What observations do we need to better constrain ERFaci in Earth system models?

The effective radiative forcing from aerosol-cloud interactions (ERFaci) remains one of the most uncertain aspects of our understanding of the Earth’s sensitivity to greenhouse gases. This uncertainty is largely due to the uncertainties in the parameterizations used to represent subgrid-scale processes in global Earth system models (ESMs). Perturbed parameter ensembles (PPEs), which vary the values of multiple parameters simultaneously, can help address this parametric uncertainty; however, ERFaci estimates are also impacted by structural uncertainties in the design choices of different ESMs. An accurate estimate of ERFaci also hinges on the availability of observations that we can use to assess the fidelity of ESMs in simulating plausible aerosol-cloud interactions.

 

Here, we focus on a PPE run in the E3SMv3 model, where 25 parameters related to aerosols and microphysics are perturbed across an ensemble of 250 nudged two-year simulations with both preindustrial and present-day aerosol forcings. In the E3SMv3 PPE, we examine which variables and which locations around the globe have the strongest correlations with the global ERFaci response to determine which measurements would be the most useful in constraining ERFaci. We also consider structural uncertainty in ERFaci by examining an opportunistic multi-model PPE consisting of a set of preindustrial and present-day PPE simulations run in different ESMs. Using the multi-model PPE, we identify the regions with the greatest disagreement between ESMs, which indicate where there are large structural uncertainties in the simulation of aerosol-cloud interactions. Together, these results highlight which regions and variables are subject to the greatest parametric and structural uncertainty related to simulating ERFaci. We argue that additional observations of key variables from these regions would have the greatest impact on reducing uncertainty in ERFaci and thus would help narrow our estimates of future projected temperature changes. This work provides a starting point for a new deployment planning framework using PPEs as part of an observing system simulation experiment (OSSE) to help improve our process understanding and simulation of aerosol-cloud interactions simultaneously.