268
M. Werner et al.
and deterioration of air quality. In such conditions it is very difficult to correctly
predict air pollution concentrations.
Many operational air-quality models are initialized using concentrations of
chemical species obtained from the previous day’s forecast, without considering
observations. Data assimilation has not been extensively applied to air quality models because chemical observations are relatively sparse compared to meteorological
observations [5]. Data assimilation in chemical models may improve initial conditions and provides tools for better estimates of emission intensity.
In this study we aim to present the influence of assimilation of ground PM2.5
measurements on forecasted concentrations of particulate matter for low air quality
episode in the year 2017 over Poland. The episode was not reproduced by forecasting
system working operationally without data assimilation. The system is based on the
Weather Research and Forecasting with Chemistry model (WRF-Chem) and produces forecasts daily, starting at 00 UTC and with 72 h lead time and uses chemistry cycling. We used Grid point Statistical Interpolation (GSI) system to assimilate
ground observations from 42 stations measuring PM2.5 concentrations at hourly
basis. The modelled concentrations of PM2.5 and PM10 were compared with observations and with the standard forecasts calculated without data assimilation.
42.2 Data and Methods
42.2.1 The WRF-Chem Model and the GSI System
In this study we used version 3.9 of WRF-Chem. We applied two one way nested
domains—the outer domain covers Europe at a 12 km × 12 km grid and the inner
domain is focused on Poland at 4 km × 4 km resolution. The mother domain has 285
and 332 points in the west-east and south-north direction, respectively. We used 35
vertical levels with the lowest layer top at about 30 m. The simulations are driven by
the GFS meteorological data, available every 3 h, at a 0.5° × 0.5° spatial resolution.
The gas phase chemistry model used in this study was the Regional Acid Deposition
Model, version 2 (RADM2, [4]) and the aerosol module included Goddard Chemistry
Aerosol Radiation and Transport scheme (GOCART, [1]). Two emission databases
were used in the simulations. For the outer domain we included TNO MACC III
database at 1/8° × 1/16°, for the inner domain for SNAP sector 2 and SNAP sector
7 up to date emissions with improved emission factors were provided by the Chief
Inspectorate of Environmental Protection.
We used the 3D-Var component of the GSI system. The system is described in
details in [3] and here only the basis is provided. Analyses are obtained by minimizing
a cost function:
J (x) = (x − x b )
T B
−1
(x − x b ) + (y − H (x))
T R
−1
(y − H (x))
M. Werner et al.
and deterioration of air quality. In such conditions it is very difficult to correctly
predict air pollution concentrations.
Many operational air-quality models are initialized using concentrations of
chemical species obtained from the previous day’s forecast, without considering
observations. Data assimilation has not been extensively applied to air quality models because chemical observations are relatively sparse compared to meteorological
observations [5]. Data assimilation in chemical models may improve initial conditions and provides tools for better estimates of emission intensity.
In this study we aim to present the influence of assimilation of ground PM2.5
measurements on forecasted concentrations of particulate matter for low air quality
episode in the year 2017 over Poland. The episode was not reproduced by forecasting
system working operationally without data assimilation. The system is based on the
Weather Research and Forecasting with Chemistry model (WRF-Chem) and produces forecasts daily, starting at 00 UTC and with 72 h lead time and uses chemistry cycling. We used Grid point Statistical Interpolation (GSI) system to assimilate
ground observations from 42 stations measuring PM2.5 concentrations at hourly
basis. The modelled concentrations of PM2.5 and PM10 were compared with observations and with the standard forecasts calculated without data assimilation.
42.2 Data and Methods
42.2.1 The WRF-Chem Model and the GSI System
In this study we used version 3.9 of WRF-Chem. We applied two one way nested
domains—the outer domain covers Europe at a 12 km × 12 km grid and the inner
domain is focused on Poland at 4 km × 4 km resolution. The mother domain has 285
and 332 points in the west-east and south-north direction, respectively. We used 35
vertical levels with the lowest layer top at about 30 m. The simulations are driven by
the GFS meteorological data, available every 3 h, at a 0.5° × 0.5° spatial resolution.
The gas phase chemistry model used in this study was the Regional Acid Deposition
Model, version 2 (RADM2, [4]) and the aerosol module included Goddard Chemistry
Aerosol Radiation and Transport scheme (GOCART, [1]). Two emission databases
were used in the simulations. For the outer domain we included TNO MACC III
database at 1/8° × 1/16°, for the inner domain for SNAP sector 2 and SNAP sector
7 up to date emissions with improved emission factors were provided by the Chief
Inspectorate of Environmental Protection.
We used the 3D-Var component of the GSI system. The system is described in
details in [3] and here only the basis is provided. Analyses are obtained by minimizing
a cost function:
J (x) = (x − x b )
T B
−1
(x − x b ) + (y − H (x))
T R
−1
(y − H (x))