Determination of the uncertainties of the German emission inventories for particulate matter and aerosol precursors using Monte-Carlo analysis

Abstract

This paper presents the application of a Monte-Carlo simulation for assessing the uncertainties of German 2005 emissions of particulate matter (PM10 & PM2.5) and aerosol precursors (SO2, NOx, NH3 and NMVOC) carried out in the PAREST (PArticle REduction STrategies) research project. For the uncertainty analysis the German Federal Environment Agency’s emission inventory was amended and integrated with a model on the disaggregation of energy balance data. A series of algorithms was developed in order to make efficient and pragmatic use of available literature and expert judgement data for uncertainties of emission model input data. The inventories for PM10 (95 %-confidence interval: −16 %/+ 23 %), PM2.5 (−15 %/+19 %) and NOx (−10 %/+23 %) appear most uncertain, while the inventories for SO2 (−9 %/+9 %), NMVOC (−10 %/+12 %) and NH3 (−13 %/+13 %) show a higher accuracy. The source categories adding the most relevant contributions to overall uncertainty vary across the pollutants and comprise agriculture, mobile machinery in agriculture and forestry, construction sites, small businesses/carpentries, cigarette smoke and fireworks, road traffic, solvent use and stationary combustion. The PAREST results on relative uncertainties have been quoted in the German Informative Inventory Reports since 2012. A comparison shows that the PAREST results for Germany are within the range of (for NH3: close below) other European countries’ results on air pollutant inventory uncertainties as reported in the 2013 Informative Inventory Reports.