Cities are severely affected by air pollution. Local emissions and urban structures can produce large spatial heterogeneities. We aim to improve the estimation of NO2, O3, PM2.5 and PM10 concentrations in 6 Italian metropolitan areas, using chemical-transport and machine learning models, and to assess the effect on population exposure by using information on urban population mobility. Three years (2013-2015) of simulations were performed by the Chemical-Transport Model (CTM) FARM, at 1 km resolution, fed by boundary conditions provided by national-scale simulations, local emission inventories and meteorological fields. A downscaling of daily air pollutants at higher resolution (200 m) was then carried out by means of a machine learning Random-Forest (RF) model, considering CTM and spatial-temporal predictors, such as population, land-use, surface greenness and vehicular traffic, as input. RF achieved mean cross-validation (CV) R2 of 0.59, 0.72, 0.76 and 0.75 for NO2, PM10, PM2.5 and O3, respectively, improving results from CTM alone. Mean concentration fields exhibited clear geographical gradients caused by climate conditions, local emission sources and photochemical processes. Time series of population weighted exposure (PWE) were estimated for two months of the year 2015 and for five cities, by combining population mobility data (derived from mobile phone traffic volumes data), and concentration levels from the RF model. PWE_RF metric better approximated the observed concentrations compared with the predictions from either CTM alone or CTM and RF combined, especially for pollutants exhibiting strong spatial gradients, such as NO2. 50% of the population was estimated to be exposed to NO2 concentrations between 12 and 38 μg/m3 and PM10 between 20 and 35 μg/m3. This work supports the potential of machine learning methods in predicting air pollutant levels in urban areas at high spatial and temporal resolutions.

This is the first study that quantifies explicitly the impact of present vegetation on concentrations and depositions, considering simultaneously its effects on meteorology, biogenic emissions, dispersion, and dry deposition in three European cities: Bologna, Milan, and Madrid. The behaviour of three pollutants (O₃, NO_2, and PM10) was investigated considering two different scenarios, with the actual vegetation (VEG) and without it (NOVEG) for two months, representative of summer and winter seasons: July and January. The evaluation is based on simulations performed with two state-of-the-art atmospheric modelling systems (AMS) that use similar but not identical descriptions of physical and chemical atmospheric processes: AMS-MINNI for the two Italian cities and WRF-CMAQ for the Spanish city. The choice of using two AMS and applying one of them in two cities has been made to ensure the robustness of the results needed for their further generalization. The analysis of the spatial distribution of the vegetation effects on air concentrations and depositions shows that they are highly variable from one grid cell to another in the city area, with positive/negative effects or high/low effects in adjacent cells being observed for the three pollutants investigated in all cities. According to the pollutant, on a monthly basis, the highest differences in concentrations (VEG-NOVEG) produced by vegetation were estimated in July for O₃ (−7.40 μg/m³ in Madrid and +2.67 μg/m³ in Milan) and NO₂ (−3.01 μg/m³ in Milan and +7.17 μg/m³ in Madrid) and in January for PM10 (−3.14 μg/m³ in Milan +2.01 μg/m³ in Madrid). Thus, in some parts of the cities, the presence of vegetation had produced an increase in pollutant concentrations despite its efficient removal action that ranges from ca. 17% for O₃ in Bologna (January) to ca. 77% for NO₂ in Madrid (July).

Diesel exhaust is hazardous to human health. In time, this has led the EU to impose on manufacturers lower and lower emission standards. These limits are very challenging in particular for nitrogen oxides (NO_x) emitted by diesel-fueled vehicles. For the town of Milan (Italy), we used a complex modeling system that takes into account the NO_x emissions from vehicular traffic and other urban sources, as well as their dispersion and chemical transformations in the atmosphere related to meteorological parameters. The traffic emissions in the Milan urban area were estimated using the geometric and structural characteristics of the road network, whereas the traffic flows were provided by the Environment and Territory Mobility Agency. Car emissions were estimated by the official European method COPERT 5. The nitrogen dioxide (NO₂) concentrations were estimated under two scenarios: the actual scenario with real emissions and the Diesel Emission Standards Compliance (DESC) scenario. Using a recent meta-analysis, limited to European studies, we evaluated the relationship between NO₂ concentrations and natural mortality. For the actual scenario, the NO₂ annual concentration mean was 44.3 µg/m³, whereas under the DESC hypothetical scenario, this would have been of 37.7 µg/m³. This “extra” exposure of 6.6 µg/m³ of NO₂ leads to a yearly excess of 574 “natural” deaths. Diesel emissions are very difficult to limit and are harmful for exposed people. This suggests that specific policies, including traffic limitations, need to be developed and enforced in urban environments.