Observation Based Constraint on Cloud-to-Precipitation Transition Deteriorates Aerosol-Cloud Interaction: Possibly a Common Problem among GCMs

The representation of cloud-to-precipitation transition process is one key source of bias that contribute largely to the uncertainty in predicted energy budget. In particular, many studies have corroborated one common bias in warm precipitation process among state-of-the-art GCMs: precipitation is triggered much too easily and too frequently relative to observations. In this study, we conduct a process-level constraint on warm precipitation process in the MIROC5.2 model, and examined the effect of this constraint on the aerosol-cloud interaction (ACI) effect between present-day and preindustrial periods. Results show that the MIROC5.2 model produces unrealistically large negative forcing due to the aerosol indirect effect when constrained with satellite observations for the warm precipitation process. This would result in a cooling that could cancel much of the greenhouse warming in the last century. Considering that precipitation may affect ACI through two pathways - the depletion of cloud water and wet-scavenging of aerosols - in an attempt to isolate the relative roles between the two pathways, a set of sensitivity tests with fixed cloud number concentrations for the wet-deposition of aerosols are conducted to turn off the feedback from aerosol-induced precipitation change on wet scavenging. The results show that the indirect forcing is less pronounced when fixed droplet number concentration is used for wet scavenging, suggesting that the ACI forcing is significantly amplified through coupling between the precipitation formation and the wet deposition. Namely, the following positive feedback is likely to operate: Aerosol-induced inhibition of rain formation causes less efficient wet deposition of aerosols, which induces larger increase in cloud droplet number concentration that further inhibits rain and causes pronounced increase in cloud water amount. This means that the “dichotomy” between the rain process constraint and the energy-balance requirement on ACI forcing is amplified through mutual coupling among aerosols, clouds and precipitation. Previous studies with the GFDL CM3 model exposed similar conclusion: the delaying of warm precipitation process (more realistic) caused a failure in reproducing the warming trend in the last century, due to overestimation of aerosol indirect effect. Given the common ‘too-easy, too-frequent’ bias of the precipitation process among major GCMs, the findings with MIROC5.2, together with those with GFDL CM3, have a broad implication for other GCMs as well: GCMs may share the common compensating biases between precipitation and aerosol-cloud interaction processes, underscoring the requirement of better constraint on key aerosol and cloud processes besides precipitation formation against observations in the future.