Speaker
Description
Southern Africa is highly exposed to multiple climate extremes, including tropical cyclones (TCs), floods, droughts and heatwaves. In particular, TCs within the Mozambique Channel (MC) domain impose recurrent humanitarian and economic costs. Moreover, objective evaluations of automated TC screening algorithms in this region are still lacking. Hence, there is a clear need to develop, implement and evaluate these algorithms over this domain. This doctoral research investigates the present variability and future projections of these systems, with particular emphasis on TCs in the MC and their implications for Early Warning Systems.
This study validates the Cyclone TRACKing framework (CyTRACK) algorithm for detecting and tracking TCs within the MC from 1999–2025. ERA5 reanalysis data are used as input, and the results are evaluated against the International Best Track Archive for Climate Stewardship (IBTrACS) dataset as a reference. Spatial validation indicates high reliability, with a mean absolute error (MAE) of approximately 42 km, well within World Meteorological Organization operational thresholds. In addition, a total of 71 reference TCs were assessed and 49 TCs were detected and tracked, with a probability of detection (POD) of 69%, and a false alarm ratio (FAR) of 0%. Therefore, this validated framework provides the basis for regional cyclone climatology and variability analysis, particularly future projections.
The research assesses regional climatology, large-scale climate signals and thermodynamics, exhibiting a pronounced seasonal TCs concentration between January and March. In the 2020s, rapid intensification (RI) events increased by 43.8 per cent, whilst landfall systems activity rose 43.5% yearly. Geographical tracking reveals that the central coast of Mozambique accounts for approximately 46.2 per cent of all historically recorded landfalls. Finally, a diagnostic framework using products relating to sea surface temperature (SST), Relative Oceanic Niño Index (RONI), Indian Ocean Dipole (IOD) indices, and compound-scenario analyses demonstrate that interacting ocean/atmospheric conditions substantially alter cyclone metrics (genesis, landfall, severity, RI, and energy dissipation), supported by robust statistical diagnostics.
| Which topic best fits your talk? | Climate and Environment |
|---|