中文摘要：

利用内蒙古锡林浩特国家气候观象台2010年3月至2011年2月全年的大气湍流观测资料,在数据质量控制的基础上,对内蒙古克氏针茅草原生态系统碳通量数据进行质量评价,并分析了日变化和季节变化特征。结果表明：在惯性副区,湍流通量的功率谱和协谱基本呈-2/3和-4/3的斜率变化。经过质量控制后的通量数据中,可用于基础研究的高质量数据约为74%,约有8%的低质量数据需要剔除。克氏针茅草原生长季中碳通量的日变化分为单峰和双峰两种类型,均有明显的不对称性,上午碳吸收强于下午。克氏针茅草原在冬季由于低温碳通量值很小,春季气温缓慢回升,草原处于早期发展阶段,表现为弱的碳汇,夏季的6月碳吸收达全年最强,7月和8月受干旱胁迫影响,碳通量逐渐减小,秋季草原开始枯黄,表现为弱的碳汇。内蒙古克氏针茅草原CO2年总量达-348 g CO2·m-2·a-1。克氏针茅草原7月夜间碳排放达最大值,6月白天碳吸收全年最强。本研究加深了对草原生态系统生长季和非生长季碳通量交换特征的理解,为陆面过程模型及相关碳模型参数修正提供了参考。

英文摘要：

Using turbulent data from Xilinhot National Climatological Observatory （Inner Mongo- lia） during March of 2010 and February of 2011, we analyzed the diurnal/seasonal variation of the net ecosystem carbon exchange （NEE） of a Stipa krylovii steppe, after data assessment and control. The results showed that the power spectrum and cospectrum in the inertial subrange could be expressed by the power law, and the power exponents were -2/3 and -4/3, respective- ly. 74% of the dataset could be classified as high-quality data for further basic research, while 8% should be removed. At the diurnal scale, the NEE in the steppe showed distinctly different temporal patterns of single and double peaks, with stronger uptake in the morning. At the season- al scale, the S. krylovii steppe was very low in carbon exchange during the cold winter, turned to be a weak sink when the temperature gradually rose in the spring, and reached its maximum in the June, the early summer. The steppe was constrained in carbon assimilation by the droughts in the months of July and August, and declined to be a weak sink when the steppe faded in the au- tumn. The annual cumulative carbon flux over the S. krylovii steppe in Inner Mongolia was -348 g CO2· m-2 · a-1. The net ecosystem exchange reached its maximum and minimum in June （daytime） and July （night）, respectively. This research provided an enhanced understanding of carbon exchanges during the growing and non-growing seasons in a temperate steppe, and would benefit the land surface process models and biogeochemical models on parameterizing and validation.