Investigating signal penetration depth in Synthetic Aperture Radar imagery over lowland permafrost

Abstract

Synthetic Aperture Radar (SAR) is an established technique to study geophysical phenomena occurring at the Earth surface. In the case of lowland permafrost monitoring, SAR is particularly interesting as it combines wave penetration capabilities within scatterers in the scene and large-scale imaging. In order to examine which ground layer the radar signal reaches, it is crucial to estimate the signal’s penetration depth within the soil, and through the vegetation covering it. We propose a study based on the PermASAR airborne campaign which was conducted by the German Aerospace Center (DLR) in the Canadian Arctic. This study focuses on a well-studied lowland permafrost site in the Canadian Artic, namely the Trail Valley Creek catchment close to Inuvik (New Territories, Canada). The catchment is characterized by tundra vegetation over continuous permafrost. Acquisitions over the test-site were performed successively in summer 2018 and winter 2019, corresponding to thawed respectively frozen state of the vegetation and active layer. The sensor was operated with signals at three different wavelengths, namely X (3cm), C (6 cm) and L- band (23 cm). Those three signals are sensitive to different natural objects and are expected to penetrate differently into frozen soil. Fully polarimetric images were acquired, allowing to separate different types of scattering. Finally, the flight geometry features several parallel tracks, allowing for interferometric processing resulting in the location of the height of the scatterer within one resolution cell. We present here a preliminary investigation towards the characterization of the penetration of SAR waves into the frozen ground by observing and analyzing the penetration effect as retrieved from the phase center height bias.