Abundance of salmon has fluctuated greatly. Figure 1 shows the fluctuation in landings of chinook (Oncorhynchus tshawytscha) and coho (O. kisutch) salmon in the United States. While we have tended to concentrate on the general decline of chinook and coho salmon from California to British Columbia, we should not lose sight of the fact that even in the 1970s there was a general perception that salmon stocks were declining, and in Canada the Salmonid Enhancement Program was begun with the objective of doubling the number of salmon.
We want to understand what causes the changes in abundance and what the impacts of alternative human actions will be. The traditional explanations for changes in salmon abundance have been the 4-Hs: habitat, harvest, hatcheries, and hydropower. We might therefore compare the trends in catch, escapement, total run, etc., to human action in one of these factors, or perhaps in several of them.
It has long been recognized that there is variability in ocean survival of salmon, and most analyses of human impacts on salmon will treat ocean survival as a form of uncontrolled noise that confounds the analysis. However, the increasing recognition of large-scale changes in ocean conditions suggests that we might need to do more than simply allow for random ocean survival, but rather try to measure it directly.
Since the early 1970s, there has been an extensive program of tagging Pacific salmon using coded wire tags. At present, approximately 30 million fish are tagged every year, and the salmon management agencies on the Pacific coast conduct an extensive tag recovery program. These data can be used to estimate the survival of tagged fish in the ocean. Coronado (1995) and Coronado and Hilborn (in prep. (a,b)) have used these data to describe the changes in ocean survival.
We used the estimated recoveries of commercially and recreationally caught fish, and the estimates of tagged fish in hatchery returns, to calculate the survival rate using virtual population analyses (VPA) (Hilborn and Walters 1992). For each species we calculated the estimated survival from time of ocean migration to age 3, the most common age of return for coho and fall chinook salmon. While VPA does require making an assumption about the rate of natural mortality for fish in the ocean after age 1, all trends and relative survival are quite insensitive to these assumptions.
Figures 2, 3, and 4 show the geographic trends in survival for coho and spring and fall chinook salmon by state and province. In each case, the vertical bars represent standard errors of the mean computed by using each tag code within the state or province for that year as a replicate sample. The shaded region represents total hatchery releases.
The trends in survival are both striking and geographically diverse. For instance, British Columbia coho salmon showed high initial survival in the early 1970s, followed by a steady decline through the 1980s to a level of about one-third of the initial survival. In contrast, Washington coho salmon showed steady survivals up until the 1987 brood year and then a decline after that. For fall chinook salmon, British Columbia and Washington showed reasonably similar trends, while Oregon and California were quite different. Spring chinook salmon showed yet a different pattern.
Within geographic regions, there tends to be considerable coherence. For instance, Figure 5 shows the trends in coho salmon survival for Columbia River hatcheries in Washington and Oregon. Even though the hatcheries are run by different agencies, the trends, particularly in the 1980s, are almost identical.
We have performed cluster analysis of survival trends for coho and fall chinook salmon, and the data tend to cluster by geographic region. We believe these trends reflect the survival in the early ocean life history of these fish, and this is, in turn, related to the general oceanographic production regime of the stock.
It is our contention that the survival rate reflects the early ocean survival, and what we see in the hatchery stocks reflects similar trends in wild fish. While the majority of coded wire tags are placed on hatchery fish, there has been some tagging of wild stocks. Figure 6 shows the trends in survival of wild and hatchery coho salmon. While far from conclusive, these data suggest to us that the trends in survival seen in the hatchery fish are likely a reflection of the survival of wild fish.
This analysis suggests that much of the fluctuation in abundance of chinook and coho salmon can be explained by changes in ocean survival. This does not mean, however, that we should ignore human impacts and simply hope for better ocean conditions. The major purpose of looking at ocean survival is to eliminate this form of variability from analysis when we consider modification of habitat, harvest, hatcheries, and hydropower.
As an example, we have constructed a simple analysis of optimal harvesting for a coho salmon population with constant freshwater conditions, but under two different ocean regimes (Fig. 7). Here we assume that there is a freshwater carrying capacity and a density-independent ocean survival. In case 1, we assume a 15% ocean survival, roughly comparable to the conditions in British Columbia in the early 1970s. The optimal escapement of this hypothetical stock is 315, and the optimal exploitation rate is 63%. The sustainable yield is about 500 fish. If the ocean then turns bad and survival drops to 5% (roughly current survival rates), the sustainable yield drops to 50 fish. The optimum escapement is now only 105, and the optimum harvest rate is 37%.
Thus, while we cannot control the ocean, we must change our management actions as ocean conditions change. We suggest that under present ocean conditions there is little if any sustainable yield for chinook and coho salmon in Oregon, Washington, and British Columbia. Harvest rates need to be drastically reduced, and we should expect the escapements to drop. We might choose to maintain escapements at the old levels, but we should recognize that the reason to do this would be to try to retain a large population size until ocean conditions improve.
Coronado, C. 1995. Spatial and temporal factors affecting survival of hatchery-reared chinook, coho and steelhead in the Pacific Northwest. Ph.D. Thesis, University of Washington, Seattle, 235 p.
Coronado, C., and R. Hilborn. In prep. (a). Spatial and temporal factors affecting survival in coho salmon (Oncorhynchus kisutch) in the Pacific Northwest.
Coronado, C., and R. Hilborn. In prep. (b). Spatial and temporal factors affecting survival in chinook salmon (Oncorhynchus tshawytscha) in the Pacific Northwest.
Hilborn, R., and C. J. Walters. 1992. Quantitative fisheries stock assessment: Choice, dynamics, and uncertainty. Chapman and Hall, New York, 570 p.