利用树木车轮资料重建西藏中部过去气候的初步尝试

A PRELIMINARY STUDY ON RECONSTRUCTING PASTCLIMATE IN THE MIDDLE XIZANG PLATEAU BYUSING TREE-RING DATA

为较可靠地利用年轮气候学方法重建过去气候，本文依据西藏中部，四个经过合适取样和精确定年的树木年轮年表，首先确立了它们各自生长状况对气候要素的响应函数，进一步明确年轮宽度与气候变化的关系。接着，选择可被重建的气候因子，并建立包括前期生长在内的、经过正交变换的转换函数，达到重建过去气温和降水的目的。此外，还划出不同的校准期和验证期，采用统计方法和其它类型代用资料进行检验。

: Based on four tree-ring chronologies which have been analysed with appropriate collec- tion and accurate dating in four sites of the middle Xizang plateau, an essential process on re- construction of past climate has been pointed out in this paper. First, the response function of each dendrochronology has been built and used to estirnate how ring-width growth res- ponds to variations in monthly climatic change. The response values to air temperature and rainfall appear significant and reliable results, and the multiple correlation coefficients of the response regression for three dendrochronologies are more than 0.85. Second, the climate fac- tors which could be reconstructed with different tree--ring series have been se1ected. The four climate predicands are precipitation in growing season, temperature in Spring, annual tem- peraturc and annual precipitation, respectively, for four tree-ring series. Then, the transfer function equation, including a new set of orthogonal variables, can be used to reconstruct local past temperature or precipitation. The orthogonal translation is adopted by Emperical Or- thogonal Function (EOF). It should be emphasized that prior growth has been considered in the relationship between climate factor and tree-ring chronology. That means the prior growth, the lag one through three years, has been put into the final f regression equation as indepen- dent factors. Besides, some different periods for calibration,and verification have been divi- ded. Some statistics and other kind of proxy data have been adopted as test approaches. Al1 statistical and other--data test has proved that the reconstruction of past climate using tree-ring data in the middle Xizang plateau is reasonable and reliable. The basic procedure could be applied to dendroclimatic analysis in other areas.