An investigation of temporal variability of CO2 fluxes in a boreal coniferous forest and a bog in central Siberia : from local to regional scale

Boreal forests and peatlands are the major types of vegetation in northern Eurasia. These ecosystems store large amount of carbon in their vegetation and soil, thus are essential ecosystems to understand the global carbon cycle. Compared to the other boreal regions, tower-based CO2 measurements in the Eurasian boreal ecosystems of central Siberia are very sparse. The objective of my thesis is to investigate temporal variability of CO2 fluxes at a boreal coniferous forest and a bog in Zotino, Russia. In Chapters 3 and 4, I present data on site-level CO2 flux measurements using the eddy covariance (EC) method. In Chapter 5, the site-level CO2 flux measurements were utilized to evaluate the reliability of regional CO2 flux estimates obtained from the modified Bowen Ratio (MBR) method. Chapter 3 investigates the diffuse radiation fertilization effect during the Siberian wildfire period from 2012-2013 using a data-driven model trained with CO2 fluxes and meteorological measurements. Key findings are: 1) Approximately 54-58% of variability in forest net ecosystem productivity (NEP) was controlled by photosynthetically active radiation (PAR), vapour pressure deficit (VPD), and diffuse fraction (fdif) of PAR during the growing season, 2) Incoming PAR decreased significantly at very high levels of fdif and high aerosol loading, 3) The diffuse radiation fertilization effect induced by clouds and aerosols increased NEP but this effect was less than a 10% increase, mainly due to sparse canopy structure and low leaf area index. Chapter 4 investigates the similarity and differences in abiotic controls of CO2 fluxes during the winter-spring transition period at a coniferous forest and bog. Key findings are: 1) Air temperature regulated CO2 flux variability in the forest, whereas surface peat temperature was the primary driver of CO2 flux variability in the bog, 2) Rapid net CO2 uptakes occurred when both air and soil temperatures exceed 5 oC, 3) Sporadic warm spells can lead to an earlier start of CO2 uptake, 4) Spring frost reduced the net ecosystem productivity in both ecosystems, however vegetation productivity increased again after frost. Chapter 5 presents the variability of regional CO2 fluxes from the profile measurements, combined using the MBR method. This study shows that diurnal cycles of profile measurements were useful for understanding the surface boundary layer structure. Since night-time mixing mostly did not extend beyond the top of the tall tower (304 m), night-time CO2 flux estimates can be reliably used without direct EC flux measurement. Despite the uncertainties in both daytime and night-time CO2 fluxes, magnitudes and patterns of diurnal cycles of regional CO2 flux estimates generally followed the EC CO2 flux measurements. Overall, this thesis shows that boreal ecosystems respond to changing environmental conditions in non-linear and complex ways. The results would be useful to evaluate CO2 fluxes from both processed-based biosphere models and inverse models. In order to characterize an annual carbon budget in Zotino, further efforts such as methane flux measurement at bog, flux footprint analysis, evaluation of the snow season CO2 flux, and uncertainty estimation in flux partitioning of the long-term net ecosystem exchange of CO2, would be necessary