Going extreme for forecasting
Classical statistical methods have limitations for characterising extremal behaviour which is of particular interest when we study natural hazards such as volcano eruptions and floods. Just as extreme value models have been used for financial risk forecasting, we seek to build extreme value models to better inform early warning systems.
This project is joint work with Prof. Almut Veraart of Imperial College London and Assoc. Prof. Benoit Taisne of Earth Observatory of Singapore, and supported by the Nanyang Technological University – Imperial College London collaboration grant.
Our research preprint and an earlier poster on the development of a dynamic extreme value model for volcanic eruption forecasting is available online!READ PREPRINTVIEW POSTER
Student engagement activities
Under our collaboration grant, we brought together motivated undergraduate students from Imperial College London as well as Nanyang Technological University to join us in our research on extreme value and spatial methods for environmental risk modelling.