An extended Kalman-filter for regional scale inverse emission estimation

50Citations
Citations of this article
61Readers
Mendeley users who have this article in their library.

Abstract

A Kalman-filter based inverse emission estimation method for long-lived trace gases is presented for use in conjunction with a Lagrangian particle dispersion model like FLEXPART. The sequential nature of the approach allows tracing slow seasonal and interannual changes rather than estimating a single period-mean emission field. Other important features include the estimation of a slowly varying concentration background at each measurement station, the possibility to constrain the solution to non-negative emissions, the quantification of uncertainties, the consideration of temporal correlations in the residuals, and the applicability to potentially large inversion problems. The method is first demonstrated for a set of synthetic observations created from a prescribed emission field with different levels of (correlated) noise, which closely mimics true observations. It is then applied to real observations of the three halocarbons HFC-125, HFC-152a and HCFC-141b at the remote research stations Jungfraujoch and Mace Head for the quantification of emissions in Western European countries from 2006 to 2010. Estimated HFC-125 emissions are mostly consistent with national totals reported to UNFCCC in the framework of the Kyoto Protocol and show a generally increasing trend over the considered period. Results for HFC-152a are much more variable with estimated emissions being both higher and lower than reported emissions in different countries. The highest emissions of the order of 700-800 Mg yr-1 are estimated for Italy, which so far does not report HFC-152a emissions. Emissions of HCFC-141b show a continuing strong decrease as expected due to its controls in developed countries under the Montreal Protocol. Emissions from France, however, were still rather large, in the range of 700-1000 Mg yr-1 in the years 2006 and 2007 but strongly declined thereafter. © Author(s) 2012.

Figures

  • Fig. 1. Time series of (a) HFC-125, (b) HFC-152a, and (c) HCFC141b measured at Jungfraujoch (black) and Mace Head (red) during the five years 2006 to 2010.
  • Fig. 2. Footprint emission sensitivity (in picoseconds per kilogram) averaged over all air masses arriving at Jungfraujoch (JFJ) and Mace Head (MHD) between February 2006 and December 2010.
  • Fig. 3. CO emissions in 2005 according to EMEP/CEIP (http://www.ceip.at/) inventory and reduced by a factor 1000 to match the range of halocarbon emissions. (a) Emis ions on regular 0.5◦× 0.5◦ grid. (b) Emissions on reduced grid with 224 cells used for the inversion.
  • Fig. 3. CO emissions in 2005 according to EMEP/CEIP (http:// www.ceip.at/) inventory and reduced by a factor 1000 to match the range of halocarbon emissions. (a) Emissions on regular 0.5◦×0.5◦ grid. (b) Emissions on reduced grid with 224 cells used for the
  • Table 1. Country emissions and aggregation errors introduced by reduced grid. For columns denoted “with sea” the sea fractions of pixels partially over sea and over land are attributed to the surrounding countries proportional to their relative areal share.
  • Fig. 4. Time series of the synthetic tracer at Jungfraujoch (a) and Mace Head (b) constructed from th original emission grid.
  • Fig. 5. Evolution of emission field for a simulation with noise-free synthetic observations. The original (true) emission field is shown in panel (f) for reference.
  • Fig. 6. Evolution of relative RMS error (RMS difference between estimated and true emissions relative to error of prior) for the case of noise free (black) and noisy (color) synthetic observations. Solid lines are for inversions with a transport uncertainty of ρsrr = 0.8, dashed lines for a lower uncertainty of ρsrr = 0.2. Inversion settings for the noisy case (blue lines) are identical to those for the noise free case (black). The red line is for an inversion identical to the blue solid line but applying the augmented state red-noise Kalman filter.

References Powered by Scopus

Optimal state estimation: Kalman, H∞, and nonlinear approaches

6052Citations
N/AReaders
Get full text

New results in linear filtering and prediction theory

4443Citations
N/AReaders
Get full text

Technical note: The Lagrangian particle dispersion model FLEXPART version 6.2

1487Citations
N/AReaders
Get full text

Cited by Powered by Scopus

Near-term climate change: Projections and predictability

945Citations
N/AReaders
Get full text

Modern inhalation anesthetics: Potent greenhouse gases in the global atmosphere

143Citations
N/AReaders
Get full text

Validation of the Swiss methane emission inventory by atmospheric observations and inverse modelling

95Citations
N/AReaders
Get full text

Register to see more suggestions

Mendeley helps you to discover research relevant for your work.

Already have an account?

Cite

CITATION STYLE

APA

Brunner, D., Henne, S., Keller, C. A., Reimann, S., Vollmer, M. K., O’Doherty, S., & Maione, M. (2012). An extended Kalman-filter for regional scale inverse emission estimation. Atmospheric Chemistry and Physics, 12(7), 3455–3478. https://doi.org/10.5194/acp-12-3455-2012

Readers over time

‘11‘12‘13‘14‘15‘16‘17‘18‘19‘20‘21‘22‘23‘24036912

Readers' Seniority

Tooltip

PhD / Post grad / Masters / Doc 21

49%

Researcher 14

33%

Professor / Associate Prof. 8

19%

Readers' Discipline

Tooltip

Environmental Science 19

50%

Earth and Planetary Sciences 13

34%

Chemistry 4

11%

Chemical Engineering 2

5%

Save time finding and organizing research with Mendeley

Sign up for free
0