CCSS Meeting #65: Neural General Circulation Model - differentiable atmospheric model for weather and climate predictions

This lecture will be held in hybrid format: the speaker Dr. Janni Yuval will be speaking online from the US, and one of the CCSS associate members will be physically present as the on-site moderator to open the session and initiate discussions. The theme of this CCSS Dinner Meeting is Neural Differential Equations and the applications in Complex Systems.

In the CCSS living room, participants can enjoy refreshments and dinner afterwards - please signup for free below.

Speaker Overview

Dr. Janni Yuval is a visiting scientist at Google Research working on the development of hybrid climate and weather models. He relies on a combination of machine learning and climate physics to enhance the accuracy of these models. Janni completed his postdoctoral research as a Lorenz postdoc fellow at MIT, where he worked with Paul O'Gorman on hybrid atmospheric models that also incorporated machine learning and climate physics. His academic background includes a PhD in Atmospheric Dynamics from the Weizmann Institute of Science, where he was mentored by Yohai Kaspi, as well as an M.Sc in Physics under the guidance of Samuel Safran. 

Lecture Overview

Recent advancements in machine learning (ML) have led to various data-driven approaches aimed at enhancing weather prediction and climate modeling. Typically, an ML-only method is used to improve weather prediction. While current state-of-the-art ML approaches achieve lower errors at medium-range lead times, physics-based models like ECMWF’s HRES/ENS demonstrate superior physical consistency and better forecast accuracy at longer lead times. In the realm of climate modeling, ML is often used in a hybrid approach, where ML components replace uncertain parameterizations while still adhering to the governing equations for the scales they resolve. Despite advancements in atmospheric hybrid models, current attempts face challenges such as instability over extended periods, limited applicability to idealized scenarios, and only modest improvements on longer time scales in realistic scenarios.

In this talk, I will explore the effectiveness of a novel approach to atmospheric modeling suitable for both weather forecasting and climate modeling. This approach involves the development of a differentiable atmospheric model, named NeuralGCM, which for the first time combines ML techniques with governing equations in an end-to-end training using ERA5 data. NeuralGCM competes with ML models for 1-10 day forecasts and matches the European Centre for Medium-Range Weather Forecasts ensemble prediction for 1-15 day forecasts. Over longer time scales, NeuralGCM displays emergent phenomena such as seasonal cycles, monsoon patterns, and tropical cyclone formation, while achieving comparable spatial bias to a global cloud-resolving model. Additionally, we present the first successful simulation of an AMIP-like experiment using a hybrid atmospheric model.

There will be 45-min lecture from the speaker, followed by a 15-min Question & Answer session. After this there will be time for networking & a pasta buffet dinner will be provided!

To attend the lecture online, please click the Zoom link at 16:00 on Tuesday May 28.

To attend the lecture & the dinner (physically), please signup below before 15:00 on Monday May 27.