1 Scripps Institution of Oceanography, University of California, San Diego, La Jolla, CA 92093-0244, USA.
2 Center for Tropical Paleoecology and Archeology, Smithsonian Tropical Research Institute, Box 2072, Balboa, Republic of Panama.
3 National Center for Ecological Analysis and Synthesis, 735 State Street, Suite 300, Santa Barbara, CA 93101, USA.
4 Archie Carr Center for Sea Turtle Research and Department of Zoology, University of Florida, Gainesville, FL 32611, USA.
5 Department of Wildlife, Fish, and Conservation Biology, University of California, Davis, CA 95616, USA.
6 Department of Anthropology, 155 Pettengill Hall, Bates College, Lewiston, ME 04240, USA.
7 Centre for Resource and Environmental Studies, Australian National University, Canberra, ACT 0200, Australia.
8 Department of Anthropology, University of Oregon, Eugene, OR 97403, USA.
9 U.S. Geological Survey, A-316 Earth and Marine Sciences Building, University of California, Santa Cruz, CA 95064, USA.
10 Center for Coral Reef Biodiversity, Department of Marine Biology, James Cook University, Townsville, QLD 4811, Australia.
11 Department of Geophysical Sciences, University of Chicago, 5734 South Ellis Avenue, Chicago, IL 60637, USA.
12 Institute of Marine Sciences, University of North Carolina at Chapel Hill, 3431 Arendell Street, Morehead City, NC 28557, USA.
13 Department of Paleobiology, National Museum of Natural History, Smithsonian Institution, Washington, DC 20560-0121, USA.
14 School of Marine Sciences, University of Maine, Darling Marine Center, Orono, ME 04573, USA.
15 Department of Ecology, Evolution, and Marine Biology, University of California, Santa Barbara, CA 93106, USA.
* To whom correspondence should be addressed. E-mail: jbcj@ucsd.edu

Few modern ecological studies take into account the former natural abundances of large marine vertebrates. There are dozens of places in the Caribbean named after large sea turtles whose adult populations now number in the tens of thousands rather than the tens of millions of a few centuries ago (1, 2). Whales, manatees, dugongs, sea cows, monk seals, crocodiles, codfish, jewfish, swordfish, sharks, and rays are other large marine vertebrates that are now functionally or entirely extinct in most coastal ecosystems (3-10). Place names for oysters, pearls, and conches conjure up other ecological ghosts of marine invertebrates that were once so abundant as to pose hazards to navigation (11), but are witnessed now only by massive garbage heaps of empty shells.

Such ghosts represent a far more profound problem for ecological understanding and management than currently realized. Evidence from retrospective records strongly suggests that major structural and functional changes due to overfishing (12) occurred worldwide in coastal marine ecosystems over many centuries. Severe overfishing drives species to ecological extinction because overfished populations no longer interact significantly with other species in the community (5). Overfishing and ecological extinction predate and precondition modern ecological investigations and the collapse of marine ecosystems in recent times, raising the possibility that many more marine ecosystems may be vulnerable to collapse in the near future.

Importance of Historical Data

We exploited data from many disciplines that span the period over which anthropogenic changes may have occurred. Because our hypothesis is that humans have been disturbing marine ecosystems since they first learned how to fish, our time periods need to begin well before the human occupation or European colonization of a coastal region. Broadly, our data fall into four categories and time periods:

1) Paleoecological records from marine sediments from about 125,000 years ago to the present, coinciding with the rise of modern Homo sapiens.

2) Archaeological records from human coastal settlements occupied after about 10,000 years before the present (yr B.P.) when worldwide sea level approached present levels. These document human exploitation of coastal resources for food and materials by past populations that range from small-scale aboriginal societies to towns, cities, and empires.

3) Historical records from documents, journals, and charts from the 15th century to the present that document the period from the first European trade-based colonial expansion and exploitation in the Americas and the South Pacific (23).

4) Ecological records from the scientific literature over the past century to the present covering the period of globalized exploitation of marine resources. These also help to calibrate the older records.

Time Periods, Geography, and Analysis

The addition of a deep historical dimension to analyze and interpret ecological problems requires that we sacrifice some of the apparent precision and analytical elegance prized by ecologists (1, 13, 14). Paleoecological, archaeological, and historical data were collected for many purposes, vary widely in methods of collection and quality, and are less amenable to many types of statistical analysis than well-controlled experiments. But none of these problems outweighs the benefits of a historical approach. Clearly, we cannot generate realistic null hypotheses about the composition and dynamics of ecosystems from our understanding of the present alone, since all ecosystems have almost certainly changed due to both human and natural environmental factors (8, 16, 27, 28). Here, we briefly review long-term human impacts in several key marine ecosystems. These reconstructions provide insight into the nature and extent of degraded ecosystems that point to new strategies for mitigation and restoration that are unlikely to emerge from modern monitoring programs.

Kelp Forests

The kelp forest ecosystem of the Northern Pacific arose during the last 20 million years with the evolution of kelps, strongylocentrotid sea urchins, sea otters, and the extinct Steller's sea cow (6). Sea cows were widely distributed across the northern Pacific Rim through the Late Pleistocene. They may have been eliminated from most of their range by aboriginal hunting at the end of the Pleistocene and in the early Holocene, because they survived thousands of years longer in the western Aleutian Islands that were not peopled until about 4000 yr B.P. (6). By the time of European contact in 1741, sea cows persisted only in the Commander Islands, the only islands of the Aleutians unoccupied by aboriginal people. European fur traders killed the last sea cow 27 years later in 1768. We have no idea to what extent abundant sea cows grazed kelp forests, although their apparent inability to dive deeply probably limited their grazing to the surface canopy of kelps and to seaweeds lining the shore (6).

Northern Pacific kelp forests presumably flourished before human settlement because predation by sea otters on sea urchins prevented the urchins from overgrazing kelp (30). Aboriginal Aleuts greatly diminished sea otters beginning around 2500 yr B.P., with a concomitant increase in the size of sea urchins (31). Fur traders subsequently hunted otters to the brink of extinction in the 1800s with the attendant collapse of kelp forests grazed away by sea urchins released from sea otter predation. Legal protection of sea otters in the 20th century partially reversed this scenario. However, kelp forests are again being depleted in areas of Alaska because of increased predation on sea otters by killer whales (33). The whales shifted their diet to sea otters from seals and sea lions, which are in drastic decline.

A similar sequence of events occurred in kelp forests of the Gulf of Maine (7, 34). Sea otters were never present, but Atlantic cod and other large ground fish are voracious predators of sea urchins. These fishes kept sea urchin populations small enough to allow persistence of kelp forests despite intensive aboriginal and early European hook-and-line fishing for at least 5000 years. New mechanized fishing technology in the 1920s set off a rapid decline in numbers and body size of coastal cod in the Gulf of Maine (7) (Fig. 2A and Table 1) that has extended offshore to Georges Bank (35). Formerly dominant predatory fish are now ecologically extinct and have been partially replaced by smaller and commercially less important species. Lobsters, crabs, and sea urchins rose in abundance accordingly (7). Kelp forests disappeared with the rise in sea urchins due to removal of predatory fish, and then reappeared when sea urchins were in turn reduced to low abundance by fishing.

The more diverse food web of southern California kelp forests historically included spiny lobsters and large sheephead labrid fish in addition to sea otters as predators of sea urchins, as well as numerous species of abalone that compete with sea urchins for kelps (Fig. 1, A and B) (36). Aboriginal exploitation began about 10,000 yr B.P. and may have had local effects on kelp communities (37). The fur trade effectively eliminated sea otters by the early 1800s (38), but kelp forests did not begin to disappear on a large scale until the intense exploitation and ecological extinction of sheephead, spiny lobsters, and abalone starting in the 1950s (8, 36) (Table 1 and Fig. 1, A and B). Subsequent fishing of the largest sea urchin species in the 1970s and 1980s resulted in the return of well-developed kelp forests in many areas that, as in the Gulf of Maine, effectively lack trophic levels higher than that of primary producers (36, 39).

Coral Reefs

Large species of branching Acropora corals dominated shallow reefs in the tropical western Atlantic for at least half a million years (44-46) until the 1980s when they declined dramatically (42, 47) (Fig. 2B and Table 1). Patterns of community membership and dominance of coral species were also highly predictable (44), so that there is a clear baseline of pristine coral community composition before human impact.

Western Atlantic reef corals suffered sudden, catastrophic mortality in the 1980s due to overgrowth by macroalgae that exploded in abundance after mass mortality of the superabundant sea urchin Diadema antillarum that was the last remaining grazer of macroalgae (42, 47). Early fisheries reports suggest that large herbivorous fishes were already rare before the 20th century (48). However, macroalgae were held in check until the last major herbivore, Diadema, was lost from the system through disease (42, 47).

Corals on the Great Barrier Reef have experienced recurrent mass mortality since 1960 due to spectacular outbreaks of the crown-of-thorns starfish Acanthaster planci that feeds on coral (49). The causes of outbreaks are controversial, but they are almost certainly new phenomena. There are no early records of Acanthaster in undisturbed fossil deposits, in aboriginal folklore, or in accounts of European explorers and fishers. Now, in recent decades, the frequency and intensity of outbreaks have exceeded the capability of longer lived species to recover as outbreaks have become more chronic than episodic (50).

One possible explanation for Acanthaster outbreaks is that overfishing of species that prevy upon larval or juvenile stages of crown-of-thorns starfish is responsible for massive recruitment of the starfish (51). The highly cryptic, predator-avoiding behavior of juvenile starfish, their formidable antipredator defenses as subadults and adults, and the reduction of some generalized predatory fishes on the Great Barrier Reef all point to such a "top-down" explanation. Commercial and recreational fishing, as well as indirect effects of intensive trawling for prawns, are likely explanations for decreased abundance of predators of crown-of-thorns starfish (52). Massive recruitment of starfish may also be due to "bottom-up" increases in productivity due to increased runoff of nutrients from the land (53). In either case, the explanation is almost certainly historical and anthropogenic, and cannot be resolved by recent observations alone.

Expeditions occurred annually to northern Australia from the Malay Archipelago throughout the 18th and 19th centuries to harvest an estimated 6 million sea cucumbers each season (54). After European colonization, industrial-scale fishing developed along the Great Barrier Reef and subtropical east Australian coast in the early to mid-19th century (55). Whales, dugongs, turtles, pearl oysters, and Trochus shell were each heavily exploited only to rapidly collapse, and all have failed to regain more than a small fraction of their former abundance (55-57). Fishing of pelagic and reef fishes, sharks, and prawns has continued to the present, although catch per unit effort has declined greatly (58).

Vast populations of very large green turtles were eliminated from the Americas before the 19th century (1, 2) (Table 1). Formerly great populations of green turtles in Moreton Bay, Australia, also were greatly reduced by the early 20th century (66). Moreover, there are no estimates of abundances of turtles in Australia at the dawn of European exploitation, so that reported reductions must be only a small fraction of the total numbers lost. All turtle species continue to decline at unsustainable rates along the Great Barrier Reef today (67).

Abundant green turtles closely crop turtlegrass and greatly reduce the flux of organic matter and nutrients to sediments (59-62, 68). In the near absence of green turtles today, turtlegrass beds grow longer blades that baffle currents, shade the bottom, start to decompose in situ, and provide suitable substrate for colonization by the slime molds that cause turtlegrass wasting disease (65). Deposition within the beds of vastly more plant detritus also fuels microbial populations, increases the oxygen demand of sediments, and promotes hypoxia (65). Thus, all the factors that have been linked with recent die-off of turtlegrass beds in Florida Bay (65), except for changes in temperature and salinity, can be attributed to the ecological extinction of green turtles (27).

European colonists did not exploit tropical American manatees as systematically as they exploited green turtles, so the data related to fisheries are poor. We know, however, that manatees were extensively fished by aboriginal people and by early colonists (68). In Australia, aboriginal people also harvested dugongs extensively long before European colonization (3), yet the numbers reported by early colonists were vast. Three- or four-mile-long herds comprising tens of thousands of large individuals were observed in Wide Bay in about 1870 (69) and in Moreton Bay as recently as 1893 (70). Widespread colonial exploitation of dugongs for their flesh and oil along the southern Queensland coast resulted in the crash of the dugong fishery by the beginning of the 20th century (3) (Table 1). Ironically, scientists recently reported the "discovery of a large population" of dugongs in Moreton Bay--a mere 300 individuals (71). Further north, numbers of dugongs in the vast southern half of the Great Barrier Reef had dwindled to fewer than 4000 when they were first accurately counted in 1986-87, with a further 50 to 80% decline in recent years (72). These increasingly fragmented populations represent the last remnants of the vast herds of the early 19th century and before.

The ecological implications of these reductions are at least as impressive as those for green turtles. Moderate sized herds of dugongs remove up to 96% of above-ground biomass and 71% of below-ground biomass of seagrasses (73). Their grazing rips up large areas of seagrass beds, providing space for colonization by competitively inferior species of seagrasses. Dugong grazing also produces massive amounts of floating debris and dung that are exported to adjacent ecosystems. The decline in seagrasses in Moreton Bay is certainly due in large part to the dramatic decline in water quality due to eutrophication and runoff of sediment (63, 64). Nevertheless, as noted for green turtles and turtlegrass in Florida Bay, the cessation of systematic plowing of the bay floor by once abundant dugongs must also have been a major factor.

Oysters and Eutrophication in Estuaries

The oldest and longest records come from cores in sediments from Chesapeake Bay (77) and Pamlico Sound (84) in the eastern United States and from the Baltic Sea (85) that extend back as far as 2500 yr B.P. (Fig. 2C and Table 1). A general sequence of ecological change is apparent in all three cases, but the timing of specific ecological transitions differs among estuaries in keeping with their unique histories of land use, exploitation, and human population growth--a difference that rules out a simple climatic explanation. Increased sedimentation and burial of organic carbon began in the mid-18th century in Chesapeake Bay, coincident with widespread land clearance for agriculture by European colonists (77). The main ecological response was a gradual shift in the taxa responsible for primary production that began in the late 18th century. Seagrasses and benthic diatoms on the bay floor declined, while planktonic diatoms and other phytoplankton in the water column correspondingly increased. However, anoxia and hypoxia were not widespread until the 1930s when phytoplankton populations and the flux of organic matter to the bay floor increased dramatically with concomitant loss of benthic fauna (75, 77) (Fig. 2C and Table 1). Similar changes began in the 1950s in the Baltic Sea, with widespread expansion of the extent of anoxic laminated sediments (74, 85), and in the 1950s to 1970s in Pamlico Sound (84).

Vast oyster reefs were once prominent structures in Chesapeake Bay (11), where they may have filtered the equivalent of the entire water column every 3 days (79). Despite intensive harvesting by aboriginal and early colonial populations spanning several millennia, it was not until the introduction of mechanical harvesting with dredges in the 1870s that deep channel reefs were seriously affected (79, 80). Oyster catch was rapidly reduced to a few percent of peak values by the early 20th century (79, 80) (Fig. 2C and Table 1). Only then, after the oyster fishery had collapsed, did hypoxia, anoxia, and other symptoms of eutrophication begin to occur in the 1930s (75, 77), and outbreaks of oyster parasites became prevalent only in the 1950s (80). Thus, fishing explains the bulk of the decline, whereas decline in water quality and disease were secondary factors (80). However, now that oyster reefs are destroyed, the effects of eutrophication, disease, hypoxia, and continued dredging interact to prevent the recovery of oysters and associated communities (86). Field experiments in Pamlico Sound demonstrate that oysters grow well, survive to maturity, and resist oyster disease when elevated above the zone of summer hypoxia--even in the presence of modern levels of eutrophication and pollution (87).

Overfishing of oysters to the point of ecological extinction is just one example in a general pattern of removal of species capable of top-down control of community structure in estuaries. Dense populations of oysters and other suspension-feeding bivalves graze plankton so efficiently that they limit blooms of phytoplankton and prevent symptoms of eutrophication (88, 89), just as occurs with grazing by zooplankton in freshwater ecosystems (90). The ecological consequences of uncounted other losses are unknown. Gray whales (now extinct in the Atlantic), dolphins, manatees, river otters, sea turtles, alligators, giant sturgeon, sheepshead, sharks, and rays were all once abundant inhabitants of Chesapeake Bay but are now virtually eliminated.

Offshore Benthic Communities

The Primacy of Overfishing in Human Disturbance to Marine Ecosystems

There are three important corollaries to the primacy of overfishing. The first is that pollution, eutrophication, physical destruction of habitats, outbreaks of disease, invasions of introduced species, and human-induced climate change all come much later than overfishing in the standard sequence of historical events (Fig. 3). The pattern holds regardless of the initial timing of colonial overfishing that began in the Americas in the 16th and 17th centuries and in Australia and New Zealand in the 19th century. The full sequence of events is most characteristic of temperate estuaries like Chesapeake Bay. Not all the human disturbances illustrated in Fig. 3 have affected all ecosystems yet. But wherever these events have occurred, the standard chronological sequence of human disturbance and modification of ecosystems is recognizable.

The second important corollary is that overfishing may often be a necessary precondition for eutrophication, outbreaks of disease, or species introductions to occur (27). For example, eutrophication and hypoxia did not occur in Chesapeake Bay until the 1930s, nearly two centuries after clearing of land for agriculture greatly increased runoff of sediments and nutrients into the estuary (77). Suspension feeding by still enormous populations of oysters was sufficient to remove most of the increased production of phytoplankton and enhanced turbidity until mechanical harvesting progressively decimated oyster beds from the 1870s to the 1920s (77, 80) (Fig. 2C).

The consequences of overfishing for outbreaks of disease in the next lower trophic level fall into two categories. The most straightforward is that populations in the lower level become so dense that they are much more susceptible to disease as a result of greatly increased rates of transmission (94). This was presumably the case for the sea urchin Diadema on Caribbean reefs and the seagrass Thalassia in Florida Bay. In contrast, among oysters disease did not become important in Chesapeake Bay until oysters had been reduced to a few percent of their original abundance (80), a pattern repeated in Pamlico Sound (86, 87) and Foveaux Strait, New Zealand(93). Two factors may be responsible. First, oysters may have become less fit owing to stresses like hypoxia or sedimentation, making them less resistant to disease (87). Alternatively, suspension feeding by dense populations of oysters and associated species on oyster reefs may have indirectly limited populations of pathogens by favoring other plankton--an explanation that may extend to blooms of toxic plankton and most other outbreaks of microbial populations (88).

The third important corollary is that changes in climate are unlikely to be the primary reason for microbial outbreaks and disease. The rise of microbes has occurred at different times and under different climatic conditions in different places, as exemplified by the time lag between events in Chesapeake Bay and Pamlico Sound (77, 79, 80, 84). Anthropogenic climate change may now be an important confounding factor, but it was not the original cause. Rapid expansion of introduced species in recent decades (95) may have a similar explanation, in addition to increase in frequency and modes of transport. Massive removal of suspension feeders, grazers, and predators must inevitably leave marine ecosystems more vulnerable to invasion (96, 97).

Synergistic Effects of Human Disturbance

A striking feature of such synergistic effects is the suddenness of the transition in abundance of different kinds of organisms and community composition due to threshold effects (99). Ecological diversity and redundancy within trophic levels is probably the most important reason for the delay or time lag between the onset of fishing and the subsequent threshold response (42, 100). The importance of biodiversity in the form of ecological redundancy is clearly apparent for the delay in the collapse of kelp forests in southern California compared with Alaska after the extirpation of sea otters. Sheephead fish, spiny lobsters, and abalone in the more diverse Californian kelp forests kept sea urchin populations in check until these predators and competitors of sea urchins had also been effectively eliminated (8, 36). Similarly, the sea urchin Diadema kept macroalgae in check long after the extreme overfishing of herbivorous fishes on Caribbean coral reefs (42).

A second potentially important mechanism for the suddenness of ecosystem collapse is the elimination of previously unfished refuges that were protected historically because of distance or expense of access. For example, reef fishes all around Jamaica in the 1960s rarely reached reproductive maturity so that the abundant recruits of fishes on Jamaican reefs at that time must have come from undiscovered populations in Jamaica or elsewhere (101). But as more and more reefs have been overfished, the potential sources of such recruits must have effectively disappeared over wider areas (102). A similar scenario has been proposed for the American lobster with regard to loss of larvae from deep-water offshore stocks (103)..

Microbialization of the Global Coastal Ocean

Nowhere is the lack of historical perspective more damaging to scientific understanding than for microbial outbreaks. Plans for remediation of eutrophication of estuaries are still based on the belief that eutrophication is caused only by increased nutrients without regard to overfishing of suspension feeders. Even more remarkable is the attribution of the rise in marine diseases to climate change and pollution (104) without regard to the pervasive removal of higher trophic levels and the asynchronous outbreaks of disease in different ecosystems that belie a simple climatic explanation.

Historical Perspectives for Ecosystem Restoration

On the other hand, recognition of these losses shows what coastal ecosystems could be like, and the extraordinary magnitude of economic resources that are retrievable if we are willing to act on the basis of historical knowledge. The central point for successful restoration is that loss of economically important fisheries, degradation of habitat attractive to landowners and tourists, and emergence of noxious, toxic, and life-threatening microbial diseases are all part of the same standard sequence of ecosystem deterioration that has deep historical roots (27). Responding only to current events on a case-by-case basis cannot solve these problems. Instead, they need to be addressed by a series of bold experiments to test the success of integrated management for multiple goals on the scale of entire ecosystems. With few exceptions, such as the Caribbean monk seal and Steller's sea cow, most species that are ecologically extinct probably survive in sufficient numbers for successful restoration. This optimism is in stark contrast with the state of many terrestrial ecosystems where many or most large animals are already extinct (28). Moreover, we now have the theoretical tools (109) to roughly estimate per capita interaction strengths of surviving individuals of now rare animals like sea turtles, sirenians, sharks, and large groupers. We can then use these data to build tentative models of the consequences of the renewed abundance of these species in their native environments that can in turn be used to design large-scale, adaptive experiments for ecosystem restoration, exploitation, and management (96, 108, 110).

One obviously timely and overdue experiment is to attempt the amelioration of eutrophication, hypoxia, and toxic blooms in Chesapeake Bay by massive restoration of oyster reefs (79). Experiments in Pamlico Sound show that this is possible (86, 87, 96), and modeling of food webs suggests that even partial restoration of oysters would reduce eutrophication substantially (110). Aquaculture of suspension-feeding bivalves like oysters might be promoted to reverse the effects of eutrophication and to restore water quality in degraded estuaries. Other important examples include the restoration of coral reefs and seagrass beds by protection of fishes, sharks, turtles, and sirenians in very large reserves on the scale of all of Florida Bay and the Florida Keys--an approach recently advocated for terrestrial ecosystems (111). Once again, small-scale grazing experiments with reef fishes (112) show that fishes could reverse the overgrowth of corals by macroalgae on a massive scale. The potential for reducing diseases of corals and turtlegrass by restoring natural levels of grazing is unproven but consistent with historical evidence (27).