Evaluation of a Cloud Nowcasting Technique Based on GOES IR Satellite Imagery

摘要

Several techniques are being developed and evaluated for Nowcasting (0-30 minutelead time) cloud and precipitation fields. Accurate forecasts of these fields are needed to support general and air terminal operations as well as to predict periods of microwave communication loss. All the methods currently under evaluation involve tracking contours of either GOES IR brightness temperature for cloud fields or radar reflectivity levels for precipitation fields. In this reprint one of the proposed forecast techniques, the Whole Contour Method, will be presented and evaluated using GOES IR imagery received. The whole contour method is an extrapolation technique that predicts the shape and location of selected contours. The shape of the contour is approximated by separating the X and Y components of the contour boundary into separate spatial functions and calculating the Fourier phases and amplitudes for each function. After a sequential number of observations (minimum of two) of a particular contour is obtained, a forecast contour is constructed based on the extrapolated values of centroid, area, and Fourier components. In its current implementation the Whole Contour Method allows the user to select the number of Fourier components used to approximate the contour shape, the number of observations upon which the forecasts are based, and the order of the polynomial extrapolation equations. (JHD)