[DLS] Tim Merlis (Princeton)

“New Approaches to Global Atmospheric Modeling”

The atmospheric component of global climate models has followed the same structural approach for more than three decades. In the last few years, new approaches have emerged that warrant thorough assessment. One approach is to simulate previously unresolved horizontal scales, of a few kilometers, that are important for cumulus convection. These kilometer-scale models avoid an aspect of the parameterization problem: unresolved flows affect the larger scales of interest. Alternatively, machine learning approaches have been introduced that replace the whole atmospheric model with neural networks. I will present the atmospheric response to ocean warming in these new approaches to global atmospheric modeling, focusing on concordance with and discrepancies from well-understood physical changes in warmed atmospheres.

EAPS Department Lecture Series —
Weekly talks aimed to bring together the entire EAPS community, given by leading thinkers in the areas of geology, geophysics, geobiology, geochemistry, atmospheric science, oceanography, climatology, and planetary science. Runs concurrently with class 12.S501.

Contact: eapsinfo@mit.edu