| description abstract | The objective of this study is to examine the broad aspects of large-scale interannual and long-term variability in the monthly mean outgoing longwave radiation (OLR) data over the global tropics. These data, derived from NOAA's polar-orbiting satellites, cover a period of more than 15 years. Rotated principal component analysis (RPCA) has been performed on monthly OLR anomalies over the global tropics (30°N?30°S) on a 10° longitude by 5° latitude grid for the period from June 1974 through March 1989, excluding calendar year 1978. The leading rotated principal components to be discussed below have been tested for robustness and reproducibility. The spatial-loading pattern and the time series for the first principal component (termed the ?canonical ENSO? mode) represent the major large-scale features in the tropics during the typical phase of the major warm and cold events in the tropical Pacific during the analysis period. The characteristics of the dramatic 1982/83 warm event that were different from the canonical ENSO mode completely dominate the second RPC (termed the 1982/83 mode). The third and fourth leading RPCs appear to describe the changes in the satellite-observing system. Specifically, the third RPC is clearly associated with the different equator crossing times of the various NOAA satellites, while the fourth eigenmode may be related to the three major changes in the spectral windows of the different NOAA satellites. Of the six leading modes considered, the ?nonphysical? modes (3 and 4) accounted for more than 40% of the explained variance over North Africa and northeastern South America. The physical modes (1, 2, 5, and 6) explained more than 70% of the variance in the central equatorial and eastern Pacific Ocean. It is demonstrated that while the eigenmodes that result from unrotated principal component analysis are sensitive to small changes in analysis domain and period, those of the rotated analysis are fairly stable. However, note that the ?1982/83 mode,? as the name implies, is unique to the analysis period (1974?89). The results of the sensitivity analysis do not provide strong support of the claim by other authors that the decade of the 1980s, as compared to the 1970s, experienced enhanced levels of convective activity in the tropical Pacific and Indian oceans. | |
