Spring upwelling in the year of ocean entry is generally recognized as an oceanic indicator of coho salmon (Oncorhynchus kisutch) recruitment (Nickelson 1986, Pearcy 1992, Lawson 1993). Recent discussions (this conference) have focused on the observation that, in the Oregon Production Index (OPI) area, south of Leadbetter Point, Washington, hatchery coho salmon smolt-to-adult survival has not been correlated with upwelling in the current period of low upwelling years. Nickelson (1986) showed that in all low upwelling years between 1960 and 1981 there had been no correlation between hatchery survival and upwelling. I suggest that upwelling, by itself, is not an adequate indicator of ocean conditions. Winter sea surface temperature in the adult return year is independent of upwelling and, in combination with upwelling, can be used to explain much of the variation in recruitment of naturally spawning coho salmon in Oregon. Survival and recruitment of hatchery coho salmon, primarily from the Columbia River, are affected by freshwater factors including variability in rearing and release practices and problems with disease. These fish enter an ocean more heavily influenced by the Columbia River freshwater plume and less strongly affected by upwelling compared with the more southerly distributed Oregon coastal natural (OCN) stocks. For these reasons, it is likely that OCN and hatchery stock groups would respond differently to changes in ocean conditions. Recruitment time series, rather than survival, are used for analysis because there are no direct estimates for naturally produced smolts from Oregon over the period from 1970 to 1996.

Table 1 presents a 25-year time series of ocean population size (adult recruits to the fishery) for public hatchery and OCN coho salmon, along with spring upwelling (UW) at lat. 42°N (Bakun 1975) and winter sea surface temperature (SST) anomalies at Charleston, Oregon. Note that the OCN recruit estimates are an index which is thought to overestimate true abundances by about 50% to 80% (Jacobs and Cooney 1994). Although the numerical estimates are biased high, they capture the interannual variations in recruitment which are the subject of this analysis. The UW, SST, and hatchery and OCN recruits time series show significant correlations with year, indicating secular trends (Table 2). The signs on all trends are negative except for SST, which has been increasing. Upwelling and SST are strongly correlated with OCN recruits (P < or = 0.0005), less strongly with hatchery recruits (0.07 > P > 0.002), but not with each other (P > 0.38, Table 2). The lack of a significant correlation between UW and SST is unusual in ocean environmental time series and supports their use as indicators of two independent oceanic processes. Autocorrelation within time series implies that the effective number of degrees of freedom is less than the number of years in the series (Kope and Botsford 1990). Employing the conservative correction described by Kope and Botsford (1990), the correlation between OCN recruits and SST remains significant (P < 0.02), and OCN recruits with UW is marginally significant (0.10 > P > 0.05).

Together, UW and SST explain most of the variability in OCN coho salmon recruitment since 1970, including years affected by El Niño. Multiple linear regression of the natural log of recruits vs. UW and SST explains about 75% of the variability in recruitment (Figs. 1 and 2; Table 3, Model 1). Although there is a secular trend in all three time series (Table 2), the residuals from the regression still show a secular decline. Accounting for this decline would explain an additional 16% of the variation. Upwelling and SST explain only about 27% of the variability in hatchery coho salmon recruitment, with UW not a significant factor in the regression (Table 3, Model 2).

Several factors may contribute to an apparent lack of correlation between upwelling and hatchery coho salmon recruitment (or survival) in recent years. Figure 2 compares OCN and hatchery coho recruits plotted against an index of ocean conditions calculated from the model parameters presented in Table 3, Model 1. The index is 0.011*UW - 0.365*SST. Average ocean conditions from 1971 to 1995 receive an index of zero. All but 2 years since 1984 show below-average ocean conditions. Although recent years have shown low recruitments for OCN coho salmon, the overall pattern of increasing recruitment with improved ocean conditions is consistent across the full range of ocean conditions. Hatchery coho salmon, on the other hand, show much more variable recruitment when plotted against this index. Years prior to 1984 fit the general pattern of higher recruitment with favorable ocean conditions. However, recent years show no pattern. Considerable experimentation with hatchery rearing and release practices has been conducted in the past 15 years, complicating interpretation of these data. Winter SST seems to be much more important than UW in explaining variability in public hatchery recruitment (Tables 2 and 3, Model 2). Spring SST in the year of ocean entry has been identified as having high correlation with hatchery survival (T. E. Nickelson, Oregon Dep. Fish Wildl., 850 SW 15th St., Corvallis, OR 97333. Pers. commun., 1996).

Time series of average weights from ocean commercial troll fishery landings, and a condition factor based on a regression of weight vs. length, show interannual variations that may provide insights into life history processes not evident from the recruit data. Chinook salmon average weights since 1950 show a decadal fluctuation, with higher weights in the 1950s and 1970s (Fig. 3a). In contrast, coho salmon show a decline in average weight since 1950 (Fig. 3b). Both species tended to be small in size during the strong El Niños in 1957-58 and 1982-83. Scale analysis of OCN coho salmon shows no evidence of a decline in size between 1950 and 1990, so this trend may be specific to hatchery reared coho salmon (D. Bottom, Oregon Dep. Fish Wildl., 850 SW 15th St., Corvallis, OR 97333. Pers. commun., 1996.). Weight:length functions (Fig. 4) for coho salmon suggest both environmental and density-dependent factors may be operating. In years with relatively normal ocean conditions (1984, 1985, and 1986), the length:weight regressions have similar slopes. Fish were shorter and in poorer condition in 1986, when ocean populations of coho salmon in the waters off Oregon were over three times as abundant as in either 1984 or 1985, indicating that growth rates may have been reduced at high ocean densities. The 2 years of strong El Niño show the highest contrast. While mean lengths and weights were identical, and the range of sizes was similar, the slopes of the regressions diverged. In 1983, longer fish were of lower weight, while in 1992 it was the shorter fish that were low in weight. The 1982-83 El Niño was unusual, with effects reaching the Oregon coast by October 1982, compared with the 1992 event, which reached Oregon in January 1992. The length:weight function in 1992, with low abundance and poor ocean conditions, was remarkably similar to 1986, with high abundance and better ocean conditions.

Mechanisms linking UW and SST with survival and growth of coho salmon in Oregon are not well known. Pearcy (1992) presents a recent review of salmonid ocean ecology, with emphasis on coho salmon. Brodeur and Pearcy (1990) and Fisher and Pearcy (1990) investigated early oceanic life history of coho and chinook salmon from 1981 to 1985. Aside from these studies, little has been done to investigate mechanisms of survival in the ocean. Upwelling in the year of ocean entry affects the nearshore oceanic environment by increasing production, lowering temperatures, and adding spatial complexity in the form of fronts and thermal breaks (J. Barth and Smith, this volume). It may also affect the movements of predators and the availability of salmon smolts to predators. Winter SST is an index of the quality of offshore oceanic conditions, which are affected by atmospheric conditions in the central North Pacific (Norton and McLain 1994; Parrish, this volume). Sea temperatures may affect predator movements, indicate nutrient regimes, or be linked to bioenergetics of salmon and their prey. Other factors potentially contributing to trends in OCN recruitment include declines in freshwater productivity (Lawson 1993), density dependent survival (McGie 1984, Emlen et al. 1990), accumulated effects of hatchery coho salmon interactions with wild fish, and a trend since 1970 of later onset of upwelling. Future investigations should focus on discovering the links between indices of ocean conditions and the ocean life history of Pacific salmonids.

Bakun, A. 1975. Daily and weekly upwelling indices, West Coast of North America, 1967-73. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF-693. 114 p.

Brodeur, R. D., and W. G. Pearcy. 1990. Trophic relations of juvenile Pacific salmon off the Oregon and Washington coast. Fish. Bull., U.S. 88:617-636.

Emlen, J. M., R. R. Reisenbichler, A. M. McGie, and T. E. Nickelson. 1990. Density-dependence at sea for coho salmon (Oncorhynchus kisutch). Can. J. Fish. Aquat. Sci. 47:1765-1772.

Fisher, J. P., and W. G. Pearcy. 1990. Spacing of scale circuli versus growth rate in young coho salmon. Fish. Bull., U.S. 88:637-643.

Jacobs, S. E., and C. X. Cooney. 1994. Improvement of methods used to estimate the spawning escapement of Oregon coastal natural coho salmon. Oreg. Dep. Fish Wildl. Ann. Prog. Rep. 24 p.

Kope, R. G., and L. W. Botsford. 1990. Determination of factors affecting recruitment of chinook salmon Oncorhynchus tshawytscha in central California. Fish. Bull., U.S. 88:257-269.

Lawson, P. W. 1993. Cycles in ocean productivity, trends in habitat quality, and the restoration of salmon runs in Oregon. Fisheries (Bethesda) 18(8):6-10.

McGie, A. M. 1984. Commentary: Evidence for density dependence among coho salmon stocks in the Oregon Production Index area. In W. G. Pearcy (editor), The influence of ocean conditions on the production of salmonids in the North Pacific, p. 37-49. Oregon State Univ. Sea Grant, Corvallis.

Nickelson, T. E. 1986. Influences of upwelling, ocean temperature, and smolt abundance on marine survival of coho salmon (Oncorhynchus kisutch) in the Oregon Production Area. Can. J. Fish. Aquat. Sci. 43:527-535.

Norton, J. G., and D. R. McLain. 1994. Diagnostic patterns of seasonal and interannual temperature variation off the west coast of the United States: Local and remote large-scale forcing. J. Geophys. Res. 99(8):16,019-16,030.

Pearcy, W. G. 1992. Ocean ecology of North Pacific salmonids. Washington Sea Grant Program, University of Washington Press, Seattle, 179 p.

Table 1. Oregon public hatchery and coastal natural coho (OCN) salmon ocean recruitment, with two environmental variables. Recruitment in thousands of fish. Sea surface temperature (SST) anomaly; deviation from mean 1971-95 January + February sea surface temperature at Charleston, Oregon, °C, upwelling (UW) anomaly; deviation from mean April + May + June Bakun upwelling index, lat. 42°N.

Table 2. Pearson correlation coefficients (P < > 0) of Oregon public hatchery and Oregon coastal natural (OCN) coho salmon recruitment and two environmental variables. See also Table 1.

Table 3. Regression statistics. See also Table 1.

Table of Contents