International Summit to Consider a Special-Purpose Computational System for Frontier Earth System Science and Climate Projection; Urbana, Illinois; Fall-2024

The scientific objectives of this project are to convene an international summit to discuss the practicability of creating a special-purpose computational system for frontier Earth system science and climate projections at kilometer-scale global resolution.

Earth System Models (ESMs) run at high resolution may help to address sources of model bias, which in some cases is significant, the pace of global warming, local and regional impacts (e.g., ecosystems, health, agriculture, inundation), and details of clouds and precipitation including the behavior of high-impact weather (severe storms, hurricanes, flash floods). Although current computers are extraordinarily powerful and theoretically able to perform a quintillion operations per second (exascale), ESMs currently may obtain only a few percent of this peak in practice.

Therefore, various types of downscaling from coarser resolution runs are required to provide information about local impacts (information which would be improved with higher resolution), and biases which cannot be addressed without higher resolution remain unresolved. Current projections suggest global ESM resolutions of 1 km will not be attainable until around 2055, which is far beyond the time needed for critical decisions to be made regarding climate mitigation and adaptation.

Consequently, it is prudent to examine the possibility of creating a special-purpose computational system for frontier Earth system science and climate projection for use by the global community. An effective first step forward in such an endeavor is holding an international summit at the University of Illinois Urbana-Champaign (UIUC) in fall 2024. The computing system would need to accommodate both ESMs and new approaches that couple or tether such models to artificial intelligence (AI).

The summit will assemble international leaders in relevant disciplines from academia, industry, government, and non-profit organizations to examine several key questions relating to ESMs and the role of AI in and around them going forward; possible structures of a special-purpose computational system and alternative pathways; strategies for creating such a system; and expected benefits.

The potential Broader Impacts (B.I.) include an opportunity to engage a diverse, international audience and provide an opportunity for practicing researchers, as well as students and post-doctoral scholars across all types and sizes of institutions, to learn not only about a special-purpose computational system for Earth system science and climate projection, but also

about the value of using fine-scale model output to inform society more broadly regarding local and regional climate change. Topics to be discussed include economic and social impacts of climate change; risk communication and management; health and epidemiology; international relations; population dynamics;infrastructure; coastal vulnerability; and national security and defense.

The summit will be free, streamed live on the Internet, and recorded for subsequent viewing. A final report containing summaries of discussions and a set of recommendations will be prepared and made freely available on the summit website.

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.