Python Strategies for NASA GEOS-5 Quasi-Operational Processing

Agencies in the Federal Government are transitioning to open source software alternatives as a cost-efficient approach to enabling scientific research. This trend is fuelling the use of the Python programming language--Python is open source, flexible, powerful, and in high demand. The question remains: is Python robust enough to meet the challenges of today's scientific research and computational environments at NASA? The use of Python in support of the Goddard Earth Observing System (GEOS-5) atmospheric data assimilation system, at the Global Modeling and Assimilation Office (GMAO) offers key insight into how Python can be used for complex processing on quasi-operational systems. GMAO staff have demonstrated that Python is sufficiently robust to support a wide range of continually evolving requirements. This study highlights how using Python aids in ensemble forecasting, multi-year reanalyses, field-campaign support, and intercomparisons between different data types for the GMAO as a precursor to the broader issue of operational programming through Python.