A Novel Algorithm for Tracking and Forecasting Convective Cells Using Satellite Image Sequences

Abstract

Accurate storm tracking and forecasting are essential parts of severe weather warning operations. The main problem of existing tracking and forecasting algorithms is unphysical split and merger of cloud clusters within the life cycle of Mesoscale Convective System (MCS). To address this issue, an automatic algorithm called TFCC (Tracking and Forecasting Convective Cells) is proposed for tracking and forecasting convective cells using infrared (IR) image sequences from geostationary meteorology satellite. In this paper, convective cells are utilized for tracking and forecasting instead of MCS because convective cells are stable portion in MCS. TFCC algorithm utilizes overlapping technique and uses a dynamic constraint technique based combinatorial optimization method. Moreover, displacement of the geometrical centroid is utilized to forecast the movement of convective cells. Case studies show that convective cells are tracked and forecasted efficiently in different phases of MCS lifecycle including genesis, maturity and dissipation using TFCC algorithm. Categorical statistics and contingency tables method applied to various case studies over China show that TFCC algorithm efficiently and accurately.