Contributions of Uncertainty in Droplet Nucleation to the Indirect Effect in Global Models

Anthropogenic aerosol perturbations to clouds and climate (the indirect effect, or AIE) contribute significant uncertainty towards understanding contemporary climate change. Despite refinements over the past two decades, modern global aerosol-climate models widely disagree on the magnitude of AIE, and wholly disagree with satellite estimates. Part of the spread in estimates of AIE arises from a lack of constraints on what exactly comprised the pre-industrial atmospheric aerosol burden, but another component is attributable to inter-model differences in simulating the chain of aerosol-cloud-precipitation processes which ultimately produce the indirect effect. Thus, one way to help constrain AIE is to thoroughly investigate the differences in aerosol-cloud processes and interactions occurring in these models.

We have configured one model, the CESM/MARC, with a suite of parameterizations affecting droplet activation. Each configuration produces similar climatologies with respect to precipitation and cloud macrophysics, but shows different sensitivies to aerosol perturbation - up to 1 W/m^2 differences in AIE. Regional differences in simulated aerosol-cloud interactions, especially in marine regions with little anthropogenic pollution, contribute to the spread in these AIE estimates. The baseline pre-industrial droplet number concentration in marine regions dominated by natural aerosol strongly predicts the magnitude of each model's AIE, suggesting that targeted observations of cloud microphysical properties across different cloud regimes and their sensitivity to aerosol influences could help provide firm constraints and targets for models.

Additionally, we have performed supplemental fully-coupled (atmosphere/ocean) simulations with each model configuration, allowing the model to relax to equilibrium following a change in aerosol emissions. These simulations allow us to assess the slower-timescale responses to aerosol perturbations. The spread in fast model responses (which produce the noted changes in indirect effect or forcing) gives rise to large differences in the equilibrium climate state of each configuration. We show that these changes in equilibrium climate state have implications for AIE estimates from model configurations tuned to the present-day climate.