Antarctic clouds and radiation in the NCAR climate models

The largest climate variance between global models is in the polar regions. An important reason is that parameterizations of clouds and cloud radiative effects are frequently more appropriate for mid-latitudes and the tropics than for high latitudes. As an example, clouds are generally much thinner over Antarctica due to the lower water vapor amounts. Furthermore, the aerosol concentration over interior Antartica is extremely low. To evaluate and improve the treatment of clouds and radiation by the climate models of the National Center for Atmospheric Research (NCAR), we examine simulations by the NCAR Community Climate Model version 3 (CCM3) as well as the recently released Community Atmosphere Model version 2 (CAM2) and Community Climate System Model version 2 (CCSM2). The parameterizations tested include the Rasch and Kristjansson prognostic cloud particulate scheme, which is now the standard scheme for CAM2 and CCSM2. Also tested is the Rapid Radiative Transfer Model (RRTM), which alleviates the deficit in downward clear-sky longwave radiation. The new radiation scheme in CAM2 and CCSM2 has a similar effect. It is found that there are serious errors in the simulation of Antarctic cloud radiative effects. The biases introduced by errors in the simulations of clouds have a much larger effect on the climatological radiation balance than the clear-sky downward longwave bias. The climate simulations show a very large cold bias in the stratosphere, especially during summer. The optical thickness of Antarctic clouds appears to be excessive. This contributes to a warm bias in surface temperature during winter and a deficit in downward shortwave radiation during summer. These biases for Antarctica are larger for CCM3 with the prognostic cloud particulate scheme than with the standard diagnostic clouds. When the threshold for autoconversion from ice cloud to precipitation is reduced towards a more realistic value, the Antarctic clouds are thinned and some of the biases are reduced. To improve the surface energy balance, not only must the radiative effects of clouds be improved, it is also necessary to improve the representation of sensible heat flux. It is likely that insufficient vertical resolution of the frequent very shallow, very stable surface boundary layer contributes to an excessively large sensible heat flux. It is found that the representation of clouds and radiation is not clearly improved in the latest NCAR climate models. For example, the surface albedo over Antarctica is decreased in CAM2 and CCSM2 simulations in comparison to CCM3 simulations. This change contributes to a warm bias in surface temperature during summer.