WRF-NMM Users Page

WRF Modeling System Updates

The WRF model tar file has been updated to Version 3.4 on April 6, 2012.

The WRF Pre-Processing System (WPS) has also been upgraded to Version 3.4.

WRF Version 3.4 new features, improvements and bug fixes

WRF-NMM specific updates

  • Changes to make the WRF-NMM released code more similar to the WRF-NMM based operational model
    • Moved existing Ferrier microphysics "etampnew" approach to be "etampold" (95) and brought in Ferrier micorphysics version used with the high-resolution window system to be "etampnew" (5). The new option (5) is tuned for finer resolution runs while the old option (95) should continue to be used for coarse resolution runs.
  • NSAS cumulus scheme is now an option for NMM
  • New QNSE and Noah-MP schemes are available for NMM
  • Restores the portion of w removed in the nonhydrostatic dynamics into a special array for output (if wp>0.001; default 0)
  • Decrease compressibility through the top 1.5% of the atmosphere to counter effects of the vertical grid increments in z becoming much larger than they are in the horiztonal (if wp>0.001; default 0)
  • Miscellaneous bug fixes:
    • Eliminated a line within module_initialize_real that was defining grid%dt to be just the integer part of the time step leading to improper definitions of some time-step dependent dynamical quantities when fractional time steps are used and causing problems (noise) with high resolution runs using a 1.x second time step.
    • Corrected uninitialized variables
    • Passed two additional arrays for RUC-LSM
    • Added checks for optional arguments in several physics schemes

HWRF specific updates

  • New Capabilities:
    • Preliminary development towards high-resolution (27-9-3km) triple nest capability. Untested additional vortex tracking algorithms.
    • High frequency (each time step) ascii output storm location, maximum surface winds, and radius of maximum winds (HWRF only)
    • Interpolate SST to a nested grid (HWRF only)
    • New fast random number generator that stores the generator state in each gridpoint, allowing each gridpoint to have its own independent random sequence. The sequence does not vary with domain decomposition, compiler or platform. This is used by the old SAS scheme (cu_opt=4).
    • New namelist variable to control the amount of momentum mixing in the HWRF SAS scheme (cu_opt=84)
    • Various updates of GFDL surface scheme, Tropical Ferrier microphysics, and HWRF SAS
    • Possibility of running NSAS cumulus and RRTMG (shortwave and longwave) radiations with HWRF
  • Miscellaneous bug fixes:
    • NMM moving nests (HWRF) - in some situations, winds on the leading edge of the moving nest would be zeroed out
    • HWRF restarts
    • Ocean coupler related to setting coupler timestep
    • Tropical Ferrier microphysics scheme to correct the use of uninitialized variables

General updates

New in Version 3.4


  • Microphysics:
    • NSSL 2-moment (Mansell, Ziegler and Bruning, JAS, 2010): The scheme is a 2-moment scheme developed at the National Severe Storms Laboratory. It predicts the mass mixing ratio and number concentration for six hydrometeor species: cloud droplets, rain drops, ice crystals, snow, graupel, and hail. A unique feature is the additional prediction of average graupel particle density, which allows graupel to span the range from frozen drops to low-density graupel. Hail is produced only by wet growth of graupel to try to represent true hail rather than merely high-density ice. An option allows prediction of cloud condensation nuclei (CCN) concentration (intended for idealized simulations). The scheme also features adaptive sedimentation to allow some size sorting but prevent spurious large particles (and radar reflectivity values) that can arise from two-moment microphysics, particularly for the larger precipitation categories (graupel, hail, and rain). Size distribution shape parameters and a number of other options can be set in the code. The scheme is intended for cloud-resolving simulations (dx <= 2km) in research applications. Thanks to T. Mansell of NSSL.
  • Land Surface Models:
    • Noah-MP (Niu et al., 2011; MP: multi-physics): It uses multiple options for key land-atmosphere interaction processes. Noah-MP contains a separate vegetation canopy defined by a canopy top and bottom with leaf physical and radiometric properties used in a two-stream canopy radiation transfer scheme that includes shading effects. Noah-MP contains a multi-layer snow pack with liquid water storage and melt/refreeze capability and a snow-interception model describing loading/unloading, melt/refreeze, and sublimation of the canopy-intercepted snow. Multiple options are available for surface water infiltration and runoff, and groundwater transfer and storage including water table depth to an unconfined aquifer. Horizontal and vertical vegetation density can be prescribed or predicted using prognostic photosynthesis and dynamic vegetation models that allocate carbon to vegetation (leaf, stem, wood and root) and soil carbon pools (fast and slow). Thanks to Zong-Liang Yang (UT-Austin) and Guo-Yue Niu (Biosphere 2, Univ. of Arizona) for the development, and Joakim Refslund Nielsen (Risø DTU), Kevin Manning (NCAR-MMM) and Michael Barlage (NCAR-RAL) for implementing it in WRF. For more information, please click here.
    • SSiB: This is the third generation of the Simplified Simple Biosphere Model (Xue et al. 1991; Sun and Xue, 2001). SSiB is developed for land/atmosphere interaction studies in the climate model. The aerodynamic resistance values in SSiB are determined in terms of vegetation properties, ground conditions and bulk Richardson number according to the modified Monin-Obukhov similarity theory. SSiB-3 includes three snow layers to realistically simulate snow processes, including destructive metamorphism, densification process due to snow load, and snow melting, which substantially enhances the model's ability for the cold season study. To use this option, ra_lw_physics and ra_sw_physics should be set to either 1, 3, or 4. The second full model level should be set to no larger than 0.982 so that the height of that level is higher than vegetation height. Thanks to Xue and de Sales of UCLA.
  • Atmospheric Radiation:
    • Fu-Liou-Gu: multiple bands, cloud and cloud fraction effects, ozone profile from climatology, shortwave can allow for aerosols. For more details and references, please click here. Thanks to Gu of UCLA.
  • Cumulus:
    • New SAS for HWRF (option 84). Not yet tested for ARW.
  • Planetary Boundary Layer:
    • QNSE-EDMF: Daytime part of QNSE PBL modified to use mass flux method with shallow convection activated by namelist option mfshconv=1 (default). Thanks to J. Pergard of Numtech, France.
  • Surface Physics:
    • A revised version of sfclay=1 to remove limits and use updated stability functions. (Jimenez et al., MWR 2012). Thanks to P. Jimenez of CIEMAT, Spain.


  • WENO (Weighted Essentially Non-Oscillatory) advection scheme: The scheme may help with inconsistencies in the microphysical moments at cloud edges. Both 5th order and 5th order with positive-definite limiter are available. Thanks to T. Mansell of NSSL.

Improvements and bug fixes


  • Microphysics:
    • Thompson: bug fix and updates
    • Morrison: minor change to the limits on autoconversion source of rain number when cloud water is depleted. Major bug fix for V3.3.1 version for ice fallspeed (posted 11/16/11).
    • WDM5/6: fix for nccn variable on lateral boundary
    • Eta scheme: updated to High-Res Window version operational at NCEP
  • Atmospheric Radiation:
    • Goddard shortwave scheme (ra_sw_physics = 1): diagnositic field RSWTOA corrected
  • Cumulus:
    • Tiedtke: A new trigger is added (controlled by cutrigger in the code; the old one is still the default). It is based on a diluted air parcel testing, while the old trigger is based on the moisture convergence. New organized entrainment and turbulent entrainment/detrainment rate for deep convection based on ECMWF method is added. Thanks to C. Zhang of Univ of Hawaii.
    • G3 + shallow: minor fix for shallow part
  • Planetary Boundary Layer:
    • YSU: improved Pr number computation and change to lower limits on diffusion
    • MYJ and MYJSFC: updated to operational version
    • MYNN and MYNNSFC: updated (see code for details). Thanks for J. Olson of NOAA.
    • ACM2: bug fix to treat convective adjustment at first time step for occasional problems for small grid sizes
    • TEMF: bug fix for occasional blowups
  • Surface Physics:
    • RUC LSM: use MODIS; capability to specify soil and land-use parameters based on mosaic approach. Two new namelist parameters are introduced: mosaic_lu and mosaic_soil, the default values are 0, but to switch to mosaic specifications these namelist parameters should be set to 1 (not necessary for both of them, could be default for soil and mosaic for land-use, or visa versa). Tuning some parameters in snow model to get more realistic snow melting rates.
    • PX LSM: MODIS support, lake ice treatment, snow albedo tretaments, bug fixes for soil moisture nudging and others.
    • tmn_update: minor fix
    • isftcflx = 2 option updated to increase drag at lower wind speed
    Nudging: Single-Column Model (SCM):
    • Spectral-nudging: fix for dtramp coefficient
    Stochastic Kinetic-Energy Backscatter Scheme:
    • Updated and bug fixes for restart
    • Default values for these namelists have been changed
      • hypsometric_opt = 2 (affecting both real and WRF model)
      • interp_theta = .false. (real only)
      • For details on hypsometric_opt and interp_theta, see paper by Wee et al. (2012)
      • lagrange_order = 2 (real only)
    • Masked SST handled better in real
    • Surface pressure recalculated in ndown
    • Fix for idealized tropical cyclone initialization
    • Fix for polar projection for grib IO
    • Fixes for binary IO
    • The restriction to compile without nesting is removed
    • A few physics sensitivity related namelist variables are removed: icloud, isfflx, ifsnow.
    • New Features:
      • A capability to interpolate a physics variable read-in or computed on the coarse grid to nest
      • Added support for SGI MPT, and NEC SX
    • Improvements:
      • pnetCDF on small-memory machine