J12A Artificial Intelligence for Actionable Insights and Applications in Climate Science

Wednesday, 31 January 2024: 4:30 PM-6:00 PM
345/346 (The Baltimore Convention Center)
Hosts: (Joint between the 23rd Conference on Artificial Intelligence for Environmental Science; the 37th Conference on Climate Variability and Change; and the Presidential Conference )
Submitters:
Marybeth Arcodia, University of Miami RSMAS, Atmospheric Science, Miami, FL; Eleanor A. Middlemas, PhD and Zane Martin, Colorado State University, Atmospheric Science, Fort Collins, CO
Cochairs:
Marybeth Arcodia, University of Miami RSMAS, Atmospheric Science, Miami, FL; Eleanor A. Middlemas, PhD; Zane Martin and Katie Dagon

Artificial intelligence (AI) applied to the earth sciences has recently been a rapidly expanding field in both academic and industry spaces, due in part to its ability to extract nonlinear relationships from noisy data. The use of AI, particularly machine learning, can lead to identification of predictable signals from purely data-driven methods. Further, explainable AI techniques allow for opening of the AI “black box” to understand the model’s decision-making strategy. AI-driven advancements in climate predictability on subseasonal through multidecadal timescales, today and in a changing climate, allows for increased predictive skill and lead time, which can improve preparation. The use of AI in climate data analysis has cultivated actionable insights for the purpose of both scientific discovery and for managing climate risk.

We invite abstracts that discuss the use of AI for actionable insights, including high resolution climate forecasting and informing of adaptation and mitigation strategies, as well as using AI to isolate at-risk areas exposed to various climate impacts. This session also welcomes AI approaches applied to climate models and observational data that can be used by decision-makers and stakeholders for planning purposes. Equal consideration will be given to reproducible novel AI techniques, explainable and interpretable AI methods for exploring the climate system, sources of predictability on subseasonal-to-multidecadal timescales, and forecasts of opportunity.

Papers:
4:30 PM
J12A.1
4:45 PM
J12A.2
A Data-Driven Approach to Identifying Key Regions of Change Associated with Future Climate Scenarios
Zachary Michael Labe, PhD, Princeton University, Princeton, NJ; and T. L. Delworth, N. Johnson, and W. Cooke

Handout (10.5 MB)

5:00 PM
J12A.3
Understanding Predictability of Extreme SEUS Precipitation Using Explainable Machine Learning
Kathleen Pegion, University of Oklahoma, Norman, OK; University of Oklahoma, Norman, OK; and E. J. Becker and B. Kirtman

5:15 PM
J12A.4
Physics-Informed Prediction of Arctic Sea Ice Drift
Heather Hunter, The Johns Hopkins University Applied Physics Laboratory, Laurel, MD; and M. R. Keller, P. Sicurello, and C. Piatko

5:30 PM
J12A.5
Assessing Global Human Impact and Climate Hazards with Landsat Data and Convolutional Neural Networks
Bryam Orihuela-Pinto, Colorado State University, Fort Collins, CO; and P. Keys, E. A. Barnes, and F. V. Davenport

5:45 PM
J12A.6
A Data-Driven Model of Urban Carbon Emissions at the Human Scale
Constantine Kontokosta, New York University, BROOKLYN, NY; and B. Bonczak

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner