Report of the Climate Variability Working Group Meeting

 Fifth Annual Community Climate System Modeling Workshop  
 E. Schneider, J. Hurrell, R. Saravanan, J. Tribbia, and J. Shukla, Co-chairs  
The Village at  Breckenridge, CO
June 28, 2000

1. PLENARY TALKS

Plenary talks were delivered for the Climate Variability Working Group by Ed Schneider and Joe Tribbia.

Ed Schneider ("Understanding Tropical Interannual Variability in CCSM") spoke on the sensitivity of the tropical coupled circulation to model parameterizations, and the manner in which atmospheric sensitivity to SST and ocean sensitivity to wind stress interact. He concluded that problems in the annual mean and annual cycle of low latitude SST in CSM probably result from CCM3 being too sensitive to SST variability, and also that the ocean might be too sensitive to atmospheric forcing. Tribbia ("Extratropical Variability in CCM3") compared estimates of external (signal) and internal (noise) variability in CCM3 and other AGCMs from seasonal predictability ensemble experiments. In midlatitudes, CCM3 has less external variability and more  internal variability  than most of the other models. The reason for the difference is not understood.

2. BREAKOUT SESSION TALKS

Lawrence Buja ("Community Data Sets and Data Access Issues") described arrangements that have been made to make CSM data accessible to the community. Results from CSM experiments are available on the web at www.cesm.ucar.edu/experiments.

Jerry Meehl ("Factors that Affect the Amplitude of El Nino in Global  Coupled Climate," with Peter Gent, Julie Arblaster, Esther Brady, Bette Otto-Bliesner, and Anthony Craig) showed that decreasing the ocean background vertical diffusivity from the CSM1.0 value to a value near 0.1 cm**2/sec leads to increased and more realistic ENSO variability in the coupled model.  The errors in the annual mean and annual cycle of low latitude SST were not improved by this change.

Andrew Robertson ("The Arctic Oscillation in the CCSM and Hamburg GCMs") found no evidence that coupling to the ocean affects the Arctic Oscillation in either the CSM or the ECHAM/OPYC models. Simulated global warming has a significant projection onto the AO in ECHAM/OPYC but not in CSM.

David DeWitt ("Developing a Coupled Prediction System at the IRI: A Work in Progress") showed results from coupling ECHAM4.5 to a near global version of the MOM3 OGCM at 1.5 degree resolution (enhanced near the equator) with IRI physics (including KPP, isopycnal, and GM). Although the equatorial zonal wind stress is realistic, a double ITCZ with an extra strong cold tongue results in the Pacific. The Atlantic climatology is poor and resembles the original CSM.

Suzana Junzueira de Camargo ("An Intercomparison Between CCM3.6, ECHAM4.5, and COLA2.0 for the Period 1979-1995 in AMIP-Type Integrations") compared precipitation variability of three AGCMs including CCM3 with observations. While each model does some things better than the others, none is always superior.

Max Suarez ("The NSIPP Model's Coupled Variability") described a new version of the NASA coupled model that has an improved annual cycle of SST in the equatorial Pacific, although ENSO variability is somewhat weak, biennial, and detached from the South American coast.

R. Saravanan ("Spiciness in Oceanic Decadal Variability") contrasted density (T'-S') and "spiciness" (T'+S') variability in CSM.

Grant Branstator ("Comparing the Structure of Internal and External  Interannual Variability") estimated that the CCM3 midlatitude response to tropical SST anomalies is about 30% too weak, the tropical response is about 10% too strong, and the magnitude of the noise is about right. He demonstrated that low frequency eddies damp the climatological stationary waves, and the internal variability can be thought of as stochastically forced but organized by the mean flow.

Jim Hurrell ("Understanding Decadal Variations in the North Atlantic Oscillation") was able to simulate the observed multi-decadal trend in the NAO using CCM3 forced by observed SST. Experiments with regional SST forcing indicated that this trend is probably a response to tropical SST trends transmitted from low latitudes via atmospheric teleconnections.

Ed Schneider ("A Numerical Experiment Concerning the Role of the Indonesian Throughflow in the Coupled Climate System," with Roxana Wajsowicz) showed that closing the Indonesian Throughflow in the COLA coupled GCM has a large effect on the climate of the Indo-Pacific region, demonstrating the potential importance of throughflow variability in the climate system and proposed that a similar experiment should be performed with CCSM.

3. DISCUSSION

A PCMDI Scientist described an experiment being carried out at PCMDI using the PCM in which SST from the coupled model is used to force the AGCM in perfect model "AMIP" mode. This experiment was proposed by the CVWG in 1999, and we are pleased to see that this recommendation is being carried out. Further collaboration and coordination with PCMDI was suggested. A comparison of the surface fluxes between the coupled and AMIP integrations should give interesting results.  Those interested in examining the results should contact Michael Wehner (mwehner@llnl.gov).

 4. RECOMMENDATIONS

 Status of old recommendations:

 Old recommendations are summarized in the report of the Natural Variability Working Group from the 1999 workshop (www.cesm.ucar.edu/working_groups/Variability/natural.html).

 Data access: improving. CSM and PCM data are available through the web. Arrangements to examine Paleo model data should be made with the Paleoclimate Working Group (ottobli@ucar.edu).

Perfect model AMIP experiment: in progress (see point 3 above).

T85 AMPI2 ensembles: The Climate Change and Assessment Working Group is planning this type of experiment.

Other long integrations: No action reported.

Improvement of tropical Pacific interannual variability: this has been an active area of research, and positive gains have been made.

Improvement of low latitude SST annual mean and annual cycle climatology and related problems: no progress reported.

OGCM intercomparison (made by Seasonal-to-Interannual WG): no progress reported.

New recommendations:

I. Coordinate evaluation of CAM 1.0 candidates prior to the final selection. The CVWG will solicit for, coordinate, and assist volunteers from the CSM community (aside from those already involved in the AMWG)  in examining results from CAM 1.0 candidates as these results are made  available. The CVWG will collect recommendations from these volunteers  and present them to both the AMWG and SSC. Input from the community is needed prior to the upcoming AMWG workshops, scheduled for September 2000 and December 2000.

 II. The body of existing coupled control integrations is sufficient to justify an effort to collect results from these integrations for comparison and evaluation. The experiments include the original CSM 300 year run, 10 sensitivity studies carried out with the PCM, and several runs made with the Paleo-CSM.

 III. The Polar Climate Working Group pointed out severe errors in the CCM simulations of polar surface winds and requested assistance in eliminating these errors. It is recommended that diagnosing the cause of these AGCM errors and eliminating them be given a high priority.

                                Participant List