Nonlinear Causal Discovery for Disentangling Drivers and Their Interactions in Marine Boundary Layer Clouds

We have known for many years that aerosols play an important role in warm cloud evolution and cloud radiative effects, but directly quantifying their influence has been hampered by a lack of understanding in the interactions between aerosols, cloud microphysics, meteorology, and cloud morphology. Disentangling the role and importance of these many different drivers has proven to be extremely difficult, especially because their strong interactions mask important processes, and strong feedbacks accentuate otherwise minor contributors.

To address this, we use a fully nonlinear causal framework which lets us attribute causal strength to individual drivers as well as to their interactions. Thus, not only can we identify important drivers, but we can also identify and qualify important interactions, such as competition, buffering, mediation, and cooperation. We have applied this framework to high-resolution time series of surface-based direct and remote sensing observations from the East North Atlantic (ENA) Atmospheric Radiation Measurement (ARM) site, contrasting precipitating and non-precipitating strato-cumulus clouds. We will discuss the importance of known processes, such as the Twomey effect, and newly identified processes and their interactions, and their dependence on cloud state and aerosol characteristics.