SCIENTIFIC COMMITTEE SEVENTH REGULAR SESSION 9-17 August 2011 Pohnpei, Federated States of Micronesia Analyses of Catch Data for Oceanic Whitetip and Silky Sharks reported by Fishery

This report presents descriptive statistical summaries and generalized linear model (GLM) analyses of catch data for oceanic whitetip shark Carcharhinus longimanus and silky shark C. falciformis in the Hawaii-based pelagic longline fishery. This paper is a collaborative effort begun at the Secretariat of the Pacific Community (SPC) in New Caledonia and completed at the NOAA Fisheries Pacific Islands Fisheries Science Center (PIFSC) in Hawaii. The data were collected by fishery observers aboard commercial vessels in 1995–2010. Oceanic whitetip shark mean annual nominal CPUE decreased significantly from 0.428/1,000 hooks in 1995 to 0.036/1,000 hooks in 2010. This reflected a significant decrease in nominal CPUE on longline sets with positive catch from 1.690/1,000 hooks to 0.773/1,000 hooks, and a significant increase in longline sets with zero catches from 74.7% in 1995 to 95.3% in 2010. Oceanic whitetip shark CPUE was standardized by delta-lognormal and zero-inflated Poisson GLM methods. The latter method was employed because 90.1% of the longline sets caught zero oceanic whitetip sharks. Four factors (16 haul years; calendar quarters; deep-and shallow-set fishery sectors; eight fishing regions) were significant explanatory variables in these analyses. Sea surface temperature was a significant continuous explanatory variable in a binomial GLM of the presence or absence of oceanic whitetip shark catches. The haul year effect coefficients from these models were used to compute indices of relative abundance. These time series were highly correlated, and each was also highly correlated with the time series of nominal CPUE. The silky shark catch data differed from the oceanic whitetip shark data in four major respects. The first was that nearly all silky sharks are caught on deep sets. The second was that most (62.5%) of the silky shark catch was taken from 0-10⁰N, although only 3.4% of the observed fishing occurred in those latitudes. The third difference was that sample sizes were very small before 2000. Finally, although 46.3% of the longline sets from 0-10⁰N caught zero silky sharks, 54.5% of the silky shark catch in these waters was taken on 11.5% of the longline sets, which caught ≥5 silky sharks. These differences led to use of the data from 0-10⁰N in the deep sector from 2000–2010 in the GLM analyses, which were fitted by delta-lognormal and quasi-Poisson (i.e., overdispersed) methods. These GLM analyses had low explanatory power. Silky shark CPUE has ranged from 0.034/1,000 hooks to 1.840/1,000 hooks, but with no significant trend. Therefore, it is concluded …