Large Aerosol Indirect Effects in Frontal Clouds in the North Atlantic Ocean

Clouds associated with mid-latitude cyclones encompass vast marine regions and likely play an important role in Earth's climate system. The effect of aerosols on marine cloud albedo is substantial and poorly quantified, and is thus a significant source of uncertainty in climate radiative forcing. We attempt to calculate the effect of aerosols on frontal clouds that pass over the Atmospheric Radiation Measurement (ARM) site at Graciosa Island in the North Atlantic during 50 distinct cyclone cases during 2017-2018. As well as surface measurements from this site, we use simulations with the UK Met Office Unified Model (UM), observations at cloud top from the SEVIRI satellite, and airborne measurements from the ACE-ENA flight campaign to try to better quantify how frontal clouds are influenced by aerosols.

We isolate frontal liquid clouds using a deep learning model which is trained on carefully selected SEVIRI satellite images using cloud optical depth and phase. We use simulations and observations from the UHSAS instrument at the ARM site to infer aerosol concentration at the cloud base, and satellite and surface remote sensing observations of cloud droplet concentrations. We then compare the Twomey effects in frontal and non-frontal clouds in observations and simulations.

We find that pristine frontal clouds in the North Atlantic are in some ways more sensitive to aerosol concentrations than non-frontal clouds. We expect aerosol effects on cloud fraction and cloud liquid water path to be muted in the strongly dynamically forced frontal clouds. However, the high updraft speeds in frontal clouds mean the albedo of these clouds is expected to be more sensitive to the number concentration of aerosols. We quantify these sensitivities, and we find Aitken-mode aerosols contribute up to ~40% of the total activated fraction in frontal clouds. Climate models generally simulate Aitken mode aerosol number concentrations inaccurately, and some widely used reanalyses consider only aerosol mass concentration as a prognostic variable. Our results suggest such models may not accurately represent how cloud albedo changes with changes in aerosol concentration, an important component of effective radiative forcing of climate. These results underscore the importance of aerosol activation diameter and Aitken-mode aerosols in aerosol-cloud interactions in frontal clouds, enhancing our understanding of their implications for climate.