Minutes of the CCSM Scientific Steering Committee Meeting
23 June 2003
The Village at Breckenridge, Breckenridge, Colorado
Attendees: Jeff Kiehl (Chair), Chris Bretherton, Cecilia Bitz, Jim Hack, Bill Collins (incoming Chair), Bill Large, Maurice Blackmon, Jay Fein, Dave Bader, Phil Merilees, Max Suarez (acting CAB Chair), Isaac Held, Eric Sundquist, Terri Paluszkiewicz, Cecelia DeLuca, Tony Craig, Peter Backlund, Susan Solomon, Dale Kellogg, Tim Killeen, and Lydia Shiver. Scott Doney and Ping Chang were unable to attend.
1. Welcome. Kiehl welcomed everyone to the meeting. He showed the composition of this year's CCSM Annual Workshop, which was 320 registrants.
2. Update on CCSM for IPCC Simulations. Kiehl stated that his goals when he became SSC Chair were to produce a CCSM brochure, business plan, science plan, CCSM version for IPCC, and a slide repository. He completed the brochure, business plan, and science plan, and the IPCC version of CCSM has started running. He stated that the next step would be developing the CCSM Implementation Plan, which probably should be a Web document, so that it could be a dynamic document.
Regarding the IPCC version of CCSM, he reported that the first iteration of all the component models was completed, and that each new component addressed previous biases in the model, such as the high latitude warm bias, cold tropopause bias, and the double ITCZ bias. He noted that there were significant differences in the version of CCSM meant for IPCC runs over CCSM2, and the largest and most aggressive changes were in the atmosphere model. There are still two candidates for the atmosphere component dynamical core: the spectral Eulerian and the finite volume. Both of these models will be run to produce AMIP simulations in an uncoupled mode, but the finite volume dynamical core model was not ready at this time to run in coupled mode.
There was much discussion about the amount of computer resources that would be needed for the IPCC simulations.
Kiehl volunteered to discuss a mini-comparison project of North American models in the U.S. and Francis Zwiers volunteered (through Kiehl) to discuss this in Canada. A central archive for data in the same format will need to be established. More and more groups, i.e., CPTs, want a common data format and common diagnostic tools.
Kiehl thanked the SSC, Jay Fein, Dave Bader, and Lydia Shiver for their help while he served as SSC Chair. Jay Fein thanked Jeff for his CCSM leadership for the past two years.
3. Multi-Tiered Climate Modeling Centers/Update on CCSP. Fein stated that NSF, DOE, and NOAA had endorsed a multi-tiered approach for climate services, but exactly how it would be implemented, especially when centers other than GFDL and CCSM are involved, had not been decided.
Fein stated that the final draft of the CCSP Plan was distributed to the agencies, but there will be more iterations. [UPDATE: The final form of the Strategic Plan for the U.S. Climate Change Science Program (CCSP) has just now (7-24-03, 11AM) been released:
A shorter, "vision" document is also available:
4. Tropical Biases Workshop Update. Hack reported that the workshop group listed the common biases of climate models as the double ITCZ, simulations of stratocumulus regimes, an excessive cold tongue in the tropical Western Pacific, weak tropical variability, and mean temperature and humidity biases. Some conclusions were that non-flux corrected models share many of the same problems; AGCM parameterization sensitivity results are generally mixed; resolution effects are mixed; and stratus sensitivity studies uniformly agree. The group summarized that all models are remarkably consistent; no potential sources of simulation biases were eliminated; and no clear consensus was reached on a strategy to sort out current biases. Hack suggested that if the workshop was repeated, that a strong focus was needed on using observationalist guidance to develop hypotheses and that a consensus on some set of hypothesis-driven experiments should be developed.
5. Status on Potential FY04 Funding for CCSM. Fein stated that he and Margaret Leinen are cautiously optimistic about the CCSM budget increase in FY04, but it will depend in large part on a substantial increase in the overall FY 2004 NSF budget. He also stated that it is usually late in a fiscal year before we know our final budgets.
At the time of the workshop, Kiehl and Hack are the PIs for the GFDL and NCAR CPT collaborations on two atmospheric sciences proposals that would fund a new associate scientist and two postdocs. Bill Large is the NCAR collaborator/coordinator for about a half dozen ocean sciences related CPT submissions, for which an NCAR postdoc or scientist will be funded to work under Bill.
6. Discussion on Changing the Name of the Ocean Model Working Group. The SSC decided to leave the formal working group name as the Ocean Model Working Group.
7. Discussion on SSC Membership. Kiehl reported that Ping Chang wanted to rotate off the SSC, and Bitz and Bretherton stated they were willing to continue for another term. The SSC discussed the need to expand the membership of the SSC, keeping the balance of NCAR and non-NCAR scientists, so that more disciplines are represented. Blackmon suggested Gordon Bonan be appointed to represent the land component, and Collins agreed to talk to Bonan about serving. Other suggestions were Inez Fung and Scott Denning for biogeoscience and a new oceanographer to replace Chang. Bader stated that he would continue as Chief Scientist of the DOE CCPP after he moves to Livermore, and he will work on getting DOE headquarters to approve his continued participation in the SSC and CAB.
8. CCSM 2004 Workshop. The SSC decided to move the 2004 workshop to the Eldorado Hotel in Santa Fe, New Mexico. Because Monday, 5 July, is a government holiday, the 2004 workshop dates will be Wednesday-Friday, 7‑9 July 2004, with a SSC meeting on Tuesday, 6 July, a SSC and Working Group Co-Chair meeting on Friday afternoon, 9 July, and a SSC/CAB meeting on Saturday, 10 July, or possibly a combined SSC/CAB meeting on Tuesday afternoon, 6 July. We will continue the poster session and forum. The working group meetings should be rotated to different days each year, and the working group meetings at the workshop may become business meetings, with science meetings at other times of the year. SSC members requested the CAB members at the meeting encourage all CAB members to attend the workshop. Lydia will finalize negotiations with the hotel in Santa Fe and announce the dates and venue via email ASAP.
The theme for next year's workshop will be biogeochemistry. Suggestions should be obtained from Doney, Fung, Mahowald, Running, and Bonan. Collins will circulate a list of plenary speakers to the SSC for consensus ASAP.
9. Discussion on CCSM Implementation Plan. Collins stated that the first priorities are to agree on the sequence of CCSM hires and how to allocate the $10 million of new funds CCSM may receive in FY04. A global aerosol associate scientist has already been hired (Conley). CCSM hires in the business plan are 5 Ph.D. scientists, 7 Project Scientists, 7 Software Engineers, and 2 Administrative Assistants. CCSM will also recruit 5 visiting scientists and several postdocs that could be located at NCAR and at universities. Collins top priorities for hiring are an ocean biogeochemical modeler, cloud microphysical modeler, and a global aerosol modeler. Large stated that careful thought needs to be given to the hiring of Ph.D. scientists regarding current needs and future needs and NCAR scientific career paths. Bretherton stated that a top priority hire should be in the dynamics/processes area. The SSC will consider CGD's hiring plans to ensure the CCSM staffing is complementary. Other priorities noted were a CCWG liaison (project scientist) and AMWG and OMWG support scientists (project/associate scientists). Doney has requested that Keith Lindsay become a full-time BGCWG liaison rather than one-half time. Bader suggested a full-time consulting desk liaison would be useful to help with the questions from the community.
The most flexible way to ramp up the hiring process would be to recruit visitors, postdocs, and graduate students from a mix of disciplines across working group's needs. It was suggested that we should link the CCSM Education and Outreach plan to the hiring priorities also. Other ramp up ideas were to recruit a very prominent scientist to join CCSM at NCAR; hire scientists to work at their home institutions on CCSM-related issues; buy out teaching time at universities, so scientists could have more time for CCSM-related research; hire postdocs and attract visiting scientists ASAP; fund additional CPTs; and acquire computing resources (purchase or lease). Fein suggested we deal with near-term needs as currently available resources allow, and carryover any new funds that may arrive in late FY 2004 to support longer term needs.
Collins will update the SSC on the progress of the Implementation Plan every 6 months and especially 1 month before SSC meetings.
10. ESMF and CCSM. DeLuca stated that CPL6 was compliant with ESMF requirements, and CCSM is running (not in production mode) using ESMF. The CGD Software Engineering Group (CSEG) is taking the lead working with ESMF. It was suggested that DeLuca give a CGD seminar on ESMF as an education and outreach session and also make presentations at working group meetings. DeLuca and Craig would also make a presentation to the CAB.
11. NCAR Earth Simulator Science Plan.
The ES Science Plan and MOU were summarized by Backlund as:
- the steering committee for this project would have two co-chairs, one from NCAR and one from CRIEPI;
- all ES simulations would be open and available;
- all results would be labeled NCAR and CRIEPI;
- this steering committee would make the decisions regarding what results would go to IPCC;
- anyone can publish results but each party must be notified;
- close coordination with CCSM development and SSC planning and management is essential;
- no co-authorship wording is contained in the plan, but Backlund can add it if it is wanted;
- define ES IPCC simulations that complement already planned CCSM simulations;
- add a software engineer to CCSM's staff to support the vectorization of the model;
- a science team needs to be formed (Bryan, Boville, Large, and others to be determined); and
- an evaluation of ES upgrades and U.S. hardware needs for the initial vectorization needs to be completed.
The next steps were defined as:
- SSC endorsement needed;
- refine the science plan and software engineering plan and carefully look at mitigating risks;
- complete discussions with agency sponsors and OSTP, CCSP, and Department of State;
- complete negotiations of the MOU;
- complete initial vectorization of the model components;
- test and evaluate the coupled model, design and implement scientific experiments, and complete the assessment of the simulations; and
- regular reporting to the SSC Chair who will report to the SSC.
Discussion of the ES Science Plan and MOU ensued. Bretherton expressed concern that the demands on the software engineers and the scientists might be underestimated. Backlund stated that the NCAR Directorate would fund one software engineer and could reconsider human resource requirements as the project evolves. Blackmon stated that NCAR could propose to Dr. Sato for ES time aside from time to perform IPCC runs. The SSC decided that the IPCC runs must be clearly distinguishable from each other (CCSM and ES). Bitz stated that university participation was not included in the plan. Backlund stated that other parties (universities) could join NCAR as a partner in the MOU, but they would have to contribute in some way. It was decided the MOU should be between CCSM and CRIEPI instead of NCAR and CRIEPI.
Killeen stated his priorities for the CCSM/CRIEPI collaboration were that use of the ES was science driven, IPCC relevant, does not disrupt CCSM work flow; SSC take ownership; and all data will be accessible.
The SSC gave provisional endorsement of the ES MOU and science plan and requested a list of specific IPCC scenarios that would be run on the ES. If the SSC agrees to the list of scenarios, the provisional endorsement would become formal. The SSC encouraged Blackmon to work with the CCWG. The SSC does want CCSM vectorized so that it will run on a Cray X1 and the ES. The SSC agreed that Craig should hire a software engineer for the vectorization task.
12. Updated IPCC Plans, including Results of the Marrakech Planning Meeting and Future Developments. Susan Solomon presented information and a timeline for IPCC. She said the IPCC report outline was due November 2003, zero draft report was due January 2005, first draft complete by August 2005, second draft complete by February 2006, and WG I panel approval due January 2007. The drop-dead date is likely to be February 2005 to May 2005 for simulations. She emphasized that the proposed outline of the report has yet to be finalized, and that the proposed outline will be submitted to the IPCC Panel for approval by governments. The talk therefore was to be considered a description of possible content, but it should be noted that the ideas presented were preliminary.
IPCC will likely consider model ability to represent variability in the climate system; ability to predict forced and unforced variability in atmosphere and ocean; climate sensitivity; and stabilization and inertia scenarios to be run by a range of models (AOGCMs, SCMs). The IPCC will also likely address paleoclimate and ice. Finally, coupling between climate and land surface, carbon, and atmospheric chemistry represents key new advances in understanding human interactions with the coupled climate system that will likely be assessed.
Some questions Solomon presented were: Can we do more to explain more linkages between climate variability and trends in observed patterns of climate change, and to identify any anthropogenic linkages? What are impacts of the NAO on climate and related variables? How much of the reported variability in ocean heat uptake is real and can models simulate it?
Regarding climate model projections, some researchers will likely be using simple models and all will be using the same emissions but different forcings, so the forcings must be tracked very carefully, and there will be a temperature range for CO2 doubling. Solomon noted the importance of addressing the fact that the transient climate response is different from steady state, and to examine observations that might help test model responses.
IPCC will sponsor a climate sensitivity workshop the last week of July 2004 in France to discuss why models produce different climate sensitivities.
The scenarios to be run with AOGCMs will likely focus both on near-term climate change and long-term climate change. IPCC expects that there will be a separate assessment of carbon cycle from assessment of GCM responses to specified forcing profiles. There is interest in increased use of ensembles for studies of predictability, and the initial set of runs requested will likely focus on A1B or A2, as well as B1 with runs being as long as possible and even out 300 years to address requests for information on stabilization by the Parties. Solomon noted that various groups are running at a range of resolutions, and there is no specified minimum for IPCC.
There is also considerable interest in carrying out so-called commitment runs in respect to the 20th and 21st centuries. That is, one fixes the level of greenhouse gases at some time and calculates the resulting climate change that results.
The impacts community would need the simulations and data sets archived by the middle of 2004 by the latest, and papers should be submitted by about May 2005 and either accepted or in print by November 2005 to be included in the IPCC fourth assessment. The model runs should be done by about September 2004 or earlier, and they must include daily values of temperature, pressure, and precipitation for specified time slices, and can extend out to 2200 or 2300.
The content of the next IPCC report will likely consider aspects of understanding of the coupled climate system that describes the emerging linkages in understanding the simulations of the fully coupled system; better identification of possible wild cards that could emerge in altered climate states; discussion of process-level understanding of soot and other aerosols, as well as precipitation pattern changes; and carbon climate futures.
In summary, Solomon said some important topics are changes in specific regions and the links to circulation; science involving ice and the oceans along with the atmosphere; information on timescales and stability as an increasing interest to policymakers; and aerosols, hydrological cycle, carbon cycle, atmospheric chemistry, land surface, and other coupling.
Solomon will give Collins feedback and information on what the other centers around the world are doing for IPCC.