Skip booklet index and go to page content

COSEWIC assessment and status report on the Sockeye Salmon Oncorhynchus nerka Sakinaw population in Canada

Population Sizes and Trends

Indices of Spawning Abundance

Because all sockeye salmon die after spawning, and always in the same year that they attain maturity, the number of mature fish in the population is usually estimated as the number of fish that survive to spawn. In most populations, this is roughly equivalent to the numbers that survive coastal fisheries and reach their natal spawning habitat (called the “spawning escapement”). Estimates of spawning escapement to Sakinaw Lake are recorded for 1947 to 2002 in DFO’s Salmon Escapement Data System (SEDS). In most years, between 1949 and 1989, the SEDS estimate is based on counts of sockeye entering Sakinaw Lake through the fishway and is considered a reasonably reliable index of both relative and absolute abundance (see discussion in Murray and Wood 2002). Enumeration at the fishway was discontinued between 1990 and 2001, but the number of sockeye spawning on beaches was estimated by various methods of unknown reliability and consistency.

SEDS estimates show no obvious trend between 1947 and 1987, fluctuating between 750 and 16 000, and averaging about 5 000. Although this population has never been “actively managed”, the target escapement suggested by DFO is 14 000 sockeye (DFO 1988). Since 1987, escapement estimates have decreased steadily (Figure 5). Summarized by 5-year intervals, reported escapements between 1988 and 1992 averaged just over 1000 (range 500 to 2500); between 1993 and 1996, less than 200 (range none observed to 250); and between 1997 and 2001, less than 50 (range 1 to 122). To some extent, enumeration effort and inconsistent methods will have affected the accuracy of the yearly estimates from 1989 through 1998, but the overall decline is undoubtedly serious. More systematic dive surveys of the spawning grounds conducted in 1999 through 2002 yielded estimates from 14 ( 23 redds) to 122 spawners (60 redds). In 2002, sockeye were enumerated both at the fishway and by dive survey, allowing a direct comparison; the fence count was 78 whereas the diver count on the spawning beaches was only 44 (G. McBain, DFO, pers. comm). These results suggest that dive surveys may underestimate true abundance, although this seems surprising in view of the excellent viewing conditions, and opportunity for inadvertently counting the same fish again on a subsequent survey. Alternatively, these results may indicate that sockeye experience significant mortality after they enter Sakinaw Lake. The latter explanation seems plausible because sockeye enter the lake several months before spawning, and are vulnerable to predation, especially by a non-anadromous parasitic lamprey (probably Lampetra tridentata). All spawning sockeye captured as broodstock in 2002 bore lamprey scars although none had fresh wounds (G. McBain, DFO, pers. comm.)

Figure 5: Trends in Number of Mature Individuals in the Sakinaw Lake Sockeye Salmon Population

Figure 5: Trends in number of mature individuals in the Sakinaw Lake sockeye salmon population.

Open circles are annual estimates of spawning escapement; filled circles are the corresponding estimates smoothed over one-generation (4 yr); line is fitted to smoothed data by LOWESS.

Annual counts of spawning fish should represent all mature fish in the population in each year, but they often fluctuate widely because of year-to-year variations in brood year (parental) abundance and survival. To remove annual “noise” unrelated to any underlying trend in population status, spawner counts should be smoothed, for example by computing a running average over one-generation. The (negative) slope of a straight line fitted to a smoothed time series plotted on a logarithmic abundance scale will provide the best estimate for a constant rate of decline caused by an underlying threatening process. This procedure also facilitates comparison with threshold rates of decline that trigger designation under the IUCN Criteria. Annual escapement data for Sakinaw Lake sockeye from 1988 to 2002 were smoothed with a 4- year running average to generate a 3-generation, 12-year time series of smoothed values, corresponding with years 1990 to 2001 in Figure 6. No spawners were reported in 1995, but this was treated as a case of missing data rather than an absence of spawners. The smoothed data were log-transformed and regressed on year (p<0.001) to estimate the rate of decline at 33% per year, or 99% decline over 3-generations. The alternative procedure suggested in the COSEWIC guidelines based on the reduction in abundance from the first to last year of the time series will be affected by annual fluctuations in individual year class strength, and thus, sensitive to the particular two years chosen for comparison. However, in this case, it yields similar estimates ranging from 93% (1990 visual estimate to 2001 dive survey) to 87% (1991 visual estimate to 2002 fence count) reduction over three generations.

Figure 6: Average Rate of Decline in Number of Mature Individuals in the Sakinaw Lake Sockeye Salmon Population

Figure 6: Average rate of decline in number of mature individuals in the Sakinaw Lake sockeye salmon population.

Estimated by regression of 1-generation smoothed escapement data (filled circles) is 33% Per Year or 99% Over Three Generations (12 yr). Open circles are annual estimates of spawning escapement; curved lines are 90% confidence intervals for regression line; dashed lines indicate IUCN thresholds of 50% and 80% decline over 12 years. Data for 1995 are treated as missing.

Indices of Juvenile Abundance

Smolts were enumerated by mark-recapture experiments at the outlet of Sakinaw Lake from 1994 to 1997. The total smolt outmigration in those years was estimated at 15 880, 12 760, 2 500 and 5 200, respectively, based on a trap efficiency of 3 to 5% (Bates and August 1997). If smolt-to-adult survival rate was 4.5%, an average value for other sockeye populations with large smolts (Forester 1968), the corresponding total adult returns before fishing mortality would have been 715, 574, 113, and 232 adults in 1996, 1997, 1998 and 1999, respectively. Of course, reported escapements in these years were considerably lower (1 to 222), probably because of underestimation by visual survey, and real losses to fishing mortality and in-lake predation. However, even if we disregard the visual estimates of spawning escapement and assume that marine survival has been favourable and fishing mortality negligible, the smolt abundance estimates indicate that total adult abundance must have declined by an order of magnitude since the more reliable counts in the 1980s.