Key points are not available for this paper at this time.
Abstract A cloud parametrization scheme which allows for low, medium, high and convective clouds has been developed from GATE data for use in the Meteorological Office 11‐layer tropical model. The problems involved in using synoptic observations to derive methods of predicting clouds are discussed. Only limited success was obtained in relating observed cloud amounts to relative humidity and atmospheric temperature structure. The restrictions imposed on the cloud scheme by the model's resolution and by its inability to produce a perfect simulation are considered. In the light of these difficulties a simple approach was adopted based on the assumption that condensation on the smallest scales is part of a larger‐scale condensation regime related to the synoptic scale situation. The scheme has been designed to reproduce the main features of a cloud field by relating the large‐scale meteorological features associated with a cloud distribution to model variables. Low, medium and high cloud amounts are determined from a quadratic relationship with relative humidity. Low cloud has also been related to the temperature lapse rate in an attempt to model the persistent areas of sub‐tropical stratocumulus occurring under inversions. A relative humidity relationship is inappropriate for convective cloud which has, therefore, been related to the convective mass flux calculated in the convection scheme of the model. The scheme has been reasonably successful in predicting the cloudiness associated with the ITCZ and the NE. and SE. trades. The cloud fields showed a good degree of coherence from day to day and there were no signs of unrealistic feedbacks between radiation, cloud and dynamics.
Julia Slingo (Wed,) studied this question.