Complex Trophic Interactions in Kelp Forest Ecosystems

The distributions and abundances of species and populations change almost continuously. Understanding the processes responsible is perhaps ecology s̓ most fundamental challenge. Kelp-forest ecosystems in southwest Alaska have undergone several phase shifts between algaand herbivore-dominated states in recent decades. Overhunting and recovery of sea otters caused the earlier shifts. Studies focusing on these changes demonstrate the importance of top-down forcing processes, a variety of indirect foodweb interactions associated with the otter-urchin-kelp trophic cascade, and the role of food-chain length in the coevolution of defense and resistance in plants and their herbivores. This system unexpectedly shifted back to an herbivore-dominated state during the 1990s, because of a sea-otter population collapse that apparently was driven by increased predation by killer whales. Reasons for this change remain uncertain but seem to be linked to the whole-sale collapse of marine mammals in the North Pacific Ocean and southern Bering Sea. We hypothesize that killer whales sequentially “fished down” pinniped and sea-otter populations after their earlier prey, the great whales, were decimated by commercial whaling. The dynamics of kelp forests in southwest Alaska thus appears to have been influenced by an ecological chain reaction that encompassed numerous species and large scales of space and time. The distributions and abundances of species and populations vary through time. Although human activities contribute to this variation, they are not solely responsible for it. Hence, managers and conservationists should be less concerned with change per se than with the extent to which change is driven by human influences and the degree to which it can be managed. Before these determinations can be made, however, the underlying mechanisms of change must be understood. Unfortunately, these mechanisms are usually poorly known because observing pattern is relatively easy whereas understanding process is much more difficult. Although the truth of this claim is evident from any number of case studies, two wellknown examples suffice to make the point. First consider the early Holocene megafaunal extinctions. Beginning about 13,000 yrs ago, more than half of medium to large-sized mammalian species abruptly disappeared from the New World (Martin and Wright, 1967). The pattern is clear, but the responsible processes are still debated (Krech, 1999), despite a substantial body of evidence that hunting by early humans was the main cause (Martin, 1973; Alroy, 2001). The problem is that the Pleistocene-Recent boundary was a time of rapid climate change, which some believe indirectly caused the megafaunal extinctions and others believe must at least have interacted with human hunting in important ways. Next consider the disappearance of large dinosaurs at the CretaceousTertiary boundary, again a period of rapid environmental change. Although the pattern is indisputable, the responsible processes are still debated, despite a wealth of evidence that the extinctions were caused by a large meteor crashing into the earth (Alvarez et al., 1980). Many other examples of the asymmetry in understanding between pattern and BULLETIN OF MARINE SCIENCE, VOL. 74, NO. 3, 2004 622 process could easily be provided, including any number of recent or ongoing trends that are concerns for conservation and management. Why are ecological processes so elusive, and how might we better understand them? The answer to the question of “why” seems clear enough. Recognizing pattern amounts to little more than identifying and enumerating species and individuals on proper scales of space and time, tasks the human body and mind are well equipped to handle. The processes that drive these changes are typically much more difficult to see and understand because of their dynamic nature and enormous potential for complexity. The present paper focuses on both the nature and the understanding of process in some North Pacific kelp-forest ecosystems. Marine ecosystems in this region have undergone a number of marked changes over the past 50 yrs but with little understanding about why (National Research Council, 1996). Our perspectives are based primarily on three decades of research on sea otters and kelp forests in the Aleutian archipelago, admittedly a marginal part of the Bering Sea–North Pacific ecosystem but one about which we have developed a mechanistic understanding for some of the observed changes. Until recently, we watched others attempt to understand the precipitous population declines of northern fur seals, harbor seals, and Steller sea lions, never imagining that the future of sea otters and kelp forests might be connected to these changes. Then, in about 1990, following nearly a century of recovery from the ravages of the Pacific maritime fur trade, sea otter populations in southwest Alaska unexpectedly plummeted (Doroff et al., 2003). We were forced to look both to the past and to the open sea for explanations and, although much remains to be learned, these retrospections have added insight to our growing view of the sea-otter–kelp-forest system. We will explain why by recounting the history of our key findings, how we have interpreted them, and how these interpretations have altered some of our views about the changeability, and hence the meaning, of foodweb dynamics.