MM5 Model Version 2

WHAT'S NEW IN MM5 VERSION 2 SINCE ITS RELEASE?

NEW SWITCHES

release-2-2 (12/2/96)

IBLTYP = 5

MRF PBL scheme. This is described in Hong and Pan (1996, MWR), but in MM5 it uses the same surface flux parameterization as HIRPBL. The PBL scheme is similar to that in NCEP's MRF model. There are two regimes, stable and unstable. The vertical diffusion is done implicitly using K theory, even in the unstable PBL where K depends upon surface fluxes as well as vertical gradients. K has a parabolic vertical profile in the PBL. The PBL depth is determined from a bulk Richardson number criterion. K above the PBL depends on local Richardson number. The scheme does not require sub-time-steps like HIRPBL does, and is therefore very efficient in comparison, especially for coarse meshes. Results appear very similar to HIRPBL. It should be used with ISOIL=1 because its long time steps can cause the old force-restore method to be unstable, while the ISOIL=1 option allows for splitting the ground temperature calculation into shorter steps when needed.

release-2-3 (2/24/97)

SVLAST

    SVLAST = .FALSE., ; T: only save the last file for restart
                      ; F: save multiple files

Default = .FALSE. When set to TRUE this option rewinds and overwrites the save-file unit(s) each time it is written. This saves disk space but eliminates the choice of restart times.

release-2-4 (4/21/97)

IVQADV, IVTADV

    IVQADV = 0, ;vertical moisture advection uses log 
                 interpolation - 0, linear - 1
    IVTADV = 0, ;vertical temperature advection uses theta 
                 interpolation - 0, linear - 1

Default = 0 : old method. When set to 1, these switches use linear vertical interpolations to calculate vertical advection instead of the old nonlinear interpolation. This option saves 5-8 per cent CPU time with very little change in results, especially if more than 20 or so levels are being used. Note that IVQADV also affects microphysical variables which are always advected the same way as water vapor.

release-2-5 (6/24/97)

IDYNIN, DTRAMP

    IDYNIN = 0, ;for dynamic initialization using a ramp-down function
                ;to gradually turn off the FDDA before the pure forecast, 
                ;set idynin=1 [y=1, n=0]
    DTRAMP =60, ;the time period in minutes over which the nudging 
                ;(obs nudging and analysis nudging) is ramped down 
                ;from one to zero. Set dtramp negative if FDDA is to be 
                ;ramped down BEFORE the end-of-data time (DATEND), and 
                ;positive if the FDDA ramp-down period extends beyond the 
                ;end-of-data time.

release-2-6 (10/10/97)

ITHADV

    ITHADV = 0, ;advection of temperature uses potential 
                 temperature - 1, standard - 0

Default = 0 : old method. When set to 1, it allows for using potential temperature instead of temperature in calculating temperature advection and adiabatic term Test results indicate that this formulation gives more accurate results in high-resolution simulation near complex topography.

release-2-11 (12/22/98)

IBLTYP=4

This is the Mellor-Yamada scheme as used in the Eta model, Janjic (1990, MWR) and Janjic (1994, MWR). It predicts TKE and has local vertical mixing. The scheme calls the SLAB routine for surface temperature and has to use ISOIL=1. Its cost is between the MRFPBL and HIRPBL schemes. Before SLAB the scheme calculates exchange coefficients using similarity theory, and after SLAB it calculates vertical fluxes with an implicit diffusion scheme.

INCTAP

  INCTAP = 1,1,1,1,1,1,1,1,1,1, ; multipliers of TAPFRQ for outputting

This switch allows output files at a lower frequency than TAPFRQ on selected domains. INCTAP(MAXSES) is an array with each element corresponding to domain id. INCTAP=1 (default) will give output at the frequency specified by TAPFRQ. INCTAP=2 will cause output at only every other TAPFRQ time, etc. Thus the model output (fort.41 etc) is every INCTAP(n)*TAPFRQ minutes.

IOVERW

    IOVERW = 1,2,1,1,1,1,1,1,1,1,

This option allows input of a TERRAIN output file at any time during a run to initialize a nest. It differs from IOVERW=1 which requires a complete analysis (MMINPUT) file. The coarser mesh does not have to have information on the fine-mesh terrain prior to the fine mesh starting. With IOVERW=2 new terrain and land-use are initialized and a vertical adjustment of only temperature and pressure are made to the interpolated coarse-mesh fields. The adjustment is automatically fed back to all levels of parent domain and affects p* fields too. To input fine mesh terrain, place the terrain output file in place of mminput file for that domain. For example, mminput_nh_domain2 may be replaced by terrain_domain2 file in the mm5 job deck.

Other parameters that may be set in PPARAM

SOLSET = -999.,

This allows setting solar constant. A value < 0 means to use standard formula.

CZO = .0032, ; Charnock constant
OZO = 1.E-4, ; background value for roughness value over water

IFDTSET=.FALSE., ; =.T. ALLOWS SETTING DTIME USED IN CALCULATING
                  LARGE-SCALE FORCING 
DTIMSET=900., 
PBCMAX=50.,      ; PBCMAX IS MAX ALLOWED DEPTH (MB) OF STABLE LAYER
                   BETWEEN LCL AND LFC
IFCDZ=.FALSE.,   ; WHETHER TO LIMIT CLOUD DEPTH IN TERMS OF HEIGHT
CDMINZ=1000.,    ; MINIMUM CLOUD DEPTH IN METERS USED IF IFCDZ=.TRUE.
CDMINP=150.,     ; MINIMUM CLOUD DEPTH IN PRESSURE (MB) USED IF 
                   IFCDZ=.FALSE.
EDTMIN=0.2,      ; LIMITS OF PRECIP EFFICIENCY 
EDTMAX=0.8, 
EDTSMN=0.3,      ; LIMITS FOR SHEAR DEPENDENCE (SETTING SAME WILL 
                   REMOVE EFFECT)
EDTSMX=0.9, 
TCRIT=273.15,    ; FREEZING IN UPDRAFT BELOW THIS TEMPERATURE 
SIGCBHI=0.7,     ; MAXIMUM CLOUD BASE HEIGHT ALLOWED IN TERMS OF SIGMA 
PKDCUT=75.,      ; DEPTH OF DOWNDRAFT DETRAINMENT (MB) 
CUQMAX=500.,     ; LIMITS ON CONVECTIVE HEATING/COOLING 
CUQMIN=-250.,

release-2-12 (04/03/99)

landuse.tbl

A new input file, called landuse.tbl, is required for release-2-12 and later versions. This file is available in the release-2-12 tar file in the directory Run/. You must either make a new deck, or add the following in your old deck in order to use this release:

ln -s landuse.tbl ${ForUnit}19

This table file contains land-surface properties (such as albedo, roughness length) that were defined through DATA statements prior to this release for the old 13-category landuse dataset. It also contains land-surface properties for the new 24-category USGS and 16-category SiB landuse datasets.

IMPHYS=8

The Schultz (1995, MWR) microphysics option is a scheme involving prediction of cloud water, rain water, pristine ice crystals, snow and graupel (or hail). Its primary distinction is that it has few parameters, and was written to minimize floating point operations. The main routine is called one grid-point at a time, so it has no loops or arrays, and therefore this code is not vectorized.

Other parameters that may be set in PPARAM

TREL = 3000., ; Relaxation time used in Betts-Miller scheme.
RAD = 1500.,  ; Cloud radius assumed in Kain-Fritsch scheme.
BFAC = 2.,    ; Constant used in computing b parameter in Anthes-Kuo
                scheme.
VCONVC = 2.,  ; Constant used in computing convective velocity
                in PBL options 1,2 and 5.