Artificial Intelligence (AI)
Image courtesy of (left) Giri Prakash & Shawn Serbin (right) Maria Zawadowicz
(Left) Modular Data Ecosystem to enable data interoperability for AI. (Right) Self-organizing maps (SOMs) used to classify aerosol measurements during the ARM TRACER field campaign. M visualization of a selection of components: organic mass, NO3 mass, hydrocarbon-line organic aerosol mass, refractory black carbon equivalent, wind speed, CO mixing ratio.
The use of AI techniques has been piloted across Biological and Environmental Research (BER) research. The AI for Earth system predictability (AI4ESP) workshop validated the scientific, academic, and commercial community interest and need for critical advancements in infrastructure, opportunity, and direction to leverage AI for BER research. Support for known shifts in how we approach co-design and cross-domain science are evident in the Earth and Environmental Systems Sciences Division (EESSD) and Advanced Scientific Computing Research (ASCR) AI4ESP workshop as well as upcoming opportunities. BER is dedicated to harnessing DOE’s computational resources to exploit ‘big data’ across many Earth system domains while maintaining FAIR and equitable access.
Press Releases and Award Lists
- Predicting the future of the Earth with artificial intelligence (October 2021)
- Atmospheric Research Turns to the Power of Computer Programs that Learn from Data (Jan 2021)
- DOE Announces Call for White Papers to Advance an Integrative Artificial Intelligence Framework for Earth System Predictability (December 2020)
Key Projects:
- From Clouds to Precipitation: Multiscale Dynamics Microphysics Interactions in Cumulus Clouds
- Understanding spatial organization during precipitation-induced convective cloud transitions
- Multiscale Aerosol Modeling Across Space and Composition
- Interplay of Gas-phase Reactions and Multi-phase Processes on Phase State and Growth Dynamics of Secondary Organic Aerosols
- Constraining Microphysical Processes of Warm Rain Formation Using Advanced Spectral Separations, an Ensemble Retrieval Framework and Machine Learning Techniques
- Advanced Precipitation and Boundary Layer Data Products Derived from ARM Radar Wind Profilers
- Classification of Cloud Particle Imagery and Thermodynamics (COCPIT): A New Databasing Tool for the Characterization of Cloud Particle Images Captured During DOE Field Campaigns
- Detection and Characteristics of Blowing Snow at ARM Sites
- Improving Parameterization of Ice Microphysical Processes in Arctic Clouds Using a Synergistic Modeling and Observational Approach
BER Workshops
- AI for Earth System Predictability Workshop (October-December 2021)
- 2020 ASR/ARM Topical Workshop on Machine Learning and Statistical Methods for Observations, Modeling, and Observational Constraints on Modeling. (October 2020)
Contacts
Justin Hnilo
Program Manager
Data Management
Justin.hnilo@science.doe.gov