Reconstructing baselines: use of habitat suitability modelling to predict pre-fishing condition of a Vulnerable Marine Ecosystem

Abstract As industrialized fishing activities have moved into deeper water, the recognition of Vulnerable Marine Ecosystems (VMEs) has become important for the protection of the deep-sea. Our limited knowledge on the past and present distribution of VMEs hinders our ability to manage bottom fisheries effectively. This study investigated whether accounting for bottom fishing intensity (derived from Vessel Monitoring System records) as a predictor in habitat suitability models can (1) improve predictions of, and (2) provide estimates for a pre-fishing baseline for the distribution and biomass of a VME indicator taxon. Random Forest models were applied to presence/absence and biomass of Geodia sponges and environmental variables with and without bottom fishing intensity. The models including fishing were further used to predict distribution and biomass of Geodia to a pre-fishing scenario. Inclusion of fishing pressure as a predictive term significantly improved model performance for both sponge presence and biomass. This study has demonstrated a way to produce a more accurate picture of the current distribution of VMEs in the study area. The pre-fishing scenario predictions also identified areas of suitable Geodia habitat that are currently impacted by fishing, suggesting that sponge habitat and biomass have been impacted by bottom trawling activities.