SUBSIDENCE DETECTION USING INSAR AND GEODETIC MEASUREMENTS IN

NORTH-WEST OF IRAN

Morteza Sedighi (1)

, Siavash Arabi(1)

, Hamid Reza Nankali(1)

, Masomeh Amighpey(1)

, Farokh Tavakoli(1)

, Ali

Soltanpour(1)

,Mehdi Motagh(2)

(1)National Cartographic Center (NCC), Meraj Ave. – Azadi Sq. - Tehran -IRAN, Email:Sedighi@ncc.org.ir (2)Tehran University, Karegar-e-Shomali St.(Kouye Daneshgah) -Tehran -IRAN

ABSTRACT

Geodetic measurements i.e., repeated leveling

measurements of first order leveling network of Iran and

continuous GPS measurements of Iranian Permanent

GPS Network of Iran (IPGN), showed that there is

subsidence in the north-west of Iran. In this paper we try

to find the area and rate of subsidence in Ghara-Geshlagh in north-west of Iran using InSAR and

geodetic techniques. The result of InSAR technique

show a better understanding on this phenomenon in

Salmas and Ghara-Gheshlagh area and have a good

consistency with accurate geodetic measurements.

1. INTERODUCTION

The subsidence of the Earth surface is a phenomenon

that occurs in some places in the world which overuse

underground sources of water. Iran has semi-arid and

arid climate and the rate of rainfall is lower than the

mean rate in the world. Now we are encountered by over-exploitation of groundwater in agricultural areas

and also for extending the cities and industrial areas.

From the past, geodetic measurements such as leveling

survey have been used for subsidence investigation to

map the height variations. It can measure the height

variations with very high precision, but only provides us

the data at limited numbers of stations along the routes.

When GPS came up with a fast and accurate technique

for measuring the earth surface displacement, we

attained a faster and cheaper measurement with respect

to leveling. The main advantage of GPS is the capability to measure real time 3D component displacements but,

similar to leveling, these measurements are also limited

to some scattered points (permanent GPS stations or

campaign GPS networks).

For studying a wide area deformation such as

subsidence phenomena, we have to design a dense

geodetic (leveling and/or GPS) network to estimate the

area and the rate of displacement. Although the geodetic

techniques have good accuracy for detecting the Earth

surface deformation but these techniques are not

increasingly competent for the large-scale deformation monitoring such as subsidence. Compared with leveling

and GPS measurements, which provide information at

the specific observation points, the Interferometric

Synthetic Aperture (InSAR) technique can provide

accurate subsidence information over widespread areas

with high spatial resolution. In the other words, the

InSAR provides an effective tool to detect and measure

the magnitude and spatial variation of the earth surface

displacement such as subsidence especially when it is combined with geodetic techniques. Indeed, InSAR only

measures displacement along the satellite Line Of Sight

(LOS) direction, and it is most sensitive to vertical

motion. Therefore it is an appropriate technique for

measuring subsidence.

In the InSAR technique two complex SAR images

(recorded by antennas mounted on satellite or airplane)

are combined to generate surface deformation maps

and/or DEM (Digital Elevation Model). The

interferograms are generated by multiplication of the

first image (Master) to the complex conjugate of the

second image (Slave). The result of this multiplication contains the phase difference of two images. This fringe

pattern reflects the ground displacement that has

occurred between the two acquisition times (Master and

Slave). The observed phase (∅int ) is the sum of several

contributions (Eq. 1):

∅int = ∅topography + ∅displacement + ∅atmospher +

∅orbit + ∅noise (1)

These components are topographic phase, surface

displacement signal (phase), atmospheric delay, orbital

phase, and noise respectively. The orbital error (affected

phase by orbital parameters) can be removed using

precise orbital parameters e.g., DEOS orbital

parameters. The topographic term is removed from each

interferogram using DEM e.g., Shuttle Radar

Topography Mapping (SRTM) DEM. The displacement

signal is the component that we attempt to extract from interferograms. Each fringe cycle in an interferogram

corresponds to half the radar wavelength (that is 28 mm

for ENVISAT satellite data) ground displacement along

the LOS direction.

2. STUDY AREA

National Cartographic Center (NCC) of Iran has

established and re-measured the first order levelling

_____________________________________________________ Proc. ‘Fringe 2009 Workshop’, Frascati, Italy, 30 November – 4 December 2009 (ESA SP-677, March 2010)

network of Iran during the past 20 years. Also NCC has

established 110 permanent GPS stations (since 2004) to

monitor ground surface movement for geodynamic

purposes (Figs. 1a, 1b). Repeated levelling (on some

levelling lines) and (some) permanent GPS

measurements showed that there are subsidences in

many parts of Iran. Salmas and Ghara-Gheshlag area as

a large agricultural zone is one of the subsiding areas located in the north-west of Iran (north-west of Urumieh

Lake) (Fig. 2).

3. DATA

The levelling line (AGAH) from Khoy to Urumieh

measured in two epochs in 1989 and 2005. These

measurements revealed subsidence in some levelling

benchmarks near Salmas. The subsidence rate

determined by levelling measurements during 17 years

is about 6 cm/year (Fig. 3). Also the GPS permanent

station in Ghara-Gheshlagh (GGSH) showed subsidence

with the same rate with nonlinear behaviour (Fig. 4).

Figure 3. Elevation changes in Levelling line (AGAH)

1989-2005

Figure 1a. First Order Levelling Network of Iran and

Detected Subsidence Areas Figure 1b. Iranian Permanent GPS Network

(IPGN)

m

Figure 4. GPS time series of GGSH permanent station –

h component

In order to study the temporal behaviour of the

deformation in high spatial resolution, the InSAR observations acquired by ENVISAT satellite are used to

investigate land subsidence in Salmas and Ghara-

Gheshlagh area caused by groundwater

overexploitation. Our dataset consists of 11 images

acquired by ENVISAT satellite chosen from descending

track 92 and ascending track 228 passes from August

2003 to November 2007. The spatial and temporal

coverage of these images are presented in Fig. 5&

Tab.1.

Figure 5. Location of ENVISAT images

Table 1. ENVISAT data used in this project No. Orbit Number Acquisition Date

1. 8162 2003-09-22

2. 9164 2003-12-01

3. 10667 2004-03-15

4. 11669 2004-05-24

5. 15312 2005-02-02

6. 16815 2005-05-18

7. 17817 2005-07-27

8. 20186 2006-01-09

9. 26198 2007-03-05

10. 28202 2007-07-23

11. 29705 2007-11-05

The selected interferograms have the perpendicular

baseline less than 400m and the time span from 70 days

to 8 months (The normal orbital cycle for each satellite

is 35 days). In this paper analysis of 6 interferograms

are presented.

4. PROCESSING

The InSAR images were acquired from ESA, are

processed using InSAR processing software "DORIS"

with Delft orbits (DEOS). The topographical effect in

the interferograms was removed from each

interferogram using a 30m SRTM-DEM. The

"Goldstien filter" used for interferogram filtering.

"SNAPHU" software was used for unwrapping the interferograms, i.e., converting from phase to

displacement along the satellite line of sight. By

considering that the displacement is a pure vertical

movement, the displacements were projected into the

vertical direction.

Fig. 6 shows six interferograms of the study area. As it

was expected, the fringes cover whole parts of plain

between Salmas and Ghara-Gheshlagh.

Figure 6. Four interferogram in Salmas area with

temporal baseline between 70-240 day

5. TIME SERIES ANALYSIS OF SUBSIDENCE

AREA BASED ON LEAST SQUARES

METHOD

For obtaining a time series of subsidence in the study

area, Least squares method is used for adjusting the

observations. The result of this adjustment can give us a

better view of subsiding (from the availability of SAR

images). Each interferogram provides the displacement

of the earth surface for a period of time between master

and slave images. The time of first image is considered

as the origin of subsidence (zero subsidence) and then

the absolute subsidence for other images are computed

with respect to the origin using Least square technique.

The result of this adjustment is represented in Fig. 7.

Also, the DEM of the area is represented in Fig. 8.

6. RESULTS

As it was expected, subsidence was observed in Salmas,

Ghara-Gheshlagh plain. Although the GPS and levelling

measurements show about 6cm/year subsiding rate for

some points in the area which is consistent with results

from InSAR for the same points, new results from

INSAR have provided an extended spatial pattern of the

subsidence with highlighting the points with maximum

rate of subsidence.

7. ACKNOWLEDGEMENT

We would like to thank the European Space Agency

(ESA) that provides SAR images (ENVISAT data)

according to the proposal 6244 (C1P.6244).

Figure 7. Adjusted subsiding maps

Figure 8. DEM map of study area