This will be an online talk using Zoom, follow the Tuesday Complexity Seminar Online link. 

The theory of critical slowing down (CSD) has been successfully used as a generic indicator of early warning signals in various fields. We propose the use of Persistent Homology (PH) as a preprocessing step to achieve a Flood Early Warning System through critical slowing down. We test our proposal on water level data of the Kelantan River, which tends to flood nearly every year. The results suggest that the new information obtained by Persistent Homology exhibits critical slowing down and, therefore, can be used as a signal for a Flood Early Warning System. We manage to establish an early warning signal for ten of the twelve food events recorded in the river; the two other events are detected on the first day of the food. Finally, we compare our results with those of a Flood Early Warning System constructed directly from water level data and find that a Flood Early Warning System via Persistent Homology creates fewer false alarms compared to the conventional technique.