Predicting environmental suitability for key benthic species in an ecologically and economically important deep-sea environment

The Great Australian Bight (GAB) is important ecologically due to a diverse benthic fauna with high endemicity, and productivity that supports marine mammals and apex predators. The region is also economically important, supporting valuable fisheries, aquaculture and ecotourism industries. Deep-sea (> 200 m depth) exploration leases for oil and gas have also recently been activated. Characterising the environment of the GAB is a priority in light of these activities, but although areas of the GAB continental shelf have been well-studied, there is a paucity of data on benthos from depths > 200 m and a virtual absence of collated benthic ecological information. Predictive ecological modelling is increasingly being used for deep-sea benthic characterisation given the large areas involved and logistic challenges of sampling these environments. For example, species distribution models can be used to predict environmental suitability in unsampled areas, and to inform spatial management for data-poor regions. We explored options and selected the maximum likelihood implementation of the popular MaxEnt method to predict environmental suitability for 96 deep-water benthic species across temperate Australia. Occurrence records were obtained from a collation of historical museum records, online databases and recent sampling in the GAB at 200-3000 m. Aggregated predictions were used to identify patterns of suitability across species. The GAB continental slope had high relative suitability across all modelled species, particularly in the phyla Chordata and Cnidaria, while predicted suitability for species in the phyla Mollusca and Arthropoda was generally higher on the east than south coast of Australia. These patterns were influenced by the suite of species used; we found predicted suitability within the GAB to be negatively influenced by the proportion of east coast records per species. Average bottom temperature, which correlated with depth and a range of water quality variables, was the most important environmental predictor of suitability. The currently poorly-sampled upper mid-slope depths (500-800 m) of the GAB had the greatest predicted suitability across the suite of modelled species; hence recent sampling of the GAB benthos may have underestimated benthic faunal abundance and diversity in this region.