Document Actions

ijege-14_01-rocca-et-alii.pdf

background image
35
Italian Journal of Engineering Geology and Environment, 1 (2014)
© Sapienza Università Editrice
www.ijege.uniroma1.it
DOI: 10.4408/IJEGE.2014-01.O-03
A
lfredo
ROCCA
(*,**)
, P
Aolo
MAZZANTI
(*,**)
, d
Aniele
PERISSIN
(***)
& f
rAncescA
BOZZANO
(*,**)
(*)
Sapienza University of Rome - Department of Earth Sciences - P.le Aldo Moro, 5 - 00185 Rome, Italy
(**)
NHAZCA s.r.l. - Spin-off Sapienza Università di Roma - Via Cori, snc - 00177 Rome, Italy
(***)
Purdue University - Lyles School of Civil Engineering - 550 Stadium Mall Drive, West Lafayette, IN 47907, USA
DETECTION OF PAST SLOPE ACTIVITY IN A DESERT AREA
USING MULTI-TEMPORAL DINSAR WITH ALOS PALSAR DATA
EXTENDED ABSTRACT
Il presente lavoro illustra i risultati di uno studio di un versante costiero nel Sultanato dell’Oman (penisola Arabica) che, tra il 2011 e
il 2012, ha subito alcuni processi d’instabilità gravitativa in corrispondenza di aree interessate dalla costruzione di una strada tra Hasik e
Ash Shuwaymiyyah (Governatorato di Dhofar). Un’analisi condotta tramite interferometria SAR (Synthetic Aperture Radar) satellitare,
ha consentito di caratterizzare la dinamica evolutiva di tale versante identificando i fenomeni più significativi di attività da un punto di
vista temporale e spaziale.
L’area di studio, localizzata nella regione di Dhofar, si trova all’interno del graben di Hasik, a nord della foce del Wadi Dahanat. La
zona è caratterizzata dalla deformazione Oligocenica del margine passivo relativo all’apertura del rift di Aden. Sovrapposte al basamento
proterozoico (~800 Ma; M
ercolli
et alii, 2006) si rinvengono nell’area diverse unità calcaree, appartenenti alla serie Eocenica del gruppo
di Hadramaut, tipiche di una deposizione marina di acque poco profonde. Dopo la deformazione Oligocenica, una nuova fase tettonica
all’inizio del Miocene ha coinvolto le unità sedimentarie definendo una linea costiera con scogliere disposte a diverse quote. In tale contesto
è localizzato il versante costiero oggetto di studio, ai piedi di una parete sub-verticale, laddove si riconosce la presenza di una spessa coltre
detritica con blocchi di decine di metri cubi e con evidenze di passati processi di instabilità gravitativa (P
lAtel
& r
oger
, 1989; r
oger
et
alii, 1989; B
echennec
et alii, 1993; W
Atchorn
et alii, 1998, r
oBertson
& B
AMkhAlif
, 2001, g
hezzi
et alii, 2012).
Le analisi SAR Interferometriche satellitari (DInSAR) (M
Assonnet
et alii, 1993; h
Anssen
, 2001; s
trozzi
et alii, 2005, 2010; J
eBur
et alii, 2013) sono state applicate al fine di acquisire informazioni quantitative sui processi deformativi avvenuti negli ultimi anni, Le
tecniche DInSAR, infatti, soprattutto nella loro declinazione avanzata “A-DInSAR” (Advanced DInSAR), consentono, di sfruttare le
immagini acquisite nel tempo dai satelliti su una stessa area per estrarre informazioni sull’evoluzione dei processi deformativi con elevata
accuratezza (f
erretti
et alii, 2001, 2011; B
erArdino
et alii, 2002; P
erissin
& W
Ang
, 2012). Le tecniche A-DInSAR, tuttavia, richiedono
la disponibilità di numerose immagini per ottenere informazioni di spostamento affidabili e con accuratezze elevate. L’area oggetto del
presente studio risulta tuttavia coperta da poche immagini d’archivio dei satelliti dell’Agenzia Spaziale Europea, insufficienti per l’appli-
cazione di metodologie A-DInSAR. Un numero maggiore di immagini è stato invece acquisito dal satellite ALOS PALSAR (Advanced
Land Observation Satellite - Phased Array type L-band Synthetic Aperture Radar), gestito dall’Agenzia Spaziale Giapponese JAXA
(Japanese Aerospace Exploration Agency). Le nove immagini acquisite da ALOS PALSAR tra il dicembre 2006 e l’aprile 2010, benché
insufficienti per eseguire analisi A-DInSAR, si sono rivelate utili per osservare la presenza di spostamenti sul versante indagato attraver-
so l’analisi manuale condotta sugli interferogrammi calcolati. Più in dettaglio, sono stati generati interferogrammi ridondanti (ciascuna
immagine è stata accoppiata con tutte le altre) e i più coerenti sono stati utilizzati per analizzare il segnale di fase interferometrica, con
l’obiettivo di stimare gli spostamenti filtrando il segnale dai contributi di fase dovuti alla topografia (quota). Tale risultato è stato ottenuto
analizzando, per ciascun interferogramma, i parametri relativi alla distribuzione temporale delle immagini e alla cosiddetta baseline nor-
male (proporzionale alla distanza tra i satelliti durante l’acquisizione delle immagini), che è direttamente correlabile con la sensibilità del
sistema di osservare e quantificare il dato relativo alla quota altimetrica.
Attraverso questa metodologia è stato possibile riconoscere sul versante tre aree circoscritte che hanno subito degli spostamenti nell’in-
tervallo di tempo compreso tra giugno 2008 e giugno 2009.
Il metodo di analisi DInSAR utilizzato in questo lavoro ha consentito di estrarre informazioni sui processi deformativi nonostante le
poche immagini disponibili. A tal proposito, è bene osservare che l’applicazione di algoritmi automatici per la stima degli spostamenti
avrebbe potuto ridurre l’osservabilità di tali processi a causa della discontinuità spaziale e temporale degli spostamenti stessi.
Tale approccio può essere considerato quindi una soluzione investigativa utile in aree remote e non urbanizzate, dove è disponibile una
limitata conoscenza sulle condizioni geomorfologiche e geologiche locali e laddove non esistono altri dati di monitoraggio sull’evoluzione
storica dei fenomeni deformativi.
background image
A. ROCCA, P. MAZZANTI, D. PERISSIN & F. BOZZANO
36
Italian Journal of Engineering Geology and Environment, 1 (2014)
© Sapienza Università Editrice
www.ijege.uniroma1.it
ABSTRACT
Coastal slope involved in the construction of a road in the
Sultanate of Oman was affected during the period 2011-2012 by
instability processes. Lithological and geomorphological evi-
dence suggested a general conformation of the slope that is very
prone to gravitational instability processes. Because satellite SAR
interferometry (InSAR) is the only technique that is able to pro-
vide quantitative information about past ground displacements,
it has been chosen to investigate the past evolution of the slope.
Nine archived SAR images acquired by the ALOS PALSAR have
been analysed using a hybrid approach based on the classical dif-
ferential interferometry (DInSAR) and Quasi-Persistent Scatter-
ers (QPS) techniques. Some ground deformation processes have
been detected and measured on three portions of the slope. One
of them has been localised within the area affected by the recent
landslide phenomenon. Thanks to this approach the deformation
processes have been defined in time and the related displacements
have been quantitatively estimated.
K
ey
words
: Oman, landslides, SAR interferometry, DInSAR, A-DInSAR,
ALOS PALSAR
INTRODUCTION
The use of Earth Observation (EO) data for the investigation
of remote areas (deserts, forests and generally areas with insuf-
ficient lines of communication) plays a key role in acquiring both
preliminary information and monitoring data for several fields of
infrastructure and natural hazard management (B
Arrett
, 2013).
Among spaceborne EO techniques, Synthetic Aperture Radar
(SAR) represents an interesting and widely used tool. Because
SAR systems are based on active sensors, they are able to provide
information both night and day and in the presence of cloud cov-
erage, thus expanding upon the opportunities offered by optical
systems. Characteristic of radar sensors is the use of coherent sig-
nals, i.e., the capability of retaining information about the phase
component of the electromagnetic signal. Therefore, the pixels
of the SAR images include both amplitude and phase terms. The
former is related to the reflection intensity of the scattering tar-
gets, and the latter is related to sensor-target distance. This pecu-
liarity has been exploited in recent decades, turning earth obser-
vation spaceborne SAR systems into tools to provide quantitative
data, using interferometric processing techniques (InSAR).
Multi-pass classical InSAR is performed by coupling SAR
images to generate a single interferogram to perform phase sig-
nal analyses (l
i
& g
oldstein
, 1987; 1990; g
ABriel
& g
oldstein
,
1988; g
oldstein
et alii, 1988; P
rAti
et alii, 1990; h
Anssen
, 2001;
s
iMons
et alii, 2002; t
ong
et alii, 2010). The phase signal of a
single interferogram, however, carries information related to sev-
eral contributions (e.g., topography, displacement, atmospheric
artefacts, and noise). If the temporal baseline (namely, the time
interval between the acquisition of the two repeat pass scenes used
to compute the interferogram) is short enough, we can assume
that no displacement occurred and thus ascribe all phase signals
to topographic contribution. In contrast, to acquire information
about displacements that occurred in the time interval between the
acquisitions of two SAR scenes, the topographic contribution can
be subtracted from the derived interferogram (e.g., by an available
Digital Elevation Model). In the latter case, this technique, known
as Differential InSAR - DInSAR, has been successfully applied in
the investigation of various ground deformation processes. DIn-
SAR provides good results for the investigation of displacement
characterised by wide, spatially smooth deformation processes,
such as coseismic and postseismic deformations (M
Assonnet
et
alii, 1993; 1994), volcanic deformation processes (M
Assonnet
et
alii, 1995) or ice and glacier dynamics (g
oldstein
et alii, 1993;
k
Wok
& f
Ahnestock
, 1996). However, landslide displacements
have also been detected and measured by DInSAR (c
Arnec
et
alii, 1996; f
runeAu
et alii, 1996; s
inghroy
et alii, 1998; s
ingh
et
alii, 2005; s
trozzi
et alii, 2005; 2010; r
iedel
& W
Alther
, 2008;
g
Arc
í
A
-d
AvAlillo
et alii, 2014; J
eBur
et alii, 2013).
As stated above, one of the primary limitations of DInSAR is
the difficulty in separating the different contributions affecting the
interferometric phase signal. Moreover, DInSAR cannot be sup-
ported by effective solutions for the detection and removal of an
atmospheric disturbance. Finally, noise caused by temporal decor-
relation (i.e., a long time interval between multi-pass SAR scene
acquisition) and/or geometrical decorrelation (i.e., long distance be-
tween positions occupied by satellite during scene acquisition) can
seriously prevent the attainment of useful interferometric results.
To compensate for these limitations, several data processing
approaches have been proposed, namely Advanced DInSAR (A-
DInSAR). All of them exploit long time series of SAR images of
the same area to attain some common results: i) estimation and
removal of atmosphere artefacts (i.e., Atmosphere Phase Screen -
APS); ii) contemporary estimation of several phase contributions
(at least, topography and displacement); iii) capability of providing
time series of deformation along the whole time interval. The de-
velopment and diffusion of Advanced Differential SAR Interferom-
etry (A-DInSAR) methods over the last decade have significantly
increased the range of applications of SAR data for past-oriented
investigation and for future monitoring of ground displacements
(f
erretti
et alii, 2001; 2011; B
erArdino
et alii, 2002; h
ooPer
et
alii, 2004; l
AnAri
et alii, 2004; l
Auknes
, 2004; k
AMPes
, 2006; v
An
l
eiJen
& h
Anssen
, 2007; P
erissin
, 2008; s
trAMondo
et alii, 2008;
P
erissin
et alii, 2011; 2012; B
ozzAno
& r
occA
, 2012).
From this perspective, if DInSAR is able to provide a “snap-
shot” of a given deformation process in the time interval between
the acquisition of two scenes, A-DInSAR approach is actually
able to provide displacement information over time, such as long-
term time series of displacement.
background image
DETECTION OF PAST SLOPE ACTIVITY IN A DESERT AREA USING MULTI-TEMPORAL DINSAR WITH ALOS PALSAR DATA
37
Italian Journal of Engineering Geology and Environment, 1 (2014)
© Sapienza Università Editrice
www.ijege.uniroma1.it
However, some conditions need to be satisfied to perform
A-DInSAR analysis, such as the availability of a large quantity
of SAR images (conventionally, greater than 20) acquired with
the same geometric and radiometric features over the same area.
However, especially in remote areas, this condition cannot be al-
ways satisfied by space agency archives.
For the area investigated in this chapter, only few archive imag-
es from the ALOS PALSAR, ERS and Envisat sensors were availa-
ble; hence, DInSAR analyses were performed using a redundant ap-
proach suitable for increasing the reliability of the attained results.
The study presented herein is of a coastal slope in the Dhofar
region in the Sultanate of Oman involved in the construction of
a major road (Hasik - Ash Shwaimiyah Project). As the slopes
experienced some gravitational instability processes during con-
struction activities during the period August 2011-June 2012, our
objective was to better understand the stability conditions of such
a slope before the above-mentioned time interval. To attain quan-
titative information related to past displacements, InSAR based
approaches represent the only available solutions.
STUDY AREA
The study area is in the Dhofar region (Oman, in the east-
ern Arabian Peninsula) (Fig. 1), located inside the Hasik graben,
north of the riverbed of Wadi Dahanat (almost always dry). The
area under investigation is a steep barren slope under a high cliff
facing the Kuria Muria Bay. The slope, affected by instability is-
sues, has been involved in the construction of a road to connect
Hasik to the northern village of Ash Shuwaymiyyah.
GEOLOGICAL AND GEOMORPHOLOGICAL SETTING
The Hasik Graben is an Oligocene onshore deformation zone
of the passive margin related to the oblique opening of the oce-
anic basin of the Aden Rift. The Hasik Graben (10 km wide and
30 km long) is bounded to the North and South by two master
normal faults with a N45°E to N75°E trend and presents an axial
dip towards the east (Fig. 2) (f
ournier
et alii, 2004).
Several paleo-environments have characterised the depo-
sition during the Tertiary period. Transgressive cycles with the
deposition of shallow marine formations and regressive cycles
evidenced by the deposition of gypsum began in the Late Palae-
ocene (60 Ma) and ended in the Late Eocene (38 Ma) (r
oger
et
alii, 1989; W
Atchorn
et alii, 1998).
A carbonate succession up to 1000 m thick is exposed, charac-
terised by three sedimentary groups (P
lAtel
& r
oger
, 1989; r
oger
et alii, 1989; B
echennec
et alii, 1993; r
oBertson
& B
AMkhAlif
,
2001), which correspond to pre-rift, syn-rift and post-rift stages of
deposition, the Hadhramaut, Dhofar and Fars groups, respectively.
The study area is dominated by the Eocene series of the Hadhramaut
group, whereas the Dhofar group (syn-rift) locally outcrops in the
northern part of this sector. The post-rift Fars group, instead, does not
outcrop in the study area and is restricted to the Salalah plain (Fig. 2).
Fig. 2 - Structural map and sedimentary units
of Southern Dhofar, Sultanate of Oman,
derived and simplified from the 1/250,000
geological maps of Salalah and Hawf
(P
latel
et alii, 1992; R
ogeR
et alii, 1992);
in addition, the main structures of the
Gulf of Aden and the on-land study area
are present. Abbreviations are as follows:
AF Tf: Alula-Fartak Transform Fault;
So Tr: Socotra Transform Fault; Ow
Tr: Owen Transform Fault; ShR: Sheba
Ridge; CaR: Carslsberg Ridge. The red
box identifies the study area (from l
eP
-
vRieR
et alii, 2002, mod.)
Fig. 1 - Geographic location of the study area
background image
A. ROCCA, P. MAZZANTI, D. PERISSIN & F. BOZZANO
38
Italian Journal of Engineering Geology and Environment, 1 (2014)
© Sapienza Università Editrice
www.ijege.uniroma1.it
Focusing on the study area (Fig. 3), the Hadhramaut group,
which in turn overlies the Proterozoic basement (~800 Ma; M
er
-
colli
et alii, 2006), consists of carbonate units typical of shallow-
water conditions, including the massive limestone of the Umm Er
Radhuma Formation and the Rus Formations (up to 600 m thick,
late Palaeocene/Thanetian - early Eocene/Ypresian), the Dam-
mam formation (middle Eocene/Lutetian-Bartonian), consisting
of yellow shale and chalky marl with interbedded argillaceous
limestone (thickness 43 meters) and fine to medium grained
creamy limestone, sometimes marly and chalky, nodular or dolo-
mised (thickness approximately of 200 meters), and finally the
Aydim formation (late Eocene/Priabonian), consisting of lime-
stone, chalky marl and calcarenitic deposits characterised by rich
macrofauna and banks of corals.
The syn-rift Dhofar group, unconformably deposited on the
Hadhramaut group, consists of lacustrine limestone at the base
(100 m), overlain by platform limestone (Ashawq Formation,
600 m), which passes laterally at the top to the overlying, chalky
calci-turbidic deposits of the late Oligocene to early Miocene
Mughsayl Formation (700 m thick beneath the Salalah plain; P
lA
-
tel
et alii, 1992). The slope deposits of the Mughsayl Formation,
which include megabreccia, debris flows, and olistolitic material
transported from the adjoining shelf, result from the collapse and
subsidence of the margin and correspond to deeper depositional
environments.
After the Oligocene tectonic phase, a tectonic readjustment
occurred in the Early Miocene so that the sedimentary package
extended to the sea of the Kuria Muria Bay with a coastline of
cliffs at different elevations, defined ridges tilting to the North-
West, stacked as tumbled dominos one on top of another, creating
overhangs and ledges (g
hezzi
et alii, 2012).
At the end of the Early Miocene, the sedimentary conditions
become similar to the present, in accordance with the general
emersion of the South Arabian plate in that period. Thus, sub-
sequent retrogressive-transgressive marine cycles of the Middle
Miocene and the youngest of the Pliocene and Pleistocene com-
bined with a possible late tectonic can be responsible for erosion-
al and depositional phenomena involving the tertiary formations.
Finally, not depicted on the map in Fig. 3 for scale reasons,
two quaternary terms outcrop, related to colluvial and alluvial de-
posits. The quaternary colluvium is particularly interesting for the
objective of the present work because it is located on the coastal
slope under investigation. It is composed of rock debris accumu-
lated at the foot of the cliffs caused by landslides below the fault
scarps and consists of limestone blocks of up to 30 cubic meters
in average size with marl and gypsum inside a silt matrix, extend-
ing towards the sea. The thickness is several tens of meters (more
Fig. 3 - Geological map of the “Hasik Graben” area. The red box identifies the study area (from F
ouRnieR
et alii, 2004; modified)
background image
DETECTION OF PAST SLOPE ACTIVITY IN A DESERT AREA USING MULTI-TEMPORAL DINSAR WITH ALOS PALSAR DATA
39
Italian Journal of Engineering Geology and Environment, 1 (2014)
© Sapienza Università Editrice
www.ijege.uniroma1.it
than 20 meters). The screes formation is still on-going from the
Pleistocene. To conclude, there are also very thin (20-50 cm) al-
luvial deposits on the Wadi Dahanat floor consisting primarily
of boulders emplaced during the last pluvial phase at the end of
Pleistocene (g
hezzi
et alii, 2012).
From the geomorphological perspective, the slope, which is
primarily involved through the presence of the above-mentioned
colluvial deposit, presents evidences of phenomena of past gravi-
tational instability. According to g
hezzi
et alii (2012), these geo-
morphologic features seem linked to the following causes: i) dif-
ferent phases of the sea level in different eras caused some terraced
planes located at different elevations. Consequently, the movement
of the collapsed deposits increased, generating new collapses; and
ii) the strong supply of sediments eroded by the Wadi Dahanat and
its minor tributary are transported to the sea over time.
These characteristics are located particularly in the southern
half of the study area, north of the mouth of the Wadi Dahanat.
In the northern part of the sector under investigation, evidence
of past instability processes involving the quaternary colluvium
seem less severe; this can most likely be related to the decreased
thickness of the colluvium layers (g
hezzi
et alii, 2012).
During the period August 2011-June 2012, slope instability
processes affected the area involved in the construction of the
Hasik - Ash Shwaimiyah road. Because we were interested in
the spatial and temporal evolution of the slope dynamics before
the beginning of the construction, past information needed to be
acquired about the slope under investigation. To reach this aim,
InSAR-based techniques have been selected as the most effec-
tive for the presented purpose because they are the only tech-
niques able to provide quantitative information on past ground
displacements
.
MULTI-TEMPORAL DINSAR ANALYSES
C-BAND ARCHIVE SAR IMAGES
The SAR data archives contain very few images related to
the study area. In particular, only four multi-temporal images
acquired by the ERS-1 and ERS-2 satellites (European Space
Agency - ESA) between 1992 and 1996 were available in the
ESA archive (Tab. 1). More recent data have been acquired by
ESA using the Envisat satellite; however, the best interferometric
stack reached only five scenes for the 2003-2004 period (Tab. 2).
L-BAND ARCHIVE SAR IMAGES
More archival images have been acquired by the ALOS satel-
lite (Advanced Land Observation Satellite) using the PALSAR
sensor (Phased Array type L-band Synthetic Aperture Radar) and
archived by JAXA (Japanese Aerospace Exploration Agency).
Nine images, acquired in the period between the end of 2006
and the middle of 2010, were available. We used 4 Fine Beam
Single Polarisation (FBS, bandwidth: 28 MHz) images and 5
Fine Beam Double polarisation (FBD, bandwidth: 14 MHz)
images. The pixel spacing of the fine beam single polarisation
(FBS) data is 4.68 m in the range direction and 3.17 m in the
azimuth direction, whereas the pixel spacing of the double po-
larisation (FBD) images is 9.36 m in the range direction by 3.17
m in the azimuth direction (c
hen
et alii, 2012) (Tab. 3). Using
the same central frequency, the range bands of the FBS and
FBD images are fully overlapped. This allowed interferometric
processing to be performed, combining FBS and FBD images in
the same interferometric pair using the common HH polarisation
[for this objective, the FBD data needed to be doubly oversam-
pled in range to ensure the same resolution for all images used in
the dataset (c
hen
et alii, 2014)].
The primary difference between the data from the ESA satel-
lites and those acquired by the ALOS PALSAR is represented
by the different wavelengths. L-band SAR images from ALOS
are characterised by much longer wavelengths (more than 23
cm), four times longer than the C-band data. Using longer wave-
lengths, the interferograms generated are characterised by higher
spatial coherence. Decorrelation is much less severe for temporal,
geometrical and atmospheric effects. The so-called “critical base-
line” (namely, the value of the Bn above which an interferogram
Tab. 1 - Available dataset acquired by ERS satellites
Tab. 2 - Available dataset acquired by Envisat satellite (ASAR sensor)
Tab. 3 - Available dataset acquired by ALOS satellite (PALSAR sensor)
background image
A. ROCCA, P. MAZZANTI, D. PERISSIN & F. BOZZANO
40
Italian Journal of Engineering Geology and Environment, 1 (2014)
© Sapienza Università Editrice
www.ijege.uniroma1.it
is totally decorrelated) is much higher (several kilometres) (Tab.
4). For this reason, a given data-stack whose normal baselines of
image pairs are usually largely below this threshold can be fully
exploited to obtain useful information.
The critical baseline is given by:
where
B
c
is the critical baseline,
λ
is the sensor wavelength, R is
the slant range distance from the satellite to the reflector on the
Earth surface,
θ
is the incidence angle, and
Δρ
is the radar range
resolution.
Moreover, the same consideration can be applied to the risk
of losing information from SAR images acquired with long tem-
poral baselines. Comparing C-band and L-band data, the latter
appear considerably resilient to temporal decorrelation effects
over the same areas, as quantitatively demonstrated by W
ei
&
s
AndWell
(2010). Finally, it is notable that the available images
belong to an ascending geometry stack and are thus able to pro-
vide more appropriate LOS orientation along the observed slope
compared with that achievable by descending ERS and Envisat
stacks, as illustrated below.
ANALYSES OF ARCHIVED SAR DATA
Regarding C-band images, with such a poor data stack, only a
few differential interferograms could be computed, even assum-
ing complete connection between the images to generate maxi-
mum redundancy of interferograms (images are connected to one
another). In this case, considering N(N - 1)/2 interferograms to
have a complete connection (where N is image number), we had
only six and ten interferograms for the ERS and Envisat stacks,
respectively (Fig. 4). Moreover, most of the computed interfero-
grams were characterised by long normal baselines (Bn), which
negatively influence the spatial coherence.
Most of the generated interferograms were extremely noisy,
primarily because of geometrical decorrelation effects. For this
reason, only one ERS and four Envisat interferograms were char-
acterised by an acceptable coherence for the analysis of past dis-
placements. Considering such conditions and the small quantity
of available data, the application of A-DInSAR methods was not
feasible. Moreover, the analysis of differential interferograms
characterised by higher spatial coherence was insufficient to de-
tect and measure the presence of displacements during the period
of investigation.
Furthermore, both the ERS and Envisat datasets were ac-
quired along descending orbital paths so that the LOS was char-
acterised by a stronger component perpendicular to the slope un-
der investigation (i.e., the worst condition under which to observe
possible displacements along the slope direction) (Fig. 5).
With the L-band SAR signal (23.6 cm wavelength) and the
greater number of available images, the interferometric results
were significantly better than those attained from the C-band
data processing (ERS and Envisat). However, even the ALOS
PALSAR stack with its potentially high coherent interferograms
contains few images, insufficient to perform reliable A-DInSAR
Fig. 4 - Graphs showing images connections related to ERS (A) and Envisat (B) data-stacks. Each dot represents an image in accordance with numeration
of Tab. 3.1 and Tab. 3.2. Images are plotted in time (X axis) and space, represented by normal baseline (Y axis). Each line represents an interfero-
gram. Colorbar is related to average spatial coherence of interferogram represented by the line connecting the two dots (i.e SAR images)
Tab. 4 - Comparison between ERS/Envisat and ALOS PALSAR data
(from W
ei
& S
andWell
, 2010 mod.)
background image
DETECTION OF PAST SLOPE ACTIVITY IN A DESERT AREA USING MULTI-TEMPORAL DINSAR WITH ALOS PALSAR DATA
41
Italian Journal of Engineering Geology and Environment, 1 (2014)
© Sapienza Università Editrice
www.ijege.uniroma1.it
analyses. To attain good information from such images, a DIn-
SAR-based approach was adopted, starting from single interfero-
grams. Considering a complete image connection graph (to gen-
erate high redundancy), 36 interferograms were created, coupling
all images to each other. As can be observed in Fig. 6, using the
L-band signal, very highly coherent interferograms have been ob-
tained, even with the high temporal and normal baselines, com-
pared with those in Fig. 4. Furthermore, the topographic phase
has been subtracted (using an available 20 m resolution DEM),
thus providing 36 Differential Interferograms that allowed for the
observation of the displacement phase contributions.
Because no A-DInSAR analysis was feasible and the area was
relatively small, we performed a manual analysis of the best inter-
ferograms. The first step to attain reliable results regarding dis-
placement occurrences was the detection and subtraction of the
residual topographic component on differential interferograms. In
this case, the residual heights have been estimated using a multi-
image approach, more specifically, the Quasi-Persistent Scat-
terers (QPS) method (P
erissin
& W
Ang
, 2012). Using the QPS
approach, a sub-selection of points characterised by a very high
and stable backscattering signal has been selected, thus estimat-
ing the phase related to the residual heights, exploiting all gener-
ated interferograms and weighing such pixels on the basis of their
spatial coherence value. Local height information, derived from
InSAR data, has been used to refine the original DEM to remove
this re-estimated topographic component from the newly gener-
ated interferograms.
Starting from this redundant dataset, a quality threshold based
on a spatial coherence value (>0.5) has been applied, thereby se-
lecting the 24 best interferograms to perform the displacement
investigation. In Tab. 5, the interferograms used for the further
investigation stages are listed. Clearly, despite the selection per-
formed to keep only the best interferograms, no images have been
discarded, and all of them have been used to provide the highest
possible temporal coverage.
INTERFEROMETRIC RESULTS
The 24 differential interferograms shown in Tab. 5 have been
analysed to investigate the interferometric signals to detect and
quantify any possible displacements.
We removed the topography as finely as possible using the
QPS approach. However, a small residual topographic compo-
nent may persist. To distinguish and separate that component
from the displacement component, a criterion based on a double
check was applied:
i) The phase signal related to the topography is proportional to
B
n
; thus, if a given interferometric feature varies proportio-
nally with B
n
, it is reasonable to assume that it is due to resi-
dual topography;
ii) Once a given interferometric signal is hypothesised to be in
relation with the displacement, it is possible to observe the
Fig. 5 - Difference between ascending (left) and descending (right) orbits to observe displacements over the studied slope
Fig. 6 - Graphs showing image connections related to ALOS PALSAR
data-stack. Each dot represents an image in accordance with
numeration of Tab. 3.3. Images are plotted in time (X axis) and
space, represented by normal baseline (Y axis). ). Each line
represents an interferogram. Colorbar is related to average
spatial coherence of interferogram represented by the line con-
necting the two dots (i.e SAR images)
background image
A. ROCCA, P. MAZZANTI, D. PERISSIN & F. BOZZANO
42
Italian Journal of Engineering Geology and Environment, 1 (2014)
© Sapienza Università Editrice
www.ijege.uniroma1.it
persistence of the signal using the redundant interferograms
because the signal should be temporally congruent with the
interferogram data-set, generated by crossing all images. This
also aided in understanding the time interval when the displa-
cement processes occurred.
The examination of such interferograms allowed for the detec-
tion of some interferometric features (identifi ed as “anomalies”)
recursively present in some of them. Such anomalies, whose pres-
ence along the interferometric set was apparently random, have
been properly interpreted by applying the principles described
above. In this way, it was possible to detect three small areas of
the investigated slope that were characterised by interferometric
features that can be interpreted as displacements that occurred
within the examined period. By ordering the interferograms tem-
porally (with the master images sorted in temporally ascending
order) and also considering the B
n
values, it was possible to iden-
tify the interferometric features related to the displacements by
discarding those related to the topographic residual phase; moreo-
ver, it was possible to defi ne the time interval when the detected
displacements occurred.
The three small areas are localized in Fig. 7, and they are re-
lated to portions of the geocoded wrapped interferograms shown
in Fig. 8, Fig. 9 and Fig. 10. Once a given interferometric feature
was identifi ed (red boxes inside the fi gures), it was “tracked” on
all of the interferograms, marking those where it appeared. The
fi rst analysis of the detected phase signals was to distinguish the
residual topographic features from the displacements. To attain
this objective, as explained above, we checked the presence/ab-
sence of such anomalies in relation to the B
n
values. As you can
see, the presence of these anomalies is not directly linked to B
n
;
interferograms 1, 2, 8, 22 and 23 in Fig. 8, Fig. 9 and Fig. 10 are
characterised by large baseline values, but they do not show the
presence of such anomalies. In contrast, the detected phase signals
are present in the other interferograms independently of B
n
; for
example, for all areas, interferograms 10 and 15 are characterised
by the above-mentioned anomalies, even with small values of B
n
.
This approach allowed for the defi nition of these phase sig-
nals due to displacements. Red borders surrounding some fi gures
in Fig. 8, Fig. 9 and Fig. 10 identify interferograms with phase
signals caused by displacements.
To confi rm this deduction, selected interferograms have also
been interpreted from a temporal point of view. In this perspec-
tive, redundant use of SAR images allowed the observation of
the temporal congruency of the interferometric signals. Because
the data used for this analysis are simple wrapped differential
interferograms, atmospheric artefacts, decorrelation effects and
general noise could still be present. For this reason, the use of the
same image for several interferograms reduces the risk of losing
information, which could be masked by such disturbances affect-
ing any single image. Furthermore, as explained in section 4, the
use of the same SAR image in more than one interferogram has
also been crucial in defi ning with reliable accuracy the time inter-
val when the detected processes occurred.
Tab. 5 - ALOS PALSAR interferograms used for the investigation of
past displacements
Fig. 7 - Localization of three small areas affected by instability proc-
esses detected by ALOS PALSAR interferograms (imagery from
Google Earth V. 7.1.2.2041, DigitalGlobe, 2014)
background image
DETECTION OF PAST SLOPE ACTIVITY IN A DESERT AREA USING MULTI-TEMPORAL DINSAR WITH ALOS PALSAR DATA
43
Italian Journal of Engineering Geology and Environment, 1 (2014)
© Sapienza Università Editrice
www.ijege.uniroma1.it
In Tab. 6, all interferograms used for phase signal analysis
and interpretation are reported as crosses between available im-
ages. Dark grey cells indicate an interferogram discarded because
of the coherence below the threshold value; light grey cells in-
dicate a null interferogram (i.e., master and slave images are the
same); green cells indicate the absence of a phase signal related to
a displacement; and red cells represent a displacement detected in
the three areas where interferometric anomalies have been recog-
nised. This graphic representation also helps to temporally define
the instability phenomena, which occurred in a well-identified
time interval between June 27
th
2008 and June 30
th
2009.
None of the generated interferograms computed by pairing
images acquired before June 27
th
, 2008 show any anomalies that
may be related to displacements. The same results is obtained
for images acquired after that date. However, by pairing images
acquired before and after this date, we obtained interferograms
Fig. 8 - Interferometric anomaly detected in the area 1 (inside red boxes). The red bounds indicate the presence of anomaly, interpreted as displacements,
in the given interferogram
background image
A. ROCCA, P. MAZZANTI, D. PERISSIN & F. BOZZANO
44
Italian Journal of Engineering Geology and Environment, 1 (2014)
© Sapienza Università Editrice
www.ijege.uniroma1.it
characterised by the presence of observable displacements in the
phase signal. These results are confirmed by all of the interfero-
grams computed by using all possible combinations. Therefore,
in the time interval between June 2008 and June 2009, some in-
stability processes occurred in the three identified areas.
Considering the results shown in Fig. 8, Fig. 9 and Fig. 10, the
phase signals can be considered to quantitatively estimate detected
displacements starting from the angular values of the wrapped phase.
Because the PALSAR sensor operates with a 236 mm wavelength
(Tab. 4), the maximum detectable phase difference (equal to 2π and
represented by the transition from red to blue colour in the inter-
ferograms) corresponds to a displacement equal to λ/2 (118 mm)
along the LOS. Using the very high coherence attained, these weak,
small and localised indications of movements were clear enough to
be quantitatively estimated. The interferometric phase difference
between the anomalies and the surrounding pixels has been consid-
Fig. 9 - Interferometric anomaly detected in the area 2 (inside red boxes). The red bounds indicate the presence of anomaly, interpreted as displacements,
in the given interferogram
background image
DETECTION OF PAST SLOPE ACTIVITY IN A DESERT AREA USING MULTI-TEMPORAL DINSAR WITH ALOS PALSAR DATA
45
Italian Journal of Engineering Geology and Environment, 1 (2014)
© Sapienza Università Editrice
www.ijege.uniroma1.it
50x50 m) is very near the coast and is located on a portion of the
slope involved in the landslide that occurred in the period August
2011-June 2012. Because of the small dimensions of the area affect-
ed by displacements in the period under investigation and because
the detected displacements are very small (approximately 35 mm
along the LOS in the direction away from the satellite), it is not pos-
sible to directly relate the previously observed displacements with
the wider instability process that affected the slope in 2011-2012.
ered. Area 1 (Fig. 8), in the northern part of the study area, is ap-
proximately 90x70 m, located near the tectonic line coincident with
the coastal rock cliff. Considering all of the analysed interferograms,
a 40-45 mm LOS displacement away from the satellite was detected.
Area 2 (Fig. 9), smaller than the previous area, is very localised (ap-
proximately 40x40 m), showing slightly smaller displacements (ap-
proximately 30 mm along the LOS). In this case, the displacements
are in the direction away from the satellite. Area 3 (approximately
Fig. 10 - Interferometric anomaly detected in the area 3 (inside red boxes). The red bounds indicate the presence of anomaly, interpreted as displacements,
in the given interferogram. The belt with casual coloured pattern on the right is due to the sea surface, which is characterized by null coherence
background image
A. ROCCA, P. MAZZANTI, D. PERISSIN & F. BOZZANO
46
Italian Journal of Engineering Geology and Environment, 1 (2014)
© Sapienza Università Editrice
www.ijege.uniroma1.it
The low temporal resolution of the available dataset did not
allow better definition of the temporal development of the detect-
ed phenomena; thus, we have to accept a 1-year approximation
for such events. Furthermore, we do not know if the detected dis-
placements developed as single, impulsive events, or if they are
the result of slow processes that developed over a longer period
(several weeks or months). At the timescale of the adopted da-
taset, the detected displacements can be considered “impulsive”
because they were not distributed throughout the entire period of
the images. In addition, for this reason, such phenomena would
most likely not be detected by an automated A-DInSAR analysis,
even with a larger dataset.
The use of the L-band data, acquired by the ALOS PALSAR
satellite, also played a key role in the attainment of the results.
In addition to the advantages in terms of the high coherence dis-
cussed in section 1.3.3, the use of these data has the additional ad-
vantage of being less dependent on phase ambiguity due to high,
rapid displacements. In such circumstances, with displacements
that occurred in a time interval not covered by many images, or
with rapid, impulsive movements, the risk of losing information
due to wavelength limits was very high. Other data, such as C-
band (wavelength: ~56 mm) or even X-band (wavelength: ~30
mm) data, which are more affected by phase ambiguity problems,
would most likely not be able to provide such information with
λ/2 values equal to ~28 mm and ~15 mm, respectively.
DISCUSSION
The methodology used to investigate past displacements for
the present case study is based on a classical approach of SAR
interferometry, which is the direct examination of computed
interferograms. Many automated algorithms to extract informa-
tion from SAR data-stacks exist; however, they are likely not
capable of achieving the results herein shown. A-DInSAR meth-
ods could not provide trustable and reliable results by using such
poor data-stacks, especially in the very difficult conditions of this
case, with very small displacements circumscribed in restricted
areas, the presence of non-linear displacements, low temporal im-
age frequency, and the absence of further supporting information.
Detailed work manually analysing phase signal contributions
through a slope-oriented approach allowed us to extract more in-
formation than expected from the available data.
We began by distinguishing the topographic components
from the displacement phase contributions manually. It could
be argued that the height estimation (to refine the DEM used to
compute the differential interferograms) was partially performed
using automated methods, but that was not the case for the dis-
placement estimation. This is because also with the few avail-
able images, the estimation of heights remains a linear problem
dependent on the normal baseline and thus more easily accom-
plished. In contrast, the estimation of displacements is much
more controlled by the number of images.
Tab. 6 - Temporal distribution of detected interferometric anomalies along interferograms time series. Green boxes identify absence of anomalies; red
boxes identify presence of anomalies; dark grey boxes identify discarded interferograms in accordance with Tab. 3.5. Light grey boxes identify null
interferograms because of coincidence between master and slave images
background image
DETECTION OF PAST SLOPE ACTIVITY IN A DESERT AREA USING MULTI-TEMPORAL DINSAR WITH ALOS PALSAR DATA
47
Italian Journal of Engineering Geology and Environment, 1 (2014)
© Sapienza Università Editrice
www.ijege.uniroma1.it
the interferograms have been analysed in terms of their relation
with the normal baseline and the congruency with their temporal
distribution along the whole dataset.
The slope instability processes have been defined in time
(occurrence between June 2008-June 2009), and the related dis-
placements have been quantitatively estimated. The proposed
approach demonstrates the capability to derive information
about past displacements by satellite SAR images also in case
when advanced interferometric techniques are not feasible due
to the limited number of images in the dataset. Therefore, it can
be considered a useful investigative solution also in remote and
not urbanized areas where a limited knowledge on geomorpho-
logical and geological settings is available and where monitor-
ing data about past evolution of displacement phenomena do
not exist. It can also provide an early prior indication that can
be deepened during in-situ investigations related to major con-
struction on large areas.
ACKNOWLEDGMENTS
The present work has been carried out thanks to the European
Space Agency in the frame of the Cat-1 project “Landslides fore-
casting analysis by time series displacement derived from Satel-
lite and Terrestrial InSAR data” (Id 9099) and Rocksoil S.p.A.,
which provided basic geological data.
CONCLUSIONS
The evolution of instability processes affecting a coastal
slope in the Sultanate of Oman has been defined by using sat-
ellite SAR interferometry. Thanks to archive data provided by
the Japanese ALOS PALSAR satellite, an investigation based
on the combination of DInSAR and QPS techniques allowed
us to achieve quantitative results about surface displacements
occurred within the time-span covered by the SAR images (De-
cember 2006-April 2010).
The approach used in this work allowed us to overcome the
limitation caused by the limited number of images in the area,
which was not sufficient to perform A-DInSAR analyses. The
QPS method has been used to obtain a more accurate DEM to
better discriminate the displacement phase signal from the topo-
graphic information.
A manual interpretation of the generated interferograms has
been carried out to properly observe the occurred displacements.
The application of an automated algorithm, in fact, could eas-
ily disguise them because they were not continuous in space
and time. More specifically, the used approach was carried out
by analysing temporal (i.e. acquisition dates of the images) and
geometrical (i.e. normal baseline) parameters characterising all
interferograms that were computed with the highest possible re-
dundancy. The phase signals related to displacement detected in
REFERENCES
B
Arrett
E.C. (2013) - Introduction to environmental remote sensing. Routledge.
B
échennec
f., l
e
M
étour
J., P
lAtel
J.P. & r
oger
J. (1993) - Geological map of the Sultanate of Oman, scale 1:1,000,000. Oman Ministry of Petroleum
and Minerals, Directorate General of Minerals.
B
erArdino
P., f
ornAro
g., l
AnAri
r. & s
Ansosti
e. (2002) - A new algorithm for surface deformation monitoring based on small baseline differeential SAR
interferograms. IEEE Trans. Geosci. Remote Sensing. 40: 2375-2383.
B
ozzAno
f. & r
occA
A. (2012) - Remote monitoring of deformation using satellite SAR interferometry. Geotechnical News, 30(2): 26.
c
Arnec
c., M
Assonnet
d. & k
ing
c. (1996) - Two examples of the use of SAR interferometry on displacement fields of small spatial extent. Geophysical
Research Letters, 23(24): 3579-3582.
c
hen
f., l
in
h., l
i
z., c
hen
Q. & z
hou
J. (2012) - Interaction between permafrost and infrastructure along the Qinghai-Tibet Railway detected via jointly
analysis of C-and L-band small baseline SAR interferometry. Remote Sensing of Environment, 123: 532-540.
c
hen
f., l
in
h. & h
u
X. (2014) - Slope superficial displacement monitoring by small baseline SAR interferometry using data from L-band ALOS PALSAR
and X-band TerraSAR: a case study of Hong Kong, China. Remote Sensing, 6(2): 1564-1586. doi:10.3390/rs6021564
f
erretti
A., P
rAti
c. &. r
occA
f. (2001) - Permanent scatterers in SAR interferometry. IEEE Trans. Geosc. and Remote Sens., 39(1): 8-20.
f
erretti
A., f
uMAgAlli
A., n
ovAli
f., P
rAti
c., r
occA
f. & r
ucci
A. (2011) - A new algorithm for processing interferometric data-stacks: SqueeSAR.
Geoscience and Remote Sensing, IEEE Transactions on, 49(9): 3460-3470.
f
ournier
M., B
ellAhsen
n., f
ABBri
o. & g
unnell
y. (2004) - Oblique rifting and segmentation of the NE Gulf of Aden passive margin. Geochemistry,
Geophysics, Geosystems, 5(11).
f
runeAu
B., A
chAche
J. & d
elAcourt
c. (1996) - Observation and modelling of the Saint-Étienne-de-Tinèe landslide using SAR interferometry.
Tectonophysics, 265: 181-190.
g
ABriel
A.k. & g
oldstein
r.M. (1988) - Crossed orbit interferometry: theory and experimental results from SIR-B. Int.J. Remote Sensing, 9(5): 857-872.
g
ArcíA
-d
AvAlillo
J. c., h
errerA
g., n
otti
d., s
trozzi
t. & Á
lvArez
-f
ernÁndez
i. (2014) - DInSAR analysis of ALOS PALSAR images for the assessment
of very slow landslides: the Tena Valley case study. Landslides, 11: 225-246. doi:10.1007/s10346-012-0379-8.
g
hezzi
g., P
erAltA
J.c., B
iAnchi
A., s
Artini
s., A
nfuso
A., c
reAtini
f., c
Assitelli
M., g
hezzi
r., P
ellegrini
M., P
orsiA
d. & r
izzA
l. (2012) - Geophysical
site investigation for Hasik landslides stretch from km 7+000 to km 8+500 (Unpublished technical report).
background image
A. ROCCA, P. MAZZANTI, D. PERISSIN & F. BOZZANO
48
Italian Journal of Engineering Geology and Environment, 1 (2014)
© Sapienza Università Editrice
www.ijege.uniroma1.it
g
oldstein
r., z
eBker
h.A. & W
erner
c.l. (1988) - Satellite radar interferometry: two-dimensional phase unwrapping, Radio Science, 23(4):713-720.
g
oldstein
r., e
ngelhArdt
h., k
AMB
B. & f
rolich
r. (1993) - Satellite radar interferometry for monitoring ice-sheet motion: application to an Antarctic
ice stream. Science, 262: 1525-1530.
h
Anssen
R. (2001) - Radar interferometry, data interpretation and error analysis. Kluwer Academic Publishers.
h
ooPer
A., z
eBker
h., s
egAll
P. & k
AMPes
B. (2004) - A new method for measuring deformation on volcanoes and other natural terrains using InSAR
persistent scatterers. Geophysical Research Letters, 31(23): L23611. doi:10.1029/2004GL021737
J
eBur
M.n., P
rAdhAn
B. & t
ehrAny
M.s. (2013) - Using ALOS PALSAR derived high-resolution DInSAR to detect slow-moving landslides in tropical forest:
Cameron Highlands, Malaysia. Geomatics, Natural Hazards and Risk (ahead of print): 1-19. doi:10.1080/19475705.2013.860407.
k
AMPes
B.M. (2006) - Radar interferometry persistent scatterers technique. (Springer Ed.) - Dordrecht, The Netherlands.
k
Wok
r. & f
Ahnestock
M.A. (1996) - Ice sheet motion and topography from radar interferometry. IEEE Transactions on Geoscience and Remote Sensing,
34(1): 189-200. doi:10.1109/36.481903.
l
AnAri
r., l
undgren
P., M
Anzo
M. & c
Asu
f. (2004) - Satellite radar interferometry time series analysis of surface deformation for Los Angeles, California.
Geophysical Research Letters, 31(23).
l
Auknes
T.R. (2004) - Long-term surface deformation mapping using small-baseline differential SAR interferograms. Unpublished Master Thesis,
Department of Physics and Technology, University of Tromsø.
l
ePvrier
c., f
ournier
M., B
érArd
t. & r
oger
J. (2002) - Cenozoic extension in coastal Dhofar (southern Oman): implications on the oblique rifting of the
Gulf of Aden. Tectonophysics, 357(1): 279-293.
l
i
f. & g
oldstein
r. (1987) - Studies of multi-baseline spaceborne interferometric synthetic aperture radar. In: International Geoscience and Remote
Sensing Symposium, Ann Arbor, 18-21 May 1987: 1545-1550.
l
i
f.k. & g
oldstein
r.M. (1990) - Studies of multibaseline spaceborne interferometric synthetic aperture radars. IEEE Transactions on Geoscience and
Remote Sensing, 28(l): 88-97.
M
Assonnet
d., r
ossi
M., c
ArMonA
c., A
drAgnA
f., P
eltzer
g., f
eigl
k. & r
ABAute
t. (1993) - The displacement field of the Landers earthquake mapped
by radar interferometry. Nature, 364(6433): 138-142.
M
Assonnet
d., f
eigl
k., r
ossi
M. & A
drAgnA
f. (1994) - Radar interferometric mapping of deformation in the year after the Landers earthquake. Nature,
369(6477): 227-230.
M
Assonnet
d., B
riole
P., & A
rnAud
A. (1995) - Deflation of Mount Etna monitored by spaceborne radar interferometry. Nature, 375(6532): 567-570.
M
ercolli
i., B
riner
A.P., f
rei
r., s
chönBerg
r., n
ägler
t.f., k
rAMers
J. & P
eters
t. (2006) - Lithostratigraphy and geochronology of the Neoproterozoic
crystalline basement of Salalah, Dhofar, Sultanate of Oman. Precambrian Research, 145: 182-206.
P
erissin
D. (2008) - Validation of the submetric accuracy of vertical positioning of PSs in C-band. Geoscience and Remote Sensing Letters, IEEE, 5(3): 502-506.
P
erissin
D. (2009) - SARPROZ Manual http://ihome.cuhk.edu.hk/~b122066/index_files/download.htm.
P
erissin
d., W
Ang
z. & W
Ang
t. (2011) - The SARPROZ InSAR tool for urban subsidence/manmade structure stability monitoring in China. Proc. of ISRSE
2010, Sidney, Australia, 10-15 April 2011.
P
erissin
d. & W
Ang
t. (2012) - Repeat-Pass SAR Interferometry with partially coherent targets. IEEE Trans. on Geosc. and Remote Sens. 50(1): 271, 280.
P
erissin
d., W
Ang
z. & l
in
h. (2012) - Shanghai subway tunnels and highways monitoring through Cosmo-SkyMed Persistent Scatterers. ISPRS Journal
of Photogrammetry and Remote Sensing, 73: 58-67.
P
lAtel
J.P. & r
oger
J. (1989) - Evolution geodynamique du Dhofar (Sultanat d’Oman) pendant le Crétacé et le Tertiaire en relation avec l’ouverture du
golfe d’Aden. Bull. Soc. Geol. Fr., 2: 253-263.
P
lAtel
J.P., r
oger
J., P
eters
t.J., M
ercolli
i., k
rAMers
J.d. & l
e
M
e
´
tour
J. (1992) - Geological map of Salalah, Sultanate of Oman; sheet NE 40-09, scale
1:250000, Oman Ministry of Petroleum and Minerals, Directorate General of Minerals.
P
rAti
c., r
occA
f., M
onti
g
uArnieri
A. & d
AMonti
e. (1990) - Seismic Migrataion For SAR Focussing: Interferometrical Applications. IEEE Transactions
on Geoscience and Remote Sensing, 28(4): 627-640.
r
iedel
B. & W
Alther
A. (2008) - InSAR processing for the recognition of landslides. Advances in Geosciences, 14(14): 189-194.
r
oBertson
A.h.f. & B
AMkhAlif
k.A.s. (2001) - Late Oligocene-early Miocene rifting of the northeastern Gulf of Aden: basin evolution in Dhofar (southern
Oman). In z
iegler
P.A., c
AvAzzA
W., r
oBertson
A.h.f. & s. c
rAsQuin
-s
oleAu
S. (e
ds
) - Peri-Tethys Memoir 6: Perihal- 00021620, version Tethyan
Rift/Wrench Basins and Passive Margins. Mém. Mus. Natn. Hist. Nat., 186: 641-670.
r
oger
J., P
lAtel
J.P., c
Avelier
c. & B
ourdillon
de
g
rissAc
c. (1989) - Donntes nouvelles sur la stratigraphie et l’histoiregtologique du Dhofar (Sultanat
d’Oman). Bull. Sot. GCol. Fr. 8(2): 265-277.
r
oger
J., P
lAtel
J.P., B
erthiAuX
A. & l
e
M
e
´t
our
J. (1992) - Geological map of Hawf with Explanatory Notes; sheet NE 39-16, scale 1:250000, Oman
Ministry of Petroleum and Minerals, Directorate General of Minerals.
s
iMons
M., f
iAlko
y. & r
iverA
l. (2002) - Coseismic deformation from the 1999 Mw 7.1 Hector Mine, California, earthquake as inferred from InSAR and
background image
DETECTION OF PAST SLOPE ACTIVITY IN A DESERT AREA USING MULTI-TEMPORAL DINSAR WITH ALOS PALSAR DATA
49
Italian Journal of Engineering Geology and Environment, 1 (2014)
© Sapienza Università Editrice
www.ijege.uniroma1.it
GPS observations. Bulletin of the Seismological Society of America, 92(4): 1390-1402.
s
ingh
l.P., W
esten
c.J., c
hAMPAti
r
Ay
P.k. & P
AsQuAli
P. (2005) - Accuracy assessment of InSAR derived input maps for landslide susceptibility analysis:
a case study from the Swiss Alps. Landslides, 2(3): 221-228. doi:10.1007/s10346-005-0059-z
s
inghroy
v., M
AttAr
k. & g
rAy
A. (1998) - Landslide characterisation in Canada using interferometric SAR and combined SAR and TM images. Advances
in Space Research, 21(3).
s
trAMondo
s., B
ozzAno
f., M
ArrA
f., W
egMuller
u., c
inti
f.r., M
oro
M. & s
Aroli
M. (2008) - Subsidence induced by urbanisation in the city of Rome
detected by advanced InSAR technique and geotechnical investigations. Remote Sensing of Environment, 112(6): 3160-3172.
s
trozzi
t., f
ArinA
P., c
orsini
A., A
MBrosi
c., t
hüring
M., z
ilger
J., W
iesMAnn
A., W
egMuller
u. & W
erner
c. (2005) - Survey and monitoring of landslide
displacements by means of L-band satellite SAR interferometry. Landslides, 2(3): 193-201. doi:10.1007/s10346-005-0003-201
s
trozzi
t., d
elAloye
r., k
ääB
A., A
MBrosi
c., P
erruchoud
e. & W
egMüller
u. (2010) - Combined observations of rock mass movements using satellite
SAR interferometry, differential GPS, airborne digital photogrammetry, and airborne photography interpretation. Journal of Geophysical Research:
Earth Surface, 115(F1): 2003-2012.
t
ong
X., s
AndWell
d.t. & f
iAlko
y. (2010) - Coseismic slip model of the 2008 Wenchuan earthquake derived from joint inversion of interferometric
synthetic aperture radar, GPS, and field data. Journal of Geophysical Research: Solid Earth, 115(B4): 1978-2012.
v
An
l
eiJen
f. J. & h
Anssen
r.f. (2007) - Persistent scatterer density improvement using adaptive deformation models. In Geoscience and Remote Sensing
Symposium, 2007. IGARSS 2007. IEEE International: 2102-2105. IEEE.
W
Atchorn
f., n
ichols
g.J. & B
osence
d.W.J. (1998) - Rift-related sedimentation and stratigraphy, southern Yemen (Gulf of Aden). In: P
urser
B.h.,
B
osence
D.W.J. (e
ds
.) - Sedimentation and Tectonics of Rift Basins: Red Sea- Gulf of Aden. Chapman & Hall, London: 165-191.
W
ei
M. & s
AndWell
D.T. (2010) - Decorrelation of L-band and C-band interferometry over vegetated areas in California. Geoscience and Remote Sensing,
IEEE Transactions on, 48(7): 2942-2952.
Received February 2014 - Accepted May 2014
background image
Statistics