Impact of mixing layer height variations on air pollutant concentrations and health in a European urban area: Madrid (Spain), a case study.

The occurrence of local high-pollution episodes in densely populated urban areas, which have huge fleets of vehicles, is currently one of the most worrying problems associated with air pollution worldwide. Such episodes are produced under specific meteorological conditions, which favour the sudden increase of levels of air pollutants. This study has investigated the influence of the mixing layer height (MLH) on the concentration levels of atmospheric pollutants and daily mortality in Madrid, Spain, during the period 2011-2014. It may help to understand the causes and impact of local high-pollution episodes. MLH at midday over Madrid was daily estimated from meteorological radio soundings. Then, days with different MLH over this urban area were characterized by meteorological parameters registered at different levels of an instrumented tower and by composite sea level pressure maps, representing the associated synoptic meteorological scenarios. Next, statistically significant associations between MLH and levels of PM, PM, NO, NO, CO and ultra-fine particles number concentrations registered at representative monitoring stations were evaluated. Finally, associations between all-natural cause daily mortality in Madrid, MLH, and air pollutants were estimated using conditional Poisson regression models. The reduction of MLH to values below 482 m above-ground level under strong atmospheric stagnation conditions was accompanied by a statistically significant increase in levels of NO, NO, CO, PM and ultra-fine particle number concentrations at urban-traffic and suburban monitoring sites. The decrease of the MLH was also associated to a linear increase of the daily number of exceedances of the UE NO hourly limit value (200 μg/m) and levels of air pollutants at hotspot urban-traffic monitoring stations. Also, a statistically significant association of the MLH with all-natural cause daily mortality was obtained. When the MLH increased by 830 m, the risk of mortality decreased by 2.5% the same day and by 3.3% the next day, when African dust episodic days were excluded. They were also higher in absolute terms than the increases in risk of mortality that were determined for the exposition to any other air pollutant. Our results suggest that when the prediction models foresee values of MLH below 482 m above-ground level in Madrid, the evolution of high-contamination episodes will be very favourable. Therefore, short-term policy measures will have to be implemented to reduce NO, NO, CO, PM and ultra-fine particle emissions from anthropogenic sources in this southern European urban location.