河西走廊沙尘暴频数的时序分形特征

Time-Series Fractal Characteristics of Sandstorm Frequency in Hexi Corridor

对1961—1990年河西走廊沙尘暴频次资料进行的分形研究发现,沙尘暴发生的时序分布具有分形特征,计算所得其相空间的关联维数为5 52,表明要恰当地描述河西地区沙尘暴变化特征,需构造至少6个状态变量的动力学系统;在分析了沙尘暴频数时间序列的Kolmogorov熵为0 18、平均可预测时间尺度为5～6a的基础上,运用R/S分析方法对1961—1990年中各时段沙尘暴的变化趋势进行了预测,预测结果与实际沙尘暴的变化趋势十分吻合,表明上述方法是进行河西地区沙尘暴时间序列分形研究的有效途径。

Many researches have been made about sandstorm based on correlative data, and the work mainly focused on the history datas arrangement and statistical analysis, but the inherent nonlinear trait of sandstorm is studied less.Based on the annual frequency data of sandstorm occurring from 1961-1990 in Hexi corridor, the timeseries phase space is reconstructed and Kolmogorov entropy is calculated. The results show that sandstorm timeseries has fractal characteristics and its correlative dimension is 552, which indicates that there should have at least 6 independent variables for describing sandstorm change characteristics and reconstructing a dynamic system. Kolmogorov entropy of sandstorm timeseries is 018, which indicates that sandstorm dynamic system is chaotic and the forecasting timescale is about 5～6 years.Based on R/S analysis of every sandstorm period from 1961-1990, all the periods have persistence change except the period of 1973-1978, and the results are quite consistent with actual changing trend of sandstorm timeseries. The method is effective for studying fractal characteristics of sandstorm timeseries in Hexi corridor.