Does Changing Atmospheric Model Resolution Affect Atmospheric Feedbacks?

Abstract:

Simulations of the Community Atmosphere Model version 5.1 (CAM5.1)
at horizontal resolutions of approximately 2, 1 and 0.25 degrees
driven with climatological sea surface temperatures (SST) and
1990 forcings were carried out. The 1 and 2 degree CAM5.1
configurations used the default parameter values with the 0.25
degree CAM5.1 using the 1 degree configuration except the physics
timestep being halved. Perturbed experiments, using CAM5.1, in
which either SST is uniformly increased by 2K or CO$_2$ doubled were
also carried out using the same configurations.

A ``Cess'' type feedback analysis (twice change in 2xCO$_2$/change in
2K simulations) was used to diagnose a ``Cess'' sensitivity. This
sensitivity increased slightly with resolution due to changes in
both the response to the uniform SST increase and to doubling
CO2. This appears to arise from differing changes in tropical clouds
as resolution increases.

Our results suggest that uncertainty in climate sensitivity is not
strongly impacted by changing horizontal resolution up to 25
km. Thus, uncertainty in parameterisation likely remain the leading
source of uncertainty in climate sensitivity.