Effective Strip Widths for Ship-Based Line-Transect Surveys of Cetaceans

摘要

Effective strip width is a key parameter in estimating abundance and density from linetransect surveys. Here we estimate effective strip widths for 58 categories (genera, species, subspecies, stocks, or other groups) of cetaceans based on 13,500 sightings from 32 line-transect surveys conducted in the eastern Pacific Ocean by the Southwest Fisheries Science Center from 1991 to 2008. Generalized linear models (GLMs) are used to first identify factors that are important in determining the perpendicular sighting distances using stepwise model selection based on AIC. Six species groups of similar taxa are created and modeled separately. Important factors for most species groups include Beaufort sea state, swell height, visibility, group size, species, and a survey-specific categorical variable (Cruise No.). Interactions between species and the other factors generally do not improve GLM fits, indicating that the effects of those factors are relatively consistent for species within a species group. Factors selected for the best-fit GLMs are included as potential covariates in a line-transect model fit to a subset of the same data, again using stepwise model selection based on AIC. The best-fit line-transect models do not include Cruise No. and are generally simpler than the GLMs, likely because distant sightings were eliminated by truncation. Species-specific differences in ESW are seen within the species groups indicating that species of cetacean do differ in the distances at which they can be detected, even after accounting for the effects of group size and other covariates that affect sighting distances. Results from this analysis of multiple surveys can be used to improve estimates of effective strip widths for any survey using the same methods and similar vessels. This is especially true for seldom-seen species whose abundance is difficult to estimate from a single survey.