Treating coral bleaching as weather: a framework to validate and optimize prediction skill.

Thomas Mario De Carlo

Research output: Contribution to journalArticlepeer-review

20 Scopus citations


Few coral reefs remain unscathed by mass bleaching over the past several decades, and much of the coral reef science conducted today relates in some way to the causes, consequences, or recovery pathways of bleaching events. Most studies portray a simple cause and effect relationship between anomalously high summer temperatures and bleaching, which is understandable given that bleaching rarely occurs outside these unusually warm times. However, the statistical skill with which temperature captures bleaching is hampered by many "false alarms", times when temperatures reached nominal bleaching levels, but bleaching did not occur. While these false alarms are often not included in global bleaching assessments, they offer valuable opportunities to improve predictive skill, and therefore understanding, of coral bleaching events. Here, I show how a statistical framework adopted from weather forecasting can optimize bleaching predictions and validate which environmental factors play a role in bleaching susceptibility. Removing the 1 °C above the maximum monthly mean cutoff in the typical degree heating weeks (DHW) definition, adjusting the DHW window from 12 to 9 weeks, using regional-specific DHW thresholds, and including an El Niño threshold already improves the model skill by 45%. Most importantly, this framework enables hypothesis testing of other factors or metrics that may improve our ability to forecast coral bleaching events.
Original languageEnglish (US)
Pages (from-to)e9449
StatePublished - Jul 21 2020

Bibliographical note

KAUST Repository Item: Exported on 2020-10-01
Acknowledgements: The author received no funding for this work.


Dive into the research topics of 'Treating coral bleaching as weather: a framework to validate and optimize prediction skill.'. Together they form a unique fingerprint.

Cite this