1. Beijing Climate Center, China Meteorological Administration
The Madden-Julian Oscillation simulated by the Beijing Climate Center’s AGCM
Dong Min Wu Tongwen
Wang Zaizhi Zhang Fang
Beijing Climate Center, China Meteorological Administration
Beijing, PRC

2. Contents Introduction Model and data Results spectral structure
　　　MJO variance distribution
　　　Annual cycle of the MJO (Wavelet
analysis)
　　　Propagating speed
Discussion

3. Model and Data The Beijing Climate Center Model (BCC_AGCM2.0.1)
It is developed based on NCAR CAM3.0 (released on )
A refference Atmosphere is introduced to the model, The prognostic variable T and Ps become T´ and P´s
A new Zhang_Mcfarlane convection scheme (2005) replaced the original one (1995)
Latent heating calculation modified over ocean
Snow cover representation slightly changed
NCAR CAM3 (Collins et al. 2004)

4. Model run and validation data
　Monthly SST from 1949,9-2001,10 (52 years)
Validation data：
　NCEP( ),ECMWF( )
CMAP ( )
Variables：U 850 hPa; Precipitation

5. The spectral structure
The U850 and PP for each year (365 days) at each latitude are used to do time-space spectral analysis by FFT
Averaged over 10S-10N
Averaged over 51 year ( )

6. Comparison of time-space spectrum of U850

7. Comparison of time-space spectrum of precipitation

8. Comparison of time-space spectrum of precipitation (pentad data)

9. Variance of MJO
After the spectral Analysis
Take the Waves of wave number 1 – 5 and period days to reconstruct a MJO time series
Plot the variance of the MJO series in the tropics (20S-20N) for each season

10. Comparison: variance of U850 MJO；ECMWF

11. Variance of U850 MJO；NCEP

12. variance of U850 MJO；BCC2

13. Comparison:variance of U850 MJO;CAM3

14. Annual cycle of MJO for U850 ；ECMWF

15. Annual cycle of MJO for U850 ;NCEP

16. Annual cycle of MJO for U850 ;BCC2

17. Annual cycle of MJO for U850 ；CAM3

18. The propagating speed of MJO for U850； ECMWF & NCEP

19. The propagating speed of MJO for U850；models

20. Discussions It can reproduce the MJO variance space distribution
The BCC AGCM can simulate the MJO prety well. The ability of BCC model in simulating the
MJO is much improved than CAM3.0
It can reproduce the time-space structure of the U850 and precipitation In the tropics
It can reproduce the MJO variance space distribution
The simulated annual cycle of MJO is close to the observation.
The convection scheme is a key issue in the simulating the MJO

21. The problems need to be solved
The simulated MJO is not concentrate on the days. It slightly moved to high frequency
The simulated MJO space distribution has some difference with the observation.
The annual cycle of the simulated MJO also has some difference from the observation, especially, the simulated MJO is too strong in summer.
The simulated MJO has more fast propagating speed than the observation

22. Future Work
Analysis the mechanism of how can the convection scheme influence the MJO
The impact of the surface heating flux on the MJO simulation.

23. Thanks！