Bird mortality due to the Deepwater Horizon oil spill: Comment on Haney et al. (2014a, b)

摘要

Haney et al. (2014a, b; Mar Ecol Prog Ser 513:225-237, 239-252) developed probability models to estimate seabird mortality from oil exposure during the Deepwater Horizon oil spill. Although frequently used to characterize avian mortality following oil spills, probability models often yield uncertain results when developed without spill-and/or region-specific data. Models based on observations of beached carcasses or exposure/mortality scenarios are sensitive to variations in assumptions and methods used to summarize data sets for model parameterization and validation. Here we present alternative parameter estimates derived from spill-and Gulf of Mexico (GoM)-specific data, and offer suggestions for reducing model uncertainty. As a primary example, we evaluate the carcass transport probability to shorelines using GoM-specific data collected in 2011 to show that Haney et al. underestimated this probability by more than an order of magnitude, thus inflating mortality estimates.