- Extreme weather events, such as heavy precipitation or strong winds, can occur simultaneously, resulting in larger socio-economic impacts compared to the linear sum of their components. These are termed “compound extremes” and are challenging to predict even just a few days in advance.
- Elucidating the nature of compound extremes is a critical step in advancing our understanding of the climate system and a pertinent societal goal.
Our technology aims to improve the forecasts of compound extremes in a changing climate. We process in two steps:
- Characterization of compound extremes from both traditional and dynamical systems perspectives.
- Improvement of the prediction of compound extremes with an AI algorithm grounded in dynamical systems theory.
- Use of dynamical systems (dealing with the change of any system that evolves in time), efficient for chaotic systems like Earth’s atmosphere and providing a highly informative, cheap, and efficient way of characterizing weather systems and extreme temperature events.
- Combined information from the dynamical systems metrics and ensemble model (using a set of perturbed initial conditions to produce a range of probable future atmospheric states) predictions, giving better forecasts. Thus, more effective preventive measures can be taken, mitigating the storm consequences.
- Flexible analysis algorithm, applicable to various regions and weather extremes worldwide
- The applicability lies particularly in improving the ability to forecast compound extremes at short-medium range time scales to issue timely weather warnings.
- The group seeks to collaborate with entrepreneurs and stakeholders that wish to integrate early weather warnings in their products and services.