Estimating flash flood surface flow velocity in heritage city using citizen-recorded videos

Urban flash floods pose significant threats to public safety, infrastructure, and urban resilience. Accurate and rapid assessment of flood flow velocities is essential for effective hazard monitoring and response. However, public awareness of flood risks often remains low, especially regarding the rapid impact of shallow waters on human safety. To address this challenge, this study leverages citizen-recorded videos to estimate flood flow velocities, supporting improved hazard assessment. Focusing on Matera, a UNESCO World Heritage city in southern Italy (Basilicata Region), known for its steep topography and historic “Sassi” districts, the study analyzes five recent flood events (August 24, 2018; November 12, 2019; June 2, 2023; July 2, 2024; and July 21, 2024) using eight citizen-sourced videos, primarily from social media platforms. These videos captured the surface flow along Via Buozzi, a key historic drainage path in the downtown area. Flow velocities were estimated using Fudaa-LSPIV, an image-based particle image velocimetry method adapted for video analysis. The accuracy of these estimates was validated against measurements obtained with SSIMS-Flow, an optical flow software tool, as well as benchmark data derived from floating debris, such as leaves and trash, visually detectable in the videos. The comparison revealed a strong correlation and close agreement between all methods. Results demonstrate the effectiveness and reliability of the proposed approach for estimating surface velocities during urban flash floods. The accompanying datasets, included both in the manuscript and in the supplementary material, offer valuable resources for calibration, validation hydrodynamic modelling and, ultimately, advancing flood risk assessment and management in urban contexts such as heritage city.