hide
Free keywords:
-
Abstract:
No-tillage is often suggested as a strategy to reduce greenhouse gas emissions. Modeling tillage effects on nitrous oxide (N2O) emissions is challenging and subject to large uncertainties, as the processes producing the emissions are complex and strongly non-linear. Previous findings have shown deviations between the LPJmL5.0-tillage model and results from meta-analysis on global estimates of tillage effects on N2O emissions. Here we tested LPJmL5.0-tillage at four different experimental sites across Europe and the USA, to verify whether deviations in N2O emissions under different tillage regimes result from a lack of detailed information on agricultural management and/or the representation of soil water dynamics. Model results were compared to observational data and outputs from field-scale DayCent simulations. DayCent has been successfully applied for the simulation of N2O emissions and provides a richer data base for comparison than non-continuous measurements at the experimental sites. We found that adding information on agricultural management improved the simulation of tillage effects on N2O emissions in LPJmL. We also found that LPJmL overestimated N2O emissions as well as the effects of no-tillage on N2O emissions, whereas DayCent tended to underestimate the emissions of no-tillage treatments. LPJmL showed a general bias to over-estimate soil moisture content. Modifications of hydraulic properties in LPJmL in order to match properties assumed in DayCent, as well as of the parameters related to residue cover, improved the overall simulation of soil water as well as the N2O emissions simulated under tillage and no-tillage separately. However, the effects of no-tillage (shifting from tillage to no-tillage) did not improve. Advancing the current state of information on agricultural management as well as improvements in soil moisture highlight the potential to improve LPJmL5.0-tillage and global estimates of tillage effects on N2O emissions.
