The AMIP simulation by the UIUC (University of Illinois at Urbana-Champaign) 7-layer tropospheric general circulation model (GCM) did not reproduce the well-organized, intraseasonal oscillations which were observed in the tropics. Because this failure to simulate tropical intraseasonal oscillations is common to many GCMs, we began an investigation of its possible cause and correction. For this investigation we raised the top of our model from 200 mb to 50 mb, and added 4 layers. The resulting 11-layer lower-stratospheric/tropospheric GCM also did not simulate well-organized tropical intraseasonal oscillations. We then focussed our investigation with the 11-layer GCM on the parameterization of cumulus convection. Three different parameterization schemes are being tested: (1) the UIUC GCM's original cumulus-convection parameterization, which is a modified version of the Arakawa-Schubert (1974) parameterization, 1995: (2) the parameterization of Kuo (1974), 1995: and (3) the moist-convective-adjustment (MCA) parameterization of Manabe et el. (1965). The GCM simulations for (1) and (2) have been completed, while that for (3) is in progress.
With its original cumulus-convection parameterization, the 11-layer GCM does not simulate well-organized intraseasonal oscillations. However, when the parameterization is modified such that penetrative convection can occur only when the relative humidity at the top of the planetary boundary layer exceeds a threshold value, RHc, well-organized, eastward-propagating global-scale waves are simulated with periods of 31 to 35 days and amplitudes that increase with increasing RHc. A similar result is obtained for the Kuo parameterization, with the GCM simulating well-defined, eastward-propagating, global-scale waves with periods of 45 to 60 days when a threshold relative humidity is imposed for the onset of cumulus convection.
The possible physical reasons for this dependence of the simulated intraseasonal
oscillations on the threshold relative humidity for the onset of convection
will be discussed in the presentation.