Advancements in methods for estimating the abundance of marine megafauna using novel sampling techniques

Abstract

Population size (‘abundance’) is a key parameter for wildlife conservation. As technology advances, new sampling techniques become available for which established estimation methods are extended and new methods developed. We explore, extend, and test new methods for estimating abundance of large marine megafauna that utilise novel sampling techniques: passive acoustic monitoring (PAM), modern genetic analyses, and baited video systems.

The PAM study is motivated by acoustic recordings of bowhead whales (Balaena mysticetus) from multiple sensors in the Beaufort Sea. A high proportion of detections on only one sensor (‘singletons’) are false positives, resulting from the automated detection algorithm. We extend acoustic spatial capture recapture methods to enable singletons to be discarded without biasing the abundance estimate. In our model, detection probability is dependent on received sound level, but we also propose an extension where detection probability is a function of signal-to-noise ratio, enabling the use of detections at lower received levels.

In the genetic study, we test the recently-developed close-kin mark-recapture (CKMR) method in the situation where the ages of sampled animals are estimated with error. Our work is motivated by a case study of grey reef sharks (Carcharhinus amblyrhynchos) at Palmyra Atoll, Pacific Ocean. CKMR relies on knowing animal age, and for many species this is inferred from measured length. We use simulation to test the effects of misspecifying the age-length relationship (‘growth curve’) and the amount of length measurement error on abundance estimates. A misspecified growth curve produces substantial bias while misspecifying measurement error does not.

Lastly, we explore the potential of extending random encounter modelling (REM), which is used to estimate population size from remote wildlife cameras, for the situation where baited cameras are used as is common in shark studies. We believe that potential exists for REM with baited cameras if bait plume intensity is considered.