Interactions of Large‐Scale Dynamics and Madden‐Julian Oscillation Propagation in Multi‐Model Simulations

Abstract

AbstractThe underrepresentation of the Madden‐Julian Oscillation (MJO) in climate models remains a challenge, limiting our ability to improve medium‐ to extended‐range atmospheric prediction. Motivated by recent work identifying the importance of the ratio of equatorial Rossby (ER) to Kelvin wave circulations in MJO propagation, this study examines MJO dynamics in 25 climate model simulations. We find that poor MJO models simulate anomalously large ER wave circulations to the west and small Kelvin and ER wave circulations to the east of the convective center. To quantify the role of circulation asymmetries in MJO propagation, we formulate a new west/east (W/E) zonal wind speed ratio. Our W/E ratio differs from other similar metrics in that it implicitly accounts for the ER wave gyres and it can be applied at all levels. Poor model ER wave biases are associated with excessive 700–1000‐hPa convergence, convection, and vertical moisture advection co‐located and west of the convective center while Kelvin and ER wave biases to the east are associated with a weaker dry anomaly and smaller horizontal moisture advection at 450–750‐hPa. Together, these biases help explain the stationary and slight westward MJO propagation in poor models. Space‐time spectral analyses of the zonal wind and precipitation confirm that good models produce realistic power, coherence, and phase for the MJO while poor models vastly underrepresent Kelvin waves and the MJO. Even though Kelvin waves are more realistic in good models, there are still model‐wide biases in circulation‐convection coupling for ER waves and Kelvin waves.