Barnes, Brian B.; Hallock, Pamela; Hu, Chuanmin; Muller-Karger, Frank; Palandro, David; Walter, Cory; Zepp, Richard (2015). Prediction of coral bleaching in the Florida Keys using remotely sensed data. CORAL REEFS, 34(2), 491-503.

Abstract
Shallow water tropical coral reefs may bleach due to extremes in a variety of environmental factors. Of particular concern have been temperature, ultraviolet radiation, and photosynthetically available radiation. Satellite observation systems allow synoptic-scale monitoring of coral environments that can be used to investigate the effects of such environmental parameters. Recent advancements in algorithm development for new satellite data products have made it possible to include light availability in such monitoring. Long-term satellite data (2000-2013), in combination with in situ bleaching surveys (N = 3,334; spanning 2003-2012), were used to identify the environmental factors contributing to bleaching of Florida reef tract corals. Stepwise multiple linear regression supports the conclusion that elevated sea surface temperature (SST; partial R (adj) (2) = 0.13; p < 0.001) and high visible light levels reaching the benthos (partial R (adj) (2) = 0.06; p < 0.001) each independently contributed to coral bleaching. The effect of SST was modulated by significant interactions with wind speed (partial R (adj) (2) = 0.03; p < 0.001) and ultraviolet benthic available light (partial R (adj) (2) = 0.01; p = 0.022). These relationships were combined via canonical analysis of principal coordinates to create a predictive model of coral reef bleaching for the region. This model predicted 'severe bleaching' and 'no bleaching' conditions with 69 and 57 % classification success, respectively. This was approximately 2.5 times greater than that predicted by chance and shows improvement over similar models created using only temperature data. The results enhance the understanding of the factors contributing to coral bleaching and allow for weekly assessment of historical and current bleaching stress.

DOI:
10.1007/s00338-015-1258-2

ISSN:
0722-4028