The ecosystems that support fisheries, together with other economic activities, are subject to a number of alterations of significant relevance to their functioning and resilience and to the goods and services they can provide. Because of our imperfect understanding of ecosystem structure and functioning, as well as the inherent difficulty of distinguishing between natural and human-induced changes, the latter are not always perfectly predictable and/or reversible (Christensen et al., 1996). The following sections elaborate briefly on some types of alteration.
Impacts from fisheries on the environment have been abundantly described and reviewed (Dayton et al., 1995; Goñi, 1998; Kaiser et al., 2003; Gislason, 2003; Agardy, 2000). More specifically, capture fisheries impact target resources. They reduce their abundance, spawning potential and, possibly, population parameters (growth, maturation, etc.). They modify age and size structure, sex ratio, genetics and species composition of the target resources, as well as of their associated and dependent species. When poorly controlled, fisheries develop excessive fishing capacity, leading to overfishing, with major ecosystem, social and economic consequences.
Fishing may also affect ecological processes at very large scale. The overall impact has been described as comparable, in aquatic systems, to that of agriculture on land in terms of the proportion of the system's primary productivity harvested by humans (Pauly and Christensen, 1995). Overfishing transforms an originally stable, mature and efficient ecosystem into one that is immature and stressed. This happens in various ways. By targeting and reducing the abundance of high-value predators, fisheries deeply modify the trophic chain and the flows of biomass (and energy) across the ecosystem (e.g. Pauly, 1979). They can also alter habitats, most notably by destroying and disturbing bottom topography and the associated habitats (e.g. seagrass and algal beds, coral reefs) and benthic communities.
The alteration of the habitat by various human activities may be physical (e.g. by adding artificial structures like artificial reefs, oil rigs, aquaculture installations), mechanical (e.g. through the "ploughing" effect of dredges and trawls), or chemical (e.g. through injection of nutrients, pesticides, heavy metals, drugs, hormones). Fishing may result in changes in productivity of resources (some positive and some negative) and affects associated species. Some aspects of fisheries can have significant and long-lasting effects, e.g. destructive fishing techniques using dynamite or cyanides or inadequate fishing practices (e.g. trawling in the wrong habitat); pollution from fish processing plants; use of ozone-depleting refrigerants; dumping at sea of plastic debris that can entangle marine animals or be swallowed by turtles; loss of fishing gear, possibly leading to ghost fishing; lack of selectivity, affecting associated and dependent species, resulting in wasteful discarding practices, juvenile mortality, added threat to endangered species, etc. Poorly-managed large-scale mariculture can damage coastal wetlands and nearshore ecosystems, often used as nurseries by key capture fishery resources, and contribute to ecosystem contamination with food residues, waste, antibiotics, hormones, diseases and alien species.
The Law of the Sea provides that fisheries management must take care also of associated and dependent species. The impact of fishing on these species has been documented in some areas (Goñi, 1998; Gislason, 2003) but is still frequently unknown or only partly understood. The decline of primary productivity consumers low in the food chain removes important forage species needed higher in the food web, with cascading effects for the ecosystem. Conversely, the removal of top predators such as mammals, tuna or sharks, may release an unusually large abundance of preys at lower levels with cascading and feedback effects on the food chain and species composition (Trites, 2003; Cury et al., 2003). For example, as most sharks and some batoid fishes (angel fishes) are predators located at or near the top of marine food webs, their depletion modifies the intricate trophic interactions of their ecosystems (Pauly and Murphy, 1982; Jackson et al., 2001). The removal of predators through fishing in Kenyan reefs resulted in the expansion of sea urchin population, which apparently led to a decrease in live coral and to loss of topographic complexity, species diversity and fish biomass (McClanahan and Muthiga, 1988). Goñi (1998) reports that the hunting of sea otters (Enhydra lutris) in the Northeast Pacific caused a large-scale expansion of sea urchins, the increased grazing of which caused the decline of the important kelp forest. She also reports that, in the Bering Sea, the expansion of the fisheries on pollock (Theragra chalcogramma) during the 1970s has been considered as a probable cause of the decline of several populations of marine mammals, e.g. sea lions (Eumetopias jubatus) by 76%, seals (Callorhinus ursinus) by 60% and (Phoca vitulina) by 85%, as well as the decline of several seabird populations (Urea algae, U. lomvia, Rissa brevirostris, R. tridactyla). All these non-fish species compete directly with the Pollock fisheries since the target species represent 21-90% of their diet. Impacts can be particularly serious on cartilaginous fish populations, which have a lower productivity and resilience than bony fishes. As a consequence, fisheries targeting shark have a low record of sustainability and some species of skates, sawfish and deep-water dogfish have been virtually extirpated from large regions (Garcia and Majkowski, 1990; Stevens et al., 2000).
A well documented example of direct impact on benthic species is that of modern towed gear (trawls and dredges) which caused, inter alia, long-term changes in abundance and species composition in the Wadden Sea (Goñi, 1998) and Australia (Moran and Stephenson, 2000). The mortality of benthic species associated with or preyed upon by target bottom fish resources resulting from the use of trawls can vary greatly, depending on how the gear is built or rigged (Moran and Stephenson, 2000). For example, the addition of a tickler chain on a commercial beam trawl will allow it to catch more of its bottom-fish target but, at the same time, will detach and uproot more benthic species (as bycatch) (Kaiser et al., 1996 and 1998; Dayton et al., 1995). Likewise, some gear modification can reduce fishing mortality. For example, operating a semi-pelagic trawl 15 cm above the sea bottom has no measurable effect on the benthos community, while the standard demersal trawling dragging on the bottom reduces benthos density by 15.5% (Moran and Stephenson, 2000). Finally, fishing can have significant impact on the genetic diversity of resources and can permanently change populations characteristics (Kenchington, 2003).
Fishing gear can change the living and non-living environment within which the target and other related resources live. Environmental damage may come from the very nature of the fishing technology (e.g. in the use of dynamite or poison) or from the inappropriate use of an otherwise acceptable gear (e.g. using trawls in coral reefs or seagrass beds).
The use of dynamite and other explosives for "blast fishing" is still common in parts of Asia, Africa, Caribbean and South Pacific. A relatively small explosive (beer-bottle size) is capable of destroying a three-metre circular area of stony corals (Goñi, 1998). These practices are generally officially banned by fisheries regulations and laws but often persist because the people involved have little, if any, alternative livelihood.
The impact on the habitat depends on the gear and sediment type. Highly dynamic, soft bottoms (e.g. coarse sand, hydraulic dunes) may suffer limited damage even when exploited by heavy (including hydraulic) dredges. On the contrary, stable, hard, and highly structured habitats (such as coral reefs, seagrass beds, sponge beds) will be easily damaged. One well-documented example is the use of modern towed gear (trawls and dredges) which caused, inter alia, destruction of seagrass beds (Posidonia oceanica) in the Mediterranean and destruction of the oyster (Cassostrea virginica) habitat in Chesapeake Bay (Goñi, 1998). Damage is also related to fishing frequency, gear weight and rigging. Addition of heavy tickler chains to the trawl ground rope increases bottom abrasion and turbidity (Kaiser et al., 1996 and 1998; Dayton et al., 1995) while adding rollers reduces it. The use of sodium cyanide in the Philippines to catch marine tropical fishes for the aquarium trade has led to the destruction of the coral reef habitat and decline of aquarium and food fish.
Fishing generates bycatch and discards. A first attempt to address the issue at global level was made in the late 1990s by FAO (1997c) and the first estimate of the global extent of the problem (about 27 million tonnes of resources dumped per year) was published by this Organization (Alverson et al., 1994). A recent review of the issue has been undertaken by Cook (2003). Most fishing activities are not selective enough to remove from the ocean only the desired targets and will probably never be. This leads to accidentally catching other species (bycatch), part of which has little or no use (at least in the local context) and will be dumped overboard (as discards) together with the offal from fish processing at sea. The effect is to increase availability of food to scavenger species (including sea birds) and, when concentrated over time, may cause local anoxia of the seabed environment. The resulting amount of organic material may not be negligible. On the North Pacific shelf and upper slope, the amount of offal generated by at-sea production of surimi (a protein extract of fish) is very significant, since the technology used extracts less than 50% of the wet weight from the total catch, the rest being dumped. In the North Sea, 6.5 to 12.5% of the groundfish caught is dumped at sea. Some of this is consumed by sea birds, but a certain amount of offal becomes available to benthic scavengers. The increase in abundance of dogfish (Scyliorhinus canicula) in northern Spain fisheries and that of Raja radiata in Greenland shrimp fisheries has been associated with increased discards. Oxygen depletion due to excess organic loading from discards has been recorded in the New Zealand fisheries for hoki (Macruronus novaezelandie) as well as in the North Eastern Atlantic (Goñi, 1998).
Bycatch mortality also affects many non-fish species which are relevant to the functioning of the overall ecosystem. For example, surface and sub-surface driftnet and long-line fisheries have serious negative effects on populations of sea birds, e.g. albatrosses and petrels in long-line fisheries in the North Pacific and in the Southern Ocean. High seas drift nets have had a considerable impact on sea birds in the northern Pacific as have gillnets in southwest Greenland, eastern Canada and elsewhere.
Voluntary dumping or loss of fishing gear may lead to ghost fishing. The scale of the impacts of ghost fishing is basically unknown but there are indications that the effects are not negligible (Goñi, 1998). Species affected by discarded gear include not only teleost fish but sea birds, marine mammals and turtles. For example, the incidence of entanglement of marine mammals in floating synthetic debris in the Bering Sea has been related to growing fishing effort and increased use of plastic. Fowler (1987) concludes that entanglement is the principal cause of the current decline in the fur seal population of the Pribilof Islands, accounting for 15% of the mortality of youngsters. As an average, a northern fur seal (Callorhinus ursinus) is expected to encounter 3-25 pieces of net debris along the 800-km yearly migration in the Northeast Pacific. Fish traps, unless made of biodegradable material, contribute to the problem. To illustrate this problem, 31 600 pots were lost in the Bristol Bay crab fishery in a period of two years (Kruse and Kimber, 1993; Goñi, 1998).