User's Guide to NCAR CCM3.6 Search Page

1. Introduction

1.1 What is new with CCM3.6.6

• Physics changes (no significant change to the climate):

 
• Added capability to calculate earth's orbital parameters.  Changed algorithm to calculate solar zenith angle (Berger, 1977).
• Modified surface stress calculation.
• Solar radiation calculations now include 19 bands.  This allows comparison to Nimbus-7 spectral albedos (Collins, 1998).
• 2m-reference height temperature computed everywhere, both in flux-coupled mode and stand-alone mode.
• Snow capped to 1m over land ice.
• Negative runoff in LSM is allowed for glaciers, lakes, wetland, and irrigated land so that the model conserved water globally.

 
• Changes to input/output files:

 
• The ozone dataset is now interpolated horizontally and vertically during the model run.
• All input files are now netCDF format rather than ccm history file format.
• LSM and CCM generate initial conditions files during model runs.
• Input and output LSM datasets are in netCDF format.
• LSM can now generate its own initial dataset at model startup.
• Off-line programs interpic, and definesurf are now provided to interpolate initial conditions to a different resolution.
• An off-line program interpsst is now privided to interpolate Sea Surface Temperature (SST) data to a different resolution.

 
• Changes to user interface:

 
• Added capabilities to run model either adiabatically or with prescribed heating and temperature profiles.
• Solar constant added as a namelist input variable.
• Volume mixing ratios for greenhouse gases added as namelist input.
• Output primary history files and restart datasets renamed to correspond to model date.
• Added a "restart-pointer file" to keep track of the latest restart file written. This way the user doesn't have to figure out the name of the last restart file written to restart a run.
• Code and Makefile mods were added to enable the CCM to be run on a PC running LINUX . This option is not officially supported, but is made available for those that might make use of it. This option assumes you are using the portland group FORTRAN-77 compiler (pgf77).
     http://www.cgd.ucar.edu/cms/ccm3/news/ccm3.6.6.shtml.
• Installed the SOM option back into the public distribution. This option is still not officially supported and datasets are not provided. However, the code is now availbe for those who need to use this configuration. For more information see:
     http://www.cgd.ucar.edu/cms/ccm3/news/ccm3.6.6.shtml.

 
Bug fixes (no significant change to the climate):

• Calendar no longer off by 1 day.
• Fixed vector temporary that was improperly used outside of loop in radiation code (routine radabs).
• Corrected pressure levels used in routine difcor.
• Corrected minor problem with top of atmosphere radiation calculation (routine radtpl).
• Corrected minor problem with water vapor paths used in trace gas overlap (routine trcabn).
• Fixed potential out of bounds memory reference in buoyancy calculation (routine bouyan).
• Workaround added to LSM to address Cray compiler bug. See the CCM3 web-site for more information on this at:
     http://www.cgd.ucar.edu/cms/ccm3/news/ccm3.2_fix.shtml.
• Fixed spectral filtering of PHIS when using interpolated initial condition datasets. For more information see the news page regarding CCM3.6.6 at:
     http://www.cgd.ucar.edu/cms/ccm3/news/ccm3.6.6.shtml.
   
• Bug fix plus tuning modification to restore the documented climate
  • Overestimate of subcloud consumption of CAPE by moist convection. (routine closure). This error had the effect of suppressing moist convection, particularly shallow convection. The parameter adjustment required to restore the simulated climate changes the minimum CAPE for initiating convection from zero to 70 J/Kg (routine conv_ccm). Read the following section (1.1.1) for more information on how this bug-fix (along with the other changes) effect the climate simulation.
• Known bugs:
  • Simulation can't continue more than 68 years on IEEE machines. More information and a patch to fix this option is provided at:
    http://www.cgd.ucar.edu/cms/ccm3/news/year68.shtml
  • Rare Cray compiler bug that causes LSM to blowup: There is a but with the Cray compiler that sometimes (only rarely) causes the code to blowup. The fix doesn't change answers, but allows the simulation to continue.
    http://www.cgd.ucar.edu/cms/ccm3/news/ccmpr19.shtml
  • Problem with packing on IEEE machines. Packing does not work correctly on IEEE machines with CCM3.6, if packing is set history files will be created but they can't be read! This is fixed in CCM3.6.6, however. More information on this bug is provided at:
    http://www.cgd.ucar.edu/cms/ccm3/news/packing.shtml
  • Problem with multitasking for more than 32 processors on RS6000. Information and a bug-fix given at.
    http://www.cgd.ucar.edu/cms/ccm3/news/rs6k_32proc.shtml

The CCM3 news page contains the code-patches for CCM3.2 for all the bug-fixes. The code patches by nature show the code changes in CCM3.6 that fix the above problems.
   http://www.cgd.ucar.edu/cms/ccm3/news/bug_fixes_in_ccm3.6.shtml.

1.1.1 CCM3.6 versus CCM3.2 climate simulations

Most globally and zonally averaged quantities are indistinguishable from the original CCM3 control simulations. The low frequency behavior of the the revised model is also the same as in the original control simulations. There are, however, some notable regional differences in the simulations involving land surface climate and precipitation. In particular, precipitation is slightly suppressed in the western Indian Ocean and tropical west Pacific during DJF, as is the Indian monsoon during JJA. The spatially coherent differences in the land surface climate include colder temperatures and a wetter climate over the JJA southern great plains.

The important point is that although most aspects of the simulated climate are virtually identical between the original (CCM3.2) and revised (CCM3.6) control simulations, there are some notable regional differences. Users should be aware that if they use the new CCM3.6 library they should also use the new CCM3.6 control simulations, especially if examining regional response. In most cases regional differences between the original and new simulations are negligible, although for certain applications the more robust regional differences should be considered in sensitivity analyses. See the CCM3 web page for information on the new control simulations.

Code and documentation for the Slab Ocean Model (SOM) and the Semi-Lagrangian Dynamics (SLD) are provided as an unsupported feature of the distribution. These features will be supported in future releases.

    1.1.1 CCM3.6 versus CCM3.2 climate simulations

 Search page

Questions on these pages can be sent to... erik@ucar.edu .