摘要
近年来分形理论在水文、股票及气象等行业中应用较多,在时间序列分析的非线性特征方面展示出强大魅力。采用重标极差方法分析西部地区6个站ETo时间序列的变化趋势和分形特征,通过对比发现：①所研究各站气象要素及ETo年变化趋势比较表明,所有站相对湿度均呈降低趋势,年积温均呈增加趋势,日照时数均呈降低趋势,年ETo值总体也呈降低趋势。②年尺度下各站点Hurst指数均大于0.5,同时分维数D均小于1.5,意味着ETo年序列在将来一段时间仍然保持与过去相一致的变化趋势,也即未来的总体趋势仍为减少趋势,各地区ETo年序列具有持续性。③各站点各月Hurst指数同样都大于0.5,分维数D均小于1.5,意味着1～12月ETo序列在将来一段时间仍然保持与过去相一致的变化趋势。此外,各站点年Hurst指数均大于月Hurst指数,表明年ETo序列比月ETo序列具有更长时间的持续性。
Fractal theory was wildly applied in many specialties,such as hydrology,stock and meteorology,in recent years,which shows great strengths in description of non-linear characteristics in time series analyses.In this paper,the rescaled range technique was used to analyze the variation trends and fractal characteristics of ETo time series for 6 stations located in west regions of China,the compared results showed：①according to the study on the annual change trends of climatic elements and ETo of each station,for all the stations,the relative humidity showed decrease tendency,the annual cumulative temperature showed increase tendency,the cumulative sunlight hours showed decrease tendency,and the annual ETo value in a year showed increase tendency.②on year scale,Hurst indexes were all larger than 0.5 for all the stations,meanwhile the fractal dimensions were less than 1.5 for all the stations,which indicated that ETo series would keep the consistent trends with the past years in some time of the future,the whole trends of ETo series would be decrease in the future,and ETo series in all the stations had continuity.③Hurst indexes of each month in all the stations were larger than 0.5,and the fractal dimensions were less than 1.5,which indicated that ETo series from January to December in a year would keep the consistent trends with the past in some time of the future.Moreover,yearly Hurst indexes for all the stations were larger than monthly Hurst indexes,which indicated that yearly ETo series had longer continuity than monthly ETo series.
出处
《节水灌溉》
北大核心
2008年第4期4-7,共4页
Water Saving Irrigation
基金
国家自然科学基金项目(50709028)
西北农林科技大学科研启动费
