loading page

Mitigation of Decorrelation and Atmospheric Noise in InSAR Time-series, with Application to Postseismic Deformation following the 2021 Haiti Earthquake
  • +1
  • Sanaz Vajedian,
  • Jeremy Maurer,
  • Harriet Zoe Yin,
  • Jennifer S. Haase
Sanaz Vajedian

Corresponding Author:[email protected]

Author Profile
Jeremy Maurer
Author Profile
Harriet Zoe Yin
Author Profile
Jennifer S. Haase
Author Profile


Geodetic analysis of postseismic responses to major earthquakes offers insight into potential subsequent seismic activities and aseismic strain release. Interferometric Synthetic Aperture Radar (InSAR) provides a particularly high-resolution imaging capability for such transient events, particularly in regions without dense Global Navigation Satellite Systems (GNSS) networks. However, InSAR suffers from decorrelation errors and atmospheric noise, which can distort the interpretation of deformation patterns. To take a step towards addressing these challenges, this study introduces an advanced InSAR processing workflow and applies it to Sentinel-1 observations of postseismic deformation following the 2021 Haiti earthquake. We address decorrelation errors by employing phase linking through the Fine Resolution InSAR with Generalized Eigenvectors (FRInGE) method. We compare three methods for mitigating atmospheric effects, including a grouped Independent Component Analysis (ICA) method, and find that ICA performs best in removing atmosphere. Without applying these methods most of the signal is lost or hidden in the noise; after processing transient postseismic deformation, likely related to shallow fault creep, can be observed over ~3 months following the earthquake on the eastern EPGFZ. We compare the estimated cumulative slip to that obtained from ALOS-2 observations and find a good match, with ~3 cm of differential displacement on either side of the EPGFZ east of the rupture area. Our workflow provides a method for more precise characterization of localized transient deformation signals using C-band InSAR.
21 Apr 2024Submitted to TechRxiv
29 Apr 2024Published in TechRxiv