Japanese longline CPUE for yellowfin tuna in the Indian Ocean up to 2012 standardized by generalized linear model

Japanese longline CPUE (quarterly and annual) for yellowfin tuna in the main fishing ground and whole Indian Ocean, as well as CPUE in each area in each of five areas for SS3 and Multifan-CL, was standardized up to 2012 by GLM (CPUE-LogNormal error structured model). Number of hooks between float (NHF) and material of main line and branch line were applied in the model to standardize the change of the catch rate which has been derived by fishing gear configuration. In order to avoid the bias of CPUE trend which may be caused by critical decrease of effort in the northwestern Indian Ocean, scenarios without Area 2 was also applied.
Basically, two series of standardized CPUEs including and excluding Area2 showed similar trend. In the main fishing ground, CPUE continuously decreased from around 15 (a nominal scale) in early 1960s to around 5.0 in 1974, and was kept in same level until 1990 with jump to 12.0 in 1977. Thereafter, it declined to about 3.0 in 1994 and has been kept in a low level with fluctuation between about 2 and 3 until 2007. After that, the CPUE declined to historical low level, 1.18-1.58 during 2008-2012. As the declining trend in the resent years was detected in both models including and excluding Area 2, where the piracy activity has been increasing since 2007, the resent declining trend would be reflecting actual change in abundance rather than change in CPUE derived from shift of fishing ground and/or decreased effort caused by increased piracy activity. The trend of standardized CPUE for whole Indian Ocean was similar to that of main fishing ground. Quarterly CPUE trends for main and whole Indian Ocean were similar to that of annual CPUE.
Trends of CPUEs were relatively similar among areas, i.e. large decline to middle 1970s, relatively stable until around 1991 and steadily declining thereafter. Applying LT5LN5 factor (five degree block) in the model showed relatively large effect on the CPUE trend for Area 3 and 4 in which the declining trend until around 1990 was steeper in the model without LT5LN5. Then, the CPUE trend derived from the model with LT5LN5 caused relatively flat trend throughout period analyzed.