Long‐range persistence in global Aerosol Index dynamics
Long‐range persistence in global Aerosol Index dynamics
复制标题
DOI：
10.1080/01431160600617236
复制
发表时间：
2006-08
影响因子：
3.4
通讯作者：
C. Varotsos;J. Ondov;A. Cracknell;M. Efstathiou;M. Assimakopoulos
中科院分区：
文献类型：
--
作者：
C. Varotsos;J. Ondov;A. Cracknell;M. Efstathiou;M. Assimakopoulos
将去趋势波动分析 (DFA) 应用于从 1979 年至 2003 年期间卫星观测得出的区域平均日气溶胶指数 (AI) 值，以寻找自相似性。结果表明，对于长于约 4 天且短于约 2 年的时间尺度，两个半球和全球的去趋势和去季节性的人工智能波动都遵循持续的长程幂律相关性。这表明小时间间隔内的 AI 波动与较长时间间隔内的 AI 波动以幂律方式相关（当时间间隔从约 4 天到约 2 年变化时）。换句话说，人工智能在一个时间范围内的异常现象会持续到下一个时间范围，表现出幂律演化。讨论了年度和半年周期对两个半球人工智能时间序列标度行为的影响。 AI时间序列中大约2年的时间尺度的一个合理机制可能是纬向风中赤道平流层准两年一次的振荡对布鲁尔-多布森环流的调制。天气尺度的气象系统可能产生大约 4 天的时间尺度。这些发现可能有助于测试现有模型的结果，应对现有模型进行检查以确定它们是否表现出上述的缩放行为。
Detrended fluctuation analysis (DFA) was applied to zonal mean daily Aerosol Index (AI) values derived from satellite observations during 1979–2003 to search for self‐similarity properties. The results show that the detrended and deseasonalized AI fluctuations in both hemispheres and globally obey persistent long‐range power‐law correlations for time scales longer than about 4 days and shorter than about 2 years. This suggests that the AI fluctuations in small time intervals are related to the AI fluctuations in longer time intervals in a power‐law fashion (when the time intervals vary from about 4 days to about 2 years). In other words, an anomaly in AI in one time frame continues into the next, exhibiting a power‐law evolution. The influence of the annual and semiannual cycles on the scaling behaviour of the AI time series in both hemispheres is discussed. A plausible mechanism for the time scale of about 2 years in AI time series could be the modulation of the Brewer–Dobson cell by the quasi‐biennial oscillation at the equatorial stratosphere in the zonal wind. The synoptic‐scale meteorological systems probably give rise to the time scale of about 4 days. These findings could prove useful in testing the results of existing models, which should be examined to determine if they demonstrate the scaling behaviour mentioned above.