Improving The Capability Of DInSAR Combined With Offsettracking For Monitoring Glacier Velocity

Author links open overlay panelXiaomanFeng1ZhuoqiChen1GangLiPersonEnvelopeQiJuZhibingYangXiaoCheng

Show moreNavigate DownHighlights
•Subtracting offset-tracking estimate phases helps in phase unwrapping.

•Our proposed method can be applied in monitoring polar glacier dynamic.

•InSAR based method are accurate in monitoring glacier flow.

Abstract
Offset-tracking is more commonly used in monitoring glacier velocity at Greenland Ice Sheet (GrIS) than the D-InSAR technique because of its robustness in handling low coherence and unnecessary of phase unwrapping. Large deformation rates of polar glaciers can cause a deformation gradient which exceeds the theoretical limitation of phase unwrapping. There is acceptable coherence for Sentinel-1 interferograms with very short temporal baselines, however, traditional D-InSAR still fails at the phase unwrapping step. Here, we propose a method to reduce the phase gradient of the D-InSAR interferogram via subtracting the estimated deformation phase to obtain a residual phase. Such rough deformation can be acquired by performing offset-tracking to the current pair of SAR images or can be calculated by any other glacier velocity products, which are usually obtained by the offset-tracking technique. Phase unwrapping is then performed on the residual phase and then the rough deformation phase is added to yield the final deformation fields. Four study sites located in northern Greenland are selected to test our proposed method, which improved the mean coherence of the residual phase by 0.037, 0.032, 0.037, and 0.039 in spring, summer, autumn, and winter, as compared to traditional D-InSAR interferometry. The modulus of the phase gradient is also reduced, by 0.054, 0.044, 0.056, and 0.064 rad in each season. The coherence improvement is more sensitive to the temporal baseline in summer than other seasons, and our method is less effective. Seven unwrapped phase profiles of the glaciers find that our method performs better in spatial and temporal phase continuity than traditional D-InSAR. The maximum deformation detected during 6-day baseline increases from ∼1.4 m by D-InSAR to ∼3.6 m by the proposed method. Our approach extends the area that can be unwrapped correctly, especially for the downstream zone of glaciers, where there are higher flow rates, in turn, this increases the robustness of the phase unwrapping and the possibility of applying the InSAR method for monitoring fast-moving glaciers in the GrIS.

* Navigate LeftPrevious article in issue
* Next article in issueNavigate Right

Data availability
Data will be made available on request.

1The first two authors contributed equally to this research.

View full text© 2022 Elsevier Inc. All rights reserved.