Marine Salmon Interactions Program
Adult Pink Salmon Forecasting Shows Promise Using Early Marine Parameters in Southeast Alaska
Pink salmon are the most abundant salmon species in Alaska and have historically shown remarkable
interannual fluctuations in adult returns. From the 1900s to 2003, annual pink salmon harvests have
fluctuated between about 10 million and 150 million fish. Pink salmon are also a highly valued resource,
with exvessel commercial values ranging from $20 million to $40 million annually over the period from
1976 to 2003. Accurate forecasts of salmon returns are important because they enable surplus fish
production to be matched with fish harvesting and processing capabilities. This leads to a greater
economic return of the resource to the harvesters, while allowing wild salmon escapement goals to be
met and sufficient adult returns at hatcheries.
In Southeast Alaska, pink salmon production is mainly based on wild stocks, and record harvests
(more than 72 million in 1999) have occurred within the last decade. Understanding factors that
influence marine survival of pink salmon (and other salmon species) is a primary focus of ABL’s
Southeast Alaska Coastal Monitoring (SECM) project. Over the past 7 years, SECM has collected a
time series of data incorporating biophysical parameters and indexes of juvenile salmon abundance
in neritic habitats of the northern region of Southeast Alaska. This research has been conducted
annually at monthly intervals during May to August, from 1997 to 2003.
Annual indexing of juvenile pink salmon under SECM shows promise as a forecasting tool for adult
pink salmon harvests in Southeast Alaska. We examined the correlation of annual harvests with data
from the juvenile marine period and found a significant correlation (r=0.94, P<0.01) of juvenile pink
salmon abundance in July (based on trawl catch-per-unit-effort (CPUE)). We found no significant
association of harvest with measures of temperature, salinity, zooplankton standing crop, or pink salmon
growth. The relationship between the index of abundance of juvenile salmon in neritic habitats and the
corresponding year class of adult pink salmon supports the hypothesis that year-class strength of pink
salmon is established early on in the marine environment.
We used two methods to develop a general linear model for forecasting the 2004 return of pink salmon
from these data: 1) stepwise regression, considering all variables, and a decision criteria of P < 0.1
for a variable to enter or remain in the model; and 2) the Akaike Information Criterion, comparing models
with CPUE and growth parameters only. For both methods, the only independent variable in the model for
forecasting harvest was CPUE. We then used the 2003 juvenile CPUE data to forecast a harvest of 46 million
pink salmon in Southeast Alaska in 2004, with an 80% confidence range of 41 million to 52 million. The
Alaska Department of Fish and Game (ADF&G) forecast for 2004, based on an exponential smoothing model of
historical harvests and forecasts, was 50 million, with an 80% confidence range of 24 million to 76 million.
The 2004 harvests will test the use of juvenile salmon sampling as an index of pink salmon year-class
strength. In the context of linking marine biophysical factors to salmon production and survival, our
SECM data point to the importance of establishing annual time series data over a long enough period to
identify critical periods for a particular species (i.e., pink salmon juveniles May–July). One goal of
the SECM project is to identify these periods for all salmon species during their juvenile life history
phase. Understanding how resources are partitioned during these critical periods for an individual salmon
species between hatchery and wild stocks, and among ecologically related species, is key to answering
questions about the marine carrying capacity of salmon and the implications of climate change on these species.
By Joseph Orsi and Alex Wertheimer
Annual Auke Creek Cooperative Research and Planning Meeting
The Auke Lake system in Juneau supports endemic populations of pink, chum, sockeye, and coho salmon,
cutthroat and steelhead trout, and Dolly Varden, with continuous data collected on some populations
since the early 1960s. In 1983, an interagency cooperative agreement relating to the Auke Creek
fish-counting weir was established between NMFS, the University of Alaska Fairbanks (UAF), and ADF&G.
An interagency meeting of participating interagency personnel is held annually, and a report of fish
counts from daily weir operations, and other information related to salmonid research at Auke Creek
is prepared by the NMFS project leader. The Auke Creek weir annual report is available on the ABL
web site (http://www.afsc.noaa.gov/abl/).
The 2004 Auke Creek interagency meeting was held in March and was split into morning and afternoon
sessions. The morning session focused on biological research projects: summaries and operational plans
were presented on projects for 2004 at Auke Creek weir. There are currently 19 projects scheduled at
Auke Creek for the 2004 season. In general, NMFS will continue long-term data collections on all species,
with emphasis on the marine survival and freshwater production of pink, coho, and sockeye salmon. The
ADF&G will continue with their long-term research programs on Dolly Varden, cutthroat trout, and coho
salmon; UAF will resume the outbreeding depression study on pink salmon in 2004 and is actively
recruiting a graduate student for the project.
Of note was Leon Shaul’s report (ADF&G) on Southeast Alaska coho salmon index stocks. Marine survival
is decreasing in all areas except at Auke Creek; however, smolt production from fresh water is decreasing
at Auke Creek, which is one of the ADF&G index coho stocks. Rick Focht of the Macaulay Hatchery, Douglas
Island Pink and Chum Corporation, presented a summary of the 2003 return of chinook released in Auke Bay
near Auke Creek, and discussed plans for the release of chinook juveniles in Auke Bay in 2004. The afternoon
session discussed multi-agency cooperative agreements relating to work at the NMFS weir and experimental fish
hatchery at Auke Creek. The current cooperative agreement expires at the end of the 2005 field season; all
cooperators expressed a strong interest in continuing cooperative salmonid research at Auke Creek.
By Jerry Taylor
Auke Creek Fish Counting Weir Operations Begins for 2004
The key to salmonid research projects at Auke Creek is the fish counting weir. The weir is a permanent
structure capable of capturing all emigrant and immigrant salmonids and can operate during extreme water
flows. The annual weir schedule of operation is developed at the annual Auke Creek cooperative research meeting.
The weir was installed in the downstream capture mode on 2 March 2004. Stream flow was moderate, and the
weir was operating by late afternoon. Continued rainfall and snow melting during the first weeks of
operation maintained streamflow at moderate levels. Often there are major icing problems on the weir
during March, but higher than average water and air temperatures prevented that this year. Auke Lake
remained ice covered throughout March, with water temperatures between 1.8° and 3.1°C.
Usually during March, pink and chum salmon fry dominate the number of emigrant salmonids, with an
occasional Dolly Varden or cutthroat trout captured. Through March, 22,950 pink salmon fry were counted
at the weir. Daily counts this year throughout March were greater than the historical daily averages.
This was the fourth highest number of pink salmon fry leaving Auke Creek in March. The average number
of pink salmon fry leaving Auke Lake in March is 8,714, and the highest count during any March from 1973
to 2003 was 45,000 in 1984. Pink salmon fry emigrations usually reach the midpoint of migration by 22
April, and the earliest midpoint of emigration was 1 April 1998. Only 904 chum salmon fry were captured
in March, less than average for Auke Creek. Six Dolly Varden and one cutthroat trout were captured.
Numbers of all fish typically increase during April.
By Jerry Taylor
AFSC Quarterly Research Reports Jan-March 2004