To what extent airborne particulate matters are influenced by ammonia and nitrogen oxides?

Abstract

Intensive farming is known to significantly impact air quality, particularly fine particulate matter (PM$_{2.5}$). Understanding in detial their relation is important for scientific reasons and policy making. Ammonia emissions convey the impact of farming, but are not directly observed. They are computed through emission inventories based on administrative data and provided on a regular spatial grid at daily resolution. In this paper, we aim to validate \textit{lato sensu} the approach mentioned above by considering ammonia concentrations instead of emissions in the Lombardy Region, Italy. While the former are available only in few monitoring stations around the region, they are direct observations. Hence, we build a model explaining PM2.5 based on precursors, ammonia (NH3) and nitrogen oxides (NOX), and meteorological variables. To do this, we use a seasonal interaction regression model allowing for temporal autocorrelation, correlation between stations, and heteroskedasticity. It is found that the sensitivity of PM2.5 to NH3 and NOX depends on season, area, and NOX level. It is recommended that an emission reduction policy should focus on the entire manure cycle and not only on spread practices.