Marine ecosystems respond quite differently to environmental changes over various time scales, and they are different enough from terrestrial ecosystems that much of our understanding of terrestrial ecosystems does not apply to marine ones (Steele 1985, 1991a, 1991b, 1991c, 1996, 1998). Oceanic environmental changes occur diurnally (tides), over periods of several days (storms), over periods of months to years (upwelling, eddies, warm-core rings [Hofmann and Powell 1998]), over several years (El Niño-Southern Oscillation, North Atlantic Oscillation, variations in sea-ice cover in the Barents and Bering seas and Southern Ocean), and up to a century or more (the conveyor-belt, a large-scale ocean-circulation system involving all major oceans from arctic to Antarctic latitudes [Broecker 1991]). In addition, human activity other than fishing has affected the marine environment and marine and anadromous fishes at a variety of time and space scales.
Natural4 environmental changes have been implicated in several examples of changes in marine ecosystems, some of them described earlier in this chapter (e.g., the Bering Sea, the Barents Sea). Indeed, large fluctuations in population densities (or at least distributions) of marine fishes have been documented from periods long before fishing could have been a factor, most notably the very large fluctuations in the densities of scales of hake, anchovy, and sardine in sediment cores off California over the past 2,000 years (Soutar and Isaacs 1974) (Figure 3-1). Because they are so widespread and have been widely reviewed elsewhere (e.g., Wooster 1983, Wooster and Fluharty 1985, Laevastu 1993, Everett et al. 1996), only a few examples of environmentally related fluctuations in fish populations are provided here, both human caused and natural. We caution that it is often difficult to disentangle the effects of environmental changes from those of fishing on fish populations; often they are both important factors. The recent study by Polovina and Haight (in press) of spiny lobsters in a protected and an unprotected area in Hawaii is an excellent example of how experimental or observational controls are needed to prevent the effects of fishing and environmental changes from being confounded.
Salmon populations and fisheries in the Pacific Northwest have declined severely over the last century (NRC 1996b). Salmon have disappeared from 40 percent of their historical breeding ranges in the continental United States and continue to decline despite a public and private investment of more than $1 billion in the last decade for protection and enhancement of salmon populations. The frequency of populations experiencing the greatest difficulty increases in a