Aim Functional diversity metrics inform how species’ traits relate to ecosystem functions, useful for quantifying how exploitation and disturbance impact ecosystems. We compare the functional diversity of entire fish communities in a shallow‐water region with a deep‐sea region for further insight into the differences between these ecosystem types. Location The regions compared in this study were selected to represent a shallow‐water coastal region, Tasman and Golden Bays (TBGB), and a deep‐sea region, Chatham Rise (CR), in New Zealand. Methods Functional diversity was assessed using four metrics: functional richness, evenness, divergence and dispersion. We compared these metrics across four key functions: habitat use, feeding, locomotion and life history. Results Our results showed that overall, the shallow‐water and deep‐sea ecosystems had equal diversity. When focusing on the four ecological functions, the two ecosystems exhibited equal diversity metrics across most analyses. Of the significantly different results, the deep‐sea had higher functional richness for habitat use and locomotion traits, lower functional dispersion for feeding and lower functional evenness for life history. Main conclusions Differences across the functions highlight higher diversity of habitat utilization by deep‐sea fish, while lower diversity in feeding suggests deep‐sea fish tend towards generalist diets, likely driven by low food availability. Deep‐sea fish displayed an increased range of locomotive traits in our analyses, but this conflicts with existing evidence and warrants further study. Life‐history results suggest deep‐sea fish exhibit higher clustering of traits, indicating potential under‐utilization of life‐history strategies in the deep‐sea. Our results demonstrate that although deep‐sea fish communities have similar levels of diversity to shallow‐water communities, the traits that structure this diversity differ, and therefore, the systems may respond to exploitation differently.

Coral reef habitat quality is declining in many locations because of heatwaves (coral bleaching) and other disturbances. Reef‐building corals provide refuge from predation, so declining complexity restricts the transfer of energy to upper trophic levels. We use a size‐based ecosystem model to test systematically the effects of refuge at specific sizes and the impacts of distributions of refuges that vary in size. Outcomes are assessed in terms of ecosystem service provision. Simulations indicate that refuge sheltering fish between 5 and 10 cm in length enhances fish biomass and fisheries productivity, for example, increasing herbivore biomass by up to 80%. We identify and present a range of refuge profiles that can theoretically improve ecosystem service provision by as much as sixfold. Profiles where refuge availability decreases with increasing fish size consistently outrank other designs. Synthesis and application: These results highlight how critical small‐scale refuge is for coral reefs, and how we might design artificial reefs to maximise service provision.