From NCL to Python: The Triumphs (and struggles) of Upgrading a Tropical Monitoring Page for Air Force Operations.

NC State’s North Carolina Institute for Climate Studies (NCICS) has developed a research tool for monitoring tropical conditions around the globe: ncics.org/mjo. The website brings together a unique blend of observational and forecast data to analyze and predict subseasonal changes in the tropical climate system. Variables used include zonal and meridional wind, velocity potential.stream function, and outgoing longwave radiation (OLR). The novel aspect of the tool is applying Fourier filtering to a combination of observation and CFSv2 forecasts in real-time to identify, track, and predict wave features, including the Madden Julian Oscillation (MJO), equatorial Rossby waves, and Kelvin waves. The website includes longitude-time Hovmollers along with global and regional maps, depicting past and future conditions. Since its release in 2011, it has served hundreds of unique users every month, including the 14th Weather Squadron, the Climate Prediction Center, the National Hurricane Center, the private sector, and meteorological agencies in more than 20 countries.

The 14th Weather Squadron, located alongside NCICS in Asheville, North Carolina, is the Air Force’s only climate operations unit that collects, protects and exploits authoritative climate data to optimize military and intelligence operations. Interest grew in the tropical monitoring product, and tasks were set to push the code base from a research environment at NCICS to Air Force operations. One of the requirements was for the code to be compatible with a Python 3 distribution supported by Enthought. Since the code was initially written in the NCAR Command Language (NCL), the entire system needed to be rewritten to match the requirements. This presentation describes the process of converting the product from research to operations, including decisions made and packages explored to match results.