Mitigating Topographic Residual Effects in InSAR-Based Landslide Detection: An Enhanced Stacking Approach

Interferometric Synthetic Aperture Radar (InSAR) has emerged as a powerful tool for landslide hazard detection, yet topographic residuals arising from outdated Digital Elevation Models (DEMs), dynamic terrain changes, and unknown scatterer positions pose significant challenges. These residuals, scaled by perpendicular baselines, can introduce substantial biases in deformation rate estimates, leading to overlooked hazards in techniques such as Stacking, Small Baseline Subset (SBAS), and Persistent Scatterer (PS)/Distributed Scatterer (DS) InSAR.

We present an enhanced Stacking methodology that eliminates topographic residual contributions through baseline normalization without directly estimating DEM errors. By leveraging the linear relationship between DEM error phase and spatial baseline, our approach performs phase normalization by baseline magnitude and applies sign-balancing transformations to ensure equal numbers of positive and negative perpendicular baselines. This preserves the simplicity, efficiency, and robustness of traditional Stacking while significantly improving deformation velocity estimation accuracy.

Additionally, we discuss complementary strategies including near-zero baseline InSAR approaches through interferogram integer combination and non-parametric Independent Component Analysis (ICA) methods for enhanced topographic residual estimation under complex deformation scenarios.
This work provides practical solutions for improving InSAR-based landslide hazard identification in dynamic terrain environments, with significant implications for geological disaster monitoring and early warning systems.