Quantifying spatial characteristics of the Bowhead Whale Aerial Survey Project (BWASP) survey design

Abstract

The Bowhead Whale Aerial Survey Project (BWASP) has been conducted annually since 1979 in the Alaskan Beaufort Sea to monitor the distribution and relative abundance of the Bering-Chukchi-Beaufort (BCB) stock of bowhead whales (Balaena mysticetus) during their autumn migration. BWASP was created to specifically address broad-scale research and management questions related to bowhead whale ecology, with particular interest in the potential effects of oil and natural gas exploration, development and production activities on the BCB bowhead whales. With elevated concerns about climate change, increasing oil and gas activities and the forecasted increase in vessel traffic, it is expected that interest in the BWASP dataset will also increase in order to evaluate effects of these anthropogenic activities on BCB bowhead whales and indigenous whaling. The following analysis quantified the spatial characteristics of the BWASP survey design and provided guidelines for the types of investigations that the BWASP data can potentially address. Sampling lags (transect spacing) in the BWASP survey design of approximately 20km along the east/west axis of the study area limit the spatial scale of phenomena that can be detected using data from a single BWASP survey. Therefore, BWASP data are relatively uninformative for studying variability in distribution or relative abundance along the east/west axis over short time scales (one survey) and within small areas measuring less than approximately 20km. In addition, computer simulations showed spatial heterogeneity in the long-term survey coverage probability (the probability that a given location will be included in a survey having an assumed effective search width under the BWASP survey design). Pooled transects created from simulated surveys resulted in a repeating diamond pattern in which coverage probability was low. Analyses incorporating data from many BWASP surveys should account for this spatial heterogeneity, via either the survey coverage probabilities or quantification of survey effort; otherwise estimates of variables such as relative density, density, or habitat use may be biased. The BWASP surveys have increased understanding of the broad-scale patterns of bowhead distribution, relative abundance and behaviour. The utility of this dataset in informing other questions will depend upon the scale of the ecological phenomena under investigation and the analytical scales used to address the questions.