What question did this study set out to answer?

This study examines the relationship between ground ice content and InSAR-derived seasonal displacements in permafrost.

February 19, 2026Open Access

InSAR sensitivity to active layer ground ice content in Adventdalen, Svalbard

Key Points

This study examines the relationship between ground ice content and InSAR-derived seasonal displacements in permafrost.
Collected sediment cores to analyze volumetric ground ice and excess ice contents.
Estimated active layer thickness by measuring thaw depth post-thawing season.
Derived vertical displacements using SBAS processing of Sentinel-1 images.
Conducted Monte Carlo simulations to account for uncertainties in subsidence measurements.
Observed a mean coefficient of determination of 0.72 between measured and expected subsidence.
Mean absolute error of 14 mm in correlation between InSAR displacements and in-situ measurements.
Noted high variability in excess ice as a major factor for observed subsidence.
Identified a weak relationship between expected subsidence and active layer thickness due to complex ice distribution.

Abstract

Abstract. Interferometric Synthetic Aperture Radar (InSAR) remote sensing of surface displacement in permafrost environments has the potential to resolve ground ice dynamics and potentially active layer thickness, yet field validation is sparse. Here we present a comparison between in-situ ground ice contents and the seasonal InSAR displacements of the following thawing season at 12 coring sites in Adventdalen, Svalbard. The study is focused on the year 2023, where frozen sediment cores were collected at the end of spring from the active layer and the uppermost permafrost. The sediment cores were analyzed with high resolution for volumetric ground ice and excess ice contents. The active layer thickness was estimated by probing the thaw depth at the end of the thawing season 2023, allowing estimation of the amount of expected subsidence from seasonal ground ice melt. The InSAR vertical displacements for the thawing season were derived from Small Baseline Subset (SBAS) processing of Sentinel-1 imagery. The expected subsidence from ground ice melt within the measured active layer aligned well with the seasonal InSAR maximum vertical displacement. Monte Carlo simulations were performed to include uncertainties in the expected and measured InSAR subsidence, leading to a mean coefficient of determination of 0.72 and a mean absolute error of 14 mm for the correlation between InSAR subsidence and expected subsidence from in-situ ground ice melt. Excess ice is highly variable and is the main source of the expected subsidence during this thawing season, which was exceptionally warm. The expected subsidence and active layer thickness show only a weak relationship due to the observed complex ice content distribution in the active layer and uppermost permafrost. Our results show the significant potential of InSAR for mapping ground ice variability; however, they also suggest that estimating active layer thickness using InSAR requires careful consideration of the complex occurrence of both pore and excess ice in the active layer and uppermost permafrost.

Bookmark

View Full Paper