The reason this distinction—top-down versus bottom-up control—is so important is that it drastically affects predictions about the impact of fishing. If control is from the bottom up, as oceanographers and fishery scientists have generally supposed, then the shock waves of our yearly 80 million ton scoop may propagate a short ways upward, to hapless animals like seabirds and seals, but will mostly be confined to the populations we actually fish. If control is top-down, however, our impact will cascade in complicated ways, causing floors below to become alternately crowded and empty, potentially compromising the food supplies of some species we fish, and broadly rearranging ecosystems. For any manager, then, the bottom-up consensus was enormously comforting.
And yet, despite all the good reasons trophic cascade ought not transpire in the sea, that's just what seems to be happening. Reading the literature and talking with marine scientists from a variety of disciplines, one gets the sense that the acknowledged domain of trophic cascades—the geographic area in which, according to most scientists, they are in fact occurring—has been spreading slowly but steadily outward, from small confined systems, like ponds and streams, out onto the littoral, across shallow reefs, and finally into the deeper waters of the ocean.
The very first demonstration of trophic cascade occurred in 1960, in little ponds, where the Czech ecologist Jaroslav Hrbacek observed that phytoplankton, the tiny photosynthetic organisms that fuel the food web in every body of water, actually increased their abundance in the presence of small fish. The fish, Hrbacek realized, were suppressing zooplankton, which occupy floor two in underwater ecosystems. In the late sixties and early seventies, top-down control was documented in several studies at the ocean's edge, and in 1974, a clear case of trophic cascade was detected in the Aleutian Islands. Sea otters had been extirpated from the islands about two centuries before, by Russian fur traders. Around islands where otters were now recovering, sea urchins were held in check, allowing a lush kelp forest to regrow. Where otters had yet to return, urchins kept the stone bare.
Even by the late nineties, however, top-down control was not generally considered relevant off the coast. John Steele, an eminent scientist at Woods Hole Oceanographic Institution, registered the view of most marine scientists when he wrote, in 1998, "The massive perturbations in marine fish stocks might be expected to have impacts on their food supplies, yet I do not know of any cases where this has been demonstrated in the sea."
But then, over the following decade, cascades came to light all over the oceans. The Aleutian cascade stretched out into open water when ocean-roving killer whales abruptly and mysteriously changed their behavior: they began eating sea otters, and just as one would have expected, based on the earlier findings, the effects cascaded down to kelp forests, many of which soon gave way to urchin barrens. Other cascades were found in the North Pacific, involving salmon, and then on both sides of the North Atlantic, involving cod. Ransom Myers, a prescient and industrious marine biologist who had for years been assembling a database tracking the world's fisheries, discovered that, off the east coast of the United States, fishing for large sharks had unleashed their prey, including rays, which were now demolishing shellfish beds. His study received much attention, not only because it was scientifically compelling, but also because the plague of rays had just brought an end to North Carolina's century-old scallop fishery.
Three days before the study was published in Science, Myers died, aged 54, of a brain tumor, but his ideas and database fed into many more publications. Among the most important was a synthesis led by Boris Worm, a protégé of Myers' at Dalhousie, and Ray Hilborn, a fisheries scientist at the University of Washington. Published in 2010, the work showed that, worldwide, the large deepwater fish that occupy the top of marine food webs had declined by 56%, while their prey—smaller, schooling fish—had actually increased by 143%, likely because they were released from top-down control. "Top-down interactions cascading from fishers to predators and their multiple prey species," the authors concluded, are "important structuring forces."