The purpose of the work reported here was to produce a videotape of Hurricane Ignacio (1979) that resembled time-lapse photography, except that the successive images were in the Lagrangian frame instead of the usual Eulerian frame. Procedures and programs are described for taking the original satellite images of the storm, finding the storm, centers, rotating the images the proper amount, and displaying them in succession for copying to videotape using special hardware. (Author).

Linear regression estimates of radiosonde observed vertical wind shears between mandatory levels were made, using radiance gradients obtained via the TIROS-N satellite as predictors in a pilot experiment based on about 200-500 observations. The best results were obtained for shears between 850 mb and a level in the upper troposphere, such as 300 mb. Results were similar for 850-400 or 850-200 mb shears, but notably inferior for 850-500 mb shears. The linear correlation coefficients between estimated and smoothed observed components of shear was about 0.70, corresponding to about 50 percent explained variance in independent test samples. Somewhat higher explained variances (about 0.60) were achieved in a limited trial where the data-set was stratified according to subjectively-judged trajectory curvature. Estimates of the wind itself in the upper troposphere were nearly as skillful as the best shear results. Overall, with the possible exception of curvature-stratified estimates in clear regions, both wind and wind shear estimates are not likely to be superior to methods that generate RMS wind errors of order 5m/s. Most of the information was contributed by the linear combination of microwave channel 2 minus microwave channel 3. Some additional information may be obtainable from the infrared (HIRS-2) channels 3, 4 and 5 (especially 5). (Author).

The development of a model to estimate 30 and 50 kt wind probabilities from tropical cyclone forecasts is described. Some examples to illustrate the use of probabilities are discussed. The model is based on position forecast errors, which are used to determine the probability of a cyclone occupying a particular geographical position, and on wind profile errors. Wind profile errors consist of errors in the forecast maximum wind and errors in the forecast radius of 30 and 50 kt winds. The profile errors are used to estimate the probability of 30 an 50 kt winds occurring at a point, given that the cyclone occupies a particular position. These position and wind probability elements are combined by using an assumption of independence which is supported by correlation coefficients. The model, which includes features of the earlier strike probability model, is tested on independent data. Test results are shown to illustrate good agreement between forecast probability and the frequency of occurrence of 30 and 50 kt winds. (Author).