Prioritisation of ocean biodiversity data collection to deliver a sustainable ocean.

Fundamental ecological questions about the distribution of ocean life remain unanswered, hindering both the effective management of the ocean, and our comprehension of life on this planet. The benthic and pelagic realms are subject to different methods of study, and to understand where to best focus effort, a thorough understanding of existing information is required, allowing identification of critical knowledge gaps. Open-access data repositories provide a valuable means to identify such gaps; however, these repositories are subject to challenges in separating benthic from pelagic data. Here we demonstrate an automated data pipeline for extracting and separating benthic from pelagic data in open-access repositories. By stratifying data against essential ocean variables in a critical gap analysis, we show that large spatial and taxonomic biases exist in both the benthic and pelagic global datasets, favouring depths shallower than ~100 m, the northern hemisphere, and vertebrate species. The newly compiled, cleaned, and classified dataset is used to identify areas of chronic under sampling and high-priority regions for exploration. We argue that coordinated strategic prioritisation of sampling is needed to support modelling and prediction, enabling us to better manage our oceans and comprehend life on Earth.