WRF Model Update
WRF model tar file has been updated to Version
3.2.1 on August 18, 2010. For known problems in V3.2.1, please click here.
WRF Pre-Processing System (WPS) has been updated to Version 3.2.1.
Note: Same as for V3.1, the V3.2 WRF can use V3.0 wrfinput and wrfbdy files as long as you don't use the gravity wave drag option which requires new data from geogrid, and add a new namelist variable, use_baseparam_fr_nml in &dynamics. V3.2 can also use V3.1 input, again as long as one does not try to use the new capablities in V3.2. In any case, rerunning the new programs is recommended. Be cautioous when using new options.
If you are interested in seeing
how V3 has been tested, click here.
Improvements and Bug fixes: (with the significant ones highlighted in bold font)
Physics:
- adapt several microphysics schemes to work at very low pressure (< 5 haPa): wdm5, wdm6, morrison, wsm3, wsm5, wsm6,
- myjsfc, myjpbl: changes the computation of virtual potential temperature so that it does NOT include cloud condensate.
- Thompson: Fixes a problem with tile artifacts appearing in RAINNC field and probably other model fields when Thompson microphysics used with numtiles set to a high value
- YSU: minor bug fixes for PBL Prandtl number calculation in stable and unstable conditions
- MYNN level 3 scheme: fix restart problem
- new namelist tice2tsk_if2cold to control unrealistic surface temp under special circumstance
- new namelist to control soil moisture nudging with PX LSM option: pxlsm_soil_nudge. This option used to share the switch with surface FDDA option
- URBPARM.TBL change: road width, roof level, etc.. Some values in this table were wrong in the 3.2 version.
- Initialize variable used in BEP and BEM urban options
- RRTMG LW reverse a change made for 3.2 that can cause warming spike
dynamics:
- tracer option can be used without compiling with chemistry
- fix for tracer across nest domain boundaries
- curvature term correction for rotated lat-lon projection, small effect unless a mass point falls on the pole
- pressure iniitialization at a domain start time is now consistent with the equation used during model integration
Initialization/IO:
- fix pressure calculation in real so that it is consistent with that in model. The problem was introduced in V3.2 due to the change of using vertically interpolated RH, rather than mixing ratio
- capability to use p_interp generated metgrid output data (use with caution, since the data at top of the model will be extrapolated)
- fix new option surface_input_source = 3 introduced in 3.2, so that dominant categories will not be recalcuated
- fix a problem with initializing complicated nested domains
- fix error with fields being incorrectly zeroed out in the real program when writing FDDA files (posted since 4/20/2010)
- fix error for writing FDDA files when compiling WRF-Chem
- fix uninitialized ProcOrient field in linked lists of state variables, which could lead to unexpected stop during a run (posted since 5/5/2010)
- fix for runtime IO for 4D arrays
- binary: add get/put functions to ensure that the sr_x and sr_y attributes are set properly for binary I/O
- no leap year change for all programs (thanks to T. Otte of EPA), and now the compile option -DNO_LEAP_CALENDAR works
- pnetCDF: Syntax error in a READ format clause in pNetCDF interface
- Performance fix for reading boundary conditions from a long series
ndown/restart
- ndown: use base state parameters written in coarse grid output. Make sure to use the same base state parameters to generate nest input
- fix problems with either wrflowinp or wrffdda not being updated right after restart if the restart time is a data time - the error was introduced in 3.2 when the problems with not being able to restart at non-data times was fixed. If the restart time is a not a data time, then it is ok in 3.2.
tc bogussing:
- Multi tropical storm bogusing
-
Finite difference using WRF horizontal staggering
-
New namelist variable to restrict the height of the rankine vortex (rankine_lid)
namelists:
- Add io_form_auxinput3 = 2 in scm namelist
- namelist.input.chem change
- More namelists cross checks
- README.namelist update
Software, configure and compile:
- default no of processors used in parallel make is now 2 (instead of 6)
- add start and end time for compile to show how it takes
- clean up DFI_RADAR compile logic
- parallel compile for chemistry when KPP is excluded
- Make the timing routine available even when the code is not compiled for DM_PARALLEL
- some fixes for GPU compile
