Novel use of hook timers to quantify changing catchability over soak time in longline surveys

Abstract

Fishery-independent survey sampling programs frequently undergo changes in operational procedures, which have the capacity to alter the catchability coefficient, q. To preserve the continuity of the time series, changes in sampling protocol must be accounted for within the raw data. We used data from a long-standing shark longline survey as a case-study to demonstrate a method of estimating changing catch over variable soak times. Catches of longline sets with and without hook timers were modeled using generalized linear models (GLMs) to estimate catch conversion factors over varying soak times. Estimated conversion factors were used to correct the raw catch data, which were then analyzed with delta-lognormal GLMs to estimate indices of relative abundance. Uncertainty in conversion factor estimation was calculated via bootstrap resampling and propagated through to annual indices by correcting raw data using resampled conversion factors. Added variation introduced by implementation of correction factors was relatively small compared to the magnitude of the observation error of the resulting indices of relative abundance. In species where catch rate declined over soak time, the expected CPUE of shortened soak times increased relative to standard soak times. Contrarily, if catchability increased over soak time, expected CPUE decreased in shortened soak times. Thus, we showed that the predominant practice of treating each unit of sampling effort as equal in fixed, baited gear is not appropriate, and changes in soak time should be accounted for to preserve the longevity of the time series.