Modeling of Rain-Related Inundation Areas in Small Rural Catchments Using the Direct Rainfall Method

The model-based, spatially distributed determination of storm hazards is important for a variety of planning tasks in the field of water resource management. On the basis of flood maps, protective measures can be planned, storm risk management plans can be drawn up, emergency plans and evacuation strategies can be developed, or urban and spatial planning decisions can be made. Especially for small, upstream catchments, there is no standardized method for locating areas at risk from heavy rainfall (storm hazard maps). The scientific investigations presented herein address the question of whether the Direct Rainfall Method (‘DRM’) is a suitable methodological approach for identifying inundation areas in small, rural catchments. For this purpose, a case study is conducted using the internationally applied 2d model HEC-RAS as a Hydrological-Hydrodynamic Rainfall-Runoff Model (‘HHDRRM’). The study area is the Fischbach catchment (38 km²) in Hesse (Germany), a subcatchment of the Gersprenz river system (485 km²). The catchment is part of the field laboratory of the Chair of Engineering Hydrology and Water Management (Fachgebiet Ingenieurhydrologie und Wasserbewirtschaftung – ‘IHWB’) at TU Darmstadt. In addition to addressing the general applicability of HEC-RAS as a HHDRRM, the conducted research focuses on examining the representation of hydrological components and topographic data in the model. To this end, the study is divided into three research questions (sub-questions) concerning the 1. General Applicability, 2. Spatial Resolution, Topography and Sensitivity and 3. Catchment Hydrological Processes and Calibration. These questions are addressed in separate methodological approaches developed in sub-studies. The results of the sub-studies formed the basis for addressing the main objective of the study, namely the applicability of the DRM in the identification of storm hazards. The results obtained with this approach demonstrate that watershed-wide floodplain identification is possible by using the DRM in conjunction with HEC-RAS. It is shown that the proposed method offers great potential for storm hazard analysis in small rural catchments. The model results display good to very good agreements with the hydrological measurements, depending on the event. Through a detailed analysis of the sensitivity of the model, recommendations for the spatial discretization of the topography are derived. Based on an observed storm event it is shown which influence the integration and resolution of catchment hydrological processes (‘CaHyPro’) into the model domain has on the model results. It was shown that a higher spatiotemporal resolution of CaHyPro led to an improvement of the modeling results, independent of a model calibration. At the same time, model limitations and challenges of the DRM in the context of event-dependent hydrological-hydrodynamic modeling are identified. These arise mainly from the computationally intensive application of the DRM compared to conventional rainfall-runoff models, the implemented equation approaches for surface roughness coming from channel hydraulics, and the integration of further runoff-relevant hydrological components. The dissertation iteratively develops modeling solutions to address the identified model deficiencies and identifies further research needs in the application of the DRM.