Improved MJO‐simulation in ECHAM6.3 by coupling a Stochastic Multicloud Model to the convection scheme

Abstract

Abstract We implement a Stochastic Multicloud Model (SMCM) in an observation‐informed configuration into the convection scheme of the state‐of‐the‐art GCM ECHAM6.3. The SMCM configuration we use here has been tuned to represent observed tropical convection by associating the occurrence and strength of deep convection to midtropospheric vertical velocity and relative humidity. We show that compared to the ECHAM6.3 standard model, the SMCM‐modified version shows improved capacity to simulate features of tropical intraseasonal variability, including MJO‐like disturbances, without significantly distorting the mean model climate. This improvement goes in hand with ameliorated coupling of atmospheric convection to tropospheric moisture and spatiotemporal coherence of tropical convection compared to reanalysis and observations. We attribute these effects to (i) improved coupling of triggering and suppression of deep convective events to the model's large‐scale environment and (ii) the observations‐informed closure formulation which leads to an overall reduction of deep convective mass fluxes. Sensitivity tests show that while (ii) improves the convection‐moisture relationship, it is (i) which improves the spatiotemporal coherence of tropical rainfall and is important for MJO simulation. Further, the simulated spatiotemporal coherence of tropical rainfall is an intrinsic property of the convection schemes themselves and not of their parameters. We stress that this study serves as a proof‐of‐concept and motivates further efforts towards building a novel convection parameterization with the SMCM as a central element.

Keywords

• tropical convection
• Madden‐Julian‐Oscillation
• convection parametrization
• stochastic parametrization
• convective organization
• ECHAM6.3