Global Sensitivity Analysis of the WetSpa model for the Ve river in Vietnam

Doldersum, T. (2009) Global Sensitivity Analysis of the WetSpa model for the Ve river in Vietnam.

Abstract:Global climate change has been widely perceived as one of the main reasons leading to an increase in frequency and magnitude of hydro-meteorological extreme events. These extreme events can lead to flooding. Flooding hinders the socio-economic development on both national and global scale. Flood prediction is an important instrument. The Hanoi University of Science’s goal is to contribute to raise the degree of accuracy in flood forecasting for the Ve River Basin. This research is done within the framework of the Bachelor Thesis. The assignment was to calibrate only for floods the WetSpa model. The WetSpa model is a GIS based-distributed model for flood prediction and water balance simulation on catchment scale. The study area for this thesis is the Ve river basin located in central Vietnam near Quang Ngai. First there is done a preliminary calibration to get a feeling with the model and the study area. Furthermore this calibration has to facilitate determining the ranges for the sensitivity and uncertainty analysis. The second topic of the research was to execute a sensitivity analysis with the method introduced by Morris (1991) to find out the most influential inputs and parameters with regard to the simulated discharges. For calibrating and validating the model three datasets were available. Of which two datasets have been used for calibrating the model. The results for the calibration were rather good, both datasets resulted in a Nash-Sutcliffe coefficient of 0.88. The preliminary verification of the model has been done with the split sample test, by using the third dataset. The result was rather poor, the Nash- Sutcliffe coefficient was equal to 0.57. Moreover the hydrograph of the simulated discharges was also rather poor, only the pattern of the simulated discharges was nearly the same. But the peak values of the simulated flood did not make sense at all, comparing them with the observed values. There are a couple of reasons for this poor verification. First some Vietnamese data is converted into Arcview standards, for instance in Arcview there are 14 land use classes used but in the Vietnamese data there were just 7 classes. Second the warming-up time could be too short. Third there were some struggles with Arcview resulting in not changeable initial conditions, which could have a negative effect on the results. Fourth the Ve river basin becomes an open-basin during an extreme rainfall event, this effect could not be taken into account within this version of the WetSpa model. The method introduced by Morris (1991) is a sensitivity analysis which takes each run one parameter or input variable into account. The main advantage of this method is the relatively low computational cost, the number of runs are needed is equal to a linear function of the number of examined factors. The main disadvantage is that the method cannot estimate individual interactions among factors. The most important results of the sensitivity analysis were that the WetSpa model is not sensitive for the time till the peak discharge is reached, which was also found during the calibration process. Further the most sensitive parameter for the value of the peak discharge and the total volume of the flood is the groundwater recession coefficient. The most important conclusions were that for better results it is important to have more data and more hourly data. For further research it is also recommended to integrate the steps of Arcview and the WetSpa model into one new program. So the whole program can run automatically which increase the calibrating process. Moreover there could be done a sensitivity analysis which takes all parameters and input variables into account.
Item Type:Essay (Bachelor)
Faculty:ET: Engineering Technology
Subject:56 civil engineering
Programme:Civil Engineering BSc (56952)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page