Towards a Cloud Computing Environment for Global Near Real-Time Satellite Cloud Processing for GMAO GEOS-5 Data Assimilation

Satellite ClOud and Radiative Property retrieval System (SatCORPS) processes and derives near real-time (NRT) global observations of clouds, aerosols, and surface properties from operational geostationary (GEO) satellites. These products are being used in NRT to improve climate and weather applications, forecast model, aircraft icing warnings, and to support scientific field campaigns.

Next generation satellites, such as the Japanese Himawari-8 and the upcoming NOAA GOES-R(-16), present challenges for NRT data processing and product dissemination due to the increase in temporal and spatial resolution. The volume of data is expected to increase to approximately 10 folds. This increase in data volume will require additional IT resources to keep up with the processing demands to satisfy NRT requirement to process and deliver data products in an operational 24/7 environment where timely delivery, reliability, and availability are very critical.

To anticipate and meet these computing requirements, the NASA Langley Research Center Atmospheric Science Data Center (ASDC) is building and testing a private and public facing cloud for users in the Science Directorate to utilize as an everyday production environment. This paper will focus on the use of the ASDC private cloud computing platform to produce high temporal and spatial SatCORPS products for use in the evaluation of NRT data assimilation to improve the Global Modeling and Assimilation Office (GMAO) GEOS-5 system. We will describe this processing system workflow for ingesting, processing, and dissemination of SatCORPS products in this hybrid processing system. Lessons learn from working on both AWS GovCloud and ASDC Cloud will be discussed and presented including the benefits, similarities, and differences that could impact decision to use cloud computing.