CHARACTERISING KRILL FISHERY DYNAMICS USING A RANDOM WALK MODEL

Krill fishery dynamics was analysed by using Fine-scale haul-by-haul data. Between-haul-distances were well described by a Levy Flight type random walk model. The inferred distribution pattern of fishable target concentrations showed differences between Subareas with targets in Subarea 48.1 being most sparsely distributed compared to Subareas 48.2 and 48.3. Commercially viable CPUE was estimated to be 12 tonnes/hour and 7 tonnes/hour in autumn in Area 48.1. Fishers seemed to be forced to make a number of long distant movements at the beginning of the fishing season, and autumn was suggested as optimal season since generally less searching seemed to be required due to higher target density in the fishing grounds.