Air quality in enclosed railway stations: Quantifying the impact of diesel trains through deployment of multi-site measurement and random forest modelling

Abstract

Concentrations of the air pollutants (NO₂ and particulate matter) were measured for several months and at multiple locations inside and outside two enclosed railway stations in the United Kingdom - Edinburgh Waverly (EDB) and London King's Cross (KGX) - which, respectively, had at the time 59% and 18% of their train services powered by diesel engines. Average concentrations of NO₂ were above the 40 μg m^-3 annual limit value outside the stations and were further elevated inside, especially at EDB. Concentrations of PM_2.5 inside the stations were 30-40% higher at EDB than outside and up to 20% higher at KGX. Concentrations of both NO₂ and PM_2.5 were highest closer to the platforms, especially those with a higher frequency of diesel services. A random-forest regression model was used to quantify the impact of numbers of different types of diesel trains on measured concentrations allowing prediction of the impact of individual diesel-powered rolling stock.

Keywords: Diesel exhaust; Diesel trains; Enclosed railway stations; Random forest.