Spatiotemporal dynamics of pedestrian flood hazard in urban areas: beyond peak inundation maps

Extreme urban floods increasingly threaten pedestrians, yet hazard assessments often emphasise peak inundation maps and overlook the duration and overlap of instability conditions that drive real-life exposure and operational decisions. We evaluate the spatiotemporal evolution of pedestrian hazard during a 3-h, 100-year design storm simulated with a coupled rainfall–flood (1D–2D) model for Cúcuta, Colombia (130.5 ha). Every 5 min, gridded flow velocity (V) and water depth (h) were extracted and translated into four hazard levels using widely adopted pedestrian stability indicators (V, h, and V·h). We quantify (i) the fraction of wet area in each hazard class through time (normalised by the instantaneous wet area and by the event’s maximum wet footprint) and (ii) persistence (time above thresholds per cell/sector) and simultaneity (co-occurrence of medium-high/high classes among the three indicators).
The wet footprint expands rapidly and peaks at ~60 min before draining incompletely. Velocity shows an impulsive response, with high-V corridors emerging near the rising limb and collapsing shortly after the peak, while hazardous depths persist longer and concentrate in low-drainage sectors. The combined indicator V·h delineates a critical hazard window (~40–120 min), when threshold exceedance and indicator overlap are maximised, identifying recurrent hotspots and the time intervals most relevant for pedestrian management. The proposed curve-plus-persistence framework complements peak hazard mapping by providing quantitative criteria to prioritise interventions and define operational time windows for closures and warning measures.