Climate prediction and risk: state estimation and foundational AI methods for climate model evaluation and emulation

Seminar by Dr. Terence O'Kane from CSIRO Oceans and Atmosphere

11 November 2024
KST 14:00 – 15:00

The Seminar is being held in Room 1010 (Jasmin) – Integrated mechanical engineering building. Click here for the campus map.

In this seminar I will describe recent and ongoing work from CSIRO’s “Modelling the earth system” group. Specifically, I describe the Climate ReAnalysis and Forecasting Ensemble (CAFÉ) system, it’s development and available data products. Next, I discuss “ClimRisk”, an AI climate model emulator featured in Nature Vol 628 and in applications to the projection of climate and technology impacts on crops key to food security.  

The second part of the presentation describes Bayesian methods for structure learning in application to climate model evaluation and bias estimation.

Finally, we report on novel Entropy-optimal AI methods for classification and regression learning (over unknown feature probabilities) as applied to ENSO prediction, including comparison to “state of the art” deep learning methods.