Adapting density surface models for complex survey areas: application to humpback whale (Megaptera novaeangliae) abundance and distribution estimation in Southeast Alaska

Abstract

Population abundance and distribution estimates are foundational to understanding and managing populations. Increasing accuracy and resolution of such estimates increases their utility for researchers and managers. Line transect surveys analyzed using both traditional design-based distance sampling (dbDS) and spatially explicit density surface modeling (DSM) are used across systems and taxa to estimate abundance and distributions of populations. DSM has the advantages of being able to increase the resolution of density estimates and evaluate the correlation between density and environmental factors as well as accepting data from a wider variety of surveys. In this thesis I used a soap film smoother to apply DSM in a more complex habitat than it has previously been used for. I compared abundance and distribution estimates from dbDS and DSM, using line transect survey data of humpback whales (Megaptera novaeangliae) in southeast Alaska, to explore and evaluate the applicability of DSM in extremely complex habitats. I found that DSM estimated distribution at a more useful higher resolution than dbDS. Abundance estimates from dbDS were approximately 50% higher than for DSM, and while further comparisons are needed to clarify relative accuracy, the abundance estimates are reasonable compared with other sources. The results demonstrate that DSM is applicable in habitats with extremely complex boundaries.