A Convoluted Tale: How Air-Sea Interaction Improves the MJO Simulation in Three GCMs

The role of air-sea interaction in the simulation of the MJO has been the focus of modeling studies for over a decade. The problem is vexing because the atmosphere does not respond directly to sea surface temperature (SST), but to its effects on surface latent and sensible heat fluxes, which lag MJO convection. The fact that MJO convection leads maximum surface fluxes presents a challenge to understanding how the latter can impact the former. However, the majority of modeling studies demonstrate improved simulation of the MJO when an atmosphere-only GCM (AGCM) is coupled to an ocean model (CGCM).

In this study, we examine the role of coupling in three GCMs: CAM3, CAM4, and the "super-parameterized" SPCAM3. Each model is run as a CGCM, and as an AGCM using SSTs derived from SPCCSM3 (the coupled version of SPCAM3), which simulates a particularly realistic MJO. The sensitivity of latent and sensible heat fluxes to SST anomalies is largest in the coupled models, and smallest in AGCMs. Local SST variance explains only ~10% (<5%) of latent heat flux variance in coupled (uncoupled) simulations; the rest is explained by wind speed variance. Sensible heat flux variance is more sensitive to local SST variance, where 15~25% of the variance is explained in coupled simulations.

We turn to the moisture budget for insight into how coupling increases the wind speed variability associated with improved MJO simulation. When MJO convection is located in the eastern Indian Ocean, vertical profiles of total moistening and moistening by horizontal and vertical advection reveal that air-sea interaction enhances midlevel moistening by horizontal advection over the Maritime Continent. The apparent moisture sink, Q2, reveals that this elevated moisture advection anomaly is supplied by convective moistening that initiates at low levels in the eastern Pacific, and gradually deepens toward the west. Our analysis suggests that Indian Ocean air-sea interactions increase the strength of convection and the associated convectively-driven circulation anomalies, which ultimately drive the moistening processes further east. Finally, we present results suggesting that the location of horizontal advective moistening--east or west of MJO convection--is sensitive to model physics, and determines the direction of propagation.