Justin Monroe, from the ARM Data Quality Office at the University of Oklahoma, led a tutorial on ARM NetCDF files and how to navigate the ARM Python notebook during the 2014 Joint User Principal Investigator Meeting.
Most ARM data are made available as time-series data in netCDF (network Common Data Form), which is a set of libraries and self-describing, platform-independent data formats that support creating, accessing, and sharing array-oriented scientific data. It is relatively compact, appendable, and capable of storing descriptive “metadata” along with measurement data.
NetCDF includes data access libraries for a wide variety of programming languages. Example software programs for reading and writing netCDF files using C, C++, Fortran, IDL®, JavaTM, MATLAB®, PythonTM, and Perl, as well as more information about netCDF, including downloads, documentation, and frequently asked questions, can be found on the Unidata website.
Some data files also contain measurements distributed over a region from a single time (e.g., satellite images), which are stored in HDF (Hierarchical Data Format) files. Raw model data files may also be stored in GRIB (GRIdded Binary) and GRIB2 file formats. For more information, see Reading netCDF, HDF, and GRIB Files.