Satellite SST-Based Coral Disease Outbreak Predictions for the Hawaiian Archipelago

• Published in: Remote sensing (Basel, Switzerland) Vol. 8; no. 2; p. 93
• Main Authors: Caldwell, Jamie M, Heron, Scott F, Eakin, C Mark, Donahue, Megan J
• Format: Journal Article
• Language: English
• Published: Switzerland MDPI AG 01.02.2016
• Subjects: boosted regression trees; cold snaps; corals; disease outbreaks; Hawaiian archipelago; hot snaps; MPSA; SST metrics; winter condition; models; corals; hot snaps; MPSA; boosted regression trees; cold snaps; disease outbreaks; winter condition; Hawaiian archipelago; SST metrics
• ISSN: 2072-4292; 2072-4292
• DOI: 10.3390/rs8020093

Predicting wildlife disease risk is essential for effective monitoring and management, especially for geographically expansive ecosystems such as coral reefs in the Hawaiian archipelago. Warming ocean temperature has increased coral disease outbreaks contributing to declines in coral cover worldwide. In this study we investigated seasonal effects of thermal stress on the prevalence of the three most widespread coral diseases in Hawai'i: white syndrome, growth anomalies and tissue loss syndrome. To predict outbreak likelihood we compared disease prevalence from surveys conducted between 2004 and 2015 from 18 Hawaiian Islands and atolls with biotic (e.g., coral density) and abiotic (satellite-derived sea surface temperature metrics) variables using boosted regression trees. To date, the only coral disease forecast models available were developed for white syndrome on the Great Barrier Reef (GBR). Given the complexities of disease etiology, differences in host demography and environmental conditions across reef regions, it is important to refine and adapt such models for different diseases and geographic regions of interest. Similar to the white syndrome models, anomalously warm conditions were important for predicting white syndrome, possibly due to a relationship between thermal stress and a compromised host immune system. However, coral density and winter conditions were the most important predictors of all three coral diseases in this study, enabling development of a forecasting system that can predict regions of elevated disease risk up to six months before an expected outbreak. Our research indicates satellite-derived systems for forecasting disease outbreaks can be appropriately adapted from the GBR tools and applied for a variety of diseases in a new region. These models can be used to enhance management capacity to prepare for and respond to emerging coral diseases throughout Hawai'i and can be modified for other diseases and regions around the world.