Since 2021, GFS is running at very high horizontal resolution, so we need to treat the 
GAVA, LTHAD, LAD, and LDIA terms of the approximate P-S Development Equation differently
than we would for GFS thinned case studies before 2021.

   [1] for case studies before 2021, refer to this web page

http://www.atms.unca.edu/dmiller/atms410/fnlpro_garpfuncs2020.txt

for the proper way to handle the GAVA, LTHAD, LAD and LDIA terms.

-------------------------

   [2] for newer case studies (2021 to present)...

*** Do NOT smooth geopotential height contours (no "sm9s( )" operators around HGHT) ***

Geostrophic Relative Vorticity (GRV)...    [SF = scale factor]
     sm9s(sm9s( VOR(GEO) )), SF=5

700 hPa vertical velocity (this is NOT omega, but w [m s^-1]), related to LDIA...
     sm9s(sm9s( VVEL@700%PRES )), SF=2

Geostrophic Absolute Vorticity (GAV)...
     sm9s( sm9s( AVOR (GEO) )), SF=5

horizontal divergence (needed at the 300 hPa level)...
     sm9s( sm9s( DIV(OBS) )), SF=5

Geostrophic Absolute Vorticity Advection (GAVA, needed at the 500 and 300 hPa levels)...
     sm9s( sm9s( ADV ( AVOR (GEO), GEO ) )), SF=9

horizontal temperature advection (needed at the 300 hPa level)...
     sm9s( sm9s( ADV ( TMPK, WND ) )), SF=3

nearly LTHAD (nLTHAD)...
     sm9s( sm9s( LAP( ADV (SUB(HGHT@500%PRES,HGHT@1000%PRES), GEO@1000%PRES )) )), SF=13

nearly LAD (nLAD)...
     sm9s( sm9s( LAP( MUL( VVEL@700%PRES, MUL(DDEN(PRES,TMPC), -1.))) )), SF=13