Drought and heat: Performance of seasonal predictions on global and regional scales. Key takeaways from the UCT-UCR seminar
By Senani Mamba | WPE
This month's UCT-UCR Seminar Series, "Water Economics and Politics," featured Prof. Dr Harald Kunstmann from the Karlsruhe Institute of Technology (KIT), the University of Augsburg and an honorary research associate at the University of Oxford. The session, moderated by Prof. Djiby Thiam, focused on “Drought and heat: Performance of seasonal predictions on global and regional scales.”
The discussion was a mix of theoretical and empirical perspectives on the estimation procedures for drought and heat, with real-life examples of how proper estimation techniques of drought, precipitation and heat could influence productivity at a regional and global scale based on projects he had worked on in Europe, the Horn of Africa and the Blue Nile region.
Key takeaways:
- Kunstmann showed that estimation of seasonal forecasts could be improved as the current ones being used “have issues”, using the forecast raw SEAS5S LM1, significant deviations from reference were noted and often below climatological forecasts as they currently use the lead forecast period.
- Other issues with current estimation techniques Kunstmann noted were bias and drift, which influenced the forecasts taken from Germany. He showed results from sub-seasonal and seasonal forecasts and their initial uncertainty conditions, as well as their probabilistic nature and how these have contributed to the climatological forecasts, which are not precise.
- Kunstmann recommends approaches to correct the suspected estimation procedures by using a statistical BCSD and AI corrections, with AI corrections showing a much more precise estimate within Europe, the greater Horn of Africa and the Blue Nile region.
He summarised that these estimation procedures were good news for early prediction indicators of weather patterns. He demonstrated that it provides better temperature estimates than precipitation. Drought Hazard indicators have more skill than precipitation, with the first month working best, but six months performing weaker, though still better than climatology estimates.
This presentation is part of the UCT-UCR Seminar Series: Water Economics and Politics, a joint initiative between the University of Cape Town (UCT) and the University of California, Riverside (UCR). About this presentation: Drought and heat: Performance of seasonal predictions on global and regional scales.