Mapping Earth surface elevation changes using offsets between Sentinel-2, Landsat, and Sentinel-1 orthoimages from adjacent orbits

Abstract. Accurate digital elevation models (DEMs) are essential for orthorectifying satellite imagery, but any deviation between the orthorectification DEM and the true surface elevation at acquisition time produces horizontal offsets in the resulting orthoimages. These offsets are typically an unwanted artefact. Here, we exploit cross-track disparities between overlapping orthoimages from adjacent orbits opportunistically to retrieve elevation changes at the Earth surface. We develop the geometric framework for this orthoimage-disparity approach for both optical pushbroom sensors (Sentinel-2, Landsat 8/9) and synthetic aperture radar (Sentinel-1), and present several implementation strategies, including single differences against the orthorectification DEM, double differences between two image epochs that eliminate the reference DEM, mixed-mission combinations (e.g. Sentinel-2 with Landsat 8/9), and joint retrieval of vertical change together with horizontal surface displacement. A broad set of case studies illustrates the method across geoscientific applications: glacier thickness changes on Aletsch (Switzerland), Disappointment (Yukon), Sermeq Kujalleq (Jakobshavn isbræ, Greenland), and Southern Patagonian Ice Field; surging glaciers in Svalbard; outlet dynamics in West Antarctica; a subglacial lake drainage on Ellesmere Island; the Aru glacier detachments and the Sedongpu erosion canyon on the Tibetan Plateau; the Karrat fjord landslide (Greenland); permafrost thaw slumps in northern Canada; topographic change at Kilauea and Nishinoshima volcanoes; snow and ice change in Iceland; changes in level and volume of a reservoir (Australia), co-seismic uplift from the 2016 Kaikōura earthquake (New Zealand); and Amazon forest clear-cuts. Single orthoimage offsets reach an accuracy of roughly 1/10 of a pixel, resulting in effective elevation uncertainties of around 3–5 m. Metre-scale precision seems achievable through stacking. The main limitations are cloud cover, sharp topographic edges, and the quality of the original orthorectification DEM. The approach is highly scalable, builds on freely available long-term operational missions, and offers strong potential for a so far unexploited global monitoring of surface elevation change at medium resolution.

Authors

Institutions

Publication Details

Published
2026-11-17
DOI
https://doi.org/10.5194/egusphere-2026-6066
Primary Topic
Satellite Image Processing and Photogrammetry
Type
preprint

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Mapping Earth surface elevation changes using offsets between Sentinel-2, Landsat, and Sentinel-1 orthoimages from adjacent orbits

Étienne Berthier, Andreas Kääb
Satellite Image Processing and Photogrammetry
preprint

Mapping Earth surface elevation changes using offsets between Sentinel-2, Landsat, and Sentinel-1 orthoimages from adjacent orbits

Étienne Berthier, Andreas Kääb
preprint en

Abstract

Abstract. Accurate digital elevation models (DEMs) are essential for orthorectifying satellite imagery, but any deviation between the orthorectification DEM and the true surface elevation at acquisition time produces horizontal offsets in the resulting orthoimages. These offsets are typically an unwanted artefact. Here, we exploit cross-track disparities between overlapping orthoimages from adjacent orbits opportunistically to retrieve elevation changes at the Earth surface. We develop the geometric framework for this orthoimage-disparity approach for both optical pushbroom sensors (Sentinel-2, Landsat 8/9) and synthetic aperture radar (Sentinel-1), and present several implementation strategies, including single differences against the orthorectification DEM, double differences between two image epochs that eliminate the reference DEM, mixed-mission combinations (e.g. Sentinel-2 with Landsat 8/9), and joint retrieval of vertical change together with horizontal surface displacement. A broad set of case studies illustrates the method across geoscientific applications: glacier thickness changes on Aletsch (Switzerland), Disappointment (Yukon), Sermeq Kujalleq (Jakobshavn isbræ, Greenland), and Southern Patagonian Ice Field; surging glaciers in Svalbard; outlet dynamics in West Antarctica; a subglacial lake drainage on Ellesmere Island; the Aru glacier detachments and the Sedongpu erosion canyon on the Tibetan Plateau; the Karrat fjord landslide (Greenland); permafrost thaw slumps in northern Canada; topographic change at Kilauea and Nishinoshima volcanoes; snow and ice change in Iceland; changes in level and volume of a reservoir (Australia), co-seismic uplift from the 2016 Kaikōura earthquake (New Zealand); and Amazon forest clear-cuts. Single orthoimage offsets reach an accuracy of roughly 1/10 of a pixel, resulting in effective elevation uncertainties of around 3–5 m. Metre-scale precision seems achievable through stacking. The main limitations are cloud cover, sharp topographic edges, and the quality of the original orthorectification DEM. The approach is highly scalable, builds on freely available long-term operational missions, and offers strong potential for a so far unexploited global monitoring of surface elevation change at medium resolution.

Centre National de la Recherche Scientifique (FR), Université Fédérale de Toulouse Midi-Pyrénées (FR), University of Oslo (NO), Institut de Recherche pour le Développement (FR)
Schmidt Sciences, European Space Agency, Centre National d’Etudes Spatiales, HORIZON EUROPE European Research Council
Climate action
Satellite Image Processing and Photogrammetry
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon