Environmental Engineering Reference
In-Depth Information
inorganic aerosols, showing also in this case the overall behaviour of a better
performance of the ensemble, than of any single model.
The reason of this result is the cancelling out of uncorrelated uncertainty and
errors. Where it is difficult to improve a single model, the improvement in
modelling can be achieved by using the ensemble approach.
Future work of the ensemble approach could be in the direction of weighting
the different models based on their separate performance, and by analysing in
more detail the resulting probability density functions, the spread of the results of
the different models.
4. Data Assimilation as Aspect of Model Evaluation
Although by the ensemble approach the overall modelling performance can be
improved, aspects like the underprediction of PM can not be removed. To improve
the performance of CTM's, the combination of observations and model results by
data assimilation is a challenging possibility. Active data assimilation, like 4D-var
and Kalman-Filtering combines in fact information from three different sources:
from the model, from observations and from the current best estimate, including
errors, of the concentration field (Carmichael et al., 2007). Although by data
assimilation models are in principle not improved, the sensitivity of the different
processes and parameterisations can be determined, and the model performance is
improved by incorporating the knowledge and information contained in observations.
Data assimilation is a valid approach to create a better model system (Denby
et al., 2008).
5. Concluding Remarks
Chemistry Transport Models, their meteorological drivers and emission input data
will always be an imperfect representation of the real atmosphere and its chemical
composition. Although some consider models as a second reality, they are just a
model. Models will never determine what is happing in the real world. In other
words, climate change is not caused by climate models.
Acknowledgments Part of this study had been performed in the framework of COST-728.
References
Baklanov, A., Korsholm, U.: On-line integrated meteorological and chemical transport
modelling: advantages and prospectives. In: Proc. 29 th ITM on Air Pollution Modelling and
its Application. Aveiro. Portugal.
Search WWH ::




Custom Search