Performance of a Reconfigured Atmospheric General Circulation Model at Low Resolution

Abstract: Paleoclimate simulations usually require model runs over a very long time. The fast integration version of a state-of-the-art general circulation model (GCM), which shares the same physical and dynamical processes but with reduced horizontal resolution and increased time step, is usually developed. In this study, we configure a fast version of an atmospheric GCM (AGCM), the Grid Atmospheric Model of IAP/LASG (Institute of Atmospheric Physics/State Key Laboratory of Numerical Modeling for Atmospheric Sciences and Geophysical Fluid Dynamics), at low resolution (GAMIL-L, hereafter), and compare the simulation results with the NCEP/NCAR reanalysis and other data to examine its performance. GAMIL-L, which is derived from the original GAMIL, is a finite difference AGCM with 72×40 grids in longitude and latitude and 26 vertical levels. To validate the simulated climatology and variability, two runs were achieved. One was a 60-year control run with fixed climatological monthly sea surface temperature (SST) forcing, and the other was a 50-yr (1950--2000) integration with observational time-varying monthly SST forcing. Comparisons between these two cases and the reanalysis, including intra-seasonal and inter-annual variability are also presented. In addition, the differences between GAMIL-L and the original version of GAMIL are also investigated. The results show that GAMIL-L can capture most of the large-scale dynamical features of the atmosphere, especially in the tropics and mid latitudes, although a few deficiencies exist, such as the underestimated Hadley cell and thereby the weak strength of the Asia summer monsoon. However, the simulated mean states over high latitudes, especially over the polar regions, are not acceptable. Apart from dynamics, the thermodynamic features mainly depend upon the physical parameterization schemes. Since the physical package of GAMIL-L is exactly the same as the original high-resolution version of GAMIL, in which the NCAR Community Atmosphere Model (CAM2) physical package was used, there are only small differences between them in the precipitation and temperature fields. Because our goal is to develop a fast-running AGCM and employ it in the coupled climate system model of IAP/LASG for paleoclimate studies such as ENSO and Australia-Asia monsoon, particular attention has been paid to the model performances in the tropics. More model validations, such as those ran for the Southern Oscillation and South Asia monsoon, indicate that GAMIL-L is reasonably competent and valuable in this regard.

Keywords:
• GCM,
• GAMIL-L,
• atmosphere,
• model validation