基于信息熵的河南省年降水量时空变化研究

第30卷第11期K江科学院院报Vol.30 No. 112013年11月Journal of Yangtze River Scientific Research InstituteNov.2013DOI:10. 3969/j. issn. 1001 - 5485.2013. 11. 0042013 ,30(11):16-19基于信息熵的河南省年降水量时空变化研究王振亚'，吴德波“,朱余生(1.河南省气象台，郑州450003; 2.黄河水利委员会水文水资源局,郑州450003; 3.机械工业第六设计研究院有限公司,郑州450003; 4.中国气象局河南省农业气象保障与应用技术重点实验室,郑州450003)摘要:选取河南省1960- 2006 年共83个站年降水量,将信息熵理论应用在河南省年降水量时空变化研究上,计算每站的年降水量时间序列和年降水量与其极小值的相对量序列的熵估计值全省各年年降水量分布和年降水量与其多年平均降水量的相对值分布的熵估计值。计算结果表明:河南省年降水量的年际变化自西北东南方向增大,自南向北减小，自西向东增大,在西南东北方向则没有明显的规律;河南省年降水量相对值的年际变化自西南东北方向增大,在其他方向上则没有明显的规律;1966,1976,1999和2001年的河南省年降水量的空间分布较均匀,2000年的年降水量的空间分布不均匀;1986年的年降水量与其多年平均的相对值空间分布较均匀,2000年的年降水量与其多年平均的相对值空间分布不均匀。关键词:信息熵;年降水量;时空变化;不确定性中图分类号:TV125文献标志码:A文章编号:1001 - 5485 (2013)11 -0016-04降水量始终是气候变化研究的一个重要方量,逐站求取各站年降水量时间序列。面"。降水量时空分布的不确定性易形成旱涝灾1.2 研究理论和方法害,对水资源、工业、农业和生态环境等方面产生重物理学中,熵是热力学系统的某种状态函数,它大影响。C. E. Shannon把Boltzmann的概念引人信是对系统混乱度的度量,熵增使系统的混乱度增加,息论中,把熵作为-个随机事件的不确定性或信息熵减使系统的有序性增加'”。随着信息论的发展量的度量(2]。这种不确定性应用在年降水量的时与应用,信息熵问题得到了不少学者的积极研究,取间序列上可以研究降水量的年际变化规律,应用在得了大量的成果,其具有代表性的模型有Renyi的年降水量的空间分布上可以研究年降水在空间上分信息熵、模糊集合的信息熵、高艾期的有效信息熵。布是否均匀。张继国'31将信息熵的理论与方法应1948年,C.E.Shannon提出了基于离散型变量用在淮河流域降雨的时空分布研究上,取得较好的的信息熵。1994 年,N. Ebrahimi等(8)提出了熵估计研究结果。的2个改进式,可以由观测数据直接得到X的熵估河南省地处北亚热带向暖温带过渡的大陆性季计,并且证明了这2个估计式分别以概率收敛到风气候区内(4),境内地理条件复杂,降水量的水平H(X)。蒙特卡洛模拟显示这2个改进的估计式有较及垂直差异显著。--些学者基于统计学方法对河南小的偏差和平均平方误差[”)。本研究采用其中的省降水量的时空变化进行了相关研究[5-6),其主要一个改进式作为计算熵的估计值。基于定性描述,笔者将信息熵理论应用在河南省年设x,x,..x。是来自总体X的一组观测值，降水量时空变化研究方面,该方法将时空变化规律则所求熵的估计值为转化为对信息熵的定量计算,理论依据明确,计算方H.(m,n) = -SInYitm = Yi-m。(1)n台c:m/n法简单。其中1≤i≤m ;1资料和方 法C; =2m +1≤i≤n-m ;1.1 研究资料n-m+1≤i≤n。选取河南省1960- -2006 年共83个站日降水中国煤化工收稿日期:2012 -09 -18;修回日期:2012 -10-22作者简介:王振亚( 1981 -) ,男,河南正阳人,工程师,硕士研究生,主要从事水文气象方面..---_”信箱)huwzy1981.JYH@ 163. com。.第11期王振亚等基于信息熵的河南省年降水量时空变化研究19(10): 84 - 86. (in Chinese) )in Spring: An Analysis Based on GIS[J]. Joumal of Natu-[5]张红卫,陈怀亮. 基于地理信息的河南省3 ~4月降水时ral Disasters, 2012, 21(1): 170 - 173. (in Chinese)) .空分布及变化[J].安徽农业科学,009,37(24);11643夏军. 灰色系统水文学:理论方法及应用[ M].武汉:- 1644. (ZHANG Hong-wei, CHEN Huai liang. Tempo-华中理工大学出版社, 2000 (XIA Jun. Grey System Hy-ral- Spatial Distribution Analysis of Precipitation in Marchdrology: Theory, Methods and Applications[ M]. Wuhan:and April in Henan Area Supported by Geographic Infor-Huazhong University of Science and Technology Press,mation System [J]. Journal of Anhui Agricultural Sc2000. (in Chinese) )ences ,2009 ,37(24) :11643 - 1644. (in Chinese))[8] EBRAHIMI N, PFLUGHOEFT K, S0OFI E s. Two Meas-[6] 张红卫,陈怀亮,张 宏.河南省春节降水与温度变化的ures of Sample Entropy [J]. Statistics & Probability Let-时空分布-基于地理信息系统的分析[J].自然灾害学ters. 1994 ,20(3) :225 -234.报,2012,21(1): 170 - 173. (ZHANG Hong-wei, CHENHuai-liang, ZHANG Hong. Spatiotemporal Distribution of(编辑:赵卫兵)Precipitation and Temperature Change in Henan ProvinceResearch on Spatio-Temporal Variation of Annual Precipitation inHenan Province by Information EntropyWANG Zhen-ya'4 , WU De-bo2 ,ZHU Yu-sheng'(1. Henan Meteorological Observatory ，Zhengzhou 450003 ，China; 2. Hydrology Bureau of Yellow RiverConservancy Commission, Zhengzhou 450003, China; 3. SIPPR Engineering Group Co.，Ltd. ,Zhengzhou450003, China; 4. Henan Provincial Key Laboratory of Agrometeorological Ensuring and Applied Techniqueunder China Meteorological Administration, Zhenghou 450003 , China)Abstract :The temporal and spatial variation of annual precipitation in Henan province was researched based on thetheory of information entropy. The annual precipitation time series at 83 stations from 1960 to 2006 was selected.The estimated information entropy of the annual precipitation time series and the relative value of annual precipitati-on to its minimum at each station, as well as the estimated information entropy of the annual precipitation' s spatialdistribution and the relative value of annual precipitation to its multi-year average in each year , were calculated.Results show that the interannual variation of annual precipitation increases from northwest to southeast and fromwest to east, and reduces from south to north, but has no obvious regularity from southwest to northeast; while theinterannual variation of the relative value of annual precipitation to its minimum increases from southwest to north-east but has no obvious regularity in the other directions. The spatial distribution of annual precipitation was uni-form in 1966, 1976, 1999 and 2001 and nonuniform in 2000; the spatial distribution of the relative value of annualprecipitation to its multi-year average is uniform in 1986 and nonuniform in 2000.Key words : information entropy ; annual precipitation; temporal and spatial variation; uncertainty.....................................................央(上接第15页)tree -ring index sequence. The results show that there are three positive phase stages and four negative phase stagesin the low-frequency information of reconstruction sequence. Except for transitory positive phase in mid-1970s andaround 1990s , the overall trend of Ispo is negative phase from 1914 to 1948 , positive phase from 1948 to 1965, andgenerally negative phase since the mid-1960s. Comparative analysis shows that the reconstructed summer IApo se-quence is well synchronic with another summer IAPO reconstruction sequence.Key words: summer Asian-Pacific Oscillation index (Ipo); tree-ring中国煤化工is; XiaowutaiMountainMYHCNMHG