HR: 08:45h
AN: A11A-02    [Abstracts]
TI: Estimating Emissions of Ozone Precursors in Europe Through Inverse Modeling of Ozone and Nitrogen Dioxide Observations
AU: * Velders, G J
EM: guus.velders@rivm.nl
AF: Netherlands Environmental Assessment Agency, Netherlands Institute of Public Health and the Environment (RIVM), PO Box 1, Bilthoven, 3720 BA Netherlands
AU: Hanea, R G
EM: r.g.hanea@ewi.tudelft.nl
AF: Netherlands Environmental Assessment Agency, Netherlands Institute of Public Health and the Environment (RIVM), PO Box 1, Bilthoven, 3720 BA Netherlands
AU: Hanea, R G
EM: r.g.hanea@ewi.tudelft.nl
AF: Faculty of Electrical Engineering, Mathematics and Computer Science, Department of Applied Mathematics, Delft University of Technology, PO Box 5031, Delft, 2600 GA Netherlands
AU: Heemink, A
EM: heemink@its.tudelft.nl
AF: Faculty of Electrical Engineering, Mathematics and Computer Science, Department of Applied Mathematics, Delft University of Technology, PO Box 5031, Delft, 2600 GA Netherlands
AB: We report on the first inverse modeling study for estimating emissions of ozone precursors in Europe. A Kalman smoother was applied in combination with the atmospheric chemistry transport model, EUROS, and rural observations of O₃ and NO₂ to estimate emissions of NO_x and VOCs for several groups of countries. The study focused on the middle-west of Europe, i.e. around the Netherlands, Germany, the UK, and France, simply because this region shows the most observations of O₃ and NO₂ at rural sites. Emissions of the ozone precursors are assessed by European countries and reported to EMEP. The NO_x emissions derived with inverse modeling are about 30-40% below the EMEP emissions for the Belgium-Netherlands-Luxembourg and Germany-Denmark regions, and about 60% for the UK-Ireland region. The NO_x emissions in France are slightly higher than the EMEP emissions, but this difference is almost within the uncertainty range. The VOC emissions derived with inverse modeling agree with those from EMEP within the estimated uncertainties. The NO_x and VOC emissions reported by the countries to EMEP show a downward trend for the 1999-2002 period. This downward trend cannot be confirmed by the derived inverse modeling emissions. Extensive simulations with the Kalman smoother show that the uncertainties in derived emissions are dominated by the choice of observations used in the assimilation (O₃, NO₂, or both) and by the choice of noise parameters (NO_x, VOC, or NO_x + VOC emissions). The overall uncertainty in NO_x emissions, estimated as the sum of different components, is estimated at 12-23% for NO_x emissions and at 23-56% for VOC emissions for the Belgium-Netherlands-Luxembourg, Germany-Denmark, France, and UK-Ireland regions. These uncertainties do not take into account the uncertainties associated with possible systematic errors in the physical and chemical parameterizations in the EUROS model. The simulations with different combinations of noise and observations showed that inverse modeling with EUROS yields consistent NO_x emissions if using NO₂, or NO₂ and O₃, observations simultaneously and consistent VOC emissions if using NO_x and VOC emissions simultaneously as noise parameters. Application of noise in VOC emissions and NO_x emissions simultaneously affects the derived NO_x and VOC emissions significantly when compared to the application of noise to these emissions individually.
DE: 0322 Constituent sources and sinks
DE: 0365 Troposphere--composition and chemistry
DE: 3337 Numerical modeling and data assimilation
SC: Atmospheric Sciences [A]
MN: 2005 Joint Assembly