hide
Free keywords:
biomass
carbon balance
heterotrophic respiration
net primary
production
Siberia
soil carbon
russian forests
CO2
flux
productivity
transport
budget
fire
northern
land
respiration
Abstract:
Northern Eurasia is the largest terrestrial reservoir of carbon, and its dynamics and interactions with climate are globally significant. We present five independent estimates of the contemporary carbon balance of central Siberia using three different methodologies: a landscape-ecosystem approach (LEA) that amalgamates comprehensive vegetation, soil, hydrological and morphological information into a Geographical Information System, linked to regression-based estimates of carbon flux; two Dynamic Global Vegetation Models (DGVMs); and two atmospheric inversions. Apart from one of the DGVMs, all methods produce estimates of the net biome productivity (NBP) that are consistent both amongst themselves and with a range of other estimates. They indicate the region to be a carbon sink with a NBP of 27.5 +/- 7.2 g C m-2 yr-1, which is equivalent to 352 +/- 92 Mt C yr-1 if considered representative for boreal Asia. This is comparable with fossil fuel emissions for the Russian Federation, currently estimated as 427 MtC yr-1, and implies that boreal Asia does not play the major role in the northern hemisphere land sink, typically estimated to be of magnitude 1.5-2.9 Gt C yr-1. The LEA and DGVM approaches produce very different partitioning of NBP into its component fluxes. The DGVMs find net primary production (NPP) to be nearly balanced by heterotrophic respiration, disturbance being a relatively small term pushing the system closer to equilibrium. In the LEA, heterotrophic respiration is significantly less than NPP, and disturbance plays a much larger role in the overall carbon balance. The use in the LEA of observationally based estimates of heterotrophic respiration and fire disturbance, along with a more complete description of disturbance fluxes, suggests that the partitioning derived by the LEA is more likely, and that improved process descriptions and constraints by data are needed in the DGVMs.