2D FLOOD MODELLING AND SENSITIVITY ANALYSIS: A CASE STUDY OF TOUS DAM BREAK

Abstract

Worldwide Floods has been deemed as the most dangerous natural calamity impacting adversely the environment and society, claiming millions of lives and damaging the properties. Compared to other natural disaster like (forest fire, land slide, volcanic eruption and drought) flooding has the most devastating impacting on human being worldwide. Across the globe flooding is responsible for approximately 50% of water related disaster, more than 90 countries are exposed to dreadful flooding effecting on average 196 million people. Flooding generally occurred due to heavy precipitation or melting of snow, and dam or dyke breaks. The measure to mitigate flooding can be structural (dams, dykes and levees etc.) and non-structural (Numerical model). On October 20 and 21, 1982, an exceptional meteorological condition led to extremely heavy rainfall in the hinterland of the central Mediterranean coast of Spain. The Júcar river was badly affected and the Tous dam, only a few kilometers upstream of Sumacàrcel towns, failed on October 20 at about 19:00 hour with destructive effect downstream. This research was carried out to better understand 2D flooding in rivers and urban areas, as well as to test the sensitivity of SRH-2D model by employing different hydraulic parameter using the Morris method of global sensitivity analysis. The Tous dam break event is successfully simulated using SRH-2D hydraulic model with both 1982 and 1998 DTM. It is concluded that the coarse mesh can produce better results in case of both DTM with larger time step and less computational time as long as the code does not diverge. Sensitivity analysis is extensively used and accepted as a good modelling practice. It is used as a useful tool to study which input influences the output most. Five parameters are selected as potentially important parameters for this analysis. Time Step, Manning value for valley (River N), Manning value for orange trees (Tree N), Depth of water on downstream boundary condition, and PARA (depth averaged parabolic turbulent model). The Morris method is most suitable for complex and nonlinear models to evaluate the sensitivity and interaction of factors. From the graph plotted between standard deviation and mean of elementary effect its evident that the Depth of water on downstream boundary condition is the most important parameter with interaction effect and Tree N has overall influence on model results with higher mean values. Our findings indicate that the sensitivity of input factors differ depending on the model output assessed, and the location and time of when and where this output is most relevant.