Comparison of 30-Day Integrations with and without Cloud-Radiation Interaction

Abstract

A parameterization package for cloud-radiation interaction is incorporated into a spectral general circulation model (GCM). Fractional cloud amount is predicted quasi-empirically; cloud optical depth is specified for warm clouds and anvil cirrus, but depends on temperature for other subfreezing clouds; the long- and shortwave cloud optical properties are linked to the cloud optical depth. The model's time-mean clouds and its radiative, thermal, and dynamical response to cloud-radiation interaction are investigated for the extended forecast range, primarily by performing two sets of 30-day integrations from real initial conditions for three Northern Hemisphere (NH) winter and three NH summer cases: (i) CLDRADI, with cloud-radiation interaction; and (ii) LONDON, with this GCM's traditional specification of climatological zonal-mean cloud amount and global-mean cloud optical properties. The 30-day mean CLDRADI fields of total and high cloud amount and corresponding outgoing longwave radiation (OLR) fields are plausible in many respects, especially in the tropics where the latter exhibit South Pacific convergence zone (SPCZ)-like and some intertropical convergence zone (ITCZ)-like features, in qualitative agreement with Nimbus-7 and Earth Radiation Budget Experiment (ERBE) observations. Also, the predicted monthly mean OLR anomalies (relative to model climatology) respond to interannual variations in sea surface temperature. Cloud amount and cloud optical depth are apparently underestimated, however, over the higher-latitude oceans, especially over the Southern Hemisphere (SH) circumpolar low pressure belt and Antarctica. The zonal mean bias in shortwave and net radiation remains large at high latitudes in the summer hemisphere, despite the improved longitudinal structure in the tropics. Cloud-radiation interaction elicits a cirrus warming response, which reduces the tropical upper-tropospheric cold bias by ?1?2 K. Over Antarctica, the warm bias in SH summer and cold bias in SH winter are both considerably reduced. During NH winter, the tropical upper troposphere experiences a significant westerly acceleration, including a sign reversal of the zonal-mean zonal wind. By being more conducive to meridional propagation, CLDRADI's tropical westerlies may contribute to the amplification of the quasi-stationary planetary waves in the SH summer extratropics. Otherwise, the impact of cloud-radiation interaction on extratropical geopotential height is generally minimal at extended range.