Fitting Euphausia superba into Southern Ocean food-web models: a review of data sources and their limitations
This paper aims to provide the overview needed to include krill in food-web models and to guide modellers to key sources of data. It describes the strengths of each method of sampling krill, i.e. with nets (for historical time series, demographic information and live krill), acoustics (distribution, time series, biomass and swarm-scale information), the fishery (sustained sampling in one place and wide area and time coverage) and via predators (long time series, demographic indices). Each data source has caveats and more efforts to combine them are recommended. Observations that krill occupy the underice layer, the 0–10 m layer, the deeper water column and the benthos have fundamental implications, both for assessing biomass and for modelling the food web. Temporally, the intense (order of magnitude) interannual variability in krill population size within the southwest (SW) Atlantic sector is a major scale of variability, driven by sea-ice and climate effects on recruitment. This variability masks top–down predation controls that may operate over multi-decadal scales. Growth in spring, summer and autumn is now fairly well quantified, but mortality remains an enigma. We are still not yet confident which are the major predators of krill but studies increasingly suggest that they are not currently birds or mammals. Krill feed across three trophic levels and can control food populations through locally high grazing impact and nutrient regeneration. They also have fundamental regional differences in overwintering strategies, on-shelf/off-shelf distributions, relationships with sea-ice and diet. Whether this reflects ‘subpopulations’ with regionally specific life cycles is still unclear. However, caution is urged when scalingup food-web models and their parameterisations, either from individual to schooling krill, or from one region to another.
Authors: Atkinson, A., Nicol, S., Kawaguchi, S., Pakhomov, E., Quetin, L., Ross, R., Hill, S., Reiss, C., Siegel, V., Tarling, G.