During the 2017 CCAMLR WG-EMM Norway was requested to act towards the industry to ensure catch reporting in accordance with the present conservation measure (CM 21-03) and to investigate potential errors in historical catch report data. As a response, the Norwegian delegation presented a list of actions that would facilitate further consideration and clarification of these technical issues, and these actions were endorsed by the Scientific Committee (SC-CAMLRXXXVI/11). This document presents the outcome of the action steps including onboard experiments and analyses of historical catch data.
Experiments on board ‘Saga Sea’ showed that there is an average delay of approximately 9 minutes between the time krill enters the trawl mouth and the time the same krill enters the holding tanks where catch weight can be first estimated and recorded. Further experiments onboard ‘Saga Sea’ and ‘Antarctic Sea’ to estimate variability in the conversion from krill holding tank volumes to krill biomass indicated that a relative standard deviation of 8-29% should be expected depending on vessel and specific holding tank features.
A previous working document indicated that there have likely been additional delays between the reported catch time and actual catch time since the krill catch is kept in holding tanks before being weighed on the flow scales. Our investigations of historical catch report data showed patterns which likely reflect reporting routines rather than true catch variability and also indicate that the reporting routines have varied between vessels and among officers on board.
The various historical reporting routines are difficult to retrace, and therefore, in order to get an idea about the uncertainty in reported catch weight and position from historical data, we used information about the maximum total capacity of the holding tanks and assumed that all catch records have been estimated based on flow scale measurements. We could then estimate a maximum residence time before the accumulated krill catch is recorded on the flow scales. The maximum residence time decreased with increasing catch rates, ranging from average values of ca. 18 hours for low catch rates to ca. 3 hours for high catch rates. The deviance in catch between original reports and reports reallocated according to estimated krill residence time was approximately normally distributed around a mean value of 1.6 tonnes with ca. 90% of the deviances contained within ±25 tonnes. These results suggest that the reporting delays cause additional uncertainty to any given reported catch value but not major bias. The geographical distribution of reported catch at different spatial scales showed only minor deviances between what was previously reported to CCAMLR and the catch reallocated in the present study using estimated krill residence times.
In summary, the results indicate that the uncertainty associated with the historical reports of catch from these fishing vessels with continuous pumping systems is higher than typically assumed. These data should therefore be used with caution, especially when conducting fine-scale retrospective analyses. When data are aggregated at temporal and spatial scales more pertinent to advisory processes and management, our results suggest that the impacts of the uncertainties in catch weight and spatial distribution are likely to be minor.