Deep Learning for Global Climate Monitoring and Predictions

Explore deep learning applications in global climate monitoring and predictions in this 53-minute lecture from the ICBS2024 conference. Delve into the paradigm shift in climate prediction, examining state-of-the-art deep learning models for El Niño forecasting that outperform conventional dynamical approaches. Investigate the challenges posed by limited reanalysis data and discover innovative solutions, including few-shot learning and deep learning-based data assimilation techniques. Learn how these methods significantly improve mid-range climate prediction performance and enhance ocean reanalysis accuracy by optimally blending observational data with short-term deep learning model forecasts. Gain insights into the development of nonlinear observation operators using partial convolution and recurrent generative models for sequential data assimilation.

YOO-GEUN HAM: Deep learning for global climate monitoring and predictions #ICBS2024