Analysis Ready SST Data for the Oceans

Abstract:

Ocean scientific research and environmental management using observational data face a growing challenge with ever increasing volumes of global and regional data of broadening diversity and complexity. Ongoing efforts to address this complexity are primarily aimed at the scalability of infrastructure to support these large and diverse data, which is converging toward a cloud-based infrastructure. But having data in the cloud only solves one part of a very complex challenge. Often a need arises to access data in a fit-for-purpose manner that is typically synonymous with refactoring oceanographic data into an analysis ready data (ARD) format. Existing examples created by the oceanographic community for the public domain include the MUR Sea Surface Temperature (SST) dataset, and Sentinel-3 datasets as part of the Amazon cloud-based Open Data Registry that provide access to Zarr/COG formatted satellite observations via Jupyter notebooks and binders.

The NASA JPL Physical Oceanography Distributed Active Archive Center (PO.DAAC) has developed a cloud-based data management, services and access infrastructure in support of the upcoming launch of NASA Surface Water and Ocean Topography mission. The service infrastructure includes a suite of discovery, subsetting and access services that lend themselves to restructuring and repackaging input data. As part of the growing PO.DAAC cloud-based dataset inventory, the GHRSST MODIS Aqua SST L2P, a 40 TB dataset, represents an 18 year times series of sustained daily global SST observations, and is thus a prime candidate for ARD.

This presentation will describe such an application for MODIS SST; first using a trade space analysis to determine appropriate ARD content and subsequently using the PO.DAAC infrastructure to develop workflows for input data refactoring and access via community tools including Pangeo and Jupyter notebooks. Besides the motivation described earlier, additional stakeholders for these ARD datasets and workflows include the CEOS SST Virtual Constellation, a working group within CEOS that seeks to improve coordination and production of SST data on an international scale, and the CEOS COAST initiative that seeks to build a fit-for-purpose applications for environmental management at the land/sea boundary and already includes several ARD datasets.