From dive slates to digital imagery: an evaluation of the effect of coral reef assessment methods on benthic and substrate compositions and an alignment between methods

Abstract Various underwater survey methods have been used for coral reef assessment and long-term monitoring; however, comparing data across different methods remains challenging and can obscure ecological patterns. This study compared and aligned three commonly used survey methods in Indonesia, including line-intercept transects (LIT) and point-intercept transects (PIT) recorded on dive slates, and underwater photo transects (UPT), to improve data consistency and comparability. Surveys were conducted at 56 sites across four locations between 2013 and 2014. Two Bayesian Dirichlet models were fitted. The first model assessed the relative performance of the methods while accounting for random effects (location and site within location). The results indicated that UPT generally produced higher estimated per cent cover across most benthic and substrate categories; PIT was typically intermediate; and LIT tended to yield lower estimates for most categories, while relatively overestimating hard coral cover compared with UPT. Because of these discrepancies, the second model was developed to convert LIT or PIT data into UPT-equivalent compositions. While the predictive models have limitations, the predictions for most categories were relatively reliable and, in general, statistically indistinguishable from the observed UPT data, especially at the community level. These models can be iteratively refit to improve regional generalisability by adding observations from different sites and locations, or applied to new sites and locations by excluding the random effect at the cost of increased uncertainty. Overall, this approach improves data comparability and consistency, supporting long-term monitoring and conservation efforts by making conventional and modern methods more directly comparable.