Earthquake Engineering Research - HKU Libraries - The University ...

150
In Japan, on reflection of the Hanshin-Awaji Earthquake Disaster 1995, national and local
governments have developed damage estimation systems using GIS, which are based on social
inventory database and observed seismic information. These systems provide real-time disaster
information and enable efficient disaster responses. But only a few countries in the world have such
damage estimation systems as developed in Japan. International rescue and research team should be
deployed properly to the impacted areas in order to make their assistance effective. In addition,
humanitarian aids by various organizations should be distributed properly and fairly to those who
need such aids. Thus, it is desirable for international disaster response and relief committees to
identify the location of impacted areas as soon as possible as the task for their mobilization.
For the purpose of supporting relief and recovery activities by the governments or other organizations,
the damaged areas were estimated using nighttime images observed by the Defense Meteorological
Satellite Program's Operational Linescan System (DMSP/OLS). As examples two large earthquakes,
the Turkey Kocaeli Earthquake (Mw7.4) on August 17 1999 and the India Gujarat Earthquake
(Mw7.7) on January 26 2001 were picked up.
2. DAMAED AREA ESTIMATION BASED ON 2 NIGHT LIGHT IMAGES BEFORE AND
AFTER EARTHQUAKE (BASIC METHOD)
It can be expected that city lights will observably decrease after a large earthquake due to various
reasons such as electricity failure, building collapses, evacuation to shelters or the suspension of
commercial activities. Therefore, the significant reduction in nighttime lights can be an indication of
possible impacted areas due to earthquake disasters. The satellite images observed by the DMSP/OLS
are suitable for the early identification of the damaged areas for following reasons:
a) Due to the sensitive scanner, nighttime images are available.
b) The nighttime images are observed twice a day by two DMSP satellites.
These mean that we could detect significant reduction in nighttime lights at any day on a daily basis.
The DMSP/OLS imagery has spatial resolution of 2.7 km, and the resolution is not as high as that of
the Landsat/TM or the SPOT. But the recurrent periods of the satellites with high-resolution sensors
are more than two weeks and the chances to observe the image immediately after a disaster are very
low. Consequently, the short recurrent period is necessary for emergency use of the observed imagery.
The steps of damaged area estimation are shown in Figure 1. Each pixel in the visible-near infrared
(VNIR) images has digital number (DN) ranging from 0 to 63. We calculated the differences of DNs
on a pixel basis between before and after the earthquake in the sampled area. The cloud influences
were checked using the thermal infrared (TIR) images and the stable light images. Base on the
histogram of the differences, the areas that show the reduction in nighttime lights withp > 0.995 were
determined as significant reduction due to the earthquake disaster (Figures 2). The result maps were
disseminated to the world through the Web page of EDM. The series of the estimation method is
automatically executed in the server machine of EDM.