Distribution, density and relative abundance of Antarctic krill estimated by maximum likelihood geostatistics on acoustic data collected during commercial fishing operations

Abstract

There is a substantial harvest for Antarctic krill in the Southern Ocean, but little regular scientific monitoring of the resource. Recently, however, the Commission for the Conservation of Marine Living Resources (CCAMLR) has initialised a process to make use of acoustic data from commercial fisheries to increase the amount of relevant information available for making management decisions. We here provide an example where 34 days of acoustic data, collected during commercial krill fishing operations on the vessel ‘Saga Sea’ were processed to produce probability of presence, conditional density and relative abundance estimates on monthly, weekly and daily basis. Data were analyzed using a maximum likelihood time-series and geostatistical approaches, selected to account for the lack of sampling design, and likely correlation in space and time. The applied method showed low sensitivity of monthly estimates to different repeated measure criteria and location sub-settings. Most weekly estimates, but the last one, were also consistent with the full data (monthly) estimate. Highly variable and lower estimates were obtained, however, from daily data sets. Although our results suggest the method had provided an adequate treatment for time and space correlation, we were not able to evaluate potential bias due to preferential sampling of high density krill aggregations and/or limited area coverage within short time periods. The results suggest that this method, combined with some additional design based coverage by the fishing vessels, can be useful to obtain quantitative evaluations of krill density and distribution for management purposes.