 |
| Titel |
Data Assimilation of surface air pollutants in a high resolution air quality model AURORA |
| VerfasserIn |
Ujjwal Kumar, Koen De Ridder, Lefebvre Wouter, Stijn Janssen |
| Konferenz |
EGU General Assembly 2011
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 13 (2011) |
| Datensatznummer |
250052134
|
|
|
|
|
|
| Zusammenfassung |
| In the present work, the optimal interpolation in conjunction with Hollingsworth-Lonnenberg
method to estimate background error covariance matrix has been applied as data assimilation
algorithm in a high resolution air quality model AURORA. This has been done in order to
assimilate the ground level O3 and NO2 concentrations over the Flanders region ([49.8782Ë
N, 1.8841Ë E] x [51.7942Ë N, 7.1159Ë E] ). The grid size was 3x3 km2. Data assimilation
has been carried out in the post-processing offline mode for representative months of two
different seasons: summer and winter. Observations have been provided by AIRBASE data
archive. Since the air quality model AURORA is presumed to represent background
stations more effectively, only the background stations within the domain have been
taken into account while carrying out the data-assimilation process. Because of the
high resolution nature of the AURORA output, the observations of background
stations have been directly assimilated and the validation has also been carried out
accordingly. The validation of the proposed method has been done by leaving out
observations of 10 monitoring stations in one run of data-assimilation process. The
proposed method has been evaluated in both the spatial as well as the temporal
domain. For the month of June-07, temporal correlation coefficients for O3 were
significantly improved at all the observation stations, ranging from 0.70 to 0.95 after
data-assimilation, whereas the same range for raw model output was from 0.35 to 0.66.
Similar improvements have been found for MAE and RMSE which were substantially
reduced after data-assimilation. The index of agreement (IOA) was also significantly
improved ranging from 0.70 to 0.98 after data-assimilation, whereas the same range for
raw model-outputs was from 0.50 to 0.70. In the spatial domain, MAE reduced
from 8.9 to 2.8 while IOA improved to 0.85 from 0.56. Similar improvements have
also been observed during the winter (Dec-07) season. Similarly for NO2, the data
assimilated outputs have been found to be in a much better agreement with the
observations as compared to raw model output for both seasons in both the temporal as
well as the spatial domain. The results clearly indicate that the data-assimilation in
conjunction with Hollingsworth method is a very promising candidate for the statistical
correction of high resolution deterministic air quality model such as AURORA. |
|
|
|
|
|
|