Antarctic krill Euphausia superba are key to most Antarctic ecosystems, particularly at South Georgia, where penguins and seals rely on them for their breeding success each summer. Changes in krill behaviour and local-scale distribution during this time can have large consequences for these predators; however, we currently have little understanding of the behaviour of Antarctic krill in response to changing conditions of predation and food. We used a stochastic dynamic programming model to

more »

... tigate the behaviour of krill within a region close to South Georgia that is known to be heavily foraged by penguins and seals. In the model, krill responded to changing conditions by adjusting their depth, density of swarm and swimming behaviour. We have shown the optimal behaviour of krill in 3 biologically distinct regions: the on-shelf region, shelf-break region and off-shelf region. We predict significantly higher concentrations of krill will result at the shelf-break region from krill choosing to swim slower and turn more often in a favourable zone. In addition, we predict a diel pattern in swarm density in most conditions of the model, with small krill generally forming lower density swarms than large krill, particularly on-shelf. This work is the first prediction of the effects of krill swarming and swimming behaviour on local-scale distribution. KEY WORDS: Euphausia superba · Behaviour · DVM · Swarming · SDP · Advection · Stochastic dynamic programming · Mortality Resale or republication not permitted without written consent of the publisher