USING SATELLITE DATA TO REDUCE SPATIAL EXTENT OF DIAGNOSED ICING

Overprediction of the spatial extent of aircraft icing is a major prob lem in forecaster products based on numerical model output. Dependence on relative humidity fields, which are inherently broad and smooth, i s the cause of this difficulty. Using multispectral satellite analysis based on NOAA Advanced Very High Resolution Radiometer data, this pap er shows how the spatial extent of icing potential based on model outp ut can be reduced where there are no subfreezing cloud tops and, there fore, where icing is unlikely. Fifty-one cases were analyzed using two scenarios: 1) model output only and 2) model output screened by a sat ellite cloud analysis. Average area efficiency, a statistical validati on measure of icing potential using coincident pilot reports of icing, improved substantially when satellite screening was applied.