Synthetic Aperture Radar TOPS-mode Interferometry for Ice Velocity Retrieval

Jonas Kvist Andersen

Research output: Book/ReportPh.D. thesis

68 Downloads (Pure)

Abstract

Mass loss from the world’s ice sheets and glaciers is one of the largest contributors to ongoing sea level rise. For both the Greenland and Antarctic ice sheets, a significant part of the mass loss stems from changes in ice dynamics, with many marine-terminating outlet glaciers consistently accelerating and retreating. Ice velocity is an essential variable in monitoring the state of ice sheets and glaciers. Over the past several decades, a revolution in the quality and availability of Synthetic Aperture Radar (SAR) satellite data has allowed frequent ice velocity retrievals to be carried out over major parts of the ice sheets. Particularly, the EU/ESA Sentinel-1 SAR satellites have been widely utilized in the generation of seasonal, annual, and multi-annual velocity retrievals, owing to their extensive coverage, facilitated by the Terrain Observation by Progressive Scans (TOPS) acquisition mode. So far, however, routine Sentinel-1 ice velocity measurements have relied solely on amplitude-based methods, which produce measurements of significantly lower accuracy and resolution than phase-based Differential SAR Interferometry (DInSAR). The main reason for this discrepancy is the added complexity introduced to interferometric processing by the TOPS acquisition mode, where along-track motion is coupled to the interferometric phase.

In this thesis, a refined image coregistration approach is developed, which alleviates the TOPS-related challenges and allows interferometric ice velocity retrieval from the extensive Sentinel-1 polar archive. A demonstration is provided of a combined DInSAR and amplitude offset tracking 2D velocity retrieval, which exploits the high accuracy and resolution of DInSAR in slower-moving inland regions and the ability of offset tracking to retrieve measurements from fast-flowing glacier outlets, highlighting the synergy between the two techniques. In the context of ice velocity retrieval, particularly in downstream regions, phase unwrapping errors is the most prominent error source in DInSAR measurements. Here, a method is designed to detect and discard measurements affected by unwrapping errors. The method is based on an estimate of pixel connectivity and is particularly effective at detecting high magnitude errors.

Finally, several demonstrations are provided of the capabilities of Sentinel-1 DInSAR time series measurements in monitoring changes in ice dynamics, with a focus on specific regions of the Greenland ice sheet. On longer time scales (multiple years), the high sensitivity of DInSAR allows for estimating trends in flow speed, even in slow-moving regions, where only subtle changes occur. On shorter time scales (months or even weeks), the DInSAR time series may reveal complex motion patterns and transient dynamic events. Multiple instances of hydrology-dynamic events, related to the drainage of both supraglacial and subglacial lakes, are documented in the thesis. Observations of such events are crucial for improving our understanding of hydrological processes in glaciers and ice sheets and the extent of their impact in the future. Collectively, these demonstrations highlight the potential of applying Sentinel-1 DInSAR measurements in investigating various types of ice dynamic changes on multiple time scales.
Original languageEnglish
Place of PublicationKgs. Lyngby
PublisherTechnical University of Denmark
Number of pages203
Publication statusPublished - 2022

Fingerprint

Dive into the research topics of 'Synthetic Aperture Radar TOPS-mode Interferometry for Ice Velocity Retrieval'. Together they form a unique fingerprint.

Cite this