Fig. 3

In situ explanatory variables of bleaching and their standardized logit coefficients with greatest predictive power. a ∆AIC_C, computed as AIC_C – min(AIC_C), values of all 10,367 runs of an ordinal logistic regression model, where models within ∆AIC_C ≤ 2 (dashed line and gray shaded region) are statistically indistinguishable, of which there were 20. b The best model (i.e. ∆AIC_C = 0) included six variables, of which high-frequency temperature variability was the absolute most influential and also greatest mitigating factor to bleaching prevalence. c Summing across 20 indistinguishably good models (i.e. within ΔAIC_C ≤ 2), high-frequency temperature variability was consistently most influential. Variable categories are shown in Table 1. Delete-1 jackknife standard error bars are shown in (b), while the standard error bars shown in (c) were obtained by summing in quadrature the individual standard errors from each of the 20 models computed after delete-1 jackknife resampling