Estimating cloud top height and spatial displacement from scan-synchronousGOES images using simplified IR-based stereoscopic analysis

An efficient method for estimating cloud top heights and correcting cloud i mage spatial displacements was developed. The method applies stereoscopic a nalysis to a pair of scan-synchronous infrared cloud images received from t wo GOES satellites using a piecewise linear approximation of the relationsh ip between height and infrared brightness temperature of top of the cloud e lement. The algorithm solves for cloud top heights and subsequently calcula tes the spatial displacements of cloud images. Optimal parameterization of the piecewise linear approximation is achieved using the shuffled complex e volution (SCE) algorithm. Because the proposed method simplifies the stereo scopic analysis, it allows for an easy implementation of stereoscopic techn ique on desktop computers. When compared to the standard isotherm matching approaches, the proposed method yielded higher correlation between GOES 8 a nd GOES 9 scan- simultaneous images after the parallax adjustment. The vali dity of the linear approximation was tested against temperature profiles ob tained from the multiple ground sounding measurements from the Tropical Rai nfall Measuring Mission/Texas and Florida Underflights (TRMM/TEFLUN) experi ments. The results of this comparison demonstrated good fit, particularly w ithin the troposphere, between the optimized relationship and atmospheric s ounding measurements. The data produced by this method, including cloud top temperatures and heights, atmospheric temperature profiles for cloudy sky areas, and spatial displacement-adjusted cloud images, can be useful for we ather/climate and atmospheric studies. In particular, the displacement-adju sted cloud images can be critical to develop high-resolution satellite rain fall estimates, which are urgently needed by mesoscale atmospheric modeling and studies, severe weather monitoring, and heavy precipitation and flash flood forecasting. Limitations of the proposed method are also identified a nd discussed.