Automatic Recognition of Supercells Thunderstorms in Numerical Weather Data

Members - Faculty, students, and collaborators
Research - Details about our current research projects
Theses and Dissertations - Publications and code releases for student theses and disserations
Publications - Recent technical papers and presentations
Software - Recent software releases
 

Abstract

The Tornado Alley is a geographic region in central United States that is home to the largest number of violent tornadoes in the world. The cloud structures that produce the most violent tornadoes are known as supercells. Being able to predict the formation and the evolution of supercell thunderstorms can improve the forecasting of such severe weather events, thereby preventing human and resources loss. Currently, most of the detection of weather events such as supercells are done using radar data. Radar data only provides limited information and it is presently impossible to possess a full set of fundamental variables in real time. However, in the near future, it will likely be possible to use data assimilation to infer those fundamental variables. Updraft helicity is one such non-radar related measure. Updraft helicity is a measure of the amount of an updraft rotation. The goal of this study is to automatically detect at which timestep a simulated thunderstorm evolves into a supercell using exclusively fundamental variables such as updraft helicity and vertical velocity. By solving this problem we are able to identify supercells among nonsupercell thunderstorms using Support Vector Machine as a classifier.

 

Thesis

Thibault Lucidarme (2014). Automatic Recognition of Supercells Thunderstorms in Numerical Weather Data. Master's Thesis, School of Computer Science, University of Oklahoma.

 


Created by amcgovern [at] ou.edu.

Last modified July 24, 2014 5:14 PM