AquaAuma
commented
1 year ago For DATRAS surveys, we always recalculated weights based on abundance at length data, and compared it to the reported taxa-haul abundance/weight when available. However, DATRAS holds reported weights that are not in the right unit, and this is not clearly indicated in the data, making it difficult to know what's "right". So cpue recalculated from the abundance at length are all estimations, not measurements, for ICES surveys. We're working on the QAQC this month on this specific aspect. I'll try to make that more clear in the MS. For DATRAS, swept areas are never measured, but always recalculated (some directly, some via model predictions). I like the idea of making that explicit in the dataset, I'll think how
CPUE/weight records. Can you clarify how you have defined CPUE/weights where unrecorded in the initial data? My understanding from your manuscript (L204–211) is that if weight was not recorded then it has been estimated (from a length-weight relationship). I am not sure if there are any surveys that do not provide counts of the number of individuals? But, if so, my understanding is that you have not calculated CPUE and it is thus defined as NA. Is that correct? While the onus is on researchers using these data to make sure they are using them properly, it might help to have a summary table that defines which surveys record individual counts and/or weights, and/or an additional field in the processed dataset that defines which ‘observations’ are real or calculated (as the uncertainty may differ). There are other data fields where this kind of additional information could be valuable too (e.g., I think swept areas in your compiled dataset variously represent the areas swept by the trawl doors or wings, depending on the survey, and are variously observed quantities, fixed/assumed values or estimated values?). Reporting these differences in the compiled dataset would greatly help researchers propagate uncertainty.