Investigating different methods for obtaining hake survey probability distributions for length, and an inspection the sub sample for which biologicals were measured

The hake cannibalism and inter-species predation model that is currently being developed by Ross-Gillespie and Butterworth (2014) has thus far failed to reflect both a biologically feasible estimate of daily ration and the proportion of hake in the diet of hake predators indicated by the diet data obtained during surveys. Andre Punt made a suggestion to investigate the sampling strategy used to obtain the diet data, in order to ascertain whether this strategy might be giving rise to biases in the estimates for the population as a whole in terms of both the length distributions and the proportions of hake in the diet of hake predators. This has led to an examination of the raw survey catch-at-length data and the methods used to analyse these data, and a few suggestions have been made for alternative approaches to weighting the length probability distributions from individual trawls in order to obtain aggregated distributions for each stratum.

This document gives the details of the survey sampling procedures, the equations for the methodology currently in place for calculating probability distributions for length, as well as equations for suggested alternative methods for weighting.