Annual cycle, quasi‐biennial oscillation, and southern oscillation in global precipitation

Abstract

The spatial and temporal variations in global precipitation associated with the annual cycle (AC), the tropospheric quasi‐biennial oscillation (QBO), and the Southern Oscillation (SO) are studied using 80 years (1901–1980) of global station rainfall data. The spatial scale and migration of the major climatological precipitation regimes are identified from the zeroth and first harmonic of the rainfall data. From the annual mean anomaly it was found that the dominant global precipitation pattern fluctuates irregularly, with 2‐ to 5‐year periods associated with the SO. This pattern is characterized by a center of action in the form of an east‐west dipole located over the equatorial central Pacific and the maritime continent of Indonesia. The occurrence of an El Niño/Southern Oscillation (ENSO) corresponds to an extreme phase of this dipole anomaly. Viewed in terms of the total rainfall amount (including the AC), the large rainfall anomalies that occurred during ENSOs represent major dislocations of the normal global rainfall pattern, affecting regions of the western Pacific/maritime continent and, to a lesser extent, China, Brazil, Argentina, southeast Africa, and southeast United States. Using the monthly rainfall anomaly, the same pattern again emerges as the dominant mode of variation. However, the temporal variation shows, in addition to the SO time scale, a distinct QBO signal. Using various time series analyses techniques, it was found that the QBO appears to be well correlated with the square of the SO signal. This result is also confirmed by bispectral analyses. The best correlation appears to occur during ENSO, with an apparent phase locking between the QBO and SO. A nonlinear interaction mechanism consistent with these results is illustrated by a conceptual model. The model suggests that the QBO may be an even more fundamental long‐term oscillation of the tropical ocean‐atmosphere system than the SO. ENSO‐like cycles can arise from a basic 2 year oscillation in the tropical ocean‐atmosphere system through its lagged effect from one year to the next. In addition, long‐term trends and amplitude modulation of the QBO can also arise from this effect. The role of the possible influence of the AC and other factors involved in this QBO–SO interaction is also discussed.