Estimating the size distribution of reported catches on-board factory vessels – Issues with using data from the production process

Abstract

High-grading is the practice of discarding lower value catches to make space for catches with higher value. It is necessary to understand the extent and variation in these unreported discards to improve stock assessments and management decisions. Where discards are not directly observed, a proposed methodology for estimating high-grading involves comparing size distributions of total catches (before sorting) and landed catches (after sorting), but we have yet to identify a suitable data source for the landed portion. This report presents data from two pilot studies exploring the suitability of data gathered during the onboard factory production process for describing the size distribution of reported catches. We received these data in a summarised report, where individual fish weights are aggregated into coarse, overlapping size grades. This summarised form results in a large loss of information. Applying the necessary statistical procedures to get a more detailed picture of fish size distributions would introduce even more uncertainty into an already uncertain estimation, potentially leading to non-signficant results, and can introduce unknown biases. We conclude that it is necessary to use the raw data behind summarised reports which provide data on individual fish. To this end, we address the logistical and statistical issues posed by production data including individual fish observations.