中文摘要：

利用采自尕海-则岔自然保护区的岷江冷杉树轮样本,建立树轮晚材最大密度年表.通过相关分析发现,研究区岷江冷杉树轮晚材最大密度与暖季（5～8月）温度变化存在显著正相关关系.其中,与5～8月平均最高温度（1967～ 2008年）的相关最高（r=0.635,p＜0.001）.在此基础上,利用树轮晚材最大密度重建了藏东北玛曲地区过去284年来的5～8月平均最高温度变化,重建方程稳定可靠,重建方程的方差解释量为40.4％（R2adj=38.9％,N=42,F=27.09）.重建结果显示,在过去的284年中玛曲地区5～8月平均最高温度的异常高温年份为38年,异常低温年份为42年,且存在5个温度偏低时段（1725 ～ 1746年、1769～1819年、1834～ 1845年、1857 ～ 1869年和1965 ～ 1988年）和5个温度偏高时段（1747～ 1758年、1820～1833年、1846～1856年、1870～ 1964年和1989～2008年）.通过对比分析,发现玛曲温度重建序列与昌都温度重建序列具有良好的同步性,它们之间的第一主成分序列能够准确指示公元1725年以来青藏高原东部大范围暖季温度变化.在与海温数据的公共时段内（1854～ 2000年）,青藏高原东部两个温度序列的第一主成分序列和海温场的相关分析显示,北印度洋孟加拉湾和西太平洋中低纬度的海温对于青藏高原东部暖季温度变化影响显著;同时,青藏高原东部地区暖季温度与中东太平洋海温的负相关说明其变化还可能受到ENSO影响,研究还发现青藏高原东部地区暖季温度与印度夏季风存在关联.

英文摘要：

Understanding of past climatic change over the Tibetan Plateau is still limited because of the shortness of available meteorological records. However, high-resolution climate proxy records from the Tibetan plateau are scarce and of limited spatial representativeness. In this study we developed a maximum latewood density （MXD） chronology of Abiesfaxoniana from a sampling site （34°26′N, 102°43′E, site code ZC） of the northeastern Tibetan Plateau and used it to reconstruct the history of temperature variation for the region. Twenty-three trees were sampled. Two 10mm cores were taken from each tree with increment borers at breast height. In the correlation analysis, the local climate data （Maqu, 34°00′N, 102°05′E; 1967-2008） along with the MXD chronology was examined from the previous July to the current September. Response analysis shows that warm season （May-August） temperature is the main factor limiting the MXD variability of fir trees in the northeastern Tibetan Plateau. Based on growth-climate analyses, we reconstructed mean May-August maximum temperature during the past 254 years for the study area. The reconstruction explained 40. 4% of the instrumental temperature variance during the period 1967 - 2008 （F = 27.09, R^2adj = 38.9% ）. In the past 284 years, there were five cold periods （ 1725 - 1746, 1769- 1819, 1834- 1845, 1857- 1869 and 1965 - 1988） and five warm periods （1747- 1755, 1520-1833, 1846- 1556, 1570- 1964 and 1989-2008）, The values beyond the standard deviation （± 1 SD） indicate warm and cold years. The warm years accounted for 28.0% （35 warm years and 42 cold years） of the years during the whole reconstruction. There is a reasonable agreement with cold and warm periods previously estimated from tree-ring records from Changdu in the southeastern Tibetan Plateau. Correlations of the first principal component （PC1） of the two temperature sequences with May-August gridded temperature data （Climatic Research Unit, CRU） were calculated to investi