Satellite Interferometry Data Interpretation and Exploitation: Case Studies from the European Ground Motion Service (EGMS)

By Michele Crosetto and Lorenzo Solari

Satellite Interferometry Data Interpretation and Exploitation: Case Studies from the European Ground Motion Service (EGMS) focuses on the interpretation and exploitation of data obtained from InSAR, thus enabling millimeter-scale deformation measurements from space. The most emblematic InSAR service, the European Ground Motion Service (EGMS), opens a wide range of new applications. However, for effective use of raw data, interpretation techniques and methods are required. This book presents interpretation protocols that can be applied to any InSAR data, as well as the most relevant technical aspects and boundaries of measurement points. Detailed case studies are reviewed to demonstrate points.

This book will be a valuable resource for remote sensing specialists, as well as non-specialists in geotechnics, geology and other geosciences who are looking to apply InSAR data techniques in their research.

• Provides user-oriented techniques for better understanding of the applications of InSAR and the European Ground Motion Service (EGMS)
• Features case studies based on manipulation of EGMS data, thus showing new applications of InSAR data interpretation
• Details InSAR and EGMS potential and limitations for the exploitation of InSAR data

• Language: English
• Publisher: Elsevier Science
• Release date: Jun 15, 2023
• ISBN: 9780443133985

Michele Crosetto

Michele Crosetto is head of the Geomatics Division at the Centre Tecnològic de Telecomunicacions de Catalunya, Spain. He previously worked in the Joint Research Centre of the European Commission in Italy and as a researcher at the Cartographic Institute of Catalonia. His main research activity is related to the analysis of spaceborne, airborne, and ground-based remote sensing data and the development of scientific and technical applications using active sensor types. In the last few years he has been involved in several projects of the 5th, 6th and 7th, and H2020 Framework Programmes in the EU. In addition, he has been involved in different projects funded by ESA including Terrafirma Validation Project for which he was the technical coordinator. He is the coordinator of the Advisory Board of the European Ground Motion Service and is associate editor of the ISPRS Journal of Photogrammetry and Remote Sensing (Elsevier).

Book preview

Chapter 1

Introduction

Abstract

This chapter introduces the content of the book. It starts with the focus and motivation of this work, which is devoted to the correct interpretation and exploitation of the data obtained from the interferometric synthetic aperture radar (InSAR) technique. The focus is on the InSAR data coming from the Copernicus European Ground Motion Service (EGMS). The book contributes to understanding what InSAR and EGMS offer, makes the reader aware of their potential and limitations and illustrates case studies based on the EGMS data. The second section addresses the range of intended book readers. This chapter finalizes with the description of the content of the book.

Keywords

Radar; InSAR; deformation; European Ground Motion Service; interpretation; exploitation

1.1 Motivation

This book is about interferometric synthetic aperture radar (InSAR), a remote-sensing technique to retrieve information from multiple satellite radar images acquired over the same area, yielding millimeter-scale ground deformation measurements from space. In the book, InSAR encompasses the terms differential InSAR (DInSAR), advanced DInSAR, time series InSAR, multitemporal InSAR, persistent scatterer interferometry, etc. In addition, the term deformation is used as a synonym of displacement.

To be more precise, the book is focused on the interpretation and exploitation of the data obtained from InSAR. What is the main driver behind this book? First, InSAR is a clear example of technology push. Since its first description (Gabriel et al., 1989), the technique has undergone intense research and development that has produced advanced data processing and analysis tools. At the same time, several space agencies, especially the European Space Agency, have launched and operated missions carrying SAR sensors, guaranteeing the constant availability of InSAR primary data. In parallel to this, the computational capability has grown significantly. All these factors make InSAR mature enough to deliver deformation measurement on which to build up monitoring services that cover wide areas. The most emblematic of such services is the European Ground Motion Service (EGMS), which is the protagonist of this book. As it will be described later, the service opens a wide range of new applications based on EGMS products. It is worth noting that these products are solely InSAR deformation data, that is, they do not provide any interpretation of the causes or effects of the motion observed. In other words, they act as the starting point for investigations into the underlying causes of movement. This involves added-value activities based on such products, which in turn require correct InSAR data interpretation and exploitation. This book aims to contribute to this effort.

Therefore this text will attempt to meet these objectives:

• Understand what InSAR and EGMS offer.

• Realize their potential and limitations.

• Illustrate case studies based on the EGMS, promote data interpretation and exploitation, and ease the development of new applications.

1.2 Target readers

The potential audience is broad, encompassing those who are connected to known InSAR applications, as well as those related to new potential applications of InSAR and EGMS. For this reason, the book does not require any specific a priori knowledge or technical background. In the following section, we list some of the most relevant types of readers. The first group would be the scientists and technicians working in the field of deformation monitoring of land, structures, and infrastructures. The second includes a wide spectrum of civil engineers, structural engineers, architects, and technicians working in the management of buildings, cultural heritage, linear infrastructures (roads, highways, railways, canals, dykes, levees, pipelines, etc.), bridges and viaducts, dams, airports and ports, industrial installations, etc. A third class includes technicians involved in the planning and management of major construction works, for example, tunnels and large excavations. A fourth and still significant group comprises the scientists and technicians working with phenomena that cause subsidence and uplift: hydrogeologists, people involved in gas and hydrocarbon extraction, mining engineers, geotechnicians, geologists, and geoscientists. A fifth class would be made up of professionals involved in the detection and monitoring of landslides and unstable slopes: geotechnicians, engineering geologists, geologists, geoscientists, etc. A sixth area covers people involved in the insurance industry, where the motion of buildings and industrial assets matters a great deal. The seventh group consists of scientists from different fields of geophysics, like tectonics, crustal deformation, vulcanology, glaciology, etc. Last, an open and potentially wide-ranging class encompasses anyone working with any other application that is directly or indirectly concerned with the deformation of land, structures, infrastructures, assets, etc.

1.3 Content of the book

The book is organized as follows. It includes two main parts. The first part introduces the key concepts related to InSAR and EGMS, while the second part discusses several case studies based on the EGMS products.

More in detail, Section 2.1 starts with the InSAR basics and the main InSAR observation equation. It then describes the InSAR measurement points, which are classified into two families: the point-like and the distributed scatterers. Then, Section 2.3 explains how the deformation is estimated with InSAR. This section refers to a general approach that is related to the experience of the authors, which includes input data, image coregistration, interferogram generation, selection of the measurement points, phase unwrapping, atmospheric component estimation, densification of the measurement points, and geocoding. Section 2.4 outlines the two main InSAR products: the deformation velocity and the deformation time series. Section 2.5 describes the main advantages and disadvantages of the InSAR technique. Finally, Section 2.6 lists some of the most important InSAR applications.

Chapter 3 discusses the most important InSAR technical aspects. These are needed to perform a correct interpretation and exploitation of the InSAR products. The chapter starts with acquisition of the SAR data, and then it discusses the nature of the measurement points, and the density of such points. Section 3.4 describes the SAR geometric effects, while Section 3.5 discusses the implication of the line-of-sight InSAR measurement. Section 3.6 describes the relative nature of the InSAR results, and the issue related to the selection of the reference point. Section 3.7 outlines the characteristics of the InSAR results in the presence of nonlinear and fast deformation. Section 3.8 discusses the content of the deformation time series, while Section 3.9 describes the thermal expansion component of such time series. Section 3.10 reports the issue of positioning of the InSAR measurement points (geolocation). Section 3.11 details the quality of the InSAR estimates, and Section 3.12 addresses the validation of the InSAR results. Section 3.13 deals with the use of artificial reflectors, while Section 3.14 discusses the comparison of InSAR results and data coming from in situ measurements. Section 3.15 introduces some important postprocessing tools to analyze the InSAR results. This chapter ends with a description of available open-source InSAR software.

Chapter 4 is entirely devoted to the EGMS. It starts with the main features of the service. It then discusses in detail the three main EGMS products, that is, Basic, Calibrated, and Ortho. Section 4.2 describes the Basic product, its main characteristics, and the key issue of the spatial and temporal reference for the deformations. Section 4.3 reports on the Calibrated product, which is considered the star product of the Service. The section introduces its main characteristics and outlines the calibration procedure based on InSAR and Global Navigation Satellite System (GNSS) data. Section 4.4 describes the Ortho product, its features, and the estimation of the deformation components. Section 4.5 outlines the EGMS validation activities, while Section 4.6 discusses the applicability of the EGMS results. Section 4.7 introduces the EGMS explorer, which is the platform to view and distribute the EGMS products. It includes a WebGIS and an interface to search and download the EGMS products. Finally, the chapter finalizes by addressing the EGMS dissemination.

Chapter 5, which is the first chapter of the book dedicated to real data examples, discusses some case studies related to subsidence and uplift. Section 5.1 analyzes the subsidence related to groundwater exploitation in the Firenze–Prato–Pistoia basin (Italy). It reports the deformation studied with old SAR imagery (ERS and Envisat) and complements this with the results from the EGMS. It analyzes in depth some areas that are particularly relevant for the high deformation rates and the temporal evolution of the motion. Section 5.2 considers the mining subsidence in the Upper Silesian Coal Basin (Poland and Czech Republic). It includes the discussion of deformations measured with old SAR imagery and with EGMS. It performs the analysis focused on specific areas of interest. Section 5.3 provides tips and tricks to interpret interferometric data in mining areas. Finally, Section 5.4 treats a miscellaneous of subsidence cases, which include gas extraction, an airport area, other types of mining activity, and a geothermal field.

Chapter 6 is devoted to landslides. Section 6.1 analyzes a case of evaluation of a landslide state of activity in the Granada coast (Spain). Section 6.2 considers landslide mapping in the Troms og Finnmark county (Norway). The analysis discusses in depth different deformation areas, which include some fjords. This section specifically addresses the potential limitation in the usage of calibrated data in areas with strong GNSS signal. The chapter finalizes with a discussion on the use of EGMS data for landslide studies.

Chapter 7 presents some examples of EGMS data in the context of volcanoes and earthquakes. Section 7.1 is devoted to volcanoes. It describes in detail the emblematic case of Mount Etna (Italy), the largest active onshore volcano in Europe. Then it considers the Campi Flegrei (Italy), a peculiar and unique active volcanic system in a densely urbanized area. Section 7.2 is focused on earthquakes. It describes the 2016 Central Italy seismic sequence and shows how InSAR can measure the coseismic deformation of an earthquake and how major earthquakes impact a time series of deformation.

Chapter 8 aims at demonstrating the usage and drawbacks of EGMS products for urban area investigation, with a focus on infrastructure, groups of buildings, and cultural heritage. This chapter only includes local-scale examples. It starts in Section 8.1 with the dam and reservoir of Rules (Spain). Section 8.2 is focused on the airport of Nice Côte d’Azur (France). Section 8.3 is devoted to the railway bridge of Blackfriars, London (United Kingdom), complemented by a railway case study in Finland. Section 8.5 analyzes the levees of the Rhine River in Bregenz (Austria). Section 8.6 considers the Port of Antwerp (Belgium), while the following section is focused on the fishing town of Thyborøn (Denmark). Section 8.8 shows an example of cultural heritage, the UNESCO world heritage of the historic city center of Sighişoara (Romania). The chapter concludes with a second example of cultural heritage, which considers the archeological site of Solnitsata-Provadia (Bulgaria).

The book finalizes with the conclusions that sum up the book and provide a list of lesson learnt and take-home messages from the EGMS use cases.

Disclaimer

The views expressed in book are solely those of the authors and its content does not necessarily represent the views or position of the European Environment Agency.

Reference

Gabriel et al., 1989 Gabriel AK, Goldstein RM, Zebker HA. Mapping small elevation changes over large areas: differential radar interferometry. Journal of Geophysical Research: Solid Earth. 1989;94(B7):9183–9191.

Chapter 2

Synthetic aperture radar interferometry

Abstract

This chapter introduces the basic concepts of interferometric synthetic aperture radar (InSAR) for ground deformation measurement. Although often the text refers to the Sentinel-1 onboard sensors, most of the concepts in this chapter are valid for any type of SAR data. This chapter starts with the InSAR basics. Then it discusses the InSAR measurement points. The third section describes a generic step-by-step procedure to derive deformation estimates, starting from the input data up to the final geocoded InSAR products. The fourth section describes the two main InSAR products: the average deformation velocity and the deformation time series. The fifth section discusses the InSAR pros and cons. The chapter ends with a concise description of the more important InSAR applications.

Keywords

InSAR; fundamentals; deformation; estimation procedure; pros and cons; applications

The goal of this chapter is to introduce the basic concepts of InSAR. Although the Sentinel-1 sensors are mentioned in several sections, most of the concepts in this chapter are valid for any type of SAR sensor.

2.1 Interferometric synthetic aperture radar basics

As stated in Chapter 1, interferometric SAR (InSAR) provides deformation measurements from a set of SAR images. How is this done?

A SAR is an active radar sensor, which emits microwave signal and records the signal backscattered by the illuminated area. A complex SAR image contains several million picture elements (pixels), each of which contains two values. The first is the amplitude, which is related to the electromagnetic energy backscattered toward the radar by the given pixel footprint on the ground. The second is the phase , which is related to the distance between the sensor M and the same pixel footprint P along the radar line-of-sight (LOS):

(2.1)

where MP is the sensor-to-footprint distance, is the radar wavelength, and is a phase component introduced during the interaction between the microwaves and the footprint P.

Let us assume that the footprint P moves from P to P′ (see Fig. 2.1). SAR interferometry requires at least a second acquisition of the same scene. Let us assume, the footprint P′ is observed from a second viewpoint S:

(2.2)

Figure 2.1 Scheme of the DInSAR deformation measurement. DInSAR, Differential InSAR.

The phase difference, which is called interferometric phase, is given by:

(2.3)

Eq. (2.3) can be written as:

(2.4)

The first term is , which depends on the topography of the observed scene. The second term is the deformation phase component , where . Let us assume that the last two terms cancel each other out. The component can be simulated using a digital elevation model (DEM) of the observed scene, obtaining . This can then be subtracted from the interferometric phase, obtaining the so-called differential interferometric phase:

(2.5)

According to this equation, the differential interferometric phase can be directly exploited to estimate the deformation PP′. This is a simplified equation. In fact, a more comprehensive equation includes:

(2.6)

where is the residual topographic component due to error in the computation of , is the atmospheric phase component for M and S, due to the propagation of microwaves through ionosphere and troposphere, and is the phase noise. In Eq. (2.6), includes the phase component due to orbital error of each image. The last term is due to the ambiguous nature of the observed phases, that is, the fact that they are bounded in the range (-π, π]. is an integer value called phase ambiguity. All the components are measured in the radar LOS, which comprises the imaginary line that connects the sensor and the footprint on the ground.

Eq. (2.6) is the main InSAR observation equation.¹ Starting from the SAR images, for each pixel is obtained. From this observed value, the estimation of is derived. To do so, this component must be separated from the others. Estimating deformations is not a straightforward task. It can only be resolved by making assumptions and deploying appropriate estimation procedures. This is a relevant factor that must be considered when exploiting any InSAR result. The estimation procedures are discussed in Section 2.3.

2.2 Interferometric synthetic aperture radar measurement points

Let us call measurement point (MP) a pixel where Eq. (2.6) can be resolved to allow to be estimated. A SAR system performs a regular and dense sampling of the observed scene.² However, the MPs are usually irregular and much less dense than the original SAR images. In other words, not all the pixels of a SAR image correspond to a measurement point. This is an important aspect. Let us work through an example to illustrate this property: a Sentinel-1 image has a pixel footprint of approximately 14 by 4 m: this equals 17860 pixels/km². A deformation map can typically have 5000–8000 MPs in urban areas, and fewer than 1000 MPs in agricultural areas (Larsen et al., 2020).

Why is there such a significant reduction in MP density? The reason can be found in the term of Eq. (2.6). This component is the main source of ambiguities of the phase unwrapping operation, thus the can only be estimated if is small enough.

What are the typical InSAR MPs? They belong to two main families of pixels. All of them have a constant component over time, hence providing a coherent response, that is, they are coherent targets.

• The first family is given by the point-like scatterers (PS), where the response to the radar wavelengths is dominated by a strong reflecting object located within the pixel footprint, that is constant over time (Ferretti et al., 2000, 2001). The same PS acronym is used to refer to such scatterers as permanent or persistent scatterers, emphasizing their constant response over time. Typical examples of PS are poles, antennas, fences, metallic objects in general, or objects with sharp edges, such as parts of buildings and man-made structures, rock outcrops, etc. These objects typically make good MPs.

• The second family is made up of the so-called distributed scatterers (DS). They maintain a constant response over time, which is due to different small scattering objects distributed within the pixel footprint, without the presence of a dominant scatterer (Ferretti et al., 2011). The response of DS is weaker than that of PS; however, their information content can be improved by spatially averaging neighboring pixels that show similar properties. DS can be found in bare soil, homogeneous ground, debris, and desert areas.

Many InSAR approaches only exploit PS. However, the most advanced techniques can exploit both PS and DS. It is worth noting that adding DS to PS offers an improvement of the MP density, which depends on the land cover. For instance, this improvement is rather modest in urban areas, but it can be remarkable in different types of nonurban areas.

2.3 A deformation estimation procedure

How is deformation estimated with InSAR? There is not a straightforward approach to doing it. In fact, during the last two decades, there has been intense research and development in this field, which has yielded several InSAR approaches. To cite just a few: Ferretti et al. (2000, 2001), Berardino et al. (2002), Mora et al. (2003), Crosetto et al. (2005), Costantini et al. (2008), Hooper (2008), Ferretti et al. (2011), Perissin and Wang (2011), and Devanthéry et al. (2014). The different approaches are not discussed in this book. Rather, a general approach is discussed that is related to the experience of the authors. A similar approach that includes the main characteristics of the InSAR techniques used in the production of the European Ground Motion Service (EGMS) is described in Ferretti et al. (2021). The flow chart of the general InSAR approach is shown in Fig. 2.2. Each step is discussed