First attempt of stock assessment using Stock Synthesis III (SS3) for the Indian Ocean albacore tuna (Thunnus alalunga)

Executive summary Stock assessment for the Indian Ocean albacore tuna based on the Stock Synthesis III (SS3) was attempted with consideration of available information on catch, abundance indices and length frequency up to 2010. A total of five fisheries were defined. Three independent CPUE series (Japan, Taiwan and Korea) were available for tuning indices. Quarterly length compositions from the different fisheries contributed to the likelihood in addition to the populations indices. The Several biological and ecological parameters such as the growth curves, length-weight relationship, natural mortality and steepness were assumed to be known (see main text of this paper) when optimizing the likelihood. Results based on the three independent incompatible CPUE series produced unreasonable results. Therefore, a CPUE series for Taiwanese longline fishery, in which fishing operation and fishing ground have been stable and its catch accounts for a large extent of total catch in the Indian Ocean, was employed in the base case scenario. Also, the weighted average of Japanese and Taiwanese CPUEs was also used in sensitivity tests. In a base case scenario (here natural mortality=0.2207, steepness=0.8 and SD of recruitment deviation=0.2), the current F/FMSY and SSB/SSBMSY were estimated as 1.657 and 0.844, respectively. A current SB(2010)/SB(1950) ratio was at 0.186, and the level of MSY was assessed as 28,093 (t). The figures become more optimistic when the abundance index was replaced with the averaged one; the current F/FMSY and SSB/SSBMSY were respectively 1.242 and 1.268 and MSY was 30,296 (t). The estimated stock indices were sensitive to changes in the value of steepness, natural mortality, the extent of recruitment deviation and abundance indices, but the outcome of this exercise provides some implications to the interpretation of the albacore stock, in which some of key biological parameters are unknown and have uncertainty, because other stock assessment methods failed to provide comprehensive sensitivity tests due to lack of convergence in optimization.