A hybrid Eulerian-Lagrangian-statistical approach to evaluate air quality in a mixed residential-industrial environment

Abstract

The aim of this work is to evaluate the air quality in a mixed urban-industrial environment. Concentration gradients can often be strong, owing to the complex orographic and meteorological context. It can be therefore misleading to rely only on the monitoring network. We used a Eulerian chemistry-transport model (CTM) to simulate the regional background concentrations of PM10, and a Lagrangian model (LM) to simulate the local impact of the emissions from an iron production plant. Firstly, regional scale background concentrations were estimated with a data fusion approach, interpolating the observations of the background stations using concentrations simulated by the CTM as a proxy. Secondly, the residuals over the urban stations were interpolated using concentrations simulated by the LM as a proxy. The performance of this modelling approach was evaluated using independent data and through a leave-one-out cross validation. The effectiveness of different strategies for impact mitigation was also assessed.