Mesoscale Parallel Experiment Change Log

Experiment Name:  North American WRF-NMM with various physics and/or analysis changes

When Who What
2008/07/17/18 Rogers, Wu
  • Due to a mirroring glitch, the observation error file used by the new GSI analysis was not mirrored to mist until this cycle. For all parallel runs on mist (12z 7/9/2008 - 00z 7/16/2008) no conventional data except for surface pressure data was used.
  • 2008/07/10/18 Wong, Mitchell
  • Reinstated LSM change which related uptake water from roots with root zone soil temperature, but modified to only be applied at grid points with deciduous broadleaf forest
  • 2008/07/08/18 Wong, Mitchell, Rogers
  • Removed LSM change (added on 7/3) which related uptake water from roots with root zone soil temperature.
  • Restarted cycled land-states with those from NDASY parallel
  • 2008/07/03/12 Janjic
  • New shallow convection modified to use "swapsoft2" clouds. For soft2 clouds:
    1. RHref=RHmean[2w(p-ptop)/(pbot-ptop)+(1-w)], where
      1. w=.05
      2. RHmean=mean RH w.r.t. pressure between cloud base and cloud top
    2. "swapsoft" clouds : the new cloud is only used at swap points, at straight points the old cloud is used
  • 2008/07/01/12 Ek, Ferrier, Gayno, Janjic, Lin, Pyle, Rogers, Wong, Wu
  • First NAMR run of partial cycling parallel, start NDASR with land-states from NDASY parallel
  • Changes to WRF-NMM forecast model:
    1. The model for shallow convection is modified to be a plume, or an ensemble of plumes extending from cloud base to cloud top that are determined in the same way as in the current scheme. The air in the cloudy grid boxes is mixed in such a way that RH does not change with height. The temperature gradient within the cloud is a weighted mean of the ambient temperature gradient and the moist adiabatic temperature gradient, where the weights depend on RH. Note that with a cloud like this we cannot have almost saturated air with almost dry adiabatic temperature lapse rate that we occasionally see with the current scheme. As before, the enthalpy is conserved.
    2. The WRF MYJPBL, PBL driver, and the NMM's TURBL subroutine were modified to mix each hydrometeor species separately in the vertical. The current production code only mixes cloud liquid water.
    3. To apply vertical diffusion for separate water species, the subroutine VDIFH has been changed so that (a) it can handle an arbitrary number of species, (b) the counter gradient option can be applied to some or all of the species if desired, and (c) option to set to zero some or all of the surface fluxes is also made available.
    4. Changes to land-sfc physics:
      1. Let ETP (potential evaporation) decrease linearly with Bulk Richardson # under stable condition, and weighted by snow coverage.
      2. Let DQSDT2 (slope of saturated humidity function wrt temperature) decrease linearly with snow coverage.
      3. Relate uptake water from roots with root zone soil temperature.
  • Changes to GSI analysis
    1. Use latest (4Q 2008) version of the GSI analysis code
    2. Assimilate METOP radiance data
    3. Assimilate TAMDAR/AMDAR aircraft data
  • Use AFWA 1/16 bedient snow depth analysis and a new version of the NAM_SNO2MDL code to process this data.
  • Use WPS (instead of WRF-SI) codes to process GDAS first guess GRIB files into a WRF input file used as a first guess to the first (tm12) GSI analysis in the NDAS
  • In the calculation of the long-term water budget used to bias-adjust the precipitation analyses used in the NDAS (as input to WRF-NMM land-sfc model), begin use of the CPC daily gauge analysis in /dcom. The new script converts it to GRIB on the fly, then checks for size of the GRIB'd file to prevent a faulty analysis from being used in long-term water budget calculation. Using the original CPC daily analysis, rather than the pre-GRIB'd version (the conversion to GRIB is done twice a day on a fixed schedule), will allow late-arriving or re-made CPC analyses to be used in the budget calculation. The additional screening performed in the new script will prevent bad analysis files (nearly all of them are less than 95,000 bytes when GRIB'd) from entering the long term budget calculations. These two changes will result in a more accurate long-term budget array and a more precise adjustment to hourly precipitation input used in the NDAS data assimilation.