Geophysics - Geohazards - AI

We aim to transform our understanding of geohazards by combining big data, advanced seismological tools, and insights from climate and hydrological sciences. Our research not only uncovers the processes that drive earthquakes, landslides, and other hazards, but also reveals how a changing climate reshapes Earth’s subsurface and impacts society and the environment.

Central to our mission is the development of low-compute, massively scalable, fully reproducible research workflows that facilitate seamless collaboration across scientific domains. Our most active group development now lives at Gaia HazLab, with open code and workflows shared through the Denolle Lab GitHub. We are committed to fostering a culture of continuous learning and innovation, empowering our members to embrace modern technologies and AI-aided software practices.

Our newest directions push into agentic scientific research and physics-grounded digital twins:

  • Climate-compounded geohazards — through Gaia HazLab we pursue multi-modal data assimilation and foundation models across modalities to understand the hydromechanical changes in the subsurface driven by fluid migration, and the coupled tectonic–volcanic–hydrological hazards that follow.
  • Agentic pipelines for science — AgentSeismo, the Gaia multi-agent system, and FrugalMind for rigorously evaluating agents against real geoscientific data.
  • Digital twins & surrogate models — built where communities need them most, anchored to ground-truth and derived data products across scales, and grounded in the physics they are meant to emulate.

On software. Open-source science has always been at the core of what we do — our tools and workflows live openly on the Denolle Lab GitHub and Gaia HazLab. What is changing is how we build: increasingly with agents in the loop, and always with an emphasis on personalization and reproducibility — software that adapts to a scientist’s workflow and yields the same answer every time. A good example is CodaMeter (code), where we are transforming operational seismology into reproducible, shareable practice.

Our mantra encapsulates our commitment to:

  • Discovering and characterizing geohazards through big data analysis.
  • Transforming seismology to address pressing societal challenges.
  • Empowering our members through modern AI-aided practices and open-source workflows.

Research Preview

Environment

Environment

Ambient seismic noise reveals soil hydromechanics for water management and geohazard mitigation.

Know more →

Earthquakes

Earthquakes

We detect and characterize earthquakes using novel observations, large-scale data, and physics-based models.

Know more →

Offshore Seismology and Geophysics

Offshore Seismology and Geophysics

We study offshore earthquakes, seafloor processes, and submarine hazards using ocean-bottom seismometers and DAS.

Know more →

Geoscience and AI

Geoscience and AI

Physics-guided AI fuses simulations and sensing to forecast climate and compounding hazard impacts.

Know more →

We are located in Seattle, WA, at the University of Washington, in the heart of the Pacific Northwest (PNW). The PNW is a wonderful natural laboratory to monitor earthquakes, landslides, volcanoes, glaciers, and more broadly speaking the subduction-zone environment.

We are looking for new graduate students, postdocs, and undergraduate researcher (more info) !

0 Citations
0 h-index
0 Publications
0 Team Members
0 GitHub Stars
0 Merged PRs
0 Open Repos

Citations & h-index via OpenAlex; code metrics across Denolle-Lab, gaia-hazlab & mdenolle. Updated 2026-07-15.

Recent papers

All publications →

Latest Updates

June, 2026
Active group cyberinfrastructure development is now centered at Gaia HazLab.
March 19, 2026
New paper out in Science! Shi et al. introduce agroseismology — using fiber-optic DAS to track how farming practices reshape soil hydromechanics.
February, 2026
Cool Paper Alert!
Aug, 2025
Yiyu and Marine featured by eScience
July, 2025
Qibin's recent papers on offshore DAS featured by UW News
Jan, 2025
Congrats Qibin and Ethan on your MuxDAS paper being accepted!

Lab Activity

Latest Authored or mentioned
Marine Denolle LinkedIn activity

LinkedIn blocks unauthenticated feed scraping, so this card links to Marine's activity feed where the newest post she authored or was mentioned in appears first.