Inter-Research >  > Prepress Abstract

MEPS prepress abstract   -  DOI:

Shining a light on the composition and distribution patterns of mesophotic and subphotic fish communities in Hawai`i

Mariska Weijerman*, Arnaud GrĂ¼ss, Dayton Dove, Jacob Asher, Ivor D. Williams, Christopher Kelley, Jeff Drazen

*Corresponding author:

ABSTRACT: As agencies shift from single-species management to ecosystem-based fisheries management, ecosystem models are gaining interest for understanding species dynamics in relation to oceanographic and ecological processes and human marine uses. However, information on community structure or distribution of many species that occupy deep (> 30m) waters is largely unavailable. We amassed a total of 24,686 fish observations of 523 species/taxa for the 30-410m depth areas surrounding the main Hawaiian Islands (MHI). We also obtained estimates of geomorphological variables, including substrate type, slope, rugosity, and ridge-like features. Using these two data sources, we (1) identified distinct fish communities along the 30-410m depth gradient, and (2) generated relative biomass maps for fish functional groups. We showed that the mesophotic zone ranges between 30m and 129m, with a fish faunal break at 60m. Beyond this zone, four subphotic zones were identified: upper rariphotic (130-169m), mid rariphotic (170-239m), lower rariphotic (240-319m), and upper bathyal (320-410m).We assigned fish species to functional groups partly based on identified depth ranges and fitted general additive models (GAMs) integrating geomorphological covariates to the functional group relative biomass estimates to determine the environmental variables that best predict the probability of encounter and relative biomass of each fish functional group. Finally, GAM predictions were employed to map functional group relative biomass distributions. These distribution maps showed a high relative biomass of many groups in the center of the MHI chain. This study contributes to a better understanding of fish community structure around the MHI and will inform ecosystem model parameterization.