Near-Real-Time InSAR Phase Estimation for Large-Scale Surface Displacement Monitoring

Abstract

Operational near-real-time monitoring of Earth's surface deformation using Interferometric Synthetic Aperture Radar (InSAR) requires processing algorithms that efficiently incorporate new acquisitions without reprocessing historical archives. We present sequential phase linking approach using compressed single-look-complex images (SLCs) capable of producing surface displacement estimates within hours of the time of a new acquisition. Our key algorithmic contribution is a mini-stack reference scheme that maintains phase consistency across processing batches without adjusting or re-estimating previous time steps, enabling straightforward operational deployment. We introduce online methods for persistent and distributed scatterer identification that adapt to temporal changes in surface properties through incremental amplitude statistics updates. The processing chain incorporates multiple complementary metrics for pixel quality that are reliable for small SLC stack sizes, and an L1-norm network inversion to limit propagation of unwrapping errors across the time series. We use our algorithm to produce OPERA Surface Displacement from Sentinel-1 product, the first continental-scale surface displacement product over North America. Validation against GPS measurements and InSAR residual analysis demonstrates millimeter-level agreement in velocity estimates in varying environmental conditions. We demonstrate our algorithm's capabilities with a successful recovery of meter-scale co-eruptive displacement at Kilauea volcano during the 2018 eruption, as well as detection of subtle uplift at Three Sisters volcano, Oregon -- a challenging environment for C-band InSAR due to dense vegetation and seasonal snow. We have made all software available as open source libraries, providing a significant advancement to the open scientific community's ability to process large InSAR data sets in a cloud environment.

Figures (12)

• Figure 1: Batch processing of $N=12$ SLC images in mini-stacks of size $M=4$. Squares indicate input SLCs, wavy lines indicate multi-looked interferograms, rounded squares indicate compressed SLCs. (a) The first mini-stack (SLC images 1-4) is phase-linked with reference index $r=1$. Input SLCs are compressed to produce $\kappa^{(1)}_{{1}\to{4}}$, and interferograms are formed with SLC 1 as reference. (b) The second mini-stack uses $\kappa^{(1)}_{{1}\to{4}}$ as a reference to create $\kappa^{(1)}_{{5}\to{8}}$ and to produce interferograms relative to day 1. (c) Subsequent mini-stacks use the most recently formed compressed SLC as reference. For mini-stack 3, this means that the mini-stack uses reference $r=2$, which is $\kappa^{(1)}_{{5}\to{8}}$. The outputs are interferograms relative to day 1.
• Figure 2: Synthetic demonstration of sequential phase linking for a $1500\times1500$ pixel scene, processed in mini-stacks of 15 SLCs. (a) Simulated temporal correlation matrix. (b) Example of simulated "true" phase, consisting of a deformation bowl plus atmospheric noise. (c) Conventional interferogram formed from two real SLCs ($s_1 s_{20}^*$) after multi-looking $(15,15)$. (d) Interferogram for the same time span using the compressed SLC $\kappa^{(1)}_{{1}\to{15}}$ as the reference, showing improved coherence. (e) Comparison of RMSE versus time for different estimators: near-real-time phase linking (blue), datum-adjusted phase linking (orange), multi-looked interferograms (green), pure noise (pink), and the Cramér–Rao lower bound (yellow).
• Figure 3: Summary of the modules used in the displacement processing workflow. Letters refer to subsections of Section \ref{['sec:proc-details']} providing details.
• Figure 4: Calibration and Validation (CalVal) sites covering diverse thematic domains, including those related to volcanoes, tectonics/earthquakes, landslides, sinkholes, and land subsidence. These CalVal sites are selected to capture a range of environmental complexities and terrain types representative of the North America geographic scope.
• Figure 5: Average LOS surface velocity from Nov. 2017 to Nov. 2024 for the ascending Sentinel-1 geometry. Red indicates average motion toward the satellite, blue indicates motion away. Velocities are relative to an internal average rate per frame. For wide-area visualization continuity, a planar ramp was removed from each frame, and a constant shift was estimated for each frame using along-track overlapping regions. Descending LOS velocities over the same time period are shown in Supplement Figure \ref{['fig:supplement-descending-conus-7year-velocity']}.