Excess coral-family change at 90% cover return was uncertain on 13 matched Great Barrier Reef reefs, 2005–2025

We tested how hard-coral family cover shares changed when total cover first returned after a large loss. We used 1,573 AIMS LTMP reef surveys at 99 Great Barrier Reef reefs through report year 2025. The first selected loss per reef was at least 50%, from baseline cover of at least 10%, across a survey gap of at most three years. The first later survey at 90% or more of baseline cover within ten years was a cover return. Of 59 first-loss events, 23 had an observed return. Mean total variation distance between baseline and return shares was 0.164 (95% sector bootstrap interval 0.092 to 0.257). Twelve returns had distance greater than 0.10. Thirteen return reefs had a same-reef reference interval with similar duration, similar cover at its two ends, and no single observed loss of at least 50%. Mean distance was 0.184 at return and 0.107 in reference intervals. The paired excess was 0.076 (95% sector bootstrap interval -0.032 to 0.172). Exact sector resampling gave a similar interval. Thus the matched test did not clearly separate excess change at cover return from background change. It did not show equal composition. The reference sample was small, nearby reefs shared large changes, and the return definition affected the result. We publish code, aggregate results, figures, source hashes and instructions. The source data remain under their original non-commercial terms in a retained snapshot.

coral recovery · Great Barrier Reef · community composition · coral cover · spatial bootstrap · AIMS LTMP · reproducible ecology