Preliminary Assessment Of Indian Ocean Yellowfin Tuna 1950-2018 (Stock Synthesis, V3.30)

This paper presents a preliminary reference model for the assessment of yellowfin tuna (Thunnus albacares) using the age and length structured integrated assessment model Stock Synthesis (SS) version 3.30.09. In this document we review the reference model that was used for the 2018 assessment as part of the 2019 workplan for yellowfin. The main features of the new model are a proposal for reducing or removing the influence of tagging data and for a reduced number of areas. The analyses that led to this proposal are explained throughout the document. In brief, the analyses and diagnostics of the model suggest that tagging data and environmental data do not contain enough information to estimate the movement between the 4 areas defined within the model: western-tropical, western-temperate, eastern-tropical and western tropical, and that these data make the model unstable. Therefore, we analyzed and compared three spatial configuration options: two area model defining East and West regions, three area model aggregating regions 3 and 4 of the 2018 model, and a four-area model comparable to the last year reference model but with a different version of (v3.30). The results suggest that the 2-area model is the most stable model and therefore, the 2-area model is proposed as the spatial structure of the reference model. From this, we update the model with the latest data available and analyse a number of sensitivity analyses. The catch and length frequency data were updated until 2018, new estimates of the joint index were introduced and two other additional indices were considered for inclusion as sensitivity runs; an acoustic index derived from echosounder receivers placed on FAD buoys prior to fishing and a purse seine free school index that improves upon the definition of effort in the purse seine fishery. However, the 2018 reference model was proven to be very sensitive to the new length frequency data. Similar conclusions arise from the reference model proposed in this paper. The model is very sensitive to the new length frequency data of longlines and the reference model proposed here uses length frequency data until 2014. This document is a draft and there are still analyses to complete before the 2019 WPTT. This document does not contain results on stock status nor reference points and these will be presented and discussed during the WPTT.