History of Upper Air definition ------------------------------- The first meeting focused on defining what data falls into this category and how the data are stored with regards to time. Of course, there are many types of upper air data including RAOB, CLASS, profiler, satellite sounders, etc. Our focus turned to specifying a definition for RAOB and CLASS data with the understanding that other upper air data definitions could be derived using the upper air definitions as a starting point. We also decided that the unlimited dimension would a record number which allows for multiple reporting times in one file. Several strawmen netCDF definitions were distributed and discussed and the members of NUWG decided to start with the definitions proposed by FSL. The major points of discussion included how to dimension the levels and storing of surface and derived data. It was decided that the number of levels will be defined by the user. Casual observation has shown for RAOB data, mandatory levels to 21 and significant levels up to 130 for sigT and 65 for sigW. For CLASS data, the mandatory level will contain the surface data (making the dimension 1) and the significant levels can be up to 200 levels. It is imperative for the user of a data set created with this definition to check the level dimensions explicitly. It was also agreed that only the reported data would be stored and commonly calculated fields such as computed sigT level heights would be optional. The surface data will be stored as the first mandatory level and is followed by the standard isobaric levels as defined by the WMO. The definition reflects the general conventions followed by the NUWG group in regards to the naming of variables, use of attributes, the representation of time, and station identification. These are considered the minimum varialbes required in files used for upper air data exchange. ------------------------------------------------------------------------------- by Susan Jesuroga for NUWG nuwg@comet.ucar.edu updated on January 4, 1994