Assessment of relative performance of a square-mesh codend design across multiple vessels in a demersal trawl fishery

Abstract

Implementing gear-based solutions to reduce bycatch in trawl fisheries can be difficult owing to fishers’ concerns over the fleet-wide applicability of results from discrete, highly controlled experiments associated with the development of such innovations. In New South Wales, Australia, a codend comprising a posterior section made from 35-mm square-shaped mesh (i.e. hung on the bar) was developed and subsequently recommended for partial implementation in a multi-species demersal trawl fishery (i.e. for use when eastern king prawn –Melicertus plebejus – is the primary target species). However, industry concerns over this design have to date prevented its implementation. Here, using direct twin-tow comparisons onboard a subset of the fleet, we quantitatively assessed the performance of the square-mesh codend, relative to the range of diamond-mesh codend designs currently being used, in terms of reductions in bycatch rates and shortfalls in catch rates of retained animals. Generalised linear mixed-effects modelling across all vessels showed that when eastern king prawn was the target species, the square-mesh codend significantly reduced bycatch rates (by ∼48%), but with an accompanying significant shortfall in catch rates of eastern king prawn (∼7%) and a substantial (though not statistically significant) shortfall in catch rates of an important component of the retained byproduct (whiting, Sillago spp., ∼36%). Partitioning of data by vessel revealed that such shortfalls in retained catch were evident for only some vessels, while the square-mesh codend performed as intended for the others. When school prawn (Metapenaeus macleayi) was the target species, substantial shortfalls in catches of school prawn (∼54%) in the 35-mm square-mesh codend confirmed that a codend of a smaller square mesh might be a more appropriate design under those circumstances. Potential strategies to address intra-vessel inconsistencies in performance of the square-mesh codend and thereby facilitate implementation of square-mesh codends in the fishery are discussed.