APPLICATION OF C-BAND SYNTHETIC APERTURE RADAR DATA AND DIGITAL ELEVATION MODEL TO EVALUATE THE CONDITIONS OF FLOOD-AFFECTED PADDIES: CHI RIVER BASIN, THAILAND

摘要

With the world's population growing rapidly, the global demand for rice exceeds production. A substantial amount of rice in South and Southeast Asia, where the world's leading rice-exporting countries such as Thailand are located, is damaged by monsoon flooding annually. In order to reduce the loss of rice, an accurate assessment of flood-affected paddies is essential. Taking the 2001 flooding that hit the Chi River Basin, Thailand as an example; the objective is to develop a method for accurate assessment of flood damages to rice paddies. A RADARSAT-1 image acquired during the flood was combined with a 30-m digital elevation model to estimate the floodwater depths. The water depth of 80 cm, which is based on the elongation capability of rice varieties, was used to separate non-damaged from damaged paddies. We evaluated the effectiveness of our method by comparing with the previous method, which utilized multi-temporal (3) RADARSAT-1 scenes of pre-, peak and post-flooding for the supervised classification. Our new method for assessing flood-affected paddies achieved the overall classification accuracy of 87% with a Kappa index of agreement of 0.73, while with the supervised classification method the overall accuracy was 77% with a Kappa index of 0.53. A test of significance for the Kappa indexes shows that our method was more accurate than the previous method, indicating that a single-date SAR scene of the peak flooding can be effectively used for assessing flood damages when it is integrated with a good DEM.