On the Identification and Tracking of Mesoscale Convective Systems Over Southern Africa

The automated tracking of weather phenomena is established in meteorology, particularly for severe weather such as tropical cyclones and mesoscale convective systems (MCS). Typically, weather systems are identified using a set of objective criteria. However, challenges arise when similar characteristics are shared by different weather features. MCSs are one system that is challenging to track since tropical cyclones or cold fronts can exhibit similar features. Although methods for removing tropical cyclones from MCS databases are relatively well established, equivalent approaches for frontal systems remain limited, particularly over subtropical and midlatitude regions such as southern Africa. Here, a new approach is implemented for removing frontal systems from a southern African MCS database. The database was developed using the PyFLEXTRKR algorithm for 2019–2024. Following the removal of tropical cyclones, frontal systems were identified using a shape complexity index (SCI) that quantifies the organization and compactness of tracked cloud systems. Sensitivity analyses based on case studies show that a SCI threshold of 0.5 exceeded for more than 50% of a system's lifetime provides the most effective distinction between organized and disorganized systems. This methodology preferentially removed systems south of 25° S, where frontal systems are common, eliminating 61.5% of tracked systems in this region. The methodology represents a balanced approach that reduces frontal contamination while preserving compact organized MCSs. The resulting SCI‐filtered dataset preserves the principal regions of MCS activity across southern Africa and provides a robust database for investigating MCS characteristics, variability and impacts there.