High-resolution spatial patterns and drivers of terrestrial ecosystem carbon dioxide, methane, and nitrous oxide fluxes in the tundra

Arctic terrestrial greenhouse gas (GHG) fluxes of carbon dioxide (CO2), methane (CH4), and nitrous oxide (N2O) play an important role in the global GHG budget. However, these GHG fluxes are rarely studied simultaneously, and our understanding of the conditions controlling them across spatial gradients is limited. Here, we explore the magnitudes and drivers of GHG fluxes across fine-scale terrestrial gradients during the peak growing season (July) in sub-Arctic Finland. We measured chamber-derived GHG fluxes and soil temperature, soil moisture, soil organic carbon and nitrogen stocks, soil pH, soil carbon-to-nitrogen (C/N) ratio, soil dissolved organic carbon content, vascular plant biomass, and vegetation type from 101 plots scattered across a heterogeneous tundra landscape (5km(2)). We used these field data together with high-resolution remote sensing data to develop machine learning models for predicting (i.e., upscaling) daytime GHG fluxes across the landscape at 2m resolution. Our results show that this region was on average a daytime net GHG sink during the growing season. Although our results suggest that this sink was driven by CO2 uptake, it also revealed small but widespread CH4 uptake in upland vegetation types, almost surpassing the high wetland CH4 emissions at the landscape scale. Average N2O fluxes were negligible. CO2 fluxes were controlled primarily by annual average soil temperature and biomass (both increase net sink) and vegetation type, CH4 fluxes by soil moisture (increases net emissions) and vegetation type, and N2O fluxes by soil C/N (lower C/N increases net source). These results demonstrate the potential of high spatial resolution modeling of GHG fluxes in the Arctic. They also reveal the dominant role of CO2 fluxes across the tundra landscape but suggest that CH4 uptake in dry upland soils might play a significant role in the regional GHG budget.