Assessing the population-level impacts of North Pacific loggerhead and western Pacific leatherback turtle interactions in the Hawaii-based shallow-set longline fishery

Citation
Siders Z, Eguchi T, Langseth B, et al (2020) Assessing the population-level impacts of North Pacific loggerhead and western Pacific leatherback turtle interactions in the Hawaii-based shallow-set longline fishery. NOAA Technical Memorandum NMFS-PIFSC-95
Abstract

This population assessment is for the North Pacific (NP) loggerhead turtle (Caretta caretta) Distinct Population Segment (DPS) and the western Pacific (WP) leatherback turtle (Dermochelys coriacea) nesting population, with the sole purpose of evaluating the population- level impacts of a single U.S. commercial fishery, the Hawaii -based shallow -set longline (SSLL) fishery on these two populations. Both populations are listed as Endangered under the Endangered Species Act (ESA), NP loggerheads as a DPS and WP leatherbacks as a global species. This assessment was performed in a Bayesian framework with four main components: nesting data imputation (leatherbacks only), nesting trend analysis, population viability analysis (PVA), and incorporation of direct SSLL fishery take into the PVA (i.e., adding a “take model” component). Conducting PVAs under scenarios with and without future takes by the SSLL fishery allowed for evaluation of the impact of the fishery on the population s tatus (e.g., abundance relative to pre-determined thresholds) and trends for the two populations. Trends and abundance for nesting females were estimated using a Bayesian state- space model (Boyd et al. 2017) with exponential population growth as the underlying process. Time series data used in the model originated from long-term nesting beach monitoring programs in Japan (loggerheads) and Indonesia (leatherbacks). The loggerhead data from Japan were provided as annual nest counts from three index beaches (Maehama, Inakahama, and Yotsusehama) from 1985 to 2015. The leatherback data came from two index beaches in Indonesia (Jamursba Medi and Wermon) from 2001 to 2017. The leatherback data set contained months with no monitoring effort; thus, we developed a model to impute the missing data in order to produce a time series of annual nest counts. The imputati on model was autoregressive with a lag of one month (AR1 model) where the relationship between the numbers of nests in two months was modeled by the Fourier series.