Using a mark-recapture model to estimate beaching probability of seabirds killed in nearshore waters during the Deepwater Horizon oil spill

During a marine oil spill, injured birds often die on the water, some eventually washing ashore, but others becoming waterlogged and sinking or being scavenged before reaching the shoreline. Birds that disappear before they can be deposited on the shoreline are difficult to enumerate, but they commonly represent a large fraction of total oil spill-related mortality. As part of the process of quantifying the overall impact to seabirds resulting from the Deepwater Horizon oil spill, it was necessary to estimate the number of dead birds lost at sea. We conducted a study to estimate the beaching probability of birds that died at sea in the Gulf of Mexico in the areas most heavily used by seabirds and impacted by the spill. Using a mark-recapture analysis to derive the beaching probability from our field study data, we estimated that dead birds afloat at sea had about a 0.1414 probability of beaching in areas searched during the Deepwater Horizon oil spill. Sensitivity analyses of our model and requisite assumptions suggested that if our assumptions were violated, the "true" beaching probability could be anywhere between 0.11 and 0.16. These estimates are much lower than beaching probabilities estimated for seabirds killed during the Exxon Valdez oil spill in the waters of Alaska, for example, likely reflect higher rates of decomposition and scavenging in the warmer waters of the Gulf of Mexico. Our estimate suggests that bird carcasses that washed onshore during the Deepwater Horizon oil spill represented only 14% of those killed at sea during the spill.