Quantifying fish escape behaviour through large mesh panels in trawls based on catch comparision data – model development and a case study from Skagerrak In: ICES (2012) Report of the ICES-FAO Working Group on Fishing Gear Technology and Fish Behaivour (WGFTFB), 23-27 April 2012, Lorient, France. ICES CM 2012/SSGESST:07

Based on catch comparison data, it is demonstrated how detailed and quantitative information about species-specific and size dependent escape behaviour in relation to a large mesh panel can be extracted. A new analytical model is developed, applied, and compared to the traditional modelling approach for such data. As a case study, we used data collected with a twin trawl setup. The only difference between the two 120 mm trawls was that a 12 meter long section in the upper panel was replaced with 800 mm diamond meshes (LMTP) in one of them. Based on this very large mesh size, we assumed that all individuals that contacted the panel also escaped through it. The new analytical method was applied to quantify escape behaviour for cod, haddock, saithe and Nephrops of different sizes. There was a need to include the full gear selec-tivity in the trawl, and we show how this selectivity can bias the interpretation of the length based escapement behaviour over the large mesh panel. Our length based behavioural description is in good agreement with direct observations of the same species in the trawl cavity reported in literature.
Fish behaviour understanding is essential. Observations are often difficult using opti-cal devices such as cameras. The alternative is to use catch data to reconstruct behav-iour. Every fish in every haul counts. Bootstrapping can be used. An experimental catch comparison index was calculated. Length frequency distributions and catch comparison rates are not suitable to infer behavioural patterns. A full gear selectivity model was developed in which data of low lengths was deleted and double boot-strapping is used. Catching is a sequential process. Panel contact was assumed lead-ing to escape. Flounder shows strong length dependent escape behaviour. Curves were presented for COD, HAD, LEM, POL, and WTH. Comments made on knife-edge curves found for cod using stochastic simulation, apparently caused by data weakness. LMTP affects fish but not Nephrops