This event is part of Confronting Global Climate Change View Details

Machine Learning for Climate and Weather Applications

Description

Back to top

The Earth’s climate system is a classical example of a multiscale, multiphysics dynamical system with an extremely large number of active degrees of freedom, exhibiting variability on scales ranging from micrometers and seconds to thousands of kilometers and centuries.  Machine learning approaches present a timely opportunity to leverage the information content of large datasets generated by observational systems and models to improve scientific understanding and prediction capability of weather and climate dynamics. The workshop will bring together an interdisciplinary group of researchers in applied mathematics, climate science, and data science to discuss recent advances and future perspectives on machine learning for weather and climate applications, including feature extraction, subgrid-scale modeling, and statistical prediction. 

This workshop will include a poster session on Wednesday, November 2. In order to propose a poster, you must first register for the workshop, and then submit a poster proposal using the form that will become available on this page after you register.  The registration form should not be used to propose a poster. The deadline for proposing a poster is October 25.

Organizers

Back to top
D G
Dimitris Giannakis Dartmouth University
V H
Vera Hur University of Illinois at Urbana-Champaign
J W
Jonathan Weare New York University

Speakers

Back to top
D A
Dorian Abbot University of Chicago
E B
Elizabeth Barnes Colorado State University
T B
Tom Beucler University of Lausanne
F B
Freddy Bouchet Centre National de la Recherche Scientifique (CNRS)
N B
Noah Brenowitz NVIDIA
N C
Nan Chen University of Wisconsin-Madison
O D
Oliver Dunbar Caltech
J F
Justin Finkel MIT
G F
Gary Froyland University of New South Wales
A G
Auroop Ganguly Northeastern University
P G
Pierre Gentine Columbia University
I G
Ian Grooms University of Colorado
P H
Pedram Hassanzadeh Rice University
K K
Karthik Kashinath NVIDIA and Lawrence Berkeley National Laboratory
V L
Valerio Lucarini University of Reading
P M
Peetak Mitra Xerox Palo Alto Research Center
D Q
Di Qi Purdue University
T S
Themistoklis Sapsis MIT
A S
Aditi Sheshadri Stanford University
M S
Maike Sonnewald Princeton University
M Y
Minah Yang New York University
J Y
Janni Yuval MIT

Poster Session

Back to top

The posters that have been submitted for the poster session are available on the poster session page.

Schedule

Back to top
Monday, October 31, 2022
9:00-9:45 CDT
Combining Stochastic Parameterized Reduced Order Models with Machine Learning for Data Assimilation and Uncertainty Quantification with Partial Observations

Speaker: Nan Chen (University of Wisconsin, Madison)

10:30-11:15 CDT
Explainable AI (XAI) for Climate Science: Detection, Prediction and Discovery

Speaker: Elizabeth Barnes (Colorado State University, Fort Collins)

11:30-13:00 CDT
Lunch
13:00-13:45 CDT
Artificial intelligence with uncertainty quantification can plug gaps in climate science and inform multi sector resilience

Speaker: Auroop Ganguly (Northeastern University)

14:15-15:00 CDT
Using autoencoders as generative models to create forecast ensembles for data assimilation

Speaker: Ian Grooms (University of Colorado, Boulder)

15:30-16:15 CDT
Accelerated Parametric Uncertainty Quantification and Optimal Data Acquisition in an Idealized Global Atmosphere Model

Speaker: Oliver Dunbar (California Institute of Technology)

Tuesday, November 1, 2022
9:00-9:45 CDT
Metastability of the Climate System

Speaker: Valerio Lucarini (University of Reading)

10:30-11:15 CDT
Smashing Cities and Planets with Rare Event Sampling and AI

Speaker: Dorian Abbot (University of Chicago)

11:30-13:00 CDT
Lunch
13:00-13:45 CDT
Physics to Machine learning and machine leaning back to physics

Speaker: Pierre Gentine (Columbia University)

14:15-15:00 CDT
Modeling our future: Advancing climate research using AI and planning intervention scenarios

Speaker: Peetak Mitra (Xerox Palo Alto Research Center)

Wednesday, November 2, 2022
9:00-9:45 CDT
Probabilistic forecast of extreme heat waves using convolutional neural networks and rare event simulations

Speaker: Freddy Bouchet (CNRS and Ecole Normale Supérieure)

10:30-11:15 CDT
Sampling Strategies for Training Machine Learning Emulators of Gravity Wave

Speaker: Minah Yang (New York University)

11:30-13:00 CDT
Lunch
13:00-13:45 CDT
Building Digital Twins of the Earth for NVIDIA’s Earth-2 Initiative

Speaker: Karthik Kashinath (NVIDIA and Lawrence Berkeley National Laboratory)

14:15-15:00 CDT
Extracting climate cycles from spatiotemporal data and detecting emergence and disappearance of coherent phenomena across multiple dynamic regimes

Speaker: Gary Froyland (University of New South Wales)

15:30-16:30 CDT
Poster Session/Social Hour
Thursday, November 3, 2022
9:00-9:45 CDT
Improving Climate Models with Machine Learning

Speaker: Noah Brenowitz (NVIDIA)

10:30-11:15 CDT
Revealing the statistics of extreme events hidden in short weather forecast data

Speaker: Justin Finkel (University of Chicago)

11:30-13:00 CDT
Lunch
13:00-13:45 CDT
Neural-network parameterization of subgrid momentum transport learned from a high-resolution simulation

Speaker: Janni Yuval (Massachusetts Institute of Technology (MIT))

14:15-15:00 CDT
High-resolution climate modeling using coarse-scale models and reanalysis data

Speaker: Themistoklis Sapsis (MIT)

15:30-16:15 CDT
Elucidating drivers of Southern Ocean circulation change: A blueprint for interpretable and explainable machine learning

Speaker: Maike Sonnewald (Princeton University)

Friday, November 4, 2022
9:00-9:45 CDT
Systematically Generating Hierarchies of Machine-Learning Models, from Equation Discovery to Deep Neural Networks

Speaker: Tom Beucler (University of Lausanne)

10:15-11:00 CDT
Integrating the spectral analysis of neural networks and nonlinear physics for explainability, generalizability, and stability

Speaker: Pedram Hassanzadeh (Rice University)

11:30-12:30 CDT
Statistical reduced-order models and data-driven closure strategies for turbulent systems

Speaker: Di Qi (Purdue University)


Videos

Back to top

Combining Stochastic Parameterized Reduced Order Models with Machine Learning for Data Assimilation and Uncertainty Quantification with Partial Observations

Nan Chen
October 31, 2022

Explainable AI (XAI) for Climate Science: Detection, Prediction and Discovery

Elizabeth Barnes
October 31, 2022

Artificial intelligence with uncertainty quantification can plug gaps in climate science and inform multi sector resilience

Auroop Ganguly
October 31, 2022

Using autoencoders as generative models to create forecast ensembles for data assimilation

Ian Grooms
October 31, 2022

Accelerated Parametric Uncertainty Quantification and Optimal Data Acquisition in an Idealized Global Atmosphere Model

Oliver Dunbar
October 31, 2022

Metastability of the Climate System

Valerio Lucarini
November 1, 2022

Smashing Cities and Planets with Rare Event Sampling and AI

Dorian Abbot
November 1, 2022

Physics to Machine learning and machine leaning back to physics

Pierre Gentine
November 1, 2022

Using data-informed methods towards an improved understanding and representation of atmospheric gravity waves

Aditi Sheshadri
November 1, 2022

Probabilistic forecast of extreme heat waves using convolutional neural networks and rare event simulations

Freddy Bouchet
November 2, 2022

Sampling Strategies for Training Machine Learning Emulators of Gravity Wave

Minah Yang
November 2, 2022

Building Digital Twins of the Earth for NVIDIA’s Earth-2 Initiative

Karthik Kashinath
November 2, 2022

Extracting climate cycles from spatiotemporal data and detecting emergence and disappearance of coherent phenomena across multiple dynamic regimes

Gary Froyland
November 2, 2022

Revealing the statistics of extreme events hidden in short weather forecast data

Justin Finkel
November 3, 2022

Neural-network parameterization of subgrid momentum transport learned from a high-resolution simulation

Janni Yuval
November 3, 2022

High-resolution climate modeling using coarse-scale models and reanalysis data

Themistoklis Sapsis
November 3, 2022

Systematically Generating Hierarchies of Machine-Learning Models, from Equation Discovery to Deep Neural Networks

Tom Beucler
November 4, 2022

Integrating the spectral analysis of neural networks and nonlinear physics for explainability, generalizability, and stability

Pedram Hassanzadeh
November 4, 2022

Statistical reduced-order models and data-driven closure strategies for turbulent systems

Di Qi
November 4, 2022