Skip to main content
U.S. flag

An official website of the United States government

Publication Date
9 April 2024

Decoding the Dynamics of Atmospheric River Clusters

Discover the characteristics of back-to-back atmospheric rivers and the risks they pose in a changing climate.
Print / PDF
Powerpoint Slide
Image Credit

Yang Zhou, Lawrence Berkeley National Laboratory. 


Atmospheric rivers are like rivers in the sky that carry vast amounts of water vapor. When these rivers meet land, they often result in heavy rains and flooding. Recently, researchers used a machine learning tool to study these phenomena, particularly focusing on groups or 'clusters' of these rivers. They found that clusters with densely packed atmospheric rivers are more likely to lead to more severe weather events. This study also discusses how these clusters will change with the warming climate.


Understanding atmospheric river clusters is crucial for preparing for and mitigating future climate risks. This research helps predict when and where such cluster of weather extremes might hit, allowing for better flood defenses and emergency responses. As our climate changes, this cluster may occur more often and become even more damaging. Therefore, insights from this study become essential to protect communities and infrastructure, especially in regions prone to extreme weather.


This study utilized advanced machine learning algorithms to analyze the characteristics and impact of atmospheric river clusters—groups of intensive moisture transport systems that can significantly increase flood risk. By examining these clusters, the research revealed that denser clusters are associated with more frequent high-category atmospheric river events, leading to heightened risks of extreme precipitation and harsh land responses. The study connects cluster activity to large-scale atmosphere circulations and identifies a large-scale pattern that is favorable for clusters to occur. The study demonstrates that the atmospheric river clusters will be more frequent and impactful in the future. Such findings are vital as they enhance our ability to predict such weather extremes and to better prepare for potential natural disasters, contributing to improved resilience strategies.

Point of Contact
Yang Zhou
Lawrence Berkeley National Laboratory
Funding Program Area(s)
Additional Resources:
NERSC (National Energy Research Scientific Computing Center)