| Summary: | The increase of the CO<sub>2</sub> content in the atmosphere caused by anthropogenic emissions from the territories of large cities (~70%) is the critical factor in determining the accuracy of emission estimations. Advanced experiment-based methods of anthropogenic CO<sub>2</sub> emission estimation are based on the solution of an inverse problem, using accurate measurements of CO<sub>2</sub> content and numerical models of atmospheric transport and chemistry. The accuracy of such models decreases the errors of the emission estimations. The aim of the current study is to adapt numerical weather prediction and atmospheric chemistry model WRF-Chem and validate its capability to simulate atmospheric CO<sub>2</sub> for the territories of the two large coastal cities of the Gulf of Finland—St. Petersburg (Russia) and Helsinki (Finland). The research has demonstrated that the WRF-Chem model is able to simulate annual variation, as well as the mean seasonal and diurnal variations of the near-surface CO<sub>2</sub> mixing ratio, in Helsinki, at a high spatial resolution (2 km). Correlation between the modelled and measured CO<sub>2</sub> mixing ratio is relatively high, at ~0.73, with a mean difference and its standard deviation of 0.15 ± 0.04 and 1.7%, respectively. The differences between the WRF-Chem data and the measurements might be caused by errors in the modelling of atmospheric transport and in a priori CO<sub>2</sub> emissions and biogenic fluxes. The WRF-Chem model simulates well the column-averaged CO<sub>2</sub> mixing ratio (XCO<sub>2</sub>) in St. Petersburg (January 2019–March 2020), with a correlation of ~0.95 relative to ground-based spectroscopic measurements by the IR–Fourier spectrometer Bruker EM27/SUN. The error of the XCO<sub>2</sub> modelling constitutes ~0.3%, and most likely is related to inaccuracies in chemical boundary conditions and a priori anthropogenic CO<sub>2</sub> emissions. The XCO<sub>2</sub> time series in St. Petersburg by the WRF-Chem model fits well with global CAMS reanalysis and CarbonTracker-modelled data (the differences are less than ~1%). However, due to much higher spatial resolution (2 vs. over 100 km), the WRF-Chem data are in the best agreement with the ground-based remote measurements of XCO<sub>2</sub>. According to the study, the modelling errors of XCO<sub>2</sub> in St. Petersburg during the whole simulated period are sufficiently minimal to fit the requirement of “Error ≤ 0.2%” in 60% of cases. This requirement should be satisfied to evaluate properly the anthropogenic CO<sub>2</sub> emissions of St. Petersburg on a city-scale.
|
|---|