Comparing Drone and Satellite DEMs for Hydrodynamic Flood Modeling in a Rural Brazilian Catchment

ABSTRACT: The rural region of the municipality of Bananal (SP, Brazil) experiences recurrent flooding events associated with rising water levels in tributaries of the Bananal River, especially during periods of intense rainfall. This study aimed to compare the performance of different Digital Elevation Models (DEMs), one derived from NASA orbital data and another generated from drone-based aerophotogrammetric surveys, in identifying and mapping flood-prone areas. The objective was to assess whether drone field campaigns are essential for this type of analysis or whether orbital DEMs are sufficient for the hydrodynamic characterization of the area. Hydrodynamic models were developed using the software QGIS, HidroFlu—for watershed parametrization and inflow estimation, and MODCEL—for hydrodynamic simulation, with spatial resolutions of 10 m, 30 m, and 50 m, in order to analyze the impact of topographic detail on simulation results. Two approaches were tested for defining boundary conditions: one based on precipitation data with a 25-year return period, and another based on the Bananal River discharge estimated from the watershed. The results indicated that the model based on the drone-derived DEM, with 10 m resolution and boundary conditions defined by river discharge, showed the best performance in representing floodable areas. However, the findings also highlight that high-resolution DEMs entail higher operational costs, due to the need for field activities and greater computational capacity to run the simulations.