Abstract
Various aspects of the intraseasonal variability (ISV) of tropical convection are studied using outgoing longwave radiation (OLR) data. From empirical orthogonal function analysis, it was found that the most dominant mode of the ISV of tropical convection consists of an east-west oriented dipole that propagates eastward at a speed of approximately 4–5 m s−1 over the equatorial Indian/Western Pacific Ocean but stalls and intensifies as it reaches cast of Indonesia and the equatorial central Pacific. When one center of the dipole is established over Indonesia, the associated anomalous convection extends from the equator southeastward over Northern Australia to the Southern Hemisphere midlatitudes. The oscillation covers a wide range of periods ranging from 30–60 days and is likely to affect strongly the transient fluctuation of the Pacific Walker circulation and the South Pacific Convergence Zone. The foregoing features are identified with the so-called “40–50” day oscillation in the tropics. Distinct patterns of extratropical OLR anomalies are found to evolve coherently with this “40–50” day oscillation. When the dipole centers are situated over the Indian Ocean and the equatorial western Pacific, major anomalous convection centers are found over the Asian sector. When the centers are shifted about 3000–4000 km (about one-quarter wavelength) eastward to Indonesia and the equatorial central Pacific, the most pronounced extratropical feature is found over the eastern North Pacific. These features reflect possible connection between tropical and extratropical convection/circulation systems. Because of the seesaw nature in the variation of a tropical heat source/sink, we suggest that extratropical anomalies arising from tropical forcing are associated with changes in the overall tropical beating distribution rather than just from a local source. We also present some evidence of a plausible connection between the “40–50” day mode and the monthly and interannual variability in tropical convection. It is shown that large fluctuations in the monthly averages of OLR over the western and central Pacific are mainly due to the modulation by the “40–50” day oscillation over these regions. A possible relationship between this oscillation and the El Niñ/Southern Oscillation is discussed.