[gradsusr] GrADS and wgrib2 temperatures in disagreement

[Ivan Toman]

Hello dear GrADS users,

I have a problem. If I extract 2m temperature for a single lat/lon point 
using GrADS, it is in slight disagreement compared to same record 
obtained from grib2 file using wgrib2.

The difference is usually around 1C but can be larger from time to time. 
I'm almost 100% sure I did not overlook anything obvious like selecting 
different time or diffrent lat/lon coordinates, but I leave something 
like that as a theoretical possibility if there is no better 
explanation. Below is an example how I extract temperatures using GrADS. 
Grib2 is WRF file, Lambert projection, and ctl/idx files are created 
using latest g2ctl and gribmap utilities. I note this because it might 
be that projection gets somehow skewed and that GrADS takes data from 
incorrect positions. Is that possible at all?

It seems very important to add that wgrib2 method is in exact agreement 
to the tslist output from WRF, so for now I take that one as "correct" 
one and GrADS figures as "incorrect".

Simple extraction method in GrADS:

'set lat 45.822'
'set lon 16.034'
'set string 1 bl 3 0'
'set t 13'
'd tmp2m-273'
ans = sublin(result,2)
temp = subwrd(ans,4)


Is this difference normal/expected? If it is, what's the reason for 
that? If not, is there anything else where I could search for user error?

Thank you,

Best regards,

Ivan Toman