An incubation study of temperature sensitivity of greenhouse gas fluxes in three land-cover types near Sydney, Australia

Greenhouse gas (GHG) fluxes play crucial roles in regulating the Earth surface temperature. However, our understanding of the effect of land-cover and soil depth on the potential GHG fluxes and their temperature sensitivities (Q₁₀) is limited, which consequently increases the uncertainty to predict GHG exchange between soils and the atmosphere. In the present study, we sampled soils with contrasting characteristics from three land-cover types (wetland, grassland, and forest) and soil depths (0-10, 10-20, and 20-30 cm) from the Cumberland Plain near Sydney, Australia, and incubated at optimal (60%) water holding capacity at three temperatures (15, 25, and 35 °C). Overall, GHG fluxes and Q₁₀ values differed significantly among land-cover types and soil depths. CO₂ and N₂O emissions were highest in wetland followed by grassland and forest soils, and they decreased with soil depth. In contrast, CH₄ uptake was highest in grassland followed by forest and wetland soils, and it increased with soil depth. Combining the three major GHGs, the global warming potential in soil from wetland was higher than that from grassland and forest. Moreover, Q₁₀ values of CO₂ and N₂O emissions were: wetland > grassland > forest, while Q₁₀ value of CH₄ uptake showed the opposite pattern. Q₁₀ values of CO₂ and N₂O emissions and CH₄ uptake all increased with soil depth, demonstrating that subsoil has a higher potential for CO₂ and N₂O emissions and CH₄ uptake in a warming climate. While these experiments were conducted under ideally controlled laboratory conditions, results suggest that the large carbon stocks in wetland soils are vulnerable to loss and thus may amplify climate warming; upland soils are crucial CH₄ sinks and thus potentially mitigate climate change. In addition, the greater temperature sensitivities of CO₂ and N₂O emissions and CH₄ uptake in subsoil should be accounted for in carbon and nitrogen cycling models.