Co-Registration Methods and Error Analysis for Four Decades (1979–2018) of Glacier Elevation Changes in the Southern Patagonian Icefield

Abstract

The main goal of this paper is to compare two co-registration methods for geodetic mass balance (GMB) calculation in 28 glaciers making up the Upper Santa Cruz River basin, Southern Patagonian Icefield (SPI), from 1979 to 2018. For this purpose, geospatial data have been used as primary sources: Hexagon KH-9, ASTER, and LANDSAT optical images; SRTM digital radar elevation model; and ICESat elevation profiles. After the analyses, the two co-registration methods, namely M1, based on horizontal displacements and 3D shift vectors, and M2, based on three-dimensional transformations, turned out to be similar. The errors in the GMB were analyzed through a k index that considers, among other variables, the error in elevation change by testing four interpolation methods for filling gaps. We found that, in 63% of the cases, the relative error in elevation change contributes 90% or more to k index. The GMB throughout our study area reported that a loss value of −1.44 ± 0.15 m w. e. a−1 (−3.0 Gt a−1) and an ice thinning median of −1.38 ± 0.11 m a−1 occurred within the study period. The glaciers that showed the most negative GMB values were Upsala, with an annual elevation change median of −2.07 ± 0.18 m w. e. a−1, and Ameghino, with −2.31 ± 0.22 m w. e. a−1.

Keywords

• satellite data
• co-registration methods
• geodetic mass balance
• Southern Patagonian Icefield