Mesoscale Parallel Experiment Change Log

Experiment Name:  Eta-32 with new radiance processing and modified cloud microphysics, EDAS precipitation assimilation, gauge adjustment of Stage IV precip analyses, and new radiation physics

When Who What
2003/10/31/00 Ferrier
  • Fixed bug in parallel and control eta-32 (ETAV) to use the correct zenith angle for outgoing solar radiation at the surface, and outgoing solar radiation at the top of the atmosphere.
    • 2003/10/24/00 Ferrier, H.-M. Lin
  • Routine GSMCONST was changed by modifying two tunable parameters that act to reduce the amount of supercooled liquid water.
    • 2003/10/21/12 Ek, Mitchell
  • Changed parameter CZIL (in routine SFCDIF) from 0.20 to 0.10
    • 2003/10/21/00 Ferrier
  • Fixed bug in control eta-32 (ETAV) to use the correct zenith angle for the incoming solar fluxes at the surface. Fix was already in ETAY parallel.
    • 2003/10/10/12 Ferrier, H.-M. Lin
  • New radiation phsyics added to parallel, see main page for details
    • 2003/10/07/00 Ek, Mitchell
  • Changes to the land-surface physics:

    In the case of snow, surface moisture flux now accounts for snowcover patchiness, with evaporation (sublimation) from the snowpack (weighted by the snowcover fraction), and surface evaporation (canopy evap, bare soil, transpiration) elsewhere (weighted by 1-snowcover fraction).

  • Restarted cycled EDAS from ETAV control run
    • 2003/10/06/12 Rogers
  • Changed eta forecast model parameters for all eta-32 runs so that physics is called every 3 time steps (270 sec) instead of every 6 time steps (540 sec) to make 32-km runs more similar to the production eta-12
  • Changed the horizontal diffusion coefficient COAC to 0.1 (was 0.25) to make the amount of diffusion applied to the eta-32 consistent with that applied to the operational eta-12
    • 2003/10/04/00 Y.Lin
  • Control 32-km EDAS (ETAV) restarted from GDAS atmospheric states and ops eta-12 land states at 12Z 3 October 2003 due to the following bug (which was only in ETAV):

    At each physics time step, ADJPPT computes the beginning and ending time of the time step to determine how much of the observed hourly rainfall should be from this time step. If this timestep falls entirely within an hour, then this amount is simply

    hourly rainfall * (length of phys time step in sec)/3600.

    The (length of physics time step in sec)/3600. is the 'time step as fraction of the hour'.

    If this timestep straddles two different hours, we need to compute the amounts of time from each of the two hours that this time step is consist of, i.e. 'fraction1' and 'fraction2', then use these two amounts to compute how much each hour's observed precipitation contributes to the amount of precipitation falling into this time step.

    In the 32-km control EDAS, fraction2 was not computed correctly because the fix which was placed in the production 12-km EDAS was inadvertently left out of the 32km version of the new Eta forecast code implemented in July 2003. In the case of the 32km eta parallel, where the physics time step is 540 seconds, during the timestep that straddles the 1st and 2nd hours (3,240 sec to 3,780 sec) fraction1 and fraction2 should be 0.1 and 0.05, respectively, but computed by the old executable, fraction2=0.90. In the 2003092812 ETAV run, during the first EDAS segment, precip obs at the 'test point' was 4.6mm and 10.4mm, respectively, so the amount falling between this timestep (3,240 sec to 3,780 sec) should be 0.98mm; but with the incorrect 'fraction2', the amount allocated to this timestep was 9.82mm.

  • Parallel (ETAY) 32-km EDAS restarted from the corrected 32-km control starting at 00Z 4 October 2003.

    • Page Last Modified: October 30, 2003