Flood Forecasting: Disaster Risk Management Initiative

1271 Words3 Pages

As a non-structural measure, flood forecasting (such as discharge, water level, or flow volume) is a crucial part of flow regulation and water resources management. Worldwide, flood disasters account for about one-third of all natural disasters in terms of number and economic losses (Berz 2000). As stated by Dutta and Herath (2004), out of the total number of flood events in the world during the past 30 years, 40% occurred in Asia and Southeast Asia countries stand for the second worst region in Asia. ASEAN Disaster Risk Management Initiative (2010) reported that a catastrophic 200-year flood (0.5 percent annual probability of exceedance) would have a major impact on the economies of the Southeast Asian countries, including Myanmar, which are already fragile. The process of floods is basically complex, uncertain and unpredictable, due to its nonlinear dependency on meteorological and topographic parameters (Thirumalaiah and Deo 1998). While distributed hydrological modeling involves multidisciplinary and complex issues, simple, robust and sustainable approaches in flood forecasting system are needed, without much effort in continuous updating such models. For flood forecasting to be effective, it must provide flood warnings with a reasonable lead time. Furthermore, for real time operation, the authorities may require to access the gauges of significant predictors (Corani and Guariso 2005), thus saving considerable costs, a critical issue in developing countries.
Since a flood warning and forecast system does not primarily aim at providing explicit knowledge of rainfall–runoff processes, black-box models have been widely used in addition to the physical based models (Abudu et al. 2010; Magar and Jothiprakash 2011). The main focus o...

... middle of paper ...

...dy has not only presented the robustness of ANN models in multi-step flood forecasting with limited data types, but also assessed their clear-cut superiority to regression models, for the conditions under which the regression technique has the best performance. To establish the true merits of ANNs relative to conventional statistical techniques, comparisons are made between the forecasting performance of ANN and stepwise multiple linear regression (SMLR) models. Two conditions are addressed in the comparison of forecasting skills: (a) using a-site data only and (b) using at-site and upstream data. This paper is an effort to improve national flood forecasting systems in Myanmar by applying ANN models which offer more advantages than the conventional regression models. Additionally, the results of this study can be applied to similar basins and further researches.

More about Flood Forecasting: Disaster Risk Management Initiative

Open Document