Scaling up our understanding of fish-habitat associations on shallow reefReef-associated Biodiversity

Elizabeth Oh¹, Rick Stuart-Smith¹

¹University of Tasmania, Australia

In this modern era of globalisation and unprecedented widespread pressures on marine ecosystems, conservation management and policy decisions often occur over large jurisdictions. This has increased demand for assessments of ecosystem health in key habitats, like reef, that cover national, even global scales. Synthesis of data at these scales risks an over-simplification of dynamic and complex reef communities into meaningless or abstract indicators. Furthermore, unstandardised selection of indicators across monitoring projects either prevents data synthesis or forces it to occur at lower resolutions of taxonomy or morphology. This issue is exacerbated for shallow reef along the temperate Australian coast, where there is no standardised reporting for the kelp forests of the “Great Southern Reef”. The taxonomic diversity of habitat-formers in these reefs is high, presenting a challenge for data synthesis, but standards for grouping habitat-formers are potentially oversimplified. For example, “large canopy-forming algae”, is a commonly used indicator for monitoring, despite encompassing large differences typical height, thickness, complexity and thus potential to provide structural amenity to other reef life. This project investigated reef fish-habitat associations from standardised biodiversity survey data across 362 sites in Temperate Australia to distinguish what habitat species may influence reef fish assemblages, and how and if these can be summarized into meaningful indicators for reporting across large scales. Structural Equation Models evaluated the direct effects of key environmental drivers on both groups of taxa, as well as indirect links between environmental drivers, habitat-formers and fish. Data were available at species-level for fish, macroinvertebrates and macroalgae, allowing us to compare models generated from taxonomic data with those generated from data grouped by potentially important morphological features of habitat.