Data Discoverability and Accessibility Report from July 2019 Workshop on Marine Imagery.

Abstract

There are increasing incentives for marine researchers to share their data, but the will of the marine community to share data has often not yet caught up with our capabilities. Marine imagery and associated annotation, for example, can be collected and analysed with various gear and digital platforms, and there is a large body of legacy imagery and an increasing rate of image collection due to technological advances. As the volume of marine imagery grows, so has the need to establish a national workflow for making it discoverable and accessible. To meet this need, a series of workshops on data discoverability and accessibility were coordinated by the NESP Marine Hub in 2018 and 2019. This report focuses on the Marine Imagery Discoverability & Accessibility Workshop II held on 25 July 2019 at CSIRO, Hobart. The overarching aim of the 2019 marine imagery workshop was to 1) assess the progress made in the past year regarding the discoverability and accessibility of marine imagery and 2) to refine the 2018 recommendations to specify priority, feasibility, and responsibility. The workshop included a range of presentations, activities, and discussions designed to shift participants’ thinking to end users, rather than their own perspectives based on their roles in the marine imagery pipeline. Workshop participants identified the top five barriers to making marine imagery discoverable and accessible: • Limited institutional support or long-term funding for some digital platforms, • Lack of a centralised image and annotation repository or tracking system to ensure FAIR (findable, accessible, interoperable, reusable) data, • No governance or oversight for the entire marine imagery community, • Bottlenecks during processing, imagery upload, and annotation in digital platforms, • Limited communication between major Australian marine imagery groups. For each of the challenges, a set of revised recommendations and actions was developed. The highest-priority actions were to 1) establish a governance body or oversight group to provide broad strategic direction as related to the general marine science community, and 2) establish an ongoing marine imagery node to develop a national workflow to ensure the discoverability and accessibility of marine imagery (i.e. progress the actions listed in the current report). All other recommendations listed in this report are underpinned by the establishment, operation, and collaboration between an oversight group and an implementation group. Importantly, marine imagery and annotation are means to an end, and the primary focus needs to be on understanding and meeting requirements for science and management, not on the sampling gear or digital platforms themselves. There appears much to be gained by AIMS and the IMOS community (which includes AIMS) working closely together to ensure that workflows and infrastructures across their initiatives (e.g. Squidle+, Benthobox/ReefCloud) are compatible and interoperable as required. It is now evident that marine imagery acquisition and annotation, for still and video and for both mono and stereo imagery, is reaching a level of maturity within Australia that would benefit from a more facilitated national approach. The recommendations listed in this report provide such a way forward, but they will require sustained effort and drive to progress, at both the individual and organisational level.