Using a 2D shallow water model to assess Large-Scale Particle Image Velocimetry (LSPIV) and Structure from Motion (SfM) techniques in a street-scale urban drainage physical model

Abstract Physically-based numerical modelling of surface processes in urban drainage, such as pollutant wash-off or the assessment of flood risks, requires appropriate calibration and terrain elevation data to properly simulate the overland flows and thus to achieve useful results. Accordingly, this study aims to obtain an accurate representation of the runoff generated by three different rain intensities, 30, 50 and 80 mm/h, in a full-scale urban drainage physical model of 36 m 2 . The study focuses firstly on applying the Structure from Motion (SfM) photogrammetric technique to carry out a high-resolution and accurate topographic survey. This topography was implemented in a 2D shallow water model and the results were compared with those obtained using traditional data point measured topography. Negligible differences were found when comparing the two models with measured discharges at the physical model gully pots. However, significant differences were obtained in the velocity distributions, especially in the shallowest flow areas where drainage channels of a few millimeters’ depth appeared in the high resolution topographic survey. Results from the numerical model were compared with overland flow velocities, determined by applying a modified Large Scale Particle Image Velocimetry (LSPIV) methodology using fluorescent particles. With the SfM topography, the 2D model was able to obtain a better representation of the experimental data, since small scale irregularities of the pavement surface could be represented in the model domain. At the same time, LSPIV was presented as a very suitable tool for the accurate measurement of runoff velocities in urban drainage models, avoiding the interference of raindrop features in the recorded images and with overland water depths in the order of few millimeters.