Advanced interferometric SAR processing techniques are able to detect and monitor various surface displacements caused by e.g. soil compaction, soluble and swelling rock formations, groundwater extraction, natural gas extraction, mining, cavern storage operation and landslides. These processes can cause damage to buildings and infrastructure, affect ecosystems and agriculture or affect the economic use of the geological underground e.g. by influencing the hydrogeological setting. Despite the maturity of the PSI technology it is rarely used in operational workflows of the German user community (e.g. from responsible authorities). Nevertheless, SAR data availability is a precondition for the operational use. From this perspective, Copernicus Sentinel-1 is a game changer, ensuring SAR data availability for almost the entire Earth, until at least 2030. In order to support the operational use of this technique the “Ground Motion Service Germany” is currently established by the Federal Institute for Geosciences and Natural Resources (BGR).

The service uses Persistent Scatterer Interferometry (PSI) datasets of the entire nation (approx. 360,000 km²). The PSI datasets are based on Sentinel-1 ascending and descending acquisitions. Continuous Global Navigation Satellite System (cGNSS) measurements are used to calibrate and validate the PSI datasets. In order to cover the entire nation, the mapping coverage is increased by mosaicking PSI datasets from adjacent SAR image stacks. Numerous applications, pilot studies and in depth case studies are performed to verify the PSI datasets and to improve the usability. This presentation gives an overview of the Ground Motion Service Germany, shows the characteristics of the nationwide PSI datasets and finally highlights the main findings of selected case studies. With respect to the characteristics the precision of the Line of Sight velocity, the geocoding precision and the spatial sampling density are shown. The presented case studies involve the monitoring of potential landslide areas on a regional scale and subsidence processes in coastal lowland areas. Value added products are generated to support the usability of the Line-of-Sight measurements. For this reason 2D-motion decomposition to estimate the vertical and East-West velocity vectors and cluster analysis to automatically detect displacement areas are applied. These products are demonstrated and discussed with respect to the case studies.

3vGeomatics (3vG) has developed a wide area landslide monitoring system using interferometric Synthetic Aperture Radar (InSAR) techniques. This system utilizes Sentinel-1 data to cover large areas and RADARSAT-2 data for high resolution follow up. 3vG has analyzed the entirety of the Canadian province of British Columbia (BC), a roughly 1 million square kilometer area containing numerous different topographical areas, including mountain ranges, dry valleys and discontinuous permafrost.
A historical landslide inventory of BC was created using ALOS-1 data. In 502 individual stacks, 6675 individual ALOS images were automatically processed at 18 m resolution with only 2 weeks of processing time. Over 3500 motion areas were detected with sizes greater than 6500 m² and are shown in Figure 1.
Sentinel-1 data for 2016 and 2017 were downloaded and separated into almost 400 stacks. Using a fully automated chain, 3vG processed all descending data at 15 m resolution, resulting in 48 Terabytes of data. Motion areas are currently being compiled and 3vG expects to identify a list of over 10 000 areas within the province.
Sentinel data provides full coverage of British Columbia with relatively short, 12 day, repeat times and thus provides the opportunity to not only detect historically moving areas but to also provide rapid detection of new moving areas and accelerating motion areas. With the Canadian Space Agency, 3vG is developing the basis for a landslide alerting system capable of sending out automated alerts upon the detection of an increasing risk of a landslide anywhere in the province, with minimal turnaround time from data acquisition.
To provide greater detail in targeted displacement regions, higher resolution SAR satellites such as RADARSAT-2 are used to further investigate motion areas detected by the Sentinel-1 processing chain. To demonstrate this capability, RADARSAT-2 stacks with 4 m resolution have been processed over four confirmed motion areas within BC.
Overall, 3vG has a demonstrated capacity to survey large areas and identify tens of thousands of small landslide prone areas from historical data. 3vG can currently provide high resolution monitoring of select hazards with reports generated within 24 hours of data acquisition and is rapidly developing the capability to generate alerts for landslide areas becoming active shortly after data acquisition. With Sentinel, this alerting system can be deployed for areas around the globe.

Figure 1: Confident motion areas (in red) detected with ALOS data overlaid onto an OpenStreetMaps background, showing the relation between discovered displacement areas, and areas where infrastructure may exist.

Located on a hot spot and the Mid-Atlantic Ridge, Iceland experiences many deformation processes. Eruptions are frequents in Iceland, the last one resulting in 60 m caldera collapse in the Báðarbunga volcano in 2014, and many volcanoes shows signs of activity. Slope instabilities leading to landslides are also a major concern, with large rockslides/landslides occurring in 2014 and 2018. Earthquake of magnitude 6 or more can occurs like in 2000 and 2008 in south Iceland. Therefore having country-wide observations of deformation processes in Iceland is extremely valuable for monitoring all these potential natural hazards. Sentinel-1 interferometry is used to derive approximate velocity fields 2015-2018 for east and vertical components of ground deformation (referred to as near-East and near-Up components), over 81% of Iceland, in about 33 millions ground pixels of 100x100 m size, that remained coherent over the study period . Sentinel-1 images of six different tracks were analyzed: three ascending and three descending, selected so that every part of Iceland is at least covered by one ascending and one descending track. We only used images acquired in summer to avoid snow cover, amounting for about 50 images per track for 2015-2018. We estimated the average line-of-sight (LOS) velocity fields for 2015-2018 for each track using a simple approach: single master interferometry time-series together with an estimate of linear component for each coherent pixel. Single interferogram were formed using the InSAR Scientific Computing Environment (ISCE) software developed by jointly by the Jet Propulsion Laboratory, Caltech and Stanford University. The time-series analysis was done using in-house tools. There is a good agreement between inferred LOS velocity fields and deformation measured by the continuous GPS network of Iceland. The LOS velocity fields are combined to extract near-East and near-Up velocities. Only pixel where the average coherence is over 0.4 in all interferograms except the ones where the slave image is in the same year than the master (2017) are kept. The 19% of missing coverage is due to large surface changes between summers. In Iceland, areas subject to such changes are mostly glaciers, agricultural land, and unconsolidated sand deserts. The near-East velocity field is compared to plate spreading models and the near-Up velocity field to glacial isostatic adjustment (GIA) models of response to ice retreat since 1890. Good agreement between the data and the models is generally found, but residuals suggest some imperfection in the models, potentially related to their assumption a uniform rheology in these models instead of rheology. Improved models could use rheology allowing for lateral variations, relating to hot spot – oceanic ridge interaction in Iceland. Numerous smaller scale deformation signals are clear in the velocity fields: post-rifting viscous relaxation following the 2014 Bárðarbunga rifting events, volcano uplift and subsidence, slope instabilities, and signals related to geothermal processes and fluid extraction. Sentinel-1 provides an invaluable dataset to monitor natural hazard, offering possibilities to observe deformation over all of Iceland every 6 days.

The June 2017 landslide and consequent tsunami in the Karrat Fjord, western Greenland, spurred the Greenlandic and Danish authorities to undertake the first large-scale efforts to map slope-instabilities and assess the risk of landslides, in particular concerning their tsunamigenic potential. In this context the Geological Survey of Denmark and Greenland (GEUS) has been leading a multi-disciplinary project over the past year, which included assessing the contribution of the rich Sentinel-1 data archive routinely acquired over the Greenland ice-sheet margin.

In this work we describe an observation strategy based on deformation measurement and change detection algorithms to map the evolution of slope-instabilities in Greenland, and discuss its application to the Karrat fjord area, using Sentinel-1 data acquired from Oct. 2014 to the present date.

Coherent and incoherent change detection methods are successful in flagging the occurrence of catastrophic events, e.g. rockfalls, which are associated with local loss of interferometric coherence and appreciable changes in topography and thus mean backscattered intensity.

Concerning deformation, the approaches, which are currently used at mid-latitudes for landslide studies are found to be only partially applicable. The latter are based on the application of multi-temporal SAR data processing algorithms, followed by post-processing steps to identify measurement clusters, which are consistent in space (hotspots) and persistent in time. In central-western Greenland the application of these methods is typically restricted in time to the snow-free season (June to October) and in space, to ice- and snow-free areas exhibiting strain rates small enough to be measured with interferometric techniques. Surprisingly however, owing to the short 6-day repeat cycle of the Sentinel-1a and 1b satellites, some coherent time-windows can be found also in the winter season. This allows for the generation of single-pair differential SAR interferograms in which local deformations are often clearly distinguishable from larger-scale phase patterns due to atmospheric propagation, in particular stratified tropospheric delay. Finally, intensity offset-tracking can be attempted all year round and used to quantify large incoherent deformations, providing their spatial scale is sufficiently broad.

An additional challenge, common to mid-latitude SAR-based landslide applications, is related to the SAR viewing geometry with respect to slope and aspect angles. For a given area of interest and acquisition time, we implement an efficient procedure to map the resolution index (slant- to ground-range resolution ratio) and a sensitivity index defined as the projection of the SAR line-of-sight vector onto the steepest slope direction. These indices provide an a-priori indication of the feasibility and sensitivity of the SAR-based analysis from different tracks, and are later used to cull and/or project the line-of-sight deformation measurements along the steepest slope.

Natural disasters have increased in frequency and intensity on a global scale and have affected a large part of the world’s population, with consequent human and economic loss, and deterioration in quality of life. Countries in Latin America and the Caribbean have also been increasingly affected by natural phenomena such as earthquakes, landslide, volcanic eruptions, tsunamis, and hurricanes. Ecuador, country analyzed in the present work, is defined as one of the highest-vulnerable countries, being constantly challenged by destructive natural phenomena such as landslides and earthquakes, floods and droughts and volcanic eruptions. Remote sensing technologies are nowadays widely implemented for the monitoring and the investigation of natural hazard phenomena. The resilience of physical infrastructure after such events is thus a function of the level of design, construction, prevention and mitigation efforts. The present work illustrates the results of a large-scale landslides analysis of Cuenca city, located in the highlands of Ecuador, in the province of Azuay, along with the characterization of a specific slow-moving landslide, affecting the buildings of the University of Azuay, by integrating Very High Resolution Synthetic Aperture Radar (SAR) data with building structural condition data, in order to define their vulnerability. The city of Cuenca is constantly expanding, and for this reason the local people build in areas with steep slopes, at times within high-risk landslide areas. The high slopes, heavy rainfall, low resistance of the ground below the city of Cuenca, associated with a high seismicity of the city, makes its built scenario highly vulnerable to catastrophic events. Landslides affecting the area and their interaction with human structures and infrastructures, as evidenced by the damage observed to buildings, make the study of the inhabited center against the gravitational phenomenon particularly interesting. As regards the DInSAR analysis, COSMO-SkyMed images were acquired within the project related to the First Announcement of Opportunity Cosmo-SkyMed. The image dataset is referred to different frames, of at least 79 images, with a footprint of 40 × 40 km, acquired along ascending track between 2016 and 2018, and processed by means of Coherent Pixels Technique (CPT) (Mora et al., 2003; Iglesias et al., 2015). Along with field surveys, many unstable areas, previously unrecorded, were recognized, providing a first updated landslide inventory map. A test area was selected aiming at the analysis of the cause-effect relationship between landslide kinematic and building damage. The university of the Azuay is located in the south-east area of Cuenca, built around the 80’s on an unstable ground, affected by a slow movement that starts from the north of the study area. The analysis of the information collected has allowed to obtain general information about geology, geomorphology and structural properties. An innovative approach, in which the results of landslide movements analysis have been combined with a numerical ad analytical approach to evaluate the vulnerability of buildings, has been developed. The results allowed to generate empirical fragility curves for buildings on the study area, tools which can be used to forecast the future likelihood of the damage occurrence. Therefore, remote sensing tools, coupled with other kind of instruments can fully support the decision-making process when dealing with risk management operations.

The aim of this work is to present the ongoing project “U-Geohaz, Geohazard impact assessment for urban areas”. U-Geohaz is a two-year project, started the 1st of January 2018, co-funded by the European Commission, Directorate-General Humanitarian Aid and Civil Protection (ECHO). The general aim of the project is to continuously assess the potential impact of geohazards on urban areas and critical infrastructures, using SAR images acquired by Sentinel-1 (S-1). The launch of the satellites S-1 A and S-1 B, in 2014 and 2016 respectively, granted an important improvement in the use of SAR interferometry for risk management and reduction. Ensuring a SAR acquisition every 6 days above the entire Earth and making the images available few hours after the acquisition without costs and limitations, S-1 represent a reliable and low cost tool for the remote monitoring of local to national-scale areas. These characteristics are improving the exploitation of SAR interferometry not only by the scientific community but also by administrative entities and Civil Protection authorities. U-Geohaz builds on the experience and the results obtained in the previous ECHO project SAFETY (2016-2017), and it makes a step forward by evolving from the periodically updated maps to a near-real time monitoring and mapping of geohazards. In particular, the project will provide tools to support early warning systems for landslide and volcanic geohazards and to evaluate the expected damage. The U-Geohaz consortium is composed of 18 partners, from 11 European countries, including 12 Geological Surveys and 3 Civil Protection Authorities that will support the implementation of tools to be operationally useful in risk management. The use of all the developed products will be implemented in the activities of the Civil Protection Authorities involved in the project. The first results achieved over the two test-sites, Canary Island (Spain) and Valle d’Aosta region (Italy), and the future steps of the project will be presented.