Version 1.3.2  February 2012

Send comments and questions to Steve at [email protected]

wrf2gem reads WRF output in netCDF format and produces a GEMPAK file according
to the specifications in convert.nl.  Some diagnostics can be output on
isobaric surfaces, and GEMPAK can handle other diagnostics or interpolations.
wrf2gem is written in standard Fortran 90.


* Compilation Instructions *
Hopefully, it's just a matter of typing:
make

However, you may want or need to make changes to the Makefile.  Things to watch
out for:
1) The FC setting should match your Fortran 90 compiler.
2) You may want other compiler flags (FFLAGS).
3) Your libnetcdf.a and gemlib.a may be in different locations than mine.
4) Depending on your GEMPAK version, you may need to link additional libraries
   in addition to gemlib.a.  I'm currently using version 5.7.2p2.
5) The module files created during compilation of your netCDF library need to
   be findable by your compiler.


* Description of convert.nl *
Here are the variables you can set in the namelist:
numFiles  -- number of WRF output filenames you would like to process
		(i.e., the number of entries in fName)
fName     -- WRF output filename(s) from which to read data
[&INITTIME no longer needed.]
gFName 	  -- What your GEMPAK file is or should be called
maxGrd    -- The maximum number of grids allowed in a new GEMPAK file
pack      -- Store grids using 16-bit GRIB packing (.true.) or do not (.false.)
overwrite -- Add to and/or overwrite existing GEMPAK files (.true.) or do not
	     (.false.).
num       -- The number of variables/diagnostics to output (i.e., the number of
        	entries in outFields)
outFields -- Which variables/diagnostics to output.  The IDs are as follows:

Variables
 1  - HGT: Terrain height (m)
 2  - T2: 2-m temperature (K)
 3  - Q2: 2-m water vapor mixing ratio (kg/kg)
 4  - ZNU: Eta coordinate on mass (half) levels
 5  - PB: Base state pressure (Pa)
 6  - P: Perturbation pressure (Pa)
 7  - T: Perterbation potential temperature (K)
 8  - U: U component of wind (m/s)
 9  - V: V component of wind (m/s)
 10 - W: W component of wind (m/s)
 11 - U10: U at 10 m (m/s)
 12 - V10: V at 10 m (m/s)
 13 - QVAPOR: Water vapor mixing ratio (kg/kg)
 14 - QCLOUD: Cloud water mixing ratio (kg/kg)
 15 - QICE: Ice mixing ratio (kg/kg)
 16 - RAINC: Accumulated cumulus precip (mm)
 17 - RAINNC: Accumulated grid-scale precip (mm)
 18 - PH: Perturbation geopotential (m^2/s^2)
 19 - PHB: Base-state geopotential (m^2/s^2)
 20 - ZNW: Eta coordinate on w (full) levels
 21 - LU_INDEX: Land use category
 22 - MU: Perturbation dry air mass in column (Pa)
 23 - MUB: Base-state dry air mass in column (Pa)
 24 - TH2: 2-m potential temperature (K)
 25 - QRAIN: Rain water mixing ratio (kg/kg)
 26 - QSNOW: Snow mixing ratio (kg/kg)
 27 - QGRAUP: Graupel mixing ratio (kg/kg)
 28 - PBLH: PBL height (m)
 29 - SST: Sea surface temperature (K)
 30 - TSK: Surface skin temperature (K)
 31 - SWDOWN: Downward short wave flux at ground surface (W/m^2)
 32 - GLW: Downward long wave flux at ground surface (W/m^2)
 33 - GRDFLX: Ground heat flux (W/m^2)
 34 - HFX: Upward heat flux at the surface (W/m^2)
 35 - QFX: Upward moisture flux at the surface (W/m^2)
 36 - LH: Latent heat flux at the surface (W/m^2)

Some of these variables (like 4 and 20) don't make sense to output to a GEMPAK
file and will be ignored.

Diagnostics
 101 - Surface pressure (hPa)
 102 - Pressure at 2 m (hPa)
 103 - Sea level pressure (hPa)
 104 - Convective precipitation accumulated over one hour (mm)
 105 - Total precipitation accumulated over one hour (mm)
 106 - Total precipitation accumulated over three hours (mm)
 107 - Total precipitation accumulated over six hours (mm)
 108 - Total accumulated precipitation (mm)
 109 - Pressure (hPa)
 110 - Potential temperature (K)
 111 - Unstaggered U component of wind (m/s)
 112 - Unstaggered V component of wind (m/s)
 113 - Cloud water + Ice mixing ratio (kg/kg)
 114 - Unstaggered W component of wind (m/s)
 115 - 2-m temperature including at initial time (K)
 116 - 2-m water vapor mixing ratio including at initial time (kg/kg)
 117 - U at 10 m including at initial time (m/s)
 118 - V at 10 m including at initial time (m/s)
 119 - Lifted Index using lowest 4 levels (K)
 120 - Mean relative humidity (%) in 850-500-hPa layer
 121 - Geopotential height (m) on w (full) levels
 122 - 2-m specific humidity (kg/kg)
 123 - Specific humidity (kg/kg)
 124 - Temperature (K)
 125 - Geopotential height (m) on half levels
 126 - Dry air mass in column (Pa)
 127 - Most unstable CAPE w/ virt. temp. correction (J/kg)
 128 - Lifted parcel level of the most unstable parcel (m)
 129 - Surface-based CAPE w/ virt. temp. correction (J/kg)
 130 - Surface-based CIN w/ virt. temp. correction (J/kg)
 131 - Mixed-layer CAPE w/ virt. temp. correction (J/kg)
 132 - Mixed-layer CIN w/ virt. temp. correction (J/kg)
 133 - Mixed-layer LCL (m)
 134 - Mixed-layer LFC (m)
 135 - U component of Bunkers storm motion (m/s)
 136 - V component of Bunkers storm motion (m/s)
 137 - 0-1-km storm relative helicity (m2/s2)
 138 - Column integrated precipitable water (cm)
 139 - Reflectivity (dBZ)
 140 - Composite reflectivity (dBZ)

Diagnostics interpolated to isobaric levels
 501 - Temperature (K)
 502 - U component of wind (m/s)
 503 - V component of wind (m/s)
 504 - W component of wind (m/s)
 505 - Geopotential height (m)
 506 - Specific humidity (kg/kg)

pb        -- Bottom pressure level for isobaric diagnostics
pt        -- Top pressure level for isobaric diagnostics
dp        -- Pressure interval for isobaric diagnostics


* Adding Diagnostics *
The conversion code is designed to be extended fairly easily (I hope).  Here's
what you should do if you'd like to add a diagnostic or variable.
1) Determine if you're adding a variable (i.e., simply moving grids from the
   WRF output to the GEMPAK file with no calculation whatsoever) or a
   diagnostic (everything else).

2) The variable case is easy.  Modify registry.f90 as follows:
   a) Add a comment after this line:
        ! 36 - LH: Latent heat flux at the surface (W/m^2)
   b) Increment the value of LastWRF in this line:
        integer, parameter :: LastWRF = 36, FirstDiag = 101, LastDiag = 140,  &
   c) Insert an identifier for your variable after this line:
       LaHeat    = 36
   d) Add an appropriate initialization statement after this line:
        wrfVar(LaHeat)    = wrf_var(1, 0, -1, None, "LH", "LH", .false.)
      The wrf_var type definition (around line 195) will be helpful here.

3) For the diagnostic case, make modifications to registry.f90 as follows:
   a) Add a comment after this line:
        ! 140 - Composite reflectivity (dBZ)
   b) Increment the value of LastDiag in this line:
        integer, parameter :: LastWRF = 36, FirstDiag = 101, LastDiag = 140,  &
   c) Insert an identifier for your diagnostic after this line:
        CompRefl  = 140
   d) Add an appropriate initialization statement after this line:
        diagVar(CompRefl)  = diag_var(1, 0, -1, None, CompRefl, "CREF")
      The diag_var type definition (around line 205) will be helpful here.
   e) Add your diagnostic to the "select case (var%diagID)" block in subroutine
      diag_out.
   f) Write a simple function for your diagnostic and add it before the
      "end module registry" line.  Call it diag_identifier to match the others.

4) If your diagnostic is particularly complex, you may need to add functions
   to diagnostics.f90 as well.  If you do this, you'll need to make these
   additional changes:
   a) Add any new functions you call from your diag_identifier function to the
      list of functions in the following statement in registry.f90:
  use diagnostics,  only: vint, hydro_interp, mix_ratio_to_spec_hum,  &
                          temp_from_theta_p, surface_pres, density, sea_pres, &
                          lifted_index, mean_rh, unstag_hght, cape,  &
                          sfcape => sbcape, mlcape, lcl_lfc, storm_motion,  &
                          storm_rel_hel, calcdbz
   b) Add your function name to the following statement in diagnostics.f90:
  public :: vint, hydro_interp, mix_ratio_to_spec_hum, temp_from_theta_p,  &
            surface_pres, density, sea_pres, lifted_index, mean_rh,  &
            unstag_hght, cape, sbcape, mlcape, lcl_lfc, storm_motion,  &
            storm_rel_hel, calcdbz
      Notice how this statement matches the corresponding statement from
      registry.f90.
   c) Add your new function(s) to diagnostics.f90.  Public functions go first,
      then private functions.

5) Recompile, and see if it works!

6) If you think your diagnostic should be a part of the "official" version of
   wrf2gem, send me your new functions and I'll incorporate them.

7) Adding a diagnostic for output on isobaric surfaces is a bit more
   complicated, but it follows a pattern similar to that for regular
   diagnostics.  Examine the pres_diag_out subroutine if you want to make an
   addition here.


* Limitations *
1) The polar stereographic projection is not yet supported.
2) The x-hr precipitation accumulations (104-107) will be set to zero for times
   at which the accumulation interval spans more than one history file.
   Exception: If each history file contains only one output time, these
              diagnostics work fine.
   Workaround: Output total accumulated precip (108), then use, e.g., 
               gfunc = sub(ptot^f27,ptot^f21) to get a 6-hr accumulation valid
               at forecast hour 27 if each history file contains a day's worth
               of data.