Comparison of Squall Line Positioning Methods Using Radar Data
Squall lines are strong indicators of potential severe weather. Yet, automated positioning and tracking algorithms are not common. We propose three different ways to model and identify squall lines using radar images. The three methods are ellipse fitting, Hough transform, and the use of a genetic algorithm-based framework. They model a squall line as an ellipse, a straight line, and adjoining segments of arc respectively. We compare the advantages and limitations of each method in terms of speed, flexibility, stability and sensitivity to parameter settings. It is found that ellipse fitting is the most efficient, followed by Hough transform. Both methods lack flexibility and stability. The genetic algorithm-based framework is stable, has flexibility in modelling and analysis, but comes with a cost of efficiency. The proposed methods provide independent and objective information sources to assist weather forecast.
KeywordsRadar Data Radar Image Weather System Hough Transformation Squall Line
Unable to display preview. Download preview PDF.
- 1.Ahrens, C.D.: Meteorology Today: An Introduction to Weather, Climate, and the Environment. West Pub. Company (2000)Google Scholar
- 2.Troxel, S., Pughe, W.: Machine Intelligent Gust Front Algorithm (MIGFA) for the WSP. Lincoln Laboratory, Massachusetts Institue of Technology (2002)Google Scholar
- 5.Wong, K.Y., Yip, C.L., Li, P.W.: Identifying weather systems from numerical weather prediction data. In: Proc. 18th ICPR, Hong Kong, 841–844 (submitted, 2006)Google Scholar
- 6.De Jongh, I., De Lannoy, G., Verhoest, N., De Troch, F.: Rainstorm characteristics derived from weather radar images. In: Proc. of Euro. Conf. on Radar Meteor., pp. 222–226 (2002)Google Scholar