Conditions for the production of robust deep sea science to inform and support ocean action

To effectively support ocean action, it is crucial to make the most of cutting-edge science. However, this "frontier science" comes with significant challenges, because of its high cost and limited deployment capabilities. While technologies such as remotely operated vehicles (ROVs), smart cables, and ocean observatories can provide detailed data, these tools are typically deployed at only a few fixed points. As the data remains geographically restricted it becomes difficult to generalize the findings to broader ocean processes. These limitations pose a challenge in validating general hypotheses about the biological, physical, oceanographic, and geological processes that occur at varying spatial and temporal scales. While knowledge of local dynamics is essential for effective decision-making regarding ocean monitoring, conservation, and resource management, the absence of broader hypotheses complicates our ability to distinguish between context-specific and universal processes. To address these gaps in knowledge, some argue for further technological advancements at fixed points, while others advocate for more extensive exploration across larger spatial and temporal scales. The key distinction between these approaches lies more in budget allocation than in technological development. The debate ultimately centres around whether to invest in replicating observations over broader geographic ranges or to focus on enhancing technologies. An effective exploratory approach must use a variety of cost-efficient, easy-to-deploy sampling and observation tools—such as trawls, dredges, Niskin bottles, eDNA samplers, towed cameras, and CTDs—across multiple spatial and temporal scales. Developing advanced technology and having an exploratory approach are thus complementary rather than conflicting. Both approaches should aim to integrate spatial and temporal scales and balance the cost-benefit equation of at-sea operations. Scientific exploration of the deep sea is inherently cumulative, requiring long-term strategies for data collection and storage. Repositories that store accumulated data and samples are vital for making the information accessible and for supporting future discoveries. Fixed-point observatories contribute to this effort by providing standardized, repeatable data. If deployed across various regions in a coordinated, multiscale framework, they can contribute to building a comprehensive understanding of deep-sea processes. To achieve generalizable and reproducible results about deep-sea dynamics, it is essential to: (i) maximize opportunities for discovering unknown species, habitats, mineral deposits, ecosystems, and processes, and (ii) enhance the spatial coverage of exploration to establish a robust baseline that allows new findings to be interpreted in a global context. A good example of this approach can be seen in physical oceanography, where systematic data from distributed Argos buoys have been collected globally, creating a reference grid despite some gaps in the Southern Oceans. In contrast, biological and geological knowledge still lacks such a global framework, largely due to the complex and heterogeneous nature of the deep-sea environment. Given these insights, the priority should be to establish a multiscale, imbricated strategy that provides a comprehensive baseline for every fields of deep-sea research. This baseline will be essential for designing experiments that inform better ocean governance and conservation practices. Here, we present the elements gathered within the framework of the IRD collective assessment on the deep-sea knowledge and governance.