Wish this email find you well!

I used the default 'regrid' function to the RegCM output (wind.nc, which is merged from ua.nc and va.nc), I got wind_lonlat.nc. But when comparing wind.nc and wind_lonlat.nc, I found large biases between them, especially the va component. Did I use the correct method to regrid the wind (vector variables)? If not, could you help me with some solutions to this problem? Thank you so much for your help in advance!

Here are major steps of the analysis progress:

>../RegCM-4.3.5.6/bin/regrid wind.nc -89.75,89.75,1.5 -179.75,179.75,1.5 bil

  Then I got the output: wind_lonlat.nc

>cdo info wind.nc
   -1 :  Date     Time   Level Gridsize    Miss :     Minimum        Mean     Maximum :ParameterID
     1 : 2009-12-31 00:00:00     850    15087       0 :     -8.9881      12.456      30.786 : -1        
     2 : 2009-12-31 00:00:00     850    15087       0 :     -10.104     0.44314      17.616 : -2        

>cdo info wind_lonlat.nc
 -1 :     Date     Time   Level Gridsize    Miss :     Minimum        Mean     Maximum :ParameterID
     1 : 2009-12-31 00:00:00     850    28800   27137 :     -8.9192      13.116      30.496 : -1        
     2 : 2009-12-31 00:00:00     850    28800   27137 :     -9.1465    0.078703      17.495 : -2        

Look forward to your reply!

Best regards!

Lang