Validation of satellite imagery for quantitative population assessment and monitoring of Atlantic walrus

Abstract

Monitoring animal population abundance and distribution is essential for conservation and management. Recent advances in very high-resolution (VHR) satellite imagery allow for unprecedented detection and identification of animals, offering a potential alternative to conventional methods of population assessment such as aerial surveys. Here, we compare satellite and aerial survey images of a terrestrial walrus haul-out site in the Canadian High Arctic taken 1.5 h apart. While the 30-cm resolution of the commercially available satellite image was insufficient to reliably count individual walruses visually, a manual digitization method successfully identified and delineated image areas confirmed visually to have walrus. Estimates of numbers of hauled out walrus obtained by multiplying the detected area by a range of densities in the aerial image (231) agreed well with actual counts in 35 and 55 mm aerial images (217–240). Potential applications of this approach include quantitative abundance estimation or more simple monitoring of presence/absence over large expanses of remote areas that are logistically difficult to cover using survey aircraft.