LONG TERM CONTINUOUSLY DInSAR FOR VOLUME CHANGE ESTIMATION OF LAND DEFORMATION

Josaphat Tetuko Sri Sumantyo (1), Masanobu Shimada (2), Pierre-Phillippe (3) and Hasanuddin Zainal Abidin (4)

(1) Center for Environmental Remote Sensing (CEReS), Chiba University, Chiba, 263-8522 Japan e-mail: jtetukoss@ faculty.chiba-u.jp

(2) Earth Observation Research Center, Japan Aerospace Exploration Agency (JAXA), 350-8505, Tsukuba, Japan. e-mail: shimada.masanobu@jaxa.jp

(3) Earth Observation Science & Applications, European Space Agency (ESA), Frascati (RM), Italy. e-mail: Pierre.Philippe.Mathieu@esa.int

(4) Department of Geodesy, Institute of Technology Bandung, Bandung, Indonesia e-mail : hzabidin@gd.itb.ac.id

ABSTRACT

In this research, a technique to estimate the volume change by using Differential Interferometric Synthetic Aperture Radar (DInSAR) is proposed to retrieve the volume change of long-term continuously land deformation (subsidence or uplift). This technique is employed to investigate subsidence of Bandung city, Indonesia to demonstrate the capability of this technique by assessing two Japanese L band Spaceborne SARs (JERS-1 SAR and ALOS PALSAR) in continuously periods of 1993-1997 and 2007-2008. This result shows the subsidence occurred concentrating seriously at industrial complexes and settlements of Bandung area. The result is confirmed by using 2000-2008 period of GPS observation data, statistical data of 1991-2008 published by The Indonesian Statistics Bureau (BPS), three campaigns of ground surveys to compiling the recent condition of buildings and other infrastructures, population, ground water level data and geological formation of this area. This research shows a capability of this technique to investigate the volume change caused by land deformation, especially subsidence. 1. INTRODUCTION

Japan Aerospace Exploration Agency (JAXA) or former National Space Development Agency of Japan (NASDA) already has been operating two Synthetic Aperture Radar onboard satellites, there are Japanese Earth Resources Satellite Synthetic Aperture Radar (JERS-1 SAR) and Advanced Land Observation Satellite - Phased Array type L-band Synthetic Aperture Radar (ALOS PALSAR). Partially, the JERS-1 SAR had operated within six years since 15 April 1992 and recorded huge number of satellite imageries (a scene covers 75 km x 75 km). SAR is known well as an active sensor that could observe the object in day and night. Various methods to analyze SAR image also already have been developed to extract the physical information as interferometry, soil moisture, biomass, soil type etc. The applications of Japanese SAR sensors also have been proposed by some researchers to observe the Earth

surface, as land deformation, Cryosphere, agriculture, forestry, ocean dynamics etc. Previously, some methods already have developed to extract volume change by using digital elevation model (DEM), map-based topographic approach, generalized tri-prism (GTP), interferometric coherence etc. However, it is difficult to obtain high precision observation to measure the accurate change of the Earth surface from the space by employing these methods. In this paper, the authors propose a technique to extract the volume change on the Earth surface by improving the Differential Interferometric SAR (DInSAR) technique. 2. VOLUME CHANGE ESTIMATION OF LAND

DEFORMATION

In this research, we will improve the implementation of phase information of two complex data well-known as Differential Interferometric Synthetic Aperture Radar (DInSAR) for long-term continuously land surface deformation by proposing a technique to retrieve the volume change caused by land deformation, especially subsidence or uplift. Firstly, by modifying some equations in Masanobu Shimada’s thesis for DInSAR (Shimada 1999), we could obtain the equations below.

sindxcosdzdc j,ij,ij,i (1)

j,ihp

m dcBsinr

z

(2)

where is wavelength. i, j (=1, 2, …) shows sub-pixel position. Then Bh and Bp are horizontal and perpendicular baseline distance. is off-nadir angle. is incidence angle on a deformed pixel. j,idx and j,idz

are the displacement of a sub-pixel or a small area at

range and elevation directions. j,idc is the total

displacement of a sub-pixel. rm is slant range distance. We assume the incidence angle on a deformed sub-pixel is same with the incidence angle at whole area

within a pixel. Hence the phase difference shows the

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

Fig. 1. Study area : (a) Study area covers part of Kotamadya Bandung, Kabupaten Bandung and Kotamadya Cimahi

(b) Study area is determined by industrial area and settlement areas that effected by the subsidence (c) False image of ASTER VNIR (3 February 2006) which white area shows the industrial complex (IC)

TABLE I. CODE, COVERAGE AND POPULATION OF THE INDUSTRIAL AND SETTLEMENT AREAS OF THE STUDY AREA

Area Code

District Code and Name Coverage

(km2)

Population in 2008 (people)

Population Density in 2008

(people/km2)

Ground Water Level in meters (number of surveyed boreholes)

Number of arthesis well

in 1995 1980s 1990s 2000s

Industrial Areas

IC1 IC11 IC12 IC13 IC14 IC15 IC16 IC17 IC18

Cimahi Utara Cimahi Tengah Cimahi Selatan Andir Batujajar Bandung Kulon Cicendo Sukajadi

13.32 10.10 16.94

3.71 83.68

6.46 6.86 4.30

137,622169,484229,637103,975106,724125,36999,452

100,244

10,33216,78113,55628,025

1,27519,40714,497

2,382

9.5 (3)5.5 (12)

12.5 (24)27.2 (6)

- (2)13.5 (16)16.0 (13)

- (0)

23.0 26.0 44.0 31.2 42.3 41.8 27.4

-

25.0 28.0 48.0 35.0

- 43.0 31.0

-

25,00117,32523,73552,757

4,00045,21252,75742,838

IC2 IC21 IC22 IC23

Cililin Soreang Margaasih

81.55 67.37 17.97

182,455147,584119,442

2,2372,1916,647

- (0)- (0)- (5)

- -

29.0

10.0 -

30.0

9,8228,334

35,431IC3 IC31

IC32 Margahayu Katapang

10.54 21.16

114,510122,038

10,8645,767

4.8 (3)5.75 (4)

8.15 30.85

12.0 32.0

7,6117,438

IC4

IC41 IC42 IC43 IC44 IC45 IC46

Dayeuhkolot Bojongsoang Bandung Kidul Bojongloa Kidul Astanaanyar Regol

11.03 27.34

6.06 6.26 2.89 4.30

113,08278,95150,10979,47870,64883,713

10,2522,8888,269

12,69624,44619,468

8.5 (16)4.0 (3)4.1 (4)

- (1)- (1)

5.3 (1)

34.0 19.0 27.9 17.9 5.4

-

36.0 21.0 29.0 20.0 7.0 7.0

7,1783,217

17,54239,14827,82729,243

IC5 IC51 Baleendah 41.82 178,060 4,258 12.8 (6) 45.7 47.0 8,786IC6 IC61

IC62 IC63

Arjasari Pameungpeuk Banjaran

64.98 14.62 42.92

87,19462,634

161,906

1,3424,2843,772

- (0)6.82 (2)

- (0)

- - -

- - -

3,3464,8338,551

IC7 IC71 IC72

Majalaya Ciparay

25.36 46.18

225,794142,008

8,9043,075

- (3)- (1)

7.0 -

8.0 7.0

4,51011,145

IC8 IC81 Rancaekek 45.30 155,004 3,421 4.3 (4) 18.2 20.0 6,158

Settlement Areas

ST1 ST11 ST12

Babakan Ciparay Bojongloa Kaler

7.45 3.03

137,392118,898

18,44239,240

14.0 (1)4.7 (3)

- 17.1

16.0 18.0

53,25437,426

ST2 ST21 ST22 ST23 ST24

Cileunyi Cibiru Ujung Berung Rancasari

31.58 10.81 10.34 13.17

125,58089,20184,93172,309

3,9778,2528,2145,490

- (1)11.0 (4)

9.1 (9)- (4)

- -

20.4 13.3

10.0 13.0 23.0 15.0

3,96832,55939,92730,927

ST3 ST31 ST32 ST33 ST34 ST35 ST36 ST37

Sumur Bandung Lengkong Batu Nunggal Kiara Condong Cicadas Arcamanik Margacinta

3.40 5.90 5.03 6.12 8.66 8.80

10.87

39,38372,059

122,345128,121108,24568,860

112,325

11,58312,21324,32320,93412,499

7,82510,333

15.2 (9)18.0 (3)

9.5 (5)4.5 (4)

15.5 (4)12.5 (5)11.0 (7)

30.2 26.6 32.2 22.5 22.9 28.8 18.5

44.3 28.0 35.0 25.0 25.0 30.0 21.3

6,38713,74052,75738,18534,11526,24325,776

topographical information ( z ) and total surface deformation (dci,j) in sub pixel as shown in Eq. (2). If we consider only the land deformation at vertical direction (uplift or subsidence) and assume the

horizontal deformation could be neglected or j,idx in

(1) to be zero. The vertical deformation could be derived by employing the Eqs. (1) and (2) as

h

m

pj,i B

sinr

Bz

cosdz

4

1 (3)

Then the volume unit of a sub-pixel of land deformation in vertical direction is defined by using change difference at range, azimuth and elevation directions as

j.ij,ij,ij,i zqrV (4)

where j,ir and j,iq are sub-pixel resolution in

ground range and azimuth direction. Then j,iz

(= j,idz ) is a change unit in elevation or z direction. As

shown in next equation, the positive and negative

sign( j,iz ) show uplift and subsidence, respectively.

subsidencezsign

upliftzsign

ji,

ji,

j,iz (5)

Total volume of land deformation in a pixel (Vpixel) can be calculated by using

N

i

N

jj,ij,ij,i

N

i

N

jj,ipixel zqrVV

1 11 1

(6)

N

i

N

jh

m

pj,ij,i B

sinr

Bz

cosqr

1 1 4

1

Actually, the size of a pixel surface area is shown as the multiplication of spatial resolution of ground range and azimuth direction.

N

i

N

jj,ij,i qrS

1 1

(7)

To simplify the calculation, we employ conditions

j,ij,i q,r Q,R (spatial resolution of a pixel

in ground and azimuth directions), then the incidence

angle and slant range of master rm of sub-pixels

within a pixel is assumed as the same value. Finally, the total deformed volume of a pixel is acquired as

h

n,m

n,mm

pn,m

n,mn,m B

sinr

Bz

cosSV

4

1

(8) where m and n show line and row of the pixel in the SAR imagery. This equation shows that the volume

difference of each pixel in an SAR imagery has a

function of elevation n,mz , nadir angle n,m , slant

range distance rm, perpendicular and horizontal

baselines Bp and Bh, incidence angle n,m , and phase

difference n,m (post unwrapping phase).

In this research, the raw SAR data were processed using the JAXA/SIGMA SAR software which developed by Shimada (1999) to derive the interferograms and other parameters employed in Eq. (8), where the interferograms of JERS-1 SAR pairs and ALOS PALSAR pairs discussed in this paper are resampled to 12.5 m resolution. Base on the definition to derive the equations in this Section, the newest image is employed as slave and the latest one is as master. Preconditioned conjugate gradient (PCG) method is used for unwrapping technique and noise in the interferogram was filtered twice or three times by using Goldstein and Werner filter until the satisfied interferogram is obtained. Geocoded-image generation is done by using cubic convolution interpolation in UTM projection. The GIS processing is also employed in this research to process the interferogram, GPS observed data, statistics, geological data, ground survey, digital map manipulation etc to analyze the subsidence and its impact on the study area. The detail result is discussed in next Section. 3. STUDY AREA : BANDUNG CITY,INDONESIA

Bandung city is located in region of west Java and capital of province of West Java, Indonesia. It is located at longitude of 107o10’48”- 108o13’12”E and latitude of 6o10’48”- 7o44’24”S, see Fig. 1(a). Topographically, the center of Bandung city is located at 791 m above the sea level; the highest point area in the north is 2,084 m and the lowland in the south which is around 64 m above sea level. The land in south Bandung is relatively flat, whereas Bandung area is surrounded by mountain creating Bandung basin. The study area consists of Kotamadya (Municipality) Bandung, Kabupaten (Regency) Bandung and Kotamadya Cimahi as shown in Fig. 1 (a) with governmental boundary as blue, yellow, and sea blue lines respectively. The study area covers 345 km2, see Fig. 1(b). Fig. 1(c) shows the false color composite of the study area that is derived from ASTER VNIR acquired on February 3, 2006. The industrial areas (ICs) is shown clearly as the white color areas. The industrial area is distributed at Cimahi, Dayeuhkolot, Cibiru, Rancaekek, Cilampeni, Balaendah, Arjasari, and Majalaya. The settlement area (ST) is distributed at Babakan Ciparay, Cileunyi, and Sumur Bandung areas. Babakan Ciparay and Bojongloa Kaler areas have the highest number of arthesis well and population. The distribution of surveyed arthesis well or borehole is shown in Fig. 2.

Fig. 2. Geological formation of the study area and location of borehole and GPS observation points

1991 1992 1993

1994 1995 1996

1997 1998 1999

2000 2001 2002

2005 2006

2007

2003 2004

2008

Legend : Population

0 - 25,000

25,001 - 50,000

50,001 - 75,000

75,001 - 100,000

100,001 - 125,000

125,001 - 150,000

150,001 - 175,000

175,001 - 200,000

200,001 - 225,000

225,001 - 250,000

1991 1992 1993

1994 1995 1996

1997 1998 1999

2000 2001 2002

2005 2006

2007

2003 2004

2008

Legends : Number of Industries

0 - 50

51 - 100

101 - 150

151 - 200

201 - 250

251 - 300

Fig. 3. Population of study area from 1991 to 2008 Fig. 4. Number of industries of study area from 1991 to 2008

As shown in Fig. 1(b) and Table I, we classified eight industrial complexes (ICs) and three settlement area (STs) to analyze the effect of human and industrial activities, and to investigate the land deformation (subsidence) in the study area. Area code, detail district code, coverage, population in 2008, population density of each area is shown in Table I that was compiled from BPS statistics (1990-2008). Table I also shows the number of manmade arthesis well of the study area in 1995. People in this area use ground water from arthesis well for daily consumption and many industries are consuming ground water for industrial purposes, i.e. textile, food etc. The trend of ground water level is surveyed in 1980s, 1990s, and 2000s as shown in Table I. The location and numbers of borehole are shown in Table I and Fig. 2. This data shows that present ground water level is deeper comparing to the condition in decades ago. It is assumed that the increasing number of population and number of industries in this area affect the changing of ground water level. Population trend of the study area from 1991 to 2008 is shown in Fig. 3. The number of population is proportional to number of arthesis well in the study area, therefore the ground water consumption is assumed proportional to the number of population in each area. The overload consumption of ground water is considered to increase the subsidence of this area. The study area is mainly also composed by pumiceous tuff and lake deposit (sediment) formed in Plistocene period, as shown in Fig. 2. This type of lithology is more unstable compared to Breccia rock. Fig. 4 shows distribution of industry in the study area. We determine the number of industry as the total number of middle and large industries. 4. CONTINUOUS SATELLITE IMAGERIES

In this research, we propose a technique to retrieve the volume change of the study area caused the land deformation in long-term and continuously observation period by improving the DInSAR technique. For this purpose, we employ the Japanese L band SAR sensors: JERS-1 SAR (1993-1997) and ALOS PALSAR (2007-2008). L band microwave is well-known as capable to penetrate into the vegetation canopy to the land surface , therefore land surface deformation is more accurate to be observed by using the sensor. By refering the definition of each parameter of Eq. (8) in Section 2, we employ the JERS-1 SAR pairs (pair no. 1 to 5) and ALOS PALSAR pairs (pair no. 6 to 7) as shown in Table II. The pairs show a continuous periods during each mission (JERS-1 SAR: 19930408-19940326-19950128-19960115-19961118-19970809, ALOS PAL-SAR:20070114- 20080117-20081204). Therefore the accumulation of continuous volume change of subsidence could be retrieved here by employing Eq. (8).

The phase difference derived by using DInSAR, slant range distance, elevation, nadir angle, incidence angle and baseline information of each pixel is employed as the input of this equation to derive the volume change of each pixel in a pair of imagery. We also employed the 1:25,000 and 1:50,000 topographic map and old Japanese Military maps (Gaihozu 1884-1945) of our archive to analyze (geometric correction, geological information extraction, environmental change confirmation etc) using these imageries. 5. RESULT AND DISCUSSION

Fig. 5 shows interferogram of long term continuously DInSAR results acquired by employing JERS-1 SAR pairs and ALOS PALSAR pairs. The legend of interferogram with cyan-blue-black and cyan-green-yellow-red show the land surface movement is respectively leaving and approaching the spacecraft. These interferogram shows that the subsidence centered in Cimahi Selatan industrial complex (IC13) affected surrounding wide area. It covers industrial complexes as Cimahi Utara (IC11), Cimahi Tengah (IC12), Andir (IC14), Bandung Kulon (IC16), Cicendo (IC17), and Sukajadi (IC18); and settlement area as Babakan Ciparay (ST11). Husein Sastranegara Military Airport is also affected by the subsidence. The value of maximum subsidence of each district is derived by unwrapping each interferogram and shown in centimeters in Table III. The interesting phenomenon is found at Batujajar (IC15), Cililin (IC21), Soreang (IC22) and Margaasih (IC23) areas or south area of Cimahi Selatan. It is considered that the subsidence phenomenon is cut clearly by strong geological layer composed lithology by tuffaceous breccia, lava, sandstone and conglomerate, even many industries and dense population that consumed much ground water are also settled in this south area (see Figs. 1 (c) and 3). The largest subsidence at Cimahi area in observed periods shows 57 cm subsidence occurred between January 28, 1995 and January 15, 1996. Recently, the subsidence is still occurring in Cimahi Selatan and surrounding area, but the affected area is narrower comparing to decades ago and about 14 cm subsidence happened in 2008.

Six GPS surveys have been conducted by the author (H.Z. Abidin) to investigate the subsidence at the study area on 21-24 February 2000, 21-30 November 2001, 11-14 July 2002, 1-3 June 2003, 24-27 June 2005, and 2 0 - 2 3 A u g u s t 2 0 0 8 w i t h t h e o b s e r v e d points as shown in Fig. 2. These surveys employed dual-frequency geodetic-type GPS receivers with reference point was GPS station inside ITB. The length of each session was 10 to 12 hours with 30 second interval of data collection and 15o of elevation mask. The height differences between two consecutive survey epochs were calculated base on the estimation of ellipsoidal heights obtained from GPS processing. The

TABLE II JERS-1 SAR AND ALOS PALSAR IMAGERIES EMPLOYED IN THIS STUDY

Sensor Name (Master) Center

date/time

(Slave) Center

date/time A/DPath/Row (Frame)

B (km)

Bp (km)

(deg)

dh (km)

dh (km)

dhpr (km)

Pair No. Center coordinate

Center coordinate

JERS-1 SAR (1)

19930408 03:07:30.900

19940326 03:04:00.1 D 106/312 1.700 -1.434 34.784 0.101 -1.434 -0.874

-6.884 /107.480 -6.884 /107.498

JERS-1 SAR (2)

19940326 03:04:00.100

19950128 03:08:48.1 D 106/312 0.925 0.653 35.104 0.170 0.649 0.637

-6.884 /107.498 -6.882 /107.522

JERS-1 SAR (3)

19950128 03:08:48.100

19960115 03:13:42.5 D 106/312 0.290 -0.116 35.219 -0.259 0.084 -0.377

-6.882 /107.522 -6.884 /107.493

JERS-1 SAR (4)

19960115 03:13:42.500

19961118 03:13:54.6 D 106/312 0.023 -0.044 35.225 0.046 -0.045 0.025

-6.884 /107.493 -6.883 /107.506

JERS-1 SAR (5)

19961118 03:13:54.600

19970809 03:13:41.1 D 106/312 0.573 -0.262 35.329 -0.329 -0.271 -0.595

-6.883 /107.506 -6.884 /107.495

ALOS PALSAR (6)

20070114 15:40:58.378

20080117 15:39:52.428 A 436/7040 1.549 1.179 34.385 0.189 1.172 1.030

-7.027/107.649 -7.019/107.658

ALOS PALSAR (7)

20080117 15:39:52.428

20081204 15:40:10.788 A 436/7040 0.848 0.776 34.393 -0.156 0.787 0.351

-7.019/107.658 -7.026/107.666

Fig. 5. Interferograms of the study area

Fig.6. Subsidence rate derived by DInSAR at GPS points

Fig. 7. Photographs collected in ground surveys

result of GPS observation of each station is shown in Table IV. This observation shows the high subsidence rate occurred at CMHI, DYHK, RCK2, GDBG, BNJR and BM18L with total subsidence between February 2000 and August 2008 were 60.5, 60.1, 46.4, 32.0, 27.9, and 34.4 centimeters, respectively.

Fig. 6 shows the subsidence rate of each GPS observation point derived by using DInSAR with observation periods of 8 April 1993 - 9 August 1997 (JERS-1) and 14 January 2007 - 4 December 2008 (ALOS). This observation shows the total subsidence rate of CMHI, DYHK, RCK2, GDBG, BNJR and BM18L also 53.8, 19.9, 4.8, 6.4, 9.7, and 27.3 centimeters, respectively. The subsidence velocity of Cimahi, Dayeuhkolot, and Baleendah areas are 7.7, 7.5, and 4.3 cm/year by using GPS observation, and 9.0, 3.9, and 4.6 cm/year by using DInSAR observation. Even the observation period of GPS and DInSAR observation is different; both shows the subsidence rate at Cimahi, Dayeuhkolot and Baleendah areas are very high. This confirmation result also shows the result derived by DInSAR is close to the result of GPS observation.

In this research, we will assess the proposed technique to retrieve the volume change of the study area by using each pair of SAR data. For example, let’s consider the district of Cimahi Selatan and its surrounding areas. The phase difference derived by using DInSAR, elevation incidence angle and other parameters of each pixel are substituted to Eq. (8) to derive the volume change of each pixel in a pair of imagery. The extraction of target district area is based on the distribution of subsidence as shown in Figs. 5 and 1(b). Therefore subsidence at Cimahi Selatan and its surrounding area within April 8, 1993 to March 26, 1994 affected about 29.5 km2 and 8,112,500 m3 of soil volume. The maximum volume of subsidence happened in period of January 28, 1995 to January 15, 1996 about 8,635,500 m3 and affected 30.3 km2 area. The same manner of assessment of the proposed technique for the other areas, especially Dayeuhkolot (IC41), Baleendah (IC51), Majalaya (IC71) and Rancaekek (IC81) with detail result is shown in Table III. Base on the interferograms shown in Fig. 5, the subsidence phenomenon is continuosly happening with center of industrial area at Cimahi Selatan, Dayeuhkolot, Majalaya and Rancaekek industrial complexes, even the subsidence activities depicts decreasing, see Table III and IV. The subsidence phenomenon at Baleendah, Majalaya and Rancaekek area is observed actively in monitoring period of January 14, 2007 to December 4, 2008 by using ALOS PALSAR data. Some subsidence activities are also observed at Cilampeni industrial complex at Katapang (IC32) and Margahayu (IC31) area in 2007. It is assumed as the effect of human and industrial activities in this area as shown in Figs. 3 and 4. Geologically, Dayeuhkolot, Baleendah, Katapang, Majalaya and Rancaekek areas have pumiceous tuff and

lake deposits, therefore the subsidence distribution in clearly order ring of subsidence phase distribution. It is assumed that this geological structure is very weak for industrial activities in these areas. On the other hand, the geological formation at Batujajar, Margaasih and small part in southwest Cimahi Selatan areas that is composed by tuffaceous breccia, lava, sandstone and conglomerate called breccia block (see Fig. 2) has a very strong structure for vertically land deformation, especially subsidence. Hence the subsidence phase distribution is blocked in this area.

Based on the result derived by DInSAR and proposed technique, we held three ground survey campaigns on 13-14 December 2008, 25-28 February 2009, and 4-10 August 2009 to collect the building and infrastructure condition with about 800 buildings were surveyed and recorded. The sampled photograph of the study area is shown in Fig. 7. Fig. 7(a) shows the condition of wall of settlement around the center of subsidence at Cimahi Selatan. The wall of some buildings around this area is cracking in horizontal (blue arrow) and vertical (red arrow) direction. The land deformation at horizontal and vertical direction is considered causing this condition, especially this phenomenon is conspicuously occurring at the buildings at center of subsidence at Cimahi Selatan Fig. 7 (a) and Dayeuhkolot Fig. 7 (d) and (e). Some houses have been found in subsided condition (about 1 meter), one of example is shown in Fig. 7 (b) that is close to center of the subsidence at Cimahi Selatan district. During the ground survey, we identified the soil type of this area is unstable fumiceous tuff and swamp, therefore the settlement around this area is very risky for the subsidence damaging and other land deformation. Some public infrastructure as irrigation network, road etc at Cimahi Selatan, Dayeuhkolot and Balaendah were also found in wrecking condition as shown in Fig. 7 (c). The railway infrastructure is also found in horizontally folded as shown in Fig. 7 (f). This condition makes the urban problems in this area worsened due to the flood in rainy season.

6. CONCLUSIONS

This research proposed a technique to derive volume change on the Earth surface caused by long-term continuously land deformation. Basically, this technique is the improvement of Differential Interferometric Synthetic Aperture Radar (DInSAR) to estimate the volume change for long-term continuously land deformation (uplift or subsidence) observation from the space. To confirm the capability of this technique, we employ it to investigate subsidence of Bandung city, Indonesia by using two Japanese L band Spaceborne SARs (JERS-1 SAR and ALOS PALSAR) in periods subsidence occurs concentrating at industrial complexes 1993-1997 and 2007-2008. This result shows the and

settlements of Bandung, especially at Cimahi, Dayeuhkolot, and Baleendah districs seriously. The analysis result of ALOS PALSAR by using proposed technique shows new subsidence areas at industrial and settlement areas of Majasari, Majalaya, Cilampeni and Rancaekek districs. The result is confirmed by using statistical data of 1991-2008 published by Statistics Bureau of Indonesia (BPS), three campaigns of ground surveys to collect the recent condition of buildings and other infrastructures, ground water level data and analysis of geological formation of this area. This research shows potential capability of the proposed technique to investigate the volume change caused by land deformation, especially subsidence. The population, number of industries, ground water level has a strong agreement or relationship with the distribution of subs idence in the s tudy area , Bandung c i ty .

ACKNOWLEDGMENT The Authors thank to JAXA SIGMA-SAR software to generate our interferograms (Masanobu Shimada, 1999); Pandhito Panji Foundation for ground survey guide and statistics data collecting; The present work was supported by the European Space Agency (ESA) Earth Observation Category 1 (No. 6613) in the framework of the project “Long Term Continuously DInSAR for Land Deformation Monitoring”; the Japan Society for the Promotion of Science (JSPS) for Grant-in-Aid for Scientific Research - Young Scientist (A) (No. 19686025) and Grant-in-Aid for Scientific Research (No. 19-07023); National Institute of Information and Communication Technology (NICT) for International Research Collaboration Research Grant 2008 and this work was supported in part by the research grant for Mission Research on Sustainable Humanosphere from Research Institute for Sustainable Humanosphere (RISH), Kyoto University, Japan.

TABLE III ESTIMATION OF SUBSIDENCE (MAXIMUM VALUE IN CENTIMETERS) AND VOLUME CHANGE (IN 1000 CUBIC METERS) OF THE STUDY AREA

District Codes

Sensor Name and Data Combination (Master-Slave)

JERS-1 19930408- 19940326

JERS-1 19940326- 19950128

JERS-1 19950128- 19960115

JERS-1 19960115- 19961118

JERS-1 19961118- 19970809

ALOS 20070114- 20080117

ALOS 20080117- 20081204

Industrial Complexes IC1 55 (8,113) 45 (4,433) 57 (8,635,500) 47 (4,935) 36 (3,852) 21 (1,701) 14 ( 662) IC2 12 ( 918) 0 ( 0) 0 ( 0) 0 ( 0) 0 ( 0) 12 ( 222) 3 ( 152) IC3 0 ( 0) 0 ( 0) 0 ( 0) 0 ( 0) 0 ( 0) 9 ( 81) 0 ( 732) IC4 9 ( 513) 31 (1,535) 25 (1,100,000) 16 ( 840) 20 (1,280) 6 ( 75) 8 ( 362) IC5 12 ( 312) 6 ( 114) 6 ( 255,000) 6 ( 102) 12 ( 336) 27 ( 351) 9 ( 255) IC6 6 ( 225) 0 ( 0) 6 ( 330,000) 6 ( 708) 3 ( 134) 12 ( 150) 6 ( 96) IC7 0 ( 0) 0 ( 0) 0 ( 0) 0 ( 0) 0 ( 0) 9 ( 85) 12 ( 168) IC8 12 ( 252) 0 ( 0) 0 ( 0) 0 ( 0) 9 ( 221) 9 ( 198) 12 ( 174)

Settlement AreasST1 12 ( 174) 6 ( 66) 6 ( 132,000) 3 ( 29) 3 ( 90) 3 ( 105) 0 ( 0) ST2 9 ( 716) 6 ( 204) 9 ( 67,500) 4 ( 26) 4 ( 94) 4 ( 78) 0 ( 0) ST3 9 (1,125) 9 ( 783) 9 ( 495,000) 4 ( 344) 4 ( 278) 4 ( 420) 0 ( 0)

Remarks : The value in the bracket shows the volume change in cubic meters. TABLE IV. GPS OBSERVATION OF THE STUDY AREA – SUBSIDENCE RATE

Location Codes

Location Name Coordinate Period of Observation

Latitude Longitude 200002-200111

200111-200207

200207-200306

200306-200506

200506-200808*)

BNJR Banjaran 7.040 107.591 8.7 3.8 3.3 6.9 5.2 (4.0) BM9L Bojong Malaka 9 6.985 107.598 - - 19.1 - - (3.8) BM13L Bojong Malaka 13 6.978 107.614 - - 5.7 10.3 - (3.1) BM18L Bojong Malaka 18 6.991 107.626 - - 15.9 10.0 8.5 (7.2) BM19L Bojong Malaka 19 6.992 107.630 - - 5.6 - - (7.5) BJNS Bojongsoang 6.993 107.652 9.2 4.2 3 - - (3.3) BRGA Braga 6.919 107.610 - 11.9 - - - (1.7) CMHI Cimahi 6.909 107.557 22.8 17.6 15.7 4.4 - (8.7) CPRY Ciparay 7.024 107.669 - 3.0 - - 1.7 (1.4) DYHK Dayeuhkolot 6.965 107.623 18.8 18.2 4.1 8.1 10.9 (2.5) GDBG Gedebage 6.964 107.688 - 1.1 16.8 4.0 10.1 (4.3) KPO1 Kopo 1 6.951 107.587 - - 7.0 1.3 7.1 (2.4) KPO2 Kopo 2 6.985 107.563 - 0.1 6.1 5.1 10.5 (3.5) MJL1 Majalaya 1 7.051 107.742 8.1 2.1 8.2 - 2.2 (1.1) MJL2 Majalaya 2 7.011 107.752 - - 3.4 3.7 6.5 (1.2) RCK1 Rancaekek 1 6.950 107.752 12.1 5.4 - 3.2 0.5 (0.4) RCK2 Rancaekek 2 6.960 107.804 18 14.8 0.8 6.9 5.9 (2.1) UJBR Ujungberung 6.914 107.690 3.1 1.7 6.1 - 0.7 (0.4)

*) Value inside the bracket shows the subsidence rate derived by DInSAR within 14 January 2007 and 4 December 2008