Geohazards Monitoring in Roma from InSAR and In Situ Data: Outcomes of the PanGeo Project

Abstract

Within the PanGeo project (financed by the European Commission under the 7th Framework Program), the Geological Survey of Italy (ISPRA) and the Urban Planning Department of the City of Roma developed a geodatabase and map of the geological hazards for the territory of Roma, integrating remotely sensed data (PSInSAR—Permanent Scatterer Interferometry Synthetic Aperture Radar) and in situ geological information. Numerous thematic layers, maps and inventories of hazards (e.g., landslides, sinkholes, cavities), geological and hydrogeological data added to historical and recent urbanization information were compared to the permanent scatterer (PS) data from the European Remote Sensing satellites (ERS-1/2, 1992–2000) and ENVISAT (2002–2005) descending scenes, in order to produce a ground stability layer (GSL). Based on the PS data, most of the territory appears stable (almost 70 % of PS velocities are within ±1 mm/year). About 14 % of the PSs show positive line-of-sight (LOS) velocities (measured along the LOS of the satellite) between 1 and 3 mm/year and more than 2 % exceed 3 mm/year; more than 11 % of PSs show negative LOS velocities between −1 and −3 mm/year, while about 3 % exceed −3 mm/year (with tens of the PSs showing velocities over −20 mm/year). The GSL is comprised of polygons or multi-polygons (multipart polygons grouping individual polygons under a single identifier geohazard) enclosing areas where geohazards have been pointed out by PS data and/or in situ surveys (observed instabilities), and by polygons enclosing areas potentially affected by geohazards, based on the available knowledge of the territory (potential instabilities). In Roma’s GSL, 18 multi-polygons (covering ca. 600 km2) related to observed instabilities have been outlined, where ground movements could be detected through InSAR data or where landslides and sinkholes are known to have occurred. Other 13 multi-polygons (covering nearly 900 km2) concern areas where the potential occurrence of geohazards was inferred by combining geological and/or geothematic data (potential instabilities). The geohazards mapped in Roma have been: landslides, collapsible grounds, compressible grounds, groundwater abstraction, mining, man-made ground, tectonic movements, and volcanic inflation/deflation. The lattermost is the likely cause of the significant uplift observed in the Alban Hills area. However, this paper focuses on two more currently impending hazards: subsidence and sinkholes. In general, sinkhole-prone areas (areas of dense underground cavities) are hard to discern from satellite data, but can be revealed by ruling out other potential causes of observed ground movement based on in situ data. Subsiding zones are effectively detected by the available PSInSAR dataset over a total extent of about 60 km2, mostly overlapping the recent alluvial areas of the Tiber and its tributaries. PSs show a very different behaviour inside and outside the historical centre. Inside, loading by anthropogenic construction and man-made ground since ancient times has led to an almost complete consolidation of the recent river deposits, marked by modest to absent subsidence. In contrast, outside, subsidence clearly stands out, with negative LOS velocities that, although generally within several mm/year, can locally exceed −25 mm/year. PS data have provided motion information at both a regional and local scale (up to the scale of a single building). Closer to the sea, in the Tiber delta area, velocities increase, especially above recently reclaimed marsh areas, rich in peat and organic clays. Velocities can change significantly over short distances, as in the international airport area, reflecting the local stratigraphic setting. The same occurs in the two subsiding areas located within the Alban Hills volcanic complex, which is generally affected by ground uplift. As a whole, PSInSAR ground motion velocities offer a significant contribution to susceptibility and hazard recognition studies. In particular, such a method provides a fast and effective tool available to local authorities to monitor ground and building behaviour, possibly allowing for timely prevention activities, especially when coupled with appropriate in situ knowledge.