RICHARD SOMERVILLE
Scripps Institution of Oceanography
University of California, San Diego
9500 Gilman Drive, Dept. 0224
La Jolla, California 92093-0224, USA
Davis, CA,
http://www.ucsd.edu/
somerville@ucsd.edu
Abstract:
Our best estimates of future climate change come from numerical climate simulations, but these suffer from serious physical and numerical limitations. Often the physical and numerical aspects of climate models are closely linked. Affordable global horizontal resolutions (typically a few hundred kilometers) mean that detailed simulations of regional climate change are impossible with standard numerical approaches, and even continental-scale phenomena may be severely distorted. Recently, modelers have come to realize that vertical resolution is also critical. Numerical experiments with varying vertical resolution demonstrate that typical models with 20 or fewer atmospheric layers are inadequate for the realistic treatment of cloud-radiation processes, a critical area of climate physics. Cloud-radiation interactions are responsible for most of the factor-of-three difference among models in the sensitivity of simulated global mean surface temperature to an increase in atmospheric carbon dioxide content. We have recently implemented new cloud algorithms and shown that the introduction of more comprehensive cloud microphysical processes alters not only the global sensitivity to atmospheric carbon dioxide concentrations, but also the regional distribution of climate changes.