Final Report Summary - DORIS (Ground Deformations Risk Scenarios: an Advanced Assessment Service)
Executive Summary:
DORIS is an advanced downstream service for the detection, mapping, monitoring, and forecasting of ground deformations caused by natural and human-induced phenomena. Natural phenomena investigated by DORIS include landslides, riverbank erosion, rock glaciers, and the melting of permafrost. Human induced causes include coal mining, the exploitation of groundwater for human and agricultural consumption, and construction works. DORIS integrates traditional and innovative Earth Observation and ground-based data and technologies to improve our understanding of the complex phenomena that result in ground deformations, at different temporal and spatial scales, and in various physiographic and environmental settings. DORIS integrates state-of-the-art technological and scientific capabilities with existing services to deliver innovative, user-driven services and products expected to be beneficial for different communities. Primary users of the downstream are Civil Protection authorities, and environmental and planning agencies operating at different geographical and organizational scales, from the local to the national scale. Secondary users of DORIS include the mining industry, builders, utilities, road and rail companies, and other businesses involved in the assessment or the mitigation of the risks posed by the ground deformations, including the insurance industry. DORIS advanced the state-of-the-art of the science and technology used to detect, map, monitor, and forecast ground deformations. New techniques were developed to exploit fully the unique ESA ERS-1/2 and ENVISAT C-band, Synthetic Aperture Radar (SAR) archives, obtaining time-series of ground deformations, and associated maps, of unprecedented length. The high-resolution X-band SAR data brought by COSMOSkyMed and TerraSAR-X satellite sensors were used to produce deformation maps and time series of unparalleled spatial and temporal densities. DORIS moved forward the multi-frequency joint analysis of DInSAR products obtained processing data captured by multiple sensors (C, X, L-band) for the same area, and the combined application of satellite and ground-based differential SAR interferometry, coupled with in situ thematic data and geophysical probing, for improved monitoring and modelling of landslides. DORIS processed more than 2000 SAR images in 13 study areas in Hungary, Italy, Poland, Spain and Switzerland. Collectively, the study areas represent a wide range of physiographical and environmental settings, and include all types of ground deformations for which the service is suited. This guarantees that the downstream service will work in Europe, and in other areas. Based on a thorough evaluation of the potential markets, and on a realistic sustainability assessment, DORIS designed a business model and an appropriate strategy for the long-term selfsustainability of the downstream service.
Project Context and Objectives:
Landslides and ground subsidence are serious and largely underestimated problems, in Europe and in other areas of the world. Where they occur, landslides and subsidence pose a major threat to public and private properties, structures and infrastructures, and the population, with significant economic, environmental, and societal consequences. A “landslide” is the movement of a mass of rock, debris, or earth down a slope, under the influence of gravity. Different phenomena cause landslides, including intense or prolonged rainfall, earthquakes, rapid snow melting, volcanic activity, and multiple human actions. Landslides can involve flowing, sliding, toppling, or falling, and many landslides exhibit a combination of two or more types of movements, at the same time or during the lifetime of a landslide. Landslides are present in all continents, and play an important role in the evolution of landscapes. They also represent a serious threat to the population. The International Disaster Database EM-DAT reports as many as 20,000 people killed in the last 20 years by major landslide events, with a major event killing ten or more persons. Every year, there are between six and 20 major landslide events in the world, with at least one of these major landslides occurring in Europe, on average. These figures are known to be vastly underestimated, and the number of landslides causing less than 10 fatalities and providing significant damage and disruptions is several orders of magnitude higher. In Italy, a country for which detailed information exists on the societal and economic impact of landslides, at least 5190 people were killed, went missing, or were injured by landslides in the 50-year period 1963–2012. In the same period, the number of evacuees and homeless people caused by landslides exceeded 150,000. The cost of geo-hydrological events in Italy, including landslides, in the period 1950–2012, was estimated at € 60-billion, with an average yearly cost exceeding € 2-billion in the last years. Although Italy is a country particularly affected by landslides and related geo-hydrological hazards (e.g. debris flows), the extent of the landslide problem is severe in several other European countries, including Austria, Czech Republic, Croatia, France, Greece, Hungary, Norway, Poland, Portugal, UK, Slovakia, Slovenia, Spain, and Switzerland. Ground subsidence has many natural and human-induced causes, and it is most severe and widespread in areas where mining and other sub-surface exploitation activities take place. According to Swiss Re, in some parts of Europe subsidence is now the costliest natural hazard, comparable to flooding. In Great Britain, over the last ten years subsidence has cost an estimated € 3.6-billion making subsidence the most damaging geo-hazard in the region. In the Ruhr region of Germany, a large number of litigation cases towards mining companies have emerged, with more than 35,000 damage claims filed each year. The claims are in the hundreds of millions of euros. In France, subsidence-related claims have risen by more than 50% in the past 20 years, costing the affected regions € 340-million every year, on average. In 2003 a single mining subsidence accident caused by a French mining company’s caused damage assessed at € 10-million in France and Germany. In Poland, coal mining causes vast subsidence and produces significant damage to public and private properties and the infrastructure. In Italy and Spain, excessive water withdrawal for human consumption and agriculture uses produce subsidence of large to very large geographical areas, producing widespread damage and considerable costs.
In this phenomenological and economic context, DORIS is an advanced downstream service for the detection, mapping, monitoring, and forecasting of ground deformations, including landslides and ground subsidence, at different temporal, spatial and organizational scales, and in various physiographic, climatic, and environmental settings. The downstream integrates traditional (consolidated) and modern (innovative) Earth Observation (EO) and ground based (non-EO) data and technologies to improve the understanding of the complex phenomena that result in ground deformations, chiefly landslides and ground subsidence, and to foster the ability of Environmental and Civil Protection authorities to manage the risks posed by ground deformations. DORIS delivers innovative products tailored on the needs of national, regional, and local authorities integrating state-of-the-art national technological and scientific capabilities. DORIS complies with general guidelines provided by the EU Emergency Response Core Services, and benefits from a unique partnership of leading research institutes, commercial providers, and experienced endusers.
The mission of DORIS was to establish an operational Copernicus (formerly GMES) downstream service capable of:
• Strengthening the operational exploitation of satellite data and technology for the detection, mapping, monitoring and forecasting of ground deformations caused by landslides and subsidence phenomena.
• Integrating effectively satellite and ground-based data and technologies to improve the understanding and forecasting of ground deformations caused by landslides and subsidence.
• Raising the awareness of landslide and subsidence and their actual and potential impacts, and of the possibility of using space data and technology to monitor and manage the ground deformations.
Specific goals of DORIS included:
• To design and deliver innovative products relevant to Civil Protection and Environmental authorities, for all the phases of a typical risk management cycle.
• To exploit the unique ESA ERS-1/2 and ENVISAT, C-band Synthetic Aperture Radar (SAR) archives, to generate very-long time-series of ground deformations, and associated deformation and deformation velocity maps.
• To evaluate quantitatively the improvements brought by the high-resolution COSMO-SkyMed and TerraSAR-X, X-band SAR sensors, for the production of innovative DInSAR products.
• To explore the combined application of satellite and ground-based DInSAR technologies and products for the improved monitoring of local-scale landslides and subsidence phenomena.
• To define and test innovative, quantitative validation procedures for an improved exploitation of DInSAR services and products. Lacks of standards for the assessment of the quality of DInSAR products, including deformation maps and time series, limits the use of the services and products, and the thrust of the endusers on the consolidated and the innovative services and products.
• To design an efficient interface between the “user domain” (represented by a diversified audience of end users with different levels of knowledge and expertize), the European level core services, and the DORIS downstream, exploiting national resources and specific technical expertise on ground deformations.
• To execute innovative research on the use of remote sensing technologies for the detection, mapping, monitoring and forecasting of landslides and subsidence.
• To evaluate private and public markets for the DORIS services and products, and to design an appropriate business model for the long-term sustainability of the downstream. DORIS has obtained a number of significant results, which are of potential interest to a large and diversified audience of service providers, operational users, and research scientists.
• For 13 test sites in Europe, DORIS produced accurate deformation and deformation-velocity maps, and associated time series of unprecedented length obtained processing very-large stacks of C-band SAR data captured by the ESA ERS-1/2 and ENVISAT satellites. We expect that service providers can use the new DInSAR techniques experimented by DORIS to offer innovative services and products to a range of costumers. Operational users can exploit the accurate deformation maps and the very long time series to better understand the ground deformations and the drivers of the deformations. Scientists have new information to study the long-term behaviour of landslides and subsidence. This is particularly relevant to investigate the relationships between precipitation and ground movements, and the expected impact of the on going or the predicted climatic and environmental changes on the pattern and rate of the natural and the human induced deformations.
• Extensive processing of X-band SAR data taken by the Cosmo-SkyMed and TerraSAR-X satellites, exploiting the significantly higher spatial resolution and the considerably shortened repeat time of the new advanced sensors. We expect that service providers will expand their portfolios offering new services and products based on the advanced processing of the high resolution X-band SAR data. Operational users will benefit from the improved temporal and spatial resolutions of the X-band sensors, being able to use significantly denser and more accurate deformation maps and time series. This will prove particularly useful in urban and sub-urban areas and along major infrastructures. Scientists can use deformation maps and associated time series of unmatched spatial density to study in greater detail the kinematics of landslides and subsidence phenomena. We further expect that operational users and scientist will use the denser maps to constrain and/or validate numerical (e.g. FEM) models of landslides and subsidence phenomena.
• Innovative use of SAR interferometry to investigate fast moving slope processes, including landslides, glaciers, rock glaciers, areas with permafrost, and human induced subsidence in active mining areas. The service providers can offer new services experimented by DORIS to monitor rapid ground deformations. These previously unavailable services and products are of potential interest to operational users and scientists that can use new tools to detect, monitor, and investigate active landslides and subsidence phenomena moving at fast to very-fast rates that previously could not be investigated effectively using SAR data and related processing technologies.
• Novel validation procedures for DInSAR services and products. Service providers can offer better-validated products, and operational users benefit from the use of validated products. We expect that these results will contribute to expand further the use and the confidence on DInSAR-based services and products by a large audience of end-users.
• Innovative framework for the semi-automatic recognition and mapping of fast moving landslides exploiting optical (multispectral) images. Service providers can now offer innovative rapid mapping services, which are particularly effective for mapping event landslides. Operational users and scientists can obtain accurate landslide maps shorty after an event. The capabilities for the semi-automatic recognition and mapping of event landslides are further expanded in the FP-7 LAMPRE project (www.lampre-project.eu/ ).
• An assessment of the sustainability of the downstream service has shown that, if certain conditions were met, DORIS could evolve towards a successful operational service. The DORIS services and products were initially planned for a few categories of users, chiefly Civil Protection and Environmental authorities, local government agencies, national and regional governments and municipalities. These remain the primary users of DORIS. There are other entities potentially interested in the DORIS benefits, including e.g. mining industry, builders, utilities, road and rail companies. These are secondary users in the DORIS value chain.
Project Results:
1.3.1 DORIS service & products
DORIS integrates consolidated and modern Earth Observation (EO) and ground based(i.e. in situ, non-EO) data and technologies to offer a number of innovative services and products for the detection, mapping, monitoring, and forecasting of ground deformations. The services & products were designed based on the requirements provided by a broad and heterogeneous community of users and experts, and the current GMES operating structure. The DORIS services & products include:
• InSAR and DInSAR processing of SAR data captured by all the available satellite SAR sensors, including C-band ERS-1/2 & ENVISAT data and RadarSAT-2 data, L-band ALOS data, X-band Cosmo-SkyMED and TerraSAR-X data.
• Multi-frequency DInSAR processing of ERS-1/2 & ENVISAT data, to produce very-long time series of deformations, spanning almost two decades. The time series of deformation and the deformation velocity maps are useful to determine the long-term behaviour of the deformations forced e.g. by climatic and environmental changes, or by human activities.
• Advanced DInSAR processing of X-band data, to produce dense deformation maps and deformation velocity maps, and associated time series, in urban and sub-urban areas, and of critical / relevant structures and infrastructures.
• Numerical modelling of ground deformations caused by landslides and subsidence phenomena, exploiting DInSAR deformation maps and surface and sub-surface morphological, geological and geotechnical data and information.
• Advanced InSAR processing for the detection and measurements of large and fastmoving ground deformations. The service proves useful to study e.g. the motion of rock glaciers and the deformations caused by the melting of the permafrost in highmountain areas, or the large deformations caused on the surface by active underground mining or the exploitation of reservoirs (oil, gas, water).
• Systematic update of existing landslide inventory maps, exploiting DInSAR deformation velocity maps obtained processing SAR data with different wavelength (C-band, X-band, L-band). The combined analysis of DInSAR products obtained using different bands allows for a better determination of the state of activity of the landslides.
• Advanced validation methods to improve the exploitation of multi-sensor DInSAR techniques for landslides and subsidence mapping and monitoring. The methods target the generation of landslide and subsidence activity maps, and are based on the combination of remote sensing data and independent in situ, non-EO data used as “ground truth”. The scope of the new methods is to demonstrate the reliability of the measured DInSAR displacements as reliable indicators of landslide or subsidence dynamics.
• Ground-based radar interferometry, exploiting terrestrial radar and SAR proprietary systems. Production of deformation maps, deformation velocity maps, and time series of deformations for specific sites subject to slow to fast moving terrain deformation phenomena (e.g. landslides, glaciers, rock glaciers, melting of permafrost). The combination of satellite and terrestrial DInSAR products proves very useful to investigate deformations caused e.g. by active landslides in mountain areas. It is also useful to complement damage assessment map, and for decisions on residual risk.
• Rapid mapping of event landslides, and production of event landslide inventory maps obtained by processing pre-event and post-event HR & VHR multispectral satellite images.
1.3.2 Geographical domain
DORIS was tested in different physiographical, administrative, and organizational settings in Europe (FIGURE 1). The DORIS test sites were selected to cover all the physiographical types where landslides, subsidence, and other surface slope processes, are common and relevant in Europe. The test sites were located in Hungary (3 sites), Italy (4), Poland (1), Spain (3), and Switzerland (2). TABLE 1 summarizes the main natural and human induced phenomena that are present in each study area, and that were investigated by DORIS. TABLE 2 lists the main services & products tested/delivered in the different study areas.
Given the geographical coverage and the variety of the test sites, we expect that DORIS downstream will work effectively in Europe, and outside of Europe in similar physiographical settings.
1.3.3 Technological effort
The technological effort in DORIS was considerable. We tested the DORIS service & products using all available SAR satellites, covering a range of wavelengths (C, X, and L bands), temporal periods, and spatial resolutions. TABLE 3 lists the types and temporal coverage of SAR data used to investigate landslides, subsidence, and other slope processes in the different DORIS test sites.
For the long-term (multi-decadal) analyses, we processed ERS-1/2 & ENVISAT data in 13 test sites, for about 2000 ERS-1/2 & ENVISAT scenes. We further processed 24 RadarSAT-2 scenes in Switzerland, and 150 L-band ALOS scenes in Spain, Switzerland, Hungary and Poland. For the advanced monitoring of urban and suburban areas, we processed 300 X-band scenes captured by the Cosmo-SkyMED constellation in Italy, Spain, and Switzerland, and 250 X-band scenes captured by TerraSAR-X in Hungary, Italy, Poland, and Switzerland. We estimate that the cumulated computational effort to process this large amount of SAR data was between 8,000 and 10,000 hours. This is between 333 and 417 days of computer time. This is an unprecedented effort for a downstream project.
1.3.4 Technological, methodological and scientific highlights of DORIS Through the massive processing of more that 2700 SAR scenes (TABLE 3), we prepared a total of 42 deformation maps, deformation velocity maps, and associated time series for 13 test sites, delivering more than 30-million point-scatters on the ground, with an average density of 830 points per km2. This is an unprecedented result for a project of this class. By itself, this is a significant result of the DORIS project.
In the following, the main highlights of DORIS are described. The description focuses on the technological, methodological and scientific impact of the results obtained. A first highlight consists in the production of very-long time series of ground deformation & associated maps obtained processing the unique archives of the ESA ERS-1/2 & ENVISAT SAR data. For the purpose, partners have developed specific, innovative Persistent Scatter Interferometry (PSI) techniques.
The original SBAS technique developed by CNR IREA was extended to exploit SAR data captured by different radar systems that acquire with the same illumination geometry, including the ERS-1/2 & ENVISAT systems. The new multi-sensor approach considers the images acquired by the ERS-1/2 & ENVISAT sensors as belonging to independent subsets, and cross-interferograms (ERS/ENVISAT) are not generated. The integration of ERS/ERS and ENVISAT/ENVISAT sequences of interferograms is performed at the stage of the generation of the time series of deformations, merging the two independent time series obtained by processing the ERS and the ENVISAT acquisitions. As a result, the SBAS technique can now be used to obtain deformation maps and associated time series spanning very long periods (decades), providing unprecedented information for studying long-term ground displacements.
The original PSInSAR technique developed by TRE was also extended in DORIS to analyse ERS-1/2 & ENVISAT data. In this case, the approach focuses on point targets that are expected to remain coherent in the ERS/ENVISAT interferograms, and whose phase history can be stitched coherently. To obtain this result, the technique estimates the term referred to as the Location Phase Screen (LPS), depending on the slant range sub-pixel position of the scatterers and the difference between the ERS-1/2 and the ENVISAT carrier frequencies (31 MHz). Given the frequency shift, the LPS changes by almost two cycles across a single slant-range resolution cell, and it is therefore extremely sensitive to the target slant-range position within the cell. The evaluation of the LPS contribution for each scatterer, and the deformation velocity and DEM error contribution, allows to remove it from the ERS/ENVISAT interferograms, allowing for the generation of time series of displacements spanning the entire period covered by the ERS-1/2 and the ENVISAT datasets.
The ability to produce very-long (> 10 years) time series of deformations is of particular interest for landslide investigations, as it allows preparing very-long time series of displacements for single or multiple points on the landslide topographic surface. This premium information is seldom available to landslide investigators, and proves very valuable. In the following, we describe an example of the generation and the exploitation of very-long time series of deformations for a landslide area.
For the Ivancich study area, in the Assisi Municipality, Italy (G in FIGURE 1), DORIS used 91 ERS-1/2 and 39 ENVISAT acquisitions taken between 21 April 1992 and 12 November 2010 to generate 347 interferograms, including 230 ERS/ERS and 117 ENVISAT/ENVISAT interferograms. Collectively, the 347 interferograms allowed determining the history of deformation in the landslide area, and in the surrounding regions, for a period of 18.6 years. Parts of the results are shown in FIGURE 2. The long time series of deformation allowed for important considerations on the history of deformation and the kinematics of the landslide.
Rainfall is a known driver for the Ivancich landslide. DORIS performed an analysis of the temporal relationship between the local rainfall record and the history of ground deformations in the Ivancich landslide area (FIGURE 2 B,D). We considered the surface displacement measurements obtained through DInSAR processing of C-band ERS-1/2 and ENVISAT images from April 1992 to September 2010, and rainfall measurements acquired by two rain gauges located 5.5 km W and 9 km SSW from the Ivancich landslide, for the same period (FIGURE 2D). The cross-correlation analysis was based on monthly-cumulated rainfall and revealed a lack of temporal correlation between the surface landslide deformations and the rainfall. The analysis was extended to consider the rainfall cumulated over different time periods (7, 15, and 90 days), obtaining crosscorrelation values < 0.2 confirming the lack of a clear relationship between landslide displacements and local rainfall history. Analysis of piezometer data acquired in the landslide deposit revealed that the groundwater regime affects moderately the landslide kinematics. In particular, the ground water surface was in general only a few meters above the known landslide shear zone. This resulted in very low piezometric heights, when compared to the total stress levels. Also, the piezometric surface was observed to be approximately constant in time, with limited seasonal fluctuations. This was a further indication of the limited influence of the rainfall pattern and of the slope groundwater regime on the kinematics of the Ivancich landslide, which is characterized by displacement rates approximately constant in time over long periods (FIGURE 2B).
Very-long time series of deformation were also prepared for a large sector of the Messina Province, NE Sicily, Italy (H in FIGURE 1), an area where show moving landslides are abundant, and active. ERS-1/2 & ENVISAT data collectively covering the 18-year period from 1992 to 2010 (FIGURE 1) were processed exploiting the stitching PSInSAR technique to produce deformation maps and average deformation velocity maps for an area of about 275 km2. For each persistent scatter, a very long time series of deformation spanning the period covered by the ERS-1/2 & ENVISAT data was produced, and exploited to investigate active landslides in the area. The geographical distribution and density of the measured points in the study area proved to be similar for the ERS-1/2 and the ENVISAT data, with a minor reduction of coherent points due to temporal decorrelation effects. The retrieved long-term deformation behaviour allowed for important geomorphological considerations on the history of deformation and the kinematics of several landslides in the study area.
Long time series of deformations were also prepared for other DORIS study areas, and specifically for the Dunaszekcső, Rácalmás and Hollóháza study areas in Hungary, for the Silesian Coal Basin, Poland, for the Tramuntana range, Mallorca, Spain, and for the Matter valley and the St. Moritz test sites, Switzerland (D and E in FIGURE 1).
However, for these areas the ERS-1/2 and the ENVISAT data were processed independently, obtaining separate deformation maps, deformation velocity maps, and associated time series. The maps and the times series proved valuable to investigate the long-term behaviour of the slope processes and the subsidence phenomena active in the different study areas.
A second highlight consists in the detection of very fast-moving, non-linear deformation patterns caused e.g. by active mining, and by seasonal movements of rock glaciers and the melting of permafrost in mountain areas. Typically, the rates of deformation of such phenomena are too large (mm to dm per day) to be measured by standard DInSAR techniques. DORIS has experimented the use of standard DInSAR techniques adapted to different natural and human induced phenomena to overcome this limitation, and to measure very-fast moving, non-linear patterns of deformation. To do this, DORIS benefited from the advanced X-band missions that, due to their higher temporal sampling of TerraSAR-X and Cosmo-SkyMED constellations, are better suited than the old-generation C-band sensors (i.e. ERS-1/2, ENVISAT, RadarSAT-2) to analyse the evolution of spatially and temporally highly non-linear ground deformations. This facilitates the detection of the trends induced by active mining and by seasonal movements of rock glaciers in mountain areas, by properly adapting to site-specific deformation signals the DInSAR techniques.
In the Upper Silesian basin, southern Poland, DORIS exploited TerraSAR-X data taken in the period 2011-2012 (TABLE 2) to study very-fast moving, non-linear patterns of deformation caused by extensive underground coal mining that produces surface subsidence in the range from a few centimetres per month to one centimetre per day, with strong accelerating and decelerating patterns. Short-term interferograms, spanning 11 or 22-day periods, were used to limit the temporal decorrelation, and to measure the very-fast moving, non-linear deformations. As a result, a significantly larger density of information – compared to ERS-1/2 & ENVISAT results (FIGURE 3) – WAS obtained, with more than 1.7-million measurements in the 950 m2 study area.Use of the X-band data confirmed the presence of high subsidence rates concentrated in rounded areas located above active mines. The information given by PSI processing was also supplemented by traditional differential 11-days interferograms. Combined data provided the possibility to make full interpretation of non-linear changes on the surface.
Rock glaciers are distinctive landform typical of cold regions and high mountain areas, including the Swiss sites of DORIS. They consist of either a glacier covered by a layer of talus or debris, or by a mass of coarse rock debris frozen in interstitial ice. Deformation of the surface of rock glaciers depends on the temporal (e.g. seasonal / decadal) cycles of deformation of the ice, in the glaciers or in the interstices and crevices. In high-mountain environments, the cyclic melting of ice and the freezing of liquid water, controlled by temperature, cause deformation of the permafrost i.e. of soils at or below the freezing point of water. Typically, the rates of deformation of rock glaciers and permafrost are too large (mm to dm per day) to be measured by standard DInSAR techniques. DORIS benefited of the shorter revisit time of X-band SAR sensors to produce small temporal baseline interferograms. Existing DInSAR techniques were extended to use only snow-free SAR scenes acquired chiefly during the summer season to limit the temporal decorrelation effects.
In the Matter Valley test site, in the Swiss Alps, (D in FIGURE 1), DORIS prepareddifferential interferograms using a large number of image pairs, including TerraSARX and COSMO-SkyMed pairs with time intervals of 4-33 days. Analysis of the single interferograms allowed detecting and measuring the activity of rock glaciers and landslides in the periglacial area (FIGURES 4 & 5). The fast-moving non-linear deformations could not be detected and measured by the traditional DInSAR techniques that are based on standard PSI analyses. DORIS showed that single interferograms, and series of single interferograms, obtained processing very-high resolution X-band SAR sensors, and particularly COSMO-SkyMed images with a spatial resolution of three meters, are well suited for the detection and monitoring of rock glaciers, and of permafrost sites. These unstable landform types have a typical size in the range from a few tens to several hundreds square meters, and require veryhigh resolution sensors to be detected and measured.
A third highlight consists in the extensive use of the high spatial and temporal resolution X-band data for the detection and mapping of deformations in urban and sub-urban areas with unprecedented spatial detail. As an example, for the Ivancich study area, in the Assisi Municipality, Italy (G in FIGURE 1), DORIS used 39 COSMO-SkyMed acquisitions taken between December 2009 and February 2012 along descending orbits, to generate 110 interferograms collectively covering the 26- month period, and 38 TerraSAR-X acquisitions taken between July 2011 and January 2013, also along the descending orbit, to generate 108 interferograms collectively covering a 19-month period. Visual inspection of the results obtained processing the X-band, COSMO-SkyMed data (FIGURE 6A), and the results obtained processing the C-band, ERS-1/2 & ENVISAT data (FIGURE 6B), reveals that the two analyses resulted in very similar spatial patterns of deformations, but a significantly larger density of coherent points detected by the COSMO-SkyMed data, compared to the ERS-1/2 & ENVISAT data. Processing of the shorter X-band data series resulted in the identification of > 30,000 coherent points in the area covered by FIGURE 6, corresponding to an average density of about 15,000 points/km2. This compares to approximately 2000 coherent points detected in the same area by processing the significantly longer C-band data series, for an average density of 1000 points/km2 (FIGURE 6B). The 15-time increase in the spatial density of coherent points was particularly significant to investigate the Ivancich landslide, and we maintain is significant to investigate similar landslides affecting urban and sub-urban areas.
Close inspection of FIGURE 6A reveals trends in the pattern of the deformation (shown by the spatial variation in the red and orange colours) that are not visible in FIGURE 6B. This is emphasized in FIGURE 7 where the displacements measured along three profiles are compared. The A-B profile shows a clear variation in the displacement rate across the landslide deposit, and outlines a distinct band of deformation of about 50 min width along the NW boundary of the landslide. The sparse ERS-1/2 & ENVISAT scatter points (purple triangles) did not detect the variation in the surface velocity across the landslide deposit. The C-D profile also shows distinct variations in the surface displacement rate, with faster movements (red) in the upper part of the landslide and slower movements (orange) in the lower part of the investigated area.
The considerably larger density of coherent points detected processing the COSMOSkyMed data is particularly significant in urban areas, where buildings and other structures and infrastructures are abundant. In FIGURE 7, the E-F profile crosses longitudinally a set of two-story, masonry buildings typical of the Ivancich neighbourhood. Close inspection of the E-F profile reveals variations in the rate of displacement between buildings, and even within individual buildings. This characteristic, distinctive of the high-resolution X-band data, is particularly relevant for the definition of the spatial pattern of the deformations. Such characteristic is also important to support the design of effective remedial measurements, and to monitor their long-term efficiency. It is worth emphasizing the possibility offered by the DInSAR analysis of the X-band COSMO-SkyMed data to detect effects of ground deformations not revealed by the ERS-1/2 & ENVISAT processing. This is demonstrated in FIGURE 8 that shows a building located on the toe of the Ivancich landslide. The building is clearly affected by displacements that ere detected and measured by processing the X-band COSMOSkyMed data (FIGURE 8A), and were not detected by processing of the C-band ERS- 1/2 & ENVISAT data (FIGURE 8B).
A fourth highlight consists in multi-frequency analysis i.e. joint analysis of the results of the processing of SAR data captured by multiple sensors for the same area.
Multi-frequency analysis was possible because of the availability of multi-band datasets for several of the DORIS test sites (TABLE 3), and was based on the following general strategy:
• C-band (ERS-1/2, ENVISAT, RadarSAT-2) data were used to measure accurately small displacement rates, whereas L-band (ALOS) data, being more sensitive to large deformation rates, were used to measure large displacement rates and to perform analyses in vegetated areas, where the use of C-band data is hampered by decorrelation phenomena, as shown in the Tramuntana Range (A in FIGURE 1) test site.
• Large archives of historical C-band data allowed for the application of consolidated and innovative PSI techniques for the production of very-long time series of deformations. Conversely, the reduced geometrical decorrelation typical of L-band SAR sensors allowed for the exploitation of conventional differential interferometry to large areas characterized by the presence of a limited number of reflectors on the ground (including structures and infrastructures). In these areas, the PSI techniques typically fail to detect ground movements, as shown in the Upper Silesian coal basin (L in FIGURE 1), and in St. Moritz (E in FIGURE 1) test sites.
Use of X-band data resulted in a higher density of measured points, and improved significantly the level of detail of the analyses, as shown in the Assisi (E in FIGURE 1), the Messina Province (H in FIGURE 1), and the St. Moritz (E in FIGURE 1) test sites. Use of X-band data further extended the applicability of space-borne SAR interferometry to faster ground movements, due to the higher spatial resolutions (up to 1 m) and the reduced repeat cycles of the X-band satellites, compared to the medium resolution C-band sensors.
The strategy was applied in the St. Moritz test area (E in FIGURE 1), where a large active landslide threatens the village (FIGURE 9). In this alpine area, DORIS performed standard PSI and single interferograms analyses, using multiple SAR sensors (FIGURE 10). Standard PSI processing of C-band ERS-1/2 & ENVISAT data allowed measuring accurately the surface deformation IN the slow-moving parts of the active landslide, including the toe of the landslide that affects the town, partially, and where the rate of movement was < 2 cm/yr. The central part of the landslide, characterized by higher rates of deformation, was analysed with COSMO-SkyMed data images covering the period 2009–2011. The results obtained processing the X-band COSMO-SkyMed data and the C-band ENVISAT data are similar, but the significantly larger number of measured points, and the reduced revisit time of the COSMO-SkyMed sensors, allowed for a more detailed investigation of the movement of single buildings, and for measuring rates of surface deformation of 20 cm/yr. In vegetated areas, single ALOS PALSAR interferogram analyses, with a time interval of 90 days, were used to measure rates of movement of about 20 cm/yr, and demonstrated that at the L-band coherence in vegetated areas is preserved even after three months.
In FIGURE 10, the different colours show the mean annual velocity determined processing data acquired by ENVISAT, COSMO-SkyMed, and the ALOS satellites.
FIGURE 11 shows similar results obtained for the eastern slope of the upper Matter Valley (D in FIGURE 1). In this area, and potentially in similar physiographical areas, application of multi-frequency interferometry allows for innovative considerations on the multiple slope processes that are active in the area. The widespread use of multifrequency interferometry not only teaches us that the landscape changes (deforms) more extensively and more rapidly than previously thought, but allows us to measure the deformation accurately and in an unprecedented range of deformation velocities.
A fifth highlight of DORIS consists in the development of methodologies to improve the exploitation of multi-sensor DInSAR techniques for landslides and subsidence mapping and monitoring. The scope of the methodologies is to generate landslide and subsidence activity maps, combining remote sensing and in situ (ground truth) data.
DORIS generates landslide activity maps in three steps (FIGURE 12). First, a preprocessing step assesses beforehand the performance of the DInSAR techniques for the targeted test sites i.e. the visibility of a specific location of interest from the radar system. This can be a cost-effective method to select the most adequate Earth Observation service & product for the problem at hand. Second, consecutive postprocessing steps are applied to combine multi-sensor DInSAR displacement measurements with landslide and subsidence inventories, to obtain terrain motion activity maps. Finally, DORIS proposed a validation methodology to demonstrate the reliability of conventional and advanced DInSAR displacement measurements to represent the “true” landslide dynamics.
Input data for the methodology include morphological and satellite PS data, a preexisting landslide inventory (LSI), and ground truth data e.g. a damage database, field survey of the study area. In the PSI post processing stage (FIGURE 12a), the visibility of the targeted study area is assessed, and the displacements along the LOS are projected downslope. The visibility assessment, calculated through the computation of two indexes (R and LUs), is useful to identify slopes, PS targets, or landslides with an adequate geometry for satellite measurements (i.e. slope direction parallel to the satellite LOS direction), to be distinguished from those where geometrical distortions will be introduced in the measurement. The geometrical distortions, typical in mountain areas, come out from the combination of the local topography with the satellite acquisition parameters, and can be reduced through the projection of satellite Line Of Sight velocities (VLOS) along the local steepest slope (VSLOPE). In a second stage (FIGURE 12a), PSI data and post-processing derived products are integrated with the landslide inventory to generate the “landslide activity map”. For each landslide in the inventory, the landslide visibility assessment and the displacement measurements are given. Separately, moving PS that were first discarded and that are not related to any specific landslide of LSI, are examined, since they could represent unknown landslides or other kinds of terrain motion processes. These clusters of moving PS are shown in a “ground motion activity map”, and need to be further investigated in the field. In the last step (FIGURE 12C), a confidence degree of the activity map is determined, in three classes (high, medium, low). The evaluation aims at assessing how reliably the PSI displacement represents the landslide movement. The reliability on the PSI measurement itself is not evaluated, but only whether the measurement is related to the landslide dynamics, or else.
The generation of subsidence activity maps is performed in four steps (FIGURE 13). In the first step, a spatial analysis is performed between the PS and the available geothematic layers, to determine the variables that should be considered for the interpolation (FIGURE 13). Next, the normalized cumulative displacements along the satellite Line of Sight (LOS) are interpolated, considering the previously identified PSs (FIGURE 13). For each PS, the experimental variograms and the fitted models are analysed. The parameters of the variogram models are used as input for the Conditional Sequential Gaussian Simulation (SGS). The retrieved percentile maps (mean, variance, 68th and 95th percentiles) are analysed to prepare the subsidence activity map (FIGURE 13), consists of an interpolated value with a confidence level on the interpolation, and an estimate of the spatial uncertainty for each pixel. In the third step, the subsidence activity map is used to identify structures (e.g. buildings) that can be damaged by ground subsidence, according to a serviceability limit state criterion (SLS). The presence of damage depends mainly on the type of structure, and on the magnitude and distribution of the settlements. SLSs are those conditions that make the structure unsuitable for its projected use. In foundations design, the most common serviceability limit states are differential settlements (δs) and angular distortions (βmax), which must be less or equal than the corresponding stated limiting value.
A significant highlight of DORIS is the extensive exploitation of terrestrial, or Ground Based Radar Interferometry. Ground Based Radar Interferometry (GBRI) exploits images captured by terrestrial radars. Both interferometers and synthetic aperture radar (SAR) systems are used. The area covered by a GBRI system depends largely on the distance from the location of the instrument (i.e. the observation point).
In practice, it is usually in the range from a few hundreds of square meters to a few square kilometres. This range is well suited to investigate single landslides and large and complex unstable slopes. Modern GBRI systems can acquire images very frequently, up to every 40 ms. This potentially very high sampling frequency makes GBRI systems well suited to measure (and monitor) fast to very fast slope movements, in addition to slow movements.
DORIS used two different GBRI systems (i) the Gamma Portable Radar Interferometer (GPRI), and (ii) the interferometer designed and built by Ellegi-LiSAlab s.r.l. & JRC. The GPRI system was used in the Matter Valley, Switzerland (D in FIGURE 1), and in the Dunaszekcső, Hungary (J in FIGURE 1) test sites. The Ellegi-LiSAlab interferometer was used in the Messina Province, NE Sicily, and specifically in the San Fratello (H in FIGURE 1) test site.
The potentialities of GBRI systems for the operational monitoring of ground deformations were demonstrated vividly during the second year DORIS annual meeting. From the panoramic view point offered by the Gornegraat peak, 3130 m elevation, in the late afternoon of 19 September 2012, the Gamma Portable Radar Interferometer (GPRI) measured surface deformation of glaciers mantling the northern slopes of Monte Rosa (FIGURE 14, from left to right the Grenzgletscher, Schwärzegletscher and Breithorngletscher glaciers). In a period of 1 hour and 33 minutes, the GPRI performed 32 successive scans, with an interval of 3 minutes between each scan. Analysis of the images allowed determining the average displacement rate of the moving glaciers, expressed in FIGURE 14 by the daily rate of movement. The red colours identify the surface movements, which are very clear for the Schwärzegletscher glacier.FIGURE 15 shows a different application of the GBRI system, mounted in the period from 1 to 3 March on the East bank of the Danube River to monitor slope instabilities along the West bank of the river that threaten the town of Dunaszekcső. Applying the PSI-like processing of the images described above, the history of displacements was obtained for all coherent pixels. FIGURE 15 shows the GBRI results together with the TerraSAR-X and the in-situ displacement measurements along the forming crack obtained through vertical and horizontal daily reading with a tape measurement.
Colours of the GPRI pixels represent the velocity of the movement toward the radar, considered positive. Pixels with velocities between +1 – -1 mm/day are not shown to highlight the rapid-moving pixels. The highest measured velocity values with theGPRI are around 1 cm/day. Motions are detected on the Várhegy area, and below the Várhegy block near the ferry station. Slower motion was detected all along the edge of the high riverbank to the N.
A particularly important highlight of DORIS is the synergic exploitation of advanced DInSAR-derived products, with in situ and thematic data and information, for improved analysis and modelling of landslides and subsidence phenomena.
DORIS experimented multiple integrated analyses of DInSAR-derived products and in situ and thematic information. In this section, three particularly relevant examples are presented and discussed.
For the Ivancich landslide (G in FIGURE 1), DORIS exploited the spatially dense COSMO-SkyMed DInSAR measurements (FIGURES 5, 6, 7), in conjunction with independent geological, geomorphological, geotechnical, and inclinometric information, for a two-dimensional modelling of the active landslide. Results are portrayed in FIGURE 16A that shows the two-dimensional displacement field obtained through Finite Element Modelling of the landslide. The numerical modelling was constrained by the available COSMO-SkyMed DInSAR measurements. FIGURE 16B shows a comparison between the modelled (red triangles) and the measured (black triangles) rates of deformation along the topographic profile. The matching is remarkably good. The result proves that high-density DInSAR measurements obtained processing the high-resolution data captured by modern X-band SAR sensors can be used, in combination with geological and geotechnical information, to explore the slope stress-strain state and the kinematic evolution of unstable slopes using Finite Element Method (FEM) modelling. This opens new possibilities for the innovative use of advanced DInSAR-based services & products.
On 14 February 2010, following a period of prolonged precipitation, a large and deepseated landslide moved along the eastern slope of the San Fratello village, in NE Sicily (H in FIGURE 1). Movement of the landslide caused extensive damage to public and private properties, and caused more than 2000 people to leave the area. For the San Fratello active landslide DORIS used satellite-based DInSAR products, terrestrial DInSAR monitoring, and independent geological, geomorphological and geotechnical information, to investigate the state of activity of the active landslide, and to determine the residual risk affecting public and private properties in the San Fratello village. DORIS exploited the historical archives of ERS-1/2 & ENVISAT data, collectively covering the period from 1992 to 2010, and recent RadarSAT-2 in the period from 2005 to 2010 to prepare DInSAR deformation maps, deformation velocity maps, and associated time series.
High-resolution X-band data taken by the COSMO-SkyMed constellation along ascending and descending orbits, and by TerraSAR-X along descending orbits, in the period from 2011 to 2012, were used to obtain spatially dense deformation maps and deformation rate maps. These maps proved particularly important to detect differential movements in the urban area. In addition, the slope that failed in February 2010 was monitored continuously between March 2010 and April 2011 using a ground based radar interferometer installed on the other side of the valley. Field surveys were conducted in the area affected by the landslide, to determine the type and extent of the damage caused by the deep-seated slope failure. FIGURE 17 shows deformation velocity maps obtained through DInSAR processing of X-band data taken by COSMO-SkyMed between 2011 and 2012. Close inspection of the map reveals multiple sites affected by ground deformation.
FIGURE 18 shows a comparison of a damage assessment map prepared through extensive field surveys (left) with a map showing post-event rate of deformation of individual buildings affected by the San Fratello landslide (right), detected processing Cosmo-SkyMED data. The matching between the most damaged buildings (red, in the left) and the buildings that suffered the largest rates of deformation (blue, in the right map) is remarkable. This result opens to the possibility of using advanced DInSARbased services & products to assess the residual risk in urban and suburban areas affected by slope deformations.
Kulcs, Rácalmás, Táborállás (I in FIGURE 1) and Dunaszekcső (J in FIGURE 1) are small touristic and residential villages located along the West banks of the Danube River, south of Budapest, Hungary and are part of the steep, high bank ridge, which can be followed down to the country border. These villages are affected by slope instabilities that cause the sudden failure of significant portion of the riverbanks, threatening the villages, the inhabitants, and private properties. For this villages, DORIS performed a multi-frequency DInSAR analysis exploiting: (i) the ESA archives of C-band ERS-1/2 data in the period from 1992 to 2000, and ENVISAT data in the period from 2002 to 2010, (ii) X-band TerraSAR-X data for 2012 (for Dunaszekcső), and for 2011 and 2012 (for Rácalmás), (iii) L-band ALOS data in the period from 2006 to 2010 for Dunaszekcső, and in the period from 2007 and 2010 for Rácalmás, and (iv) GBRI data acquired between 1 and 3 March 2013 for Dunaszekcső. In addition, (i) systematic visual interpretation of stereoscopic aerial photographs was used to complete and refine the existing landslide inventory maps for the river banks, and (iii) geophysical surveys were performed to monitor the propagation of the crack system at depth.
Very long (1992-2012) InSAR deformation time series were produced for a landslide area where only localized and after-event ground-based deformation measurements were previously available. The InSAR datasets revealed very-high confidence degree with the known landslide events on each affected village. New moving areas were detected outside the known event boundaries. On the pre-event phase, we detected significant velocity change on the area, where the damages later occurred, as an example shows below. FIGURE 19 shows a smaller part of the Rácalmás test site (I in FIGURE 1), the area around the Ferry station (Hajóállomás) of Kulcs, the village in the Northern part of the area.
The repeatedly re-deposited debris slope of Kulcs was landscaped with terraces in the 1970’s. Parcels were sold to build weekend and summerhouses. Since then the area was densely built. In the 1990’s water supply network was constructed, drainage, however, has not been resolved. The first recorded landslide with damages in the infrastructure was in 1977 March and several events occurred after (known events from the existing landslide inventory and the aerial photo interpretation are marked with colours regarding to the occurrence date on the FIGURE 19 (a) – (f) maps). The first event near Kulcs and Hajóállomás during the time period covered by InSAR datasets (24.11.1992-22.05.2012) occurred on May 2010, therefore the deformations before considered as the pre-event phase (shown on FIGURE 19 (g) graph). We selected the nearest ERS-1/2 and ENVISAT PSs, however they are on a relative higher positions, from an area not affected by the latest deformations causing damages. The ALOS and the TerraSAR-X PS coverage of the area is much dense, therefore we could select PSs from the landslide affected area (FIGURE 19 (f) map). The average VLOS velocity of the ERS-1/2 PS is 0.1 mm/yr (calculated VSLOPE = -0.33 mm/yr), which can be considered stable. The selected ENVISAT PS shows similar low mean velocity (VLOS: 0.3 mm/yr, VSLOPE = -1 mm/yr). The ALOS PS dataset shows significantly higher mean velocities (Average VSLOPE = -11.1 mm/yr) during the pre-event phase.
The selected TerraSAR-X PSs shows similar VSLOPE values (TSX-01 = -10.7 mm/yr, TSX-02 = -13.9 mm/yr) as the ALOS PS. It indicates that the area hasn’t consolidated during the last radar observation period.
A last highlight of DORIS consists in the modification of the original Multiple Change Detection (MCD) technique for the detection and mapping of eventtriggered landslides using VHR images captured by optical (multi-spectral) satellite sensors. The original MCD technique exploits VHR panchromatic and HR multispectral satellite images, and was developed and tested in a 9.4 km2 area in NE Sicily, Italy (H in FIGURE 1), where on 1 October 2009 a high intensity rainfall event caused abundant shallow landslides, soil erosion, and inundation. Pre-event and postevent images of the study area taken by the QuickBird optical satellite, and information on the location and type of landslides obtained in the field and through the interpretation of post-event digital aerial photographs, were used to validate the classification model. In particular, the pre-event image was taken on 26/09/2006 10:06:20 with an off-nadir viewing angle of 12.9°, and the post-event image was taken on 08/10/2009 09:53:31 (eight days after the event) with an off-nadir viewing angle of 2.4°. The models classify each image element (pixel) based on the probability that the pixel contains (or does not contain) a new rainfall induced landslide. To detect the landslides and to prepare the event inventory map, the MCD procedure operates following three main steps (FIGURE 20). First, the pre-event and the post-event images are pre-processed. The pre-processing includes (i) pan-sharpening and check of the result using the Walt protocol, (ii) orthorectification, (iii) geometric co-registration, (iv) radiometric correction, and (v) images co-registration. Next, variables describing changes between the pre-event and the postevent images attributed to the presence of landslides are estimated, and stacked. Next, in the original MCD, three (or more) classification models are calibrated in a training area using different multivariate statistical techniques. The calibrated models are then applied in a validation area using the same set of independent variables, and the same statistical techniques. Lastly, combined terrain classification models are prepared for the training and the validation areas. DORIS replaced the last step of the procedure with a maximum likelihood supervised classifier to cluster the indices of changes extracted by the satellite images. The classifier is driven by training areas selected by an operator on the stack of the indices. The modified MCD procedure is faster than the original procedure, because the new procedure does not need a pre-prepared inventory map to train the classifier. The modified procedure proved to be less biased by features present in the landslide map prepared manually, and that cannot be detected in the satellite images. The inventory obtained through the exploitation of the modified MCD showed about 5% less false positives, compared to the original MCD procedure. We maintain that this is a significant improvement.
The original and the modified MCD techniques were capable of detecting and mapping the rainfall induced landslides in the Giampilieri study area (FIGURE 21). We expect the method to be capable of detecting analogous shallow landslides caused by similar (rainfall) or different (e.g. earthquake) triggers, provided that the event slope failures leave discernable features captured by the post-event satellite images, and that the terrain information and satellite images are of adequate quality. The original and the modified MCD techniques can facilitate the rapid production of accurate landslide event-inventory maps, and improve the ability to map landslides consistently over large areas. Systematic application of the MCD techniques can foster the understanding of the evolution of landscapes shaped by mass-wasting processes.
The obtained results indicate that satellite passive optical sensors can be exploited successfully to detect and map event landslides. The original and the modified MCD techniques are more efficient where landslides have areas larger than a single pixel, or of clusters of two or three pixels. Use of very high resolution (VHR) multi-spectral images (e.g. IKONOS, Quickbird, World-View-1/2, Pleiades, Geoeye-1, Rapid-Eye) that are characterized by a very high spatial resolution (up to few centimetres) and by very reduced revisiting periods (down to a few days), are best suited for mapping event triggered landslides of small to medium size (AL > 10 m2). Images captured by medium resolution passive optical satellite sensors (e.g. Landsat, Aster, Spot, Formosat, E0-1, and possibly the upcoming Sentinel-2 satellite) can be used to map landslides having AL > 200 m2. The exploitation of optical remote sensing data, and related processing techniques, is better suited for the rapid mapping of populations of event-induced landslides immediately after the event, and it is feasible only where a good post-event satellite image is available in due time. The techniques are not suitable for the identification and mapping of slow-moving landslides (few centimetre/year) because, in general, slow-moving landslides do not leave clear sings that can be captured by the optical satellite sensors.
1.3.5 Problems encountered
As explained before (e.g. see Table 3), DORIS used a very large number of satellite data, chiefly SAR images, captured by different sensors. The massive use of SAR data in DORIS allowed highlighting operational constraints in the delivery of the DORIS services & products. The main problems experienced in the different study areas are listed in the following Table 5. DORIS experimented a variety of solutions to mitigate the problems encountered. The main problems were related to the collection of the necessary EO data. For several test sites, DORIS experienced insufficient and/or inadequate gathering of data, tailored to the local problems and needs. This was due primarily to satellite constraints that did not allowed to properly plan the needed acquisitions. This risk for the service and products was reduced partially exploiting archive data. In some cases, the delay in the delivery of the satellite data hampered the collection of a number of scenes sufficient to deliver the services and products.
DORIS experienced issues related to the low density of measured points in the investigated areas. The problems resulted from the local characteristics of the test sites, located in vegetated, rural, and mountain regions, where noise effects are particularly relevant. In the Swiss and the Polish sites, the low density of coherent pixels was also due to the presence of non-linear and very rapid (from centimetre to meter per year) deformation phenomena. These issues were tackled in various ways:
• L-band DInSAR data were used. The long L-band wavelength allowed to mitigate decorrelation effects and to detect larger movements, resulting in larger measured point density.
• Single interferograms with short perpendicular baselines and acquisition time intervals were considered, instead of deformation time series, allowing for an effective description of the ground movements in the studied areas.
• Where available, images taken along ascending and descending orbits were used to overcome the geometrical SAR distortion (layovers and shadows) problems.
• Terrestrial interferometry was used. GBSAR complemented effectively space-borne SAR measurements, allowing to acquire relevant information on rapid (cm per day) deformation phenomena.
Potential Impact:
1.4.1 Socio-economic impact
In large regions of Europe, and in other continents, landslides and ground subsidence are serious, partly unknown and largely underestimated phenomena that pose a threat to public and private properties, structures and infrastructures, and the population, with relevant environmental and societal consequences, and large economic losses. In some areas, mass movements represent the primary cause of death due to natural hazards, and subsidence is the most damaging geo-hazard.
DORIS is an advanced downstream service for the detection, mapping, monitoring, and forecasting of ground deformations caused by landslides and ground subsidence, at different temporal, spatial and organizational scales, and in various physiographic, climatic, and environmental settings. The services and products experimented in DORIS can contribute significantly to the detection and monitoring of ground subsidence phenomena, to the understanding of the natural and the human-made causes of the deformations, to the design of effective remedial measures and the monitoring of the mitigation measures, and to the reduction of the environmental, societal and economic consequences of the phenomena.
The benefits of DORIS include:
• An improved capacity to detect and monitoring accurately ground deformations caused by subsidence and a variety of slope processes, including landslides of different types, riverbank erosion, glacier and rock glaciers, and the melting of permafrost. This will help civil protection, environmental, transportation and planning agencies and organization, at different organizational and geographical levels – from the local to the national levels – to cope with the long-term problems posed by the ground deformations.
• An improved capacity to model the natural phenomena and the human induced causes of the ground deformations, using satellite data and products, in combination with ground-based monitoring and thematic and environmental information. This will help user of different types to predict the spatial and temporal behaviour and the effects of the ground deformations, to design effective defensive measures, and to monitor the efficacy of the defensive measures.
• The ability to provide validated results (services and products) through extensive testing and demonstration. This will benefit service providers that can now deliver validated services and products. End users of the DORIS downstream service will also benefit from the possibility of using validated services and products.
• The services and products offered by DORIS were tuned through a strong interaction with civil protection authorities, geological surveys, research institutions, and local authorities. This, and the fact that DORIS was tested in a large number of test sites representative off most of the physiographical settings where landslides and subsidence are abundant in Europe, assures that the service will operate effectively in Europe, and in other similar physiographical settings.
• The services and products offered by DORIS shall broaden the capacity and the geographical range of action of existing GMES services.
DORIS exploits fully European satellites, and related processing and interpretation technologies. DORIS will contribute to export European satellites data and technologies in the global market.
1.4.2 DORIS sustainability assessment and business model
To determine its long-term sustainability, DORIS performed a comprehensive analysis of the potential markets for its services and products. FIGURE 22 summarizes the sustainability requirements for DORIS.
The added value proposition of the services and products offered by DORIS were identified, and can be summarised as follows:
• DORIS is fully integrated with National Focal Points ensuring a consolidation of needs at various organizational levels, from the national to the local levels, and a single reliable point of contact for Civil Protection organisations.
• Satellite images and derived products are treated and interpreted by experienced, professional research geological organizations, and are adapted to local and regional geographical settings, and the specific needs of the end users.
• DORIS is uniquely positioned to bring continuous improvements of its products and services, based on its R&D capabilities. DORIS is capable to respond to ever changing user needs, on short notice.
• DORIS can provide an umbrella and/or showcase for different data providers to promote the advantages and availability of the ground detection and monitoring services to a broader audience.
• As an organisation, DORIS is in the position to negotiate input data images in bulk at a preferential rate when not free, reducing the cost of the service and products offered by the downstream.
DORIS has unique processing chains of EO and non-EO input data for subsidence, landslide, and other slope processes monitoring products, on demand.
• DORIS provides methods and procedures to ensure a reliable chain of service from raw data to decision makers.
• DORIS is a distributed centre of competence on landslide and subsidence detection, mapping and monitoring, and can contribute significantly to hazard assessment and to the design of effective risk mitigation and adaptation strategies.
The value chain of DORIS (FIGURE 23) involves a large number of participants along several steps, from EO data providers to final users. Value Added Service providers are involved to process the data using state of the art techniques. The data is then interpreted by specialised entities that have knowledge of local/regional conditions and settings. The products are then disseminated via a technical interface set up in the frame of DORIS.
DORIS services are planned primarily for Civil Protection authorities, regional and local government agencies, and municipalities. These users are identified as primary users in the value chain. There could be a higher number of entities potentially interested in DORIS benefits, including mining industry, builders, utilities, road and rail companies, etc. These users are identified in the value chain as secondary users.
Secondary users could form an important part of the potential DORIS services market, and can contribute to the sustainability of DORIS, i.e. ensure a broad user base, and contribute to the financing of products.
Not all countries are organised for civil protection in the same manner and up to the same high level standard. The complexity of the Civil Protection Capacity landscape across countries creates both an opportunity and a challenge for DORIS. The challenge is to find a way to operate in such diverse environment. The opportunity is to help achieve more homogeneous and more cost effective approaches to monitoring and assessing ground deformation problems in Europe, and elsewhere.
For the extraction industry, there is a need for independent third party assessments.
The fixed location of mines and long-term nature of mining projects encourage stakeholders to use experts and state of the art techniques to assess the ground deformation risks, and consolidate an understanding they can capitalize on. On the other hand, builders and contractors are likely to use ready products rather than engage in vast geological investigation of ground displacements for areas they are only concerned with at the time of the construction. In most countries, there is not a systematic use of EO products by utilities, beyond initiatives with universities and PhD thesis. Convincing utility operators is a marketing challenge for GMES downstream services. Road maintenance companies are interested in susceptibility maps that cover only the road, not the build-up area. Similarly to utility operators, wide and practical availability of DORIS products may incentivise them to integrate these products in their monitoring processes. In particular those that operate large portions under concession of the European motorway network in mountainous areas such as APPR/AREA. The number of laws, regulations and policies on dealing with landslide hazards are growing rapidly. The landslide hazards are getting wider recognition among European Member States. This should encourage parties to rely on more precise information for preventative risk assessment and litigations.
The potential market of DORIS was estimated for three markets of relevance:
Environmental monitoring (municipalities, ministries, R&D institutes [e.g. industry, cultural heritage, tourism]), Disaster response (civil protection), and Engineering (e.g. road, rail, mining, construction). The market is rapidly developing as is showing the double-digit sustained growth of large players like Astrium Services and DigitalGlob, in particular in value add services. There will be a growing demand from commercial entities in the next 5 years. The uptake of DORIS services by potential users will depend on a number of factors, such as (i) Service availability, (responsiveness to requests for service), (ii) Service accessibility (ease of use of the web interface, retrieval of information), (iii) Quality of service (delivery time, periodicity accuracy of the processing and interpretation), (iv) Perceived added value of service (characterisation and communication on DORIS benefits), and (v) available budgets (budgetary constraints at institutional level). Setting up Service Levels would help DORIS develop a reputation of trusted downstream service provider.
A business model for DORIS was designed, considering different operational scenarios, with different costs structures, and demand levels. On the supply side, the access to data was found to be critical to ensuring a good quality of service.
Procurement processes can translate into delays of up to several months in the acquisition of the images. This can compromise the ability of DORIS to generate and deliver DORIS products in time for the users, and to comply with its service level agreement. The redundancy of sources must be ensured, and the processing must be able to accommodate alternate sources of data / missions.
DORIS intends to coordinate the procurement of EO data through national focal points to ensure maximum use and re-use of satellite EO data, as permitted by the end user agreements of EO data providers. These sources provide prices and rules of purchasing data. The price lists have been used in the DORIS financial model to estimate cost of data. Different pricing models were evaluated: (1) a cost based pricing model where each product is sold at cost plus margin. (2) Portfolio pricing model (3) a subscription based pricing model implying different cost benefits levels for the beneficiaries of DORIS services but involving different risk profiles for the DORIS organisation.
A simple cost model structure was developed (FIGURE 24). The analysis showed that the introduction of Sentinel and free data policy opens up an opportunity for DORIS to adapt its service delivery model to deliver very small areas products and to ensure pricing at the level of square kilometre, or even 0.1 square kilometre, to match some of its customer expectation, e.g. road and rail infrastructure operators interested in small areas surrounding long stretches of land. This is likely to fuel demand for the products.
This delivery model would be greatly facilitated by a common interface based on cloud technology and shared with other Copernicus services. DORIS know-how and methodology could create business cases also outside Europe and if successfully applied, it could stimulate the European market.
The Non-Recurrent Costs (NRC) incurred during the pre-operational phase were estimated based on the cost structure of the pre-operational phase FP7 costs. Then, DORIS expected costs and revenues were estimated for three scenarios looking at servicing the regions around the tests site, the entire countries where test sites are located, and in the third scenario the entire Europe. The assessment identified that additional investment would be required for a billing platform, possibly adapting the process to the sentinel format, and for producing the first set of products for new areas of intense landslide activity in Europe. In case of cooperation among several Copernicus services, a common cloud interface and billing engine would help reduce set-up costs.
Cash flow projections were established for the different scenarios. With no surprise, the costs are largely influenced by the cost of EO data, which should go down after Sentinel become operational, under the assumption that all COSMO-SkyMed and TerraSAR-X data can be replaced by Sentinel data. The analysis showed that DORIS is likely to face a financing issue in the first years of operations before demand becomes sufficient to finance operations. DORIS will need to build time series products before being able to sell the recurrent product at affordable recurrent costs. In the meantime, the participants are likely to be required to co-finance the initial operations, unless institutional co-financing can be arranged (EC or country financing). Finally the SWOT showed that the success of DORIS would hinge on continuous improvement of the processing through R&D and adaptation to the evolving state of art and expectation of the user base.
In terms of organisation, several options were considered. DORIS could remain operating as a consortium of institutional and commercial organisations. Alternatively, DORIS could evolve into a public-private partnership where commercial players are enticed to continue contribute to R&D the time the product becomes distributed on a common platform. The public contribution of institutional organisation would be for instance a guarantee of demand for the next two years at negotiated prices. The private entities would recoup their investment through increase and new demand from the private sector. Also, DORIS could operate as a stand-alone spin-off in which all IPR owners would have a share. Different organisation templates were elaborated. The proposed organisational structure take into account the characteristics of the consortium participants, and the role they could play when, after the pre-operational period, DORIS transitions to sustained operations.
1.4.3 DORIS Spatial Data Infrastructure
Most of the DORIS services and products are “geographical” in nature, or they have a geographical component. To facilitate the exploitation of the geographical products of DORIS, a specific Spatial Data Infrastructure (SDI) was designed and implemented. The DORIS SDI exploits innovative open source software technology to facilitate access, use, and dissemination of the downstream services and products. Site specific, thematic, and environmental information is stored in dedicated geo-databases, where all the geographical products of DORIS are also archived. A modern and intuitive interface allows users to access, display, overlay, and interrogate the thematic and the environmental layers, and to use effectively the DORIS products. FIGURE 24 shows an example of the Web interface of the DORIS SDI (http://doris-sdi.irpi.cnr.it).
The DORIS SDI contains more than 160 thematic layers for six test sites. To provide an efficient Web access to the DORIS SDI, the standard Geonode Web portal was extended to include a view service client, and a discovery service client. Access to the different layers in the SDI is granted to “authenticated” users, based on their individual rights.
The SDI infrastructure is composed of:
• A View service User Interface i.e. a WebGIS that shows the DORIS served by a Web Mapping Service (WMS). The WebGIS is dynamic, with the list of the managed layers updated regularly by the server. Thus, new layers (new products) become available shortly after they are uploade in the system. The WebGIS exploits AJAX technology based on the ExtJS, OpenLayers, and GeoExt Javascript libraries. To be INSPIRE compliant, the WebGIS comprises the following components: (i) interactive map with pan, zoom, etc., (ii) list of layers for final products, (iii) list of base layers, (iv) legend (layer style), Get Feature Info tool, and (v) link to metadata.
• A Discovery service User Interface, based on the GeoNtwork project (http://geonetwork-opensource.org/) that allows searching the metadata using the following criteria: (i) geographic area, (ii) keywords, (iii) metadata change date, and (iv) temporal extent.
• A View service i.e. a WMS server developed according to INSPIRE specifications for the publication and the configuration of the WMS instance. The module exploits GeoServer (http://geoserver.org) and represents the back-end for the View service User Interface.
• A Discovery service i.e. a Catalog Service for the Web (CSW) server developed according to INSPIRE specifications. The interface is used to query the existing metadata catalogue to obtain the requested records, and to store new metadata records. Based on the GeoNetwork project, it represents the back-end for the Discovery service.
• A Download service i.e. a Web Feature Service (WFS) server developed according to INSPIRE specifications using GeoServer. The interface is used to query the list of DORIS products and to download the products, or parts of them, in GML or Shapefile format.
• A Geodatabase module i.e. a relational database with geographic extensions to handle geographic objects and classic relational objects. The database exploits PostgreSQL (http://www.postgresql.org/) and PostGIS (http://postgis.refractions.net/) technologies, and represents the back-end for the Discovery, Download and View services. The module stores all the DORIS products and their associated metadata.
1.4.4 Training activities
DORIS organized dedicated (targeted) training to inform National, Regional and Local Civil Protection authorities and environmental agencies in four European Countries of the results of the project, and specifically of the services and products offered by DORIS. The meetings were held in Mallorca, Spain, on 27 January 2011, Rome, Italy, on 4 May 2012, in Palermo, Italy, on 16 January 2013, in Budapest, Hungary, on 27 February 2013, in Katowice, Poland, on 16 April 2013, in Foligno, Italy, on 20 June 2013, and in Madrid, Spain, on 3 September 2013. More than 80 people pertaining to 50 different national, regional and local organizations informed about DORIS services & products. DORIS received interest in all countries were it was presented. FIGURE 25 shows pictures taken during some of the meetings.
1.4.5 Outreach and dissemination
DORIS performed general and specific actions to advertise its innovative services and products to a wide and diversified audience of Regional stakeholders in Europe. This was achieved particularly with a dedicated meeting held in Brussels, on 26 September 2013, and hosted by NEREUS, the Network of European Regions Using Space Technologies. More than 50 people attended the workshop.
On the request of the Research Executive Agency (REA), DORIS participated to internal and/or closed-doors meetings organized by the European Commission Directorates and the European Space Agency, including (i) the Downstream Service Meeting for Data Access, held at ESA ESRIN, Frascati, on 25 January 2010, (ii) the Supersites Coordination Workshop, held in Brussels on 10 June 2013, and (iii) a meeting on the prospective of Core and Downstream Services organized by the REA, in Brussels, on 21 November 2013.
DORIS prepared different brochures to advertise the scopes, the advancements, and the outcomes of the projects. The brochures were disseminated to key stakeholders, and are available for download through the DORIS web site (http://www.DORISPROJECT.eu/).
Scientists of the DORIS team presented the results obtained during the lifetime of the project at a number of international and national conferences and workshops, including International Geoscience and Remote Sensing Symposium (IGARS, in 2011, 2012, 2013), the General Assemblies of the European Geosciences Union (EGU, in 2011, 2012, 2013), the American Geophysical Union Fall Meetings (AGU, in 2011, 2012, 2013), the Asia Oceania Geosciences Society (AOGS, in 2012, 2013), and the “Santorini Conference”, organized by ESA in Santorini, Greece, on 21-23 May 2012.
Key scientific, technological, and methodological results of DORIS were published – or are accepted for publication – in key, peer reviewed scientific journals, including Earth Sciences Reviews, Engineering Geology, Environmental Earth Sciences, Geomorphology, Natural Hazards and Earth System Sciences, Remote Sensing, Remote Sensing of Environment, Tectonophysics. We expect that other papers describing the achievements and results of DORIS will be published in peer-reviewed journals in the near future. FIGURE 26 shows the number of papers, conference proceedings and book chapters published by DORIS scientists during the lifetime of the project.
DORIS designed, implemented and maintained a project website (http://www.DORIS-PROJECT.eu). Hosted by the project Coordinator in Perugia, the website exploits Joomla(R) open-source content management software technology. Sections of the DORIS web site include: FIGURE 27 shows the homepage of the DORIS web site.
List of Websites:
1.5 General contacts for the DORIS project
• The Coordinator of the DORIS project was the Italian Consiglio Nazionale delle Ricerche (CNR).
• The Project Manager for the DORIS project was Fausto Guzzetti, CNR IRPI, via Madonna Alta 126, 06128 Perugia, Italy.
• The Project Office for the DORIS project was composed of Frida Clerissi (CNR IRPI, Perugia, Italy), Francesca Di Matteo (CNR IREA, Naples, Italy), Angela Perrone (CNR IMAA, Potenza, Italy) and Monica Proto (CNR IMAA, Potenza, Italy). The email address of the project office is: project.office@doris-project.eu.
• The person in charge of all EO data requests and handling, including relationships with the European Space Agency, was Michele Manunta, CNR IREA, Naples, Italy.
• The mailing list of the project partners is: partners@doris-project.eu.
• The mailing list of the project website administrator is: web@doris-project.eu.
1.6 Project partners
The partners of the DORIS project are:
(1) The Italian Consiglio Nazionale delle Ricerche (CNR), the largest public research organization in Italy with the mission of designing, executing and promoting research, and of transferring and disseminating knowledge, to foster scientific, technological, economic and social development in Italy. An official Centre of Competence for the Italian Civil Protection system. [http://www.cnr.it]
(2) The Earth Sciences Department of the University of Florence (UNIFI), with the Engineering Geology Group - Geohazard Research Lab. Since 2004 an official Centre of Competence of the Italian Civil Protection for Remote Sensing and Geohazards. [http://www.unifi.it]
(3) The Italian Space Agency (ASI), a national public Institution established in 1988 to coordinate national efforts and investments in the space sectors. [http://www.asi.it]
(4) The Italian National Department for Civil Protection (DPC), a public administration office under the Office of the Prime Minister, with the mission of safeguarding the integrity of life, goods, buildings and environment from natural and human made disasters. [http://protezionecivile.gov.it]
(5) Tele-Rilevamento Europa (TRE), based in Milano, Italy, is a leading international company providing remote sensing services, and particular Synthetic Aperture Radar (SAR) processing using proprietary software. [http://treuropa.com/it/]
(6) Altamira Information (ALTAM), based in Barcelona, Spain, is a leading international company in Earth Observation using radar technology, provides customised solutions for the detection and measurements of ground movement and mapping solutions. [http://www.altamira-information.com/]
(7) Gamma Remote Sensing (GAMMA), a Swiss SME with extensive expertise in the use of space-borne and terrestrial radar technology for land surface deformation monitoring. [http://www.gamma-rs.ch]
(8) The Spanish Instituto Geológico y Minero de España (IGME), the Geological Survey of Spain, was established in 1849 by a Royal Decree with the original denomination of “Commission for the Geological Chart of Madrid and the Kingdom”. [http://www.igme.es/]
(9) The Hungarian Magyar Földtani és Geofizikai Intézet (MFGI), formerly Eötvös Loránd Geofizikai Intézet, was established in 1907 as the first applied geophysical institute in the world. It is part of the Hungarian Office for Mining and Geology (MBFH). [http://www.mfgi.hu/]
(10) The Swiss Federal Department for Environment Transports Energy and Communication (FOEN) is the Swiss federal government’s centre of environmental expertise and is part of the Swiss Federal Department of the Environment, Transport, Energy and Communications.
(11) The Polish Panstwowy Instytut Geologiczny (PGI), Polish Geological Institute, National Research Institute, is the largest R&D state institution conducting studies of geological setting of the country under the general supervision of the Ministry of the Environment. [http://www.pgi.gov.pl/]
(12) Booz & Company (BOOZ), a premier global management consulting firm launched in 2008 following the separation of Booz Allen Hamilton’s global commercial business from its U.S. Government services business. [http://www.booz.com/]
(13) The Italian Consortium Tecnologie per le Osservazioni della Terra e i Rischi Naturali (TERN), was constituted on 2005 by public and private partners, including the E-Geos SpA (Finmeccanica Group), and the CREATEC SMEs Consortium.
DORIS WEBSITE:
www.doris-project.eu
