MJO change with A1B global warming estimated by the 40-km ECHAM5

This study estimates MJO change under the A1B greenhouse gas emission scenario using the ECHAM5 AGCM whose coupled version (ECHAM5/MPI-OM) has simulated best MJO variance among fourteen CGCMs. The model has a horizontal resolution at T319 (about 40 km) and is forced by the monthly evolving SST derived from the ECHAM5/MPI-OM at a lower resolution of T63 (about 200 km). Two runs are carried out covering the last 21 years of the twentieth and twenty-first centuries. The NCEP/NCAR Reanalysis products and observed precipitation are used to validate the simulated MJO during the twentieth century, based on which the twenty-first century MJO change is compared and predicted. The validation indicates that the previously reported MJO variances in the T63 coupled version are reproduced by the 40-km ECHAM5. More aspects of MJO, such as the eastward propagation, structure, and dominant frequency and zonal wavenumber in power spectrum, are simulated reasonably well. The magnitude in power, however, is still low so that the signal is marginally detectable and embedded in the over-reddened background. Under the A1B scenario, the T63 ECHAM5/MPI-OM projected an over 3 K warmer tropical sea surface that forces the 40-km ECHAM to produce wetter tropics. The enhanced precipitation variance shows more spectral enhancement in background than in most wavebands. The zonal winds associated with MJO, however, are strengthened in the lower troposphere but weakened in the upper. On the one hand, the 850-hPa zonal wind has power nearly doubled in 30–60-days bands, demonstrating relatively clearer enhancement than the precipitation in MJO with the warming. A 1-tailed Student’s t test suggests that most of the MJO changes in variance and power spectra are statically significant. Subject to a 20–100-days band-pass filtering of that wind, an EOF analysis indicates that the two leading components in the twentieth-century run have a close structure to but smaller percentage of explained-to-total variance than those in observations; the A1B warming slightly increases the explained percentage and alters the structure. An MJO index formed by the two leading principal components discloses nearly doubling in the number of prominent MJO events with a peak phase occurring in February and March. A composite MJO life cycle of these events favors the frictional moisture convergence mechanism in maintaining the MJO and the nonlinear wind-induced surface heat exchange (WISHE) mechanism also appears in the A1B warming case. On the other hand, the Slingo index based on the 300-hPa zonal wind discloses that the upper-level P. Liu T. Li B. Wang International Pacific Research Center, SOEST, University of Hawaii at Manoa, Honolulu, HI, USA P. Liu (&) M. Zhang School of Marine and Atmospheric Sciences, Stony Brook University, 199 Endeavour Hall, SoMAS, Stony Brook, NY 11794-5000, USA e-mail: [email protected] J. Luo Y. Masumoto Research Institution for Global Change, Japan Agency for Marine-Earth Science and Technology, Kanagawa, Japan J. Luo Centre for Australian Weather and Climate Research, Bureau of Meteorology, Melbourne, Australia X. Wang The State Key Laboratory of Numerical Modeling for Atmospheric Sciences and Geophysical Fluid Dynamics (LASG), Beijing, China E. Roeckner Max Planck Institute for Meteorology, Hamburg, Germany