1. Charles University in Prague
Faculty of Mathematics and Physics
Department of Atmospheric Physics
V Holešovičkách 2, Prague 8, Czech Republic
Jan Karlický, Peter Huszár, and Tomáš Halenka
Validation of gas phase chemistry modules used in WRF/chem model on domain including Europe
1. Objectives
O3 [μg/m3] in Suburban stations
O3 [μg/m3] in Urban stations
Evaluate the ability of atmospheric model WRF/chem to capture temporal and spatial distribution of short-lived gas concentrations
Compare gas-chemistry modules CBMZ and RADM2
Study links between emissions, amount, and concentrations of some gases
2. Methods
a) Model setup
Model: WRF/chem v , simulation year 2008, 5-days spin-up
Domain: ENSEMBLES 190x206 points, 25 km resolution, 30 vertical levels
Boundary conditions: ERA-interim data, MOZART-4/GEOS-5, Clim. UBC
Parameterization: Radiation RRTM(LW) + Goddard(SW), Morrison double-moment scheme (Microphysics), Noah Land Surface Model, MM5 surface layer, PBL – Yonsei University scheme, convection – Grell-Freitas scheme
Emission inputs: TNO MACC-II anthropogenic emission data, MEGAN biogenic emissions, FINN SAPRC99 biomass burning emissions
Chemical modules: CBMZ and RADM2 (gas only)
BIAS, CC, and SLOPE for all gases and st. types
b) Reference data
E-OBS v.11 (physical part of model), gridded data with 0.25° resolution
AirBase database – The European air quality database (v. 7, chemical part), European station data
c) Links between biogenic emission of Isoprene, total column of Formaldehyde, and NOx total column
Isoprene is the main component of biogenic emissions for Formaldehyde production
But there is not direct relation between Isoprene and Formaldehyde amount, NOx is needed for decomposition of Isoprene [Wolfe et al., 2015]
The resulting images are composed by localities with prevailing biogenic emissions (less NOx) and urban emissions (more NOx and Formaldehyde, but less Isoprene)
Chosen AirBase background stations in (Rural – green, Suburban – orange, Urban – red) O3
NO2
Relation between biogenic em. of Isoprene (moles/km2), Formaldehyde total column (1019 molec/m2), and NOx total column (1019 molec/m2)
(Blue – CBMZ, Red – RADM2; in the middle and on the left – CBMZ only) CO
SO2
4. Summary, Discussion
Differences between simulation with CBMZ and RADM2 modules in terms of meteorology are negligible – there is only low impact from gas-phase chemistry to meteorology (direct to radiation transfer)
Impact of higher precipitation or lower daily amplitude of temperature to chemistry could be stronger – through wet scavenging or less sunshine
From presented gases, simulated Ozone concentrations make the best results. Probably reason is that Ozone is secondary pollutant, so it is not directly emitted by sources and its spatial distribution is not as strongly varying near emission sources as in term of CO or SO2, so it could be captured by model with 25 km resolution much better than primary pollutants CO or SO2
Formaldehyde total amount depends strongly on NOx total column, which could be seen on middle image, except urban NOx localities we can see ‘triangle’ with base of low and top of higher NOx total column.
c) Statistical quantities
BIAS = Model deviation from measurement normalized by reference state (%)
CC – Standard Correlation Coefficient
SLOPE – Standard Regression Coefficient
3. Results
a) Validation of Physical state of model WRF/chem
Full-domain season average biases for temperature and precipitation
Average biases for temperature (°C) and rainfall (mm)
5. References
Haylock, M.R., N. Hofstra, A.M.G. Klein Tank, E.J. Klok, P.D. Jones, M. New. 2008: A European daily high-resolution gridded dataset of surface temperature and precipitation.
Peckham, E., et al., 2013: WRF/Chem Version 3.5 User’s Guide
Wolfe, G.M. et al., 2015: Formaldehyde production from isoprene oxidation across NOx regimes
b) Correlation between model and station gas concentrations
Scatter plot of model and station daily O3 values over 2008 (rural background st.)
CBMZ v. RADM2
O3 [μg/m3] in Rural stations
Acknowledgement
This work is supported by the project GAUK No / Authors wish to express their thanks to WRF team community for their development of the model WRF and chemistry modules used in the study. Further, our thanks go to the ECA&D team for E-OBS gridded data and European Environment Agency for AirBase data used for validation and result analysis. Authors also thanks to MACC organisation for TNO emission data, and also to NCAR for MOZART-4/GEOS-5 and Biomass Burning emission data, which are all used in model simulations.