Scientists need access to research data from model generated gridded products that are increasing in number and continue to grow in volume. This trend can create non-science challenges for the users that include lengthy download times, the need for greater amounts of local storage and computing resources, and complexities that arise from differing data formats. Recent improvements in gridded data preparation service, through the Research Data Archive (RDA) at NCAR, are designed to address these challenges.

New IT infrastructure has made this new service possible by co-locating the RDA data with a data analysis cluster and super computing resources. Together the hundreds of terabytes of storage and computing capacity uniquely equip NCAR with the capability to do server-side data preparation from the RDA collections. Operationally updated metadata databases complement the computing infrastructure to support standardized subset-request web interfaces, always up to date information for the users, and efficiently driven backend data processing tools.

Currently these capabilities furnish users with the options to select individual parameters, temporal and spatial domains, and standard NetCDF output formats from several of the RDA's large suite of global gridded reanalyses, and operational model forecast datasets. Planned future improvements include the addition of selectable grid-interpolation, faster turn around time on requests, a command line interface for NCAR HPC users, and implementation across more datasets. This presentation will discuss how this scalable system works and what datasets are currently available, as well as what other collections are slated to be made available in the future.