Effects of the exxon valdez oil spill on habitat use by birds in prince william sound, alaska

Oil spills may affect species through direct effects on population size and structure and direct and indirect (toxicological) effects on reproduction. Spill effects on the habitats these organisms occupy have received less attention, but they are no less important. For 2.5 yr following the Exxon Valdez oil spill in Prince William Sound, Alaska, we studied the use of oil‐affected habitats by 42 species of marine‐oriented birds. On 11 survey cruises, we surveyed bays that had received different levels of initial oiling. We related the abundance of individual species in the bays to the oiling gradient, using regression models that included habitat measures to control for variations among the sites in features other than oiling level. We defined a spill‐induced impact as a statistically significant relationship between the abundance of a species and values along the oiling gradient, after accounting for the effects of variations in habitat features. We used among‐year comparisons of regressions between oiling levels and abundance, controlled for season, to assess recovery. We concluded that recovery from a spill‐induced impact had occurred when we no longer could detect a significant relationship between a species’ abundance and oiling levels. Overall, 23 (55%) of the 42 species exhibited no initial negative impacts on their use of oil‐affected habitats. Of the 19 species that did exhibit negative impacts, 13 (68%) showed evidence of recovery within 2.5 yr (the final survey in 1991). Six species (Horned Grebe, Red‐necked Grebe, Barrow’s Goldeneye, Bufflehead, Mew Gull, and Northwestern Crow) showed no clear evidence of recovery by our final survey. The proportion of species recorded on individual surveys that exhibited negative impacts at that time declined over the study, from 54% on the first survey after the spill in 1989 to 10% in late 1991. A principal components analysis revealed extensive ecological overlap between species that were negatively impacted in their use of oil‐affected habitats and those that were not. The six species that had not recovered by late 1991 tended to be intertidal feeders and residents, but these traits also characterized some species that did not exhibit initial impacts and some species that subsequently recovered from impacts. We detected no obvious ecological differences between species that suffered spill impacts on habitat use and those that apparently were not affected, or between impacted species that later recovered in their use of habitats and species that had not yet recovered. These results indicate that the Exxon Valdez oil spill had clear initial negative impacts on habitat use by nearly half of the species examined, suggesting substantial initial effects on habitat suitability for these species. These impacts persisted for <2.5 yr for most affected species. This rate of recovery in habitat use parallels the rapid recovery (usually <2 yr) documented for other oil‐affected communities (e.g., intertidal invertebrates, fishes, and birds) that have been studied in Alaska and elsewhere.