摘要
多尺度的气候变化具有显著的长期记忆性或持续性.本文利用分数阶微积分、功率谱、自相关函数等方面的知识,推导出导致气候变化长期记忆性的理论基础:气候变化显著的长期记忆性是其变化分数阶累积的结果,其变化的功率谱指数是分数阶积分阶数的二倍.最后利用去趋势涨落分析方法从实际的观测序列分析证实了这一关系的合理性.
The multi-scale climatic changes take the dominant long-term memory or long-term persistence. In this paper, based on the knowledge from fractional integral, power spectrum and self-correlation function, the mechanism resulted in this feature has been derived that the exponent of power spectrum is twice the order of fractional integral. Finally, this theoretical relation has been confirmed by the results from the observational temperature record analysis by the method of detrended fluctuation analysis.
出处
《中国科学：物理学、力学、天文学》
CSCD
北大核心
2013年第10期1327-1331,共5页
Scientia Sinica Physica,Mechanica & Astronomica
基金
国家自然科学基金资助项目(批准号:41175141)
关键词
长期记忆性
气候变化
分数阶积分
观测序列
long-term memory, climatic changes, fractional integral, observational confirmation
