Write script to convert wrf output to CF-convention about wrf-downscaling HOT 5 CLOSED

If handling multiple domains is tricky, start with the inner domain (d01 I believe).

We only need the 2m / near surface (above the canopy) variables for the crop model

from wrf-downscaling.

HI David,
looking at wrf file/d01 - need to isolate the required vars: candidates are:
air_temperature-> T2:description = "TEMP at 2 M"
air_pressure -> PSFC:description = "SFC PRESSURE" ;

specific_humidity -> MIDRH_MIN:description = "Min Mid-level relative humidity" ;
( nb all values set to zero in d01 ? )

surface_downwelling_shortwave_flux_in_air->
SWDOWN:description = "DOWNWARD SHORT WAVE FLUX AT GROUND SURFACE"

precipitation_flux -> choose from :
RAINC:description = "ACCUMULATED TOTAL CUMULUS PRECIPITATION" ;
RAINSH:description = "ACCUMULATED SHALLOW CUMULUS PRECIPITATION" ;
RAINNC:description = "ACCUMULATED TOTAL GRID SCALE PRECIPITATION" ;
SR:description = "fraction of frozen precipitation" ;
PREC_ACC_C:description = "ACCUMULATED CUMULUS PRECIPITATION OVER prec_acc_dt PERIODS OF TIME" ;
PREC_ACC_NC:description = "ACCUMULATED GRID SCALE PRECIPITATION OVER prec_acc_dt PERIODS OF TIME" ;

Also the "Times" needs to be converted from date stamps to "days since 2008-01-07_00:00:00" (fro example in d01)

from wrf-downscaling.

• precipitation_flux = RAINC+RAINNC
  • ref 1
  • "RAINC includes all phases of convective precip, and RAINNC includes all phases of non-convective precip. So total precip is given by RAINC + RAINNC" ref 2
• Specific humidity = Q2?

from wrf-downscaling.

HI David,
below are the required vars to be extracted:

     float T2(Time, south_north, west_east) ;
     float PSFC(Time, south_north, west_east) ;
    float Q2(Time, south_north, west_east) ;
    float SWDOWN(Time, south_north, west_east) ;
    float RAINC(Time, south_north, west_east) ;
    float RAINNC(Time, south_north, west_east) ;

the coordinate vars :
XLAT(Time, south_north, west_east) ;
XLONG(Time, south_north, west_east) ;
Are overly complicated and don t vary with respect to "Time" , So we can simply use the first frame. ie
XLAT=XLAT( 0,:,:);
XLONG=XLONG( 0,:,:);

In the file wrf/brazil/wrfout_d01_2008-01-07_00:00:00 , there are 9 time frames in 3 hour intervals. Have you got an estimate of the number of time frames in later output , and will the interval be 3 hours ?
I am assuming that you wanna calculate the vars at a given lat,lon ?
If the time var is large then the dimensions will need repermuting from (Time, south_north, west_east) to (south_north, west_east,Time)

from wrf-downscaling.

see new comments on github !

On 17/10/14 05:04, David LeBauer wrote:

• precipitation_flux = RAINC+RAINNC (ref 1
  http://www.meteo.unican.es/wiki/cordexwrf/OutputVariables)
• "RAINC includes all phases of convective precip, and RAINNC
  includes all phases of non-convective precip. So total precip is
  given by RAINC + RAINNC" ref 2
  http://mailman.ucar.edu/pipermail/wrf-users/2012/002781.html
• Specific humidity = Q2?

—
Reply to this email directly or view it on GitHub
#2 (comment).

from wrf-downscaling.