Downscaling Ecological Trends from the Spatially Randomized Datasets of the National Coral Reef Monitoring Program

Our goals in this analysis are to (1) identify spatial sectors that are smaller than our current survey design scale while retaining robust, responsible statistical sampling, (2) ensure that our sectors are as ecologically homogenous as possible, and (3) extend our temporal coverage and sampling density by responsibly comparing similar metrics across multiple methods within the coherent statistical framework of a mixed-effects model. Specifically, we identify clusters based on NCRMP benthic cover data from four different survey methods, and then use mixed model analysis to assess both the optimal number of spatial sectors and the significance of long-term temporal change in hard coral cover in each newly created sector. Interesting patterns in Hawaiian percent cover data are revealed with significant long-term trends that are obscured at the island level: 13 out of 63 sectors show long-term decline, and only 4 out of 63 show an increase in coral cover.