B063-0006
Improved calibration of absolute InSAR deformation estimates in support of permafrost active layer retrievals

Friday, 11 December 2020
Poster
Yuhuan Zhao1, Richard H Chen2, Roger J Michaelides3, Taylor D Sullivan4, Andy Parsekian4, Howard A Zebker5, Kevin M Schaefer6 and Mahta Moghaddam1, (1)University of Southern California, Ming Hsieh Department of Electrical and Computer Engineering, Los Angeles, CA, United States, (2)Jet Propulsion Laboratory, Pasadena, CA, United States, (3)Colorado School of Mines, Department of Geophysics, Golden, CO, United States, (4)University of Wyoming, Laramie, WY, United States, (5)Stanford Univ, Stanford, CA, United States, (6)National Snow and Ice Data Center, Cooperative Institute for Research in the Environmental Sciences, University of Colorado at Boulder, Boulder, Colorado, U.S.A, Boulder, CO, United States
Abstract:
Active layer thickness (ALT) and soil moisture are key variables in monitoring the warming of permafrost in the Arctic-Boreal region. Interferometric synthetic aperture radar (InSAR) can measure seasonal deformation due to active layer thaw, from which ALT, the maximum thaw depth of the year, can be derived if subsurface soil moisture is known. Polarimetric SAR (PolSAR) time-series can also be used to retrieve soil moisture and ALT simultaneously but the ALT estimates are limited by the radar sensing depth (40-60 cm at P-band). We have integrated the complementary strengths of both methods and developed a joint approach for active layer retrievals using L-band InSAR and P-band PolSAR data acquired during the 2017 Arctic-Boreal Vulnerability Experiment (ABoVE) airborne campaign.

Compared to PolSAR backscatter, InSAR seasonal deformation is the dominant factor in estimating ALT so its calibration accuracy is critical to the overall retrieval performance. After phase unwrapping, InSAR phase measurements are still relative and must be calibrated to absolute deformations. Typical calibration techniques include referencing the entire deformation distribution to the reference points where zero deformation is expected or whose deformation can be inferred by in-situ thaw depth and soil moisture measurements. However, we observed that the distribution of deformation estimated from a single airborne SAR interferogram (UAVSAR) presents long tails at both ends, resulting in deformation estimates that are inconsistent with previous spaceborne InSAR (ALOS-2) estimates and field measurements. This suggests that the distribution of relative deformation estimates from UAVSAR needs to be empirically calibrated to more closely resemble independent measurements. In this work, we proposed to use reference points that represent different deformation levels to tighten the distribution to have a more realistic dynamic range as compared to the in-situ data. The calibrated deformation is then injected into the joint retrieval to enable more accurate active layer estimates. We will demonstrate the calibration techniques using the SAR data over Barrow and Yukon-Kuskokwim Delta, Alaska. The results show that the RMSEs of retrieved ALT and soil moisture are improved approximately by 40% and 20%, respectively.

<< Previous Abstract | Next Abstract