Overly frequent light precipitation (<5 mm d-1), or drizzle, is a persistent problem in many climate models, including the ACME model. Errors in the intensity and frequency of precipitation have a large impact on cloud evolution and latent heat distribution within the atmosphere and therefore have a potentially important impact on a model’s climate and the general circulation. Past studies have noted that most of the drizzle in models is produced by convective parameterizations and have suggested that the frequent drizzle problem arises from a propensity of convection schemes to trigger too easily.

We characterize more precisely the conditions under which these convective drizzle events occur and investigate how a simple modification to the ACME v1 model, which alleviates convective drizzle, impacts the model’s climate. We find that drizzle occurs globally, but primarily in oceanic trade cumulus regions, where daily drizzle frequencies can exceed 90%. About 80% of this drizzle comes from the deep convective scheme. As expected, drizzle falls primarily from clouds in subsiding regions with modest precipitable water. With the modified simulations, we examine the dynamic and thermodynamic response of the climate to reduced drizzle to test whether suppressing convective drizzle leads to (a) increased cloud water path, (b) strengthened and improved intraseasonal variability in the tropics, and (c) an improvement in the double-ITCZ bias of the model by suppressing the early triggering of convection upstream from the observed regions of deep convection.