Ground vertical movements in urban areas of the Veneto Region
(Italy) detected by DInSAR
T. Strozzi, U. Wegmuller
Gamma Remote Sensing, Muri (BE), Switzerland
L. Carbognin, L. Tosi
Istituto per lo Studio della Dinamica delle Grandi Masse, CNR, Venezia,
Italy
R. Rosselli
Consorzio Venezia Nuova Servizio Informativo, Venezia, Italy
P. Teatini,
Dept. Mathematical Methods and Models for Scientific
Applications, University of Padova, Padova, Italy
EXTENDED ABSTRACT
Remote-sensing techniques represent suitable tools for geological
cartography and sustainable management of natural resources. Spaceborne
Synthetic Aperture Radar (SAR) systems offer the possibility, through
differential SAR interferometry (DInSAR), to map surface displacements at
mm to cm resolution. Spectacular results are obtained for geophysical
sciences with earthquake displacement, volcano deformation, glacier dynamics
and land subsidence being mapped. With regard to land subsidence,
SAR interferometry exhibits complementary characteristics to the leveling
surveys, because it has the capability to map large urban areas at low
cost and high spatial resolution. The high precision leveling surveys,
on the other hand, are used outside of the cities and to set up a
reference point for the SAR subsidence values. In the case of the
south-eastern Veneto region, where till to 2000 high precision leveling
surveys are available only (i) around the Lagoon margin and along few lines
from (ii) Venezia to Rua di Feletto (Treviso), (iii) Mestre to Padova,
(iv) Padova to Rovigo, and (v) Chioggia to Adria, SAR interferometry has
the capability to monitor the vertical displacements of all the built-up
or sparsely vegetated areas (i.e. where stable structures permit the
formation of a coherent phase signal over time) not fully covered with
leveling results.
DInSAR makes use of two SAR images acquired from
slightly different orbit configurations and at different times to exploit
the phase difference of the signals. The phase signal derived from
an interferometric image pair relates both to topography and line-of-sight
surface movement between the acquisitions, with atmospheric phase
distortions, signal noise and inaccuracy in the orbit determination as
main error sources. The basic idea of DInSAR is to subtract the topography
related phase (for instance simulated from a Digital Elevation Model, DEM)
from the interferogram to derive a displacement map. In this study a
time series of six interferometric radar images of the European Remote
Sensing Satellites ERS-1 and ERS-2 from 1993 to 2000 was used. In order
to generate a single subsidence map with reduced errors, the interferometric
radar images were combined. In the following, displacement in the vertical
direction was assumed, the land subsidence map was transformed to the
Italian cartographic system with a spatial resolution of 30 m, and the
pixel corresponding to the benchmark Nodale 63 (ex 24'') in Treviso
(that has been already considered stable for the leveling surveys)
was considered the stable reference. The high accuracy of the 1993-2000
DInSAR survey is confirmed by a quantitative validation with the
1993-2000 leveling data. For the 87 benchmarks where values from both
surveying techniques are available a zero average difference of the
vertical displacement velocity is prescribed and a standard deviation of 0.9
mm/year is found; the minimum and maximum differences are -3.1 mm/year
and +2.5 mm/year, respectively. From this number and previous works
we conclude that the accuracy of the SAR interferometric subsidence rates
is on the order ±1 mm/year, which is also the expected accuracy of the
leveling surveys.
DInSAR is very useful to perform detailed investigations in coastal areas.
Chioggia, Jesolo and Caorle, Bibione and Lignano reveal the presence of
a significant seaward gradient in land subsidence practically impossible
to be detected by other traditional monitoring techniques like
geometric leveling and DGPS. Hence, the proposed methodology provides
a fundamental support for coastal management projects and environmental
risk analyses.
SAR interferometry has been applied in other minor cities of the
Veneto plain characterized by very different conditions from the
displacement point of view. Castelfranco and Conegliano, located
close to the Alpine foothills, appear quite stable. Land settlement
in Rovigo appears very changeable with high values (2-3 mm/year) downtown
and lower rate (about 1 mm/year) in the outskirts. In Cavarzere,
Oderzo and San Donà the subsidence velocity ranges between 3 and
4 mm/year. The more critical situation is found in Adria, Taglio di Po,
and Portogruaro where the sinking velocity reaches 5 mm/year
in large portion of the towns.
DInSAR mapping in large cities such as Padova, Treviso, and Mestre
is shown. These applications indicate the potential of the radar
satellite analysis in built-up areas characterized by a
general stability, with displacement rates within the range
±1 mm/year.
A figure shows an example of the DInSAR capability
to monitor the evolution of land subsidence in a relatively small
urban area (Abano - Montegrotto Terme). The stack of ERS-1/2 SAR
images previously used has been increased to cover also more
recent years. For this particular example the point labeled as A
was considered the stable reference, following the indications
provided by historical leveling surveys. The influence of the
drastic reduction of groundwater pumping used for geothermal purpose
in recent years is clearly shown. The DInSAR results match very well
the available leveling data.
Although in 2000 the old leveling network has been extended within
the framework of the ISES Project (the present overall length is
about 1000 km), a detailed monitoring of several villages is impossible
through leveling and DGPS. Considering the high hydrogeological
risk in these urban areas due to the presence of rivers and the critical
land elevation (large portions of this territory lie below the mean
sea level), the areally distribution of the subsidence rate has to be
well known. Results shown by the present work prove that this
integration in the knowledge of the subsidence process can be efficiently
obtained by DInSAR. This technique represents an effective methodology
for both a back-analysis during the last decade and the future
monitoring in the portion of territory covered by large and small built-up areas.
