Abstract
Significant relationships between atmospheric circulation anomalies in the Northern Hemisphere, in terms of 5-day mean 500 hPa geopotential height (Z500) and sea level pressure (SLP), and nonseasonal variations of 5-day rainfall (R5) over eastern China, where the influence of typical Asian monsoon prevails, are statistically examined by canonical correlation analysis (CCA) with time lags of 0 to a half year (36 Pentads). At a confidence level of 99. 99%, the CCA yields eight significant pairs of patterns that describe simultaneous responses of the precipitation field to anomalies in the middle troposphere. For the SLP-R5 fields, seven pairs of simultaneous canonical correlation patterns are found at the same confidence level. Most of these CCA pairs of patterns are in good agreement with well-known characteristic large scale patterns of circulation responsible for main rainstorms over China. From the physical point of view, these statistically significant correlation pairs of patterns are very reasonable. Possible synoptic processes associated with such relationships are figured out. In particular, the relationships represented by the first four robust concurrent CCA modes of the Z500-R5 fields, have been discussed in great detail and confirmed by a composite analysis. Also, significant (99.99%) canonical correlation pairs of patterns with lags up to 3 pentads for the Z500-R5 fields or the SLP-R5 fields are detected, presenting some antecedent circulation modes associated with certain rainstorms over China. Interestingly, some CCA pairs of patterns at different time-lags display the same or quite similar rainfall distributions over China, therefore exhibiting the evolutionary process of circulation systems responsible for the corresponding rainfall pattern. Those patterns are thus of crucial importance for long-range forecasting of Asian monsoon activity and hence of precipitation over China. Moreover, it is worth pointing out that some CCA patterns of sea level pressure (SLP) field for different time lags are significantly correlated with a rainfall distribution that is also robustly related with certain midtropospheric circulation regimes. These are indicative of the associated vertical structure of the circulation systems in question.
