E3SMv2, like various other global climate models that include representations of aerosol-cloud interactions, uses an empirically chosen lower bound on the simulated in-cloud cloud droplet number concentration (CDNC) to help constrain the effective radiative forcing of anthropogenic aerosols, ERFaer. This work uses sensitivity experiments combined with online conditional sampling and offline interactive data exploration to identify when and where ultra-low CDNCs are simulated by E3SMv2 and which of the ultra-low values have the strongest impact on ERFaer. Mass and number budget analyses are conducted to reveal the characteristics of the cloud droplet formation processes when impactful ultra-low CDNCs occur.

Our analyses and sensitivity experiments lead to two main conclusions: (1) Ultra-low CDNCs and strong ERFaer in E3SMv2 are related but also distinct issues. Attempts to address one issue might lead to degraded results in the other aspect. (2) While the Ghan and Abdul-Razzak (2000) aerosol activation parameterization might have room for improvement, it will be worth revisiting how that parameterization is used in E3SM, i.e., how it is connected with the other parameterizations in the model, in order to have a more physically realistic representation of droplet number formation.