In the dry season, the model successfully simulates the stratus cloud deck over the southeastern Pacific off South America, the surface wind field as well as precipitation when compared with observations (TRMM cloud liquid water, QuikSCAT surface wind and CMAP precipitation). A sensitivity experiment was carried out by removing the Andean mountain range. Comparison of the control with the No-Andes experiment indicates that the presence of the mountains substantially enhance cloud deck over the southeastern Pacific. It is found that the presence of the Andes blocks low-level warm easterly winds from South America, leading to the development of divergence and an enhanced temperature inversion off South America that favor the formation of stratus clouds. In the wet season of southeastern Pacific, on the other hand, the presence of the Andes not only changes the precipitation over the South America but also affects the precipitation in the eastern Pacific ITCZ. The Andes act to enhance the northern ITCZ and weaken the southern one. This suggests that the poor representation of the steep and narrow Andes in current global GCMs is a likely cause of persistent double ITCZ in their simulation.