How to Reconstruct Fault Movement and Assess Tsunami Risk After a Giant Earthquake: A Step-by-Step Guide

Introduction

On July 29, 2025, a magnitude 8.8 earthquake struck near the Kamchatka Peninsula, ranking as the sixth-largest ever recorded by modern instruments. This event provided a unique opportunity for researchers at Tohoku University's International Research Institute of Disaster Science (IRIDeS) to refine methods for understanding tsunami risks. By combining multiple datasets, they reconstructed the fault rupture process and gained insights that can help protect coastal communities. This guide walks you through their methodology—a proven approach for analyzing giant earthquakes and assessing tsunami hazards. Whether you are a researcher, disaster manager, or advanced student, these steps will help you replicate the analysis and apply it to other subduction zones.

What You Need

Before starting, gather the following materials, data, and tools:

• Seismic waveform data from regional and global networks (e.g., IRIS, GEOSCOPE, local arrays)
• Geodetic data – GPS/GNSS time series near the epicenter and InSAR (Interferometric Synthetic Aperture Radar) images if available
• Tsunami records – tide gauge and deep-ocean bottom pressure recorder (e.g., DART buoy) data
• Bathymetry and topography digital elevation models (e.g., GEBCO, SRTM)
• High-performance computing resources or a powerful workstation
• Fault modeling software – for elastic dislocation calculations (e.g., Okada's model, Coulomb 3)
• Tsunami simulation code – open-source options like GeoClaw, COMCOT, or MOST
• Data analysis and visualization tools – Python with NumPy, SciPy, ObsPy; MATLAB; GIS software (QGIS, ArcGIS)

Steps to Reconstruct Fault Rupture and Tsunami Risk

Follow these numbered steps in order. Each builds on the previous one.

1. Step 1: Collect and Preprocess Seismic and Geodetic Data

   Start by gathering seismic waveform data from multiple stations within 10° of the epicenter. Remove instrument response, filter between 0.01–0.5 Hz for the teleseismic band, and cut windows around P-wave and S-wave arrivals. For geodetic data, process GPS time series to obtain cosesimic offsets using a reliable inversion approach (e.g., estimate static offsets before and after the event). If InSAR is available, unwrap interferograms to get displacement maps. Ensure all data are in a consistent coordinate system (WGS84) and time (UTC).

2. Step 2: Perform a Joint Fault Rupture Inversion

   Use an inversion algorithm that can simultaneously fit the seismic waveforms and geodetic displacements. Define a fault plane based on moment tensor solutions and aftershock locations. Discretize the fault into subfaults (e.g., 10×10 km). For each subfault, compute Green's functions for seismic (using a 1D velocity model) and static displacements (using Okada's point source formulas). Invert for slip amplitude, rake angle, and rupture time on each subfault while smoothing constraints. A common approach is the linear least-squares with positivity constraint. The Tohoku team used a multi‑time window method to capture the rupture propagation. Validate by checking the fit to data and the resolution of slip patches.

3. Step 3: Validate the Slip Model with Aftershock Locations

   Compare your inverted slip distribution with the spatial pattern of aftershocks. Typically, high‑slip areas correlate with gaps in aftershocks (they are stress shadows) while edges of slip zones show more aftershocks. If your model fails this qualitative check, revise the fault geometry or smoothing parameters. This step ensures the rupture model is geologically plausible.

4. Step 4: Run Tsunami Propagation Simulations

   Input the seafloor deformation computed from the slip model (using Okada or equivalent) as the initial condition for tsunami simulation. Set up a computational grid that covers the entire source region and coastal areas of interest. Use a nested grid approach to resolve nearshore bathymetry. Run the tsunami code (e.g., GeoClaw with adaptive mesh refinement) for at least 12 hours of simulated time. Record water elevations at virtual tide gauges near the Kamehatka coast and at DART stations.

5. Step 5: Compare Simulated vs Observed Tsunami Heights

   Extract synthetic waveforms from your simulation at the locations of actual tide gauges and DART buoys. Compute goodness‑of‑fit metrics: correlation coefficient, root‑mean‑square error, and Aida index. The Tohoku researchers found that their model reproduced the relatively modest tsunami heights (~5 m maximum) despite the large magnitude because the rupture was deep and had a large component of strike‑slip motion. If your simulation overestimates the tsunami, revisit the slip distribution and consider using tsunami inversion to refine the model.

6. Step 6: Identify Areas of High Residual Risk

   Map the difference between predicted and observed tsunami runup along the coastline. Areas where the model underpredicts the observations may indicate that additional tsunami sources (e.g., submarine landslides) were present. Even if the tsunami was smaller than expected for an M8.8 earthquake, there can be pockets of elevated risk from local bathymetric effects or complex rupture geometry. Perform a sensitivity analysis by perturbing the slip model and re‑running the tsunami simulation. This helps delineate zones where the community should remain vigilant for future events.

Tips for Success

• Integrate all available data. The power of the Tohoku team’s analysis came from combining seismic, geodetic, and tsunami records. Never rely on a single dataset alone.
• Use near‑field stations if possible. For giant earthquakes, near‑field data (within 500 km) provide crucial constraints on shallow slip and rupture timing.
• Check for tsunami source complexity. If the tsunami is smaller than expected, consider deep slip, oblique faulting, or a segmented rupture. The Kamchatka event featured a combination of these factors.
• Automate the workflow. Write scripts for data download, preprocessing, and inversion to reduce human error and speed up iteration.
• Communicate results clearly. Use maps of slip and tsunami hazard zones to inform local emergency managers and planning agencies.
• Stay updated. The field of source inversion is rapidly evolving. New methods like full‑wavefield inversion or machine‑learning acceleration can improve accuracy.

By following these six steps, you can reconstruct the fault rupture of a giant earthquake and assess tsunami risk with the same rigor as the IRIDeS team. Their work on the 2025 Kamchatka earthquake demonstrates that even an M8.8 event can produce a manageable tsunami—if the rupture is deep and strike‑slip dominated. However, the analysis also highlights lingering uncertainties that warrant continued monitoring and preparedness. Use this guide as a starting point, and adapt it to the specific tectonic setting of your region.