The Madden-Julian Oscillation (MJO) is a convective disturbance that initiates over the Indian Ocean and propagates slowly eastward along the equator every one to three months. The MJO has significant impacts on global weather and climate extremes and is of interest to extended-range weather forecasting. However, many studies still find the MJO is underrepresented in general circulation models. Recent investigations of general circulation models have identified that the horizontal advection of seasonal moisture by the intraseasonal (one--three month) large scale circulation is critical to the simulation of coherent eastward MJO propagation.

This study looks into the anomalous circulations associated with the November--April MJO in 20-year climate simulations for over 20 climate models using a new, precipitation-based MJO index. One of the main ingredients separating good and poor MJO Task Force (MJOTF) models is their representation of the large scale westerly vs. easterly wind amplitude at many atmospheric pressure levels, especially below 500 mb. A new climate model metric is devised called the westerly/easterly ratio, which is strongly correlated with model MJO propagation skill and straightforward to calculate as it 1) does not involve spectral filtering and 2) is a function of pressure, allowing for a better understanding of what levels should be targeted for improving model MJO dynamics.

Furthermore, a space-time spectral analysis further clarifies many of the shortcomings of the precipitation and low-level zonal winds for the MJOTF models. For the poor models, the low-level westerly wind biases are associated with a biased equatorial Rossby wave filtered response while the low-level easterly wind biases are associated with a weak MJO filtered response.