bartbrashers
New member
This is a very obvious use-case to me, but I can't find any evidence that someone has done this yet.
For hindcasting (simulating the past, for multi-year periods) we often use analysis nudging. For domains with finer grid spacing than the input dataset (which is most domains for me) we objectively re-analyze the met_em files using OBSGRID, which produces metoa_em files. These are then fed to REAL, which produces the desired wrffdda_d0? files.
OBSGRID hasn't been updated in years. WRFDA should do a much better job even using the relatively simple 3DVAR option. Can it be used to provide better wrffdda_d0? files, or even metoa_em-style files?
For hindcasting (simulating the past, for multi-year periods) we often use analysis nudging. For domains with finer grid spacing than the input dataset (which is most domains for me) we objectively re-analyze the met_em files using OBSGRID, which produces metoa_em files. These are then fed to REAL, which produces the desired wrffdda_d0? files.
OBSGRID hasn't been updated in years. WRFDA should do a much better job even using the relatively simple 3DVAR option. Can it be used to provide better wrffdda_d0? files, or even metoa_em-style files?