Sensitivity Analysis of Flash-Flood Modelling in Grenada, As A Small Island Caribbean States

摘要

Flash-flood is an unpredictable event, has a great and destructive power. Because of these characteristics, flash-flood gives the extremely dangerous impacts to society, infrastructures, and economic losses. In order to reduce those impacts, it is required to understand the uncertainty conditions in flash-flood susceptibility areas and prediction of the flash-flood impacts by conducting flash-flood modelling. One of the appropriate software to develop this model is OpenLISEM. This modelling is required in Gouyave and St. John's watershed, as the higher susceptible areas to flash-flood in Grenada. In order to get the best model's performance, an appropriate configuration of the model is required. Therefore, it is important to analyze the parameters that influence this model by conducting sensitivity analysis. This paper aims to find out the most influential parameter to the result of flash-flood modelling. Parameters that are analyzed, such as saturated hydraulic conductivity (Ksat), initial soil moisture, surface roughness (manning's n), random roughness, and fraction coverage of vegetation. The result shows that Ksat is the most sensitive parameter of flash-flood modelling. When these parameters values increase or decrease, they always give the different values of flash-flood characteristics. It implies that these parameters have significant influences to flash-flood characteristics that are generated by this model.