树木年轮记录的祁连山东段公元1895年来的气温变化

利用祁连山东段吐鲁沟的油松树轮宽度序列，与气象要素进行了相关分析，并重建了公元1895年来 1～8月平均气温，转换方程方差解释量52.6%(R2adj＝48.7%，F＝13.589, p<0.001)。重建结果表明研究区近120年来经历了19世纪末到20世纪20年代、20世纪20年代到40年代末、20世纪 40～70年代3个明显的升降温过程和20世纪70年代以来的升温过程。重建序列与研究区以西的祁连山中段、天山西段、阿尔泰山中段温度的变化基本一致； 重建序列显示的 1923～1929年的高温时段与中国20世纪20年代大范围的干旱事件一致，而且与北大西洋涛动、北极涛动的变化遥相呼应，说明祁连山东段的树轮资料不仅记录了区域气候变化信号，而且对于全球变化响应敏感。

RECONSTRUCTION OF MEAN TEMPERATURES OF JANUARY TO AUGUST SINCE A.D.1895 BASED ON TREE-RING DATA IN THE EASTERN PART OF THE QILIAN MOUNTAINS

The tree­ring samples of Pinus tabulaeformis studied in the paper were from Tulugou(36 °41′N,102°43′E; 2400～2500m a．s.l.) of the eastern part of the Qilianshan Mountains. We collected 17 trees by growth cone and two cores per tree were selected randomly for each plot.
According to the international standard，the tree­ring samples were dried，pasted，polished and measured with LINTAB with resolution of 0.01mm. Using programme ARSTAN，STD chronology was finally set up by adopting the negative exponential function fitting the tree growth trend and by eradicating the growth trend generated by genetic factors and the interference between trees.
The monthly precipitation(1951～2005)and the monthly mean temperature(1951～2007)were adopted from the Wushaoling Meteorological Station(37°12′N,102°52′E; 3045m a．s.l.). The relationship between the index series of tree­ring width and climatic elements was analyzed. The results show that the tree­ring width is more sensitive to temperature than precipitation and the tree­ring width index is negatively correlative to monthly temperature，but significantly correlative to the temperature of January，June and August. Moreover，the tree­ring width is correlative significantly to the annual average temperature and the mean temperature of January to August，with its correlation coefficient to the mean temperature of January to August to reach 99% confidence.
A transformation equation was developed between the tree­ring width and the mean temperature of January to August. The multiple correlation coefficient of this equation is R＝0.725; the explained variance 52.6% and after adjusting,48.7%; F is 13.589，greatly exceeding 0.001 confidence.
The Csss of subsample signal strength can be used to evaluate the reliable length of chronology. We know that when Csss≥0.85，the reliable reconstruction interval is from 1895 to 2007.According to the transformation equation，the mean temperature of January to August was reconstructed within 120 years in the eastern part of the Qilianshan Mountains.
The reconstruction results show such temperature rising and declining times as 1890s～1920s,1920s～1940s and 1940s～1970s and a temperature rising time since 1970s. The climate change in the eastern part of the Qilianshan Mountains was corresponding to that in the middle part of the Qilianshan Mountains，in the western part of Tianshan and in the middle part of Altai. The high temperature period in 1923～1929 was corresponding to the serious and large scale drought happened in 1920s. Moreover，the reconstruction results are closely to NAOI and AOI. All these show that the tree­ring in the eastern part of the Qilianshan Mountains should record the signals of regional climate change and be sensitive to the global change.