Variation of the Southern Oscillation (SO) is, though observed at low latitudes, an indication of anom- alies of the global atmospheric circulation. In this study, we found that the dryness/wetness in China is influen...Variation of the Southern Oscillation (SO) is, though observed at low latitudes, an indication of anom- alies of the global atmospheric circulation. In this study, we found that the dryness/wetness in China is influenced significantly by variation of the SO, and that monsoon rainfall, prior to ENSO maybe affect large scale atmosphere motion at lower latitudes. Since the dryness/wetness grades appear to be responsive to climate anomalies related to the SO, it should be possible to reconstruct past value of SO directly from grade indices. Thus we attempted to reconstruct the seasonal SO indices back to 1472 A. D. using canonical correlation technique. The best cstimates were derived from the model in which the period 1851--1962 is used for calibration and then is tested over an independent period 1963--1986. These estimates calibrate about 37 percent of the SO variance with about 17 percent of the variance verified. Significant peaks are apparent at period in 2.5--3, 5--6, 10.3--25 year range in the spectrum of the reconstructións of the SO which seems to have relationship with QBO and solar activity. It is encouraging that ENSO event in 1982-- 1983 is very well identified, especially in three seasons JJA, SON and DJF.展开更多
文摘Variation of the Southern Oscillation (SO) is, though observed at low latitudes, an indication of anom- alies of the global atmospheric circulation. In this study, we found that the dryness/wetness in China is influenced significantly by variation of the SO, and that monsoon rainfall, prior to ENSO maybe affect large scale atmosphere motion at lower latitudes. Since the dryness/wetness grades appear to be responsive to climate anomalies related to the SO, it should be possible to reconstruct past value of SO directly from grade indices. Thus we attempted to reconstruct the seasonal SO indices back to 1472 A. D. using canonical correlation technique. The best cstimates were derived from the model in which the period 1851--1962 is used for calibration and then is tested over an independent period 1963--1986. These estimates calibrate about 37 percent of the SO variance with about 17 percent of the variance verified. Significant peaks are apparent at period in 2.5--3, 5--6, 10.3--25 year range in the spectrum of the reconstructións of the SO which seems to have relationship with QBO and solar activity. It is encouraging that ENSO event in 1982-- 1983 is very well identified, especially in three seasons JJA, SON and DJF.
