Abstract:

Climate projections continue to be marred by large uncertainties, which originate in processes that need to be parameterized in models, such as clouds, turbulence, and ecosystems. But rapid progress is now within reach. New tools from data assimilation and machine learning make it possible to integrate global observations and local high-resolution simulations in an Earth system model (ESM) that systematically learns from both. This talk proposes a blueprint for such an ESM and outlines the research program needed to realize it. It discusses how ESMs can learn from global observations and targeted high-resolution simulations, and what computational and mathematical challenges have to be confronted to develop learning algorithms that are suitable for ESMs, given their large computational expense. While challenges remain to realize it, the proposed framework offers an opportunity for dramatic improvements in the accuracy of ESMs.

Bio:

Tapio Schneider is the Theodore Y. Wu Professor of Environmental Science and Engineering at Caltech. He researches how fundamental aspects of how the climate of Earth and other planets come about and may change. He was named one of the "Top 20 Scientists under 40” by Discover Magazine, is a recipient of a David and Lucile Packard Fellowship in Science and Engineering, of an Alfred P. Sloan Research Fellowship, and of the inaugural James R. Holton Award of the American Geophysical Union. He received his PhD in 2001 from Princeton University.