Preliminary Field Report Post-Tsunami Urban Damage Survey in Thailand by eld18221


									Preliminary Field Report: Post-Tsunami Urban Damage Survey in Thailand, Using the
VIEWS Reconnaissance System

       Shubharoop Ghosh, Charles K. Huyck, Beverley J. Adams and Ronald T. Eguchi
                           ImageCat, Inc. 400 Oceangate, Suite 1050, Long Beach, 90802
                                              Fumio Yamazaki
                        Chiba University, I-33 Yayoi-cho, Inage-ku, Chiba 263-8522, Japan
                                           and Masashi Matsuoka
      Earthquake Disaster Mitigation Research Center (EDM), National Research Institute for Earth Science and
            Disaster Prevention (NIED, 1-5-2 Kaigandori, Wakinohama, Chuo-ku, Kobe 651-0073, Japan
1.0      Introduction
On 26 Dec 2004, 00:58:53 UTC, a magnitude 9.0 earthquake occurred off the west coast of
northern Sumatra, Indonesia. The epicenter was located beneath the Indian Ocean at 3.307°N
95.947°E. The earthquake triggered a series of tsunamis in the region that devastated
communities along thousands of miles of exposed coastline. Ranked as one of the most
catastrophic events in recent times, the earthquake and subsequent tsunami caused significant
damage in 11 countries and resulted in over 200,000 deaths. Countless others were rendered
homeless, or left without basic lifelines and amenities.

The Multidisciplinary Center for Earthquake Engineering Research (MCEER) joined a multi-
lateral Thailand-Japanese reconnaissance team to investigate the effects of the Tsunami disaster
in Thailand. Of the affected nations, Thailand was selected as the destination based on a
preliminary assessment of potential casualties in the region by ImageCat using remote sensing
data1 and media reports of >5,300 deaths, destruction of multiple tourist destinations, and
>US$500 million damage to the nations shrimp industry. The reconnaissance team was led by
Professor Fumio Yamazaki of Chiba University in Chiba, Japan and Dr. Pennung Warnitchai of
the Asian Institute of Technology in Bangkok, Thailand. Shubharoop Ghosh of ImageCat, Inc.
represented MCEER. Other researchers in the team included Dr. Masahiko Honzawa of Japan
International Cooperation Agency (JICA), and Dr. Masashi Matsuoka of Earthquake Disaster
Mitigation Research Center (EDM), Kobe, Japan.
The objective of this field deployment was to collect perishable information about building and
lifeline damage characteristics. The field-based damage assessment was conducted using the
VIEWS (Visualizing the Impacts of Earthquakes With Satellites). These ground-based
observations can be later used to validate damage characteristics identified on satellite imagery.
It is envisioned that such perishable data will be invaluable for future research in evaluating
damage from Tsunami hazards.
VIEWS is a notebook-based system, which integrates GPS-registered digital video footage,
digital photographs and observations with high-resolution satellite imagery collected before and
after a disaster. VIEWS was previously used in reconnaissance activities following the 2003
Bam, Iran earthquake, Hurricane Charley and Hurricane Ivan that hit the Gulf coast in 2004, and
the Niigata, Japan earthquake in October 2004.

 Areas of potential devastation was identified by cross referencing population data from NOAA’s DMSP sensor
with wave height modeled by Vasily Titov at NOAA and proximity to the coastline.

The following preliminary field report begins with a brief overview of the field study sites. It
goes on to document damage survey activities that were conducted, together with methodologies
employed for data collection. It includes a discussion on the satellite imagery and other data that
were available from various sources. The resulting data sets from the survey are presented. The
report summarizes preliminary findings through a selection of illustrative examples, which were
extracted using the MCEER funded D-VRS system.

2.0    Survey Sites

The post-tsunami damage assessment was conducted in three south-western provinces of
Thailand (Figure 1): Phang Nga; Krabi; and Phuket. Survey site selection was made in
consultation with Dr. Pennung Warnitchai of the Asian Institute of Technology, and was based
on media reports of areas experiencing severe damage (particularly where the key industries of
tourism and fisheries were hard hit), and areas for which “before” and “after” Quickbird satellite
imagery was available.


                                                PHANG NGA PROVINCE AREAS SURVEYED
                                                Dates: 8 Jan, 10 Jan 2005
                                                Available Imagery*: Ikonos, Landsat

                                                Ban Pracho in Ko Prathong
                                                Ban Nam Khem Area
                                                Le Meridian Khaolak
                                                Bamboo Orchid Resort
                                                Similana Resort

       Date: 11 Jan 2005
       Available Imagery*: Quickbird, Landsat

       Bang Tao Bay                                    KRABI PROVINCE AREAS SURVEYED
       Kamala Beach                                    Date: 9 Jan 2005
       Patong Beach                                    Available Imagery*: Landsat
       Karon Beach
       Kata Beach                                      Phi Phi Island

                                                                   * See Section 3.2 for imagery sources

Figure 1: Field Study regions of Phang Nga, Phi Phi Island, and Phuket, which sustained
heavy tsunami damage after the Sumatra Earthquake of 26 December, 2004

3.0    Damage Survey using VIEWS Reconnaissance System

3.1    Data Collection

The field survey was conducted using the VIEWS system from a moving vehicle and on foot
(see Figure 2), depending on vehicular access to impacted areas. Access to selected study areas
in general did not prove to be a significant limitation; admittance to the heavily damaged sites
did not require special authorization. However, the areas surveyed within the three provinces
were spread wide apart, and traveling to and from the study sites proved to be time consuming.

To conduct survey from a moving vehicle, a van was selected by the team, since it provided
increased elevation above street level, and better coverage by avoiding obstructions in the
foreground. The vehicle was driven at around 10-15 mph, as this speed allowed the video
coverage and stills to be obtained with a lesser degree of aberration, while enabling a large
geographic extent to be covered. More in-depth damage assessments were conducted on foot.

For this deployment, a new data collection approach was adopted (Figure 2) by deploying three
video cameras that simultaneously captured footage for three directions (front, left, and right) in
some heavily affected areas. This streamlined the video collection process and provided a wider
view of the area. An alternative handheld-based approach was also tested for collecting GPS
points. A Garmin iQue® Handheld with built-in GPS receiver was used to collect GPS points in
the survey sites where VIEWS was deployed. 11 hours of digital video footage was recorded
during the reconnaissance route that covered >50 kms. A library of about 1500 digital
photographs was also collected by the team.

(a) VIEWS deployment with one camera            (b) VEIWS deployment with three cameras

Figure 2 Field deployment of the VIEWS system following Tsunamis in Thailand
(a) Using a single video camera (b) Using multiple video cameras

3.2      Satellite Data and Other Base Data Available for the Survey

The screen shot of the VIEWS User interface in Figure 3 demonstrates how the satellite images
serve as the mapping base layer. Through the real-time GPS feed, routes taken around the
damaged areas were logged and overlaid on a vector-based street map. Georeferenced building
damage and impacted area observations were recorded using the GPS-linked digital video
recorder. A georeferenced photographic record was also collected, illustrating in detail, damage
characteristics from the Tsunami.

Figure 3 User interface for the VIEWS system, deployed to collect building damage data in Patong
Beach, Phuket, Thailand

Satellite imagery was made available to the field investigation team from multiple sources.
ImageCat obtained high resolution “before” and “after” Quickbird imagery from Digitalglobe for
Patong Beach area of Phuket, Thailand. Geo-Informatics and Space Technology Development
Agency (GISTDA), Thailand provided the following datasets in ECW format for visualizing
impacts and general navigation purposes:

-     7m Landsat data for southern Thailand
-     1m Ikonos data for Khaolak Area in Phang Nga, Patong Beach area of Phuket
-     Quickbird data for Phuket International Airport area
-     IRS satellite data for southern Thailand

GISTDA also provided street and province boundary data in shape file format for field
investigation team.

4      Preliminary Findings

4.1    Phang Nga Province

Areas in Phang Nga province surveyed included tourist resort areas in Khao Lak, and
fisherman’s villages in the Ban Nam Khem area and Ban Pracho in Ko Prathong. With its long
and narrow stretch of coastline, Khao Lak was the one of the hardest hit tourist destinations in
Thailand. Resort locations surveyed included: the Similana Resort; the Bamboo Orchid Resort;
and Le Meridian Khaolak. Tablamu Pier in Khao Lak was also surveyed as a part of the
reconnaissance mission. Survey efforts also focused on collecting damage information within
fishing industry centers of Ban Nam Khem and Ban Pacho.

Figure 4 (a) through (c) depicts the routes through Khao Lak area in Phang Nga province along
which GPS readings, georeferenced video coverage and photographic record were collected.
Reportedly, these are areas recorded tsunami wave heights of up to11 meters. Most of the resorts
surveyed sustained heavy to very heavy damage.

                                 1                            2

                                 3                             4


                                 5                             6


      Approximate location
      of still photographs

Figure 4 VIEWS reconnaissance routes in Khao Lak (a) Pakarang Cape routes and the
devastated Bamboo Orchid Resort, (b) Mid-Khao Lak routes and photos of destroyed buildings, (c)
Lower Khao Lak routes and Similana Resort photographs

      4.2       Phi Phi Island in Krabi Province

      Figure 5 depicts the locations on the Phi Hi Island where GPS readings, georeferenced video
      coverage and photographic record were collected including photographs of damage sustained.
      Phi Phi Island is located about 48 kms southeast of Phuket in the Anadaman Sea. The island of
      Phi Phi suffered extensive damage, with most resorts in Ton Sai bay closed in the aftermath of
      the huge waves that swept across the island. The reconnaissance team surveyed two locations on
      the island including the Ton Sai Bay area where most damage was reported.

                         Laem Tong

                                                         Phi Phi


                              b c
                              Ton Sai Bay

(a)                                     Approximate location of still

      Figure 5 Locations on Phi Phi Island where VIEWS was deployed. (a) Phi Island Survey Routes,
      (b) Destroyed shopping street, (c) Showing water level inside a completely destroyed shop, (d)
      Destroyed resort in Ton Sai Bay

      4.3         Phuket Province

      Areas in Phuket province surveyed included tourist resort areas in Bang Tao Bay, Kamala Beach,
      Patong Beach, Karon Beach, and Kata Beach. Phuket is primarily a tourist destination, with
      beaches lined with hotels and shopping. Damage observed in Bang Tao Bay and Kamala Beach

was heavy, following reported wave heights of 5-6m. Shops along the beach front in Patong
beach were observed to have sustained moderate to heavy damage with majority of those closed
for repair or debris cleanup. Damage to buildings along the beaches in Karong and Kata beaches
appeared was less than for the beaches further north e.g. Patong beach, Kamala beach. Figure
6(a) below depicts the routes along which GPS readings, georeferenced video coverage and
photographic record were collected. Figures 6(b) through 6(d) are a selection of digital
photographs showing damage in areas surveyed in Phuket.

         Bang Tao Bay

                                              Phuket                                (b)

      Kamala Beach



         Patong Beach


         Karon Beach

  Approximate location of still photographs

Figure 6 VIEWS reconnaissance routes in Phuket. (a) Phuket Survey Routes, (b) Kamala Beach
damage, (c) Alley near Patong beach with debris brought by the waves, (d) Cordoned off shopping
areas along Patong beach to help cleanup operations

5.0      Preliminary Damage Visualization using D-VRS

In order to integrate, share, visualize, and ultimately analyze post-disaster reconnaissance field
data collected using VIEWS, MCEER funded the development of tandem internet- and desktop-
based ‘virtual reconnaissance systems’, referred to as VRS and D-VRS. Figure 7 shows a screen
grab from D-VRS, which provides researchers with easy access to the satellite imagery, GPS
readings and georeferenced video and photographic records for the survey sites. The User has an

option to toggle between multitemporal and multi-source satellite images, and to explore these
images in detail using zoom and pan functions. These images are overlaid with GPS routes
collected during the survey. By selecting a GPS point, the User can view corresponding video
footage and scroll through the photographic archive in the adjacent windows. The photographic
library can be augmented with stills captured directly from the video as it plays, each of which is
output to a new georeferenced file.

Figure 7 Screen shot from the D-VRS virtual reconnaissance system, showing satellite
imagery, GPS readings, video footage and digital photographs collected in Khao Lak,

For this study, a multi-look angle version of visualization interface was developed by ImageCat.
From Figure 8, three windows (front, left, and right) play video footage from the three cameras
that were simultaneously deployed within heavily affected areas. This video collection and
visualization process provides an enhanced field of view.

Figure 8 Screen shot from the D-VRS virtual reconnaissance system, showing satellite
imagery, GPS readings, video footage from three video cameras collected in Khao Lak,


Adams, Beverley J., Huyck, Charles K., and Mansouri, Babak (2004). Streamlining Post-
Earthquake Data Collection and Damage Assessment in Bam, Using VIEWS (Visualizing the
Impacts of Earthquakes with Satellite Images).

Adams, Beverley J., Womble, J. Arn, Mio, Michael Z., Turner, John B., Mehta, Kishor C. and
Ghosh, Shubharoop (2004). Collection of Satellite–referenced Building Damage Information in
the Aftermath of Hurricane Charley.

Huyck, C., Scawthorn, C., Bardet, Jean-Pierre, Kayen, R., Kawamata, Y., Olshanky, R.,
Somerville, P., Mori, J., Rathje, E., Bay, J., Jibson, R., Kelson, K., Pack R., Nishi, N. (2005).
Preliminary Observations on the Niigata Ken Chuetsu, Japan. Earthquake of October 23, 2004.
EERI Newsletter, January 2005 Volume 39, Number 1.


To top