Effects of LIDAR DEM resolution in flood modelling : a model sensitivity study for the city of Tegucigalpa, Honduras

An important aspect of hydraulic flood modelling deals with representing topography of river and floodplain. Commonly, flood model applications are reported successful in topographically simple areas were topography only changes gradually and where topography is simulated by DEM’s of relatively low resolution. Such DEM’s are particular useful for flood simulation in rural areas, although important topographic features and properties are not simulated explicitly. In urban areas, however, features like roads, buildings, river banks and dykes have great effect on flow dynamics and flood propagation and as such must be accounted for in the model set-up. This is possible by means of high resolution input data that relates to the systems topography as well as to the identified features. By frequent urban floodings over the past decades, an urgent need is identified to improve and increase our modelling efforts and to address more explicitly the effect model input data has on the simulation results. Society demands accurate and detailed information on magnitude and likeliness of hazardous flood events for design of flood mitigation measures. This also is the case for the city of Tegucigalpa in Honduras that severely has been affected by floodings as caused by extreme rainfall events. For the study area, a DEM with grid size of 1.5 m. is generated from LIDAR data and served as a base line case for various flood simulations. In this study, this DEM is re-sampled and DEM’s of resolutions up to 15 m. are created and serve as input to the flood simulations. By the re-sampling to courser grid elements, averaging across increasingly larger domains is realized and has resulted in an increased loss of detailed topographic properties that affect flood simulations. The original DEM is also used to extract buildings by using geomorphologic filters and other GIS operations. For the simulation, buildings are represented as solid, partially solid and hollow objects by varying the surface roughness value. The sensitivity analysis to DEM resolution revealed that topographic representation is critical and that model output is significantly affected by the resolution of the DEM.