Environmental Engineering Reference
In-Depth Information
One of the goals of the evaluation was to test the ability of the multi-model
ensemble to (a) reproduce the specifics of the episodes (the actual level of
concentrations), preferably, better than the individual models do, (b) reflect the
uncertainty of the predictions represented via an ensemble spread.
In-situ air quality
measurements
In-situ air quality
measurements
Soundings
(meteo)
Soundings
(meteo)
Satellite
data (AOD)
Satellite
data (AOD)
Spatial statistics
Spatial statistics
Time-series analysis
Time-series analysis
Profile analysis
Profile analysis
Modelling
ensemble
Modelling
ensemble
ENSEMBLE system
ENSEMBLE system
Model-
Measurement
Assessment
Software
Model-
Measurement
Assessment
Software
Simple
model ensembles,
Joint model
analysis
Simple
model ensembles,
Joint model
analysis
Google earth
visualization tools
Google earth
visualization tools
Ensemble
Creation
Software
Ensemble
Creation
Software
Timeseries analysis:
Multiple statistics for
individual models
Timeseries analysis:
Multiple statistics for
individual models
Integrated view:
Model results,
measurements,
Geo-information on
station location
Integrated view:
Model results,
measurements,
Geo-information on
station location
Simple and dynamic
ensembles
Simple and dynamic
ensembles
Fig. 1.
A scheme of evaluation of the model ensemble
The current paper presents the details of evaluation of the SILAM model, which
was one of the members of the multi-model ensemble.
For the SILAM model (Sofiev et al., 2006, 2008) evaluation, the input meteoro-
logical data were taken from ECMWF operational archives for the case of 2003
and from FMI-HIRLAM v.6 for both 2006 cases. Anthropogenic emissions of CO,
NH3, NOx, SOx and primary PM from TNO and EMEP emission inventories
were used. The wild-land fire emissions were derived from the MODIS fire radiative
power product using FMI Fire Assimilation System FAS-FRP. Separate computations
were also performed for sea salt. For 2006 cases daily MODIS Collection 5 AOD
(Kaufman et al., 1997) was generated at Giovanni website http://disc.sci.gsfc.nasa.
gov/techlab/giovanni/. Hourly in-situ observations of a few main pollutants were
collected from the Airbase archive of European Environment Agency and included
about 150 sites in 2003 and over 300 for 2006.
2. Results of the Computations and Comparison
with Observations
The episode of February-March 2003 was a continuation of long-lasting poor air
quality conditions
(Fig. 2)
, with the bulk of pollution seemingly originating from
the anthropogenic sources (Stern et al., 2008). The end of the series was put by a
frontal precipitation 7.3.2003, which cleaned the air and promoted better
ventilation due to substantially increased wind speed.
