of the CCSM Scientific Steering Committee Meeting
CCSM 7th Annual Workshop, The Village at Breckenridge
24 June 2002
Kiehl (chair), Chris Bretherton, Ping Chang, Jim Hack, Peter Gent, Maurice
Blackmon, Cecilia Bitz, Danny McKenna, Scott Doney, Steve Meacham, Phil
Merilees, Ed Sarachik, Isaac Held, Dave Bader, Lydia Shiver, and Jay Fein
Participants: Jeff Kiehl (chair), Chris Bretherton, Ping Chang, Jim Hack, Peter Gent, Maurice Blackmon, Cecilia Bitz, Danny McKenna, Scott Doney, Steve Meacham, Phil Merilees, Ed Sarachik, Isaac Held, Dave Bader, Lydia Shiver, and Jay Fein via teleconference
1. Welcome. Kiehl welcomed Phil Merilees as the new CCSM Coordinator, Ed Sarachik, Chair of the CCSM Advisory Board (CAB), and Isaac Held, CAB member, as the CAB representatives, and Steve Meacham, Associate Program Director, Climate Dynamics Program, working with Jay Fein to this SSC meeting.
2. Status of CCSM2 Effort. Kiehl reported that the CCSM2 standard model was released to the community on 17 May 2002, along with the data from the control simulation and model documentation and user's guides. The CCSM2.0 Public Release Home Page is at http://www.cesm.ucar.edu/models/ccsm2.0/index.html. It is anticipated that the PaleoCCSM (PCCSM) will be released in July 2002 with a T31, x3 configuration. The CCSM2.0 control simulation is running at NERSC and has run out to year 800. We plan to continue the run to year 1000. It was later discussed that the 1000-year run should go to year 1350 since the last code change was made at year 350.
Kiehl expressed his appreciation and thanks to Dave Bader for his excellent program management at DOE and his collaborations with CCSM and NCAR.
Kiehl reported that CCSM2 simulation successes are improved variability on intra-annual to decadal timescales, realistic teleconnection patterns, much improved ocean circulation in both tropics and polar regions, much reduced deep ocean drifts, realistic salinity structure, and much improved sea ice distribution.
The current model issues are the mean SST biases are similar to ones in CSM1, double ITCZ persists, tropical tropopause is too cold, land temperatures are too warm, and sea ice is too thin.
Kiehl reported that the Annual CCSM Workshop continues to grow every year, and this year there is a formal poster session scheduled to coincide with the reception. There are several subgroup meetings that also take place during the workshop that involve the collaborations of many groups on projects related to the CCSM, such as SciDAC.
The future SSC foci are to coordinate updating the CCSM Plan (2000-2005), to encourage diagnosis of the system (e.g., to address tropical biases), to decide how to develop CCSM in the future, to ascertain how CCSM will fit into the Climate Change Research Initiative (CCRI), and help to plan how CCSM will participate in the IPCC FAR.
Kiehl stated that Merilees would lead updating of the CCSM Plan. He also announced that the Working Group Co-chairs would outline their working group's future plans at a joint meeting with the SSC later in the day. It was reported that the CCRI still has the two-center approach in the conceptual framework document, and IPCC runs need to be started in fall 2003 to contribute to the report that will be published in 2007.
Kiehl stated that an updated version of CCSM will probably be released during the next six months with a few of the major problems fixed, such as land temperatures, cloud scheme problems, etc., and the new coupler will be released.
3. Coupled Model Diagnostic Effort. This subject was not discussed formally.
4. GFDL-NCAR Collaborations on Atmospheric Models. Held reported on the progress of the new coupled model being developed at GFDL and showed the teams that are working on the different models. The GFDL atmosphere team hosted a meeting and invited NCAR atmosphere modelers to focus on the GFDL model and CCSM. AMIP runs were done before the meeting to intercompare the two models. One outcome was that the two groups discussed swapping cloud schemes to see the results and try to pinpoint what makes the two models give different results. The atmosphere collaboration is ongoing, and in the future other GFDL and CCSM model component groups will meet and collaborate.
5. Status of Working Groups and Liaisons. Kiehl reported that the Ocean Model Working Group liaison, Matthew Hecht, has taken a position at Los Alamos, and he will be replaced. The Atmosphere Model Working Group (AMWG) has been searching for a liaison for 10 months and the group is interviewing prospective candidates now. Dave Randall will step down as one of the AMWG co-chairs at the end of July, and Leo Donner has accepted that position. Paul Houser will step down as one of the Land Model Working Group co-chairs at the end of July, and Steve Running has accepted that position.
6. Data Distribution. Kiehl reported that the CCSM2 control simulation data is stored at NCAR and that PCMDI and NERSC have offered to distribute the data also. This solves the problem of having a computer account at NCAR to be able to access the CCSM2 data.
7. The New CCSM (ESM) Plan. Blackmon reported that the NSF Review Panel requested that CGD update its Strategic Plan and that will involve updating plans for the future of the CCSM project. NCAR will also update its Strategic Plan. Blackmon explained the NCAR Initiatives program and how that funding could be proposed for to gain additional funds for CCSM projects. The SSC members discussed that the plan's focus should be on climate change (IPCC scenarios), carbon cycle, paleoclimate, climate variability and its impact on the hydrologic cycle, the role of the tropics in climate variability, and anthropogenic forcing of the climate system. Doney suggested using the CSL proposed experiments as a first cut, then identify working group overlaps, then identify common goals and differences/conflicts, and define any gaps. Bader stated that the IPCC and NRC have posted information on the topic issues and that information should be reviewed. The new CCSM plan should be a community plan not an NCAR plan.
8. IPCC Timetable. Kiehl suggested that the SSC members meet with Susan Solomon at their next meeting to learn about her ideas and plans for the IPCC and where the process stands. Solomon will be invited to the next SSC meeting.
To produce the next IPCC report in 2007, the report has to be final in 2006 for the review process that takes about 1 year. Lead authors for each chapter are being sought at this time. Simulations need to be completed around 2004 for time for analysis and writing of chapters. Computer time needs to be found to do the IPCC scenarios. Kiehl will check with Solomon about specific definitions of scenarios that should be produced and communicate that information to the SSC and Working Group Co-chairs.
The SSC will meet with the Working Group Co-Chairs to define CCSM's role in IPCC and outline plans, i.e., who involved, what runs will be done, and what resources will be used to do the runs. Collaboration should be defined with GFDL and other modeling groups as well.
9. CAB Membership. Kiehl reported that Tom Crowley, Duke University, has joined the CAB, Susan Solomon has stepped down due to her IPCC chair position, and Steve Running has stepped down since he will become a co-chair of the Land Model Working Group.
10. CCSM Business Plan. This topic will be discussed at the next SSC meeting.
11. JOINT SSC/WORKING GROUP CO-CHAIR MEETING:
The Working Group Co-Chairs who joined the SSC meeting were Rick Smith (Ocean); Elizabeth Hunke (Polar Climate); William Collins, Dave Randall, and Leo Donner (Atmosphere); Gordon Bonan (Land); Lisa Sloan (Paleoclimate); Jim Hurrell and Mike Alexander (Climate Variability); Inez Fung (Biogeochemistry); Warren Washington, Jerry Meehl, and Karl Taylor (Climate Change and Assessment); and Tony Craig, Cecilia DeLuca, and Ricky Rood (Software Engineering).
Kiehl opened the joint meeting by thanking Dave Randall for his many years of service and contributions to the Atmosphere Model Working Group.
Kiehl requested that the working group co-chairs inform him and Lydia ASAP about invited speakers and any other invitees that will be funded to come to working group meetings and/or the annual workshop so the funding can be budgeted.
Kiehl opened the meeting for discussion. The AMWG co-chairs asked for a decision by the SSC of how long old versions of models, e.g., CCM3, will be available. The Biogeochemistry Working Group co-chairs also stated that this is an issue for their group as they are also using older versions of the model components from CSM1. The Land Model Working Group wants to use CCM3 and LSM to track differences in signals between the old LSM and the new CLM. The CCAWG is using CCM3 and the Ocean Model Working Group is still using NCOM. Craig stated that keeping all the old and new versions of the CSM and CCSM is possible but will take resources. Kiehl suggested we keep the CSM1 versions for one year, but the SSC will discuss this issue further at its next meeting.
Kiehl reported that the next focus for the SSC and Working Group Co-chairs would be updating of the CCSM Plan. Phil Merilees will lead this effort when he arrives at NCAR on 3 September 2002. The overarching issues identified are: anthropogenic forcing and climate change, solar variability, carbon cycle, extreme events, teleconnections, climate change and the hydrological cycle, role of the tropics in climate variability, and more vigorous paleoclimates activity.
Kiehl requested that the working groups identify plans and crosscutting ideas, identify overlaps between working group's plans that complement each other, and identify gaps in the plan.
The AMWG plans presented by Collins are to incorporate major aerosol species, to develop new radiation parameterizations, to develop alternative formulations of convective processes and transport, and develop new formulation for boundary layer structure and cloud systems. The AMWG will also continue working with NASA's DAO on assimilation and forecasting issues, use CAGATE to explore initial error growth and test it in a weather forecast mode, add EOS data, explore new dynamical cores (finite-volume), and explore use of isentropic coordinates. The AMWG would like to use CAM as a chemical transport model also.
The LMWG plans presented by Bonan are to focus on biogeophysics, hydrology, and river routing by continuing to improve existing physical parameterizations in the model and adding new biological and chemical processes. The short-term tasks are to incorporate new data structures (CLM2.1) and improve biogeophysics and include new biogeochemistry (CLM2.2). The long-term goals are to include terrestrial carbon cycle, prognostic biogeography, land use and land cover change, mineral aerosols, subgrid-scale atmospheric forcing, atmospheric chemistry, water and carbon isotopes, and agroecosystems and incorporate the new coupler.
The OMWG plans presented by Gent are to address biases in the ocean component of CCSM2, such as long term drifts in potential temperature in the global scale biases, incorrect mean flow in the gyre scale biases, and subsurface cold biases south of the equator in the tropics. Other areas of research will be in partial bottom cells and bottom boundary layer, mixing over rough topography, spatial dependence of the Gent-McWilliams diffusivity coefficient (KGM), new advection schemes for tracers, and natural boundary conditions for freshwater.
The PCWG plans presented by Hunke include development of a thorough sea ice model test suite, compilation of suitable forcing data for the ice model, improvement of the prescribed ice configuration and the thermodynamics-only single column model used for lower boundary conditions with the atmosphere model, implementation of the incremental remapping advection scheme in CSIM, testing of dynamics model parameterizations, and continued participation in software engineering projects related to CCSM. This group will perform ice-ocean and fully coupled simulations to test sensitivities to ice thickness distribution, CO2 doubling, Lin-Rood dynamics, sea ice model uncertainty due to forcing, variability, and physics, and to simulate the SHEBA year. They also plan to publish joint scientific papers and maintain web pages about their activities.
The BGCWG plans presented by Doney are to produce a coupled carbon cycle system model and incorporate it and other biogeochemistry into CCSM2, continue the Flying Leap experiments, and contribute to the IPCC scenarios. Doney reported that manpower resources would be needed (software engineer) to incorporate the biogeochemistry into CCSM2, and the group needs to know if the SSC will approve the new data structures in CLM.
The PaleoWG plans presented by Sloan are to continue work on CSENT (climate of 1550-1950), the climate of the past 21,000 years, and extreme warmth climates of the past 100 million years. Sloan reported that there are a number of critical components that are missing in PaleoCCSM and that a functioning and fully useful true paleoclimates CCSM2 is needed. The components needed include the ability to specify orography and topography different from the present day into CCSM2, the lack of runoff and river routing components for different orographies and topographies and T31 resolution in CCSM2, the lack of knowledge and integration of various details now required in CCSM2 (e.g., detailed soil characteristics, detailed ocean bathymetries, and oceanic heat flux values), and the lack of a functioning slab ocean model and land model relevant for paleoclimates limitations (e.g., vegetation specifications). This group will also need additional manpower resources (programmer) to help with the transition to a fully functioning CCSM2 for paleoclimates applications.
The CVWG plans presented by Alexander are to do new model integrations that will be available to the broad community. These will include experiments with CAM in uncoupled mode, as well as intermediate coupled model simulations (e.g., CAM plus an ocean mixed layer model, and POP with an atmospheric mixed layer model) for diagnosis and analysis. He stated that this group plans to jointly work with other CCSM working groups on specific science topics. They also plan to work with other modeling centers (GFDL) and organizations/programs (CLIVAR). To help improve CCSM2, the CVWG will establish a core subgroup to focus on coupled processes and errors and generate a large suite of climate variability diagnostics that will be posted on the web.
The CCAWG plans presented by Washington are to quantify uncertainty in climate change projections, and he stated they would need a "forcing repository" so that all groups can use the same forcings. Other steps to accomplish this are (a) to improve regional climate simulation and extremes by producing more coupled model outputs in time and space for extremes analyses; (b) to run probabilistic projections of climate change by running ensemble simulations with various forcings and scenarios including integrations from 1500-2000 with volcano and solar data working with the Paleoclimate Working Group, (c) to understand the model's response to changes of forcing by running sensitivity experiments with CAM only or mixed layer slab ocean coupled experiments and models from different model centers; and (d) to analyze the 1000 year runs from PCM and CCSM. Key issues for this activity are computer time and model data transfer, storage, and access. This group will produce IPCC scenarios A1F1, A1B, B1, and A2 and B2.
The SEWG plans presented by Craig are to improve CCSM extensibility, flexibility, performance, robustness, and ease of use; continue to support scientific development; improve software process and resource management; and grow knowledge and skills of staff and continue collaborations with other software engineering projects.
Kiehl and Gent thanked Tony Craig for his collaborative work and guidance of the CCSM Software Engineering Group (CSEG), and they thanked the CSEG for their efforts to release the best possible CCSM2.
12. General Discussion of New CCSM Plan. During the discussion some of the ideas were: overall architectural plan needed, NCL group and SEWG should coordinate, integrated validation plan needed with identified problems trying to address, and units in all components should be the same or universal. Merilees will take the lead on defining validation criteria and implementation. Sarachik suggested a problem list with tracking of what was tried and failed and new ideas.