CESM1.0.4 Release Notes
cesm1_0_4
Contents
Release Notes
The main purpose of the CESM1.0.4 release is to provide the community a version of CESM1 that supports the Thermosphere/Ionosphere extension of WACCM (WACCMX) and the the capability to configure cases using interannual forcing component sets (CIAF, DIAF, and GIAF). CESM1.0.4 also includes a "patch" to MPI_RSEND that improves cray-gemini interconnect performance on Cray XT6 machines, and it contains other enhancements, machine updates, and bug fixes, as noted in this section. Aside from these new features, the CESM1.0.4 release is an incremental release beyond CESM1.0.3.
What's New
CAM (Active Atmosphere Model)
- New features in CAM5.1.1
- Added WACCM-X. WACCM-X is the extension of WACCM/CAM from the lower thermosphere to the top of the thermosphere. See the WACCM-X section of the Notable Improvements document for more information.
- COSP updates. Included changes are:
- made several high-level namelist variables that control COSP.
The new options are all meant to reduce cost by running with fewer
subcolumns (10) and by only running cosp every 3rd time the
radiation is run. The new variables are:
cosp_amwg
cosp_lite
cosp_passive
cosp_active
cosp_isccp - change the default mode from cosp_runall to cosp_amwg. Add cosp_runall to the namelist so this mode can still be specified.
- changed default for cosp_histfile_num from '2' to '1'.
- changed sub-column outputs to avoid crashes when you specify outputs as 'A' on a monthly history file. The subcolumn output is only on by default when explicitly requested via setting either cosp_lfrac_out=.true. or cosp_runall=.true.
- made COSP compatible with fincl1lonlat -- so that you can run COSP and save COSP outputs locally.
- Add checks to keep rain_cv and snow_cv above 0.0
- Fixed restart bug.
- made several high-level namelist variables that control COSP.
The new options are all meant to reduce cost by running with fewer
subcolumns (10) and by only running cosp every 3rd time the
radiation is run. The new variables are:
- Code cleanup and refactoring
- A reference pressure module has been added and remaining references in the physics package to the hybrid vertical coordinates have been removed. The physics package should only need information about the pressures at layer midpoints and interfaces. It should not depend in any way on the vertical coordinate used by the dynamical core.
- The prescribed aerosol data modules were extended to be able to read in modal aerosol concentrations and aerosol deposition fluxes. This was done to facilitate development of prescribed Modal Aerosol Model configurations (in progress).
- Code refactored for memory reduction in chemistry/mozart/mo_drydep.F90 and physics/cam/radae.F90
- Removed CAM specific print_memusage messages
- CAM build system checks for the existence of a netcdff library
- mkDepends was modified to relax the restrictions on source file name matching module name, and only one module per source file.
- For WACCM-SC, removed QTHERMAL, QRS_AUR, QRS_CO2NIR, QRS_EUV, from the output files.
- Bugs fixed
- Bug fix to pass the correct variable for the molecular diffusion coefficient to the WACCM gravity wave routine.
- Bug fix to use the correct values of molecular diffusion coefficient in the vertical diffusion solver at the lowest level at which molecular diffusion is turned on (0.55 hPa) in WACCM.
- Bug fix to correct the output values of EKGWSPEC, the effective Kzz diffusion coefficient due to the gravity wave spectrum, which were low by exactly a factor of 2 in WACCM history output.
- Bug fix to make restart history files double precision.
- Fixed a restart problem that occurs when empty_htapes=.true. in some configurations. Seems to be a problem that occurs with multiple history file output.
- Correction to local-time averaged history output.
- Correction to sathist output of dynamics decomposed variables on the HOMME grid.
- Bug fixed for column data using the collect_column_output option when the same number of columns (but different locations) was used on multiple history files.
- Work around for mpi_rsend bug on the Gemini interconnect. The Gemini interconnect has a performance bug for mpi_rsend and mpi_irsend. CPP flag _NO_MPI_RSEND was added so that MPI_RSEND and MPI_IRSEND can be redefined to MPI_SEND. _NO_MPI_RSEND is only set within the CESM scripts.
- Added correction to O3 chemical loss rate diagnostic.
- Bugfix for the diagnostic output fields CLDLIQSTR and CLDICESTR. The previous version was based on state values that did not contain the tendencies from the macrophysics package.
- Allocate statement in mo_drydep.F90 not used and was removed. It could have caused problems with UNSTRUCTURED grids since the nlon_veg and nlat_veg variables are not defined in that case.
- Fixed timers in tphysbc.F90 and macrop_driver.F90
CICE (Active Ice Model)
- Supports Gregorian calendar.
CISM (Active Land-Ice Model)
- No changes.
CLM (Active Land Model) (See CLM User's Guide for details of all the following)
- The only changes were bugs fixes to CESM1.0.3 as follows
- A couple issues with CNDV were resolved.
- A issue with the configure step where it would hang with a bad user_nl_clm file was fixed.
- A few issues with history variables was fixed.
- We fixed it so that you could build with C13 on bluefire.
- Domain files and fraction files were made to be consistent.
- A problem with Gregorian input was fixed.
- Several issues with PTCLM were fixed.
- Monthly average files are no longer required to be one per month.
- CLM1PT mode was changed to "cycle" mode by default so that simulations running beyond the valid data range would NOT be invalidated.
- A couple problems with PIO were fixed.
- The I1850SPINUPCN was changed to point to an appropriate 1850 case rather than a transient case.
POP2 (Active Ocean Model)
- Minor changes and bugfixes:
- New default setting for the output restart file format. By default, POP2 restart files are now written in netCDF format. This change was made in order to preserve b4b restarts on some machines. Note that POP2 cases can restart from an existing binary restart file, regardless of the format selected for new output restart files. The POP2 scripts will automatically detect the format of the old restart file and read it accordingly.
- Correction to the units for the time-averaged quantity TPOWER.
- Bugfix for gfortran format-statement problem.
- MPI_RSEND workaround for cray/gemini performance.
- Add addtional supported decomposition for the tx0.1 resolution.
- Minor bugfix for potential problem in N_HEAT and N_SALT diagnostics computation (there was no bug in the out-of-the-box configuration; only in customized cases)
- Recharacterize some arrays in the ecosys_mod.F90 module as assumed-shape arrays, in order to avoid Intel compiler problem.
- New POP2 FAQ
Driver
- No notable changes
Scripts
- Updated machines list. Added new machines (titan,olympus_pgi, olympus_intel,janus) and deleted obsolete machines (prototype_columbia, midnight, hopp2, jaguar* (use titan instead))
- Change default DATM_CPL_CASE for I_1850_SPINUP_3HrWx_CN compset
- Change CLMQIAN directories for I compsets to under DIN_LOC_ROOT on all machines
- Bugfixes:
- Fix: compsets: E_2000_CN_GLC, E_1850_CN_GLC, F_AMIP_CAM5_CN, F_1850_CN_GLC so CN is turned on as it should
- Fix: On machine Lynx env_build.xml and env_run.xml references to /ptmp changed to /glade/scratch
- Fix: ESMF library path on bluefire no longer points to a temporary area.
Data Models
- Added support for interannual forcing (IAF) compsets.
- Update datm domain files for some grids with matching land/atm [T85_T85, T42_T42, T31_T31, f10_f10, f45_45, f19_f19, f09_f09, f05_f05, and f02_f02 resolutions] (will change answers).
- Bugfixes:
- datm: dust bin 4 wet/dry deposition bug fix -- changes answers in cases with datm dust.
- dice: bug fix for flux_Qacc -- changes answers for cases using C-type compsets.
Dead Models
- Nothing changed.
Stub Models
- Nothing changed.
csm_share Code
- Bugfixes:
- Minor Bugfix for using Gregorian input calendar
- Fix CF data string parsing bug for PGI 11.10
Answer-Changing Features
This page documents changes in cesm1_0_4 that will result in non-bit-for-bit model solutions when compared with cesm1_0_3. Directly below is a summary of compsets that are affected by changes in cesm1_0_4; details can be found in the sections that follow.
- WACCM compsets (see CAM/WACCM)
- E_2000_CN_GLC, E_1850_CN_GLC, F_AMIP_CAM5_CN, F_1850_CN_GLC (see Scripts)
- I compsets, using any of the following domains: T85_T85, T42_T42, T31_T31, f10_f10, f45_45, f19_f19, f09_f09, f05_f05, f02_f02 (see CLM)
- A,C,G,D and I compsets (see "bin 4" discussion in CLM)
- C compsets (see dice)
CAM (Active Atmosphere Model)
- Climate-preserving changes:
- WACCM:
Two bugs were fixed that affect model calculations. First, the incorrect variable had been passed into the gravity wave calculation as the molecular diffusion coefficient. Second, the diffusion solver was using the wrong value as molecular diffusion at the lowest level at which molecular diffusion is turned on (0.55 hPa, ~53 km). The correction of these two bugs produces significant answer changes above 100 km, where molecular diffusion is most significant. In particular, increases in CO2 are significant in the global average above 100 km, with changes increasing with altitude. Additionally, global average temperature decreases significantly between 100 and 120 km, and increases significantly above 120 km.
A third bug was fixed to correct the output values of EKGWSPEC, the effective Kzz diffusion coefficient due to the gravity wave spectrum, which were low by a factor of 2. This bug fix affects only model output for this variable, as the correct value of the diffusion coefficient was being used in model calculations prior to the bug fix.
- Non-climate-preserving changes:
- None
CICE (Active Ice Model)
- Climate-preserving changes:
- None
- Non-climate-preserving changes:
- None
CISM (Active Land-Ice Model)
- Climate-preserving changes:
- None
- Non-climate-preserving changes:
- None
CLM (Active Land Model)
- Climate-preserving changes:
- The domain and fraction datasets were made to be consistent so answers for the following I compset resolutions change: T85_T85, T42_T42, T31_T31, f10_f10, f45_45, f19_f19, f09_f09, f05_f05, and f02_f02.
- Dust wet/dry deposition was mislabeled for bin 4. This changes the coupler history fields and sometimes has a small impact on this DUSTDEP history field in CLM.
- Non-climate-preserving changes:
- Answers for the default I1850SPINUPCN compset changes as the MOAR forcing dataset driving this compset were changed to an appropriate 1850 spinup case.
- The history variables: CISUN, CISHA, ALPHAPSNSUN, and ALPHAPSNSHA were fixed rather than writing out gdir.
- The soil data for PTCLM US-UMB site was corrected.
- A bug in the datm forcing setting for CLM1PT mode was fixed so that it runs in "cycle" mode rather than extrapolating when beyond the available data. This problem caused simulations to be worthless.
POP2 (Active Ocean Model)
- Climate-preserving changes:
- None
- Non-climate-preserving changes:
- None
dice (Data Ice Model)
- Climate-preserving changes:
- bug fix for flux_Qacc -- changes answers for cases using C-type compsets.
- Non-climate-preserving changes:
- None
Scripts
- Climate-preserving changes:
- None
- Non-climate-preserving changes:
- A bugfix was applied to the compset definitions for the following compsets: E_2000_CN_GLC, E_1850_CN_GLC, F_AMIP_CAM5_CN, F_1850_CN_GLC. In each of these compsets, CN should have been activated, but it was not.
Machines
Users can find a list of machines which are recognized by the CESM1.0.4 scripts in the CESM1.0.4 User's Guide.
Known Problems
____________________________________________________________________________________________ Description: The CICE history files have some incorrect history fields (albpnd, albice, albsno). These should not be used. Date : originally reported 2011-05-09 Contact : Dave Bailey Information: The counters for these albedos over ponds, ice, and snow are incorrect. These should not be used. They are diagnostic only. A fix is currently being worked on and will be released at a later date.
____________________________________________________________________________________________ Description: "Claim exceeds reservation's resources" error message in ccsm.log on Crays Date : Originally reported 2011-06-10 Contact : Chris Fischer Information: On a rare occasion you can get this error message when two concurrent tasks, with different threading, are running on the same node. The fix for this is to change the PES layout so the concurrent tasks are on different nodes.
____________________________________________________________________________________________ Description: WACCM compsets and use of ESMF library Date : Originally reported in cesm1_0_3 2011-06-06 Contact : Francis Vitt Information: There is a known problem with following WACCM compsets when the ESMF time manager is used. The problem occurs when the solar_parms_file input file contains data for leap days (Feb 29) while the calendar is 'NOLEAP'. A work around to this problem is to use the following alternative datasets for solar_parms_file which do not contain leap days: B_1850-2000_WACCM_CN solar_parms_file = '$DIN_LOC_ROOT/atm/waccm/solar/proxy_solar_Solomon_Richmond_1845-2008_daily_noleap_c110526.nc' B_1955-2005_WACCM_CN F_1955-2005_WACCM_CN B_RCP2.6_WACCM_CN B_RCP4.5_WACCM_CN B_RCP8.5_WACCM_CN solar_parms_file = '$DIN_LOC_ROOT/atm/waccm/solar/wasolar_1948-2140_daily_noleap_c110525.nc'
____________________________________________________________________________________________ Description: Scripts don't support ESMF serial code Date : Originally reported on 2011-06-09 Contact : erik Information: During compilation, ESMF is trying to link in a mpi library. On bluefire, setting ESMF_LIBDIR to /contrib/esmf_5_2_0-64-O-mpiuni will allow you to run serial code.
____________________________________________________________________________________________ Description: Problem running I1850SPINUPCN with more than 82 years of data Date : Originally reported 2011-06-10 Contact : erik Bugnumber : 1339 Information: In order to run with more than 82 years of CPLHIST data the file limit in shr_stream needs to increase from 1000 to 2000 (technically 1860 would be sufficient, but might as well bump it to 2000). Index: shr_stream_mod.F90 =================================================================== --- shr_stream_mod.F90 (revision 28396) +++ shr_stream_mod.F90 (working copy) @@ -101,7 +101,7 @@ end type shr_stream_fileType !--- hard-coded array dims ~ could allocate these at run time --- - integer(SHR_KIND_IN),parameter :: nFileMax = 1000 ! max number of files + integer(SHR_KIND_IN),parameter :: nFileMax = 2000 ! max number of files type shr_stream_streamType !private ! no public access to internal components We didn't put this change in as it causes a compiler error on bluefire for AIX when compiling the dlnd component. See the CLM KnownBugs file for more information on this... http://www.cesm.ucar.edu/models/cesm1.0/clm/KnownBugs.CESM1.0.4
____________________________________________________________________________________________ Description: exact restart test on janus (cprnc) Date : 2012-02-16 Contact : jedwards Information: There is a very minor problem on janus wrt the cprnc tool: the path is incorrectly specified. You might encounter this problem if you try to set up an exact restart test. The incorrect path for cprnc on janus is: /lustre/janus_scratch/cesm/tools/cprnc The correct path for cprnc on janus is: /lustre/janus_scratch/cesm/tools/cprnc/cprnc To correct the problem, edit your $CASE/env_run.xml file so that the variable CCSM_CPRNC is set to /lustre/janus_scratch/cesm/tools/cprnc/cprnc You do not need to rebuild anything; just submit/resubmit your job.
____________________________________________________________________________________________ Description: Compset F_1850-2000*** is using the incorrect SST data Date : 2012-03-21 Contact : fischer Information: The compsets F_1850-2000*** are using an the incorrect sst dataset. There are 2 ways to fix this. ---------- 1) After you've created your case and configured it, edit Buildconf/docn.buildnml.csh Set the following For T31 set dat_datfile = sst_HadOIBl_bc_48x96_1850_2010_c110526 For 1 degree set dat_datfile = sst_HadOIBl_bc_0.9x1.25_1850_2008_c100127.nc For 2 degree set dat_datfile = sst_HadOIBl_bc_1.9x2.5_1850_2008_c100127.nc ---------- 2) Before you run create_newcase Edit scripts/ccsm_utils/Case.template/config_compsets.xml in your sandbox For F_1850-2000_CN (F20TRCN) add the following after after line 1323. <compset NAME="F_1850-2000_CN" SHORTNAME="F20TRCN" GRID_MATCH="T31_T31" DOCN_SSTDATA_FILENAME="$DIN_LOC_ROOT/atm/cam/sst/sst_HadOIBl_bc_48x96_1850_2010_c110526" DOCN_SSTDATA_YEAR_START="1850" DOCN_SSTDATA_YEAR_END="2008" /> <compset NAME="F_1850-2000_CN" SHORTNAME="F20TRCN" GRID_MATCH="f19_f19" DOCN_SSTDATA_FILENAME="$DIN_LOC_ROOT/atm/cam/sst/sst_HadOIBl_bc_1.9x2.5_1850_2008_c100127.nc" DOCN_SSTDATA_YEAR_START="1850" DOCN_SSTDATA_YEAR_END="2008" /> <compset NAME="F_1850-2000_CN" SHORTNAME="F20TRCN" GRID_MATCH="f09_f09" DOCN_SSTDATA_FILENAME="$DIN_LOC_ROOT/atm/cam/sst/sst_HadOIBl_bc_0.9x1.25_1850_2008_c100127.nc" DOCN_SSTDATA_YEAR_START="1850" DOCN_SSTDATA_YEAR_END="2008" /> Then for F_1850-2000_CN_GLF (FG20TRCN) Change the following at line 1355 <compset NAME="F_1850-2000_CN_GLC" SHORTNAME="FG20TRCN" GRID_MATCH="f09_f09" DESC="20th Century transient stand-alone cam default, prescribed ocn/ice, cism land-ice, with CN" /> <compset NAME="F_1850-2000_CN_GLC" SHORTNAME="FG20TRCN" GRID_MATCH="f19_f19" DESC="20th Century transient stand-alone cam default, prescribed ocn/ice, cism land-ice, with CN" /> <compset NAME="F_1850-2000_CN_GLC" SHORTNAME="FG20TRCN" GRID_MATCH="T31_T31" DESC="20th Century transient stand-alone cam default, prescribed ocn/ice, cism land-ice, with CN" /> -- to -- <compset NAME="F_1850-2000_CN_GLC" SHORTNAME="FG20TRCN" GRID_MATCH="f09_f09" DESC="20th Century transient stand-alone cam default, prescribed ocn/ice, cism land-ice, with CN" DOCN_SSTDATA_FILENAME="$DIN_LOC_ROOT/atm/cam/sst/sst_HadOIBl_bc_0.9x1.25_1850_2008_c100127.nc" DOCN_SSTDATA_YEAR_START="1850" DOCN_SSTDATA_YEAR_END="2008" />> <compset NAME="F_1850-2000_CN_GLC" SHORTNAME="FG20TRCN" GRID_MATCH="f19_f19" DESC="20th Century transient stand-alone cam default, prescribed ocn/ice, cism land-ice, with CN" DOCN_SSTDATA_FILENAME="$DIN_LOC_ROOT/atm/cam/sst/sst_HadOIBl_bc_1.9x2.5_1850_2008_c100127.nc" DOCN_SSTDATA_YEAR_START="1850" DOCN_SSTDATA_YEAR_END="2008" />> <compset NAME="F_1850-2000_CN_GLC" SHORTNAME="FG20TRCN" GRID_MATCH="T31_T31" DESC="20th Century transient stand-alone cam default, prescribed ocn/ice, cism land-ice, with CN" DOCN_SSTDATA_FILENAME="$DIN_LOC_ROOT/atm/cam/sst/sst_HadOIBl_bc_48x96_1850_2010_c110526" DOCN_SSTDATA_YEAR_START="1850" DOCN_SSTDATA_YEAR_END="2008" />>
____________________________________________________________________________________________ Description: List of minor issues for CLM in cesm1_0_4 Date : 2012-08-21 Contact : erik Information: There is a list of minor issues and Known limitations with CLM under the Documentation for CLM4 in cesm1_0_4 at http://www.cesm.ucar.edu/models/cesm1.0/clm/KnownBugs.CESM1.0.4 http://www.cesm.ucar.edu/models/cesm1.0/clm/models/lnd/clm/doc/KnownLimitations
____________________________________________________________________________________________ Description: CFC initialization in POP is incorrect for 20th century branch and hybrid runs Date : 2013-01-28 Contact : mlevy Information: To fix this problem, copy this version of ocn.cfc.setup.csh into $CASEROOT/SourceMods/src.pop2/
____________________________________________________________________________________________ Description: Incorrect diagnostic output for POP2 variable RHO_VINT Date : 2013-02-05 Contact : mlevy Information: There is a bug in the way POP2 computes the diagnostic variable RHO_VINT. Note that this issue does not affect any of the prognostic variables in the CESM 1.0[.x] releases.
____________________________________________________________________________________________ Description: (Bugzilla 1740) Limitations with different coupling frequencies for different models. Date : 2013-06-07 Contact : aliceb Information:
CESM allows different coupling frequencies for different models. However, there are some limitations. The coupling frequencies are set via the env_run.xml variables *_NCPL. Generally, the atmosphere, land, and ice model coupling frequencies should match. The ocean model coupling frequency should be equal to or less than the atm coupling frequency. And the runoff model coupling frequency should be greater or equal to the atm coupling frequency and less than or equal to the ocean coupling frequency. The known problem is that there are no checks at setup or run-time that these constraints are met. If they are not met, there can be problems with the simulation. In summary, users should ensure that the following is true,
ATM_NCPL = LND_NCPL = ICE_NCPL >= ROF_NCPL >= OCN_NCPLespecially for fully active configurations.____________________________________________________________________________________________ Description: (Bugzilla 1740) Limitations with different coupling frequencies for different models. Date : 2013-06-07 Contact : aliceb Information:
CESM allows different coupling frequencies for different models. However, there are some limitations. The coupling frequencies are set via the env_run.xml variables *_NCPL. Generally, the atmosphere, land, and ice model coupling frequencies should match. The ocean model coupling frequency should be equal to or less than the atm coupling frequency. And the runoff model coupling frequency should be greater or equal to the atm coupling frequency and less than or equal to the ocean coupling frequency. The known problem is that there are no checks at setup or run-time that these constraints are met. If they are not met, there can be problems with the simulation. In summary, users should ensure that the following is true,
ATM_NCPL = LND_NCPL = ICE_NCPL >= ROF_NCPL >= OCN_NCPLespecially for fully active configurations.____________________________________________________________________________________________ Description: Error when building the code DV active - bugzilla ID=1745 provides fix. Date : 2013-06-11 Contact : aliceb Information: User got this error when building the code with DV active: /cesm1_0_4/models/lnd/clm/src/biogeochem/CNDVMod.F90(160): error #6580: Name in only-list does not exist. [FRIVINP_RTM] frivinp_rtm, iulog ---------------------------------^ /cesm1_0_4/models/lnd/clm/src/biogeochem/CNDVMod.F90(295): error #6404: This name does not have a type, and must have an explicit type. [FRIVINP_RTM] if (frivinp_rtm /= ' ') then --------^ ...and I replied: Looks to me like a bug. Add the following two lines around all instances of the variable frivinp_rtm, as done in clm_varctl.F90. #ifdef RTM #endif User responded that this fixed the problem, and he was able to run.