Fish behaviour during sampling as observed from free floating buoys - application for bottom trawl survey assessment

Varying fish distribution and behaviour during bottom trawl surveys has long been considered important for the reliability of survey abundance estimates. Size and species dependent behaviour may bias the estimates and the problem may be augmented by within and between survey variation in the natural conditions. Qualitative descriptions of behavioural effects are numerous, but a quantitative methodology driven by observations is still lacking. In this paper we look into the application of data from a newly developed acoustic buoy based on experiments from saithe (Pollachius virens) off northern Norway. Time series plots of collective as well as individual fish behaviour during vessel passage show avoidance reaction. Time series plots of vertical fish velocity vectors show periods in which both speed and direction appear to be random alternating with periods with clear synchronous co-ordinated movement even when the fish are not simulated by vessel noise. The data uncover substantial variability in natural fish behaviour which is an obstacle for drawing firm conclusion as well as for modelling vessel/trawl affected behaviour for use in survey stock assessment. Bergen Acoustic Buoy (BAB) is an improved sampling tool, which has the potential to supply the data needed for resolving these problems