| Abstract: | Hat Yai, the largest commercial and tourist city in southern Thailand, is subjected to regular flood events, primarily during the northeast monsoon period. Flooding in this region is recognized as a serious disaster in terms of frequency, rate of risk, and affected areas. The monsoon of 21–25 November 2000 caused extremely heavy rain in the southern part of Thailand, resulting in a great flood occupying Hat Yai. This caused significant damage. Therefore, the use of both structural and non‐structural measures is mandatory to reduce the economic losses and the risk for society. This paper investigates two modelling approaches for flood prevention and mitigation of Hat Yai city. First, a hard computing approach by a physically distributed model was applied to study the flood behaviour in a two‐dimensional floodplain flow. Second, a soft computing approach using a neuro‐genetic algorithm was used to develop a flood‐forecasting tool. It was found that the great flood of 2000 can be simulated well by the FLO‐2D model. Computed discharges and flood level in the floodplain are close to the observed data. Countermeasures using diversion canals are guaranteed to accelerate the floodwater drainage to Songkla Lake, significantly reducing the flood impact to the people. In addition, the flood forecasting technique developed in this study can give satisfactory results. This would be very useful as a flood‐warning tool for the community Copyright © 2005 John Wiley & Sons, Ltd. |
