Extended Data Table 3.

Summary of ordinal logistic regression (OLR) models relating the per cent cover of reef-builders (hard coral + crustose coralline algae) four years following the 2015 marine heatwave to local land-sea human impacts and environmental factors

AICc, Akaike’s information criterion corrected for small sample size; ΔAICc, change in AICc across the candidate models (ΔAICc ≤ 2 of the top model are shown); McFadden’s pseudo-R², proportion of variation explained by candidate model (0.2–0.4 represent an excellent fit)⁹⁸. Significant predictors at p < 0.05 are in bold.