【作者】 宋涛；

【导师】 王盘兴； 王跃思；

【作者基本信息】 南京信息工程大学 ， 应用气象学， 2007， 博士

【摘要】 基于涡度相关技术，于2004～2006年连续3年对三江平原天然湿地、人工湿地（水稻）和旱地（大豆）生态系统CO₂、水汽和能量通量进行了观测。研究了涡度相关技术应用于三江生态系统通量长期观测中的理论问题和方法适用性，探讨了不同时间尺度上湿地生态系统碳交换的变异特征及其环境控制机制，阐述了气象要素变化对湿地生态系统碳交换的影响。通过比较天然湿地、人工湿地和旱地生态系统能量通量、水汽通量和CO₂通量的差异，初步分析了土地利用方式的变化对植被／大气间能量和物质交换的影响。功率谱和协谱分析表明，涡度相关仪器对高频湍流信号有着良好的响应能力能够满足观测的要求，与常规气象系统和气相色谱结果比较表明，涡度相关仪器在通量长期观测中性能稳定，可以满足长期观测客观需要，选取30min作为平均周期计算垂直湍流通量已基本满足实际需要。由于三江相对均一的下垫面条件，二次旋转和平面拟合方法校正通量与未校正通量间没有显著差异。稳态和方差相似性检验作为数据质量控制的方法表明，大多数低质量的数据集中在稳定边界层和边界层的交替的时刻。footprint模型分析指出，几乎100％的最大通量贡献区处于实验涉及的研究区域内。冠层CO₂存储项对夜间CO₂通量观测有着不可忽略的影响，当u*小于0.12 ms^-1时，冠层CO₂存储项占夜间估计的生态系统净交换的6％。与箱法观测的夜间生态系统净交换相比，涡度相关夜间观测结果都偏低4％～30％，并且低估的程度随季节变化而变化，在生长初期和晚期低估程度最小，而在生长旺期低估程度最大。湿地生态系统有着明显日和季节变化，生长季月平均CO₂净交换通量日变化幅度很大，从2004和2005年6，7，8月最大的夜间排放通量0.11，0.15，0.13 mg CO₂ m^-2s^-1到最大的白天净吸收通量-0.26，-0.43，-0.36 mg CO₂ m^-2s^-1。3年观测期最大的日CO₂净交换通量出现在2004年7月13日，为-8.0 g Cm^-2d^-1。白天生态系统碳交换主要受光合有效辐射控制，与高的叶面温度相联系的大的水汽压差使光合同化速率平均降低50％；夜间生态系统碳交换（夜间生态系统呼吸）主要受温度的控制，两者关系可以用简单的指数方程描述（R²范围在0.52～0.6），Q₁₀系数范围在1.9～3.3之间，而与地表水位没有明显的相关性（R²=0.04，P＞0.05）。考虑了植物叶面积指数的Wohlfahrt模型较简单指数方程更好的拟合了夜间通量（R²范围在0.6～0.75），可以用于插补夜间被剔除低质量数据。考虑生长季涡度相关和非生长季箱法观测的不确定性，3年间湿地生态系统净交换分别为-146±41.3，-44±21.9和5±14 g C m^-2a^-1，在碳汇和碳源汇近似平衡之间变化。与降水变化相对应的入射辐射的改变是影响不同年份生长季生态系统净交换量的主要因素。湿地开垦为水田增加了碳固定的能力，2005年生长季水稻累积生态系统净交换量为-35l gC m^-2，较湿地多固定218 gC m^-2。湿地开垦为旱田降低了碳固定能力，2005年生长季大豆累积生态系统净交换量仅为-61gC m^-2，这主要归咎于1)大豆地大的土壤呼吸释放；2)大豆植株高的氮素含量。更多还原

【Abstract】 Eddy covariance （EC） was used to measure the fluxes of carbon dioxide, water andenergy in different ecosystems during growing season from 2004 to 2006 in sanjiang Plain,China. Measurements were carried out at wetland, rice （reclaimed from wetland） andsoybean sites. We investigated the daily and seasonal variation of the net ecosystemexchange （NEE） and their response to environmental factors. The effects of climate changeon the inter-annual variation in growing season NEE of wetland was been examined. Theeffects of land use change from wetland to cropland on energy and mass fluxes was alsobeen discussed. The main results are as follows:Spectra analysis suggested that the instrument effects including the frequency responseof the sonic anemometer and of IRGA and sensor separation did not obviously damp thehigh-frequency fluctuations. Good agreements of air temperature and air vapor densitymeasured by EC with of that did by routine meteorological sensors, along with CO₂concentration measured by EC with that by chromatography, guarantee the satisfaction ofEC system to flux measurements. There is not significant difference of flux corrected by2D rotation and Planar fit method with the raw flux, due to the relative homogeneous andflat surface layer. As the methods of quality control, the application of stationary test andintegral turbulence test indicated that most low quality data were usually been found in theearly morning and later afternoon hours during which extremely non-stationary conditionswas observed. The estimated flux footprint showed that nearly100％Dmax （the largestcontribution to the measured flux per unit area） was located in the measuring plot wefocusing on. The night CO2 storage term （Fs） estimated by single height measurementshowed that Fs account for 6％of nighttime NEE on average and could not be ignoredwhen friction velocity was less than 0.12ms^-1. Comparison of nocturnal NEE measuredwith EC and dark chamber（DC） which was a independent measurement of CO₂ fluxshowed eddy covariance measurements were consistently lower 4％～30％than DC andthis underestimation was varied with the change of growing stage.The diurnal and seasonal variation of NEE in wetland ecosystem was obviously, withshift maximum mean monthly emission of 0.11, 0.15 and 0.13 mg CO₂ m^-2s^-1in nighttime to maximum uptake of-0.26,-0.43, -0.36 mg CO₂ m^-2s^-1in daytime from June toAugust in 2004 and 2005. The peak daily NEE of -8.0gCm^-2d^-1was observed at July 13 in2004. Daytime NEE were related to the photosynthetically active radiation （PAR）, a highvapour press deficit（≥15hPa） corresponding with high temperature was found to reducethe assimilation rate by on average 50％. Nighttime NEE（nighttime ecosystem respiration）were related to the temperature, which could be described by a simple exponentialfunction（R² varied from 0.52 to 0.6 in different growing seasons） and produced a relativehigh Q₁₀value of from 1.9 to 3.3. A significant relationship between night ecosystemrespiration and water table did not exist. Wohlfahrt model that took account of LAI couldexplain more variation of nighttime ecosystem respiration （R² varied from 0.6 to 0.75 indifferent growing seasons） than simple exponential function and could be used for gapfilling of low quality nighttime data. Combined with the daily NEE of 0.96 gCm^-2d^-1measured by chamber technique during the non growing season, the annual sums of NEEwas-146±41.3, -44±21.9和5±14g C m^-2a^-1, varied from carbon sink to neutral.The cumulative NEE of rice ecosystem was-351 gC m^-2during growing season of 2005,which reclaimed from wetland, 218 gC m^-2more than that of wetland, however, Thecumulative NEE of soybean ecosystem was only -61gC m^-2, 71 gC m^-2less than that ofwetland, it could be ascribed to 1) high soil respiration of soybean ecosystem and 2) high Ncontent of soybean plant.更多还原