Flexible preference of southern elephant seals for distinct mesoscale features within the Antarctic Circumpolar Current

The open ocean is a highly variable environment where marine top predators are thought to require optimized foraging strategies to locate and capture prey. Mesoscale and sub-mesoscale features are known to effect planktonic organisms but the response of top predators to these features results from behavioural choices and is poorly understood. Here, we investigated a multi-year database of at-sea distribution and behaviour of female Southern elephant seals (Mirounga leonina) to identify their preference for specific structures within the intense eddy field of the dynamic Antarctic Circumpolar Current (ACC). We distinguished two behavioural modes, i.e. travelling and intensive foraging, using state-space modelling. We employed multisatellite Lagrangian diagnostics to describe properties of (sub-)mesoscale oceanic circulation. Statistical analyses (GAMMs and Student's t-tests) revealed relationships between elephant seal behaviour and (sub-)mesoscale features during the post-moulting period (January-August): travelling along thermal fronts and intensive foraging in cold and long-lived mesoscale water patches. A Lagrangian analysis suggests that these water patches - where the prey field likely developed and concentrated - corresponded to waters which have supported the bloom during spring. In contrast, no clear preference emerged at the (sub-)mesoscale during the post-breeding period (October-December), although seals were distributed within the Chlorophyll-rich water plume detaching from the plateau. We interpret this difference in terms of a seasonal change in the prey field. Our interdisciplinary approach contributes to elucidate the foraging strategies of top predators in a complex and dynamic environment. It also brings top down insights on prey distribution in remote areas where information on mid-trophic levels are strongly lacking and it identifies important physical-biological interactions relevant for ecosystem modelling and management. (C) 2014 Elsevier Ltd. All rights reserved.