Forecasting the Tornadic Intensities of Thunderstorms by Multivariate Techniques
Abstract
An investigation on the feasibility of statistically characterizing tornadic and non-tornadic thunderstorms after cells develop is the basis of this research. Nine tornado outbreak cases during the years 1984 through 1988 were studied. The intensity of each tornado producing thunderstorm cell was obtained. Satellite as well as radiosonde data were used to determine the meteorological parameters describing thunderstorm behavior. One of these parameters was the downstream mass flux of an anvil outflow plume (called UMAX) and the other was the rightward deviation angle of the anvil from the storm-relative flow (called MDA). These parameters were used in a two-dimensional kinematic anvil plume model implemented on an interactive computer data system which simulated the actual anvil plume as seen in satellite imagery. A modified Fujita scale rating system was employed to characterize each outbreak cell as a weighted mean called Fw and a weighted mean square called Fw-sq. Fw was a better indicator of tornadic intensity than UMAX and MDA were used as predictors in a bivariate regression model (two-variables) to predict the tornadic intensity of a given observed cell. Linear regression yielded promising results in correlating the data, and described a rising-ridge surface that related UMAX to MDA for both F sub -w. Attempts to relate PBE and shear to Fw yielded mixed results, and only accounted for a small fraction of the error variance in the data. Theses.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 01, 1989
- Accession Number
- ADA218284
Entities
People
- Dale R. Perry
Organizations
- Air Force Institute of Technology