Abstract

Scientists are beginning to unravel the connections between patterns of climate variability, oceanic and freshwater environmental conditions, and the biotic responses of fish stocks. If these efforts bear fruit in the form of predictable relationships between the environment and stock abundance, the ability of fishery managers to forecast variation in stock recruitment should also improve. Because the data collection and preparation needed to support recruitment forecasts can be costly, it is reasonable to ask what such a forecast might be worth.

The basic framework of a decision analytic approach to assessing forecast value is adapted to a dynamic simulation of a two-stage inshore/offshore coho salmon (Oncorhynchus kisutch) fishery in Washington State. The fishery harvests intermingle stocks of wild and hatchery-reared coho. The manager attempts to meet spawning escapement goals and to maximize the economic value of the fishery by regulating catch levels sequentially in the offshore and inshore harvest sectors.

The model applies a stochastic simulation of the management process over 50-year sequences to compare outcomes without a forecast of interannual recruitment variation (assuming historical average recruitment rates) to performance under forecasts of varying accuracy.

The thesis concludes that 1) recruitment forecasts must be very accurate in order to have value, and 2) potential forecast value is limited by institutional constraints and by the predominance of hatchery production in Washington State.