Fishery selection and its relevance to stock assessment and fishery management

• Fishery selection (selectivity) measures relative fishing mortality-at-age.
• Fishing gear and the locations of the fish and the fishing influence selectivity.
• VPA estimates indicate diverse selectivity curve shapes and temporal variability.
• Selectivity influences fishery performance and stock assessments.

Fishery selection (selectivity for short) is the term often used to describe the phenomenon whereby a fish stock experiences mortality due to fishing that is age- or size-specific. Selectivity operates both at a local scale, as in the direct interactions of individual fish with the fishing gear (contact selection), and at a stock-wide scale (population selection), as evidenced by the differential rates of fishing mortality-at-age that are generally observed in stock assessment results. All age-structured stock assessment models have some form of fishery selection to modulate the impact of fishing mortality on differing age-classes, but selection coefficients, from a stock assessment viewpoint, generally are nuisance parameters rather than a focus of attention. This paper provides an overview of the three main processes that contribute to and influence population selection: (1) physical sorting by the fishing gear or differential behavioral responses of the fish to the gear produce the phenomenon of contact selection; (2) differing selection properties of different types of fishing gear (e.g., trawl versus longline) in turn generate a composite selection curve that is a weighted average of the different kinds of contact selection; and (3) when the fish are not well mixed spatially, then the spatial distribution of fishing relative to the spatial distribution of the fish also affects population selectivity. Fishing mortality-at-age estimates derived from a published Virtual Population Analysis of Scotian Shelf haddock are used to illustrate the diversity of shapes that can be seen in population selection curves and their considerable temporal variability. A spatial model for fishery age-selectivity is then used to demonstrate that the maximum relative yield harvested from a stock can be a function of both contact selection and the spatial distribution of fishing.