[an error occurred while processing this directive] [an error occurred while processing this directive]
Minutes of the CCSM SSC Meeting
17 October 2000, NCAR
Attendees: Jeff Kiehl (chair), Dave Randall, Dick Moritz, Bob Dickinson, Peter Gent, Susan Solomon, Byron Boville, Anjuli Bamzai, and Lydia Harper
Not Attending: Maurice Blackmon, J. Shukla, and Jay Fein
1. Atmosphere Model Component Status Report. Dave Randall discussed the 8 different versions of the atmosphere model component that are being compared, showed results of the simulations run so far, showed results of diagnostics performed by NCAR and PCMDI, and discussed problem areas and future plans. A copy of his presentation is available upon request.
2. Ocean Model Component Status Report. Peter Gent reported on the progress of the ocean model component, which is:
Further progress that is being worked on is:
A revised POP code should be ready by the end of December or soon thereafter. Other issues are:
A copy of Gent's presentation is available upon request.
3. Land Model Component Status Report. Gordon Bonan reported on the status of the land model component. He stated that there are many components to a land model, such as biogeochemical issues, biogeophysical issues, river routing, leaf phenology, carbon allocation, etc., and all must be integrated. The land model also must be coordinated with all the other working groups because the atmosphere model constrains the land model and the river runoff must work with the ocean model, etc.
A new Common Land Model has been developed by the Land Model Working Group, but it does not have all the requirements needed for CCSM, such as making surface data sets on any land mask at any resolution, writing history files in netCDF format with user-defined frequency, etc. For the next version of the model (CCSM2), the working group will merge the biogeophysical parameterizations of the Common Land Model and the current NCAR Land Surface Model to form the Community Land Model.
Future issues are that the Common Land Model will continue to be developed independent of the Community Land Model, so that physical components can be developed independent of the CCSM and so that development and testing of new parameters can continue using the Common Land Model.
The CCSM2 version of the model will be available by mid to late December. Bonan stressed that continued software engineering and associate scientist staff support is needed. A copy of Bonan's presentation is available upon request.
4. Sea Ice Model Component Status Report. Dick Moritz reported that new physics, such as elliptical yield curve, multi-category ITD, multi-level temperature in ice, etc., have been implemented; the sea ice model and the ocean model will be on the same grid in CCSM2; code cleanup and maintenance is underway in the sea ice model; a developer's page has been established; sensitivity studies of linear remapping of the ITD, MPDATA, and forcing versus physics is underway; documentation of the sea ice model is in progress; progress has been made in the checkerboard thick ice problem; the standalone sea ice model and the atmosphere/ice/ ocean model is running on the NCAR SGI and at LANL; and future plans are to add frazil ice, snow ice, heterogeneous snow cover, melt ponds, and collaborate with the SHEBA project. A copy of Moritz's viewgraphs is available upon request.
5. Software Engineering Status Report. Cecelia DeLuca reported that the Earth System Model Framework proposal and the NASA HPCC CAN Announcement has increased the NCAR emphasis on software engineering planning. NCAR recognizes that there is a need for a software engineering manager in SCD, and the CGD software engineers working on CCSM recommend a CGD software engineering manager to help coordinate efforts for CCSM and with SCD. The Software Engineering Working Group has posted a Software Engineering Plan for CCSM on the Web at http://www.scd.ucar.edu/css/staff/cecelia/seplan/seplanlinked.html. They ask that the SSC modify and/or approve the plan as soon as possible. Harper will send the URL to all SSC members.
6. Coupler Status Report. Brian Kauffman reported that this group is currently collaborating with the DOE/ACPI team members to upgrade the existing coupler in the near term (called Track 1) and to create a new coupler to replace the CSM and PCM couplers (called Track 2). The coupler group feels that minimal effort should be spent on Track 1 but that it must not be a significant bottleneck to running CCSM on an IBM or a SGI.
Performance tuning to the upgraded coupler (Track 1) is underway, but its exact performance requirements cannot be known until the model components are frozen. Possible problems that might arise have been discussed and solutions planned.
The coupler group feels that a new coupler could be functional by summer 2001 and completed by December 2001. This new coupler will offer a significant improvement in flexibility and performance portability. A meeting of this group will be held at NCAR on 25 and 26 October 2000. A copy of Kauffman's presentation is available upon request.
7. Discuss and Formalize Procedure to Freeze CCSM2. Lawrence Buja discussed the roadmap designed internally to freeze CCSM2. He stated that all the model components are running on the IBM; the Los Alamos National Laboratory (LANL) staff will deliver the POP code in December for the ocean component; the merged Community Land Model is a large job to complete by the end of December; the sea ice model is in the code cleanup stage; and the coupler will be ready when the components are decided. The final atmosphere model component is still in question pending further development and analysis of various proposed configurations by the AMWG. The next steps are to finalize the model components, do component acceptance testing, and complete the validation procedure.
The SSC members discussed the freezing procedure and decided to keep the freeze deadline as the end of December, 2000. CCSM2 will not have everything desired but will be a vastly improved system. At the January meeting of the SSC, the working group co-chairs will be invited to attend and present convincing evidence regarding their recommended model component configuration to the SSC and the SSC will decide the final configuration for CCSM2. It was also suggested that the working groups establish a webpage that shows diagnostic results before the January meeting, so that the SSC members can be familiar with the model experiments and results.
8. Discussion of BAMS Article on CCSM. Randall stated that the draft article still needs to be shortened and that he needed to receive comments from the SSC. He is looking for a cover figure to submit to BAMS along with the article and asked the SSC members to recommend figures that would be appropriate for the cover. He asked the SSC members to help with the reference citations throughout the text and they agreed to do so. The article will also point people to the CCSM webpage and to the bibliography. Harper will research what authors who contributed to the article are not listed and send the information to Randall. The SSC decided to add a thank you to the CAB in the Acknowledgments section and to delete the word Plan from the title of the article. A goal statement will be added regarding providing information to current and future users. Some information will be added about the status of the CCSM project now, and Randall requested that more recent information be sent to him if anyone wanted to add it. How the community participates in the CCSM project and how to get involved in the project will be added to the article. The SSC will send all changes to Randall by 1 November 2000. The SSC thanked Randall for taking the lead on writing the article. Dickinson will begin writing the EOS article.
9. Actions Items from Last Meeting (26 June 2000).