Pacific salmonids (Oncorhynchus spp.) have complex life histories. They display considerable inter- and intraspecific variability at all life stages, which results from the influence of numerous factors. Despite this variability, geographical patterns in some of the better-studied life stages can provide considerable insight into ecological processes that occur during estuarine and ocean residency, a period for which little is known. This paper describes several examples of how life history trait patterns provide information about key questions concerning estuarine and ocean residence, such as when and why significant mortality occurs, which fish survive, and how widespread particular patterns are. Estuarine and ocean residency is just one part of the life cycle, but considerable insight can be gained by viewing it in the context of the entire life cycle.
A review of data for salmonid populations along the west coast of North America (from numerous sources including Aro and Shepard 1967, Atkinson et al. 1967, Groot and Margolis 1991, Weitkamp et al. 1995) suggests that life history traits largely fall into two categories—those with clear latitudinal and altitudinal trends (e.g., smolt outmigration and spawn timing) and those without apparent trends (e.g., adult size and smolt size). Presumably, traits with strong latitudinal/altitudinal trends are most affected by factors that also have strong latitudinal trends (e.g., temperature, photoperiod), while traits that do not show strong latitudinal trends may be affected by single factors without latitudinal trends or by numerous factors working at smaller geographic scales.
One life history trait that shows a mixed pattern is freshwater age. Two species, pink (O. gorbuscha) and chum (O. keta) salmon, always leave fresh water shortly after emergence. In contrast, all other species examined (chinook (O. tshawytscha), sockeye (O. nerka), coho (O. kisutch), Atlantic (Salmo salar), and masu (O. masou) salmon and steelhead (O. mykiss)) generally have extended (> or = 1 year) freshwater residence and clear latitudinal/altitudinal trends in the duration of the residence.
The difference in patterns of freshwater age (leaving immediately vs. extended residence) suggests the existence of two general juvenile strategies used to make the transition between fresh water and salt water. Possible advantages of the pink/chum (and, to a lesser degree, ocean-type chinook and sea-type sockeye salmon) strategy include avoiding fresh water and its associated problems (such as relatively low average productivity and high mortality and low growth during winter), and taking advantage of highly productive estuarine or nearshore marine habitats, where rapid growth can be achieved to quickly outgrow high levels of size-specific predation. The other strategy (extended freshwater residence) might be driven by low predation levels in freshwater habitats or a longer period of freshwater growth to avoid high predation on small fish in estuarine and marine environments. Because fish using these two strategies enter estuarine and marine waters at dramatically different ages and sizes, they have different requirements from initial estuarine and marine habitats, and mortality patterns are likely to be quite different. Consequently, the results of studies examining populations using one type of strategy may not be relevant to populations using the other strategy.
One example of how knowledge of freshwater residency may provide insight into ocean mortality comes from freshwater age measurements from southeast Alaska populations of coho salmon (Halupka et al. 1993), a species with extended freshwater residence. Freshwater ages in these populations were measured on outmigrating smolts as well as returning adults, allowing a comparison of freshwater age of fish going to sea with those returning from sea. Although freshwater ages measured at these two stages were similar for most populations, two populations were notable exceptions. In both cases, the average freshwater age measured from the smolts was older than the freshwater age measured from the adults, suggesting that ocean mortality within these populations was not random, but rather was higher on the older juveniles. In this example, patterns of life history traits provide insight about which fish do and do not survive, and suggest processes (i.e., low freshwater growth rates) that might have some bearing on marine survival.
Another example of how life history traits provide insight into ocean residence comes from ocean migration patterns for coho salmon as inferred from coded wire tag (CWT) recovery patterns. We compiled marine (as defined in the database) CWT recovery records (PSMFC 1994) for 65 coho salmon hatcheries from southeast Alaska to central California, using recoveries expanded for sampling effort but not for untagged fish (Weitkamp et al. 1995). The proportion of recoveries occurring in each state or province for each hatchery are indicated; CWT recovery patterns are fairly consistent among hatcheries within regions (e.g., British Columbia, Puget Sound and Hood Canal, Columbia River, etc.), but there are abrupt changes between adjacent regions, rather than a gradual transition as might be expected due strictly to the geographic location of the hatcheries (Fig. 1). For example, CWT recovery patterns for coho salmon released from the Naselle Hatchery on Willapa Bay (Washington coast) and from Grays River Hatchery on the lower Columbia River are very different from each other, even though the two hatcheries are separated by less than 20 air miles. Naselle and other Washington coast hatchery coho salmon have much higher recovery rates from British Columbia and Washington and correspondingly lower recovery rates from Oregon and California than do Grays River and other lower Columbia River hatchery fish (Fig. 1). Although these patterns reflect only the last few months of a 1.5-year migration, the dramatic differences in migration patterns suggest that there also may be significant differences earlier in ocean residence as well.
Although patterns of CWT recoveries also vary over time, the responses of stocks within a region are generally strongly correlated. Figure 2 displays the percentage of total marine CWT recoveries from Oregon for fish released from Washington, Oregon, and California coastal hatcheries between 1978 and 1992. Within each of the three regions, patterns over time are very similar across populations, but they are quite different between the three regions. This suggests that fish within each region are responding in similar manner to the ocean conditions they encounter, but that conditions vary significantly between regions.
We also examined trends in coho salmon adult size from different regions and found the same pattern of high similarity within, and large differences between, regions, particularly during anomalous years (Fig. 3). For example, Oregon coast coho salmon returning to rivers north of Cape Blanco in 1983 were exceptionally small, presumably a result of the strong El Niño that year (Johnson 1988). In contrast, coho salmon from rivers south of Cape Blanco or north of the Columbia River were not unusually small in 1983, suggesting they were not experiencing the same ocean conditions as the Oregon coast fish, or, if they were, they were not affected to the same degree. Puget Sound coho salmon were unusually small the following year (1984, Fig. 3), suggesting that either 1) conditions that caused the small size for Oregon coast coho salmon also affected Puget Sound fish, but at an earlier age, or 2) marine conditions moved and affected Puget Sound coho salmon at the same age as the Oregon fish, but 1 year later.
The above discussion indicates that patterns for coho salmon ocean migration and adult size exist at relatively small spatial scales. Other species or populations may show patterns at different spatial scales. For example, Olsen and Richards (1994) compared trends in abundance for stream-type chinook salmon from the Snake River with stream-type Fraser River and ocean-type Puget Sound chinook salmon. Despite the much larger distance by water between the Snake and Fraser River chinook salmon spawning grounds compared to Snake River and Puget Sound spawning grounds, the former pair's abundance trends were very similar (r = 0.64), while those for the latter pair were not (r < or = 0.1). This suggests that the factors that determine year class strength affected both Snake and Fraser River populations in a similar fashion and were quite different than the factors that determine year class strength for Puget Sound populations. The strength of the Fraser-Snake River correlation, in spite of the presence of eight mainstem dams for the Snake River fish to navigate and none for the Fraser River fish, suggests that ocean conditions play a key role in overall mortality for stream-type chinook salmon in both systems. In this case, geographically distant populations show more similarity than do geographically close populations, which illustrates how life history trait patterns seem particularly useful at indicating how widespread ocean processes and patterns might be.
A final example of how life history traits provide insight into estuarine and ocean residence comes from the use of jack (precocious male) abundance as a predictor of the following year's adult abundance. Prior to 1983, Oregon coast and Columbia River coho salmon jack (2 years old) abundance was a good predictor of adult (3 years old) abundance (Pearcy 1992). Because jacks only spend 6 months at sea instead of 18 months for adults, this suggested that much of the mortality that determined year-class strength occurred early in the ocean residence. However, since 1983, the relationship between jack and adult abundance has weakened, and jack abundance has been a poor predictor of adult run size in some years. This change in the jack-adult relationship suggests two likely scenarios—either the timing of significant mortality events has changed to later in the ocean residency, perhaps due to low prey abundance, or the proportion of males that become jacks has become more variable, due to large smolt size or poor ocean conditions in some years. Furthermore, Puget Sound coho salmon jack run size has never been a good predictor of adult run size, suggesting a mortality schedule quite different than that for Oregon coast coho salmon prior to 1983. In this example, the relationship between jacks and adults provides information about when large mortality events occur, indicates that these patterns can change over time, and shows how widespread the patterns are.
To summarize, we have provided several examples of how life history trait information provides valuable insight into patterns and processes that occur during estuarine and ocean residency. These examples were the product of a fairly limited review of the subject, and we expect that many more such examples will be unearthed with further investigation. Although ocean residency is just one part of the salmon life cycle, considerable insight about it can be gained by viewing it in the context of the larger life cycle.
Aro, K. V., and M. P. Shepard. 1967. Pacific salmon in Canada. Bull. Int. North Pac. Fish. Comm. 23:225-327.
Atkinson, C. E., J. H. Rose, and T. O. Duncan. 1967. Pacific salmon in the United States. Bull. Int. North Pac. Fish. Comm. 23:43-224.
Groot, C., and L. Margolis (editors). 1991. Pacific salmon life histories. Univ. British Columbia Press, Vancouver, 564 p.
Halupka, K. C., J. K. Troyer, M. F. Wilson, and F. H. Everest. 1993. Identification of unique and sensitive coho salmon stocks of southeast Alaska. Draft manuscript, U.S. Dep. Agriculture, Pacific Northwest Research Station, Forest Sciences Laboratory, Juneau, AK, 97 p. (Available from K. Halupka, Biology Dep., Albertson College of Idaho, Caldwell, ID 83605.)
Healey, M. C. 1983. Coastwide distribution and ocean migrations patterns of stream- and ocean-type chinook salmon, Oncorhynchus tshawytscha. Can. Field-Nat. 97:427-433.
Johnson, S. L. 1988. The effects of the 1983 El Niño on Oregon's coho (Oncorhynchus kisutch) and chinook (O. tshawytscha) salmon. Fish. Res. 6:105-123.
Olsen, D., and J. Richards. 1994. Inter-basin comparison study, Columbia River salmon production compared to other West Coast production areas, Phase II analysis. Report to the U.S. Army Corps of Engineers, 80 p. (Available from U.S. Army Corps of Engineers, P.O. Box 2946, Portland, OR 97208-2946.)
Pacific States Marine Fisheries Commission (PSMFC). 1994. Regional Mark Information System (RMIS) Coded-wire tag on-line database. (Available from Pacific States Marine Fisheries Commission, 45 SE 82nd Dr., Suite 1000, Gladstone, OR 97027.)
Pearcy, W. G. 1992. Ocean ecology of North Pacific salmonids. Washington Sea Grant Program. Univ. Washington Press, Seattle, 179 p.
Weitkamp, L. A., T. C. Wainwright, G. J. Bryant, G. B. Milner, D. J. Teel, R. G. Kope, and R. S. Waples. 1995. Status review of coho salmon from Washington, Oregon, and California. U.S. Dep. Commer., NOAA Tech. Memo. NMFS-NWFSC-24, 258 p.
Wood, C. C. 1995. Life history variation and population structure in sockeye salmon. In J. L. Nielsen (editor), Evolution and the aquatic ecosystem: Defining unique units in population conservation. Am. Fish. Soc. Symposium 17, Bethesda, MD, p. 195-216.