The sea level is rising, possibly by one meter, by the end of this century. The Netherlands is preparing for this scenario through Kennisprogramma Zeespiegelstijging (KP-ZSS). One of the goals is to determine the hydraulic effects of sea level rise (SLR) on the current system. For the flood risk analysis within
KP-ZSS, the current strength of the flood defences needs to be described. A set of typology-defined fragility curves based on river dikes in river areas describes the current strength of all the primary flood defences in the Netherlands. However, it is unknown if these fragility curves are applicable to sea dikes
due to shorter high water and differences in soil parameters in sea/tidal areas. The research described in this thesis aims to evaluate if site-specific fragility curves for sea dikes result in significantly different predictions of future reinforcement cost for sea dikes compared to the typology-defined fragility curves. The analysis is centred around three dike sections in trajectories 29-3, 30-3 and 32-4 in the Western Scheldt. Initially, an evaluation is conducted on the fragility curves related to the geotechnical failure mechanisms of macro-instability and backward erosion piping. Subsequently, the height requirement is taken into account during the cost calculation. The typology-defined fragility curves underestimate the strength of macro-stability in two out of three
cases with a factor 104 and factor 102, while underestimating it for one section with a factor 105. For the piping failure mechanism the typology-definde fragility curves overestimate the strength in two out of the three cases with a factor 103 and factor 102, but underestimates it for one section with a factor 103.
The pre-overburden pressure appeared to be the most important factor influencing the failure probability of the macro-stability failure mechanism, with higher occurring values for the sections in the Western
Scheldt. The failure mechanism of piping was influenced most by the hydraulic conductivity, where the
encountered soil in the Western Scheldt consisted of finer soils with lower hydraulic conductivity.
From the results, it can be concluded that the use of site-specific fragility curves resulted in a decrease of 13% in net present value, averaging over all SLR scenarios for an analysis until 2200. Considering the most probable SLR scenario resulted in a 12% decrease. If the height requirement is included in the cost calculation, the use of site-specific fragility curves results in a 7% decrease in net present value on average and a 12% decrease for SLR scenario low. For the low SLR scenario, the cost of relocating the road infrastructure in and around the expansion zone of the dike is dominant. With increasing SLR scenarios, the increase in crest height becomes the most important factor, with the revetment the dominating factor cost-wise, leading to minimal differences in reinforcement cost between the fragility
curve approaches.

In the southeast of Spain heavy storms can occur in the late summer and early autumn, during which great amounts of rain pours down. During a storm that took place in September 2019, over 300 mil- limetres of rain fell in just five hours. Throughout the 2019 flooding event, the Rambla de la Carrasquilla has overflowed its banks and the adjacent agricultural lands flooded and large amounts of water ac- cumulated in the streets of Los Nietos. Since the 2019 flood, improvements to the river system of the Rambla de la Carrasquilla have been made. For instance: a relatively small culvert was replaced with a bigger one in 2021, to increase the discharge capacity. Due to global warming, the probability of occurrence of the heavy storms will increase and thus the probability of flooding. However, the lack of historical data makes it hard to assess the influence of different storms and the replacement of the cul- vert on future floods. Numerical models can be used to make predictions of floodings and gain insight in storm impacts. The aim of this report is summarized by the question: How can the flood areas and peak water levels of the Rambla de la Carrasquilla and its catchment be obtained by simulating three different scenarios based on peak discharges using HEC-RAS models?. At first a spatial analysis of the river system was done using QGIS. Hereafter a 1D and 2D model were constructed with the help of a Digital Terrain Model (DTM). QGIS and RiverGIS were used to define the geometry of the system for the 1D model, where the 2D model geometry was constructed in HEC-RAS itself. The dimensions of the structures in the river were measured using photogrammetry, due to which dimensions could be obtained with an absolute error of around 2 cm. Fieldwork was carried out to estimate the surface roughness of the main channel and parts of the flood plains.
The results consist of different floodmaps that were made using 1D and 2D HEC-RAS models with different settings and boundary conditions. For the 1D model, one run was done using a single reach for the whole river and another including a bifurcation for the downstream part of the river. It can be concluded that the 1D model approach is not very suitable for our research as it is unable to accurately construct the flooded areas. The 2D model performed better and thus was used to further investigate the influence of different model parameters and inputs. Important take-away points are that the time step is of large influence for finding a stable solution, which is found around Δ𝑡 = 1 [second] in this report. The influence of the surface roughness of the river bed and floodplains was investigated by using Manning’s n. These results showed clearly that a higher value for Manning’s n leads to a larger flooded area. When investigating the influence of structures (two bridges, a pipeline and a culvert) on the flooded area, it is found that the structures do not have a great influence. The reason for this is that at critical points, the river already overflows its banks when the flow is not hindered by these structures. Adding the structures makes this only slightly worse. Furthermore, the 2021 culvert only showed a slight improvement in flooded area with respect to the 2019 culvert.
In order to further calibrate the HEC-RAS models, a comparison was made with floodmaps that were constructed by the Spanish government. Both studies show similar results, although the government used a Digital Terrain Model with a higher resolution, which can have a large impact on the outcome. However, this government study cannot be used for validation of the models in this research, as the results are based on hypothetical situations instead of empirical data as well.
For further research it is recommended to investigate more accurate hydrological boundary conditions, based on empirical rainfall data. For both boundary conditions a constant value was taken, but the behaviour over time could change the output of the results. Besides that, it could be interesting to study the exact effect of using a Digital Terrain Model with a higher resolution for the Rambla de la Carrasquilla. As this can have a direct influence on the size of the flooded area. Furthermore, the approaches used in this research can be used for investigating different flood mitigation solutions. These solutions, for example heightening of the river banks or widening of the bed, can be implemented in the 2D model before applying them in real life.