A method to automatically determine sea level for referencing snow freeboards and computing sea ice thicknesses from NASA IceBridge airborne LIDAR

• Published in: Remote sensing of environment Vol. 131; pp. 160 - 172
• Main Authors: Wang, Xianwei, Xie, Hongjie, Ke, Yanan, Ackley, Stephen F., Liu, Lin
• Format: Journal Article
• Language: English
• Published: New York, NY Elsevier Inc 15-04-2013; Elsevier
• Subjects: Animal, plant and microbial ecology; Applied geophysics; Bellingshausen Sea; Biological and medical sciences; Earth sciences; Earth, ocean, space; Exact sciences and technology; Fundamental and applied biological sciences. Psychology; General aspects. Techniques; Ice thickness; IceBridge; Internal geophysics; Sea level reference; Snow freeboard; Teledetection and vegetation maps; Antarctic Ocean; polar regions; Antarctica; Bellingshausen Sea; IceBridge; Sea level reference; Bellingshausen Sea; Ice thickness; Snow freeboard; thickness; NASA; Space remote sensing; topography; cartography; automated analysis; Profile; snow; sea ice; imagery; observations; Lidar; methodology; sea level
• ISSN: 0034-4257; 1879-0704
• DOI: 10.1016/j.rse.2012.12.022

The NASA IceBridge flights have obtained critical observations for Earth's polar ice since ICESat stopped collecting data in 2009. This study develops an automatic method in processing IceBridge Airborne Topographic Mapper (ATM) altimeter L1B data (one elevation per 3–4m horizontally) to derive a local sea level height for referencing snow freeboards and then computing sea ice thicknesses. Four 30-km L1B profiles (A, B, C and D) flown on October 21, 2009 over the Bellingshausen Sea in Antarctica are selected. The local sea level reference is first obtained by visual examination of ATM L1B heights over leads or thin ice identified on images simultaneously acquired from the Digital Mapping System camera (called manual selection). This sea level reference is then used as ground truth to validate sea level heights derived by automatic calculations using five thresholds of 2%, 1%, 0.5%, 0.2% and 0.1% of the lowest L1B data. The L1B_0.2% method gives a similar sea level height as from the L1B manual selection, by mean (absolute) difference of −0.01 (0.06) m. The sea level heights demonstrate a near linear gradient of 0.01m/km to 0.03m/km within each ~30-km L1B profile along the flight track from section A to D. The resulting mean snow freeboards are 0.59m, 0.67m, 0.53m, and 0.60m on sections A, B, C and D, respectively. Three empirical equations and the buoyancy equation (with zero ice freeboard assumption) all give similar statistics in ice thickness estimation with mean ice thicknesses of 1.91m for section A, 2.16m for B, 1.76m for C, and 1.94m for D. The sea level over leads cannot be accurately resolved from the ATM L2 data (~60m x 80m horizontal averaging). However, by using a sea level reference obtained from the L1B data, the ATM L2 data can achieve reasonable ice thickness estimates with mean absolute difference of only 0.10m compared to thickness derived from the L1B data.
► We obtained a local sea level reference at the accuracy of centimeters using the ATM L1B data. ► The sea level reference has a spatial gradient from 0.01m to 0.03m/km along each L1B profiles. ► The mean snow freeboards are 0.59, 0.67, 0.53, and 0.60m on the four profiles, respectively. ► The mean ice thicknesses are 1.91, 2.16, 1.76, and 1.94m on the four profiles, respectively. ► The ATM L2 data can give an overall estimate of ice thickness, but can not accurately resolve leads.