Sub-seasonal extreme rainfall prediction in the Kelani River basin of Sri Lanka by using self-organizing map classification.
By J. F. Vuillaume, S. Dorji, A. Komolafe and S. Herath
- Online availability of multi-model and ensemble sub-seasonal forecasts has sparked interest in extreme rainfall prediction and early warning systems.
- Developing tropical countries like Sri Lanka face complex meteorological challenges and need effective early warning systems for flood mitigation.
- This study examines the potential benefits of the Sub-seasonal to Seasonal (s2s) database, offered by a consortium of weather forecasting institutes, using self-organizing map classification.
- Key findings of the study include:
- Establishing a connection between teleconnection indexes like the Madden–Julian Oscillation and spatiotemporal rainfall patterns.
- Demonstrating that the frequency of heavy rainfall events is influenced by the cluster type.
- Observing varying performance of s2s forecasts across different clusters.
- Introducing corrective bias coefficients for forecasting water volume in the basin for each cluster.
- The study emphasizes the value of s2s forecasts for extreme rainfall prediction and advocates for the real-time release of s2s data, particularly beneficial for early warning systems in developing countries like Sri Lanka.
