Mathematical model and computer algorithm for tracking coastal storm cells for short term tactical forecasts

Abstract

An algorithm has been developed for real-time forecasting of precipitation storm cell movement over water. The key to the algorithm is the Kalman filter tracking model which is continually updating the mean value and error covariance matrix of a cell's centroid position from past measurements. The algorithm was developed and applied to precipitation cells to evaluate the advantages of utilizing an optimal recursive processing program to assist in making short term tactical forecasts. This tracking algorithm was designed for use-in a stand alone desktop computer. All of the real-time tracking data was detected by a land-based radar system. The thesis results suggest that for short term forecasting the Kalman filter can produce some improvements over other tracking models, but further refinement in identifying the storm cell center and mathematically identifying the area of extent of each individual cell is needed. Other possible improvements to the tracking algorithm might include a methodology to identify what meteorological parameters need to be included in the motion model. Rain Cell Tracking; Kalman Filter.