Articles | Volume 8, issue 21
Atmos. Chem. Phys., 8, 6341–6353, 2008
Atmos. Chem. Phys., 8, 6341–6353, 2008

  05 Nov 2008

05 Nov 2008

Four-dimensional variational data assimilation for inverse modelling of atmospheric methane emissions: method and comparison with synthesis inversion

J. F. Meirink1,*, P. Bergamaschi2, and M. C. Krol1,3,4 J. F. Meirink et al.
  • 1Institute for Marine and Atmospheric research Utrecht (IMAU), University of Utrecht, Utrecht, The Netherlands
  • 2European Commission Joint Research Centre, Institute for Environment and Sustainability (EC JRC IES), Ispra (VA), Italy
  • 3Wageningen University and Research Centre (WUR), Wageningen, The Netherlands
  • 4Netherlands Institute for Space Research (SRON), Utrecht, The Netherlands
  • *now at: Royal Netherlands Meteorological Institute (KNMI), De Bilt, The Netherlands

Abstract. A four-dimensional variational (4D-Var) data assimilation system for inverse modelling of atmospheric methane emissions is presented. The system is based on the TM5 atmospheric transport model. It can be used for assimilating large volumes of measurements, in particular satellite observations and quasi-continuous in-situ observations, and at the same time it enables the optimization of a large number of model parameters, specifically grid-scale emission rates. Furthermore, the variational method allows to estimate uncertainties in posterior emissions. Here, the system is applied to optimize monthly methane emissions over a 1-year time window on the basis of surface observations from the NOAA-ESRL network. The results are rigorously compared with an analogous inversion by Bergamaschi et al. (2007), which was based on the traditional synthesis approach. The posterior emissions as well as their uncertainties obtained in both inversions show a high degree of consistency. At the same time we illustrate the advantage of 4D-Var in reducing aggregation errors by optimizing emissions at the grid scale of the transport model. The full potential of the assimilation system is exploited in Meirink et al. (2008), who use satellite observations of column-averaged methane mixing ratios to optimize emissions at high spatial resolution, taking advantage of the zooming capability of the TM5 model.

Final-revised paper