Spatiotemporal variability in bigeye tuna ( Thunnus obesus) dive behavior in the central North Pacific Ocean

Data from 29 pop-up archival transmission (PAT) tags deployed on commercial-size bigeye tuna ( Thunnus obesus) in the central North Pacific Ocean from 4°N to 32°N were analyzed to describe variability in their dive behavior across space and time. During the day, bigeye tuna generally spent time in the 0-50 m and 300-400 m depth ranges, with spatial and temporal variability in the deep mode. At night, bigeye tuna generally inhabited the 0-100 m depth range. Three daily dive types were defined based on the percentage of time tuna spent in specific depth layers during the day. These three types were defined as shallow, intermediate, and deep and represented 24.4%, 18.8%, and 56.8% of the total number of days in the study, respectively. More shallow and intermediate dive-type behavior was found in the first half of the year, and in latitudes from 14°N to 16°N and north of 28°N. A greater amount of deep-dive behavior was found in the regions south of 10°N and between 18°N and 28°N during the third and fourth quarters of the year. Dive-type behavior also varied with oceanographic conditions, with more shallow and intermediate behavior found in colder surface waters. Intermediate and deep-dive types were pooled to reflect the depths where bigeye tuna may have potential interactions with fishing gear. A Generalized Additive Model was used to quantify the effects of time, space, and sea surface temperature on this pooled dive type. Results from the model showed that while latitude and quarter of the year were important parameters, sea surface temperature had the most significant effect on the pooled intermediate and deep-dive behavior. Model predictions indicated that the largest percentage of potential interaction would occur in the fourth quarter in the region from 18°N-20°N, which corresponds to the time and place of the highest bigeye tuna catch rates by the Hawaii-based long-line fishery. These results suggest that a model framework using these three predictive variables may be useful in identifying areas of potentially high bigeye tuna catch rates.