Validating CFD modelling of ship plume dispersion in an urban environment with pollutant concentration measurements

Air pollution in urban areas constitutes a global environmental problem, with shipping being one major contributor to hazardous pollutants in harbour areas. This work concerns the application of a method using CFD modelling to study how ships affect the air quality of port areas at a microscale level. A steady RANS-CFD approach was applied to simulate the dispersion of shipping-emitted pollutants, and a spatial sensitivity analysis of the CFD modelling results was conducted. The port of Marseille was used as a case study, and the CFD predictions were compared with on-site observations from two monitoring stations for CO2, CO, NOx, SO2 and PM concentrations. Representative modelled and measured concentrations were considered at the location of the monitoring stations to facilitate one-by-one comparisons for all pollutants in three different test cases of departing vessels. The modelling predictions presented an 8.2% (95% CI:-9.3%, 25.7%) average deviation from the measurements. Validation metrics were included to conduct a statistical comparison between predicted and measured concentrations, with almost all metric values indicating acceptable agreement between the CFD model and measurements. From a technical perspective, this study demonstrates the reliability of the applied CFD modelling method in estimating shipping plume dispersion, while from a societal perspective, this model can serve as an advisory tool for port authorities and policy makers to reduce the impact of shipping emissions on urban air quality.