Simulations were conducted for the period January 1901 to December 2016 at a spatial resolution of 0.5 deg x 0.5 deg in latitude and longitude. Climate conditions (air temperature, precipitation, cloudiness, and humidity) were derived from global historical observation-based mesh data of CRU TS3.25 (Harris et al., 2014). Global mean atmospheric CO2 and CH4 concentrations during the simulation period were determined by historical data of the representative concentration pathways (Meinshausen et al., 2011), enabling a ready connection with future projection. Land-use conditions (fractional coverage of primary and secondary lands, pasture, and cropland and their mutual transitions) were derived from historical data in the Harmonized Land Use database (Hurtt et al., 2011). Spin-up by iterative calculation was performed at each grid cell until reaching an equilibrium state of the carbon cycle, for 300 to 4000 yr, under stationary climate and atmospheric conditions. A baseline experimental simulation was conducted following the spin-up, including the changes in climate, atmospheric greenhouse gases, and land-use conditions. This study focused on pan-Arctic (i.e., above 60degN) wetlands, as defined by the Global Lake and Wetland Dataset (Lehner and Döll, 2004). The 1-km-mesh data were aggregated into the model resolution of 0.5 deg grid cells . I assumed that land-use conversion from natural ecosystems to croplands occurs only in forests and grasslands (i.e., no landfill).

Ito, A.: Methane emission from pan-Arctic natural wetlands estimated using a process-based model, 1901–2016, Polar Science, in press. https://doi.org/10.1016/j.polar.2018.12.001