Avaliação de manchas de inundação a partir de diferentes técnicas e modelos digitais de elevação para a microbacia do córrego do Gregório em São Carlos-SP

Abstract

Hydroclimatic events in Brazil and other regions of the world have increased in both intensity and frequency, driven by climate change and urban expansion, which intensifies surface runoff due to soil impermeabilization. The Gregório Stream watershed, located in the central region of the municipality of São Carlos, São Paulo, Brazil, has experienced recurrent flooding since the 1930s, resulting in significant socioeconomic losses. Although many other studies have already investigated flooding in the municipality, this study stands out by comparing multiple elevation datasets and simulation models, aiming to identify those that yield the most accurate results and best support urban planning and the development of mitigation plans. In this context, the present study evaluated how the quality of different Digital Elevation Models (DEMs) influenced the performance of HAND and HEC-RAS models in simulating flood extents within the watershed. Global and local DEMs with spatial resolutions of 30, 5, and 1 m were used, and the simulated flood extents were compared with observed flood records provided by the Municipal Civil Defense. In addition to flood extent mapping, the potential of each DEM for delineating the watershed drainage network was also assessed. The evaluation of the drainage networks derived from the different global DEMs identified ANADEM as the most suitable model for representing the watershed hydrography, and it was therefore selected for the subsequent analyses. Regarding the flood simulation models, HAND showed a greater tendency to overestimate flood extents, although it demonstrated considerable effectiveness in identifying flood-prone areas. In contrast, HEC-RAS produced flood maps that more closely matched the observed records, despite requiring a larger amount of input data and calibration parameters. Among the evaluated DEMs, the LiDAR-derived Digital Terrain Model (DTM) yielded results that were most consistent with the flood records from the Municipal Civil Defense, highlighting the importance of high spatial resolution for accurately representing channel topography and simulating flood events. The findings demonstrate that the spatial resolution and quality of DEMs significantly influence flood inundation mapping, and this information is expected to provide a more reliable basis for developing planning and risk management strategies in areas susceptible to extreme hydrological events.