AUTHOR=Asner Gregory P. , Vaughn Nicholas , Grady Bryant W. , Foo Shawna A. , Anand Harish , Carlson Rachel R. , Shafron Ethan , Teague Christopher , Martin Roberta E. TITLE=Regional Reef Fish Survey Design and Scaling Using High-Resolution Mapping and Analysis JOURNAL=Frontiers in Marine Science VOLUME=8 YEAR=2021 URL=https://www.frontiersin.org/journals/marine-science/articles/10.3389/fmars.2021.683184 DOI=10.3389/fmars.2021.683184 ISSN=2296-7745 ABSTRACT=

Coral reefs are undergoing changes caused by coastal development, resource use, and climate change. The extent and rate of reef change demand robust and spatially explicit monitoring to support management and conservation decision-making. We developed and demonstrated an airborne-assisted approach to design and upscale field surveys of reef fish over an ecologically complex reef ecosystem along Hawai‘i Island. We also determined the minimal set of mapped variables, mapped reef strata, and field survey sites needed to meet three goals: (i) increase field survey efficiency, (ii) reduce field sampling costs, and (iii) ensure field sampling is geostatistically robust for upscaling to regional estimates of reef fish composition. Variability in reef habitat was best described by a combination of water depth, live coral and macroalgal cover, fine-scale reef rugosity, reef curvature, and latitude as a proxy for a regional climate-ecosystem gradient. In combination, these factors yielded 18 distinct reef habitats, or strata, throughout the study region, which subsequently required 117 field survey sites to quantify fish diversity and biomass with minimal uncertainty. The distribution of field sites was proportional to stratum size and the variation in benthic habitat properties within each stratum. Upscaled maps of reef survey data indicated that fish diversity is spatially more uniform than fish biomass, which was lowest in embayments and near land-based access points. Decreasing the number of field sites from 117 to 45 and 75 sites for diversity and biomass, respectively, resulted in a manageable increase of statistical uncertainty, but would still yield actionable trend data over time for the 60 km reef study region on Hawai‘i Island. Our findings suggest that high-resolution benthic mapping can be combined with stratified-random field sampling to generate spatially explicit estimates of fish diversity and biomass. Future expansions of the methodology can also incorporate temporal shifts in benthic composition to drive continuously evolving fish monitoring for sampling and upscaling. Doing so reduces field-based labor and costs while increasing the geostatistical power and ecological representativeness of field work.