Assessing the suitability of DEMs derived from Sentinel-1 for landslide volume estimation

Short Description

Starting point / motivation

Each year, landslides cause numerous fatalities and significant infrastructure damages. Landslides occur predominantly in mountainous environments and having timely, accurate, and comprehensive information on the distribution and magnitude of landslide events demands laborious and time-consuming work.

Earth observation (EO) data, such as optical and radar satellite imagery, have played an important role in assessing and analysing landslides. In particular, volume estimates of landslides are critical for understanding landslide characteristics and (post-failure) behaviour as well as potential interconnected and cascading effects, such as landslide dam outburst floods or debris flows.

Pre- and post- event digital elevation model (DEM) differencing is a suitable method to estimate landslide volumes remotely, leveraging EO techniques. However, such analyses are restricted by limitations inherent to existing DEM products, such as high costs for commercial products, limited temporal and spatial coverage and resolution, or insufficient accuracy. Therefore, there is a need for systematic generation of DEMs to facilitate volume estimations of landslides at sufficient spatial and temporal scales.

The European Union's Earth observation programme Copernicus features two Synthetic Aperture Radar (SAR) sensors, namely Sentinel-1A and Sentinel-1B. Together they provide 6-day repeat imagery, with the advantage of being weather and daylight independent. Pairs of SAR images can be processed using interferometric SAR (InSAR), where the change in phase between the images can be related to topography, and hence, DEM data can be derived.

Contents and goals

SliDEM seeks to explore the potential of Sentinel-1 InSAR for the generation of DEMs for landslide assessment, in particular for volume estimations, within a semi-automated and transferable workflow.

SliDEM aims to create an open-source Python package for automatically downloading and archiving Sentinel-1 data, assessing the suitability of each image pair, producing DEMs using InSAR, as well as performing necessary post-processing, such as co-registering pre- and post-event DEMS together to quantify the landslide volumes, as well as validating the DEMs against reference data.


We will evaluate and validate our workflow in terms of reliability, performance, reproducibility, and transferability over several key landslides in Austria and Norway. SliDEM represents an important contribution to the field of natural hazard research by developing a low-cost, transferable, and semi-automated method for DEM generation and landslide volume estimation that analyses the technical challenges related to SAR imaging and environmental characteristics of the study areas.

Expected results

Based on the outcomes of this exploratory project, we will develop a roadmap towards a larger research, development, and innovation (RDI) project.


Project Partners


University of Salzburg Department of Geoinformatics – Z_GIS

Project partner

University of Bergen

Contact Address

University of Salzburg Department of Geoinformatics – Z_GIS
Mag. Daniel Hölbling
Schillerstr. 30
A-5020 Salzburg