Acta Geodaetica et Cartographica Sinica ›› 2026, Vol. 55 ›› Issue (6): 1031-1046.doi: 10.11947/j.AGCS.2026.20260019

• Geodesy and Navigation • Previous Articles    

High-resolution surface elevation change inversion over Antarctic ice shelf using REMA strips

Zhiwen YANG1,2(), Lu AN1,2(), Yujie SUN1,2, Jing LI1,2, Zhe ZHOU1,2, Junjie ZHANG1,2, Tong HAO1,2   

  1. 1.Center for Spatial Information Science and Sustainable Development Applications, Tongji University, Shanghai 200092, China
    2.College of Surveying and Geo-Informatics, Tongji University, Shanghai 200092, China
  • Received:2026-01-14 Revised:2026-04-25 Published:2026-07-28
  • Contact: Lu AN E-mail:yangzhiwen@tongji.edu.cn;anlu2021@tongji.edu.cn
  • About author:YANG Zhiwen (1998—), male, PhD candidate, majors in cryosphere remote sensing, ice shelf basal melting and stability, the validation and application of satellite altimetry data. E-mail: yangzhiwen@tongji.edu.cn
  • Supported by:
    The National Key Research and Development Program of China(2024YFF0808302);The National Natural Science Foundation of China(42576273);The Fundamental Research Funds for the Central Universities of China;Shanghai Pilot Program for Basic Research

RichHTML

PDF

Abstract:

Precise delineation of continuous surface elevation changes on ice shelves is critical for understanding ice-ocean interactions. However, existing observation frameworks are constrained by the sparse spatial sampling of satellite altimetry and the topographic advection artifacts inherent to high-resolution optical imagery under traditional Eulerian differencing. To address this issue, this study proposes a long-term, high-resolution retrieval method for surface elevation change rates based on REMA optical stereo strips. The method uses ITS_LIVE velocity fields to drive material-point backtracking and establishes a Lagrangian spatiotemporal matching model under the absolute elevation constraints of CryoSat-2 Swath altimetry, thereby effectively suppressing geometric artifacts arising from ice-flow advection. Using the Dotson Ice Shelf as a case study, and based on REMA strips and multi-source satellite observations from 2010 to 2022, we reconstructed a time series of surface elevation change rates on a 50 m common analysis grid using a strategy that combines long-term trends with sliding windows. Results show that: ① compared with the traditional Eulerian framework, the proposed method reduces the core error metric (NMAD) by 51.6% and substantially weakens spurious elevation-change signals in fast-flowing regions; ② independent validation against ICESat-2 ATL06 data yields RMSEs of 4.64 m and 2.97 m for the long-term retrieval and near-contemporaneous single-strip comparison, respectively, confirming the stability of the framework; ③ based on the constructed thinning intensity classification system, intense thinning zones (<－1 m/a) are highly concentrated near the grounding line and channel core areas (accounting for 9.2% of the area); the central Dotson melt channel exhibits a nonlinear evolution of “acceleration (2010—2016) followed by deceleration (2016—2022)”, while the grounding-line vicinity shows a continuous thinning trend. These results indicate that, under unified physical corrections and a Lagrangian framework, REMA strip data can be used to construct continuous products of ice-shelf surface elevation change rates, thereby providing finer observational constraints for analyzing BMC-related surface responses in fast-flowing ice shelves.

Key words: Dotson ice shelf, REMA, Lagrangian framework, surface elevation change rate, high-resolution time-series reconstruction

CLC Number: