Advance ¯ood forecasting for ¯ood stricken Bangladesh with a fuzzy reasoning method Shie-Yui Liong, 1 Wee-Han Lim, 1 Toshiharu Kojiri 2 and Tomoharu Hori 3 * 1 Disaster Prevention Research Institute, Kyoto University, Kyoto 611-0011, Japan; on leave from Department of Civil Engineering, National University of Singapore, Singapore 119260 2 Disaster Prevention Research Institute, Kyoto University, Kyoto 611-0011, Japan 3 Department of Civil Engineering Systems, Kyoto University, Kyoto 606-8501, Japan Abstract: An arti®cial Neural Network (NN) was successfully applied, in an earlier study, as a prediction tool to forecast water level at Dhaka (Bangladesh), for up to seven lead days in advance, with a high accuracy level. In addition, this high accuracy degree was accompanied with a very short computational time. Both make NN a desirable advance warming forecasting tool. In a later study, a sensitivity analysis was also performed to retain only the most sensitive gauging stations for the Dhaka station. The resulting reduction of gauging stations insigni®- cantly aects the prediction accuracy level. The work concerning the possibility of measurement failure in any of the gauging stations during the critical ¯ow level at Dhaka requires prediction tools which can interpret linguistic assessment of ¯ow levels. A fuzzy logic approach is introduced with two or three membership functions, depending on necessity, for the input stations with ®ve membership functions for the output station. Membership functions for each station are derived from their respective water level frequency distributions, after the Kohonen neural network is used to group the data into clusters. The proposed approach in deriving membership function shows a number of advances over the approach commonly used. When prediction results are compared with measured data, the prediction accuracy level is comparable with that of the data driven neural network approach. Copyright # 2000 John Wiley & Sons, Ltd. KEY WORDS fuzzy logic; neural network; ¯ood forecasting; Bangladesh INTRODUCTION Bangladesh is known as the land of six seasons, the most dominant being the rainy season which lasts from May to September. The country is located on the Tropic of Cancer at longitude 908E and has a land area of 145 000 km 2 . About 60% of the land in Bangladesh is located on the world's largest delta comprising three of the world's most unstable rivers, the Ganges, the Brahmaputra and the Meghna, which annually ¯ood during the monsoon period. The rivers ¯owing into Bangladesh drain some of the wettest catchment areas on earth with average rainfalls as high as 1100 cm. The major rivers and their tributaries have their origins outside Bangladesh and only about 7 . 5% of their total catchment area of about 1 . 5 million km 2 lie within Bangladesh. Ninety percent of their annual ¯ows originate outside the country. CCC 0885±6087/2000/030431±18$17 . 50 Received 2 December 1998 Copyright # 2000 John Wiley & Sons, Ltd. Accepted 8 February 1999 HYDROLOGICAL PROCESSES Hydrol. Process. 14, 431±448 (2000) *Correspondence to Dr T. Hori, Department of Civil Engineering Systems, National University of Singapore, Singapore 119260. E-mail: hori@hydro7.kuciv.kyoto-u.ac.jp Contract/grant sponsor: National University of Singapore. Contract/grant number: RP950685. Contract/grant sponsor: Disaster Prevention Research Institute of Kyoto University.