Skip to content

EQ-INSAR

Earthquake InSAR Synthetic Data Generator

A lightweight, physics-based forward model for generating synthetic InSAR surface deformation data from earthquake sources. Designed for machine learning training, benchmarking, and sensitivity analysis.

Python 3.8+ License: MIT PyPI DOI

Why EQ-INSAR?

Generating realistic synthetic InSAR data for earthquake studies typically requires heavy finite-element codes or full elastic dislocation models. EQ-INSAR provides a fast, lightweight alternative using the Davis (1986) point source model — ideal for producing large-scale training datasets for machine learning.

Key Features

  • Fast Point Source Model — Davis (1986) elastic half-space model, accurate for small-to-moderate earthquakes (Mw < 6.5)
  • InSAR-Native Outputs — Line-of-sight displacement, wrapped/unwrapped interferometric phase
  • 9 SAR Satellites — Sentinel-1, ALOS-2, TerraSAR-X, COSMO-SkyMed, RADARSAT-2, NISAR, SAOCOM, ENVISAT, ICEYE
  • ML Training Ready — Time series with pre/co/post-event frames, binary segmentation labels, batch generation
  • Minimal Dependencies — Core computation requires only NumPy
  • Export Formats — GeoTIFF and NetCDF support

Quick Example

from eq_insar import generate_synthetic_insar

# Mw 6.0 thrust earthquake with Sentinel-1 geometry
result = generate_synthetic_insar(
    Mw=6.0,
    strike_deg=30,
    dip_deg=45,
    rake_deg=90,
    depth_km=10,
    satellite='sentinel1',
    orbit='ascending'
)

los_displacement = result['los_displacement']   # (H, W) array
wrapped_phase = result['wrapped_phase']          # (H, W) array

Physics

EQ-INSAR implements the Davis (1986) point source Green's functions for surface displacement in an elastic half-space, combined with:

  • Aki & Richards (2002) moment tensor convention (strike/dip/rake)
  • Hanks & Kanamori (1979) moment magnitude scale
  • Wells & Coppersmith (1994) fault dimension scaling relations

Installation

pip install eq-insar

See the Getting Started guide for more options.

Citation

If you use EQ-INSAR in your research, please cite:

@software{cieslik2026eqinsar,
  author = {Cieslik, Konrad and Milczarek, Wojciech},
  title = {EQ-INSAR: A Python Package for Generating Synthetic Earthquake InSAR Deformation Data},
  year = {2026},
  url = {https://github.com/kcieslik/eq-insar},
  doi = {10.5281/zenodo.18647189}
}

License

MIT License — see LICENSE for details.

Developed at the Wroclaw University of Science and Technology & trainai.io.