
adjusting to a 24-hour period,for all those days when the
CPUE exceeds an empirically derived percentage of the peak 24
hour CPUE for the entire season.This method would result in
an index that would be comparable from year to year provided
that the pre-or post-season outmigration rates did not exceed
the percentage selected.
After several years of data have been collected,the inter-
annual variance of either the adult on juvenile indices could
be estimated by using the variance of the mean of several
years of indices.This may best be approached by taking the
mean of the index from 0-10 percent,11-20 percent,etc.of
the cumulative migration.Calculation of these adult and
juvenile indices would be done for both the treatment and the
control populations;and then the one number which would be
calculated each year is:
Survival Monitoring Index =Fry/Adult Ratio for Treatment
Fry/Adult Ratio for Control
-21-
"""
.....
.....
-
-
4.Variance/Structure The ability to statistically detect change
beyond that occurring naturally is the basic goal of any
environmental monitoring program.It is hoped that it will be
possible to detect dam-caused change in the salmon populations
of the middle Susitna River with this proposed methodology.
Variance arises from both natural processes and from sampling.
We can probably assume that the sampling variance is the same
for the treatment and the control because similar methods will
be used at both areas.We also have to assume that the
factors affecting year to year natural variation work equally
on both important ratios.
The survival monitoring index (SMI),is probably under natural
conditions somewhere around one because survival in the two
areas shou 1d be rough 1y equa 1.Let I s assume for a moment that
the dam did cause a substantial decrease in the fry/adult
ratio I for the treatment.A depiction of what time series
plots of fry/adult ratios and the SMI might look like are
presented in Figure 3.
The question we would want to ask then is whether the drop in
the SMI was significant.If the fry/adult ratios for the
treatment and the control have large variances,then only a
large change over a long period of time in the SMI will be
statistically detectable.
-22-
--
-It is not possible to "cancelli these large variances by
dividing the two.So the question now is what percentage
...Jl"'dam ,rdam
..J ,control::>
0
~::;;
"-\,\~,,,"en
:>-" \'I.'..',,
~a:•..'t~eatment~
YEARS YEARS
Figure 3.Time series plots of Fry/Adult Ratios and SMIts.
change in the treatment/control ratio is required before a
change can be statistically detected?What this percentage is
could easily be calculated before the fact if one knows the
variance structure of the ratio.There is probably no good
way to make this calculation until we have a number of years
of data.We will have to calculate the index for several
years and see how much it vari es.If there is a 1arge degree
of variance between the years,the model will not accomplish
the objective.
D.Assumptions The following are assumptions which to a greater or
lesser degree need to be valid for t~e SMI approach for detecting
impacts to work.
-23-
1.Sa 1mon stocks in the Ta 1keetna and Chulitna ri vers,and tha t
portion of the Susitna River in the control are subject to the
same natural variability as the treatment portion of the
Susitna River.
2.The Cook Inlet commercial fishery does not selectively fish
for any stock or species migrating above Sunshine Station (the
control).
3.That adult fishwheel CPUE is primarily related to total
escapement regardless of the extent of milling.
4.The peak of juvenile outmigration for all species occurs
during the open-water period.
....
....E.
5.Smolt trap and fishwheel efficiency remains constant.
6.A change in the position of the fishwheels as a result of
with-project flow will not adversely affect the efficiency of
the wheels .
7.Any with-project change outside the natural variation of the
ratio would be a dam-caused effect.
Problems Based on past observation,a number of potential problems
have been identified which suggest that our model assumptions may
be violated.
-24-
•
j
1.It is not likely that the control will be subject to
exactly the same variability as the treatment.
a.The control (Sunshine)is not a true control because it
includes treatment effects.The Talkeetna River may be a
better control and should be considered.
b.The treatment and the control experience different and
varying milling rates.
c.Gear efficiency varies with discharge,debris loading,
catch rates,etc.
-
.....
2.Adult milling is not considered in the CPUE ratio equation.
3.The peak migration of chum and pink salmon may occur before it
is physically possible to place outmigrant traps in the river
(i.e.,under the ice or with break-up).
4.The variance in the natural ratio may be so large it will be
impossible to detect a with-project effect (until a
catastrophic impact has occurred)•
....
F.Options/Alternatives The following alternatives to the
proposed SMI model are being considered:
-25-
.....
-
--
.-
j
1.Time series analysis where the natural variability at one site
(Curry)is used for both the treatment and the control.This
method has a problem where several years would pass before a
change could be detected unless projected physical parameters
regimes were linked to an index (survival,adult inmigration,
juvenile outmigration)with a transfer function model .
2.Evaluate habitat for a specific 1 ife stage (for example,
incubation)in conjunction with the adult inmigration and
juvenile outmigration.
3.Use the Talkeetna River,Chulitna River,or both as the
control site.The reason Sunshine Station was tentatively
chosen was because five years of adult salmon inmigration
has been collected at this site.However,collecting both
juvenile outmigrant data and the adult data may be more
feasible on the Talkeetna or Chulitna rivers.
G.Review Team Comments
The review team discussed the conceptual basis for a Survival
Monitoring Index.Dana Schmidt,who had proposed the SMI ap-
proach when he was working on the Susitna Aquatic Studies Pro-
gram,participated in the discussions.The team voiced strong
support for conducting post-project habitat utilization moni-
toring in conjunction with population monitoring as a means to
-26-
--
-
j
verify impact projections and the effectiveness of mitigation
plans.Habitat monitoring objectives were not discussed by
the revi ew team in any deta i 1;rather,it was deci ded that
specific habitat monitoring needs would be addressed by the
divisions as part of the Department's involvement in a
mitigation plan for Susitna.There was general agreement
about the importance of not relying on either habitat or
population monitoring alone.The need for a unified
departmental position on long-term monitoring was emphasized.
It was recommended that the management divisions work closely
with Habitat Division to develop the Department's position.
The team as a whole accepted the concept of using a project
control that woul d not be affected by the construction and
operation of the dam to monitor natural variations in salmon
production.However,the review team strongly doubted the
feasibil ity of successfully implementing the SMI approach
given the difficulty of developing a sensitive indicator of
change for four independent population parameters which
individually are subject to large natural variations.
Rather than attempt to monitor environmental change with one
all encompassing parameter,the proposed SMI,the team
recommended employing a number of key parameters which could
be evaluated independently as well as interrelated when
appropriate.The reliability of using Sunshine Station as a
-27-
.....
-
control when the treatment reach (middle Susitna River)would
be included as part of the control was also questioned.In
100
90
80
z
~70
l-/,<t 60 I>
ILl Ia:;0 I
l-Iz40Iw
u ~oc
ILl 30 "/a."/"/20 "/"/"-
10
"rf
/
a
1981 1982 1983 1984 1985
YEAR
a "l(,O~VIATION FRO"5 Y~AR "~AN OF CURRY ESCAPEMENT
(:).Y.O£'VtATtON FROM e 'tEAR M,fAH OF SUNSHINE ESCAPEMENT
•-I.VARIATION BETWEEH CURR't S SUNSH'NE £SCAP£W£NTS
Figure 4.Annual variation in chum salmon escapement,percent
variation from the five year mean.
addition,the team pointed out that the population levels
being assessed at the proposed control are typically at least
an order of magnitude greater than the proposed treatment
site.Population size could have a significant effect on the
ability of the proposed sampling gear to index the populations
and hence influence sampling variance.
Dana Schmidt presented a comparison of chum salmon population
escapement estimates for Sunshine Station and Curry Station
for the five year pedod 1981-1985.
-28-
The percent variation
•,.,...
....
-
....
j
from the 5-year mean for both sites appeared to track
reasonably well (Figure 4);however,a comparison for chinook
salmon did not support the assumption that the control and
treatment are subject to simi 1ar natural trends.Compari son
of population estimates of juvenile salmon between the control
and treatment was not possible because juvenile estimates have
not been developed for a control site.Four years of juvenile
population estimates are available for the treatment reach
(Talkeetna Station site.)
It was pointed out that the proposed method of observing the
variation in the SMI index over a period of years would just
reflect inter-annual variation of its components and does very
little to help determine how precise the SMI index may be.If
an expression for the variance of the SMI index as it,is
related to its individual components is not derived the degree
of change that can be detected (and is stati sti ca lly
significant)by it will not be known.
The use of an alternative site as a control was briefly dis-
cussed.The Tal keetna River has been considered in the past
as an alternative study control.Collection of juvenile out-
migrant data may be more feasible at a Talkeetna River site
than at the Parks Highway crossing site near Sunshine Station.
The longer width of the river and high water velocities at the
-29-
....
.-
.-
....
Parks Highway site significantly increases the difficulty and
risk of operating juvenile inclined plane traps at this site.
However,Sunshine Station has proven to be an excellent
fishwheel capture site for adult salmon.The team did not
make any recommendations concerning alternate sites other than
emphasizing the need to maintain Sunshine Station for adult
monitori ng.
The use of CPUE data rather than popul ation estimates for
adults was not supported by the review team.Although
correlation between CPUE for fishwheels and population es-
timates were reasonably high for some species;correlation
based on only four data points was considered inconclusive.
The teams also expressed concern about the highly variable
catch efficiency of fishwheels.The wheels are moved pe-
riodically during the season because their catch efficiency
varies with changing flows.Size selectively is also suspect-
ed for some fishwheels sites.Milling has been identified as
a possible source of bias for fishwheels sites,especially the
Curry site where milling has been observed during
tag/recapture studies.Maintenance of constant catch
efficiencies of fishwheels would be needed to appropriately
employ a CPUE index;however,not only ;s maintenance of a
constant efficiency impractical,measurement of efficiency is
not possible without a independent estimates of population
size.
-30-
....
....
"""
.....
Although sampling design problems associated with adult salmon
populations estimates were identified (see discussion of popu-
lation estimates above)~these problems were considered solv-
able through time and size class stratification of tag de-
ployment and a more random allocation of recapture effort.
The population estimates were considered a much more reliable
monitoring parameter then CPUE data.The spawning ground
surveys which are used to recapture tags were also viewed as a
valuable habitat utilization monitoring activity which should
be continued.Pre-and post-project observations of stream
life~distributional timing~and habitat use information
obtained via the surveys may provide additional indicators of
change caused by the dams .
Inclusion of juvenile outmigrant CPUE from inclined plane
traps in the monitoring plan was discussed at length by the
revi ew team.Conceptua lly ~the outmi grant component of the
proposed SMI index should provide the most sensitive indicator
of dam caused change on fresh water production.Monitoring of
only adult salmon escapement has the risk of not accounting
for differences in the saltwater life phase of the control and
treatment populations.Some indication that treatment and
control populations are not disproportionally affected during
the ocean phase could be derived by monitoring and comparing
timing~size~and age composition of returning adult stocks .
An additional concern was that a dam-caused effect on
-31-
-
,--
freshwater production might not be observed until the adult
population returned;a period of three to five years for some
species could elapse before the problem could be detected.
The review team unanimously expressed their concern as to the
feasibility of implementing a juvenile population index.
Experience gained from work on monitoring juvenile salmon was
shared and discussed.Most review team members had been
associated with a juvenile monitoring program that had failed
to meet its objectives.The high degree of difficulty associ-
ated with effectively sampling a juvenile population in large,
complex systems where natural variations in time and space can
be extremely large was stressed.The team agreed that at
least four or five years of pre-project juvenile data would be
needed to evaluate the relationship between control and
treatment populations.
Another concern was that even if a post-project change in the
SMI was statistically identified,could the change be attrib-
uted to a dam-caused effect.For example,if juvenile chinook
salmon outmigration timing is influenced by flow,a post-
project decrease in the annual CPUE index for juvenile chinook
may not mean that their freshwater survival decreased.Under
post-project regul ated flows,a hi gher percent of reari ng
chinook may move out of the middle river prior to break-up and
consequently would be missed by the sampling program.
-32-
-
.....
The team outlined two possible courses of action to address
concerns about long-term fisheries monitoring.These are
stated as follows:1)The Department and APA should agree as
part of the settlement process that any change in fisheries
parameters beyond established limits would be treated for
mitigation purposes as a dam-caused impact and 2)a change
in population indicators would serve as a signal that a
problem may exist and detailed habitat assessment studies
would be required to identify the cause.An impact mitigation
policy that placed the burden of proof on the dam was favored.
If a negative change is detected beyond negotiated limits,it
would be assumed that the dam was responsible.The mitigation
plan should also require that studies be initiated to identify
the cause of the impact and to determine what specific
mitigation actions are appropriate.
The use of CPUE data rather than juvenile population estimates
was addressed by the team.Given the difficulty and expense
associated with capturing and marking a sufficient number of
juveniles to produce a reliable population estimate (approxi-
mately 250,000-500,000 marked chum would be needed for an es-
timate at Sunshine Station),the team suggested that a sample
design employing mobile juvenile inclined plane traps be de-
veloped to address some of the obvious sources of temporal and
spatial variance.In addition,the team felt that future
juvenile monitoring work should concentrate on chum and
-33-
-
•
chi nook sa lmon.Most pi nk salmon outmi grate duri ng break-up
and before it is feasible to deploy the traps.Sockeye juve-
niles in the control and treatment reaches are not similar.
Populations in the middle river spawn in side channel habitat
and migrate downstream soon after emergence where as most
sockeye outmigrating past the control station from the
Talkeetna River are produced and rear in nursery lakes.The
team did not recommend using juvenile coho salmon as a target
species for monitoring environmental change.The highly
variable nature of juvenile coho distribution would limit the
probability of successfully detecting a dam-caused change.
The value of monitoring the total production capability of the
Susitna River was discussed by the review team.It was sug-
gested that continued monitoring of total salmon escapement
into the River was important for stock management and also to
provide a useful comparison of middle river salmon populations
with the total production of the Susitna.The team felt that
a comparison of this type might put the magnitude of dam-
caused impacts on salmon resources into perspective.For
example,how significant would a 50%reduction in sockeye
salmon production in the middle river be without knowing that
middle river production accounts for less than five percent of
total Susitna sockeye.Conversely,the significance of a 50%
reduction in king salmon production in the middle river would
be evident if it represents 30%of the total production of the
-34-
,
I
r
-
-
f
Sus itna.Also t monitori ng tot a1 producti on of the Sus itna
would provide a very useful check on the long term health of
the river.The possibility of joint funding of total
enumeration studies on the Susitna by the Department and APA
was also discussed.
III.Resident Fish Long-Term Monitoring
A.Problem Statement Hydroelectric development may alter the
population structures of resident fish in the middle Susitna
River reach.Natural conditions are characterized by low t
clear winter flows and high turbid summer flows.With project
conditions will be characterized by high t relatively turbid
and warmer winter flows t delayed ice formation t and lower and
clearer summer flows.
B.Objective Assess the positive and negative with-project ef-
fects on resident fish populations in the Susitna River by
comparing pre-and post-project population abundance and
distribution.
C.Procedures A continuation of the basic tasks now in place is
proposed.The pre-project data will be compared to the
post-project data.Changes wi 11 be documented and reported.
The studies are geographically organized into pre-and
post-project comparisons 1)below Devil Canyon and 2)the im-
poundment area.
-35-
-
......
,....
....
1.Below Devil Canyon
(A)Pre-Project
(l)Continue to electrofish by boat at the 16 middle
river index sites to make pre-and post-project
comparisons in catch and CPUE.These sites are
composed of three major habitats and comparisons of
resident fish CPUE can also be made between
macrohabita ts.
(2)Continue to use secondary gear types such as gill
nets and hook and 1 ine to supplement boat
e 1ectrofi sh i ng da ta.A1so,cont i nue to record
fishwheel and outmigrant resident fish catches .
(3)Continue to collect biological data from resident
fish species to observe trends in age composition.
(4)Continue the tag/recapture program to generate popu-
1ation estimates and determi ne mi grationa 1 patterns
of selected resident fish species.Population
estimates woul d be made for adul t rai nbow trout,
Arct ic gray 1 i ng ,bu rbot,round wh itefi sh ,and
longnose sucker in the middle river.Population
estimates will also be made separately for rainbow
-36-
....
.....
.-
trout in Fourth of July Creek.Fish movement data
will be provided by analysis of tag recoveries.
Species that will be tagged are rainbow trout,Arctic
grayl i ng,burbot,Dolly Varden,round and humpback
whitefish,and longnose suckers .
(5)Every third year beginning in 1986,generate
microhabitat suitability criteria curves to
supplement microhabitat data gathered during
previ ous years.
(B)Post-Project
(1)Continue boat electrofishing at middle river index
sites and use of secondary gear types as described
in the pre-project task.
(2)Continue to collect biological data as described in
the pre-project task.
(3)Continue the tag/recapture program as described in
the pre-project task.
(4)The radio telemetry program should be re-instituted
for two years during construction and for at least
three years after construction to provide better
-37-
movement data on middle river rainbow trout,Arctic
grayling,and burbot.
(5)Microhabitat suitability criteria for adult middle
ri ver res i dent fi sh shou 1d be generated each yea r
during construction and for at least two years after
construction to determine how these fish have
adapted to with-project changes.
2.Impoundment Area
(A)Pre-Project
Existing data collected over the last five years would
serve as the pre-project data base.
(8)Post-Project
(1)Reinstitute the tag/recapture program during con-
struction and continue the program for at least two
years after construction to:
(a)generate population estimates for Arctic
grayling,and
-38-
r
~I
(b)determine movement patterns of selected fish
species in the eight major clear water
tributaries (i.e.,Fourth of July Creek,
Portage Creek).
(2)Continue to collect biological data from resident
fish to observe for trends in age structures.
(3)Radio tag Arctic grayling for at least two years to
provide better information than gathered by tag
recoveries on with-project Arctic grayling movement
behavior.
-39-
.....
D.Assumptions
1.Random mark-and-recapture effort.
2.Time between sampling does not affect recapture probabil-
ities.
3.The population is closed geographically.
4.Gear efficiency varies with the size of fish.
5.There is a random mixing of tagged with untagged fish.
6.Mortalities due to capture and tagging are insignificant.
7.There is little difference in behavior between tagged and
untagged fish.
8.The variability in sampling remains constant within and
between years .
.....E.Review Team Comments
The review team was provided with a summary of data collection
procedures,analysis,and results based on five years of
resident fish species studies.Based on available data,resi-
dent fi sh popul ations other than burbot primarily use the
-40-
-
"""
....
....
mainstem Susitna River as a seasonal migration route between
summer and winter rearing habitats;some use of the mainstem
for overwintering rainbow occurs near the mouths of tributary
streams.Populations of Arctic grayling and rainbow trout are
small and presently may be relatively stable.It appeared
that increased fishing pressure resulting from improved access
and/or increased awareness of fisherman about the area may
pose the largest threat to resident populations.
Due to difficulties in obtaining sufficient recaptures,past
population estimates have not been very precise.Given a
large variance associated with population estimates,the like-
lihood of detecting even major post-project changes in resi-
dent fish production appeared remote.It was also noted that
radio tagged fish have been tracked in areas being
electroshocked.When the generator was started,the radio
monitored fish left the area immediately.If this avoidance
of electroshocking occurs in the population in general,CPUE
data collected by this method may be meaningless.The review
team felt further mark-recapture studies for resident fish
species may be unwarranted considering how imprecise past
estimates have been,unless a feasible method of increasing
tag recoveries could be developed that would produce
population estimates with meaningful confidence intervals .
-41-
-
-
.....
j
The team suggested several possi b le approaches to increase
precision for resident fish populations without greatly
increasing project costs.Options included increasing
sampl ing effort by 1)reduction of species monitored to rain-
bow trout,Arctic grayling,and possibly burbot,2)reduction
of the number of fishing sites,and 3)a reduction of the
frequency of surveys from annua lly to every two or three
years.The team also recommended that one or two
representative middle river monitoring areas (e.g.,Fourth of
July Creek and Portage Creek)be selected and that sampl ing
effort should be concentrated on rainbow trout and Arctic
grayling.The possibility of selecting a resident fish
monitoring control area (e.g.,Prairie Creek on the Talkeetna
River)was also suggested.
Summary of Recommendations and Findings
1.The Department should develop a position on long-term monitor-
ing.It was recommended that Commercial Fisheries,Sport
Fisheries,and Game divisions should assist the Habitat
Division in developing a comprehensive departmental position
paper addressing Sus itnal ong-term monitoring in the near
future.The position should emphasize that a population based
monitoring program should be initiated in FY87 and continued
through construction and operation of a hydropower project on
-42-
-
,....
the Susitna.In additi on t regul atory monitori ng and enforce-
ment of mitigation measures (structual and operational require-
ments)as well as habitat utilization monitoring to assess the
reliability of impact projections would be needed during con-
struction and post-project.Details of what habitat moni-
toring might entail were not addressed.An interdivisional
effort will be needed to detail the scope and components of a
habitat/enforcemnt monitoring program in the near future so
that both habitat and population monitoring programs can be
i ntergrated.
2.As a whole,tag/recapture methodologies currently employed for
Susitna stocks provide reasonably precise estimates (tight
confidence limits)of salmon populations however accuracy of
some of the estimates were questioned due to violation of as-
sumpti ons used in developing Peterson estimates.Apparent
discrepancies in adult sockeye salmon population estimates
derived by hydroacoustic and tag/recapture methods may be ex-
plained and remedied by modification of the tag/recapture
sampling design.
3.Sampling effort could be optimized and possibly reduced with-
out a significant loss of precision in the estimate.This
must be determined by an analysis of past years data.
-4.A positive relationship between population estimates and cumu-
lative seasonal CPUE from fishwheels catches based on a very
-43-
._--------------._------------------
1imited number of data poi nts may exi st for some speci es at
some sampling locations;however,for other species this rela-
tionship can not be detected.Some of the assumptions
-
I"""
I
....
"""
5.
implicit in the use of CPUE as an index of population change
have not been tested and may not be val i d.For exampl e,
possible within season and between season variability in gear
efficiency is suspected.Insufficient data is available to
evaluate a CPUE index for juvenile outmigrant populations for
any species.No outmigrant data is available for Sunshine
Station.
Based on the last five years of baseline data collection,nat-
ural variations associated with population estimates of adult
and juvenile salmon populations utilizing the Susitna River
are 1arge and consequently detecti ng project induced effects
on these populations would be very difficult unless these
changes were of a large magnitude.The use of a control as a
means to account for natural variation was supported.
It may be assumed,based on observations from past dam
monitoring projects (the Libby Dam was specifically mentioned
by Dana Schmidt),that changes resulting from hydroprojects
were often large scale and detectable by conventional
fisheries science methodologies.More subtle change probably
can not be detected against the highly variable natural
-44-
.-
background of fish populations;state-of-the-art methods
simply may not be sensitive enough to detect small scale
impacts.
6.The use of juvenile chum and possibljL chinook salmon outmi-
grant CPUE data in addition to adult salmon population
estimates should be evaluated as long-term monitoring
parameters.The rel ationship of adult salmon enteri ng both
the treatment and control reaches and the subsequent
outmigration of juvenile salmon from these reaches should be
monitored over at least a four-year period to determine if the
assumpti on that treatment and control popul ati ons are
influenced by similar environmental processes is valid .
7.Because the freshwater survival index approach to long-term
monitoring is untested and a number of assumptions implicit in
its application are considered highly questionable,the
approach was not considered feasible.Alternative approaches
-!
'.
for long-term monitoring should be identified and considered.
Population parameters in addition to distribution and
abundance such as age structure,condition factors,freshwater
growth,and fecundity should continue to be collected during
the pre-project phase of monitoring so that trends in these
parameters for project treatment and control populations can
be compared.
-45-
......
f
/'I
I
8.The precision of population data for resident fish species
collected at established index sites should be improved to
increase the probabil ity that post-project changes in resident
populations could be detected.A number of alternative
methods for increasing precision were suggested and will be
evaluated.It was also recommended that the effect of sport
fishing on middle river resident fish populations be
evaluated.Expansion of Sport Fisheries Divisionis creel
and/or postal survey programs to include the growing middle
river sport fishery was suggested.
9.The revi ew team strongly recommend that the Susitna
Aquatic Studies Program obtain biometric support a~d approval
of all future work plans.
-
10.The team proposed the following priority (descending order)
for long-term'monitor"ing projects:1).adult salmon
monitoring,2)juvenile salmon monitoring,3)resident fish
monitoring,and 4)water quality monitoring.It was suggested
that the water quality monitoring program could be given to a
consulting firm or USGS.
-46-