The strength of the Asian-Pacific summer monsoon (APSM) is associ­ated with intensities from two major convective heat sources centered pri­marily over the Bay of Bengal (BB) and the western North Pacific (WNP). Because of the complexity of the APSM, especially over WNP, almost all of the general circulation models (GCMs) cannot simulate the climate fea­tures around the Asian monsoon region satisfactorily. This could be due to the inability of the GCMs to simulate correct cloud formations and radia­tive properties of clouds. In this study we will focus on the role of clouds and cloud radiative forcing (CRF) in the evolution of the APSM.

Our study relies on radiative fluxes at the top of the atmosphere from the Earth Radiation Budget Experiment (ERBE) and cloud data from the International Satellite Cloud Climatology Project (ISCCP-Dl) to inquire about some common properties of CRF in the four regions of East and Southern Asia: the BB, South China Sea (SCS), WNP, and Central China (CC).

During the APSM, we arrived at the following conclusions: After the onset of the summer monsoon, there is a lower OLR and a larger high cloud fraction with heavy and persistent rainfall in the BB throughout the whole summer. At the same time, extremely dense clouds (with thick optical depth) reaching into the higher atmosphere associated with strong convergence in the lower layers and divergences over the upper troposphere persistently occur in the BB. The strong dynamic force associated with closed spaced convective updrafts, as in BB, can result in a significant negative CRF (as low as -70 W m^-2). However, WNP and SCS are different from BB, not only in the CRF but also in the cloud's temporal and spatial distributions. A brief recess of deep convection occurs in mid-July in WNP and SCS, which causes the magnitude of precipitation, cloud amount, optical depth and CRF for the entire summer to average less than in BB. In both WNP and SCS, individual elements within the convective cloud system show a wide range of net radiative forcing (from the thick anvil clouds near convective cores to thin cirrus clouds near the edges of the anvil cloud) which produces a net CRF of near zero when averaged over the convective cloud system.

Through spectral analysis of high cloud amounts, we can find that over BB, the highest power is in the period around 75 days, which is the intra­seasonal timescale. While over WNP, the leading spectral peaks are at 10-20 days and 8-9 days. These two dominant frequencies are definitely re­lated to the fast annual cycle (proposed by LinHo and Wang). Failure to simulate correctly these two frequencies may be the primary reason why the GCMs exhibit great difficulty in reproducing correct precipitation in the WNP. The cloud structures and CRF in CC are very different from the oceanic regions in our analysis and where mid-low level clouds with thick optical depths play more of a role than does the high, thin cloud.