A New Method on Time Series of Weather and Climate to Diagnose the Features of Extreme Value
诊断天气气候时间序列极值特征的一种新方法

Abstract A new diagnostic method tor the extreme value of time series is presented by using the cross theory of stationary stochstic process to extreme value of weather and climate sequences. The estimational formulas for occurrence frequency of extreme event, its persislcnl lime and interval lime are derived from the normality hypothesis in weather or climate series; especially, the relationship regarding the frequency of occurrence of extreme event and the composition of frequency spectrum in time series as well as several cases are presented. Results show that consistence of theory and observations is excellent. Thus, the above method can be applied to diagnosis extreme event of weather and climate, and the problems for forecasting extremes of climate need a further research in another paper.