科尔沁沙地1826年以来榆树树轮宽度年表的建立及降水重建

ESTABLISHMENT OF RING WIDTH CHRONOLOGY AND RAINFALL RECONSTRUCTION OF Ulmus pumila L. IN HORQIN SANDY LAND SINCE 1826

利用10个采样点230个树木年轮样本,通过树轮分析程序,建立了科尔沁沙地区域1826～2008年共计183年长度的榆树树轮宽度年表。通过分析发现,水热条件与科尔沁沙地榆树生长关系密切。以标准年表及年表t+1年和t+2年序列重建了区域1826～2006年年降水。从过去181年来看,年降水重建序列具有2年、12年、18年和30年等变化准周期,分别经历了5个偏湿和5个偏于阶段,干旱和偏干年份分别占总年份的0.55%和16.02%,1950～2006年间偏干年数占1900～2006年间偏干总年数的78.57%,偏干阶段年数不断增加。降水出现了4次突变,1839年和1920年前后均发生了由少到多的突变,1861年和1958年前后均发生了由多到少的突变,这其中尤以1958年前后突变最为显著。总体上,科尔沁沙地近181年中,1826～1957年近130年间降水变化较为平缓,而后近50年,降水量趋于减少,气候趋于干旱。

The current study was done on Ulmus pumila L. in Horqin sandy land （42°36＇ -45°20＇N, 117°45＇ - 124°06＇E）. Samples were collected according to the principles of ecological environment, perception and restoration. Two hundred and thirty U. pumila L. tree ring samples were collected using the tree growth cones from 10 sites （ KLDD, MLML, KLSS, HQHST, ZZMLQ, KLWD, XAHY, ZQBGTL, KLMX, and BLYYC） in 2009. The pretreatment and dating were carried out according to the analytical procedure for tree rings. The widths of U. pumila L. tree rings were measured and the cross-dating was performed through the COFECHA procedure. The U. pumila L. growth trend was fitted to a negative exponential function or spline calculated using the ARSTAN procedure（the time step was assumed as 50 years）. The growth trend of U. purnila L. rings was removed, and all samples were used to establish a regional and comprehensive width chronology. The standard,residual,and self-regression chronologies of 183 years （1826 to 2008 ）were obtained. The results indicate that the chronologies had good characteristic parameters and included a large amount of environmental information. The U. pumila L. trees in Horqin sandy land were appropriate for the tree ring studies, and the constructed chronologies were reasonable and available. Using the average sequences of climatic factors observed at 8 climatic to gauge the regional climate factor sequences,they were mostly closed to the 10 sampling sites. The correlation coefficients for the standard chronology and the July, August, September, and annual rainfall were 0. 576, 0. 589, 0. 561, and 0. 623, respectively. Coefficients for the standard chronologies of t＋ 1, t＋2 years and annual rainfall were 0. 543 and 0. 535, respectively. All results reached the 0. 001 confidence level. Moreover, the correlation coefficients between the standard chronology and the June, July, August and September air temperature were - 0. 442, - 0. 433, - 0. 440, and -0. 545, respectively, at 0. 001 confidence level. The growth of U. purnila L. had relatively strong dependence on the July, August, September, and annual rainfall. In addition, the rainfall of one and two years before the primary year also had effects on the number of U. pumila L. However, the average air temperature presented an inhibitory effect on Ulrnus pumila L. The series of annual rainfall from 1826 to 2006 was reconstructed using the standard chronology, t＋ 1- and t＋2-year series. The testing parameters indicate that the equation for the model was reliable and credible. The reconstructed series had 2-, 12-, 18-, and 30-year quasi-periodic changes,which underwent 5 partially wet and 5 partially dry stages. The wet and dry stages accounted for 0.55% and 16.02% of the total years（ 183 years）. The dry years 1950 to 2008 accounted for 78.57% in which from 1900 to 2006. The number of partial dry years increased continuously. Four rainfall mutations occurred during this period. Mutations of increased rainfall occurred during 1839 and 1920,and those of decreased rainfall occurred during 1861 and 1958; a peak mutation occurred around 1958. Generally, the rainfall in Horqin sandy land changed relatively gentle from 1826 to 1957, and presented a decreasing trend during the recent 50 years. These indicate that the climate tended to be dry. This study provides a reference scale for further tree ring dating studies in Horqin sandy land for climate monitoring.