Entering a new era: climate science in the time of big data and machine learning
Over the last two decades, climate science has rapidly evolved into a big data research field. This development is mainly the result of ever increasing computational power, which has led to a swift expansion in the number of computer modeling projects, and parallel advances in measurement systems particularly on satellites, leading to a massive increase in data acquisition.
Here, based on several examples from my research group at Imperial College, I highlight how typical data science methods have the potential to revolutionize scientific disciplines that rely on these increasingly large datasets. I show how machine learning parameterizations of Earth system processes could improve and speed up global climate models (the fluid-dynamical, thermodynamic computer models used to make climate change projections). In addition, I introduce causal discovery algorithms as sophisticated tools for climate model evaluation and explain how unsupervised learning algorithms can help us detect and understand drivers of changes in extreme weather events such as heat waves. I further touch on how existing large datasets can be used to train fast climate model emulators and applied to machine learning applications in ocean science, air pollution modeling and seasonal weather forecasting.
Biography: Peer Nowack is a computational physicist who combines state-of-the-art numerical models, Earth observations and machine learning to address key challenges in climate science. His main research interests include the development and application of novel computational methods to understand and reduce uncertainty in regional climate change projections. More generally, he is interested in data science questions related to the Earth system, for instance in relation to seasonal weather forecasting or the modeling of global air pollution and its effect on human health and climate. Since August 2017, Peer has been an independent Research Fellow at the Department of Physics, the Grantham Institute for Climate Change and the Data Science Institute of Imperial College London, UK. Previously, he worked as a postdoctoral researcher and PhD student at the University of Cambridge, UK, after completing his undergraduate degree at ETH Zurich in Switzerland. His research received several awards from national and international institutions, including the American Geophysical Union, the University of Cambridge and the UK’s National Centre for Atmospheric Science.
This is a joint seminar with the Department of Computer Science and the School of the Environment.