Skip to article frontmatterSkip to article content
Site not loading correctly?

This may be due to an incorrect BASE_URL configuration. See the MyST Documentation for reference.

Problem Statement

The Challenge

Landslides triggered by atmospheric rivers, liquefaction amplified by saturated soils, catastrophic runoff after wildfires, severe convective storms intensified by land–atmosphere feedbacks — the most devastating geodisasters arise from cascading interactions across the ocean, atmosphere, and solid Earth. The severity of these hazards — landslides, flash floods, earthquake liquefaction, and convective storms — is profoundly shaped by the soil hydromechanical history and by land management practices that alter the critical zone over time.

Yet our current models struggle to keep pace with these nonlinear cascades. Atmosphere, hydrology, and geomechanics are typically studied and modeled in isolation, leaving critical couplings unresolved. As climate change intensifies extreme weather and shifts precipitation patterns, the gap between what we can predict and what communities need to prepare for continues to grow.


Our Approach

We take a data-driven and physics-grounded approach to monitor, characterize, and predict the susceptibility of climate-compounded geodisasters — both in real time and under future weather and climate scenarios.

We leverage:

These capabilities serve three interlinked research goals:

  1. Discovery of missing physics — Identify the governing processes and couplings (e.g., soil memory effects, ocean–atmosphere teleconnections) that current hazard models neglect.

  2. Real-time hazard prediction — Monitor and predict the susceptibility to landslides, floods, liquefaction, and severe storms as conditions evolve.

  3. Computational playgrounds for scenario exploration — Build nowcasting and forecasting frameworks that couple AI-driven weather and climate models with geohazard models to interrogate future climate and hazard scenarios.


Use Cases

We ground our research in real-world coupled natural disasters that validate our methods and drive technological development:


Backbone Technologies

We develop integrated, cloud-native infrastructure to support the full research lifecycle — from data ingestion through hazard evaluation:


Earth System Science Nexus

At the heart of our framework is the critical zone — the thin, dynamic layer from bedrock to canopy where rock, soil, water, air, and life interact. The hydromechanical state of the soil governs how water infiltrates to recharge groundwater, evaporates to feed atmospheric moisture, or runs off to drive erosion and flooding. Understanding and monitoring this state — its memory of past wetting, drying, and disturbance — is essential for predicting hazard severity under current and future conditions.

We investigate two key nexus domains: