Here we present a survey of the spatial variability in different climate zones seen from AIRS data using the spectral EOF analysis. Over the tropical and subtropical oceans, the first principal component (PC1) is mostly due to the thermal contrast between surface and thick cold cloud tops. The second principal component (PC2) is mainly due to the spatial variation of the lower tropospheric humidity (LTH) and the low clouds. The signature of dust aerosol over the Arabian Sea and the Atlantic off the coast of North Africa in the summertime can be clearly seen in the PC2. Both the PC1 and the PC2 capture the upper tropospheric water vapor variability due to the forced orthogonality of EOFs. The third principal component (PC3) is mainly due to the spatial variation of the lower stratospheric temperature. Over the midlatitude oceans, the PC1 is still due to the thermal contrast of emission temperature. During wintertime, the PC2 is mainly due to stratospheric temperature variations. In the summer, the PC2 over the southern hemisphere is still due to stratospheric temperature variations, but in the northern hemisphere it is mainly due to the variations of the LTH and the low clouds. An exploratory study using synthetic spectra based on a NCAR CAM2 simulation shows that the model could account for the essential features in the data as well as provide an explanation of the three leading PCs. Major disagreements exist in the location of the ITCZ, the dust aerosol, and the lower stratospheric temperature.