Using vegetation cover type to predict and scale peatland methane dynamics.

Abstract:

Permafrost ecosystems contain about 50% of the global soil carbon. As these northern ecosystems experience warmer temperature, permafrost thaws and may result in an increase in atmospheric methane. We examined a thawing and discontinuous permafrost boundary at Stordalen Mire, in Northern Sweden, in an effort to better understand methane emissions. Stable isotope analysis of methane in peatland porewater can give insights into the pathway of methane production. By measuring δ¹³CH₄ we can predict whether a system is dominated by either hydrogenotrophic or acetaclastic methane production. Currently, it is a challenge to scale these isotopic patterns, thus, atmospheric inversion models simply assume that acetoclastic production dominates. We analyzed porewater samples collected across a range of vegetation cover types for δ¹³CH₄ using a QCL (Quantum Cascade Laser Spectrometer) in conjunction with highly accurate GPS (3-10cm) measurements and high-resolution UAV imaging. We found δ¹³CH₄ values ranging from -88‰ to -41‰, with averages based on cover type and other vegetation features showing differences of up to -15‰. We then used a computer neural network to predict cover types across Stordalen Mire from UAV imagery based on field-based plot measurements and training samples.. This prediction map was used to scale methane flux and isotope measurements. Our results suggest that the current values used in atmospheric inversion studies may oversimplify the relationship between plant and microbial communities in complex permafrost landscapes. As we gain a deeper understanding of how vegetation relates to methanogenic communities, understanding the spatial component of ecosystem methane metabolism and distribution will be increasingly valuable.