艾比湖荒漠-湿地生态系统非生长季碳通量数据特征

选取新疆艾比湖湿地国家级自然保护区非生长季(2012年1月15日—3月15日、11月1日—12月14日、2013年11月1日—11月16日)通量观测数据,参考中国通量数据处理标准开展荒漠-湿地生态系统非生长季碳通量数据特征研究。结果表明:研究区非生长季碳通量观测数据存在大量"野点",占数据总量的37.39%,有效通量变化范围为-0.197—0.283 mg m-2s-1;平面拟合旋转(PF)校正具有较好的结果(R2=0.9349,P0.01),频率响应校正引起的碳通量增量为7.55%,水热校正影响较小;碳通量数据在不同质量等级分布较平均,大气湍流发展的充分性处于中等水平,夜间摩擦风速可划分为3级,在0 m/su*0.30 m/s内进行选择性剔除;传感器状态异常、检验分析和阈值分析剔除数据比例分别为26.34%、2.48%和8.57%;碳通量与5 cm土壤温度和太阳辐射不存在显著相关,采用线性内插和平均每日变异法可实现缺失数据的插补。

Research on data characteristics during non-growing season of desert-wetland ecosystem in Ebinur Lake

With the global climate change and the resulting frequent regional natural disasters becoming increasingly real, CO₂, H₂O, and energy transport and transformation processes of the surface atmosphere are becoming important concerns for the international community. Carbon flux in the non-growing season (Jan. 15 to Mar. 15, Nov. 1 to Dec. 14 in 2012; Nov. 1 to 16 in 2013) was measured in the Ebinur Lake Wetland National Nature Reserve of Xinjiang using the eddy covariance technique. Then, research was conducted on the carbon flux data processing method after consulting the flux data processing standards in China, and the data were further analyzed to understand the carbon flux data characteristics in the non-growing season, and to attempt to reveal the controlling factors and change law. The results indicated that there were a large number of "wild" non-growing season carbon flux observation data in the study area, which accounted for 37.39% of the total. The effective carbon flux was in the range from -0.197 mg m^-2 s^-1 to 0.283 mg m^-2 s^-1, and the average atmospheric CO₂ concentration and carbon flux were 766.905 mg/m³ and 0.018 mg m^-2 s^-1, respectively, which showed weak carbon sources. Axis-rotation correction results showed TR < DR < PF, and the planer fit (PF) correction showed better accuracy (R² = 0.9349, P < 0.01). The frequency-response correction increased carbon flux by 7.55%, but the carbon flux was less affected by the hydrothermal correction. The energy closure throughout the non-growing season in the study area was 0.62, indicating that 38% of the energy is not closed in the non-growing season. The proportion of IST classification results of carbon flux data in the desert wetland in quality levels 1-6 was 92.03%, indicating that the distribution in different quality grades of carbon flux data in the desert-wetland ecosystem was relatively uniform. The percentage of ITC classification results in the five quality levels was 91.65%, illustrating that the development of atmospheric turbulence was at a middle level of adequacy in this period. When the friction velocity was 0-0.15, 0.15-0.30, and more than 0.30 m/s, respectively, the carbon flux showed grade variation characteristics; therefore, frictional wind speed at night can be divided into the three levels, and data within the 0 m/s and 0.30 m/s can be selectively weeded out. The proportions of data eliminated through the methods of sensor abnormal state, check analysis, and threshold analysis were 26.34%, 2.48%, and 8.57%, respectively, and the vast majority showed intermittent loss characteristics. There were no significant correlations between carbon flux and solar radiation or 5 cm soil temperature. Use of the linear interpolation and average daily variation methods could realize interpolation of missing data, but these could not reflect the true flux variations in a better manner. Conducting research on carbon flux in the non-growing season in a unique ecosystem in a desert-wetland environment surrounded by arid land and probing the carbon flux data characteristics and peculiarities could help provide the needed carbon flux research data and also a basis for improving research on carbon flux in China.