*4.3. Statistical Analyses*

To assess the impact of reef site, depth, and morphology on the percentage of cover of live and recently dead coral, data were analysed using generalised linear mixed models (GLMM, R version 3.6.3 [73]). The influence of reef, morphology, and depth was investigated using a full subsets model selection approach [74]. Following convention, the simplest model with the lowest Akaike information criterion was considered optimal [74]. A null model with only the intercept and random effect was included to determine whether any of the included variables were useful predictors of coral cover. The percentage of live or recently dead hard coral was modelled in relation to reef site, coral morphology, and transect depth, with transect included as a random effect. The model used a Tweedie distribution with a log link and was modelled using the glmmTMB function within the glmmTMB package [75].

**Figure 6.** Examples of different coral health categories. The legend on each image indicates the site name (see Figure 1), transect depth (either deep or shallow), and number (T1, T2, or T3), along with the recorded coral health category. Coral health abbreviations include live branching coral (BC), bleached branching coral (BBC), recently dead branching coral (RDBC), live massive coral (MC), and recently dead massive coral (RDMC).

> **Author Contributions:** K.B. and N.H. designed the study. P.B.-B., K.B. and P.B. processed the data. P.B.-B. and H.E.E. analysed the data. All authors contributed to the writing and review of the manuscript. There are no competing or conflicting interests. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author.

**Acknowledgments:** This research was conducted with the support of The Hydrous NGO. We thank Hassan Hameez, Abdul Sattar Hassan, Kate Malmgren, Rick Miskiv, and Erika Woolsey for their help with field surveys and trip logistics.

**Conflicts of Interest:** The authors declare no conflict of interest.
