Using gis and remote sensing for the delineation of risk disaster areas in phuket, Thailand

Abstract

Severe natural disasters including landslides and tsunamis attacked Thailand from time to time, lowering the national income and increasing international panic. The worst landslide occurred in 1988 after being triggered by heavy rainstorms and many more have occurred since. More recently, the 2004 tsunami triggered by submarine earthquakes struck the west coast of Thailand, causing massive destruction and subsequently arousing national and the international concerns. This disaster has become a national agenda. Many researchers have focused their attention to the issue and urgently tried to find ways to prevent the lost of lives and properties if such phenomenon repeats itself. The objective of this study was to delineate the landslide-probable and tsunami-disaster areas in Phuket Island, Thailand which is a world-renowned tourist attraction. To deal with the problem of landslides, the Geographic Information System (GIS) and remote sensing technology were selected as the analytical tools. The areas possible to be struck by landslides were delineated using the landslide predictive model adopted by the Department of Environmental Geology for landslide prediction in Thailand. The model has eight parameters including elevation, adjusted aspect, slope, flow accumulation, flow direction, LANDSAT TM-band 4, brightness and wetness. Logistic regression was used for the statistical analysis of the model. As for the tsunami which scientifically can be triggered by earthquakes with the magnitude of over 8.2 on the Richter scale and can cause severe destruction of the coastal zone, a case study of such disastrous incident found that the use of remote-sensing data should be best in locating the tsunami-disaster areas. This study uses IKONOS data showing the effects of tsunami on the areas which, as a result, should be considered at risk. Integration of both sets of results would indicate the potential areas that should be considered risk areas of both disasters. The flat coastal areas could be affected by the tsunami while the hill slopes could be affected by the landslide phenomena. The landslide predictive model may also have useful applications in other areas which have similar geology, geography, and climate. It is particularly attractive in areas which is inaccessible or with limited data availability, as only remotely sense data (LANDSAT imagery and/or aerial photographs) and topographic map are required. In addition, the tsunami-disaster areas should be further investigated and the use of remotely sense data especially IKONOS, which has one meter resolution, seemed to be the best tool for the tsunami delineation.