CEM - Abstract

Abstract Information :

Air pollution is the first environmental health risk significantly affecting morbidity and mortality. Air quality is a key aspect of human health and ecological preservation, which has deteriorated due to the increase in anthropogenic emissions from different economic sectors. To improve the air that citizens breathe, Directive 2008/50/EC of the European Parliament and of the Council of 21 May 2008 on ambient air quality and cleaner air for Europe indicates that each member state classifies its territory into air quality zones that constitute an independent and homogeneous geographical unit for air quality management (monitoring, implementation and evaluation of air quality plans). In this context, the scientific development of air quality models is essential to perform zoning based on data, and CHIMERE complains perfectly with the requirements to perform air quality zoning. In addition, we believe that using neural networks trained with hyperlocal air quality data could improve the resolution and accuracy of the predictions of CHIMERE, allowing automatic high-resolution zoning. This has been validated in the area of the Region of Murcia (Spain), where air quality zoning studies have been performed, obtaining new zoning that has been compared to the official one. This has been used to assess if the stations complain with Directive 2008/50/EC and propose new locations for monitoring points. As a result, we have integrated 601 hyperlocal air quality measurement points, including reference air quality stations; we have developed seven deep learning models, and the CHIMERE mean performance has been improved by 22,3%. This work is the first application of the CHIMERE model in air quality zoning studies and the first integration of CHIMERE with IoT hyperlocal measurements.