Restricting 32–128 km horizontal scales hardly affects the MJO in the Superparameterized Community Atmosphere Model v.3.0 but the number of cloud-resolving grid columns constrains vertical mixing



The effects of artificially restricting the 32–128 km horizontal scale regime on MJO dynamics in the Superparameterized Community Atmosphere Model v.3.0 have been explored through reducing the extent of its embedded cloud resolving model (CRM) arrays. Two and four-fold reductions in CRM extent (from 128 to 64 km and 32 km) produce statistical composite MJO signatures with spatial scale, zonal phase speed, and intrinsic wind-convection anomaly structure that are all remarkably similar to the standard SPCAM's MJO. This suggests that the physics of mesoscale convective organization on 32–128 km scales are not critical to MJO dynamics in SPCAM and that reducing CRM extent may be a viable strategy for 400% more computationally efficient analysis of superparameterized MJO dynamics. However several unexpected basic state responses caution that extreme CRM domain reduction can lead to systematic mean state issues in superparameterized models. We hypothesize that an artificial limit on the efficiency of vertical updraft mixing is set by the number of grid columns available for compensating subsidence in the embedded CRM arrays. This can lead to reduced moisture ventilation supporting too much liquid cloud and thus an overly strong cloud shortwave radiative forcing, particularly in regions of deep convection.