Collaborative research: frameworks: developing robust cyberinfrastructure to advance earth system model hierarchies for scientific discoveries and educational outreach

The National Science Foundation Office of Integrative Activities awarded the University Corporation for Atmospheric Research (National Center for Atmospheric Research) $1,515,241 on August 1, 2026, under the Geosciences program to develop robust cyberinfrastructure for the Community Earth System Model hierarchy.

The project modernizes the Community Atmosphere Model (CAM) by establishing an extended, sustainable, and user-configurable hierarchy of idealized model configurations. Work focuses on migrating CAM's existing dynamical cores into the new CAM-SIMA (System for Integrated Modeling of the Atmosphere) GitHub framework using a common community physics package protocol. The effort lowers technical and financial barriers to participation from resource-limited institutions by allowing scientists to strip away non-essential variables and systematically isolate physical mechanisms within comprehensive Earth system models.

The project strengthens the technical workforce by incorporating cutting-edge AI models into undergraduate and graduate classrooms, hosting a Dynamical Core Model Intercomparison Project summer school, and creating free online training materials. Deliverables include transparent, accessible modeling tools that enhance scientific and AI literacy and improve society's readiness for weather-related decision-making.

Performance occurs in Boulder, Colorado. The award runs through July 31, 2031.

Abstract

This project modernizes and expands the computing tools used to simulate and predict the earth?S system. While comprehensive earth system models are essential for projecting complex interactions between the atmosphere, oceans, land, and ICE, their overwhelming complexity often makes IT difficult to isolate fundamental cause-and-effect relationships or verify underlying software flaws. This project addresses these challenges by developing a robust, user-friendly hierarchy of simpler models within the community earth system model (CESM), allowing scientists to strip away non-essential variables and systematically isolate physical mechanisms. This lowering of technical and financial barriers actively expands participation from resource-limited institutions. Furthermore, the project directly strengthens the technical workforce by incorporating cutting-edge AI models, embedding these tools into undergraduate and graduate classrooms, hosting a dynamical core model intercomparison project (DCMIP) summer school, and creating free online training materials to prepare the next generation of atmospheric scientists. Ultimately, these transparent and accessible modeling tools enhance scientific and AI literacy and improve society's long-term readiness for weather-related decision-making. The technical objective of this project is to enhance the cyberinfrastructure of CESM's community atmosphere model (cam) by establishing an extended, sustainable, and user-configurable hierarchy of idealized model configurations. A critical focus of the work is migrating cam's existing dynamical cores with very different numerical approaches into the new cam-sima (system for integrated modeling of the atmosphere) github framework using the common community physics package protocol. To eliminate rigid configuration barriers, the project implements inlined analytic initial conditions that dynamically balance across varying computational grids and unstructured meshes. Furthermore, the project introduces a pioneer testing hierarchy for AI-based global atmospheric emulators. Among them are the credit and camulator frameworks, developed by the NSF national center for atmospheric research, and the latest AI emulator innovations from the global community to rigorously evaluate their physical consistency, conservation properties, and performance against traditional fluid dynamics solvers. The project integrates its technical innovations into an extended configuration of the CESM?S graphical user interface workflow tool. Code deliverables, standard configuration templates, and quality-controlled DCMIP simulation databases are managed via public github repositories to ensure community-wide reusability and regression tracking. This award by the office of advanced cyberinfrastructure is fully supported by the directorate for geosciences. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the foundation's intellectual merit and broader impacts review criteria.- subawards are not planned for this award.