Introducing trees to agricultural lands increases greenhouse gas emission during spring thaw in Canadian agroforestry systems.

The role of agroforestry systems in mitigating greenhouse gas (GHG) emission from agricultural soils during spring thaw (early April to mid-May) has been poorly studied. Soil CO2, CH4 and N2O fluxes were measured from treed areas and adjacent herblands (areas without trees) during spring thaw in 2014 and 2015 at 36 agroforestry sites (12 hedgerow, 12 shelterbelt and 12 silvopasture) in central Alberta, Canada. Fluxes of those GHGs varied with agroforestry systems and land-cover types. We found greater CO2 emission (P < 0.001) and CH4 uptake (P < 0.05), but lower N2O emission (P < 0.01) in the silvopasture than in the hedgerow and shelterbelt systems, with no difference between the last two systems. Treed areas in general had greater CO2 emissions (P < 0.001) and CH4 uptake (P < 0.01), and lower N2O emissions (P < 0.001) than the herblands. Soil temperature, moisture content, organic C content and soil available N concentration affected GHG fluxes. The global warming potential (GWP) was greater (P < 0.05) in the silvopasture than in the hedgerow or shelterbelt systems over the two spring thaw seasons examined, and greater (P < 0.05) in the treed areas than in the herblands during the cool spring in 2015. However, the GWP per unit soil organic C was lower in the treed areas (0.004-0.101%) than in the herblands (0.005-0.225%). As compared to previously reported mean growing season GHG emission (15.4 g CO2-eq m-2 day-1), the GWP of these land uses during spring thaw was small (<5% of the annual GWP) due to the short spring period (6 weeks) and the small GHG emission (2.5 g CO2-eq m-2 day-1). Although GHG emissions during spring thaw were small compared to those in the growing season, they should not be ignored.