Urban flood modeling using hydrograph and Rain-on-Grid approaches with HEC-HMS and HEC-RAS: A case study of the Culiacán River

Abstract

Floods are devastating natural disasters, causing extensive economic losses and fatalities globally each year. In developing regions, flood modeling is often hindered by limited or inconsistent hydrometeorological data and historical records. Accurate floodplain modeling is essential for assessing flood risks and identifying vulnerable areas. Combining hydrological and hydraulic models with GIS has become a key approach for simulating and analyzing flood events, particularly in ungauged watersheds. Traditionally, flood simulation is a two-step process where a hydrological model provides flow input for a separate hydraulic model. Modern approaches, such as 'Rain-on-Grid', use a single and integrated step: rainfall is applied directly to a 2D hydraulic model's grid, which handles both runoff generation and the routing of water through the landscape, channels, and streets. This study models urban flooding by integrating a 'Rain-on-Grid' approach for surface runoff with a hydrograph, derived from SCS and Snyder methods, to simulate the river flow. Model performance was assessed against gauging station data. For the SCS model, results showed R² values of 0.927, NSE of 0.678, RMSE of 2.1, and PBIAS of 12.06 for the SCS model; the Snyder model presented an R² of 0.865, NSE of 0.75, RMSE equal to 2.3, and PBIAS of 23.84, indicating comparable performance. Hydraulic simulation was evaluated using the fraction of model performance (F⁽²⁾), with values of 0.7101 (SCS) and 0.7108 (Snyder). Hazard mapping identified rivers, irrigation canals, drains, levees, and road crossings as high-risk zones. Additionally, intermittent streams present significant flood hazards due to rapid onset and high flow velocities, posing threats to vehicles and residents.