Application of Hydro-Meteorological Model and GIS in Short Range Severe Flood Forecasting and Mapping

Floods resulting from severe seasonal monsoon rainfall are the most important significant natural disaster affecting Malaysia in terms of their impact on the economic, damage to property and sometimes-tragic losses of lives. One of the greatest deficiencies of the current flood models used in the country is the inability to provide cloud and mesoscale rainfall information in the earliest portion of 0-6h of heir forecasting period. In this study NOAA- AVHRR and GMS satellite data were processed for grid based rainfall and rainfall intensity mapping to improve short range quantitative precipitation forecasting (QPF) of severe monsoon weather and also to facilitate the assimilation of QPF into operational flood forecasting. 1D cloud model based QPF modeling process was developed, that relates cloud top temperature below 235oK, reflectance above 28% and cloud heights above 12000m with tropical rainfall formation within the range of 3-12 mm/hr. A grid based rainfall intensity map was thus produced for Langat River Basin. High correlations of R2 above 0.75 were observed for cloud top temperature processed from GOES data and recorded rainfall of severe monsoon weather of selected stations in Terengganu, Kuantan and Kota Bahru. Hydrodynamic and rainfall-runoff simulation were performed using MIKE 11 hydrological model and a suitably auto-calibrated NAM runoff model. The hydrological model was tested for rainfall runoff process using observed hourly rainfall data for the flood event of 27 Sept. to 8 Oct. 2000 as well as rainfall estimation derive from the cloud model QPF using hourly GMS temperature reading for the same storm period. The rainfall- runoff hydrographs generated for the two rainfall sources showed similarities with R2 of 0.9028. The results of the runoff modeling were integrated in MIKE11 GIS model for flood inundation mapping. Separate inundation maps were generated for the observed rainfall and the QPF derive rainfall runoff results for comparison. The accuracies of both maps were verified using grid point location data for flooded areas published in the DID Annual Flood Report. The verification results showed an accuracy of 70% for both flood maps. The methods and processes developed by this study are flexible enough to be applied in other mesoscale and severe storm forecasting particularly a tropical setting. In using NOAA satellite, AVHRR data can be received and processed in advance of 6h of the actual rainfall event. This study should thus prove very useful for the assimilation of grid based rainfall intensity into and improve short- range operational flood forecast.

Preview PDF 1600468.pdf Download (196kB)

Actions (login required)

View Item