The role of the cloud–radiation interaction in the simulation of the Madden–Julian oscillation (MJO) is investigated. A special focus is on the enhancement of column-integrated diabatic heating due to the greenhouse effects of clouds and moisture in the region of anomalous convection. The degree of this enhancement, the greenhouse enhancement factor (GEF), is measured at different precipitation anomaly regimes as the negative ratio of anomalous outgoing longwave radiation to anomalous precipitation.
Observations show that the GEF varies significantly with precipitation anomaly and with the MJO cycle. The greenhouse enhancement is greater in weak precipitation anomaly regimes and its effectiveness decreases monotonically with increasing precipitation anomaly. The GEF also amplifies locally when convection is strengthened in association with the MJO, especially in the weak precipitation anomaly regime (,5 mm day21).
A robust statistical relationship is found among CMIP5 climate model simulations between the GEF and the MJO simulation fidelity. Models that simulate a stronger MJO also simulate a greater GEF, especially in the weak precipitation anomaly regime (,5 mm day21). Models with a greater GEF in the strong precipitation anomaly regime (.30 mm day21) represent a slightly slower MJO propagation speed. Many models that lack the MJO underestimate the GEF in general and in particular in the weak precipitation anomaly regime. The results herein highlight that the cloud–radiation interaction is a crucial process for climate models to correctly represent the MJO.