Principal Investigator: Stephen Leroy, Atmospheric and Environmental Research, Inc.

GNSS radio occultation (RO) is a growing industry, which is enabling new capabilities in numerical weather prediction, climate monitoring, and atmospheric process studies. The use of RO data is hindered, however, by the volume of data involved, the latency required to obtain the data, and the organization of the data. Until now, research using RO data has been effectively limited to the few major RO processing centers worldwide. NASA has undertaken a project to migrate NASA instrument datasets to the Amazon Web Services (AWS) cloud through the Cumulus project, but the “cloudification” of big Earth data sets faces major challenges and obstacles of its own, namely that Earth data are not clearly available for analysis using AWS compute resources despite the enormous potential of cloud-native computing.

This project addresses this challenge by making RO data available through AWS, generating AWS-friendly databases and data formats for RO, and performing outreach to the global Earth science community, thereby massively expanding the RO user base. Specifically, the team will migrate all GNSS RO data as processed by three independent retrieval centers into the AWS Earth Data repository, create AWS Dynamo DB databases and cloud-streamable data formats to facilitate the analysis of all RO data (from multiple centers) in AWS compute environments, develop tutorial demonstrations that address four major research areas, and present three workshops to diverse research communities that will introduce those communities to cloud computing as it can be applied to RO data.

The RO data will be contributed by the COSMIC Program Office of the University Corporation for Atmospheric Research (UCAR), NASA’s Jet Propulsion Laboratory, and the Radio Occultation Meteorological Satellite Application Facility (ROM SAF) of the European Organisation for the Exploitation of Meteorological Satellites (EUMETSAT).

The data products to be migrated and the tutorial demonstrations to be constructed are relevant to four scientific disciplines, all of which are active areas of research:

  • The reprocessing of RO data to improve RO as a climate data record,
  • Atmospheric process studies, including those relevant to dynamics and variability of the tropopause and the planetary boundary layer,
  • The inter-comparison of RO data as it is produced by different RO retrieval centers; and
  • How RO data can best be assimilated into numerical weather prediction systems.

These demonstrations will be made available through Github and presented at an RO workshop, at a NASA center, and at a major scientific conference. The demonstrations will emphasize the tremendous capabilities of AWS cloud-native computing, freeing users from the well-known constraints of data volume and long latency associated with obtaining Earth science data sets.

Last Updated
Sep 24, 2020