【作者】 尹红；

【导师】 郭品文； 刘洪滨； Hans Linderholm；

【作者基本信息】 南京信息工程大学 ， 气象学， 2009， 博士

【摘要】 树木年轮资料以其定年准确、连续性强和分辨率高等优点,在历史时期的气候变化研究中得到了广泛的应用。对历史时期气候的重建多采用的是线性回归方法,这种方法假定树木生长和气候因子之间为线性关系,偏向对气候变化的定性分析,而实际上树木对气候的响应是一种物理的和生理的非线性关系。统计意义上的相关是否具有明确的生理意义,需要从机理方面分析。利用树木气候机理模型可以从树木生长机理方面入手,同时考虑多个气候因子,如:温度、降水、光照等来建立树木的生长过程,允许不同时期有不同的生长限制因子,从定量角度予以讨论。树木气候机理模型自建立起来得到了不断更新,其中TREE-RING模型是目前模拟树木生长比较完善的模型。本研究基于TREE-RING机理模型对小兴安岭红松树木生长对气候的响应进行了模拟分析,采用的树轮资料来自小兴安岭五营丰林国家自然保护区,模拟所用的气候资料为伊春气象站的逐日最高气温、最低气温和日降水量资料。通过对模型参数的调试,模拟结果的分析,辨别出影响树木生长的主要气候因子,并对主要气候因子进行了气候变化重建,本文得出以下主要结论:（1）TREE-RING模型对小兴安岭红松树木生长的模拟取得了较好的应用效果,生理意义明确。小兴安岭红松树木生长平均开始时间是在4月25日,结束时间是在10月9日生育期为167天;（2）小兴安岭红松树木的径向生长主要受到生长季上一年10月份气温和当年4月份气温的控制。模型分析结果认为,树木形成层开始生长平均是在4月下旬,这时水分充足,温度在光合最低温度和最适温度下限之间,温度愈高,光合速率愈大,储存的养料愈多,因此表现为与树轮宽度的正相关关系。形成层生长结束的时间平均在10月上旬,用于形成层细胞生长的光合产物的消耗减少,而光合速率随着温度的升高而增大,因此,10月份的气温越高为下一年储存的养料越多,翌年易形成宽轮;（3）对参数敏感性分析发现,光合最低温度,光合最适温度下限,最适土壤含水量下限和最大土壤含水量,是红松树木形成层生长比较敏感的参数,对这些参数的精确确定,是模型模拟准确与否的关键,气候变化的敏感性分析显示,1958—1998年若日平均温度升高1℃,树木生长开始时间会提前,生长量会降低。大气CO₂敏感性试验发现,模型对于CO₂变化敏感,模拟结果证实了CO₂浓度增加对植物生长具有助长作用,即CO₂施肥效应,CO₂对树木生长的趋势性变化产生影响。（4）基于树轮气候机理模型,对影响小兴安岭红松树木生长的主要气候因子10月份温度进行重建。重建了五营1796年以来的10月份平均气温变化,采用了逐一剔除法检验了重建序列的稳定性和可靠性,显示重建序列是可靠的,重建序列和实测序列相关系数为0.590,重建序列的方法解释量为34.8%。（5）重建序列的变化特征分析显示,在过去的209a中,五营重建气温变化大体经历了4个偏冷和4个偏暖阶段。通过与哈尔滨站实测资料、其它代用资料研究结果的对比验证可见,重建结果是可信的。18世纪末至19世纪30年代五营地区、长白山区、中国东部,乌鲁木齐河源同处于冷期,19世纪40年代末到60年代初五营地区、长白山区和乌鲁木齐河源同处于暖期,20世纪初到30年代末五营地区和乌鲁木齐河源同处于冷期。19世纪70年代初到19世纪末五营地区、长白山区、中国东部,乌鲁木齐河源同处于暖期。不同地区的冷暖发生时间同步,说明这些时间段冷暖期的发生并非局地现象,可代表一个较大的范围。（6）周期分析显示,五营温度变化存在3年、6年、20年和40年左右的周期特征,且不同时期的周期变化存在差异。突变分析显示,重建序列30年尺度的突变检验结果表明,1834年、1871年和1900年前后,五营地区10月份平均温度出现了显著的均值突变,1851年前后出现了明显的方差突变。重建序列15年尺度突变分析表明,在1822年、1847年、1858年、1871年、1900年和1987年前后发生了明显的均值突变。在1851年、1973年和1988年出现明显的方差突变。通过不同研究结果比较,发现了不同地区突变结果的一致性。更多还原

【Abstract】 Tree monitor environmental conditions that limit their biological processes and this information are stored in the structure of the annual ring.Tree-ring width data may reflect a nonlinear response to multivariate climate forcing.Single-variable paleoclimatic reconstruction via linear statistical calibration maybe represents a problem.Because there are nonclimatic influence on tree-ring records,including tree bilogy,size and age et al.Statistical relationships which may represent linearization of nonlinear processes are difficult to validate for long period processes for times outside the instrumental era.Moreover,traditional statistical techniques are inadequate when describing tree-growth responses to climate on cellular level.In this paper,we use the TREE-RING model to simulate climatic response of Pinus koraiensis in the lower mountains of XiaoXing’AnLing and discuss its hysiological response mechanism.Tree ring data were collected from Pinus koraiensis in the lower mountains of XiaoXing’AnLing.The climate data used in this study include local daily maximum temperature,minimum temperature and precipitation records from National Meteorological Information Center,China Meteorological Administration. Through model parameter debug and simulation result analysis,we can draw some main conclusions as follow:（1） A comparison between simulated and measured tree-ring series shows that the TREE-RING model could simulate the Pinus koraiensis tree-growth well.the correlation of cell number between the measured and modeled series equal to 0.62,and the value has exceed 99%degree of confidence.Pinus koraiensis grow starts average on April 25,cease on October 9,and the average growing time is 167 days.（2） The simulation results indicated that variations in tree-ring width are mainly controlled by October temperatures of the last year to growth and April temperatures of the growth year.Cambium growth begins in the last ten-day of April when there is enough water,but temperatures are low in XiaoXing’AnLing,and cambium growth ends in the first ten-day of October,but temperatures of October are between Lower optimal and minimum temperature for photosynthesis,where the photosynthetic rate increases at higher temperature.Thus,more food is stored at higher October temperature,and wider ring was easy to form the next year to grow. （3） Sensitivity analysis of parameter indicated that the tree-ring growth was sensitive to minimum temperature for photosynthesis,lower optimal temperature for photosynthesis,upper optimal volumetric water content and maximum volumetric water content.Sensitivity analysis of climate show tree growing start will bring forward and the increment will reduce if dailly temperature increases 1℃from 1958 to 1998.Sensitivity analysis of atmosphere CO₂ display the model is sensital to CO₂ variation.The result validate CO₂ contration increasing will accelerate tree grow.（4） Based on simulation record,the mean temperature of October was reconstructed for the period of 1796 to 2004 using RES tree ring chronology.The explained variance of reconstruction is 34.8%.In the past 209 year,there were 4 colder periods and 4 warmer periods in reconstructed series.（5） Through comparing with research result of different area in China,we found cold period concurrence in Wu ying,Changbai,East of China and Urumqi river source of Tianshan Mountain from 18 century end to 19 century 40’s.Warm period alse concurrence in there area from 19 century 70’s to 19 century end.（6） Periods of 3a,6a,20a and 40a were found markedly in the reconstruct series by power spectrum and wavelate analysis.The abrupt changes were also be detected with 30-year time scale and 15-year time scale by the smoothing T test,the smoothing F test and Lepage test methods.The significant abrupt changes of 30-year time scale mean value were present in around A.D.1871 and 1900 and the significant abrupt changes of standard deviation in around A.D.1851.The significant abrupt changes of 15-year time scale mean value were present in around A.D.1822,1847,1871,1900 and 1987,and the significant abrupt changes of standard deviation in around A.D.1851,1973 and 1988.更多还原