Variations Of CO₂Fluxes And The Mechanisms Of Environmental Response In The Rain-fed Wheat Ecosystem Of Chinese Loess Plateau

Study on the variation of CO₂fluxes and its mechanism of environmental response playsignificant roles in understanding the process of the carbon cycling, and constructing anevaluation model for the influence of the climate change on the agricultural production in therain-fed agro-ecosystem. Based on the eddy covariance （EC） technique, a long-termcontinuous observation was conducted in the rain-fed winter wheat ecosystem of ChineseLoess Plateau during2008^-2009and2009^-2010crop years. The main results are as follows:1. Data quality analysis of CO₂fluxesThe results of non-stationary nonstationarity test and integral turbulence characteristicsshowed that the quality of observation data were reliable. The results of power spectrum andcospectrum analyses indicated that the response ability of EC system to high frequency signalmet the requirements of the study. The analysis of the energy balance closure showed that thedegree of energy balance closure was good, and the rate of energy balance closure （slope）during the2008^-2009and2009^-2010crop years and growing seasons was0.83,0.80,0.93,and0.91, respectively. The flux footprint analysis showed that the observation data mainlycame from the area of study interest （target area）, during peak growing seasons of winterwheat, there were about90%EC data came from the target area in the daytime, whereas therewere about77%of EC data came from the target area in the nighttime. The evaluation of dataquality suggested that the EC system was in good working conditions and the quality ofobservation data was reliable during the study period.2. Diurnal, seasonal and annual variations of CO₂fluxes（1） Diurnal variation of CO₂fluxes. Diurnal variation of CO₂fluxes displayed aU-shaped curve from returning green stage to grain filling stage. The maximum uptake peaksof CO₂occurred in heading stage of2008^-2009(-0.65±0.02mg CO₂m^-2s^-1) and jointingstage of2009^-2010(-0.51±0.02mg CO₂m^-2s^-1), respectively. The maximum emission peaksof CO₂also occurred in the heading stage of2008^-2009(0.18±0.01mg CO₂m^-2s^-1) andjointing stage of2009^-2010(0.16±0.01mg CO₂m^-2s^-1), respectively.（2） Seasonal variation of CO₂fluxes. The curve of net ecosystem CO₂exchange （NEE）displayed the trend of changed slightly in the early period, and then decreased rapidly in the middle period, finally increased gradually in the late period. The maximum daily mean NEE（absolute value） occurred in heading stage of2008^-2009(-5.68±0.11g C m^-2d^-1) and jointingstage of2009^-2010(-5.26±0.10g C m^-2d^-1), respectively, the maximum daily mean grossprimary productivity （GPP） also occurred in the periods, however, the maximum daily meanecosystem respiration (R_eco) occurred in the middle and late growing seasons. Overall, thetotal NEE for2008^-2009and2009^-2010growing seasons were^-263.2±13.8and^-218.9±11.5g C m^-2d^-1, respectively, indicating that the winter wheat ecosystem was a strong carbon sink.However, after considering the carbon in the grain after harvest, the ecosystem turned into aweak carbon sink (-65.4～^-36.2g C m^-2).（3） Annual variation of CO₂fluxes. The annual NEE for2008^-2009and2009^-2010cropyears were-71.6±5.7and-65.3±5.3g C m^-2, respectively. Summer fallow periods played animportant role in annual carbon budget, the winter wheat ecosystem acted as a strong carbonsource in the period. The accumulated R_ecoduring summer fallow periods cut the amount ofcarbon sequestered during growing seasons by25%^-30%. Considering the internal andexternal factors that influenced the annual carbon balance, the paper put forward twomeasures to improve the carbon sequestration ability in the rain-fed agro-ecosystem ofChinese Loess Plateau.3. Environmental response mechanism of CO₂fluxes（1） Response of CO₂fluxes to meteorological factors. The regression coefficientsbetween NEE and photosynthetically active radiation （PAR） was0.72to0.89in the middlegrowing seasons of winter wheat （returning green stage to grain filling stage）. Furthermore,there were different trends of daytime NEE at different classes of air temperature （Ta） andvapor pressure deficit （VPD）. The net CO₂uptake rate gradually increased with the increasingof the Taunder the condition of15＜Ta＜25℃, whereas it began to decrease when Ta＞30℃.The net CO₂uptake rate increased with the increasing of the VPD under the condition of0＜VPD＜1kPa, whereas it began to decrease when VPD＞2kPa.（2） Response of CO₂fluxes to soil factors. On the one hand, the ecosystem respirationwas sensitive to the changes in soil temperature （T_s） during winter wheat growing seasons（except for the overwintering stage, grain filling stage and ripening stage） and summer fallowperiod. On the other, during the peak growing seasons of winter wheat （jointing stage toheading stage）, there were different relationships of daytime NEE-PAR, and nighttime R_eco-T_sunder the different soil moisture conditions. On the range of soil water content from0.15to0.21m3m^-3, the highest correlations of NEE-PAR, and R_eco-T_swere observed.4. Response mechanism of CO₂fluxes to rainfall events（1） The ecosystem CO₂fluxes were sensitive to the effective rainfall events （daily precipitation＞5mm） during the growing seasons. From returning green stage to grain fillingstage, the maximum half-hourly R_eco1^-3 days after the effective rainfall events were70%-630%higher than that before the effective rainfall events. The reasons for the pulseresponse of R_ecowere as follows: the effect of physical alternative, the mechanism of organicmatter mineralization, the response mechanism of microorganism to drought stress. However,during the summer fallow periods, the daliy R_ecovalues1^-2 days after heavy rainfall event（daily precipitation＞40mm） and continuous rainfall event （rainfall days＞4d） were lowerthan that before these rainfall events.（2） The effects of the antecedent soil water content. During the middle and late growingperiods, the antecedent soil water content explained about40%-42%variation in R_eco,whereas precipitation explained about25%^-27%variation in R_eco. During the summer fallowperiods, the antecedent soil water content explained about50%-63%variation in R_eco,whereas there were almost no correlation between the precipitation and the R_eco. Thissuggested that the effect of antecedent soil water content on the variation of R_ecowas higherthan the effect of precipitation on the change of R_ecofor the rain-fed winter wheat ecosystem.In summary, the rain-fed winter wheat ecosystem had strong carbon sequestration ability.Meanwhile, there were obvious diurnal, seasonal, and annual variations in CO₂fluxes. CO₂fluxes were affected by the factors such as meteorological factors, soil factors, and rainfallevents, and displayed a complex response mechanism. The research results may providetechnical support and theoretical guidance for understanding the carbon cycling processes,and constructing the technical system of carbon emission reduction in the rain-fedagro-ecosystem of the Chinese Loess Plateau.