Flash flood modeling with a specific focus on arid regions and infiltration
Motivation
This project aims to study different techniques for and applications of flash floods modeling focusing on arid areas and infiltration. Flash floods are among the most dangerous natural hazards and can occur everywhere in the world. The ongoing global warming enhances flash flood-inducing heavy rainfalls. Observations have already shown a significant intensification of heavy rainfalls over the last decades (see e.g. IPCC 2021). Depending on the scenario of Shared Socio-economic Pathways coming along with different greenhouse gas emissions, climate models projected further significant increases for almost all land areas by the middle and end of this century. Ongoing urbanization enhances the intensity of and exposure to urban flash floods due to increased percentages of sealed surfaces and higher building densities. In arid areas, flash floods in usually dry river basins, so-called wadi systems, pose infrastructure, cities, and communities at high risks.
The modeling of flash floods supports several measures of flash flood risk management. These include, on the one hand, non-structural measures such as early-warning systems as well as hazard and risk maps used for urban and landscape planning and risk communication, and on the other hand, structural measures such as retention basins and diversion channels. The combination of hydrological and hydrodynamic models is beneficial as it combines the advantages of short CPU times of hydrological models to generate runoff hydrographs from larger catchments and the high spatial resolution and detailed information of flooding areas for specific areas of interests (e.g. urban areas) provided by hydrodynamic models. For rural catchments, models should account for infiltration as it can have a significant effect on flash floods, especially in arid areas, whose soils have typically a very low initial saturation. Furthermore, the consideration of infiltration is necessary if the effectiveness of sustainable urban drainage systems in terms of infiltration basins and trenches should be investigated with hydrodynamic models.
Study Area
The main study area is the city of El Gouna, located in the Wadi Bili catchment in the Eastern Deser of Egypt. The Eastern Desert of Egypt is characterized by lots of wadi systems coming from the mountains and draining into the river Nile or the Red Sea. Rainfall in this region is rare but sometimes very heavy and can generate flash floods in the wadi systems. Many cities and settlements are located inside the catchments and deltas of the wadis. Just a few areas are protected by dams, drainage canals or embankments. In most cases rainfall data in the mountains are only available through satellites, there might be one or a few rainfall measurement stations in plane parts of the catchment close to the Red Sea or the Nile. Most of the wadis are ungauged, thus almost no runoff measurements are available. During 8 and 9 March 2014, heavy rainfalls caused flash floods in different wadis. For the event in El Gouna, observations of rainfall and runoff are available and are published in Hadidi (2016). El Gouna is a touristic town characterized by many lagoons and is located at the Red Sea coast about 20 km North from Hurghada. The main wadi affecting the city is wadi Bili, whose catchment area has an extension of approximately 880 km² coming from the Red Sea Hills and draining into the Red Sea.
Simulation Results
The hydrological model was calibrated against the observed hydrograph published by Hadidi (2016), a validation was not possible due to a lack of further measurements. The calibrated model was then used to simulate further extreme events in terms of the estimated 100-year event and an assumed worst-case scenario, both with the same temporal rainfall distribution as measured during the flash flood event in 2014. The simulation results of the hydrological model show a highly non-linear relation between rainfall and catchment response of the Wadi Bili catchment. The flash flood coming from Wadi Bili to the city of El Gouna was simulated with the 2D shallow water model hms by using the DEM AW3D30 ©Jaxa to consider the topography and by implementing the different hydrographs generated with the hydrological model as a boundary condition at the wadi outlet before reaching the coastal plain as well as using the temporal rainfall distribution for the different events as a source term. By using the Green-Ampt model, infiltration can be considered. The results of the 2D shallow water model show e.g. the propagation of the flash flood in terms of flow velocities and water depths in each cell of the computational domain for time steps between 1-3 hours over simulation periods between 43-72 hours.
Several structural mitigation measures in the upstream wadi catchment and directly around El Gouna are studied concerning their effectiveness. According to the simulation results, retention basins in the upstream wadi catchment that are designed to completely capture the event that was observed in March 2014 can successfully reduce and lag the downstream flood wave of the 100-year event but do not affect the peak discharge of the assumed worst-case scenario. In addition to reducing the flooding, such basins could retain and store the fresh water in the upstream catchment making it usable for Bedouins in that area. According to the simulation results, the reduction of runoff by retention basins in the upstream wadi catchment has only a minor effect on the flooding in El Gouna, since in the considered cases, the locally fallen rainfall contributes significantly more to the flooding in El Gouna than the flood wave from the wadi, except for the assumed worst-case scenario. In contrast, retention basins and drainage channels directly around El Gouna could significantly reduce water depths in the area under consideration, even in the assumed worst-case scenario. But it has to be aware that overloaded structural measures such as overflowing retention basins or canals could even increase flooding at other locations.
The Green-Ampt approach, which is incorporated in the robust 2D shallow water model to account for infiltration, is used for the model domain of El Gouna, as the area mainly consists of natural surfaces with mostly sandy soil. Literature values for the Green-Ampt parameters of the given soil types lead to a strong overestimation of infiltration. Different rainfall-runoff experiments from the literature, for which observed runoff data is available, are studied to shed light on the applicability and limitations of literature values for the Green-Ampt parameters.
Different sources in terms of satellite images and community-based statements and pictures serve to improve the digital surface model of the study area and validate the simulation results for El Gouna.
Outlook
In future research, the consideration of infiltration and associated processes such as surface crusting or macropore infiltration should be improved in hydrodynamic models. In addition, the reduction of computational times, the availability and usage of high-resolution data, and the establishment of measuring devices for direct measurements should be further enhanced. The growing access to satellite and crowd-sourced data provide the potential to further investigate flash floods and better validate flash flood models, especially for ungauged regions. To study urban flash floods, hydrodynamic models should account for the drainage system and urban structures. In general, the awareness of the risk of flash floods and flash flood risk management need to be further improved.
Publications
Tügel, F., Özgen, I., Hadidi, A., Tröger, U. & Hinkelmann, R. (2018): Modelling of flash floods in wadi systems using a robust shallow water model – case study El Gouna, Egypt. In: Gourbesville P., Cunge J., Caignaert G. (eds) Advances in Hydroinformatics. Springer Water. https://doi.org/10.1007/978-981-10-7218-5_41
Tügel F., Abdelrahman A.A.A., Özgen-Xian I., Hadidi A. & Hinkelmann R. (2020a): Rainfall-Runoff Modeling to Investigate Flash Floods and Mitigation Measures in the Wadi Bili Catchment, Egypt. In: Gourbesville P., Caignaert G. (eds) Advances in Hydroinformatics. Springer Water. https://doi.org/10.1007/978-981-15-5436-0_44
Tügel, F., Özgen-Xian, I., Marafini, E., Hadidi, A. & Hinkelmann, R. (2020b): Flash flood simulations for an Egyptian city - mitigation measures and impact of infiltration. Urban Water Journal. https://doi.org/10.1080/1573062X.2020.1713171
Tügel, F., Hassan, A. & Hinkelmann, R. (2021): Applicability of the Green-Ampt model with literature parameter values to account for infiltration in rainfall-runoff simulations for ungauged catchments. Environmental Modeling and Assessment. https://doi.org/10.1007/s10666-021-09788-0
Tügel, F., Hadidi, A., Özgen-Xian, I., Hou, J. & Hinkelmann, R. (2022): Validation of flash flood simulations using satellite images and community-based observations - impact of infiltration and small-scale topographical features. Springer Book Wadi Flash Floods. https://doi.org/10.1007/978-981-16-2904-4
Simulation results
