“…In this context, it was appropriate to fit the delta approach which models the probability of encounter of a fish population and the non-zero CPUE when fish are encountered (Pennington, 1983(Pennington, , 1996Lo et al, 1992). Due to the low percentage of zero catches (<10%) of the quarterly 5 • × 5 • aggregated dataset that implying zero inflation was not an issue (Ichinokawa and Brodziak, 2010), we only fit the catch rate model using the log-transformed CPUEs of albacore, with a small constant (10% of the grand mean) added to avoid log-transformation problems (Campbell et al, 1996;Howell and Kobayashi, 2006;Mugo et al, 2010). GAMs were built using the gam function of the "mgcv" package in R-language TABLE 1 | List of IPCC atmosphere-ocean general circulation models where the future environmental variables, including the sea surface temperature (SST, • C), sea surface salinity (SSS, PSU), mixed layer depth (MLD, m), chlorophyll-a concentration (CHL, kg L −1 ), and dissolved oxygen concentration under 100 m depth (DO100, mmol L −1 ), were obtained to generate potential habitat maps of south Pacific albacore under ocean warming scenarios (RCP 4.5 and RCP 8.5).…”