Abstract
The unusual heavy rainfall episodes over Kelantan River Basin in 2014 had caused massive destruction and several deaths. The unprecedented storm events at the north-eastern Peninsular Malaysia and many other places indicate the need for enhanced storm forecasting to improve disaster preparedness among the civilian. Quantitative precipitation forecast (QPF) from atmospheric model combined with geostationary meteorological satellite information as input to hydrodynamic model for flood forecasting system can potentially provide improved lead time for warning. In this study, a QPF model is developed using the multilayer neural network with data inputs from the numerical weather prediction (NWP) model products combined with the geostationary meteorological satellite infrared and visible image features to forecast precipitation for a flood-prone area in a tropical region. The results indicate that the model can satisfactorily produce 1-hour rainfall forecast with improved accuracy for larger forecast area. The R2 for areal average rainfall for Kelantan river basin is 0.674 and for Klang river basin is 0.893 whereas the R2 for point rainfall is 0.392 for Kelantan river basin and 0.495 for Klang river basin.
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call