Assessment of Madden–Julian Oscillation Simulations with Various Configurations of CESM

Xiaojing Li,Youmin Tang,Lei Zhou,Dake Chen,Zhixiong Yao,Siraj Ul Islam
DOI: https://doi.org/10.1007/s00382-016-2991-0
IF: 4.901
2016-01-01
Climate Dynamics
Abstract:This paper presents an assessment of the Madden–Julian oscillation (MJO) simulated in five experiments using the Community Earth System Model under different model settings. The analysis focused on the effects of air–sea coupling, resolution and atmospheric physics on the basic characteristics of the MJO, including intraseasonal variance, wavenumber-frequency characteristics and eastward propagation, using outgoing longwave radiation (OLR), zonal winds at 850 hPa (U850) and at 200 hPa (U200). Five experiments are conducted for this purpose including one atmospheric model—Community Atmosphere model version 4 (CAM4), two coupled models with CAM4 or Community Atmosphere model version 5 (CAM5) as the atmospheric component at a low resolution (CLP4_2d, CPL5_2d) and two the same coupled model with a high resolution (CPL4_1d and CLP5_1d). The results show that all models have better intraseasonal characteristics in U850 than in OLR. The uncoupled model CAM4 has lower fidelity than the coupled models in characterizing MJO basic features including the temporal and spatial intraseasonal variability and the eastward propagation. With ocean feedback, the coherence of convection and circulation is improved in the coupled models. The higher resolution is helpful in improving ISV spatial distribution and eliminating low frequency bias in the frequency-wavenumber spectra although it has little improvement to MJO-band variance (power) in frequency-wavenumber spectra. The new shallow convection scheme in CAM5 improves the moisture process of the lower troposphere so that CPL5_2d and CPL5_1d have more realistic eastward propagation speed in the boreal winter and better northward propagation in the boreal summer than other models. However, the strength of the convective MJO signal in CPL5_2d and CPL5_1d are weaker than other models and observations, which is probably one of the most spurious features in CPL5_2d and CPL5_1d experiments, suggesting that the CAM5 has a weaker convection activity than its predecessor.
What problem does this paper attempt to address?