Designing an effective mark–recapture study of Antarctic blue whales

To properly conserve and manage wild populations, it is important to have information on abundance and population dynamics. In the case of rare and cryptic species, especially in remote locations, surveys can be difficult and expensive, and run the risk of not producing sample sizes large enough to produce precise estimates. Therefore, it is crucial to conduct preliminary analysis to determine if the study will produce useable estimates. The focus of this paper is a proposed mark–recapture study of Antarctic blue whales (Balaenoptera musculus intermedia). Antarctic blue whales were hunted to near extinction up until the mid-1960s, when commercial exploitation of this species ended. Current abundance estimates are a decade old. Furthermore, at present, there are no formal circumpolar-level cetacean surveys operating in Antarctic waters and, specifically, there is no strategy to monitor the potential recovery of Antarctic blue whales. Hence the work in this paper was motivated by the need to inform decisions on strategies for future monitoring of Antarctic blue whale population. The paper describes a model to predict the precision and bias of estimates from a proposed survey program. The analysis showed that mark–recapture is indeed a suitable method to provide a circumpolar abundance estimate of Antarctic blue whales, with precision of the abundance, at the midpoint of the program, predicted to be between 0.2 and 0.3. However, this was only if passive acoustic tracking was utilized to increase the encounter rate. The analysis also provided guidance on general design for an Antarctic blue whale program, showing that it requires a 12-year duration; although surveys do not necessarily need to be run every year if multiple vessels are available to clump effort. Mark–recapture is based on a number of assumptions; it was evident from the analysis that ongoing analysis and monitoring of the data would be required to check such assumptions hold (e.g., test for heterogeneity), with the modeling adjusted as needed.