Application of MJO Dynamics-Oriented Diagnostics to CMIP5 Models

Juan Li,Ying Yang,Zhiwei Zhu
DOI: https://doi.org/10.1007/s00704-020-03185-5
2020-01-01
Theoretical and Applied Climatology
Abstract:Realistic simulation of Madden–Julian Oscillation (MJO) propagation in coupled global climate models remains a common problem. In this study, the ability of 20 coupled models from Coupled Model Intercomparison Project Phase 5 (CMIP5) in simulating MJO is examined using dynamics-oriented diagnostics. The diagnostics focus on dynamic and thermodynamic structures of MJOs on three dimensions, which help to identify the shortcomings of models and evaluate whether they could reproduce MJO for the right reason. According to the simulation performance of the eastward propagation of MJO, the “good” models and “poor” models are detected. The dynamics-oriented diagnostics are further applied to the good models, poor models, and all models to establish a linkage between MJO simulation skill and their dynamic/thermodynamic structures. Results show that the simulations of good models have the following common features: (1) a horizontal zonal structural asymmetry in low-level zonal wind, upper-level diabatic heating and divergence; (2) a preceding eastward propagation of boundary layer moisture convergence; and (3) a rearward-tilted vertical structure of diabatic heating, equivalent potential temperature and available potential energy generation. The poor models that fail to capture these three-dimensional structures do not reproduce the eastward propagation of MJO. More than half of these 20 CMIP5 models still have difficulties in simulating these dynamic/thermodynamic structures.
What problem does this paper attempt to address?