Abstract
The Hawaiian Islands Cetacean and Ecosystem Assessment Survey (HICEAS) 2017 was conducted in waters within the United States (U.S.) Exclusive Economic Zone (EEZ) around the Hawaiian Archipelago (henceforth “Hawaiian EEZ” for brevity) from 6 July through 1 December 2017 (Yano et al. 2018). The primary objective of this line-transect survey was to collect cetacean sighting data to support the derivation of cetacean density estimates using both design-based analyses and habitat modeling techniques. This report summarizes the results of the habitat modeling effort. The design-based estimates are described separately in Bradford et al. (in review).
Habitat models, or species distribution models (SDMs), have been recognized as valuable tools for estimating the density and distribution of cetaceans and assessing potential impacts from a wide range of anthropogenic activities (e.g., Gilles et al. 2011; Goetz et al. 2012; Hammond et al. 2013; Redfern et al. 2013). SDMs for nine cetacean species have been developed for waters in the central North Pacific, including U.S. EEZ waters around the Hawaiian Islands, from ship based, line-transect survey data collected by the Pacific Islands Fisheries Science Center (PIFSC) and Southwest Fisheries Science Center (SWFSC) between 1997 and 2012 (Forney et al. 2015). The models provided spatially explicit density predictions at a 25 km × 25 km grid resolution for pantropical spotted dolphin (Stenella attenuata), striped dolphin (S. coeruleoalba), spinner dolphin (S. longirostris), rough-toothed dolphin (Steno bredanensis), common bottlenose dolphin (Tursiops truncatus), false killer whale (Pseudorca crassidens), short-finned pilot whale (Globicephala macrorhynchus), sperm whale (Physeter macrocephalus), and Bryde’s whale (Balaenoptera edeni).
To develop improved and updated SDMs, sighting data from HICEAS 2017 were combined with previous line-transect survey data collected within waters of the Hawaiian EEZ from 2002 to 2016. The majority of these data were from the two previous HICEAS efforts, the first in 2002 (Barlow 2006) and the second in 2010 (Bradford et al. 2017). In contrast to previous modeling efforts that included survey data from a broader region of the central Pacific Ocean (Becker et al. 2012; Forney et al. 2015), the current SDMs were built only with survey data collected within waters of the Hawaiian EEZ. Habitat models were developed to derive spatially explicit estimates of species density specific to the Hawaiian EEZ based on previously established methods that allow for the incorporation of segment-specific estimates of detection probability (Becker et al. 2016). Potential habitat variables included bathymetric depth, distance to islands, and a suite of dynamic surface and subsurface outputs from an ocean circulation model. The habitat-based models of cetacean density developed in this study represent an improvement over the previous models developed by Forney et al. (2015) because they more accurately account for variation in detection probabilities, provide finer-scale density predictions (~9 km × 9 km grid resolution), and better account for uncertainty in the resulting study area abundance estimates. In addition, they include dynamic subsurface variables that were not available for the previous models. Further, increases in sample sizes allowed us to develop a new habitat model for Risso’s dolphin (Grampus griseus).