Global Modeling of Cloud Radiative Effects Using ISCCP Cloud Data

Abstract

Cloud radiative effects are represented in simulations with the general circulation model of the Navy Operational Global Atmospheric Prediction System (NOCAPS) using ingested cloud field data from the ISCCP dataset rather than model-diagnosed cloud fields. The primary objective is to investigate the extent to which the high temporal resolution ISCCP data can be used to improve the simulation of cloud radiative effects on the general circulation in GCM simulations much as observed sea surface temperatures (SSTs) have been used to avoid simulation errors resulting from inaccurately modeled SSTs. Experiments are described that examine the degree to which uncertainties in cloud field vertical structure impair the utility of the observed cloud data in this regard, as well as the extent to which unrealistic combinations of cloud radiative forcing and other physical processes may affect GCM simulations. The potential for such unrealistic combinations stems from the lack of feedback to the cloud fields in simulations using ingested cloud data in place of model-predicted cloud fields. Simulations for the present work were carried out for three April through July periods (1986?1988) using prescribed sea surface temperatures. Analysis of the model results concentrated primarily on the month of July, allowing for a 3-month spinup period. Comparisons with ERBE data show the expected improvement in the simulation of top of the atmosphere radiation fields using the observed cloud data. Three experiments are described that examine the model sensitivity to the vertical structure assumed for the cloud fields. The authors show that although uncertainties in assumed vertical profiles of cloudiness may possibly have significant effects on certain aspects of our simulations, such effects do not appear to be large in terms of monthly mean quantities except in the case of large errors in cloud field vertical profiles. Precipitation fields are particularly insensitive to such uncertainties. A preliminary investigation of potential inaccuracies in our representation of cloud radiative effects with ISCCP data resulting from unrealistic combinations of cloud radiative forcing and other physical processes is made by comparing simulations with 3-hourly and monthly mean cloud fraction data. The authors find little difference in the simulation of monthly mean quantities in spite of large differences in the temporal variability of the imposed ISCCP-based cloud radiative forcing in these simulations. These results do not preclude the importance of simulating the correct temporal relationship between cloud radiative forcing and other physical processes in climate model simulations, but they do support the assumption that a correct simulation of that relationship is not essential for the simulation of certain monthly mean quantities. The present results point favorably to the use of the ISCCP cloud data for climate model testing, as well as further GCM experiments examining the radiative effects of clouds on the general circulation.