Direct and Indirect Estimation of Gillnet Selection Curves of Atlantic Herring (Clupea harengus harengus)

Abstract

Length-specific selection curves for Atlantic herring (Clupea harengus) were calculated for a series of gillnets ranging in mesh size from 50.8 to 76.2 mm (stretched measure) using Holt's (1963) model (ICNAF Spec. Publ. 5: 106–115). These curves were than compared with direct estimates of length-specific selectivity obtained from a comparison of gillnet catch length frequencies with population length composition data as determined from acoustic surveys. Selection curves calculated indirectly using the Holt model were unimodal and congruent. The empirical selection curves however were multimodal and fishing power varied with mesh size. These differences in selectivities were due to the fact that herring were caught not only by wedging at the maximum girth but also at other body positions such as the gills and snout. Each of these modes of capture have different length-specific selectivity characteristics and, since the relative contributions of the different modes of capture varied both between nets and annually, the selection curve of herring for a particular mesh size is not unique. It can however be reasonably approximated when girth is used as the selection criterion. Direct empirical selectivities are therefore recommended when interpreting population parameters from herring gillnet catch data.