Large uncertainties exist in carbon-water-climate feedbacks in cold regions, partly due to an insufficient understanding of the simultaneous effects of climatic and biotic controls on water and carbon dynamics. The 10-year growing season flux data were analyzed to evaluate the relative contributions of climatic and biotic effects on the variability of water vapor (ET) and net ecosystem CO2 (NEE) exchanges over a humid alpine deciduous shrubland on the northeastern Qinghai-Tibetan Plateau. The results showed that the alpine shrubland ecosystem acted as a water source and a carbon sink during the growing season, and its potential ET and NEE ranged from 161.4 mm and -41.0 g Cm-2 to 408.0 mm and -278.4 gCm(-2) at a 95% confidence interval, respectively. The average 8-day ET and NEE during the early growing season (June to July) were both significantly (P < 0.05) more than those of the late growing season (August to September). And the slopes of ET and NEE against the Julian day during the two growth stages also changed significantly (P < 0.01). Such asymmetric manners of ET and NEE during the two growth stages were probably related to the seasonal variations of net radiation (Rn) and vegetation growth (satellite-derived enhanced vegetation index: EVI), respectively. The structural equation models showed that the seasonal variations of 8-day ET were jointly determined by Rn and vapor pressure deficit (VPD), as partly indicated by a modest decoupling coefficient (0.54 +/- 0.03). The seasonal variability in 8-day NEE was controlled by the combinations of EVI and growing season degree days (GDD). The standardized coefficient of the direct effect of EVI on ET was 0.16, much less than the corresponding value (0.51) on NEE, suggesting that a weak coupling between ET and NEE arose likely because water vapor loss were about half controlled by surface evaporation, whereas CO2 flux were largely regulated by vascular plant activity. Our results highlighted the asymmetric sensitivities of ET and NEE during the early and the late growing season, and the weak coupling of water loss and carbon fixation during the whole growing season. These findings would provide a new sight to understand the growth stage-dependent responses of water budget and carbon sequestration to grazing management and climate change in humid alpine shrublands.

This study investigated the CO2 exchange over a 10-year period (2005-2014) inside and above a larch-dominant forest in the central Lena river basin, eastern Siberia. A wet-soil condition, such as that found in the active layer (seasonally thawed soil layer of upper permafrost), containing unusually high soil water close to saturation and partial surface waterlogging, was prolonged during the warm season of 2005-2009. In later years, the soil layer closer to the ground surface became dry (similar to 10% volumetric water content), although the deeper part remained relatively wet (similar to 30%). We quantitatively compared the whole forest and the understory CO2 exchanges to detect the separate effects of excessive soil waters on the overstory and understory vegetation. The conventional light and temperature response functions for half-hourly CO2 fluxes, that is, the net ecosystem exchange of daytime and night-time, respectively, were applicable to the understory observations. Comparison of the fitting parameters of the light response function at two levels revealed a smaller maximum net ecosystem exchange (NEE) under light saturation with a steep response under weak light conditions for the understory. The CO2 exchanges at the understory increased from the wet-soil period to the drying soil period by 46% (1.3 g C m(-2) d(-1)) of gross primary production (GPP) and 29% (1.2 g C m(-2) d(-1)) of ecosystem respiration (ER), while no trend was found in the ecosystem scale fluxes. These increases were due to an increasing understory biomass, changes in plentiful light and soil water in the inside-canopy environments, and enhanced turbulent mixing. The decline in the larch contribution could be compensated for by the understory growth and the remaining wetness of the active layer, which indicated that the interactions between the larch and the understory supported the stability of carbon cycles in this forest ecosystem.

Climate change will have important effects on arctic productivity and greenhouse gas exchange. These changes were projected by the model ecosys under an Special Report on Emissions Scenarios (SRES) A2 scenario over the 21st century for a landscape including an upland tundra and a lowland fen at Daring Lake, NWT. Rising temperatures and precipitation caused increases in active layer depths (ALD) and eventual formation of taliks, particularly in the fen, which were attributed to heat advection from warmer and more intense precipitation and downslope flow. These changes raised net primary productivity from more rapid N mineralization and uptake, driven by more rapid heterotrophic respiration and increasing deciduous versus evergreen plant functional types. Consequently, gains in net ecosystem productivity (NEP) of 29 and 10gCm(-2)yr(-1) were modeled in the tundra and fen after 90years. However, CH4 emissions modeled from the fen rose sharply from direct effects of increasing soil temperatures and greater ALD on fermenter and methanogenic populations and from indirect effects of increasing sedge growth, which hastened transfer of CH4 through porous roots to the atmosphere. After 90years, landscape CH4 emissions increased from 1.1 to 5.2gCm(-2)yr(-1) while landscape NEP increased from 34 to 46gCm(-2)yr(-1). Positive feedback to radiative forcing from increases in CH4 emissions more than offset negative feedback from increases in NEP. This feedback was largely attributed to rises in CH4 emission caused by heat advection from increasing precipitation, the impacts of which require greater attention in arctic climate change studies.