HomeMy WebLinkAboutBradley River Salmon Escapement 1987BRA
081
FINAL REPORT
BRADLEY RIVER SALMON ESCAPEMENT
MONITORING STUDY
1987
Prepared for
BECHTEL CIVIL & MINERALS, INC.
AND
ALASKA POWER AUTHORITY
Alaska Energy Authority
LIBRARY COPY
Dames&Moore
FINAL REPORT
BRADLEY RIVER SALMON ESCAPEMENT
MONITORING STUDY
1987
Prepared for
BECHTEL CIVIL & MINERALS, INC.
AND
ALASKA POWER AUTHORITY
DECEMBER 1987
Dames&Moore
Alaska Power Authority
Bradley River Salmon Escapement
Monitoring Study
1987
BRADLEY LAKE HYDROELECTRIC PROJECT
Federal Energy Regulatory Commission
Project No. P-8221-000
Prepared by
Dames and Moore
Anchorage, Alaska
December 1987
FINAL REPORT
BRADLEY RIVER SALMON ESCAPMENT
MONITORING STUDY
1987
Prepared for:
BECHTEL CIVIL & MINERAL, INC.
AND
ALASKA POWER AUTHORITY
by
David Erikson
Mike McDowell
Anne DeToni
Mary Pearsall
DAMES & MOORE
INTRODUCTION
The Federal Energy Regulatory Commission (FERC) license granted to
the Alaska Power Authority for the Bradley Lake Hydroelectric Project
(Project No. 8221-000) stipulates that a plan be developed and imple-
mented to monitor the abundance of salmon in the Bradley River. A Salmon
Monitoring Plan was submitted to FERC in June of 1986 <Alaska Power
Authority 1986a). The intent of this monitoring plan is to provide a
yearly index of salmon abundance both during the pre-operational and
post-operational periods to allow an appraisal of project impacts to the
salmon resources of the Bradley River. This report summarizes the results
of the second year of the pre-operational studies of adult escapement to
the Bradley River per the proposed scope of work described in the Salmon
Monitoring Plan.
The salmon resources of the Bradley River have been documented in
considerable detail through a series of studies (USFWS 1982, Woodard-
Clyde Consultants 1983, 1984, Northern Technical Services 1985). The
results of these early studies indicated that the pink salmon (Oncorhyn-
chus gorbuscha was the principal salmon species utilizing the river for
spawning although smaller numbers of chum (~ keta), chinook (~
tshawytscha), and coho salmon (~ kisutch) may also spawn in the river.
The 1985 study by Northern Technical Services represented the first
year of study for this ongoing salmon monitoring program. Comparable
data were collected only during the 1986 and 1987 field seasons in
studies conducted by Dames & Moore. Thus, these are the only data pre-
sented and discussed in detail in this report. The second year of this
multi-year salmon monitoring study in 1986 confirmed that pink salmon
were the major spawners and that the river also supports small runs of
the other salmon species (APA 1986). Considerable data on the abun-
dance, distribution of spawning areas and the timing of the runs were
collected during 1986 and 1987.
The habitat available for salmon to spawn is restricted to a short
segment of river between an impassable waterfall upstream and the tide
flats downstream.
-1-
The primary objectives of the 1987 field effort were similar to
those of the 1986 study with some minor modifications to improve and
streamline the effort. The objectives were to:
0
0
Duplicate sample methods used during 1986 to compare abundance
between years.
Provide an index of abundance or estimate of the 1987 salmon
escapement to the Bradley River with emphasis on pink salmon.
Study Area
The study area was the same as 1986 and consisted of a 2011 m (6600
feet) stream segment extending from the downstream end of Riffle Reach to
the upstream end of Bear Island Slough (Figure 1). This study area
encompassed most of the known pink salmon spawning areas and was above
the primary zone of tidal influence, although Riffle Reach is affected at
most high tides. Extreme high tides influence water levels as far
upstream as the lower portions of Tree Bar Reach.
METHODS
Study Duration
The study was conducted over a nine week period from July 14 through
September 10, 1987 similar to the period covered in 1986. This sampling
period was originally selected to coincide with the period of the pink
salmon run based on the results of the earlier studies.
The field crew traveled to the site on Monday of each calendar week
and intensive sampling took place on every Tuesday, Wednesday, and
Thursday.
Hoop Net Sampling
Hoop nets were used as the primary sampling technique similar to the
1986 field effort. This technique was originally selected to standardize
catch effort which would allow comparisons between years based on a
catch-per-unit-effort (CPUE) index. Hoop nets proved to be very success-
ful for capturing adult salmon during the 1986 field season.
-2-
LONG
SLOUGH
ft\~1 STUDY AREA
3.0 etc. = RIVER MILES
0 feet 500
SCALE
Job No. 12023-028
FISH STUDY
N
LOWER BRADLEY RIVER WITH
SALMON ESCAPEMENT STUDY AREA
Dames & Moore
Figure 1
The hoop nets used for the 1987 study were the same ones used for
the 1986 study (Figure 2). These nets were made from 2. 5 inch stretch
mesh nylon with 6-foot diameter hoops. Net wings were attached to the
main frame of the net in various configurations depending on the location
of net in the river.
The hoop net sites established in 1986 were reoccupied in most
cases. However, significant changes in the course of the river had
occurred as a result of major flooding in October and December of 1986
which made it necessary to relocate two nets. Hoop net Site 5 had filled
in with gravel to the extent that it was too shallow to deploy the net at
the same site as last year and a new location for net 5 had to be
established approximately 50 m downstream. Hoop net Site 1 had to be
moved approximately 10 m downstream due to changes in the river channel.
Some minor reorientations were necessary for hoop nets 2 and 3 because of
changes in river current patterns during fall 1986. Figure 3 illustrates
the hoop nets as deployed during the 1987 field season.
During each weekly sampling period, the hoop nets were set Tuesday
morning and fished until Thursday morning for a total of 48 hours after
which they were removed from the water until the following week. Low
water levels prevented the setting of hoop net 6 in Bear Island Slough
during week eight. All other sites were fished during each weekly
sampling period.
During typical operations, each net was checked every four hours
during the daytime and then left to fish overnight. The fish were removed
at each check, identified to species, measured, and tagged using sequen-
tially numbered Flay spaghetti tags. Each salmon species was tagged
using a different colored Flay tag in contrast to last year when only
pink salmon were tagged. Since sockeye salmon (Oncorhynchus nerka) only
occassionally wander into the Bradley River, they were marked by
splitting the left pectoral fin. Sex and spawning condition were also
recorded for all salmon captured. Spawning condition codes were further
refined from the 1986 season to allow closer monitoring of the timing of
spawning.
-4-
(...
0
0'
z
0
.....
N
Q
N
(o) • Q
N
Q:l
.,
cS' c ...
tO
N
i'
~
SID
1:
l
2 112• STRETCH MESH NETTING THROUGHOUT
FLOATS
12• THROAT
LEAD WEIGHTS
6' SQUARE FRAME OPENING
1-c 16' --------1 .......... 1
BRADLEY RIVER HOOP NET DESIGN
.,.. z-----
KEY:
~ Hoop Nets
~ Seine Site
0 300
Scale in Feet
Job No. 12023-028
1987 HOOP NET SITES AND
BEACH SEINE LOCATIONS
Dames & Moore
Figure 3
Beach Seine Sampling
In 1987 beach seining was used primarily as a technique to capture
and tag adult salmon in addition to those caught in the hoop nets, and
secondarily as an index of abundance. Only one of the seine sites
established in 1986, Seine Site C, was unaffected by the severe flooding
during the fall storms. Site B in Tree Bar Reach could not be used this
year and another site had to be established approximately 50 m downstream
(Site D) (Figure 3). Seine Site A was used in 1987 but large boulders
deposited along this area from fall 1986 flood waters made it difficult
to make consistent sets.
Several other sites were seined during the study period but not on a
regular basis. An exposed gravel bar in Riffle Reach (Seine Site E)
provided an opportunity to seine in this normally difficult area during
week eight. An attempt was also made to seine in the lower portions of
the Bradley River, near Fox Farm Creek, to try and intercept tagged fish
leaving the river prior to spawning.
The beach seine utilized was similar in length (120 feet) to one
used in 1986, but the mesh size was increaced to 2. 5 inch to minimize
drag. The seine was deployed in the same manner as last year by tying
one end to the shore and feeding the remainder of the net from the bow of
the boat as it traveled in reverse making a broad sweep down the river
and back to the beach. If fish were captured, the effort was repeated up
to three times. Captured fish were processed in a manner similar to the
hoop net catch.
Electro shocker
Electroshocking was primarily used as a means of delineating areas
used by spawning fish and secondarily for capturing fish for mark and
recapture. A Coffelt backpack shocker was used for the first five weeks
until problems developed with this unit. A Smith-Root Type XI backpack
electroshocker was used both from the boat and on foot in shallow riffle
areas during the last several weeks of the study period.
-7-
Electroshocking was conducted at regular intervals throughout the
first and last three weekly sampling periods but three consecutive weeks
were missed during the peak of the run because of equipment problems.
Carcass Counts
No special effort was made this year to census carcasses since
efforts last season to find carcasses yielded very few observations even
though several thousand fish spawned in the river. All carcasses were
recorded when encountered and checked for tags.
Miscellaneous Observations
Since several pink salmon tags were found at the end of last season
in a small clear water tributary to the Bradley River, Fox Farm Creek, it
was decided to monitor this stream for salmon that had been tagged
farther upstream in the Bradley River study area. Periodic visits were
made each week to census Fox Farm Creek throughout the study period.
The presence of harbor seals and other biological observations were
noted.
RESULTS
Overall Catch
All species of North American Pacific salmon were captured in the
Bradley River during the 1987 sampling season. As was the case during
the 1986 study, pink salmon were again the most abundant species (Table
1>. Relatively few churn salmon were captured this year but were still
second in overall abundance followed by coho salmon and chinook salmon.
The total catch for sockeye salmon was only six, with two being recap-
tures. The percentage of recaptures was high for pink, chum, and chinook
salmon. This tends to indicate that the sampling methods applied were
effective in capturing a significant proportion of the spawning salmon
present. The percent recapture for coho salmon was significantly lower
than the other species which may suggest only a small portion of the
population was being captured. It is also apparent that only the early
part of the coho run was encountered during the study season, and the
-8-
Sample Week
1
July 14-16
2
July 21-23
3
July 28-30
4
Aug. 4-6
5
Aug. 11-13
6
Aug. 18-20
7
Aug. 25-27
8
Sept. 1-3
9
Sept. 8-10
Study Total
TABLE 1
TOTAL ADULT SALMON CATCH FOR ALL SAMPLE METHODS COMBINED
(Recapture Percentage in Parentheses)
Pink Churn Chinook Coho Sockeye
Salmon Salmon Salmon Salmon Salmon
I I I
3(0) 2(0) I 2(0) I 0(0) 0(0)
14(0) 6 (17) 1( 0) 0(0) 0(0)
36(6) 18 ( 11) 8(38) 0(0) 0(0)
I I I
83(16) 11( 27) 7 ( 71) 0(0) 2(0)
220(19) 2(50) 6(67) 3(0) 0(0)
261(26) 1( 0) 6(16) 11 ( 0) 1( 0)
174(27) 0(0) 5 (100) 12(0) 2 (100)
118(31) 5(20) 0(0) 1( 100) 1( 0)
35(37) 4(0) 1( 0) 17(0) 0(0)
944(20) 49 (17) 36(37) 44(2) 6(33)
Total
Salmon
I
I 7(0) I
21(5)
62<11> I
I
103 (19)
231( 20)
280(25)
193(27)
125(30)
57(23)
1079(23)
relatively low numbers of fish may be affecting recapture rates for this
species.
The only other species caught in the river this year were slimy
sculpins (Cottus cognatus) and Dolly Varden (Salvelinus malma). The
Dolly Varden utilizing the river ranged from very small juveniles to
mature adults in spawning condition. Most of the Dolly Varden escaped
through the mesh of the hoop nets and the seine making accurate counts
impossible.
Hoop Net Sampling
Hoop nets again proved to be a good standardized method of sampling
salmon in the highly turbid waters of the Bradley River. Water levels in
the river remained relatively constant throughout the study period and
allowed uniform sampling for most of the nets. One problem did occur
during the third week when a beaver repeatedly entered the nets and
chewed holes in the cod end. Fortunately, the beaver left the area by
the next week and no further problems occurred.
Pink Salmon
The hoop net catch data for pink salmon for each hoop net site by
week is given in Table 2. These data indicate the pink run started
rather slowly as only small numbers entered the river during the first
three weeks. The run in 1987 reached its peak in week six (August 18 to
20). The total catch declined substantially by week seven. By weeks
eight and nine, a large percentage (41 percent and 53 percent, respec-
tively) of the fish caught in the hoop nets were spawned out. The 1986
and 1987 pink salmon runs followed generally the same pattern based on
total catch data from the hoop nets. The main difference was the 1986
run appears to have reached its peak in the seventh week of that year's
sampling program. However, these results may have been influenced by
extremely high flows during the seventh week in 1986. It is suspected
that fish sought refuge in the lower velocity water near shore during the
flooding and therefore became more susceptible to capture. This sup-
position is born out by the almost doubled CPUE during week seven when
-10-
compared to other weeks when flows were at normal levels (see Table 2 and
Figure 4).
The CPUE and the progression of the reproductive condition of the
fish supported the conclusion that the peak occurred the third week in
August (see Figure 4).
Chum Salmon
Only a few chum salmon were present in the river during the first
week of sampling (Table 2). Chums increased only slightly with a total
catch of 18 during the third week. The catch declined significantly
after the fourth week (August 4-6). In early September, a few late chums
came into the river but it was not determined if these individuals
remained in the river to spawn. The relatively bright condition of
these fish indicated they were not part of the earlier run observed in
July. Comparison of 1986 and 1987 CPUE for chum salmon is presented in
Figure 5.
Of the 37 individual chums caught and tagged in the hoop nets, only
five were in the process of spawning and one was spawned-out. This
suggests that very little spawning activity occurred in the river this
year.
Chinook Salmon
The hoop net catch data for chinook salmon is summarized in Table 2
and compared with 1986 data. These data indicate a small number of
chinooks were in the river throughout the study period. No actual peak
was evident in the hoop net catch but based on the progression of the
reproductive condition of the fish it appeared that the peak of spawning
occurred in mid-August. However, the first two chinooks caught in mid-
July were already in advanced reproductive condition. Half of the total
catch of chinooks were recaptures which indicate that a large proportion
of the population was being sampled. Comparison of 1986 and 1987 CPUE
for chinook salmon is presented in Figure 6.
-11-
TABLE 2
BRADLEY LAKE SALMON MONITORING STUDY
1986 AND 1987 TOTAL CATCH (TC) AND
CATCH-PER-UNIT-EFFORT (CPUE) SUMMARY DATA -ALL SPECIES
CPUE DATA -1987
SAMPLE DOLLY
WEEK PINK CHUM COHO SOCKEYE CHINOOK VARDEN
TC CPUE TC CPUE TC CPUE TC CPUE TC CPUE TC CPUE
1 3 0.01 2 0.006 0 o.oo 0 0.000 2 0.006 0 0.000
2 14 0.04 6 0.02 0 o.oo 0 o.ooo 1 0.003 0 0.000
3 36 0.11 18 0.05 0 o.oo 0 0.000 8 0.024 1 0.003
4 54 0.16 11 0.03 0 0.00 2 0.006 7 0.021 3 0.009
5 205 0.62 2 0.006 3 0.007 0 0.000 6 0.018 5 0.015
6 247 0.75 1 0.003 11 0.03 1 0.003 6 0.018 1 0.003
7 154 0.46 0 o.oo 12 0.04 2 0.006 5 0.015 0 0.000
8 74 0.26 5 0.02 2 0.01 1 0.003 0 0.000 1 0.003
9 31 0.09 1 0.003 17 0.05 0 0.000 0 0.000 4 0.012
1986 DATA
SAMPLE DOLLY
WEEK PINK CHUM COHO SOCKEYE CHINOOK VARDEN
TC CPUE TC CPUE TC CPUE TC CPUE TC CPUE TC CPUE
1 5 0.03 21 0.14 0 0.00 3 0.02
2 18 0.07 57 0.22 0 0.00 7 0.03
3 95 0.31 131 0.42 0 0.00 10 0.03
4 152 0.46 95 0.29 0 0.00 13 0.04
5 143 0.41 67 0.19 14 0.04 2 0.01
6 190 0.81 19 0.08 16 0.07 11 0.05
7 236 1.51 0 0.00 1 0.01 0 0.00
8 145 0.44 1 0.005 56 0.17 0 0.00
9 63 0.18 1 0.005 38 0.11 2 0.01
;
r
!::
) -..,
::> .., .,
I ::>
" D
I
AD
flf
: Cl
1.60
1.50
1.40
1.30
1.20
1.10
1.00
0.90 w
:l 0.80 a..
0
0.70
0.60
0.50
0.40
0.30
0.20
o. 10
0.00
BRADLEY LAKE SALMON ESCAPEMENT STUDY
PINK SALMON -CATCH PER UNIT EFFORT
, 2 3 4 5 6 7 8 9
WEEKS
rs:~ 1987 [12'/1 1986
c..
0
C"
z p ...
N
0
~ I
0
~
CD
I
i
i
lit go
!U
L1l
:J
Cl.
()
0.45
0.40
0.35
0.30
0.25
0.20
0.15
0.10
0.05
0.00
BRADLEY LAKE SALMON ESCAPEMENT STUDY
CHUM SALMON -CATCH PER UNIT EFFORT
1 2 3 4 5 6 7 8 9
WEEKS rs::mJ 1 9 s 7 WJ 1986
c...
0
C'
z p
~
0
~
N
o:P
i'
I •
!If
"" :J
f;
o., 0
0.09
0.08
0.07
0.06
0.05
0.04
0.03
0.02
0.01
0.00
BRADLEY LAKE SALMON ESCAPEMENT STUDY
CHINOOK SALMON -CATCH PER UNIT EFFORT
' , 2 3 4 5 6 7 a 9
WEEKS
J:S] 1987 [22":J 1986
Of the 17 individual fish tagged over the course of the study, over
half (9) were small precocious males (jacks). These small males
accounted for most of the recaptures with only two adults being captured
more than once. There was no clear pattern of distribution in the river
this year. Several chinooks were present in Bear Island Slough during
the third week, but they moved out and were later recaptured further
downstream.
Coho Salmon
The catch data for coho salmon (Table 2) indicate this species moved
into the river during the second week of August. The CPUE steadily
increased throughout August but fell off the first week in September.
Fish began entering the river again the last week of the study period
when the largest catch occurred (see Figure 7>. The recapture rate for
cohos was very low. Only one of the 33 fish tagged was captured a second
time. The progression of reproductive condition was not evident by the
last week when 66 percent of the fish were still fresh. No spawned-out
cohos were observed by the end of the study, but three fish were in the
process of spawning when captured.
Since fish were still entering the river at the end of the study
period, little can be said from these data about total abundance or
timing of the complete run for this species in the Bradley River.
Sockeye Salmon
Five sockeye were caught in the hoop nets during the study period.
It is believed that these were strays from nearby systems.
individual was captured more than once.
Beach Seining
Only one
Beach seining proved to be unproductive in 1987 largely because of
alterations in channel configuration at the seine sites resulting from
the 1986 fall floods. A total of 77 pink salmon (21 percent recaptures)
were caught using this method. The recapture rate was reasonably close
to the overall 24 percent attained with hoop nets. The most productive
-16-
c..
0
0"
z
? ... ..,
0
~ ..,
C»
I
Hi
LLI
::J
~
0.17
0.16
0.15
0.14
o., 3
0.12
0.1,
0.10
0.09
0.08
0.07
0.06
0.05
0.04
0.03
0.02
0.01
0.00
BRADLEY LAKE SALMON ESCAPEMENT STUDY
COHO SALMON -CATCH PER UNIT EFFORT
, 2 3 4 5 6 7 8 9
WEEKS
[S~] 1987 WA 1986
site was seine Site D, located in Tree Bar Reach, which was established
during the fourth week after not being able to catch any fish at the per-
manent sites established last year. Site D yielded 86 percent of all
fish captured by beach seining. Difficulty in making standardized hauls
with the seine and the low numbers of fish in areas where seining was
possible made this method impractical as a way of determining abundance.
However, beach seining did provide an additional means of tag and recap-
ture.
No chum, chinook, or coho salmon were captured in the beach seine
hauls at the permanent sites. One coho was caught in an experimental set
near an exposed gravel bar in Riffle Reach, which was accessible only
during periods of low water.
Electroshocking
Electroshocking proved to be a useful method of delineating the
spawning areas and as a capture technique for some of the shallow riffle
areas. Low conductivity in the Bradley River made it difficult to stun
the fish sufficiently to catch them. However, many of the fish do come
to the surface where it is often possible to see if tagged fish are pre-
sent.
Problems with the shocker prevented data acquisition for three weeks
during the middle of the pink salmon run. Results of the electroshocking
indicate most of the pink salmon spawning activity probably occurred in
the Tree Bar Reach area. Considerably fewer pinks were found in the
lower spawning grounds in Riffle Reach. Only two chinooks and one chum
were located using the shocker and all of these were in the upper portion
of the river (Tree Bar Reach). Spawning distributions for the various
salmon species are delineated in Figure 8.
Carcass Counts
A total of only four pink carcasses (one tagged), one chinook (with-
out tag), and one coho <without tag) were found incidental to the weekly
sampling program. This low number of carcasses is not unusual since
-18-
.. z·----
KEY:
Spawning Areas for-
(ji$1) Pink Salmon
(/jjj!!J Pink and Chum Salmon
k-King Salmon
0 300
Scale in Feet
Job No. 12023-028
DISTRIBUTION OF
SPAWNING SALMON
Dames & Moore
Figure 8
spawned-out fish were regularly captured in the hoop nets, especially in
the latter weeks of the study period. A possible explanation is that
dead and dying fish are regularly flushed by the currents into the inter-
tidal zone and then are carried out with the daily tides. The highly
turbid waters of the river also made detection of carcasses difficult.
Tag Retention
Four individual fish caught in the hoop nets were found to have shed
their tags. Considering the number of fish tagged and recaptured, this
was considered acceptable. Special care was taken during 1987 to insert
the tags through the dorsal fin rays to help insure tag retention. It is
believed that this procedure minimized tag loss.
Stream Life
Pink Salmon
Of the 194 recaptures over the course of the study period, 90 per-
cent had been tagged during the prior two week period. Only eight indi-
vidual fish held longer than two weeks. Six held for three weeks and two
remained in the river over four weeks. The two longest river residents
were tagged on July 18 and were recaptured alive on August 25 and 26.
Chum Salmon
Only nine tagged chum salmon were recaptured and four of these were
recaptured in the same week they were tagged. The remaining chums were
caught within two weeks of their initial tagging.
Chinook Salmon
Of ten chinook salmon captured more than once, eight were recaptured
within two weeks of being tagged. One jack was tagged the first week of
the study on July 15 and recaught 27 days later on August 11. Another
jack lasted at least 21 days from August 4 to August 25.
-20-
Coho Salmon
Recaptures were insufficient to provide any insight into stream life
of cohos in the Bradley River. The only recapture was a small precocious
male tagged on August 19 and recaptured 20 days later on September 8.
Mark and Recapture Population Estimate
Pink Salmon
The methods used for estimating populations were identical in 1986
and 1987, except for use of a modified version of the Peterson formula.
Population estimates were calculated for pink salmon in the river during
each week. This method is based on a Peterson model as modified by
Chapman (Ricker 1975). The formula is presented in Table 3. This for-
mula reduces the potential bias in mark-recapture population estimates,
particularly when the number of recaptures is relatively small. The 1986
data were also recalculated using this formula and are presented in Table
3 with data from the 1987 study effort. These estimates are based on the
following assumptions:
1. Pink salmon numbers remained stable during the three-day sample
period.
2. All fish marked during the previous two sample weeks were still
present in the study area.
3. Fish marked three or more weeks prior to the sample week were no
longer present in the study area.
4. Marked and unmarked fish were equally susceptible to capture.
As evident from these assumptions,
mates should be treated with caution.
the resultant population esti-
It is likely that assumptions
regarding immigration and mortality/emigration that form the basis for
mark-recapture population estimates are violated by andromous salmonids
during a spawning run. However, the approach and sampling methods used
in these studies were specifically designed to minimize these potential
sources of bias. The nature of the spawning run is likely to cause an
-21-
underestimate of the population during the early part of the run when
immigration rates are greatest. Conversely, later in the run when mor-
tality rates increase, these methods are likely to overestimate the
actual population.
Weekly population estimates for pink salmon from both 1986 and 1987
are presented in Table 3. These estimates would suggest that the 1987
peak of the pink salmon run (3083) occurred during the seventh week of the
study.
week.
This followed the highest total catch in the hoop nets by one
The statistical reliability of the mark-recapture estimates
increases with the number of recaptures and the proportion of marked fish
in the population. Thus the best estimates are those for weeks five
through nine. The weeks prior to week five had very low recapture rates
and lower proportions of marked fish. This increases the potential for
bias in these estimates in comparison to those later in the study period.
These factors result in wider confidence intervals than in the later
weeks.
Most of the catches from beach seining and electroshocking included
insufficient recaptures to provide reliable, independent verification of
the weekly estimates. However, these alternate methods did result in
adequate recaptures in weeks seven and eight to provide at least some
comparison with the hoop net data. These estimates are within the
ballpark of estimates generated from the hoop net data (see Table 3)
while displaying wider confidence intervals as would be expected by the
lower recapture rates.
Week 7
Week 8
Total
Captured
20
44
Total
Marked
364
306
Recaptures
2
11
1 95 percent confidence interval.
-22-
Population
Estimate
2555 (935-6388)1
1151 (667-2159)
TABLE 3
BRADLEY LAKE SALMON MONITORING STUDY
MARK -RECAPTURE POPULATION ESTIMATE CALCULATIONS
1987 DATA
SAMPLE
WEEK MARK CATCH RECAP. -95% R +95% R N -95% N +95% N
1 0 3 0
2 3 14 0
3 17 36 2 7.2 0.2 222 81 555
4 46 54 2 7.2 0.2 862 315 2154
5 103 205 14 23.5 7.7 1428 874 2463
6 243 247 27 39.2 17.8 2161 1505 3219
7 357 154 17 27.2 9.9 3083 1968 5091
8 300 74 16 26.0 9.2 1328 836 2213
9 194 31 5 11.7 1.6 1040 491 2400
1986 DATA
SAMPLE
WEEK MARK CATCH RECAP. -95% R +95% R N -95% N +95% N
1 0 4 0
2 4 36 1 5.6 0.1 93 28 168
3 31 158 8 15.8 3.4 565 303 1156
4 156 259 29 41.6 19.4 1361 958 2001
5 308 270 51 65.9 37.0 1610 1252 2204
6 381 262 43 57.9 31.1 2283 1706 3130
7 426 236 36 49.8 25.1 2735 1992 3877
8 397 156 33 46.3 22.7 1838 1321 2637
9 277 68 14 23.5 7.7 1279 783 2205
Modified Peterson Formula:
Estimated Population = (Total Marked +1) X (Total Catch +1)
(Recaptures +1)
An approach consistent with that used by the Alaska Department of
Fish and Game (Pirtle and McCurdy 1980) can be used to derive a rough
estimate of the total of pink salmon escapement in the Bradley River.
Using this method weekly estimates are combined and then divided by 2.5
which would give a total escapement of 4049 for 1987. Using the 1986
data and applying the estimates derived from the Chapman modification of
the Peterson formula, provides an escapement estimate of 4706. Based on
these estimates, the 1987 pink salmon escapement may have been approxi-
mately 14 percent less than the 1986 run.
Chum Salmon
The number of recaptured chums was considered too low to provide a
meaningful estimate of the spawning population. The highest number of
recaptures of fish tagged two weeks prior to the sample week was two in
the fourth week. These data indicate there were fewer than 100 chums
in the river during 1987.
Chinook Salmon
Catch data for chinook salmon were not sufficient to calculate a
meaningful population estimate for this species. The total catches for
chinooks were low and the recapture rate was high indicating a relatively
small population. Numbers of chinooks in the river year was probably
around 50.
Coho Salmon
Little can be said about the coho population since the recapture
rate was very low and they were still coming in to the river at the end
of the study period.
-24-
Miscellaneous Observations
Fox Farm Creek
Regular census of this clear water tributary of the lower Bradley
River showed no fish in the stream until August 5 when four pink and
eight chum salmon were observed spawning in the shallow waters. One of
the pinks and one of the chums were tagged with color-coded tags. On
August 11 only three pinks and two chums were seen in the stream and none
had tags. Though other live fish were observed here for the remainder of
the study period, one pink carcass (no tag) was found on August 18.
These observations suggest that a small portion of tagged salmon may
leave the upper spawning areas after being tagged. The number of tags
recovered during 1987 was significantly less than the 15 found during
1986.
The only tag recovered from outside the Bradley River system was a
pink salmon tag found in a Bald Eagle nest near the temporary airstrip by
the Martin River delta on September 18 (Roseaneu, 1987, Alaska Research
Associates, Inc., personal communication). This fish had last been cap-
tured on August 19 in hoop net 1.
Harbor Seals
Harbor seals regularly swam up the river to feed on spawning salmon.
Seals were seen as far upstream as the lower portion of Tree Bar Reach,
but usually fed in Riffle Reach. On two occasions, seals became trapped
in the hoop nets but were released unharmed at the next check of the
nets.
-25-
DISCUSSION
Comparison with 1986 Season
Abundance Indices
The results from the 1987 season indicated a somewhat smaller pink
salmon run than occurred in 1986 from the CPUE statistics derived from
the hoop net data as well as population and escapement estimates
(Figure 9). Total sampling effort was less in 1986 because of time
involved in finding locations for the nets during the first couple weeks
and because of the high water which prevented the deployment of several
nets during week seven. The overall CPUE for pink salmon in 1987 was
down 40 percent from the 1986 season. However, this may overstate the
actual difference in the populations between years because of events
during the 1986 study season. Flooding during week seven in 1986 may
have caused fish to seek refuge along the shore as a result of high river
velocities, thus increasing their susceptibility to capture in the hoop
nets. Thus the CPUE for week seven was almost double the rate of the
next highest sample week in either year. If a value of 0.75 is used in
calculating the average CPUE instead of 1. 51, a smaller difference is
indicated between the two years. (This is probably more indicative of
normal fishing conditions close to the peak of a run based on week six
during both study years.) Using 0.75 yields a reduction of 26 versus 40
percent drop in CPUE from 1986 to 1987. This lower value is also closer
to the magnitude of reductions indicated in the population and escapement
estimates calculated separately.
The peak of the pink salmon run occurred in week six during 1986 and
1987. However, fish came into the river later, built up more quickly and
to a slightly higher peak in 1987, as illustrated by Figure 9. Pink
salmon runs throughout Kachemak Bay and lower Cook Inlet were late, and
overall numbers were lower than expected this year <Dudiak, 1987, Alaska
Department of Fish and Game, personal communication).
The CPUE statistics for chum salmon indicated chums were down
approximately 90 percent from 1986. Total catch was much lower during
-26-
BRADLEY LAKE SALMON ESCAPEMENT STUDY
POPULATION BY WEEK -PINK SALMON
6.0
5.0 0
z
0 4.0 ~~ l A
:J ., n.."O I 2 oc Q..g 3.0 o5
2.0 ~ /~\ b. ~t tJ
0 ___:/, ~ tJ t= A ~ I!]
A /
% +
1.0 -i ~ ~-+ +
+ . -0.0 .,
1 2 3 4 5 6 7 a 9 u
i WEEKS Ul
SID tJ +95" -1987 + -95% -1987 ¢ 1987 -POP. EST. E A +95% -1986 X -95% -1986 v 1986 -POP. EST. ~
1987 for chum salmon as well, indicating much lower population levels
than were encountered last year. The peak of the 1987 chum run occurred
at the same time as the 1986 season, the last week in July.
Mark and Recapture Estimates
The validity of mark and recapture techniques for spawning salmon
populations is discussed in detail in the 1986 Salmon Monitoring Report
(APA 1986b). The weekly population estimates for this year did not indi-
cate as great a decline in the pink run as the CPUE data. The calculated
estimates for this season are probaby higher than the actual number using
the river. This was attributed to lower percent of recaptures during
1987 through the peak of the run, in comparison to 1986. Since the model
is very sensitive to the number of recaptures, a lower percentage of
recaptures results in wider confidence limits for population estimates.
A proportionally larger number of tagged fish were available to be cap-
tured last season as a result of successful beach seining efforts.
Approximately 300 pinks were tagged from beach seine hauls in the first
six weeks in 1986 compared with only 47 in 1987. These additional tagged
fish provided a larger number of marked fish in the population and
resulted in a larger percent of recaptures and therefore greater con-
fidence in the 1986 estimates. Examination of Figure 9 indicates that
populations in both years were quite similar, with 1987 confidence inter-
vals generally encompassing those of 1986.
It is believed that population estimates will be useful in comparing
pre-operational and post-operational population levels in the Bradley
River. Given the shortcomings of the population estimates under the pre-
vailing conditions in the Bradley River, the methods to reduce sources of
bias and error appear to provide consistent results from year to year.
Thus population estimates, in conjunction with CPUE statistics, provide a
good measure of pre-operational pink salmon abundance for use as a bench-
mark for comparison with population levels after project operation begins.
-28-
Comparisons With Other Streams
The Alaska Department of Fish and Game monitors pink and chum salmon
escapement in five streams in the Kachemak Bay area. These include Humpy
Creek, the largest producer of pink salmon in the bay, China Poot Creek,
Barbara Creek, Seldovia River, and Port Graham River (Alaska Department of
Fish and Game 1986). Only Barbara Creek has a smaller run of pink
salmon than the Bradley River. These streams exhibit a cyclical pattern
in pink salmon escapement, with the highest escapement occurring during
the odd-numbered years. Therefore, the 1987 pink salmon run for local
systems was expected to be larger than 1986 returns. The actual escape-
ments for 1987 were significantly lower than expected and were much lower
than the 1986 escapement levels (Alaska Department of Fish and Game
1987). Humpy Creek suffered only a 47 percent decline from 1986 compared
with an average decline of over 75 percent for the other streams.
The approximate 14 percent decline in pink salmon in the Bradley
River was less than other spawning streams, but still exhibited the same
pattern of lower numbers in what normally would have been a peak year.
Factors contributing to the drop in pink salmon returns in other parts of
the bay also affected the Bradley River system.
Chum salmon are not a major species in the streams and rivers around
Kachemak Bay. Port Graham River supports the largest run which averages
approximately 2,000 spawners (Alaska Department of Fish and Games 1986).
In 1987, this river had an increase in escapement of 2.5 times the 1986
run, but was still below the long-term average. This was opposite the
trend in the Bradley River, which experienced an approximate 90 percent
drop in escapement. It is difficult to draw any conclusions regarding
these patterns since chums are not widely distributed and overall numbers
are relatively low.
-29-
RECOMMENDATIONS FOR FUTURE YEARS
1. Continue use of the hoop nets during the pre-operational moni-
toring phase as the most reliable means of indexing population
abundance. Hoop nets may not be useful for post-operational
monitoring due to the predicted reduction of flow.
2. Continue to use beach seining as a supplemental method for mark
and recapture estimates. Increased effort should be devoted in
the early weeks to maximize tagging prior to peak spawning.
3. Continue tagging all salmon species.
4. Continue to census Fox Farm Creek on a weekly basis to monitor
the presence of tagged fish.
S. Continue the use of the mark and recapture population estimates.
6. Maintain a three-person field crew.
7. Continue the study timing and duration established in 1986 and
1987 for species other than coho salmon and continue the three-
day intensive sampling period each calendar week. Lengthening
the study duration into October would aid in understanding the
timing and magnitude of the coho salmon run in the Bradley River.
8. Increase electroshocking efforts as an alternate method of mark
and recapture and as a method of determining distribution of
fish in the river.
9. A single hoop net should be set in the lower portions of the
river in the vicinity of Fox Farm Creek for approximately eight
hours per week during the peak of the run to intercept tagged
fish leaving the upper spawning area after being tagged. The
departure of tagged salmon from the study area could be an
important factor affecting the population estimates and needs to
be evaluated in future studies.
-30-
10. Equipment items and study logistics as employed in the 1987
field season were adequate to accomplish the stated tasks.
Since the hoop nets have been heavily used for two years it may
be necessary to replace some nets or sections of webbing and
purchase some new net wings.
-31-
REFERENCES
Alaska Department of Fish and Game. 1986. Lower Cook Inlet Fin Fish
Report. Commercial Fisheries Division. 440 pp.
Alaska Power Authority. 1986a. Salmon monitoring plan for Federal
Energy Regulatory Commission Project No. P-8221-000. Prepared by
Stone and Webster Engineering Corp., Anchorage, Alaska.
1986b. Bradley River Salmon Escapement Monitoring Report
Prepared under contract to Bechtel Civil & Minerals, Inc. for
Alaska Power Authority Bradley Lake Hydroelectric Project.
Anchorage, Alaska.
1986.
the
Dudiak, N. (Alaska Department of Fish and Game). 1987. Personal com-
munication with David Erikson, Dames & Moore.
Northern Technical Services, Inc. 1985. 1985 Salmon Escapement Survey
Report. Prepared for Alaska Power Authority, Anchorage, Alaska.
29 pp.
Pirtle, R.B., and M.L. McCurdy. 1980. Prince William Sound general
districts 1976 pink and chum salmon aerial and ground escapement
surveys and consequent brood year egg deposition and pre-emergent
fry index programs. Alaska Department of Fish and Game Technical
Data Report No. 51.
Roseneau, D. (LGL, Alaska Research Associates, Inc.). 1987. Personal
communication with Dave Erikson, Dames & Moore.
u.s. Fish and Wildlife Service. 1982. Appendix B: Bradley Lake
Hydroelectric Project. Homer, Alaska. Final Coordination Report.
USFWS Western Alaska Ecological Services. Anchorage, Alaska.
131 pp. In: U.S. Army Corps of Engineers. 1982. Bradley Lake
Hydroelectric Project, Alaska. Final Environmental Impact
Statement. Alaska District, u.s. Army Corps of Engineers.
Woodward-Clyde Consultants. 1983. Bradley River instream flow studies.
Prepared for Alaska Power Authority, Anchorage Alaska. 77 pp.
Woodward-Clyde Consultants. 1984. Report on salmon fry sampling in the
Bradley River. Prepared for Alaska Power Authority, Anchorage
Alaska. 6 pp.