Chris Hill led a virtual Town Hall meeting at the Ocean Sciences Meeting entitled “Open Science Analysis of Petabyte Scale Ocean and Ocean-Atmosphere Models with Open Source Cloud Tools.” If you attended the meeting, you can watch it here (login required).
Details: This Town Hall presents new software tools to catalyze the analysis of global ocean-only and ocean-atmosphere coupled models with resolutions of a few kilometers. Computational capacity has recently advanced so that multiple projects have achieved global ocean circulation model solutions that resolve down to a few kilometers. These solutions provide eye-opening perspectives on the high frequency and high wavenumber phenomena that influence planetary dynamics. Analyzing and making sense of the petabyte scale virtual datasets that emerge from these simulations is challenging, however. This obstructs a broad community of researchers worldwide from fully utilizing these data. This is especially problematic for researchers without access to large computing resources. This Town Hall will demonstrate open, accessible tools to analyze petabyte-scale ocean model solutions in the cloud and solicit feedback on them. The tools include Eulerian and Lagrangian diagnostics of common oceanographic and dynamical variables with interactive and responsive visualization capabilities for rapid exploration. The tools can be accessed via a web browser, including from low-end devices like tablet computers. They are being developed as part of a US National Science Foundation project to facilitate easy access to state-of-the-science computational oceanography for users at all levels of experience. The project also includes cyber-infrastructure to employ petabyte-scale ocean and atmosphere circulation model data with artificial intelligence and machine learning techniques. In the Town Hall we will demonstrate the analyses that are possible and include hands-on activities where participants can try out the system for themselves. The demonstrations will leverage kilometer scale ocean-only and coupled model solutions and analysis tools from the ECCO-MITgcm-GEOS-Pangeo-SciServer collaborative group that spans NASA JPL, NASA Goddard, NASA Ames, MIT, Johns Hopkins and Columbia University.
