Year-round simulated methane emissions from a permafrost ecosystem in Northeast Siberia

Wetlands of northern high latitudes are ecosystems highly vulnerable to climate change. Some degradation effects include soil hydrologic changes due to permafrost thaw, formation of deeper active layers, and rising topsoil temperatures that accelerate the degradation of permafrost carbon and increase in CO2 and CH4 emissions. In this work we present two years of modeled year-round CH4 emissions to the atmosphere from a northeastern Siberian region in the Russian Far East. We use a revisited version of the process-based JSBACH-methane model that includes four CH4 transport pathways: plant-mediated transport, ebullition and molecular diffusion in the presence or absence of snow. The gas is emitted through wetlands represented by grid cell inundated areas simulated with a TOPMODEL approach. The magnitude of the summertime modeled CH4 emissions is comparable to ground-based CH4 fluxes measured with the eddy covariance technique and flux chambers in the same area of study, whereas wintertime modeled values are underestimated by one order of magnitude. In an annual balance, the most important mechanism for transport of methane into the atmosphere is through plants (61 %). This is followed by ebullition (~35 %), while summertime molecular diffusion is negligible (0.02 %) compared to the diffusion through the snow during winter (~4 %). We investigate the relationship between temporal changes in the CH4 fluxes, soil temperature, and soil moisture content. Our results highlight the heterogeneity in CH4 emissions at landscape scale and suggest that further improvements to the representation of large-scale hydrological conditions in the model, will facilitate a more process-oriented land surface scheme and better simulate CH4 emissions under climate change. This is especially necessary at regional scales in Arctic ecosystems influenced by permafrost thaw.Additional funding support through the German Ministry of Education and Research (CarboPerm-Project, BMBF grant no. 03G0836G), and the AXA Research Fund (PDOC_2012_W2 campaign, ARF fellowship, Mathias Göckede)