Many fish stocks undergo predictable migrations on a seasonal or longer time scale, either for feeding or reproduction, or in response to changing environmental conditions or physiological needs. Such stocks can become stratified so that younger and older animals are geographically separated and the fishery may become stratified, with different groups of fishermen harvesting different segments of the stock. If this occurs, a major obstacle to effective fishery management may be resolving the allocation of harvest (and bycatch) among the geographically separated fishermen. In the case of species that migrate across national boundaries (e.g., Atlantic and Pacific salmon, Pacific whiting, and tuna and billfish species), catch allocation conflicts can become especially difficult to resolve, as vividly illustrated by the difficulties surrounding breakdown of the Pacific Salmon Treaty.
Most fish stocks in temperate seas breed seasonally and exhibit high variability in their annual production of offspring. In some stocks, the largest year classes are as much as several hundred times larger than the smallest year classes (Myers et al., 1995). This large variability in recruitment leads to great uncertainty in determining appropriate TACs, which generally are based on the notion of a well-measured, functional relationship between the size of a parent stock and its subsequent production of offspring. Because recruitment is often highly variable and seemingly independent of parental stock size, it is extremely difficult to determine how much of the stock to leave behind (i.e., how to set the TAC). Interannual variability in the growth rates of individuals is another source of uncertainty for some fish stocks. In regions of the Gulf of Alaska in 1980, for example, 12-year-old Pacific halibut were twice the weight of halibut of the same age in 1996 (IPHC, 1997). However, the presence of uncertainty and variability is not adequate grounds for rejecting TAC-based management because TACs can be designed to reflect variability and risk.
Most marine animals are strongly affected by their environment; the influence of the varying biophysical environment on stock size and the condition of individual animals has long concerned fishery scientists. Small (almost unmeasurable) changes in growth and mortality rates during early life that are attributable to environmental variability can lead to significant changes in annual recruitment and persistent impacts on stock size (e.g., Hofmann and Powell, 1998).
Although some recent studies of fish population dynamics emphasize "surprises," discontinuities, and uncertainties (e.g., Botkin, 1990; Ludwig et al., 1993; Wilson et al., 1994), the continuing dominant approach of applied fish population dynamics and bioeconomics that emphasizes deterministic, single-species linear relationships and equilibrium conditions may not account for the realities of many fisheries.
One very difficult fishery management situation, which is sometimes described as the "mixed-stock fishery" (Ricker, 1958; Paulik et al., 1967) or "mixed-species fishery" problem (Clark, 1985a), arises when biologically productive (fast-growing and fecund) and unproductive (slow-growing, late-maturing, slowly