Estimates of Detection Probability for BWASP Bowhead Whale, Gray Whale, and Beluga Sightings Collected from Twin Otter and Aero Commander Aircraft, 1989 to 2007 and 2008 to 2011.

摘要

The Bowhead Whale Aerial Survey Project (BWASP) database is a long-term (1979-2013) database on the distribution, relative abundance, and behavior of marine mammals in the northern Bering, northeastern Chukchi, and western Beaufort seas. Data from the BWASP database are analyzed to produce separate detection function models using multiple covariates distance sampling methods for each combination of species (bowhead whale (Balaena mysticetus), gray whale (Eschrichtius robustus), and beluga (Delphinapterus leucas)), aircraft (Twin Otter and Aero Commander), and time period (1989-2007 and 2008-2011) in the northeastern Chukchi and western Beaufort seas. The line-transect survey protocols used to collect the data have been remarkably consistent across the years; however, changes in protocol that could affect detection probability are discussed and accounted for in the analyses. Potential covariates evaluated for inclusion in the detection function models included group size, depth of the sea floor at the location of the sighting, Beaufort Sea State, longitude of the sighting, sea ice percentage, and observer effects. Group size variables tended to increase effective strip width (ESW) in the bowhead whale models for Twin Otters, and sea state and sea ice variables decreased ESW for beluga models for Twin Otters. ESW decreased with increasing depths and more westerly longitudes for some bowhead whale and beluga models; this result could be due to spatial heterogeneity in either animal behavior or unmeasured or unmodeled environmental conditions. Beaufort Sea State positively affected ESW for bowhead whale models for Twin Otters in the early years and gray whale models for Aero Commanders; this relationship is difficult to explain.