The use of historical information about floods is crucial for improving and reducing uncertainty in hydrological dam safety analyses. Traditionally, hydrological analyses for estimating flood frequency curves have been based on simplified hydrometeorological methods and models with limited data, leading to flood quantile estimates with great uncertainties. However, historical information such as water level marks for high-magnitude floods too old to be included in contemporary flood time series recorded at gauging stations are usually available in settlements near existing dams. Incorporating the information about such floods could improve flood quantile estimates associated with high return periods, reducing estimate uncertainties and improving hydrological dam safety analysis.
Therefore, the aim of this study entails using 2D hydrodynamic simulations to estimate flood magnitudes associated with historical floods. A given water level can be generated by a set of flood hydrographs with differing peak flows, flood volumes, and durations. Two-dimensional hydrodynamic simulations allow us to characterise the uncertainty of flood peak and hydrograph volume estimates for historical floods. Such information is incorporated in flood frequency curve estimates.
This study was carried out in the River Douro in Spain. In this area, information about historical water level marks for several floods in the past are available in the reach between the San José Dam and the city of Zamora.
This first step focuses on the calibration and validation of the hydraulic model in the current situation, which were performed using the hydrological data available. It focuses on obtaining a suitable mesh, time step, and landcover for optimal simulation.
With the validated model, simulations will be conducted on historical flood events for which flood marks are available. Given flood hydrograph shapes will be rescaled with varying peak flows and durations to determine the flood hydrographs leading to the recorded water level heights of the flood marks.
