Anthropogenic activities since Industrial Revolution have caused concurrent changes in atmospheric composition (rising greenhouse gases, i.e., carbon dioxide, methane, and nitrous oxide, reactive nitrogen) and climate (climate warming and changing precipitation regimes). Possible interactions among these driving factors pose great challenges for the projections of climate change-terrestrial C feedbacks. Manipulative Experiments with one or two driving factors have improved our understanding of the global change impacts on the terrestrial biosphere, but may not be able to reveal the terrestrial ecosystem dynamics and their underlying mechanisms under global change scenarios with multiple driving factors. Nevertheless, due to technical, financial, and other limitations, global change research has seldom examined the responses of terrestrial ecosystems to the commitment changes in the above driving factors using manipulative Experiments.
Global Change Ecology laboratory’s Global Change Impacts experiment has been established since May 2011 to manipulate four global change factors and each at two levels: atmospheric CO2 concentrations [CO2] [ambient (aCO2) and elevated [CO2] by 200 ppm (eCO2)], precipitation [ambient (aP) and 30% above the ambient precipitation (iP)], nitrogen (N) deposition [ambient (aN) and ambient plus 10 g N m-2 yr-1 (eN)], and temperature [unwarming (UW) and nighttime (18:00-06:00, local time) warming (W)]. A full factorial design was used for eCO2, iP, and W with eight treatment combinations and three replications for each treatment. Twenty-four 4-m × 4-m plots were set up and arranged into six rows and four columns, with a 4 m buffer zone between any two adjacent plots. In addition, a split-plot design, that each of the 24 plots was divided into two sub-plots, one with N addition and the other one without N addition, was used to manipulate the two-level N deposition. Therefore, the GCE Global Change Impacts experiment included 16 treatment combinations in total, including 1) control, 2) eCO2, 3) iP, 4) eN, 5) W, 6) eCO2 plus iP (eCO2iP), 7) eCO2 plus eN (eCO2eN), 8) eCO2 plus W (eCO2W), 9) iP plus eN (iPeN), 10) iP plus W (iPW), 11) eN plus W (eNW), 12) eCO2, iP, plus eN (eCO2iPeN), 13) eCO2, iP, plus W (eCO2iPW), 14) eCO2, eN, plus W (eCO2eNW), 15) iP, eN, plus W (iPeNW), and 16) eCO2, iP, eN, plus W (eCO2iPeNW).
Twenty-four octagon open-top chambers (OTCs; 4 m distance between any two parallel sides, 2 m height, and enclosing 13.2 m2 ground areas) were constructed in the 24 4-m × 4-m plots, using steel frames and optical glasses to manipulate [CO2]. In addition, three additional OTCs only with steel frames were built as ambient OTCs to estimate the effect of OTC on air and soil microclimate. In each eCO2 plot, pure CO2 was introduced into the OTC to achieve a diurnal CO2 enrichment of 200 ppm over ambient air from June-September of each year, controlled by LI-820 CO2 test system (LiCor, Lincoln, NE, USA) and an automatic control system (Luzhai Co., Beijing, China). Increased precipitation was applied with an automatic sprinkler system during each natural rain event to avoid changing rainfall frequency from June-September of each year. Nitrogen deposition was mimicked by applying NH4NO3 in the mid-June (5 g N m-2 yr-1) and mid-July (5 g N m-2 yr-1) of each year. All the 27 OTCs were divided into two sub-plots, one sub-plot with N addition and the other one without N addition. All the nighttime-warmed plots were heated continuously by 136 W m-2 of infrared radiation from mid-March to mid-November in each year, using 1.65-m (length) × 0.15-m (width) MSR-2420 infrared radiators (Kalglo Electronics Inc., Bethlehem, PA, USA) suspended 2.75 m above the ground.