Surface Effects of the 2004 Indonesian Earthquake and Tsunami from SAR data
C. Bignami(1), M. Chini(2), N. Pierdicca(1) , S. Stramondo(3)
(1)Univ. La Sapienza of Rome, via Eudossiana 18, 00184 Rome, Italy
(2) Univ. Alma Mater Studiorum of Bologna, v.le C. Berti Pichat 8, 40127 Bologna, Italy
(3) Isituto. Nazionale di Geofisica e Vulcanologia, via di Vigna Murata 605, 00143 Rome, Italy
On December 26, 2004, at 00:58 GMT a Mw 9.0 earthquake took place in the Indian Ocean, offshore the West
coast of Sumatra, at a depth of about 30 km. This earthquake is one of the largest events of the last 100 years,
comparable only to the Chile 1960 and Alaska 1964 ones. The earthquake originates in the subduction zone of the
Indian and Burma plates, moving at a relative velocity of 6 cm/year. The aftershocks were distributed along a plate
boundary about 1000 Km long, between Sumatra and the Andaman islands. GPS vectors point out a strong surface
displacements towards SW of Sumatra Island ranging one meter and more. Ground surveys in Sumatra and
Andaman Islands showed wide uplift areas and large modification of the coastline. A destructive tsunami followed
and hit, up to some hours after the earthquake, the coastlines of the surrounding regions, causing widespread
destruction in Indonesia, India, Thailand, and Sri Lanka. We used ERS SAR and Envisat ASAR satellite data to
estimate the co-seismic and tsunami effects occurred on the human settlements and on the coast. A large dataset
was made available by ESA. SAR images were co-registered each other and changes were analysed both visually
and automatically. Image difference and correlation coefficient between data before and after the destructive event
were exploited. Changes along the coastline hit by the tsunami have been detected and are reported in this paper.
1 INTRODUCTION AND METHODOLGY
The remote sensing techniques can give a powerful tool to detect surface changes caused by catastrophic events
(, ). Thanks to the capability to acquire images of large areas, the satellite remote sensing can furnish a
synoptic view of the temporal evolution of the event. In particular, the Synthetic Aperture Radar (SAR) collects
maps of the backscattering coefficient  almost independently on the meteorological conditions and sun
illumination. Therefore SAR applications can be considered a useful instrument for crisis management. In addition,
the interferometric SAR techniques (InSAR) can exploit the phase information of the echoed signal to measure
surface movements .
In this work, the surface effects of the earthquake/tsunami have been analysed by the use of SAR images,
comparing the acquisitions before and after the seismic event occurred on 26th of December 2004. The analysis has
been focused on the Andaman Islands archipelago and the North of Sumatra. Results from preliminary ground
surveys in Sumatra and Andaman Islands showed wide uplift areas and large modification of the coastline . Such
surface effects have been also measured in available GPS benchmarks reoccupied in January-February 2005 .
GPS vectors point out strong surface displacements towards SW of Sumatra Island ranging one meter and more, up
to 1-2 m vertical and 5 m horizontal in Nicobar Island and 1-2 m vertical and 3 m horizontal the Andaman Island
The dataset used for our study, provided by the European Space Agency, is composed of ERS2-PRI and Envisat-
ASAR-IMP products. The lack of Single Look Complex data prevented us to perform any interferometric
processing. Tab. 1 sums up the mentioned dataset.
Tab.1 Summary of the data used to analyse the surface effects of the Indonesian tsunami.
North of Sumatra
North of Sumatra
2 SURFACE DISPLACEMENTS BY SAR
In order to visualize the surface effects of the earthquake, RGB colour composite images have been generated.
After a precise co-registration of the ASAR data, by using the BEST (Basic Envisat SAR Toolbox) ESA software,
the pre-event acquisition (03/06/2004) has been assigned to the red component, while the post-seismic one
(30/12/2004) to the blue and green. The changes have been well identified by the cyan pixels on the resulting
image. The analysis of the pre and post seismic SAR data has pointed out the up-lift phenomenon of the west
shorelines of the Andaman archipelago (red contours superimposed on the image in Fig.1). It is easy to identify the
risen out lands, prevalently beaches and reefs, as also derived by field observations .
Fig.1 RGB colour composition of North Andaman Island: R= pre-seismic 03/06/2004, G=B=post-seismic
30/12/2004; in the yellow box a detail of the risen up lands: A) pre-seismic image, B) post-seismic image.
To assess the permanent surface effects of the earthquake we have analysed a longer series of data and used the
intensity correlation coefficient. The correlation coefficient between images has been estimated in a moving
window of 300m x 300m, so as to generate a new map, named the intensity correlation map. As an example, we
show in Fig.2 a colour composite of the intensity correlation images of Little Andaman Island. The red channel has
been assigned to the post-seismic correlation (70 days time span), the green channel to the co-seismic correlation
with large temporal baseline (280 days) and the blue channel assigned to the shorter temporal baseline (210 days)
co-seismic correlation. The red pixels along the shoreline in Fig.2 highlight the emerged land as a consequence of
the earthquake, where the high post-seismic correlation points out the presence of land, whilst the low co-seismic
correlation reveals a considerable change. This is in agreement with , where the new shorelines and reefs risen
out (about 90 cm) of sea after the earthquake has been shown. The up-lift phenomenon detected by SAR is also in
agreement with the same surface displacement modelled by the GPS measurements as shown in Fig.3a.
3 DAMAGE ASSESSMENTS: THE FLOODING
SAR intensity images have been used to analyse the coastal areas and detect those affected by the flooding. The
analysis has been performed over the Northern portion of Sumatra, the area closest to the epicenter and the most hit,
and on the Andaman islands, in particular Little Andaman. Pre and post tsunami data have been combined in RGB
false colour composite image as done in Fig.1 and the result in and around the city of Banda Aceh is shown in
Fig.4. The inundation of bare soil induces backscattering decrease in the SAR post-tsunami acquisitions .
Moreover, the same kind of behaviour can be due to damaged man made structures. Therefore, the result of such
combination of intensity images allows us to clearly point out the hit areas (red pixels in Fig.4a). The extension of
flooded area detected by the SAR images well agrees with the estimates based on the SPOT2 optical images .
Fig.2 Colour composite of the intensity correlation maps of Little Andaman Island. The yellow boxes represent a
detail of the island: A) intensity correlations combination, B) pre-seismic image, C) and D) post-seismic images
acquired on 30/12/2004 and 10/03/2005 respectively.
Fig.3 Preliminary model derived by GPS network measurements
(http://cires.colorado.edu/~bilham/IndonesiAndaman2004.htm ); (a) the model highlights an uplift up to 1-2 m in the
West coast and a subsidence to the E. (b) Horizontal displacement from GPS data.
Fig.4 a): Colour composition of Banda Aceh: R=pre-seismic, G=B=post-seismic acquired on 2004-12-30. b) The
SPOT2 optical image.
Regarding Little Andaman, the pre and post tsunami ASAR images allowed us to identify the inundation of forestal
areas (green features in the pictures in Fig.5) in the North and Western portions of the island. In this case the water
beneath vegetation cause an increase of the backscattering in the 2004-12-30 acquisition [R= intensity image (2004-
06-03) G=intensity image (2004-12-30), B=intensity image (2005-03-10)] caused by an increasing of the double
bounce backscattering mechanism . Differently from the other images used in this study, the 2004-12-30 image
has been acquired in HH polarization (the other are VV). However, in forestry regions and at the used frequencies
different polarizations (HH and VV) do not produce significative effects on intensity images .
Fig.5 Details of the inundation of forestal areas in Little Andaman Island. RGB color composition: R= intensity
image 2004-06-03) G=intensity image (2004-12-30), B=intensity image (2005-03-10). Note the green pixels
representing the inundated areas beneath the vegetation.
In this work we showed some results regarding the use of SAR intensity images to detect the onshore effects of the
Indonesian tsunami and earthquake; moreover, we investigated the SAR capabilities for monitoring large areas
during catastrophical events. The up-lift phenomena and surface changes caused by the earthquake have been well
detected in North Sumatra and in the Andaman archipelago, where the reef rising strongly modified the shoreline,
in particular in the Western coast. The SAR results and the preliminary surface displacement models retrieved by
GPS are in good agreement. We also analysed flooding under forestal areas in Little Andaman Island based on
differences of the backscattered signal resulting from surface properties changes.
1. Matsuoka M. and Yamazaki F., Image processing of building damage detection due to disasters using SAR
intensity images, Proc. of 31st Conference of the Remote Sensing Society of Japan, pp. 269-270, 2001.
2. Bignami C., Chini M., Pierdicca N., Stramondo S., Comparing and combining the capability of detecting
earthquake damage in urban areas using SAR and optical data, Proc. IEEE IGARSS, Vol.1 pp. 55-58, 2004.
3. Dobson M.C., Ulaby F.T. and Pierce L.E., Land-cover classification and estimation of terrain attributes using
synthetic aperture radar, Remote Sensing of Environment, vol.51, pp 199-214, 1995.
4. Massonet D., Rossi M., Carmona C., Adragna F., Peltzer G., Feigl K. and Rte T. , The displacement field of the
Landers earthquake mapped by radar interferometry, Nature, 364, pp. 138-142,1993.
5. Sieh K., Aceh-Andaman Earthquake: what happened and what’s next?, Nature, vol. 434, 2005.
6. Seismology Research Group, Centre for Earth Science Studies, Trivandrum, India
7. Ammon C. J., Ji C., Thio H., Robinson D., Ni S., Hjorleifsdottir V., Kanamori H., Lay T., Das S., Helmberger
D., Ichinose G., Polet J., Wald D., Rupture Process of the 2004 Sumatra-Andaman Earthquake, Science, vol. 308,
pp. 1133-1139, 2005.
8. Wheeler M., The Great Sumatra Quake, Engineering & Science, n°1/2, 2005.
9. Weydahl D.J., Flood monitoring in Norway using ERS-1 SAR images, Geoscience and Remote Sensing
Symposium,IGARSS '96, Vol., 27-31 May, pp. 151–153, 1996.
11. Townsend P.A. and Foster J. R., Assessing flooding and vegetation structure in forested wetlands using
Radarsat SAR imagery, IEEE Geoscience and Remote Sensing Symposium, IGARSS '02, Vol. 2 , 24-28 June, 1171
– 1173, 2002.
12. Ferrazzoli P., Paloscia S., Pampaloni P., Schiavon G., Sigismondi S., Solimini D., The potential of
multifrequency polarimetric SAR in assessing agricultural and arboneous biomass, IEEE Transactions on
Geoscience and Remote Sensing, Vol. 35, No. 1, January 1997.