Atmospheric Environment 36 (2002) 1195–1204 Using wind-direction-dependent differences between model calculations and field measurements as indicator for the inaccuracy of emission inventories J.A. van Aardenne a, *, P.J.H. Builtjes a , L. Hordijk b , C. Kroeze b , M.P.J. Pulles a a TNO Institute of Environmental Sciences, Energy and Process Innovation, Laan van Westenenk 501, Apeldoorn 7334 DT, Netherlands b Environmental Systems Analysis Group, Wageningen University, P.O. Box 9101, 6700 HB Wageningen, Netherlands Received 1 June 2001; received in revised form 20 September 2001; accepted 28 September 2001 Abstract In forward air quality modelling, an emission inventory is used as input into an atmospheric dispersion model to calculate atmospheric concentrations of the pollutant. Differences between calculated concentrations and concentra- tions found by atmospheric measurements can be used as an indicator for the inaccuracy of the emission inventory used in the calculations. The problem with comparing calculated and observed concentrations is that it is not easy to pinpoint the emission inventory as a single cause for the differences. One of the reasons for this is that inaccuracies exist in the model, both in measurements and in the inventory. In this paper, we argue that when wind-direction-dependent differences at several measurement stations in different countries point towards a specific region, the emission estimate for that specific region is the likely cause for the differences between modelled and observed concentrations. We have applied this methodology to study the inaccuracies of a European SO 2 emissions inventory for 1994, by plotting the calculated SO 2 concentrations from a long term ozone simulation model with SO 2 concentrations measured in the EMEP network. The results show that we were able to identify inaccuracies in the emission inventory for several areas within Europe. These areas include Sachsen/Brandenburg (Germany), Central England and the Western part of the Russian Federation. Although this type of analysis is accompanied with several limitations, it could provide the emission inventory community with a relatively simple technique to identify inaccuracies in the emission inventory. r 2002 Elsevier Science Ltd. All rights reserved. Keywords: Emission inventory; Dispersion model; Field measurement; Uncertainty analysis; Sulphur dioxide 1. Introduction Emissions of air pollutants within a country or region are the result of a variety of individual sources. Since it is not practical to measure each emission source individu- ally, the estimation of large-scale emissions is, in most cases, based on the calculation of emissions using an emission factor approach. This emission factor ap- proach aggregates information of sources in both time and space, which will (amongst other reasons, e.g. error in emission measurement) lead to an inaccurate repre- sentation of the real emission. Although we know that emission inventories are inaccurate, we do not know exactly the specific sources or the size of the inaccuracy. In other words, we are ignorant about the source or size of the inaccuracy. This is what we define as the uncertainty of the emission inventory. By performing an uncertainty assessment, an attempt is made to identify the sources of inaccuracy that we are ignorant about and to quantify their impact on the accuracy of the emission estimate. We distinguish between internal and external assessments. In an internal *Corresponding author. E-mail address: john.vanaardenne@algemeen.cmkw.wau.nl (J.A. van Aardenne). 1352-2310/02/$-see front matter r 2002 Elsevier Science Ltd. All rights reserved. PII:S1352-2310(01)00525-8