Soil is the Earth’s largest terrestrial carbon sink, playing a critical role in regulating the planet’s carbon cycle. How it’s managed could determine whether it mitigates or exacerbates climate change, but the complex interactions between soil, carbon, and climate remain a significant uncertainty in global climate models. To address this, the Department of Energy (DOE) has awarded the University of Arizona (U of A) $610,166 as part of an $8 million initiative to refine climate change models.
Researchers at U of A’s Department of Hydrology and Atmospheric Sciences are combining genomics and artificial intelligence to tackle the challenge of incorporating soil microbes—tiny organisms with outsized effects—into global climate models.
Soil and Climate Feedback: The Missing Link
Understanding soil’s role in climate feedback is critical for improving climate predictions. Microbial communities in the soil significantly influence the release of greenhouse gases like carbon dioxide (CO2) and the availability of nutrients for plants. Yet, due to their microscopic size and complexity, these communities have long been a challenge to study and integrate into models.
“Microbial communities are the main driver controlling greenhouse gas emissions from soil,” assistant professor of hydrology and atmospheric sciences at U of A Yang Song explained.
Song is the project’s lead investigator. Her team will use cutting-edge genomics data to study how environmental changes affect microbial diversity and influence soil carbon-climate feedback. For example, shifts in microbial populations could alter how much carbon dioxide is released into the atmosphere, further influencing climate change.
Leveraging AI for Climate Change Feedback
The researchers aim to integrate biological and environmental data into the DOE’s Energy Exascale Earth System Model (E3SM)—a sophisticated climate model designed to simulate Earth’s climate system with high precision. Song’s team has created AI models capable of mapping microbial communities across the U.S. and assessing how environmental changes impact their functional diversity.
They are now scaling this work globally to refine E3SM’s depiction of soil biogeochemical processes. The team’s AI climate modeling will address critical questions about how soil carbon responds to changing climate conditions, reducing uncertainties in climate risk modeling and improving predictions of future scenarios.
Implications for Global Climate Models
The project’s findings could influence major climate assessments, such as those from the Intergovernmental Panel on Climate Change (IPCC). Enhanced AI climate models could offer better insights into soil carbon dynamics, enabling scientists to predict how carbon stored in soil will affect atmospheric carbon dioxide levels and global temperatures over time. Beyond advancing science, this research could also inform policies and strategies to combat climate change. By understanding soil’s role in climate change feedback, the improved models could help identify actionable ways to manage soil carbon and mitigate climate risks.