Sunday 27 February 2011

The Exxon Valdez Oil Spill in Alaska: Acute, Indirect and Chronic Effects on the Ecosystem

Peterson, CH
Advances in Marine Biology [Adv. Mar. Biol.]. Vol. 39, pp. 3-84. 2000. 

Following the oil spill in Prince William Sound, Alaska, in 1989, effects were observed across a wide range of habitats and species. The data allow us to evaluate direct and indirect links between shoreline habitats and the coastal ecosystem in general. The intertidal zone suffered from direct oiling and clean-up treatments such as pressurized hot water, resulting in freeing of bare space on rocks and reductions in fucoid algal cover. Grazing limpets, periwinkles, mussels and barnacles were killed or removed. Subsequent indirect effects included colonization of the upper shore by ephemeral algae and an opportunistic barnacle and, in some regions, spread of Fucus gardneri into the lower shore where it inhibited return of red algae. The loss of habitat provided by the Fucus canopy slowed recovery on high shores, and lowered abundance of associated invertebrates. Abundance of sediment infauna declined and densities of clams were reduced directly. Their recovery was still incomplete by 1997 on oiled and treated shores where fine sediments had been washed down slope during treatment. Impacts in subtidal habitats were less intense than in the intertidal zone. Kelps were reduced in 1989 but recovered rapidly through re-colonization by 1990. Abundances of a dominant crab and seastar were reduced greatly, with recovery of the more mobile species, the crab, occurring by 1991. For about 4 years, there was reduced eelgrass density and hence less habitat for associated animals. Abundance of several toxin-sensitive amphipods declined dramatically and had not recovered by 1995. In general, however, many subtidal infaunal invertebrates increased in abundance, especially oligochaetes and surface deposit-feeding polychaetes. This may have resulted from increases in sediment hydrocarbon-degrading bacteria, but may also reflect reduction of predators. Along northern Knight Island, where sea otter populations had not recovered by 1997, green sea-urchins were larger, compared with those in un-oiled parts of Montague Island. This initial response from reduced predation by sea otters, if sustained, could lead to additional indirect effects of the spill. Scavenging terrestrial birds, such as bald eagles and northwestern crows, suffered direct mortality as adults and reproductive losses, although eagles recovered rapidly. Numbers of intertidal benthic fishes were 40% lower on oiled than on un-oiled shores in 1990, but recovery was underway by 1991. Small benthic fishes living in eelgrass showed sensitivity to hydrocarbon contamination until at least 1996, as evidenced by hemosiderosis in liver tissues and P450 1A enzyme induction. Oiling of intertidal spawning habitats affected breeding of herring and pink salmon. Pink salmon, and possibly Dolly Varden char and cut-throat trout, showed slower growth when foraging on oiled shorelines as older juveniles and adults, which for pink salmon implies lower survival. The pigeon guillemots that suffered from the oil spill showed reduced feeding on sand eels and capelin, which may also have been affected by the spill, and this may have contributed to failure of guillemot recovery. There was an analogous failure of harbor seals to recover. Sea otters declined by approximately 50%, and juvenile survival was depressed on oiled shores for at least four winters. Both black oystercatchers, shorebirds that feed on intertidal invertebrates, and also harlequin ducks showed reduced abundance on oiled shores that persisted for years after the spill. Oystercatchers consumed oiled mussels from beds where contamination by only partially weathered oil persisted until at least 1994, with a resulting impact on productivity of chicks. A high over-winter mortality of adult harlequin ducks continued in 1995-96, 1996-97 and 1997-98. Delays in the recovery of avian and mammalian predators of fishes and invertebrates through chronic and indirect effects occurred long after the initial impacts of the spill. Such delayed effects are not usually incorporated into ecotoxicity risk assessments which thus substantially underestimate impacts of a spill. Detection of delayed impacts requires