4. Advanced Scientific Computing Research (ASCR) Program’s Leadership Computing Facilities These facilities provide peta- & now exascale computing capabilities enabling detailed computational modeling efforts. For BER researchers in particular, these efforts include climate science modeling such as the Energy Exascale Earth System Model (E3SM)e3sm.org/ & smaller modeling efforts in biological, Earth, & environmental science. More recent AI efforts are similarly supported by ASCR facilities. science.osti.gov/User-Facili… Energy Exascale Earth System Model (E3SM) Documentation E3SM is a state-of-the-art fully coupled model of the Earth's including important biogeochemical & cryospheric processes. docs.e3sm.org/E3SM/ e3sm.atlassian.net/wiki/spac… Docs for the tools & models developed by the E3SM Project docs.e3sm.org/ E3SM User Guide E3SM is not just one earth system model but a modeling system that allows many different configurations of atmosphere, ocean, land & other components w/ both full model & data model options. Also, the configurations of model components can run at different resolutions. The configuration options are managed by the Case Control System (CCS) w/in the Community Infrastructure for Modeling the Earth (CIME). docs.e3sm.org/E3SM/user-guid… The Community Infrastructure for Modeling the Earth (CIME) github.com/ESMCI/cime The Common Infrastructure for Modeling the Earth (CIME - pronounced “SEAM”) provides 2 core features; a Case Control System (CCS) for configuring, compiling & executing Earth System Models, & a framework for system testing an Earth System Model. CIME is developed by the E3SM & CESM cesm.ucar.edu/ projects. Case Control System (CCS) There are 3 components that make the Case Control System. 1. XML files that describe the models, model configuration (components, grids, input-data, etc) & machines. 2. Python module that provides tools for users to create cases, configure their model, build & submit jobs. 3. Addition stand-alone tools useful for Earth System Modeling. esmci.github.io/cime/version… Earth System Model Computational Infrastructure Repo github.com/ESMCI E3SM AI for Model Emulation – A Pilot Study The E3SM Project plans to release AI-enabled emulators alongside each physics-based model, including both standalone component models & a fully coupled system tailored to E3SM’s scientific needs & software requirements. A new E3SM AI group will lead this pilot. Their first priority is to build a robust software framework to integrate AI-enabled emulators into the E3SM code base, followed by iterative testing & improvement. The goal: a seamless, end-to-end AI workflow that meets E3SM’s software & science requirements. Ongoing AI efforts w/in other groups will continue under their existing leadership, w/ the new AI group coordinating & tracking progress across the entire project. This work builds on E3SM’s collaboration w/ the Allen Institute for AI (AI2) & its ACE emulator,allenai.org/blog/ai2-climate… which models the global atmosphere using large autoregressive neural networks w/ a Spherical Fourier Neural Operator (SFNO) backbone. ACE incorporates physical constraints directly into its architecture, improving conservation & stability during long simulations. It will be coupled w/ Samudra,agupubs.onlinelibrary.wiley.… a AI global ocean emulator, forming the ACE–Samudra system — the starting point for this new AI initiative, which will be further customized to meet E3SM’s specific science goals. e3sm.org/e3sm-ai-for-model-e… Accelerating Earth System Modeling: AI-based land model spinup The researchers have developed a modular deep‑learning framework to solve the spinup problem (Gao, et al., 2025).arxiv.org/html/2509.23453v1 The solution combines data‑type–aware encoders for heterogeneous inputs w/ multi‑level physics‑based constraints that promote consistency from local dynamics to global system behavior. e3sm.org/accelerating-earth-…