1.     Under current trajectories, it is unlikely that the coral reefs of the future will resemble those of the past. As multiple stressors, such as climate change and coastal development, continue to impact coral reefs, understanding the changes in ecosystem functioning is imperative to protect key ecosystem services.

2.     We used a 26-year dataset of benthic reef fishes (including cryptobenthic fishes) to identify multi-decadal trends in fish biomass production on a degraded coral reef. We converted fish abundances into estimates of community productivity to track the long-term trend of fish biomass production through time.

3.     Following the first mass coral bleaching event in 1998, the abundance, standing biomass, and productivity of fish communities remained remarkably constant through time, despite the occurrence of multiple stressors, including extreme sedimentation, cyclones, and mass coral bleaching events. Species richness declined following the 1998 bleaching event, but rebounded to pre-bleaching levels and also remained relatively stable.

4.     Although the species composition of the communities changed over time, these new community configurations still maintain a steady level of fish biomass production. While these highly dynamic and increasingly degraded systems can still provide some critical ecosystem functions, it is unclear whether these patterns will remain stable over future decades.

We sampled benthic reef fishes annually from 1993-2021 (excluding 2014, 2017, and 2020) at approximately 3-5 m depth. Sampling always occurred during the Austral winter to avoid sampling during the peak summer recruitment (Bellwood et al., 2012, 2006). Each year, 2-4 coral outcrops (hereon referred to as stations) of roughly equal sizes were sampled (approximately 2 m³; Bellwood et al., 2012, 2006). Because cryptobenthic reef fishes cannot be surveyed adequately using visual surveys, we used enclosed ichthyocide stations to sample the entire small benthic fish communities (e.g. Ackerman and Bellwood, 2000; Brandl et al., 2019; Morais and Bellwood, 2019). Each station was enclosed in a 2 mm-mesh mosquito net (double bed), weighed along the perimeter by a 2.3 m long steel chain. The net was sown over the chain to avoid coral abrasion. The net was then enclosed in a 2-man tent, again weighed down by a chain. All fishes within the net were euthanized with rotenone or (after 2000) anesthetized with clove oil (cf. Ackerman and Bellwood, 2000); both rotenone and clove oil equally sample benthic fishes (Ackerman and Bellwood, 2002). After five minutes, the divers slowly removed the tent and net, searching for anesthetized fishes. All fishes were collected in plastic bags and placed on ice. In the lab, the fishes were identified to species, where possible, and counted.