Sensitivity analysis of the microphysics scheme in WRF-Chem contributions to AQMEII phase 2
The parameterization of cloud microphysics is a crucial part of fully-coupled meteorology-chemistry models, since microphysics governs the formation, growth and dissipation of hydrometeors and also aerosol cloud interactions. The main objective of this study, which is based on two simulations for Europe contributing to Phase 2 of the Air Quality Model Evaluation International Initiative (AQMEII) is to assess the sensitivity of WRF-Chem to the selection of the microphysics scheme. Two one-year simulations including aerosol cloud interactions with identical physical-chemical parameterizations except for the microphysics scheme (Morrison –MORRAT vs Lin –LINES) are compared. The study covers the difference between the simulations for two three-month periods (cold and a warm) during the year 2010, allowing thus a seasonal analysis. Overall, when comparing to observational data, no significant benefits from the selection of the microphysical schemes can be derived from the results. However, these results highlight a marked north-south pattern of differences, as well as a decisive impact of the aerosol pollution on the results. The MORRAT simulation resulted in higher cloud water mixing ratios over remote areas with low CCN concentrations, whereas the LINES simulation yields higher cloud water mixing ratios over the more polluted areas. Regarding the droplet number mixing ratio, the Morrison scheme was found to yield higher values both during winter and summer for nearly the entire model domain. As smaller and more numerous cloud droplets are more effective in scattering shortwave radiation, the downwelling shortwave radiation flux at surface was found to be up to 30 W m−2 lower for central Europe for the MORRAT simulation as compared to the simulation using the LINES simulation during wintertime. Finally, less convective precipitation is simulated over land with MORRAT during summertime, while no almost difference was found for the winter. On the other hand, non-convective precipitation was up to 4 mm lower during wintertime over Italy and the Balkans for the case of including Lin microphysics as compared to the MORRAT simulation.