 |
| Titel |
Validation of a chemical data assimilation system |
| VerfasserIn |
Jeremy D. Silver, Jesper H. Christensen, Michael Kahnert, Lennart Robertson, Jørgen Brandt |
| Konferenz |
EGU General Assembly 2014
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 16 (2014) |
| Datensatznummer |
250098364
|
| Publikation (Nr.) |
 EGU/EGU2014-14034.pdf |
|
|
|
|
|
| Zusammenfassung |
| Data assimilation can be used with air quality models to improve historical simulations or
initial conditions for forecasts. We present two data assimilation schemes coupled to the
DEHM (Danish Eulerian Hemispheric Model), a three-dimensional, off-line chemical
transport model with full photo-chemistry. The first scheme was a local ensemble transform
Kalman filter, a variant on the ensemble Kalman filter, which uses a low-dimensional
stochastic representation of the background errors. The second scheme involved the
three-dimensional variational method, with a climatological background error model, in
which correlations are assumed to be homogeneous and isotropic in the horizontal plane (i.e.,
depending only on the separation distance), and non-separable in the horizontal, vertical and
chemical dimensions.
Retrievals from polar-orbiting satellites of multiple atmospheric trace gases were
assimilated. These were: OMI tropospheric column NO2, SCIAMACHY total column CH4,
MOPITT partial column CO, and TES partial column O3, CO and CH4. Data for each species
were assimilated independently of one another. Other species were only adjusted indirectly
via the model’s chemistry and dynamics. Assimilation results were compared against
measurements from surface monitoring stations and other satellite retrievals, and preliminary
validation results are presented.
The initial results show are promising. For example, reference simulations (i.e., without
assimilation) grossly underestimate surface CO concentrations, and both assimilation
schemes eliminate this large and systematic bias. At the surface, the assimilation improved to
the spatial correlation for CO, and the winter-time temporal correlation for NO2. However,
results for O3 and CH4 suggest that further work is needed for these species, in terms of the
observation operator or the choice of which level to assimilate. Finally, the potential for
improving modelled surface concentrations is discussed in relation to the vertical sensitivity
of the retrievals. |
|
|
|
|
|
|