190-1 OneEarth: Earth Systems Analysis as a Fundamental Prerequisite for Next Generation AI Models
Session: Generative Artificial Intelligence and Other AI-Assistants in Geosciences - Today and in the Near Future
Presenting Author:
Richard ChuchlaAuthors:
Chuchla, Richard1, Xiao, Yitian2, Bouquet, Grant3, Ye, Jieping4, Chen, Hongyang5r> (1) GeoGPT Governance Committee, Hangzhou, China, (2) Deep-time Digital Earth Research Center of Excellence, Hangzhou, China, (3) Zhejiang Lab, Hangzhou, China, (4) Zhejiang Lab, Hangzhou, China, (5) Zhejiang Lab, Hangzhou, China,Abstract:
To truly understand our planet and confront the immense challenges facing society, we need a way of thinking that is both holistic and integrative. Earth system science demands that we consider the interplay of all domains: the lithosphere, asthenosphere, mantle, and core; the atmosphere, hydrosphere, and biosphere, while recognizing that Earth itself is part of a larger cosmic system (solar system and the universe). Yet despite decades of progress, our ability to weave these threads into a coherent fabric of Earth remains limited.
Paradoxically, the challenge grows as our data grows. Nearly 750 Earth‑observing satellites now generate between 100 and 150 terabytes of data every day. This torrent of information is immensely valuable, but without a unifying framework, it risks overwhelming rather than enlightening us.
Our models remain siloed and limited to only a small part of the very highly dimensional nature of earth science. Earth is not a set of compartments. Likewise, A 4D Digital Earth platform is not a climate model, a basin model, a geodynamic model or an AI chatbot with maps. It is a unified, evolving, data‑assimilating representation of the earth system across space and time, built on a first principles understanding of physics, chemistry and biology. It has spatial intelligence and is orchestrated and driven by AI. It is a single, coupled, multi-scaled dynamic system. This is the scientific frontier and this is the goal of OneEarth.
OneEarth adopts an AI-native architecture in which artificial intelligence functions as a core orchestration, integration, and reasoning layer throughout the platform. AI is intended to not only process data, but to also connect observations, simulations, scientific literature, expert knowledge, and evolving earth system models within a continuously interacting scientific platform. Among the grand challenges that we can tackle in ways we could not before include climate change and strategies for climate resilience, resourcing future generations, water scarcity and soil erosion.
OneEarth will be open sourced to encourage equitable access to data and tools and to foster co-creation and improvement through iterative user feedback. Important building blocks like the 021 Large Science Model will be released as they are developed.
Geological Society of America Abstracts with Programs. Vol. 58, No. 5, 2026
doi: 10.1130/abs/2026AM-16461
© Copyright 2026 The Geological Society of America (GSA), all rights reserved.
OneEarth: Earth Systems Analysis as a Fundamental Prerequisite for Next Generation AI Models
Category
Topical Sessions
Description
Session Format: Oral
Presentation Date: 10/13/2026
Presentation Start Time: 08:00 AM
Presentation Room: CCC, Bluebird Ballroom 3E
Back to Session