COMMENTARY Apex predators and trophic cascades in large marine ecosystems: Learning from serendipity Robert S. Steneck1 Darling Marine Center, School of Marine Sciences, University of Maine, Walpole, ME 04573 he global loss of large predators large, and important benthic predators is undeniable. However, the throughout cold temperate to subarctic T effects of predator depletion on regions of the North Atlantic (Fig. 1) (2, the structure and functioning of 3). In the western North Atlantic pre- ecosystems are far from resolved, espe- historic fishing targeted cod not only be- cially as they apply to large pelagic marine cause of their abundance and size but also ecosystems. Much of what we know about because they are easy to catch and pre- how marine predators function in ecosys- serve (2, 4). Cod became the first impor- tems comes from small-scale studies on tant export from the New World and as relatively small, slow-moving, seafloor- fishing technology and effort escalated, feeding predators that are easy to manip- serial depletion progressed from coastal ulate. Scaling up to consider pelagic regions to the final collapse of Canada and (ocean) ecosystem effects from large US cod stocks in the 1990s. The resulting predatory fish has been challenging for decline in cod predation relaxed popu- several reasons. For one thing, predators lation limitations on several invertebrate have been functionally removed from prey species that live on or near the sea many marine ecosystems due to unsus- floor such as American lobsters, large crab tainable fishing that occurred decades or species (e.g., Jonah and snow crabs), and centuries ago. Also, studies often rely on shrimp (2, 3, 5, 6). In addition to those correlations showing increases in prey benthic invertebrate carnivores, the pred- populations as predators decline, but these ator decline also relaxed demographic correlations can be confounded if coin- limits on herbivorous sea urchins in a tro- cident oceanographic factors such as phic cascading of widespread grazing ocean warming control prey abundances. down of kelp forests (7). What is needed is an ecosystem-scale, Trophic cascades are most common and large predator addition experiment for clearly evident in low-diversity benthic which most biotic components of the marine ecosystems (8, 9). Whereas top– ecosystem are monitored before and after down (consumer-driven) effects in benthic predator addition. Such an experiment Fig. 1. Ghosts of apex predators past. Photo of food webs are widely accepted, their is exactly what Casini et al. (1) report Atlantic cod from coastal Maine circa 1880. (Image is in the public domain.) application to large pelagic marine eco- in PNAS. systems is more contentious. The well- Casini et al. (1) document a rare but “ ” documented decline of predatory cod important natural experiment of preda- zooplankton from herring predation pres- along Canada’s Scotian Shelf was deter- tor addition into a relatively isolated sure. The rise in herbivorous zooplankton 2 mined to have resulted in increased her- 18,000-km branch of the Baltic Sea called resulted in a decline in phytoplankton, ring and other small pelagic fish, causing the Gulf of Riga. The apex predator in the increasing water clarity during the decade their zooplankton prey to decline, ulti- Baltic, Atlantic cod (Gadus morhua), is cod were abundant. What the authors de- mately increasing the region’s phyto- a large generalist carnivore that before “ ” scribe is called a trophic cascade in plankton (10). However, the zooplankton fishing was widespread, abundant, and fi which higher-order consumers signi cantly and phytoplankton changes were also co- possibly the most important predator affect how organisms interact at three or incident with an oceanographic regime throughout coastal regions of the North more lower trophic levels of the food web. shift resulting from climate changes in the Atlantic (2, 3). Casini et al. (1) describe The Casini et al. study illustrates and Arctic (11). That example illustrates con- a pulse infusion of juvenile and adult cod reaffirms three important points: (i) Apex founding vulnerabilities associated with into the Gulf where cod fed and grew for predators can affect large pelagic marine drawing conclusions from single-trend about a decade but, for environmental ecosystems. (ii) Cod were a foundation correlations. The Casini et al. study, how- reasons, they could not reproduce or sus- species for the North Atlantic before its ever, has a stronger case for a trophic tain their population. Importantly, cod fisheries-induced extirpation (2) and it cascade by having quantitative data for and all other key players in this pelagic hints at why this species often fails to re- ecosystem such as herring, herbivorous cover from very low population densities the entire food web before and after the zooplankton, and phytoplankton have (3). And (iii) perhaps the overarching addition of cod in this large marine eco- been continuously monitored since 1973. message is that complex ecological inter- system. Thus, evidence is growing that Thus, when Baltic cod populations swelled actions among and between managed Atlantic cod may have the unique capac- and spilled over into the Gulf of Riga in species can drive dynamics of local pop- ity to trigger large-scale trophic cascades 1977, they effectively initiated a predator ulations in demographically significant addition natural experiment into a sys- ways that profoundly affect entire ecosys- tem that had well-established baseline tems. I briefly expand on each of these Author contributions: R.S.S. wrote the paper. conditions. The predator pulse lasted a three points. The author declares no conflict of interest. decade during which time cod’s prey, There is little doubt that before fishing See companion article on page 8185. herring, declined in abundance, releasing pressure, Atlantic cod were abundant, 1E-mail: [email protected]. www.pnas.org/cgi/doi/10.1073/pnas.1205591109 PNAS | May 22, 2012 | vol. 109 | no. 21 | 7953–7954 Downloaded by guest on October 1, 2021 in both benthic and pelagic marine Such aggregations will result in higher sure, but importantly it resulted from an ecosystems. fertilization success than that of a single in-migration of juvenile and adult cod. The cod influx into the Gulf of Riga also pair mating in isolation. Numerous These older cod thus escaped the mortal- informs us about how subpopulations of spawning aggregation sites along coastal ity bottlenecks befalling those 1 y old and this species expand to colonize new areas. Maine in the 1920s were targeted by fish- less by avoiding limiting feedback mecha- New colonization by cod may seem un- ers and eliminated over the next several nisms such as the reverse predator–prey remarkable but it stands in stark contrast decades (12, 18). Today there are no dynamics described above. As a result, to most previous studies on spatial dy- the older recruits had no difficulty surviv- namics of this species that focused on how The local cod stock ing and growing in the Gulf of Riga. Re- and why cod stocks have declined, con- cruitment by older cod stands in stark tracted, or gone locally extinct (12). Fur- colonization in the contrast to the widely held concept that ther, it is particularly frustrating for fish population expansion is driven by the managers and policymakers that this spe- Gulf of Riga provides us early life sequences leading to recruitment cies, when fished to low population den- (known as “The Recruitment Limitation sities, recovers slowly or apparently not at with a clue to how cod Hypothesis,” sensu ref. 20). Perhaps cod all even after all fishing has ceased (13, uniquely establish new local stocks as 14). For example, after Atlantic cod stocks populations proliferate. larger adults. If so, this colonization pro- collapsed, both Canada and the United cess needs to be considered by managers States closed large areas to fishing in the and policymakers. Perhaps local stocks at early 1990s. Fisheries managers expected known spawning aggregation sites in low but not dysfunctional reproductive full recovery within a decade but after coastal Maine despite low or zero fishing levels should be allowed to recover enough nearly two decades cod stocks remain at pressure in many regions over the past half not only to restore the depleted local historically low levels of abundance (15). century. Thus, local reproductive and stock but also to colonize adjacent areas Such protracted lags in population or population dynamics could be extremely where past stocks have been extirpated. ecosystem recovery often result from re- important to the regional success of We learn from these recent studies that inforcing ecological feedbacks (14). this species. Atlantic cod can affect entire food webs in Several different, but not mutually ex- The local cod stock colonization in the both the benthic and the pelagic realms. clusive, feedback mechanisms reducing Gulf of Riga (1) provides us with a clue to Not only are they strong interactors capa- reproductive success or increasing mor- how cod populations proliferate. Although ble of limiting the abundance of their prey tality of eggs, larvae, or young of the year cod are broadly distributed throughout and their prey’s prey (i.e., trophic cas- cod may interfere with or prevent the re- the North Atlantic, their population cade), but also the prey themselves may covery of cod stocks. One theory gaining structure is effectively a mosaic of sub- limit the recovery of this predator. In most support in both the eastern and the west- stocks or local stocks that complete their countries where fisheries management ern North Atlantic is predator–prey re- life cycle within a relatively small area. exists, the focus is on the dynamics of versal between cod and forage fish. Forage Cod tagging studies, genetics, and long- single species and often there is no con- fish examples include herring (15) or ca- term fisheries independent monitoring in- sideration of how two or more managed plin (16) in the Atlantic and sprat in the dicate this species is a large-scale “meta- species interact or how such interactions Baltic (17).