Effective factors of tori-poles in reducing incidental catch of seabirds in the Japanese longline fishery

Effective factors of tori-poles in reducing incidental catch of albatross were examined with the data from Japanese observer program in southern bluefin tuna fishery. A total of 727 observations were used in the analysis. The data in night settings were not used because observed numbers of albatross and other seabirds were not recorded correctly during night-setting. The tori-pole specifications were categorized as follows: i) bird line material (Type I: multifilament twine, Type II: nylon code, and Type III: nylon monofilament), ii) streamer material (Type A: nylon code and urethane cube, Type B: polypropylene (PP) band, and Type C: combination of Type A and B), iii) bird line length (approx. 50 m, 100 m, 150 m, and 200 m), iv) pole height above sea surface (5 - 10 m, and 10 - 15 m). A Catch model (generalized linear model) was constructed: catch number of albatross was treated as responsible variables with a negative binomial error structure distribution; the potential factors affecting albatross catch were incorporated as explanatory variables. The model was evaluated by model selection based on Akaike's Information Criterion (AIC). In the model, bird line length was selected as explanatory variable, but such factors as bird line material, streamer material, and pole height were not. The model selected the number of albatross observed during line setting, and indicated that the catch increased with the observed number, as might be expected. Results in the model analysis suggest that: 1) the effectiveness of tori-pole in reducing incidental catch of albatross increased with longer bird line; 2) the effectiveness did not differ between Type I, II, and III in bird line material, and between Type A, B, and C in streamer material; 3) the effectiveness did not differ between 5 - 10 m and 10 - 15 m in pole height above the sea surface.