Climate informatics: Machine learning for the study of climate change
There is a huge potential for machine learning to help improve climate model accuracy and applicability. From model calibration, to model development to model applications, new approaches are being applied with varying degrees of success. I will discuss where there have been clear successes, and where there are some pitfalls that can be avoided. Overall, I will stress the need for the development of ‘physics-aware’ approaches that can bring fruitfully ML into the development paths of existing models.
Despite the scientific consensus on climate change, drastic uncertainties remain. Crucial questions about regional climate trends, changes in extreme events, such as heat waves and mega-storms, and understanding how climate varied in the distant past, must be answered in order to improve predictions, assess impacts and vulnerability, and inform mitigation and sustainable adaptation strategies. Machine learning can help answer such questions and shed light on climate change. I will give an overview of our climate informatics research, focusing on challenges in learning from spatiotemporal data, along with semi- and unsupervised deep learning approaches to studying rare and extreme events, and precipitation and temperature downscaling.