3 of them significantly, including the smooth dogfish, which increased 13-fold. However, these were deepwater species generally out of reach of trawling.

Release from predators can have spectacular consequences through the development of trophic cascades. For example, reductions of ≈90–99% for 11 large sharks that consume smaller elasmobranchs along the northwest Atlantic Coast of North America resulted in the increase of 12 of their 14 common prey species (Myers et al., 2007). Populations of one of these smaller species, the cownose ray Rhinoptera bonasus, exploded to some 40 million. Each ray can consume ≈210 g shell-free wet weight of bivalve mollusks per day, assuming that they are available. The rays migrate through Chesapeake Bay each year, where they stay for ≈100 days, which amounts to a potential consumption of 840,000 metric tons of mollusks/year. In contrast, the commercial harvest of bay scallops that peaked in the early 1980s had fallen by 2003 to only 300 metric tons. Thus, the rays could potentially consume 2,500 times the commercial harvest, and it is hardly surprising that the once prosperous clam fisheries have totally collapsed.

The canonical example of a trophic cascade involves the near extinction of sea otters by hunting in the northeast Pacific that resulted in explosions of sea urchins that in turn eliminated entire kelp forests by overgrazing (Estes and Duggins, 1995). Trophic cascades have also been documented for the formerly cod-dominated ecosystem of the northwest Atlantic (Frank et al., 2005), where removal of large groundfish resulted in large increases in pelagic shrimp and snow crabs, decreases in large zooplankton, and increases in phytoplankton. Thus, removal of top-down controls affects ecosystem structure and function of large marine ecosystems with complex food webs, as well as simpler, low-diversity systems.

The occurrence of trophic cascades is closely linked to the phenomenon of fishing down the food web (Pauly et al., 1998). Analysis of Food and Agriculture Organization global fisheries data showed a decline in mean trophic level of the global predatory fish catch of ≈0.1 per decade since 1950. There has been much discussion of the quality and suitability of the data, and whether the decline in trophic level primarily reflects serial depletion of overfished species or serial additions of lower trophic level species caused by a depletion of their predators (Essington et al., 2006; NRC, 2006). Regardless of the relative importance of these different mechanisms, however, there is increasingly reliable evidence for the pervasive decline in mean trophic level in heavily fished ecosystems. For example, mean trophic level fell from 4.06 to 3.41 in the Bohai Sea between 1959 and 1998 (Zhang et al., 2007), a decline of 0.16–0.19 per decade. This drop parallels a dramatic 95% decline in total fish catch from 190 to 10 kg per standardized haul per hour (Table 1.1) and a precipitous drop in the