User menu
This analysis compared IOTC catch data in the public domain with an alternative estimation for associated (log school) catches based on port sampling data from the European Union sampling program aggregated by 5º square or statistical area, year and quarter.
The underlying assumption is that any fleet fishing in the same spatio-temporal strata and on log schools will have, on average, the same catch composition.
Species composition distribution in the sampled strata (year, quarter and 5x5º cell or statistical area) was estimated by bootstrapping across the catch by species derived from each sample and was applied to the total catch (aggregated across flags) reported in these strata. For unsampled strata, a correction factor was estimated by comparing the species composition reported and estimated in sampled strata on a yearly basis. This correction factor was then applied to the total catch on log school in each unsampled strata.
As expected, the results indicate significant deviations in the total catch between the current estimates and the public domain data in 2018. Moreover, it also indicates some deviations in the early time series and in the most recent years. Two approaches for the catch estimation, one based on the raw samples and another one in which species composition is estimated separately for different catch categories (<10 kg, 10-30 kg and >30 kg) and adjusted according to the amount of each fraction as reported in the logbooks, were tested.
While this analysis does not provide any insight on the best procedure to estimate catch composition in the purse seine fishery, it provides a time-consistent, scientific-based, estimate of catches for the purse seine fishery that can be of use for future stock assessment.
