Collective behaviour of penned herring: Observing the collective behaviour and investigating the effect of various sound stimuli

The primary objective is to understand the interaction between direct stimuli and information transfer between individuals in schooling fish and to understand how this affects the collective behaviour, with special emphasis on noise induced behaviour. In 2012 we further developed methodology to observe fish behaviour from multibeam sonars, and it resulted in a paper in Current Biology. The paper received good coverage, with a dispatch written by Graeme Ruxton and referred to in a feature by Michael Gross. We have also written a news article on our web site, in addition to coverage through Princeton University webpages. In 2013-2014 we further developed these techniques. We are now able to extract fine scale behavioural metrics of schooling fish in net pens, enabling us to study how noise and other stressors are affecting the school structure. This approach have formed the basis of new grant applications with the objective of assessing the impact on various stressors on schooling fish. We have also developed the instrument platfrom further enabling it to be deployed from research vessels at sea. Noise propagation modelling have been carried out by the NTNU team, and we have used their hydrophone array when assessing the exposure in our net pen experiments. We have published a paper on how noise propagation models can be used to interpret the responses in fish from acoustic surveys, and we have several conference proceedings from this work on low frequency propagation models. The main net-pen experiments, including testing of the various sound sources, was conducted summer of 2012 in Austevoll. We tested the collective responses in herring to various stimuli, including playback of naturally occurring predator calls, vessel noise and various synthesized sounds. We had several guests during our experiments, bringing important competence to our project team. The tv-channel ARTE was making a documentary on Underwater noise and filmed and did interviews during our experiments. The experiments have so far resulted in 4 published papers. Spring 2013 the postdoc in the project stayed 3 months in the US working with colleagues at Princeton University and Florida International University. The PI also joined for a shorter duration. This resulted in a strengthened collaboration with the international partners. In Florida we did field work on fish distribution and collective behaviour in the field, which resulted in 2 papers. Summer 2013 the second net pen experiments was conducted. The data from this experiment is processed and we show how different stress levels in the fish alter the strength of the collective responses in herring. Combined with the paper on density dependent effects of the collective behaviour, this shows that the response in traditional stimulus-response experiments are heavily influenced by density and level of vigilance. The method was also used to assess how killer whale playback affected schooling structure in herring. To link the response seen in the net pens to the behaviour of schools in the open ocean, we participated on a fishery technology development survey, where we stimulate the herring using objects shot through the schools while simultaneously observing the behaviour using multibeam sonar. To aid interpretation of the data, we have developed a model to predict acoustic backscatter as a function of fish behaviour. This work is published in the Journal of Acoustical Society of America, and will provide the foundation for bridging the gap between the open ocean and the net pen experiments. This work is being extended to a method that can track behavioural responses that travels as waves in large schools. Using 4D sonars, we can now test the hypotheses developed in the net pen experiments in situ.

Collective behaviour is an intriguing pattern-forming process in nature, and how individual rules scale to collective phenomena are still being learned. An area of particular interest is to understand how information is transferred in animals groups, and how stimuli is amplified or dampened through social interactions. The project will perform lab experiments on schooling fish responses to external stimuli, and investigate the role of the social interactions. The fish will be observed using video cameras , and techniques to extract the behavioural information will be further developed. An important aspect of the proposal is to establish a link between the highly controlled laboratory observations to in situ experiments. The results from the laboratory w ill be used to form rigorous hypotheses on the collective behavioural rules in herring, a key species both economically and ecologically. These will be tested on herring in a net pen setup. We will use both video cameras and acoustic cameras to observe th e behaviour, and the methods developed from the laboratory experiments will be refined to fit the herring net-pen system. We are particularly interested in sound stimulated collective behaviour, since anthropogenic noise pollution is an increasing concer n worldwide. To achieve this a range of different sound sources will be used, and the different noise fields will be modelled and verified before being exposed to the fish. The role of near field effects and range effects will be investigated, and the res ults will be related to in situ sound exposure experiments, in particular where sound pressure levels has failed to explain the behaviour. Finally, the expected acoustic image on a 4D scientific multi-beam sonar will be predicted from simulated schools. These predictions will be based on the experimentally derived rules, and the result will be used to evaluate the feasibility to use sonars to observe internal school dynamics in the open ocean.