Research

Hurricane Fiona, Puerto Rico (11/05/2022)

Coupled Regional Rainfall-Induced and Seismic Slope Instability Simulations (CRISIS) Simulations

CRISIS is a state-of-the-art, open-source, physics-based model designed to simulate 3D rainfall-induced and earthquake-triggered landslides at regional scales with high detail and precision.

CRISIS can predict risks over time, enabling timely responses and effective early warning systems.

It enables the characterization of regional natural hazards, and the development of mitigation strategies to enhance community resilience and support sustainable infrastructure.

Whether you're studying landslide hazards in mountainous catchments or assessing regional susceptibility to coupled rainfall-earthquake events, CRISIS offers a robust, high-resolution solution grounded in physical process understanding.

Collaborators: Prof. Dimitrios Zekkos

Key Features of CRISIS:

1. Regional Scale Landslide Modeling Capabilities

Simulates rainfall-induced and coseismic landslides at a regional scale while preserving high-resolution inputs and advanced model capabilities.

2. Transient 3D Hydrological Model

Integrates a user-defined hydrological model—here, ParFlow—to simulate 3D transient groundwater flow in variably saturated subsurface, overland flow, and the effects of evapotranspiration and vegetation.

3. Surface Runoff

Accounts for surface runoff generated when rainfall intensity (I) exceeds the soil’s infiltration capacity, approximated as k*(dh/dl), where k is the surface hydraulic conductivity and dh/dl is the hydraulic gradient.

4. Generated Pore Water Pressures

Integrates pore pressure heads that vary spatially, temporally, and with depth across the watershed. Shown here are surface pressure heads for illustrative purposes.

5. Pseudo-3D Slope Stability Analysis

Applies (a) an infinite slope stability model driven by simulated pore pressure heads to identify triggering cells, followed by (b) a pseudo-3D procedure to form 3D landslides.

6. Predicted Landslides Across the Watershed Through Time

Simulates rainfall-induced landslides, spatially across a watershed and temporally synchronized with the storm time series. Landslides are characterized by well-define location, area, volume, triggering depth, and timing of failure.

7. Classification into Different Failure Mechanisms

Classifies landslides into 4 different failure mechanisms: Bottom-up failures driven by rising groundwater tables, Top-down failures caused by downward-propagating wetting fronts, Top-down failures triggered by the formation of perched water tables, and Failures due to water exfiltration from fractured bedrock into overlying soil layers (Figure adapted from Johnson and Sitar (1990), with updates).

8. Two Components: Back-Analysis and Forward Modeling

Offers both forward modeling (for predictive simulations) and back-analysis (to reconstruct conditions leading to mapped landslides). This dual capability allows users to both forecast future risk and refine model performance using historical data.

Other Features:

Supports spatially and depth-variable input properties for detailed representation of real-world soil and hydrological heterogeneity.
Open-source and Python-based for easy access, customization, and integration.
Written in parallel and optimized for efficient use on high-performance computing (HPC) systems.
Runs on a wide range of platforms (from laptops to supercomputers) using the same source code and input files.
Utilizes HDF5 files for efficient, scalable data management and compatibility.
Includes a suite of Matlab and Python tools for input preparation, result visualization, and diagnostic analysis.

Some Example Applications:

Click on each image to learn more!

Failure Mechanisms Simulated by the CRISIS Model

Sensitivity Analysis of CRISIS's Parameters and Features

Influence of Storm Time Series Characteristics on Landslide Predictions

Impact of Spatial Resolution on Landslide Predictions: The Tradeoff between Detail and Computational Demands

Back-Analysis of Mapped Landslide Inventories

Collaborations

Application of Virtual Reality (VR) in Geotechnical Engineering

Mapping and Characterizing Landslides After Extreme Events

Publications

Journal Papers:

Kassem M, Zekkos D (2026) Assessing rainfall-induced landslide failure mechanisms in regional hydrological and hillslope stability simulations. Landslides. https://doi.org/10.1007/s10346-025-02690-w
Kassem M, Zekkos D, Gong W, Clark M, Stanley T, Kirschbaum D. Multiscale Inter-Model Comparison of Regional Precipitation-Induced Landslide Prediction: The Example of Hurricane Maria in Puerto Rico. Journal of Natural Hazards (Springer). Under Review (second round).
Kassem M, Zekkos D. Sensitivity Analysis of a Regional Hydrological and Hillslope Stability Model: A Case Study from Hurricane Maria in Puerto Rico. Journal of Engineering Geology (Elsevier). Under Review (first round).

Conference Papers:

Kassem M, Gong W, Zekkos D, Clark M, Hughes S (2024) Pseudo-Three-Dimensional Back-Analysis of Rainfall-Induced Landslides in Utuado, Puerto Rico. In Geo-Congress 2024 (pp. 579-589). https://doi.org/10.1061/9780784485316.060
Kassem M, Zekkos D (2025) The Influence of Model Spatial Resolution on Rainfall-Induced Landslide Prediction for a Basin in Utuado, Puerto Rico. In Geo-Extreme 2025. https://doi.org/10.1061/9780784486528.008
Kassem M, Zekkos D (in press) The Influence of Storm Time Series Characteristics on Landslide Triggering within a Watershed in Utuado, Puerto Rico. In Geo-Congress 2026.

Technical Reports:

Morales A, Hughes S, Lang K, Rivera-Hernandez F, Vargas P, Mario Lozano J, Karantanellis E, Kassem M, Gomberg D, …, & Ortega V (2023) Geotechnical Impacts of Hurricane Fiona in Puerto Rico. Geotechnical Extreme Events and Reconnaissance (GEER). doi:10.18118/G6Z38B

This work is supported by

Page updated

Google Sites

Report abuse