A weather radar data processing module for storm analysis

A pre- and post-processing weather radar data module was developed in the Matlab suite of software with GIS data exchange abilities for storm event analysis. During pre-processing, each radar sweep is converted from spherical to Cartesian coordinates in the desired temporal and spatial resolution. The module's functionality in post processing includes radar data display, geo-referencing over GIS maps, data filtering with the Wiener filter and single or multiple sweep processing. The user can perform individual storm cell detection and tracking, resulting in the storm's average velocity and track length. The tested methods are modifications of the LoG (Laplacian of the Gaussian) blob detection method and a Brownian particle trajectory linking algorithm. Radar reflectivity factor (Z) data can be referenced over predefined rainfall (R) gauges in order to determine the radar Z-R equation parameters. The user can also produce spatially distributed precipitation estimates by using standard Z-R equations from the literature. The module's functionality is demonstrated using data from a rainfall event captured by the NSA Souda Bay C-Band radar during a storm in October 2006. Results show that the Rosenfeld Tropical Z-R equation is the one that gives a satisfactory description of the spatial and temporal precipitation distribution of the investigated event.