THE IMPACT OF A DIAGNOSTIC CLOUD SCHEME ON FORECASTS WITH A GLOBAL ATMOSPHERE MODEL

Experiments were conducted to test the impact of a cloud diagnosis scheme in place of prescribed zonal average cloud on medium and long range integrations with the Australian Bureau of Meteorology Research Centre(BMRC)global atmosphere model.The cloud scheme was shown to improve the temperature bias in the lower troposphere but there was deterioration in the upper troposphere,especially in the tropics,associated with underestimation of high cloud amount.Thirty day mean fields in a January integration showed greater amplitude in the Northern Hemisphere planetary waves and a deeper Antarctic circumpolar trough than the control experiment or a simulation with no cloud.The results for the diagnosed cloud case agree more closely with corresponding observed fields.There was also some reduction in the zonal average zonal wind component reflecting the additional zonal asymmetry introduced by the diagnostic cloud scheme.Similar trends were also noted in medium and long range forecasts for January and July conditions,although the impact on predictive skill was slight and in some cases detrimental.Areas for improving the diagnostic cloud scheme are noted.