Because strong absorption of infrared radiation by greenhouse gases is more significant than the cloud longwave (LW) scattering effect, most climate models neglect cloud LW scattering to save computational costs. However, ignoring cloud LW scattering directly overestimates the outgoing longwave radiation (OLR). A recent study performed slab-ocean model simulations in the Community Earth System Model and showed that such radiative flux changes due to ice cloud LW scattering can affect the polar surface climate more than other climate zones. In this study, we included the same ice cloud LW scattering treatment in the Exascale Energy Earth System Model (E3SM) version 2 and ran fully-coupled simulations to assess the impact of ice cloud LW scattering on global climate simulation. Including ice cloud LW scattering leads to ~2 W m^-2 instantaneous OLR reduction in the tropics, more than the OLR reduction in other climate zones. Strong surface warming occurs in the Arctic, which is dominantly caused by the polar amplification resulting from the radiative forcing caused by ice cloud LW scattering. In the tropics, when the ice cloud LW scattering effect is included, more liquid clouds form in the middle troposphere, high clouds in the convection zone are lifted, anvil clouds retreat, and stratiform low cloud fraction increases. Most of these effects are similar to the cloud response to the increase of well-mixed greenhouse gases. The present study suggests that the ice cloud LW scattering effect must be incorporated into climate simulations.

Impacts of small-scale surface irregularities, or surface roughness, of atmospheric ice crystals on lidar backscattering properties are quantified. Geometric ice crystal models with various degrees of surface roughness and state-of-the-science light-scattering computational capabilities are used to simulate single-scattering properties across the entire practical size parameter range. The simulated bulk lidar and depolarization ratios of polydisperse ice crystals at 532 nm are strongly sensitive to the degree of surface roughness. Comparisons of these quantities between the theoretical simulations and counterparts inferred from spaceborne lidar observations for cold cirrus clouds suggest a typical surface roughness range of 0.03--0.15, which is most consistent with direct measurements of scanning electron microscopic images. The degree of surface roughness needs to be accounted for to properly interpret lidar backscattering observations of ice clouds.