中国近530年干湿变化及其持续性特征研究

Dry/wet changes and their durative characteristics in China during the past 530 years.

采用中国530年旱涝指数序列,并将其划分为华北和西北东部地区(Ⅰ区)、长江流域地区(Ⅱ区)以及中南和东南沿海地区(Ⅲ区)3个区域。应用功率谱、滤波方法、BG算法等研究旱涝指数序列各相对平稳均值段之间的干湿转化特征。结果表明,近130年的干旱时段和历史上的干旱或偏旱时段相比,Ⅰ区干湿转化频率有所加快;Ⅱ区干湿转化频率没有太大变化;Ⅲ区干湿转化频率有所降低。并且重大干湿转折时期大多对应突变点比较集中,即这一时期气候态不稳定,容易发生突变或各种极端气候事件。结合小波系数的周期分析结果发现,Ⅰ区从1920年左右开始的干旱,在经历了20世纪70年代末以来的严重干旱以后,有可能在21世纪再持续50到70年,其后再一次发生由干旱向湿润的转型;Ⅱ区则有可能在接下来的几十年中持续湿润期相对集中的情况。此外,Ⅰ—Ⅲ区干湿变化的特征与北半球的气候变化有一定的对应关系;Ⅰ区的干湿变化与当地温度变化具有较好的正相关;Ⅱ区和Ⅲ区干湿变化与温度变化的联系较Ⅰ区差一些;太平洋年代际涛动可能对3个区域的干湿转化均有不同程度的影响。在此基础上,定义旱(涝)尺度因子,可以定量描述旱涝持续性的区域特征;滑动计算旱(涝)尺度因子,可以检测哪一时段对应有旱涝群发性事件及重大干湿期的转折。

Based on a dry-wet （DW） index series of last 531 years at 48 stations, we divided the study area into three regions： Area Ⅰ covers North China and the east of Northwest China; Area Ⅱ covers most of the Yangtze River basin; Area Ⅲ covers South China and the Southeast coastal region. The evolutional characteristics of the mean value segments of the DW index series for each area were studied using the power spectrum analysis, the filtering method, and the Bernaola-Galvan algorithm. The results show that the frequency of DW transition in the drought-intensive period of the last 130 years, has increased in Area Ⅰ , not obviously changed in Area Ⅱ , but decreased in Area Ⅲ , in comparison with the other two earlier historical drought-intensive periods. In the dryness wetness transition periods, abrupt change points were relatively concentrated, i.e., the climate was unstable, and various climatic extreme events were likely to occur. Based on the period analysis results of the DW index, we found that the current dry epoch started at about 1920 in Area Ⅰ with severe droughts recorded from the late 1970s to the early 1980s, might last for about 50 70 years in this century, and then a DW shift will take place; while the current wet epoch in Area Ⅱ may also continue in the next several decades. Comparing DW variations of Area Ⅰ , Ⅱ , and Ⅲ with the Northern Hemispheric tree ring data, and the Pacific decadal oscillation index, we also found that DW variations in the three areas were to some extent associated with Northern Hemispheric climate change. DW variations in Area Ⅰ were positively correlated with temperature changes, but the correlations in Area Ⅱ and Ⅲ were relatively worse; and the Pacific deeadal oscillations might to different extent impact the DW shift in the three areas. It was found from the above analyses that the occurrence probability of various categories of dryness duration index is an exponential function of running window length. Dryness/wetness scale factor was defined as the reciprocal of the characteristic value of the exponential distribution, and its calculations show a band-like fluctuation distribution wherein the value of the factor increases northwards, with a mean value 1. 87/1.66 in Area Ⅰ , 1.62/1.54 in Area Ⅱ , and 1. 82/1.71 in Area Ⅲ . The dryness/wetness scale factor is able to quantitatively describe the persistent character of regional dryness/wetness, therefore it can be used to effectively detect the clustering of dry/wet events and the abrupt change of DW index.