The use of Libraries of Satellite Images for Estimating Earthquake- and Tsunami Risk and Losses  


Knowledge about the built environment is essential for estimating losses in real-time, or for possible future earthquake and tsunami catastrophes. In many developing countries, the number of buildings, their size and their quality is poorly known. The task of counting the buildings, geo-referencing their position and estimating their structural quality in cities with millions of inhabitants is too large, even for an army of engineers on the ground. Therefore, we seek to improve our knowledge about building stock, using satellite images. These images also allow the mapping of lifelines and critical facilities, to some extent.

We have colleted about 1,000 high-resolution satellite images of cities in earthquake prone areas and along coasts exposed to tsunamis. As a first step to make them useful, we plan to create a library of these images, from which they can be called up immediately after earthquakes, tsunamis and floods, in order to assist the rescue teams, who often are on the scene without maps of the cities devastated by natural disasters. Figure 1 shows an example of a satellite image that would be of great value to rescuers to find their way around.

Figure 1: High resolution satellite image of a city that shows the air port (lower right), highways and roads, the coast, rivers with bridges and the nature of neighborhoods. This would be valuable information for rescuers, if this city were damaged by an earthquake or tsunami.

Counting buildings and defining the character of the built environment in relatively homogeneous neighborhoods is another task one can achieve with an image as shown in Figure 2. Instead of estimating the average loss for an entire city, relatively differentiated loss estimates can be obtained by subdividing a city in districts or neighborhoods, each containing the number and type of buildings seen in the image of Figure 2.

Figure 2: Satellite image showing neighborhoods with different types of buildings: one story single family homes, several story apartments and industrial buildings.



  The use of Satellite Images Obtained After Earthquakes and Tsunamis for Estimating Losses  



After major earthquakes and tsunamis, the extent of the disaster often needs to be estimated from outside the affected area, because no information flows from the center of the disaster. For this reason, WAPMERR issues loss estimates in real-time, based on data sets on building stock and population in its GIS data base and the calculated strong shaking due to an earthquake. If a satellite photograph for the devastated area becomes available by chance soon after an earthquake or tsunami, the damage can be assessed directly. If, in addition, a satellite photograph of the city in question exists in the image library, a comparison between the two images, before and after, can furnish a quantitative measure of the extent of destruction. An example of this is seen in Figure 1 for the case of the city of Bam, Iran, that was devastated in an earthquake.

Figure 1: High resolution satellite image of parts of the city of Bam, Iran, before (right) and after (left) of the M6.6 earthquake of 26 December 2003 that killed more than 20,000 people. The power relay station at the lower left is not damaged significantly, whereas single family homes next to it have collapsed. A row of houses near the center is much less damaged than those surrounding them.

We have proposed that another, less obvious use of satellite images can help estimating losses in real-time, if conditions are favorable. Some satellites take Synthetic Aperture Radar (SAR) images as they circle the planet. These can image the surface also through cloud cover and at night, when photographic images are not available. However, SAR images do not show the damage directly and there are other potential problems of the analysis. Nevertheless, we show in the following article, published in EOS, that highly accurate locations of the earthquakes energy release can be obtained by interferometry of pictures before and after the earthquake (InSAR). This high quality information can reduce the errors in real time loss estimates by a factor of 10, and may therefore be of great use to rescue teams and authorities attempting to bring relief to the population in the devastated area. For details click here.