Ocean oxygen data: how to measure, how to manage?

Abstract A global coordination and continuous synthesis of interoperable data related to biogeochemical Essential Ocean Variables (EOVs) is critically needed to enhance the creation of information products and services to sustainably manage the climate system and ocean health. Among the existing biogeochemical EOVs, data synthesis products—which demonstrate the immense value of data coordination—already exist for carbon-relevant data (e.g. SOCAT, Global Ocean Data Analysis Project), and for methane and nitrous oxide (MEMENTO). The roadmap for building a Global Ocean Oxygen Database and ATlas (GO 2 DAT) (Grégoire et al (2021 Front. Mar. Sci. 1638 )) provides the theoretical basis to increase the interoperability of ocean oxygen data sets, without creating yet another separate repository. The goal is now to advance from the idea of GO 2 DAT to its implementation, building a sustainable, interoperable, and inclusive digital ecosystem for all stakeholders who may use ocean oxygen data. Successful implementation will require (I) the provision of guidance on data acquisition/ocean oxygen measurements, (II) recommended practices for ocean oxygen data management, including metadata requirements, uncertainty and data quality control attribution, (III) development of the ocean oxygen data platform including data flow and application of the recommended practices introduced in I and II, as well as its deep integration with cross-domain data federations such as the Ocean Data and Information System. This document provides an outline of GO 2 DAT’s objective and progress since 2021 and contributes to addressing these three requirements, synthesizing a series of global consultations on recommended practices for marine dissolved oxygen measurements, a working definition of ocean oxygen metadata, proposed data quality control levels and flags, a described novel mechanism for uncertainty attribution to allow the determination of data suitability for different scientific applications, and it concludes with an illustration of the data flow for implementation.