Sentinel-1-Aided Mutual Calibration of TanDEM-X DEMs for the Estimation of Height and Volume Changes

Abstract

Digital elevation models (DEM) derived from TanDEM-X single-pass interferometric SAR data are widely used for a large vary of applications. One of the main application fields regards the monitoring of topographic changes, which can be properly estimated by differentiating two DEM scenes acquired at different times. In this scenario, a critical aspect is represented by the precise mutual calibration of the two DEMs, which are normally affected by residual offsets and tilts caused by uncertainties in the baseline determination. This calibration is in general performed by utilizing reference tie-points, whose location and height is known a priori, or estimated through other reference sensors, such as GPS or LiDAR data. The manual procedure is time-consuming and can be jeopardized by the absence of available external reference points. In this paper, we present a novel technique for performing an automatic selection of reliable natural tie-points from Sentinel-1 repeat-pass time-series, by exploiting persistent scatterers candidates, and we discuss their applicability to TanDEM-X data. We propose a novel method for retrieving a set of reliable natural calibration targets and for performing an effective mutual calibration of TanDEM-X time-tagged DEMs. The approach is validated by considering the estimation of lava field volumes in the test-case scenario of the volcanic eruptions in the Reykjanes Peninsula, Iceland, occurred between 2021 and 2023. Thanks to the continuous global coverage of Sentinel-1, this approach can be applied when no a priori knowledge on reference tie-points is available, allowing for the precise estimation of height and volume changes.