Test the NEXRAD with HAFS

[JingCheng-NOAA]

Started with the bufr2ioda yaml that from RDASApp issue#58 .
HAFS NEXRAD data was directly obtained from bufr tank and data were prepared with 3 hour time window. The file includes both radial wind and reflectivity data.

[JingCheng-NOAA]

The bufr2ioda processing json file contains following messages:
[ "NC006010", "NC006011", "NC006012", "NC006013", "NC006014", "NC006015", "NC006016", "NC006017", "NC006018", "NC006019", "NC006020", "NC006021", "NC006022", "NC006023", "NC006024", "NC006025", "NC006026", "NC006027", "NC006028", "NC006029", "NC006030", "NC006031", "NC006032", "NC006033", "NC006040", "NC006041", "NC006042", "NC006043", "NC006044", "NC006045", "NC006046", "NC006047", "NC006048", "NC006049", "NC006050", "NC006051", "NC006052", "NC006053", "NC006054", "NC006055", "NC006056", "NC006057", "NC006058", "NC006059", "NC006060", "NC006061", "NC006062", "NC006063" ]
Message type that larger than NC006040 are for lvl2 reflectivity data. The corresponding python script only process radial wind. Since in HAFS nexrad data, both radial winds and reflectivity data are presented. Thus, the json file used by HADSApp should remove the message types that larger than NC00604*

[JingCheng-NOAA]

Increasing the memory request in bash file to 80G helps to solve the out-of-memory issue.

[JingCheng-NOAA]

Initial test indicates a missing variable (MetaData/cosazm_costilt). Consider to add this variable into IODA nc file through bufr2ioda python script.

[JingCheng-NOAA]

Added three variables into bufr2ioda: MetaData/coszam_costilt, MetaData/sinazm_costilt, MetaData/sintilt

[JingCheng-NOAA]

Another test run with modified IODA NC file that contains updated variables can pass previous error. However, the run stopped with another message

GeoVaLs field w has no known link to fields in model state