dwfncar
commented
10 years ago Grid-Stat in met-5.0 and earlier versions employed the following logic... if both the forecast and observation fields being evaluated were precipitation (determined by calling VarInfo::is_precipitation()), fill the OBTYPE column with MC_PCP. Otherwise, fill it with ANALYS. The question is whether or not that logic should be preserved if the user has not specified an meaningful (non-NA) entry for obtype in the config file. I discussed this with Michelle and she thinks for simplicity that we should just pass through the config file entry, setting the default value to ANALYS. The mesoscale modelling group has never used the obtype column for filtering gridded verification output. I'll proceed with that advice, just passing through the config file entry. But in order to prevent the regression tests from failing, I'll edit the Grid-Stat config files for precip to set the obtype as MC_PCP. by johnhg
During John and Tara's trip to NCEP in November 2014, they saw that the VSDB files created by Ying Lin for verification of precipitation contained differing values in the OBTYPE column. For example, some lines contained Stage2 and other Stage4. There is currently no way to control the contents of the OBTYPE column in the .stat output files from Grid-Stat. This task is to add an configuration file entry named "OBTYPE" for the user to specify the type of observations used for the verification. Grid-Stat should just write the config file entry into the output OBTYPE column. [MET-466] created by johnhg