9515

• UCREFRP Catalog Number: 9515
• Author: Waples, R. S., G. A. Winans, F. M. Utter and C. Mahnken
• Title: Genetic Approaches to the Management of Pacific Salmon
• USFW Year: 1990
• USFW - Doc Type: Fisheries
• Copyright Material: YES

LAWLER, MATUSKY & SKELLY ENGINEERS

Environmental Science & Engineering Consultants
One Blue HiI Plaza. Pea~ River. New York 10965
914-735-8300

FISHERIES SCIENCE/FISH PROTECnON
. SUlveys and MOdeling
. Intake TeChnology

HYDROPOWER UCENSING/MmGAnON
. ExhibW E
.IFIM
WATER QUALITY
. MOdeling
. Toxicity Evaluation

ENVIRONMENTAL IMPACT ASSESSMENT
. EIS/EAS Preparation
. Permilling
WETlANDS
. Delineation
· M~igation Design

[LMS]

anadromous salmonids in the Pacific Northwest. Most of
the state and federally operated hatcheries were initiated
to mitigate these losses. Current hatchery operations are
conducted on a gigantic scale: in 1987, over 330,000,000
Pacific salmon from over 130 facilities were released in
Washington State alone (Abrahamson, Washington De-
partment of fisheries unpublished data). A measure of the
success of hatchery production is the observation that
hatchery fish now represent a majority of the take in many
fisheries (e.g., Mathews 1980).
However, the success of some hatchery operations creates
other problems that concern both fishery managers and
conservationists. Whereas the discreteness of most wild
stocks is guaranteed by the high rate of homing, the concept
of a stock is less well defined in the hatchery environment.
Many hatchery "stocks" were started from multiple lineages
or subsequently received substantial input from sources
other than the original seed stock. Although less common
today, the exchange of eggs or progeny still occurs in many
hatcheries, rendering largely meaningless traditional defi-
nitions of a fish stock as "one consisting of randomly
interbreeding members whose genetic integrity persists
whether they remain spatially and temporally isolated as a
group, or whether they alternately segregate for breeding
and otherwise mix freely with members of other unit stocks
of the same species" (Kutkuhn 1981, p. 1477). In addition
to questions about the status of the hatchery stocks them-
selves, there is increasing concern about the impact of
hatchery fish on wild populations. Straying rates may be
higher for hatchery stocks in general, and several studies
have shown greatly increased rates of straying in fish
transplanted offsite (Ricker 1972; Lister et al. 1981; Withler
1982).
It is clear from the above that effectively managing
anadromous salmonids in the Pacific Northwest is a large
and difficult enterprise. Commensurate with the scope of
the problems, a number of large-scale programs have been
developed to deal with specific objectives (e.g., the North-
west Power Planning Act, the U S.-Canada salmon treaty,
and various programs for habitat reconstruction and fish
passage through or around dams).
In contrast to the millions of dollars invested in the above
programs, relatively few resources have been devoted to
addressing the genetic aspects of Pacific salmon manage-
ment. It is noted above that many of the major management

20

~

\........

. ~
problems are traceable to the fact that salmon spawn In
discrete freshwater populations. Therefore, there is an
underlying genetic basis to the entire management problem,
and genetic considerations must be an integral part of any
comprehensive management plan for Pacific salmon. The
importance of genetics to fisheries management in general
(Meffe 1986; Ryman and Utter 1987; Kapuscinski and Philipp
1988) and Pacific salmon in particular (Ricker 1972; Larkin
1981; NWPPC 1987) has not gone unnoticed, but too seldom
have genetic considerations played a significant role in
management decisions. Here, we outline three specific areas
in which genetic approaches can provide information of
considerable value to those concerned with managing Pacific
salmon populations.

Genetic Stock Identification

Given Kutkuhn's definition of a unit stock, a genetic
approach to the analysis of mixed-stock catches is natural.
Grant et al. (1980) were the first to demonstrate the use-
fulness of Genetic Stock Identification (GSI) in the analysis
of mixed-stock fisheries. Application of the method involves
three steps: (1) "baseline" genotypic frequency data (typi-
cally provided by protein electrophoresis) are gathered by
sampling potential source populations; (2) comparable data
are gathered for a sample of fish from the mixed-stock
fishery; (3) a computer algorithm is used to estimate stock
composition of the fishery, given information in (1) and (2).
Because Pacific salmon stocks are not usually characterized
by unique genetic markers (rather, they differ in frequencies
of the same suite of alleles), it is generally not possible to
determine with certainty the stock of origin for each indi-
vidual fish. Computation of a single estimate of stock
composition for all fish in the mixture has been shown to
be more precise and less biased than an approach that
classifies each fish individually (Millar 1987). With adequate
mixture and baseline samples, the composition estimates
can provide very reliable information about the relative
contributions of different stocks (Milner et al. 1981; Wood
et al. 1987; Pella and Milner 1987). In a given analysis, the
degree of resolution possible depends largely on the mag-
nitude of genetic differences among stocks. In some cases,
individual stocks within a drainage can be reliably distin-
guished; in other cases, substantial differences exist between
drainages but stocks within a drainage are similar enough
that estimates cannot reasonably be made for each individual
stock.
Although not as powerful as traditional tagging methods
for detecting movements of individual fish, GSI has several
advantages in the analysis of mixed-stock fisheries. first,
genetic marks are heritable and intrinsic to the populations;
it is not necessary to undertake an expensive tagging effort
every year for which data are desired (Waples 1990). Second,
apart from the difficulty of sampling remote streams (prin-
cipally in Alaska and British Columbia), inclusion of wild
populations is generally not a problem for GSI analysis.
Because physical tagging of wild populations is difficult
and expensive, even the most ambitious coded-wire-tag
programs routinely provide catch information only for
selected hatchery stocks (Johnson, in press). Third, standard
errors are routinely available for stock contribution estimates
computed using GSt whereas there are fundamental dif-
ficulties in estimating levels of precision for composition

Fisheries, Vol. 15. No.5