Growing season CH4 and N2O fluxes from a sub-arctic landscape in northern Finland
Dinsmore, K. J., J. Drewer, P. E. Levy, C. George, A. Lohila, M. Aurela and U. M. Skiba. 2017. Biogeosciences 14:799-815
Abstract
Subarctic and boreal emissions of CH4 are important contributors to the atmospheric greenhouse gas (GHG) balance
10 and subsequently the global radiative forcing. Whilst N2O emissions may be lower, the much greater radiative forcing they
produce justifies their inclusion in GHG studies. In addition to the quantification of flux magnitude, it is essential that we
understand the drivers of emissions to be able to accurately predict climate-driven changes and potential feedback mechanisms.
Hence this study aims to increase our understanding of what drives fluxes of CH4 and N2O in a subarctic forest/wetland
landscape, exploring both spatial and temporal variability, and uses satellite derived spectral data to extrapolate from chamber
15 scale fluxes to a 2 x 2 km landscape area.
From static chamber measurements made during summer and autumn campaigns in 2012 in the Sodankylä region of Northern
Finland, we concluded that wetlands represent a significant source of CH4 (3.35 ± 0.44 mg C m-2 hr-1 during summer campaign
and 0.62 ± 0.09 mg C m-2 hr-1 during autumn campaign), whilst the surrounding forests represent a small sink (-0.06 ± <0.01
mg C m-2 hr-1 during the summer campaign and -0.03 ± <0.01 mg C m-2 hr-1 during the autumn campaign). N2O fluxes were
20 near-zero across both ecosystems and as such could not be accurately described as either consistent sinks or sources.
We found a weak negative relationship between CH4 emissions and water table depth in the wetland, with emissions decreasing
as the water table approached and flooded the soil surface. We attribute this relationship, which initially seems counter to much
of the current literature, to water tables being consistently above the level where a positive relationship would be expected.
Whilst conditions may appear optimal for CH4 production at higher water tables, reduced diffusivity may reduce the net
25 emissions, indicating a complex interaction of processes which combine to produce the net emission rate measured.
Temperature was also an important driver of CH4 with emissions increasing to a peak at approximately 12°C. Increases in
temperature beyond 12°C led to a subsequent reduction in emissions, indicating the presence of multiple interacting processes.
A multiple regression modelling approach was used to describe CH4 emissions based on spectral data from PLEIADES PA1
satellite imagery across a 2 x 2 km landscape. Our best model described 45% of spatial variability using blue and near infra30
red bands with the inclusion of the commonly described simple ratio (SR) and normalised difference vegetation index (NDVI).
When applied across the whole image domain we calculated a CH4 source of 2.05 ± 0.61 mg C m-2 hr-1. This was significantly
higher than landscape estimates based on either a simple mean or weighted by forest/wetland proportion (0.99 ± 0.16 mg C m-2 hr-1, 0.93 ± 0.12 mg C m-2 hr-1, respectively). Hence we conclude that ignoring the detailed spatial variability in CH4 emissions
within a landscape leads to a potentially significant underestimation of landscape scale fluxes.
