Analyzing and Visualizing Next Generation Climate Data

The analysis of petascale data from climate models raises a number of new challenges. One involves the development of alternatives to models based on longitude-latitude grids, which exhibit a singularity at the north and south poles. These grid singularities can cause numerical instabilities and impose severe time step restrictions. One novel solution to this problem involves the use of assemblies of subgrids (tiles) to form a so-called Mosaic grid. However, at present, no data analysis or visualization tool is capable of interpreting Mosaic data. This project will extend the Climate Data Analysis Tool (CDAT) to interpret and visualize petascale data on Mosaic grids. The approach will rely on the recently proposed Mosaic grid speci¿cation for storing connectivity information between Mosaic tiles. The need to plow through petabytes of archived data will be addressed using the parallel computing facility in IPython. A three-dimensional visualization capability will be provided by using VisIt to harness large remote visualization resources. Commercial Applications and other Benefits as described by the awardee: The extensions brought to CDAT should benefit researchers that need to apply large computational resources to climate problems. The technology also should be applicable to other DOE sponsored research, including fusion, where the magnetic axis plays a similar role to the north and south poles.