Email Print Share
NSF 23-046

Dear Colleague Letter: Advancing Research in the Geosciences Using Artificial Intelligence (AI) and Machine Learning (ML)

January 20, 2023

Dear Colleagues:

The National Science Foundation’s (NSF) Directorate for Geosciences (GEO) encourages the submission of proposals that advance our understanding of geosciences using Artificial Intelligence (AI) and Machine Learning (ML) methods.

To tackle grand challenge problems across the geosciences, researchers increasingly are turning to AI/ML methods. AI includes any computational tool that mimics human intelligence, including using logic and decision trees. ML methods use statistical techniques to enable machines to improve at tasks with experience and include neural networks that permit software to train itself to perform tasks after exposure to vast amounts of data. These are powerful tools for analyzing large and complex datasets, developing physical models, expediting computation or scaling between models, or designing and deploying sensor networks. Unique aspects of understanding Earth systems and using geosciences data can also inform and inspire new developments in AI/ML, and AI-enabled research will require a workforce prepared to understand, use, and develop appropriate AI/ML techniques.

To promote research that benefits from AI/ML and reduces barriers to its use in the geosciences, GEO welcomes proposals incorporating AI/ML methods across its broad range of programs. The geosciences collectively refers to the research supported in the Divisions of Atmosphere and Geospace (AGS), Earth (EAR), and Ocean (OCE) Sciences, and in the Office of Polar Programs (OPP). Proposals in response to this DCL must advance core geosciences program goals and use AI/ML methods toward addressing scientific problems.

Areas where AI/ML methods may be used include, but are not limited to, implementing existing AI/ML methods to address geosciences problems, developing new algorithms to build geosciences insights, and/or engaging AI/ML to explore or emulate physically based models. AI/ML methods may be posed in conjunction with other geosciences and analysis methods to address a fundamental geoscience question. Proposals may also include comparison or validation of the outputs of AI/ML techniques against outputs from other, traditional analytical methods or theoretical and experimental approaches. Activities, such as making AI/ML training datasets, software and tools openly available to the scientific community and developing a workforce trained in AI/ML techniques, may be appropriate Broader Impacts of proposals in response to this call.


This is not a special competition or new program. Relevant proposals should be submitted to an existing GEO program, according to that program's submission guidelines. Before submission, PIs should contact cognizant program directors in the program(s) within AGS, EAR, OCE, or OPP that are most relevant to their projects to discuss the appropriate mechanism for submission.

Proposals will be evaluated by the core programs, alongside other proposals submitted to those programs. Therefore, proposals should first and foremost focus on important scientific questions in the discipline of interest. Proposals must also describe the AI/ML methods and justify how the methods address a scientific challenge or question that was previously intractable. Proposers will need appropriate expertise in AI/ML methods, which can be demonstrated through previous experience with proposed methods, collaboration with relevant data science experts, and/or pathways for training students and other researchers in AI/ML.

When making investments, NSF seeks broad representation of PIs and institutions in its award portfolio, including a geographically diverse set of institutions (including those in EPSCoR jurisdictions) and PIs who are women, early-career researchers, members of underrepresented minorities, veterans, and persons with disabilities. Submissions which benefit and involve the full breadth of the geoscience research community, including undergraduates, graduate students, cyberinfrastructure professionals, and faculty at two-year and four-year institutions of higher education, including minority serving institutions and non-R1 institutions, are encouraged.

General questions about this Dear Colleague Letter may be submitted to


Alexandra R. Isern
Assistant Director for Geosciences