Employing airborne radiation and cloud microphysics observations to improve cloud representation in ICON at kilometer-scale resolution in the Arctic
Clouds play a potentially important role in Arctic climate change but are poorly represented in current atmospheric models across scales. To improve the representation of Arctic clouds in models, it is necessary to compare models to observations to consequently reduce this uncertainty. This study co...
| Published in: | Atmospheric Chemistry and Physics |
|---|---|
| Main Authors: | , , , , |
| Format: | Text |
| Language: | English |
| Published: |
2020
|
| Subjects: | |
| Online Access: | https://doi.org/10.5194/acp-20-13145-2020 https://acp.copernicus.org/articles/20/13145/2020/ |
| Summary: | Clouds play a potentially important role in Arctic climate change but are poorly represented in current atmospheric models across scales. To improve the representation of Arctic clouds in models, it is necessary to compare models to observations to consequently reduce this uncertainty. This study compares aircraft observations from the Arctic CLoud Observations Using airborne measurements during polar Day (ACLOUD) campaign around Svalbard, Norway, in May–June 2017 and simulations using the ICON (ICOsahedral Non-hydrostatic) model in its numerical weather prediction (NWP) setup at 1.2 km horizontal resolution. By comparing measurements of solar and terrestrial irradiances during ACLOUD flights to the respective properties in ICON, we showed that the model systematically overestimates the transmissivity of the mostly liquid clouds during the campaign. This model bias is traced back to the way cloud condensation nuclei (CCN) get activated into cloud droplets in the two-moment bulk microphysical scheme used in this study. This process is parameterized as a function of grid-scale vertical velocity in the microphysical scheme used, but in-cloud turbulence cannot be sufficiently resolved at 1.2 km horizontal resolution in Arctic clouds. By parameterizing subgrid-scale vertical motion as a function of turbulent kinetic energy, we are able to achieve a more realistic CCN activation into cloud droplets. Additionally, we showed that by scaling the presently used CCN activation profile, the hydrometeor number concentration could be modified to be in better agreement with ACLOUD observations in our revised CCN activation parameterization. This consequently results in an improved representation of cloud optical properties in our ICON simulations. |
|---|