The impact of mixed-phase cloud processes on radiative fluxes over the Southern Ocean in a convection-permitting model

Atmospheric and climate models have large biases in their short and long wave radiative fluxes over the Southern Ocean, leading to significant errors in their sea surface temperature, sea ice and large scale circulation. The primary cause for these biases is the representation of low-level clouds, both at the macro- and micro-scale. We assess the performance of a convection-permitting configuration of the Met Office Unified Model (MetUM) over the Southern Ocean using satellite and aircraft observations from the 2023 special observing period of the Southern Ocean Clouds (SOC) field experiment. We focus on the model’s sensitivity to the microphysics schemes. Firstly, the impact of ice nucleating particles (INP) parametrizations via sensitivity experiments using different temperature dependent INP distributions: (i) from Cooper (1986); (ii) as derived for the east Antarctic coast; and (iii) a new distribution derived from observations from the west Antarctic Peninsula during the SOC experiment. Secondly, we examine the impact of the parameterized overlap between ice and water within a grid box (the mixed-phase overlap factor), which modifies mixed-phase process rates, for example the Wegener–Bergeron–Findeisen process and riming.   Reducing the INP concentration to values observed over the Southern Ocean results in top of the atmosphere (TOA) radiative fluxes closer to observations. The lower INP concentrations result in lower ice water content and higher liquid water content, leading to brighter and more widespread cloud; this increases the albedo, resulting in a more accurate simulation of the TOA radiation. Equally, a large sensitivity in the top of the atmosphere fluxes is seen when changing the mixed-phase overlap factor. Decreasing (increasing) the mixed-phase overlap factor results in less (more) ice and more (less) liquid reducing the TOA fluxes. Decreasing the mixed-phase overlap factor results in TOA fluxes closer to the satellite observations. In summary, simulations using INP concentrations suitable for the Southern Ocean result in simulations closer to observed but other parametrizations in the microphysics scheme are equally important for accurate simulations of radiative fluxes.