A Regional Climate Model Study of the Scale Dependence of Cloud-Radiation Interactions

Abstract

The scale dependence of cloud-radiation interaction associated with the parameterizations for fractional cloudiness and radiation used in a global climate model is studied by examining the averages, for different spatial scales, of detailed structure of cloudiness and radiation simulated from a regional climate model that incorporates these parameterizations. The regional model simulation is conducted over an area about (360 km)² located on the southern Great Plains for the period 10–17 April 1994 during which both satellite and surface measurements of radiation fluxes and clouds are available from the Intensive Observing Period of the Atmospheric Radiation Measurement program. The area corresponds approximately to one gridpoint size of a global climate model with horizontal resolution T31. The regional model simulates well the overall cloud and radiation temporal features when averaged over the entire region. However, specific biases exist in the spatial patterns such as the high clouds, the TOA upwelling solar radiation under cloudy conditions, and the net longwave surface flux under clear conditions at night. The cloud and radiation parameterizations are found to be sensitive to the spatial scale of the computation. The diagnosed total cloudiness shows a strong horizontal resolution dependence that leads to large changes in the surface and TOA radiation budgets. An additional experiment, in which the diagnosed cloud at each level is held constant while the radiation parameterization is recalculated, still produces a substantial sensitivity to spatial scale in the calculated radiation quantities. This is because the nature of the cloud vertical overlapping assumption changes as the horizontal scale of the computation varies.