Quantitative Evaluation of Landsat 7 ETM+ SLC-off Images for Surface Velocity Estimation of Mountain Glaciers

Wednesday, 17 December 2014
Liming Jiang1, Yongling Sun1,2, Lin Liu1,2, Sisheng Wang1,2 and Hansheng Wang1, (1)State Key Laboratory of Geodesy and Earth's Dynamics, Institute of Geodesy and Geophysics, Chinese Academy of Sciences, Wuhan, China, (2)UCAS University of Chinese Academy of Sciences, Beijing, China
In many cases the Landsat mission series (Landsat 1-5, 7 and 8) provide our only detailed and consistent data source for mapping the global glacier changes over the last 40 years. However, the scan-line corrector (SLC) of the ETM+ sensor on board Landsat 7 permanently failed, resulting in wedge-shaped data gaps in SLC-off images that caused roughly 22% of the pixels to be missed. The SLC failure has left a serious problem for the glacial applications of ETM+ data, particularly for monitoring long-term glacier dynamics in High Asian Mountain where has few available data due to the frequently cloudy covers.

This study aims to evaluate the potential of the Landsat 7 SLC-off images in deriving surface velocities of mountain glaciers. A pair of SLC-off images over the Siachen glacier acquired in Aug 2009 and 2010 was used for this purpose. Firstly, two typical filling-gap methods, the localized linear histogram match (LLHM) and the weighted liner regression (WLR), were utilized to recover the mentioned SLC-off images. Subsequently these recovered pairs were applied for deriving glacier-surface velocities with the COSI-Corr feature tracking procedure. Finally, the glacier velocity results were quantitatively compared with that of a pair of Landsat-5 TM images acquired nearly at the same time with the SLC-off pair.

Our results show that (1) the WLR method achieves a better performance of gap recovering than the LLHM method, (2) the surface velocities estimated with the recovered SLC-off images are highly agreement with those of the TM images, and (3) the annual mean velocity of the Siachen glacier is approximately 70 m/yr between 2009 and 2010 with a maximum of 280 m/yr close to the glacial equilibrium line that are similar with the results in previous studies. Therefore, if a suitable filling-gap method is adopted, e.g. the WLR method, it is highly feasible that the ETM+ SLC-off data can be utilized to estimate the surface velocities of mountain glaciers.