Flood analysis comparison with probability density functions and a stochastic weather generator

Flood prediction has become essential to hydrology and natural disaster management due to the increasing frequency and severity of extreme hydrological events driven by climate change. This study compares two methodologies for predicting flood events in Morelia, Mexico: theoretical distribution functions and stochastic weather generators. The methodology integrates maximum runoff results for different return periods into a drainage network hydraulic model, using the Soil Conservation Service Curve Number (SCS-CN) method and a multivariate stochastic model (MASVC). Hydrodynamic modeling with HEC-RAS, incorporating two-dimensional shallow water equations, was used to simulate flood inundation areas. The study reveals that while both modeling approaches similarly replicate the system's behavior, they produce different water levels due to variations in maximum flow values. The stochastic model tends to generate higher maximum water levels. High-resolution digital elevation models (DEMs) with a pixel size of five m in urban areas and 0.5 m in drainage network zones, and land use data were crucial in improving the accuracy of the hydraulic simulations. Findings indicate that unregulated urban growth in flood-prone areas significantly exacerbates the impact of flooding. The generated hazard maps and flood simulations provide valuable tools for urban planning and decision-making, highlighting the need for strategic interventions to mitigate flood risks. This research underscores the importance of integrating advanced modeling techniques in flood risk management to enhance the precision and reliability of flood predictions.