Diff file for release-2-12
You will find descriptions of the changes to the MM5 version 2
source code in this file. For the difference file between the
current release and the last release, please see
ReleaseNotes/diff.ddmmmyy. The 'current release' refers to the newest
source code, while the 'last release' refers to the source code
with the last change. Each release will be a complete tar file
such as this. The bug fixes/changes in the diff.ddmmmyy
files will be the differences between successive releases.
The current release is 2.12 (April 3 1999).
This is the last release in MM5 Version 2. Last update provided
on Nov 12 1999.
==========================================================
Changes implemented since April 3 1999:
1. Add loader option for SGI O2000 section for BT scheme.
Affected file:
configure.user
2. Correct physical property in 24 category landuse data for
winter category 23 ('Bare Ground Tundra').
Affected file:
Run/landuse.tbl
3. Correct modification to physical properties for permanent ice
category if using USGS or SiB landuse data. Ice category was
hardwired in the code before. Affect results only if you use
USGS or SiB landuse data, and you have permannet ice in your
domain and use IFSNOW switch in the namelist.
Affected files:
domain/initial/init.F, param.F
domain/nest/initnest.F
include/landuse.incl
4. Correction to curvature and coriolis terms for w. Small effect.
Affected file:
dynamics/nonhydro/solve3.F
5. Prevent KF scheme from potential floating point violation.
Affected file:
physics/cumulus/kf/kfpara.F
6. Two corrections for shallow convection scheme (ISHALLO=1).
- Use of shallow convection scheme would cause ice process in
several explicit moisture schemes to be turned off.
Affected file:
physics/cumulus/shallow/shallow.F
- Prevent shallow clouds from getting too deep, which prevents
blow-ups in some cases.
Affected file:
physics/cumulus/shallow/entrs.F
7. Prevent Goddard microphysics scheme from getting potential
floating point error.
Affected file:
physics/explicit/gsfc/satice.F
8. Correction to Reisner I/II schemes for evaporation of freezing rain.
Affected file:
physics/explicit/reisner1/exmoisr.F, lexmoisr.F
physics/explicit/reisner2/exmoisg.F
9. Correction to bulk PBL scheme. Too few variables in argument
list to call SLAB.
Affected file:
physics/pbl_sfc/bulk/blkpbl.F
10.Prevent Eta Mellor-Yamada scheme from getting floating point
error.
Affected file:
physics/pbl_sfc/myepbl/mixlen.F
- Correct an error when using this PBL option with hydrostatic model.
Affected file:
physics/pbl_sfc/myepbl/myepbl.F
11.CCM2 radiation scheme made available in MPP MM5.
12.A correction is made to MPP tar file for IMPHYS=8 option.
==========================================================
Changes for release 2.12 (tag release-2-12)
Please see ReleaseNotes/diff.03APR99 file for actual differences.
1. Bug fixes and modifications:
There are a number of bug fixes required for release-2-11 (see items
1-6 below). These bugs were introduced when modifications were made in that
release. We are sorry if this would have caused any inconvenience for users.
1. Two constants used in the Grell scheme are fixed: SIGCBHI and KBMAX.
This won't change results unless you used them through namelist.
Affected file:
domain/initial/param.F
2. A constant used in Reisner I and II schemes is changed back from
release-2-11 value to an earlier value. This fix was broadcast in an
mesouser email dated Jan 26 1999.
Affected file:
domain/initial/paramr.F (RON=8.E6)
3. A bug fix is made for feedback in nested hydrostatic runs. This fix
was broadcast in an mesouser email dated Feb 24 1999.
Affected file:
domain/initial/feedbk.F
4. A bug is fixed for IFEED=1,2 options. This fix was broadcast in an
mesouser email dated Feb 24 1999.
Affected file:
domain/nest/feedbk.F
5. A bug is fixed for the moving nest option.
Affected file:
domain/nest/initnest.F
6. An include file dependency is added for subroutine domain/util/fill.F
Affected file:
domain/util/Makefile
- A few private variables are added in declaration lists of parallel directives.
Affected file:
domain/nest/initnest.F (OpenMP directives)
dynamics/nonhydro/sound.F (old SGI and HP directives)
- Explicit moisture schemes, Reisner I, Reisner II and Goddard, are modified
to deal with hydrostatic module. These options would not have worked
properly with hydrostatic model.
Affected file:
physics/explicit/reisner1/exmoisr.F
physics/explicit/reisner2/exmoisg.F
physics/explicit/reisner1/godmic.F
- A fix is made to avoid possible floating point exception problem when rain
water and snow fields may have negative values at the nest boundaries.
Affected file:
physics/radiation/cloud/lwrad.F
2. New features:
- A way of allowing the model to handle different number of landuse categories,
and allowing a user to change land surface properties is implemented.
A new table file, landuse.tbl, is created in the directory Run/ in which
the land surface properties are defined (which were defined through DATA
statement in the code prior to this release). Because the values in the table
are read during program execution, they may be modified without recompilation
of the code. The current table supports 13-category, 16-category SiB, and
24-category USGS landuse datasets. The last two types of landuse will be
available in next release of new terrain program (due the week of April 5,
1999), which will have 30", 5', 10', 30' and 1 degree resolution data derived
from 30" datasets. The USGS datasets are global, and the SiB datasets are
N. America.
Affected file:
All decks: need to retype 'make mm5.deck' to create appropriate deck for your
machine. This will link the landuse.tbl file to unit 19.
Run/landuse.tbl (new file)
include/landuse.incl
domain/initial/param.F
- A new explicit moisture scheme is implemented. It is referred to as the Schultz
microphysics scheme (Schultz 1995 MWR). This scheme predicts cloud/rain water,
cloud ice/ snow, and graupel. It is fairly efficient on scalar machines.
Thanks to Greg Thompson (NCAR).
Affected files:
configure.user
Util/parseconfig.c
domain/initial/param.F
domain/io/outtap.F
dynamics/hydro/solve1.F
dynamics/nonhydro/solve3.F
physics/explicit/Makefile
physics/explicit/schultz/ (new directory)
- MM5 code can now deal with 8-digit MDATE in year 2000 and later, including
runs across the millennium.
The fix however doesn't yet apply to 4DDA code.
Affected files:
domain/initial/param.F
domain/io/outtap.F
domain/util/deltatim.F
- A number of constants in Blackadar/MRF/bulk PBL scheme, Kain-Fritsch,
Betts-Miller, Anthes-Kuo schemes are changed to variables and added
to namelist. These variables can be changed through namelist but are hidden.
Unless one has a good reason to do so, we recommend the default values.
Default values will not change results.
(See http://www2.mmm.ucar.edu/mm5/whatisnewinv2-now.html for detail.)
Affected files:
include/parakcu.incl (new file)
include/parkfcu.incl (new file)
include/parbmcu.incl (new file)
domain/initial/param.F
include/param3.incl
physics/cumulus/kf/Makefile, kfpara.F
physics/cumulus/bm/Makefile, bmpara.F
physics/cumulus/kuo/Makefile, cupara2.F
physics/pbl_sfc/hirpbl/hirpbl.F
physics/pbl_sfc/mrfpbl/mrfpbl.F
physics/pbl_sfc/bulk/blkpbl.F
3. Changes (mainly code clean ups):
- The description of changes for last release is updated to include
a change in feedback.
Affected file:
CHANGES
- ReleaseNotes is updated to include latest changes to MM5.
Affected files:
ReleaseNotes/Release-note.ps
- Compiler options for SGI/DEC_Alpha/IBM are updated.
SGI: add mpio option for dealing with parallel io; Also recommended if
your compiler is 7.2.1 and newer, use the OpenMP parallel directives
instead of old doacross directives.
DEC: add -automatic to reduce memory usage
IBM: options for newer xlf compiler. No need to do 'make little_f'.
Affected file:
configure.user
- Namelist explanation for IFEED is made more clear, and IFEED option
is now printed in mm5.print.out file.
Affected file:
Templates/pparam
domain/initial/param.F
domain/nest/initnest.F
- CFL prints from sound.F is changed for clarity.
Affected files:
dynamic/nonhydro/sound.F
- Header arrays JUNKI/R, and JUNKIC/JUNKRC are put into common blocks.
This will reduce memory used.
Affected file:
include/header.incl
- The effects of ice clouds and rain water on liquid water path is
reduced. This has the effect of increasing solar radiation received
at the ground. Also the effect of graupel on liquid water path is
included.
Affected files:
physics/radiation/cloud/swrad.F
- Reisner II scheme (graupel version) is re-released in this release.
New changes include bug fixes to previous release, deposition onto
ice, snow and graupel, reduce minimum fall speed for all particles
(except cloud water), and constant ice number concentration (which
will be fixed in next release). This version produces very little
overhanging ice cloud. The Heymsfield-Donner ice sedimentation option
is removed.
Also functions are moved to a new subroutine so that MPP can work.
Affected files:
physics/explicit/reisner2/exmoisg.F, rslf.F (new file)
4. MPP related changes/bug fixes (not yet updated)
If you are running MPP MM5, you should obtain a new MPP tar file that
corresponds to this release.
- README.MPP file is updated to include information about changes made
in this release.
Affected file:
README.MPP
- Major changes for MPP MM5:
1. The new IMPHYS=8 (Schultz) microphysics option is supported in
the parallel version
2. IFEED=3 has been added as a nest feedback option in the MPP version
3. STOP statements changed to calls to MPI-Abort
4. DM/SMP operation on DEC clusters and IBM SP with SMP nodes
5. The MPP nest-forcing interpolation routines have been reworked to
give better vector performance on the Fujitsu VPP
6. Several bug fixes reported in R11
7. The MPP version reads fort.19 for the landuse table, same
as the non-MPP version of Release 12
For details on these changes, see README.MPP file.
Note: CCM2 Radiation is not yet supported in this version; however work
is in progress