ABSTRACT Flood warning systems are essential tools for reducing risks in extreme hydrological events and, when integrated with effective response actions, contribute to mitigating material damage. However, simulations are conducted with static data, disregarding the temporal dynamics of changes in land use and land cover (LULC) and precipitation and surface run‐off projections. In this study, we quantify the influence of precipitation duration and intensity and soil sealing on surface run‐off and flooded area in an urban watershed in the Brazilian Cerrado. To this end, we propose a hybrid structure that integrates precipitation forecasting (LSTM with attention mechanism and Monte Carlo dropout), land use change modelling (CA‐Markov with U‐net) and hydrological‐hydraulic simulation (HECHMS and HEC‐RAS). The results show that soil sealing has a more significant influence on simulated peak flows than projected variations in precipitation for the period from 2025 to 2040. The simulations indicate that, regardless of changes in projected precipitation intensities, the critical duration remained equal to or less than 2 h, with the connectivity of impervious areas being the dominant factor. For 2‐h events, maximum flows increased by 3. 72% with current CN values and up to 8. 78% in the upper scenario (CN + 11%). For 6‐h events, the increase reached 16. 7% in the most urbanised scenarios. These results indicate a transition from the dominant process of precipitation intensity to surface connectivity in short‐duration events, with direct implications for design standards based on IDF curves. The study supports a reorientation of urban planning towards the disconnection of impervious surfaces and the control of peak flow generation.
Building similarity graph...
Analyzing shared references across papers
Loading...
M. Schmidt
C. C Pereira
Alan Petrônio Pinheiro
Hydrological Processes
Universidade Federal de Uberlândia
Building similarity graph...
Analyzing shared references across papers
Loading...
Schmidt et al. (Wed,) studied this question.
www.synapsesocial.com/papers/69d893626c1944d70ce046d0 — DOI: https://doi.org/10.1002/hyp.70510