You may remember my interest in the “collectors' curve” as part of the equation describing the expected number of species you’ll observe in a single haul. In the past, when fish were plenty, many fisheries surveys, primarily interested in monitoring commercial species, subsampled the catch from large hauls. Being efficient and rational, you did not wish to end up sorting more fish than what you needed for a precise abundance estimate of the commercial species for which an abundance proxy was required. But although subsampling may save you time, it reduces the number of fish examined, and therefore also the possibility for recording the less common species. Taking a subsample of 100 kg of fish from a 1 ton catch with a few dominating commercial species. Discarding 900 kg, may provide a sample providing a decent estimate of abundance for the most abundant commercial species after you have multiplied the sample by 10, but an underestimate of species richness, unless you know that only one tenth of the catch, and hence of the individuals and the area swept, was examined for species occurrence/absence.
I know the ICES GOV data contain opportunities for safeguarding information about the size of subsamples. But perhaps they have not been consistently used. Please introduce information about the proportion of the survey catch that is examined for species composition in your guide and database.
From Henrik Gislason:
You may remember my interest in the “collectors' curve” as part of the equation describing the expected number of species you’ll observe in a single haul. In the past, when fish were plenty, many fisheries surveys, primarily interested in monitoring commercial species, subsampled the catch from large hauls. Being efficient and rational, you did not wish to end up sorting more fish than what you needed for a precise abundance estimate of the commercial species for which an abundance proxy was required. But although subsampling may save you time, it reduces the number of fish examined, and therefore also the possibility for recording the less common species. Taking a subsample of 100 kg of fish from a 1 ton catch with a few dominating commercial species. Discarding 900 kg, may provide a sample providing a decent estimate of abundance for the most abundant commercial species after you have multiplied the sample by 10, but an underestimate of species richness, unless you know that only one tenth of the catch, and hence of the individuals and the area swept, was examined for species occurrence/absence.
I know the ICES GOV data contain opportunities for safeguarding information about the size of subsamples. But perhaps they have not been consistently used. Please introduce information about the proportion of the survey catch that is examined for species composition in your guide and database.