[an error occurred while processing this directive] [an error occurred while processing this directive]

Report on CCSM Atmosphere Model Working Group Meeting

Hotel Boulderado, Boulder, Colorado

Co-Chairs: Phil Rasch and Leo Donner

9-10 March 2004

Approximately 100 participants attended the two-day meeting in person. We also web-cast the meeting. Both audio and video were available via a web-browser. 50 individuals and groups around the country used this method to monitor the presentations remotely. There was a very positive response to the webcast.

The agenda for the meeting can be accessed here.

There were approximately 25 talks presented at the meeting. Most of the talks are available for download here.

Session 1 focused on recent CAM developments. Collins provided an overview of the status of CAM in the context of the coupled climate model, and how the model will be used for the upcoming IPCC exercise. Rasch provided an update on the precise configurations of CAM that will be released to the atmospheric community in June 2004. Hack described the behavior of CAM at various resolutions and its interactions with physical parametrizations within the model. Kiehl described his initial exploration of the sensitivity of model climate to SST variations.

Session 2 dealt with application studies in the CAM community. These studies used older configurations of CCCM and CAM to explore studies with the current model (CAM) and older configurations (CCM). We heard about initial explorations with a spectral element dynamical core (Wang), progress reports using CAM with convective super-parameterizations (Grabowski and Khairoutdinov), and the enormous computational speedup that can be achieved by replacing the standard radiative transfer calculation with a neural net trained to replicate the behavior of the standard scheme. Initial explorations of CAM with interactive tropospheric and stratospheric chemistry were described by Lamarque and Sassi, respectively. Bergman described his generalization of the CCM single column model to parameterize the large scale forcing and allow that forcing to interact with the subgrid parameterizations.

A variety of presentations were made in which CAM simulations were strongly influenced by observations. Williamson presented results analyzing very short simulations started from analyses to identify model errors. Anderson et al., presented results in which they employed CAM in a data assimilation framework and indicated that CAM is capable of quite good forecasts. Hess presented work using CAM as an offline chemical transport model to allow simulations of trace constituents using CAM strongly constrained by observed meteorology. Xie described his exploration of variations in convection in CAM2, and Ghan described the impact that a new subgridscale orography formulation has on surface properties in the model.

Session 3 dealt with the future of CAM Physics. This session inaugurated the process of critically assessing the treatment of physical processes in CAM with an eye toward future development. We asked Adrian Tompkins of the ECMWF to provide an overview of current approaches to parameterizing cloud fraction in GCMs, their limitations, and future prospects. Of course this topic is intimately tied to representations of other processes (e.g., convection, dynamics, and cloud microphysics) and these topics were also discussed. Bretherton described the new Cloud Feedbacks Cloud Processing Team (CPT) and showed some results from their evaluation of three different models. Dickinson discussed deficiencies of CAM for current land models and suggested constraints that future versions of CAM should address. Collins discussed his ideas for future versions of the radiative transfer calculation. Bretherton presented results of simulations with CAM3 using some promising parameterizations for boundary layers and boundary layer clouds.

CAM is currently being deployed in a wide range of applications, and numerous explorations of its basic physics, numerics, and dynamics are under way. A consensus emerged from the meeting that the parameterization of cloud fraction (or, more precisely, sub-grid variations of water in its various forms, of which cloud fraction is a coarse indicator) requires significant attention. A call to develop a team to implement and evaluate new approaches to this problem was issued at the meeting. A group to develop and diagnose new approaches is being formed in response. Statistical methods will play a major role in this effort.