Large‐Scale Precipitation Systems: Essential Elements of the Madden‐Julian Oscillation.

Possible roles of large‐scale precipitation on model simulations of the Madden‐Julian Oscillation (MJO) were investigated using the Large‐scale Precipitation Tracking (LPT) method. Individual LPT systems and eastward propagating MJO LPT systems were identified in both observations and 25 general circulation model (GCM) simulations. LPT systems show dominant eastward propagation over the Indian Ocean in the observation, whereas LPT systems are too stationary in GCM simulations. The MJO statistical signals in the model simulations are represented by the occurence frequencies of LPT systems with a size of ∼4–9 × 106 km2 (∼20–30° in longitude and latitude) and a strength of ∼15–18 mm d−1. Growth rate in size of LPT systems in early stage determines the size of LPT systems, which differentiates the MJO statistical signals among the GCM simulations. Comprehensive diagnostics of MJO simulations in GCMs should include both individual LPT systems as essential elements and the MJO statistical signals. Plain Language Summary: The Madden‐Julian Oscillation (MJO) is the dominant mode of intraseasonal variabilities in the tropics and represents a major source of subseasonal predictability. A distinct feature of the MJO is its eastward propagating large‐scale precipitating systems from the tropical Indian Ocean to the western‐central Pacific. However, this feature is not well‐reproduced by most general circulation models (GCMs). In this study, we explore possible reasons for why GCMs struggle to simulate the MJO eastward propagation by investing the Large‐scale Precipitation Tracking (LPT) systems. The results suggest that the problems of models in simulating the MJO are rooted in the LPT systems. The model capability of producing sufficient LPT systems that are large in size (∼4–9 × 106 km2) and strong in strength (∼15–18 mm d−1) is essential for simulating the MJO. It is particularly important for models to capture the growth rate in size of LPT systems within the first several days after their initiation. Key Points: Madden‐Julian Oscillation (MJO) statistical signals in general circulation model simulations depend on the size of large‐scale precipitation tracking (LPT) systemsGrowth rate in the size of LPT systems at early stage determines the model simulated MJO statistical signalsComprehensive diagnostics of MJO simulations should include LPT systems as essential elements of the MJO [ABSTRACT FROM AUTHOR]