MEPS 275:275-287 (2004)  -  doi:10.3354/meps275275

ABSTRACT: Line-transect estimators were developed to assess abundance of coastal dolphins Tursiops truncatus and Stenella coeruleoalba encountered in low densities during aerial sighting surveys. The analysis improved on conventional approaches by objectively combining data from different species, survey areas and other covariates affecting dolphin detectability. Model selection and multimodel inference allowed robust estimates of precision in accounting for covariate selection uncertainty. These methods were used to estimate bottlenose dolphin abundance in NE Mediterranean waters that included a putative subpopulation in the Balearic Islands. Total abundance was estimated as 7654 (coefficient of variation, CV = 0.47; 95% CI = 1608 to 15766) and the abundance in inshore waters of the Balearic Islands varied from 727 (CV = 0.47; 95% CI = 149 to 1481) dolphins in spring 2002 to 1333 (CV = 0.44; 95% CI = 419 to 2617) dolphins in autumn 2002, with an average estimate of 1030 (CV = 0.35; 95% CI = 415 to 1849). The results do not support an exclusively coastal Balearic Island subpopulation, but they strongly indicate that the islands contain critical habitats required for the conservation of the species. Given the observed decline of the species during the last few decades, conservation-oriented management should focus on reducing or eliminating adverse fishing interactions while key areas are protected from encroachment produced by human development.

KEY WORDS: Abundance · Conservation · Aerial survey · Availability bias · Balearic Islands · Balearic Sea · Bottlenose dolphin · Covariate analysis · Line transect · Tursiops truncatus