DETECTING AND MONITORING LANDSLIDE PHENOMENA WITH ...
DETECTING AND MONITORING LANDSLIDE PHENOMENA WITH TERRASAR-X PERSISTENT SCATTERERS DATA: THE GIMIGLIANO CASE STUDY IN CALABRIA REGION (ITALY) Silvia Bianchini, Francesca Cigna, Chiara Del Ventisette, Sandro Moretti, Nicola Casagli Department of Earth Sciences, University of Firenze, Via La Pira 4, 50121 Florence (Italy) ABSTRACT This work illustrates the potential of Persistent Scatterer Interferometry (PSI) using X-band SAR (Synthetic Aperture Radar) data for a detailed detection and characterization of landslide ground displacements at local scale. We present the case study of Gimigliano, located in Calabria Region (Italy) and extensively affected by past and present landslide phenomena. 27 TerraSAR-X SAR images were collected from December 2010 to October 2011, over the Gimigliano area to perform the PSI analysis. In order to assess any spatial and temporal evolution patterns of deformation, the present X-band PS data were compared with historical motion rates derived from ERS1/2 and ENVISAT satellites, and with geo-morphological evidences resulting from auxiliary data like pre-existing landslide inventories and orthophotos referred to different dates, finally validated with recent field checks. The outcomes of this work can be taken into account for further hazard-reduction analysis and to support risk mitigation design within the investigated area. Index Terms — Landslides, Persistent Scatterer Interferometry, SAR Interferometry, landslide mapping 1. INTRODUCTION Landslide phenomena represent a major geological hazard worldwide, threatening human lives and settlements, especially in urban areas where the potential socio-economic losses and damages are stronger because of the higher value of the element at risk exposure and vulnerability. The impact of these natural disasters in highly populated and vulnerable areas can be reduced or prevented by performing a proper identification and mapping of such ground movements, in order to support an appropriate urban planning and therefore to facilitate the risk mitigation design. Detection, mapping and monitoring of active landslides and vulnerable slopes can greatly benefit from radar satellite data analysis. From an economical and practical point of view, remote sensing techniques successfully complement conventional surveying and geomorphological tools, due to
their great cost-benefit ratio, non-invasiveness and high precision. Persistent Scatterer Interferometry (PSI) is a multi-temporal interferometric technique which exploits temporal and spatial characteristics of interferometric SAR (Synthetic Aperture Radar) data, collected by point-likely targets called Persistent Scatterers or PS [1]. The PS analysis has recently demonstrated to be a valuable and useful tool to detect surface displacements with millimeter precision for mapping and monitoring of slow-moving landslide phenomena.
2. CASE STUDY: LOCATION AND BACKGROUND 2.1. Geographical and geological location The study area, extended up about 15 km2, is located in the Gimigliano municipality within Catanzaro (CZ) province, in Calabria (Italy) (Fig.1). Calabria region is characterized by one of the highest risk scenarios of the whole Italian territory and it is extremely susceptible to landsliding because of high slope gradients and presence of poorly-consolidated or weathered lithologies, associated with unsuitable urban expansion and seasonally intense precipitations [2]. We focused on 2 sites: Gimigliano town (Site 1) within the main urban centre and a periferic district, Cavorà (Site 2). The old medieval village is divided into two built-up areas (Gimigliano Inferiore and Gimigliano Superiore), placed on sloping ridges of Mt. Gimigliano, almost on the same axis NNE oriented, but at different altitudes (510 m and 620 m, respectively); the new modern Gimigliano is built-up in the eastern lower part of the urbanized area. Cavorà is a district located in the southern-eastern part of Gimigliano municipality (Fig. 1). From a geological and tectonic point of view, Gimigliano is located on a “push-up” structure, that is bounded on its NW side by a NNE-striking W-verging thrust, along which Mesozoic ophiolite rocks overthrusted Paleozoic metamorphic crystalline units, and on its SE side is delimited by NNE-striking E-verging thrust. This well-
structured fault system generates many wide cataclastic zones characterized by weathered and intensively fractured rocks that turn out to be geotechnically soft and highly erodible. 2.2. Recent landslide events occurrence Landslide phenomena are widespread across time up to nowadays, all over the Gimigliano area. In winter 2010 catastrophic rainfall-induced landslides affected Gimigliano town and surrounding districts, causing damages to structures, infrastructures and roads. In January 2010 a lanslide occurred on the slope along Corace River, reactivating a dormant deep-seated landslide, and the existing displacement scarp surface enlarged, damaging Corace bridge pylon (Fig. 2a). A similar event happened on 13th February 2010, when a landslide occurred on the metamorphic rock escarpment, threatening the buildings stability of the upper Gimigliano Superiore village, as well as the safety of Risorgimento road below (Fig. 2b). On 10th March 2010, rainfall storm with intensity climax in 1-2 hours, triggered a complex landslide on the already unstable rock slope of Mt. Gimigliano southern mountainside. At the end of January 2011, in Cavorà district, sever precipitation triggered a dormant deep-seated landslide, activating a suspended phenomenon occurred on 28th January 2010 and causing huge damages to the road network (Fig. 2c).
Fig. 1 – Location of the study area and TerraSAR-X input data.
3. METHODOLOGY AND COMBINED INPUT DATA The methodology performed is based on recent radar PS TerraSAR-X data, combined with historical SAR archives derived from ERS and ENVISAT data, in order to assess any spatial and temporal deformation patterns, prior and after known landslides occurred in 2010 in the Gimigliano area (Fig. 3 and Fig. 4). These interferometric data were properly overlapped on geo-morphological information resulting from existing auxiliary data like landslide databases (i.e. PAI and IFFI inventories, referred respectively to 2001 and 2007) and orthophotos referred to different dates, finally validated with recent field checks in the area of interest. Throughout the photo- and radar-interpretation procedures, recently exploited and validated by several scientific applications (e.g. [3], [4], [5]), the operative procedure finally led to the detailed mapping and characterization of ground movement phenomena in the Gimigliano area.
Fig. 2 - Latest landslides occurrence in winter 2010 in Gimigliano area. Explanations in the text. 3.1. TerraSAR-X data for Gimigliano case study 27 SAR images were acquired by the satellite TerraSAR-X in X-band (signal wavelength ~3.1 cm) over an area of 116.9 km2 to perform the PSI analysis. All scenes were acquired in Spotlight mode, along descending orbits, with 11 days revisiting time and look angle of 34°. The acquisition period is from 8th December 2010 to 1st October 2011. The available scenes were processed by e-GEOS by means of Persistent Scatterers Pairs approach (PSP-IFSAR analysis, [6]) and provided 50173 PS in the study area. Regarding the velocity values along the LOS (Line of Sight), mean deformation vary from -87 to +33 mm/year in the 11 months acquisition time. Within the PS velocity, a threshold range of ± 2.5 mm/yr (away or toward the satellite direction) was fixed for classifying PS rates and for distinguishing moving from not moving areas.
Fig. 3 - Site 1: TSX PS analysis within Gimigliano town.
4. DETECTING LANDSLIDE PHENOMENA 4.1 Historical scenario Historical ground motions in the Gimigliano area were analyzed through PS data from ERS and ENVISAT satellites, referred respectively to the spanning time 19932000 and 2003-2009. Within Gimigliano town (Site 1), the PS analysis highlights that the modern downhill built-up area was affected by surface ground movement since 1993 (average LOS motion rates of about 4.5 mm/yr), while the old urbanized centres of Gimigliano Superiore and Inferiore turned out to be stable. Active landslide recorded in PAI and IFFI inventories, affecting the eastern part of old Gimigliano Superiore, was confirmed to be moving with mean velocity of about 6-7 mm/yr. A significant enlargement of the instability phenomena was detected across time. Within Cavorà district (Site 2), the historical PS ERS and ENVISAT analysis evidences a ground deformation since 1993, covering an area wider than the pre-mapped geometries, with average motion rate of 6-7 mm/yr. 4.2 Updated scenario The recent PS TerraSAR-X data and in situ observations permitted to accurately update the ground movements in the two selected sites. PSI analysis allowed a detailed spatial and temporal detection of the landslides phenomena, leading to an updated scenario and new landslide inventory mapping up to 2011 (Fig. 3 and Fig. 4). Landslide boundaries of IFFI inventory were confirmed or modified, usually enlarging,
Fig. 4 - Site 2: TSX PS analysis within Cavorà district
throughout PSI data and orthophotos analysis, that permitted to better trace the geometries of the unstable areas; moreover, some phenomena not previously mapped were detected. Among the total 42 landslides in the area, 10 premapped landslides were enlarged within their boundaries and 5 new landslides were mapped. PS annual mean velocities and deformation time series were studied in depth and allowed defining an updated state of activity for each phenomenon (Fig. 5). Within Site 1, PS data evidence that the maximum deformation velocities are recorded in the southern-western Mt. Gimigliano mountainside toward Corace River, with LOS rates up to -40 mm/yr, and in the eastern part of Gimigliano Superiore old village, with mean displacement velocity reaching up values of -30 mm/yr in the radar coverage period (Fig. 3). A high instability phenomenon was also detected in the modern built-up sector of Gimigliano Superiore, where a wide landslide affecting the whole area turned out to be reactivated, with average velocity of about 8-9 mm/yr. Within Site 2, the area characterized by the highest PS velocities is close to the built-up area around the railway station where local slope failures are documented in the last years (Fig. 4). The average annual PS velocities range from 30 to -60 mm/yr over the satellite coverage period and progressively increases from February 2011 until the last acquisition (October 2011), demonstrating that the landslide phenomenon is still active and the ground movement is dangerously raising with deformation rates up to -70 mm/yr.
local scale. PS analysis in Gimigliano area was focused on two sites, Gimigliano town and Cavorà district, that showed the highest displacement rates within the present TerraSARX PS datastack. PS interpretation allowed detecting and monitoring landslide-induced phenomena in Gimigliano area and providing an updated scenario up to 2011 within the two selected sites. The applied approach demonstrated to have a successful operative usefulness to support landslide hazard and risk assessment, since PSI technique has the potential to work in synergy with conventional geo-morphological tools and to improve the quality and the temporal reference of pre-existing landslide inventories. 7. ACKNOWLEDGMENTS
Fig. 5 – Updated scenario (2011): example within Cavorà district. Each landslide is identified by a code (e.g. CAV08).
This work was carried out within the SAFER (Services and Applications For Emergency Response) project, funded by the EC-GMES-FP7 initiative (grant agreement n°218802). TerraSAR-X images, as well as ERS and ENVISAT PSI products, were processed and provided by e-GEOS.
5. DISCUSSIONS
8. REFERENCES
PS data allowed carefully mapping and measuring present and historical ground motions up to nowadays in the Gimigliano area, and therefore to obtain a spatial and temporal evolution of the investigated landslide phenomena. In particular, using and integrating different interferometric datasets referred to different time intervals (ERS, ENVISAT and TerraSAR-X), it was possible to reconstruct the dynamic of the area of interest and to provide estimates of the mean yearly velocities since 1992 up to 2011. A great amount of TerraSAR-X PS was identified in the investigated area: the number and the spatial density of detected X-band PS data are significantly higher than the C-band ones. It can also been observed that TerraSAR-X results show a higher deformation rates than those of ERS and ENVISAT: X-band PS targets with deformation greater than ± 2.5 mm/yr is 68% of the total, while 10-30% for the C-band PS. This is due to the shorter TerraSAR-X monitoring period (only 10 months) and shorter temporal sampling (11 days in respect to 35 days), enabling the detection of more coherent pixels that show displacements. On the other hand, decorrelation noises are of greater importance in X-band, because of the shorter wavelength and revisiting cycle. Moreover, due to the phase ambiguity of SAR data processing and satellite acquisition parameters in both C- and X-band, the PSI analysis of landslides is limited to phenomena ranging from “extremely slow” to “very slow”, according to the velocity classification of Cruden & Varnes, 1996, [7].
[1] A. Ferretti, C. Prati, F. Rocca, “Permanent Scatterers in SAR interferometry”, IEEE Trans. Geosci. Remote Sens, vol. 39 (1), pp. 8-20, 2001.
6. CONCLUSIONS The outcomes of this work represent a valuable example of landslides detection and characterization by PSI analysis at
[2] C. Tansi, F. Muto, S. Critelli, and G. Iovine, “NeogeneQuaternary strike-slip tectonics in the central Calabrian Arc (southern Italy)”, Journal of Geodynamics, 43, pp. 393-414, 2007. [3] F. Cigna, S. Bianchini, N. Casagli, “How to assess landslide activity and intensity with Persistent Scatterer Interferometry (PSI): the PSI-based matrix approach”, Landslides, doi: 10.1007/s10346-012-0335-7, 2012. [4] S. Bianchini, F. Cigna, G. Righini, C. Proietti, and N. Casagli, “Landslide HotSpot Mapping by means of Persistent Scatterer Interferometry”, Environmental Earth Sciences, vol online, pp. 118, doi: 10.1007/s12665-012-1559-5, February 2012. [5] G. Herrera, D. Notti, J.C. García-Davalillo, O. Mora, G. Cooksley, M. Sánchez, A. Arnaud, and M. Crosetto, “Landslides analysis with C- and X-band satellite SAR data: the Portalet landslide area”, Landslides, 8 (2), pp. 195-206, 2011. [6] M. Costantini, A. Iodice, L. Magnapane, and L. Pietranera, “Monitoring terrain movements by means of sparse SAR differential interferometric measurements”. Proc. of IGARSS 2000, 20th, IEEE Int. Geoscience and Remote Sensing Symposium, Honolulu, Hawaii, USA, 24-28 July 2000, pp. 3225–3227, 2000. [7] D.M. Cruden and D.J. Varnes, “Landslide types and processes”, Turner AK, Schuster RL (eds) Landslides: Investigation and Mitigation, Sp. Rep. 247, Transportation Research Board, National research Council, National Academy Press, Washington DC, pp 36–75, 1996.
their great cost-benefit ratio, non-invasiveness and high precision. Persistent Scatterer Interferometry (PSI) is a multi-temporal interferometric technique which exploits temporal and spatial characteristics of interferometric SAR (Synthetic Aperture Radar) data, collected by point-likely targets called Persistent Scatterers or PS [1]. The PS analysis has recently demonstrated to be a valuable and useful tool to detect surface displacements with millimeter precision for mapping and monitoring of slow-moving landslide phenomena.
2. CASE STUDY: LOCATION AND BACKGROUND 2.1. Geographical and geological location The study area, extended up about 15 km2, is located in the Gimigliano municipality within Catanzaro (CZ) province, in Calabria (Italy) (Fig.1). Calabria region is characterized by one of the highest risk scenarios of the whole Italian territory and it is extremely susceptible to landsliding because of high slope gradients and presence of poorly-consolidated or weathered lithologies, associated with unsuitable urban expansion and seasonally intense precipitations [2]. We focused on 2 sites: Gimigliano town (Site 1) within the main urban centre and a periferic district, Cavorà (Site 2). The old medieval village is divided into two built-up areas (Gimigliano Inferiore and Gimigliano Superiore), placed on sloping ridges of Mt. Gimigliano, almost on the same axis NNE oriented, but at different altitudes (510 m and 620 m, respectively); the new modern Gimigliano is built-up in the eastern lower part of the urbanized area. Cavorà is a district located in the southern-eastern part of Gimigliano municipality (Fig. 1). From a geological and tectonic point of view, Gimigliano is located on a “push-up” structure, that is bounded on its NW side by a NNE-striking W-verging thrust, along which Mesozoic ophiolite rocks overthrusted Paleozoic metamorphic crystalline units, and on its SE side is delimited by NNE-striking E-verging thrust. This well-
structured fault system generates many wide cataclastic zones characterized by weathered and intensively fractured rocks that turn out to be geotechnically soft and highly erodible. 2.2. Recent landslide events occurrence Landslide phenomena are widespread across time up to nowadays, all over the Gimigliano area. In winter 2010 catastrophic rainfall-induced landslides affected Gimigliano town and surrounding districts, causing damages to structures, infrastructures and roads. In January 2010 a lanslide occurred on the slope along Corace River, reactivating a dormant deep-seated landslide, and the existing displacement scarp surface enlarged, damaging Corace bridge pylon (Fig. 2a). A similar event happened on 13th February 2010, when a landslide occurred on the metamorphic rock escarpment, threatening the buildings stability of the upper Gimigliano Superiore village, as well as the safety of Risorgimento road below (Fig. 2b). On 10th March 2010, rainfall storm with intensity climax in 1-2 hours, triggered a complex landslide on the already unstable rock slope of Mt. Gimigliano southern mountainside. At the end of January 2011, in Cavorà district, sever precipitation triggered a dormant deep-seated landslide, activating a suspended phenomenon occurred on 28th January 2010 and causing huge damages to the road network (Fig. 2c).
Fig. 1 – Location of the study area and TerraSAR-X input data.
3. METHODOLOGY AND COMBINED INPUT DATA The methodology performed is based on recent radar PS TerraSAR-X data, combined with historical SAR archives derived from ERS and ENVISAT data, in order to assess any spatial and temporal deformation patterns, prior and after known landslides occurred in 2010 in the Gimigliano area (Fig. 3 and Fig. 4). These interferometric data were properly overlapped on geo-morphological information resulting from existing auxiliary data like landslide databases (i.e. PAI and IFFI inventories, referred respectively to 2001 and 2007) and orthophotos referred to different dates, finally validated with recent field checks in the area of interest. Throughout the photo- and radar-interpretation procedures, recently exploited and validated by several scientific applications (e.g. [3], [4], [5]), the operative procedure finally led to the detailed mapping and characterization of ground movement phenomena in the Gimigliano area.
Fig. 2 - Latest landslides occurrence in winter 2010 in Gimigliano area. Explanations in the text. 3.1. TerraSAR-X data for Gimigliano case study 27 SAR images were acquired by the satellite TerraSAR-X in X-band (signal wavelength ~3.1 cm) over an area of 116.9 km2 to perform the PSI analysis. All scenes were acquired in Spotlight mode, along descending orbits, with 11 days revisiting time and look angle of 34°. The acquisition period is from 8th December 2010 to 1st October 2011. The available scenes were processed by e-GEOS by means of Persistent Scatterers Pairs approach (PSP-IFSAR analysis, [6]) and provided 50173 PS in the study area. Regarding the velocity values along the LOS (Line of Sight), mean deformation vary from -87 to +33 mm/year in the 11 months acquisition time. Within the PS velocity, a threshold range of ± 2.5 mm/yr (away or toward the satellite direction) was fixed for classifying PS rates and for distinguishing moving from not moving areas.
Fig. 3 - Site 1: TSX PS analysis within Gimigliano town.
4. DETECTING LANDSLIDE PHENOMENA 4.1 Historical scenario Historical ground motions in the Gimigliano area were analyzed through PS data from ERS and ENVISAT satellites, referred respectively to the spanning time 19932000 and 2003-2009. Within Gimigliano town (Site 1), the PS analysis highlights that the modern downhill built-up area was affected by surface ground movement since 1993 (average LOS motion rates of about 4.5 mm/yr), while the old urbanized centres of Gimigliano Superiore and Inferiore turned out to be stable. Active landslide recorded in PAI and IFFI inventories, affecting the eastern part of old Gimigliano Superiore, was confirmed to be moving with mean velocity of about 6-7 mm/yr. A significant enlargement of the instability phenomena was detected across time. Within Cavorà district (Site 2), the historical PS ERS and ENVISAT analysis evidences a ground deformation since 1993, covering an area wider than the pre-mapped geometries, with average motion rate of 6-7 mm/yr. 4.2 Updated scenario The recent PS TerraSAR-X data and in situ observations permitted to accurately update the ground movements in the two selected sites. PSI analysis allowed a detailed spatial and temporal detection of the landslides phenomena, leading to an updated scenario and new landslide inventory mapping up to 2011 (Fig. 3 and Fig. 4). Landslide boundaries of IFFI inventory were confirmed or modified, usually enlarging,
Fig. 4 - Site 2: TSX PS analysis within Cavorà district
throughout PSI data and orthophotos analysis, that permitted to better trace the geometries of the unstable areas; moreover, some phenomena not previously mapped were detected. Among the total 42 landslides in the area, 10 premapped landslides were enlarged within their boundaries and 5 new landslides were mapped. PS annual mean velocities and deformation time series were studied in depth and allowed defining an updated state of activity for each phenomenon (Fig. 5). Within Site 1, PS data evidence that the maximum deformation velocities are recorded in the southern-western Mt. Gimigliano mountainside toward Corace River, with LOS rates up to -40 mm/yr, and in the eastern part of Gimigliano Superiore old village, with mean displacement velocity reaching up values of -30 mm/yr in the radar coverage period (Fig. 3). A high instability phenomenon was also detected in the modern built-up sector of Gimigliano Superiore, where a wide landslide affecting the whole area turned out to be reactivated, with average velocity of about 8-9 mm/yr. Within Site 2, the area characterized by the highest PS velocities is close to the built-up area around the railway station where local slope failures are documented in the last years (Fig. 4). The average annual PS velocities range from 30 to -60 mm/yr over the satellite coverage period and progressively increases from February 2011 until the last acquisition (October 2011), demonstrating that the landslide phenomenon is still active and the ground movement is dangerously raising with deformation rates up to -70 mm/yr.
local scale. PS analysis in Gimigliano area was focused on two sites, Gimigliano town and Cavorà district, that showed the highest displacement rates within the present TerraSARX PS datastack. PS interpretation allowed detecting and monitoring landslide-induced phenomena in Gimigliano area and providing an updated scenario up to 2011 within the two selected sites. The applied approach demonstrated to have a successful operative usefulness to support landslide hazard and risk assessment, since PSI technique has the potential to work in synergy with conventional geo-morphological tools and to improve the quality and the temporal reference of pre-existing landslide inventories. 7. ACKNOWLEDGMENTS
Fig. 5 – Updated scenario (2011): example within Cavorà district. Each landslide is identified by a code (e.g. CAV08).
This work was carried out within the SAFER (Services and Applications For Emergency Response) project, funded by the EC-GMES-FP7 initiative (grant agreement n°218802). TerraSAR-X images, as well as ERS and ENVISAT PSI products, were processed and provided by e-GEOS.
5. DISCUSSIONS
8. REFERENCES
PS data allowed carefully mapping and measuring present and historical ground motions up to nowadays in the Gimigliano area, and therefore to obtain a spatial and temporal evolution of the investigated landslide phenomena. In particular, using and integrating different interferometric datasets referred to different time intervals (ERS, ENVISAT and TerraSAR-X), it was possible to reconstruct the dynamic of the area of interest and to provide estimates of the mean yearly velocities since 1992 up to 2011. A great amount of TerraSAR-X PS was identified in the investigated area: the number and the spatial density of detected X-band PS data are significantly higher than the C-band ones. It can also been observed that TerraSAR-X results show a higher deformation rates than those of ERS and ENVISAT: X-band PS targets with deformation greater than ± 2.5 mm/yr is 68% of the total, while 10-30% for the C-band PS. This is due to the shorter TerraSAR-X monitoring period (only 10 months) and shorter temporal sampling (11 days in respect to 35 days), enabling the detection of more coherent pixels that show displacements. On the other hand, decorrelation noises are of greater importance in X-band, because of the shorter wavelength and revisiting cycle. Moreover, due to the phase ambiguity of SAR data processing and satellite acquisition parameters in both C- and X-band, the PSI analysis of landslides is limited to phenomena ranging from “extremely slow” to “very slow”, according to the velocity classification of Cruden & Varnes, 1996, [7].
[1] A. Ferretti, C. Prati, F. Rocca, “Permanent Scatterers in SAR interferometry”, IEEE Trans. Geosci. Remote Sens, vol. 39 (1), pp. 8-20, 2001.
6. CONCLUSIONS The outcomes of this work represent a valuable example of landslides detection and characterization by PSI analysis at
[2] C. Tansi, F. Muto, S. Critelli, and G. Iovine, “NeogeneQuaternary strike-slip tectonics in the central Calabrian Arc (southern Italy)”, Journal of Geodynamics, 43, pp. 393-414, 2007. [3] F. Cigna, S. Bianchini, N. Casagli, “How to assess landslide activity and intensity with Persistent Scatterer Interferometry (PSI): the PSI-based matrix approach”, Landslides, doi: 10.1007/s10346-012-0335-7, 2012. [4] S. Bianchini, F. Cigna, G. Righini, C. Proietti, and N. Casagli, “Landslide HotSpot Mapping by means of Persistent Scatterer Interferometry”, Environmental Earth Sciences, vol online, pp. 118, doi: 10.1007/s12665-012-1559-5, February 2012. [5] G. Herrera, D. Notti, J.C. García-Davalillo, O. Mora, G. Cooksley, M. Sánchez, A. Arnaud, and M. Crosetto, “Landslides analysis with C- and X-band satellite SAR data: the Portalet landslide area”, Landslides, 8 (2), pp. 195-206, 2011. [6] M. Costantini, A. Iodice, L. Magnapane, and L. Pietranera, “Monitoring terrain movements by means of sparse SAR differential interferometric measurements”. Proc. of IGARSS 2000, 20th, IEEE Int. Geoscience and Remote Sensing Symposium, Honolulu, Hawaii, USA, 24-28 July 2000, pp. 3225–3227, 2000. [7] D.M. Cruden and D.J. Varnes, “Landslide types and processes”, Turner AK, Schuster RL (eds) Landslides: Investigation and Mitigation, Sp. Rep. 247, Transportation Research Board, National research Council, National Academy Press, Washington DC, pp 36–75, 1996.
