From swarming behaviour to trophic interactions: forecasting dynamics of Antarctic krill in ecosystem hotspots using behaviour-based models

The fishery for Antarctic krill in the Southern Ocean is currently being carried out in very small areas. These areas seem to be preferred by both fishermen and natural krill predators (whales, penguins, seals). We have gradually gained some understanding of which areas are such "hotspots", but we lack knowledge of why, in other words: what is special about hotspots? ?Hotspot? is also not a well-defined term: we know that both fishermen and krill predators concentrate in these areas, but is this because there is a lot of Antarctic krill there, because the krill are easily accessible there, or for completely different reasons? The project "From swarming behavior to trophic interactions: forecasting Dynamics of Antarctic krill in ecosystem hotspots using behavior-based models." has studied the interaction between krill behavior and the physical environment, and has used data from echo sounders and current meters and physical models to analyze how concentrations and distribution of krill vary with current conditions, ice and other factors. The aim has been to use new knowledge and models to improve the management of krill resources: if current conditions lead to constantly new krill being added to the "hotspot", the limit for local extraction may be increased, if the amount of krill in the area is more determined by behavior local extraction may have to be curtailed. The project has built on HI's existing krill monitoring efforts in the area, by supplementing with data from acoustic rigs deployed for longer periods. Data from a hydrodynamic model that has been used show that the transport of krill into the area is strongly influenced by whether the krill use sea ice as a habitat or not. From the observation platforms, we have direct targets for transport of krill into the area, since we measure both current speeds and the amount of krill. However, the majority of Antarctic krill occur in large shoals, so there is a large natural variation in our estimates of krill transport into the area. In a normal year, HI conducts resource mapping in the area for approx. 2 weeks in the summer, thanks to boat time donated by the fishing fleet. Continuous monitoring time-series, similar to the one the SWARM project has worked up, provide a different perspective on processes in the area, and information on, among other things. the depth of the krill swarms throughout the year, how common and how large they are, etc., are important for understanding how representative the results from the resource mapping efforts are, and for the interpretation of these results. Analyzes carried out in the project show that although most of the krill occur in shoals, in fact Antarctic krill is almost always present, but most often occurs in low densities. Although this krill is hardly interesting to fishermen, it is probably important for the natural krill predators, most predators do not depend on large amounts of krill per predator. Continuous data series also give us occasional surprises: some of the largest observed changes in krill biomass in the area did not occur due to horizontal, but vertical migrations of krill: on several occasions large quantities of krill swam suddenly into the deep sea. Antarctic krill at great depths have been described before, but how the actual migration up or down takes place has been unknown until now. Our observations show that krill predators show the highest activity in the area during the summer. This applies to diving predators (e.g. penguins and seals), but also fish, which we previously thought were scarce in the area. Both components occur far more frequently in observations from the summer. If the commercial fishing for krill has any impact on the system, e.g. leads to a reduction in the amount of krill available to natural predators, then it is therefore far more likely that this will happen in summer than in winter. The observation platforms have given us a better overview of the krill's behavior, and we have identified several processes that can contribute to the high biomasses in "hotspots". Common to these processes is that they depend on both the physical environment and the krill's own behavior.

Prosjektets samfunnsnytte ligger i første omgang i å bidra til bedre overvåking og forvaltning i Antarktis. Det meste av resultater som har blitt presentert fra prosjektet frem til nå har vært direkte rettet inn mot EMM, og ansees som relevant for overvåkning og forvaltning av marine økosystemer i Antarktis. I tillegg til å videreutvikle og forbedre HIs kompetanse og evne til å sette ut og hente inn, og analysere dataserier fra, (primært) akustiske observatorier i Antarktis, har prosjektet styrket HIs (og CCAMLRs) evne til å overvåke marine ressurser rundt South Orkney Islands. 2 master oppgaver (Universitet i Oslo, Aix-Marseille University) er avlevert basert på materialet som har blitt samlet inn, i tillegg til en bachelor oppgave (DTU, Danmark).

Understanding the formation, maintenance and distribution of krill aggregations in relation to the environment is essential to comprehension of many Antarctic marine ecosystems, their function and their sensitivity to climate change and human impacts. This is particularly important as the currently expanding fishery(in terms of catches) is increasingly geographically confined to local hotspots. This project aims I) to build firmer ecological and mechanistic explanations for the distribution and dynamic patterns of krill swarms in ecosystem hotspots and ii) to use the acquired knowledge to develop a predictive behavioural-based theoretical framework of krill aggregative behaviour and distribution. Our multidisciplinary project will merge insights from advanced research in collective behaviour, behavioural ecology, marine ecology, oceanography, hydrodynamic modelling and acoustics to increase the biological and ecological understanding of the effects of environmental fluctuations and fisheries on an important aspect of krill ecology. The strength and novelty of our approach resides in the development of models of krill optimal distribution and swarming behaviour, by combining observations using state-of-the-art acoustic technology with hydrodynamic models and models of krill behaviour. Since our approach integrates historical survey and fisheries information, continuous acoustic monitoring from stationary platforms and oceanography models to a simulation-based behavioural framework, we expect to acquire important insights about ecological processes at both small and larger spatial and temporal scales. Such insights and predictive tools are essential for a sound implementation of the Feedback Management system for Antarctic krill, which is currently being developed under the auspices of the Commission for the Conservation of Antarctic Living Resources (CCAMLR).