The aim of Task 4.1 is to derive emission factors based on bottom-up land surface modelling, using two process-based dynamic global vegetation models (DGVMs), LPJ-GUESS and ORCHIDEE. In recent years, emission factors have been proposed as a useful methodology for estimating greenhouse gas emissions under different land-management strategies. Whilst emission factors are straightforward to implement and offer several practical advantages, conventional emission factors do not account for climatic variability (e.g., increased runoff following heavy precipitation in a wet year may reduce N₂O emissions) or spatial heterogeneity (differences in soil properties also impact CO2 and N2O emissions). To investigate the extent to which incorporating spatially and temporally explicit processes improves the robustness of GHG emission estimates, we assess the added value of representing spatial and temporal variability associated with land management and extreme weather events using two case studies: carbon emissions from forestry in Sweden and nitrogen emissions from agriculture in Italy. Model simulations were evaluated against available observational data. The simulations demonstrated the usefulness of DGVMs for estimating spatial and temporal changes in carbon stock (forest in Sweden) and N2O emissions (agriculture in Italy), as influenced by weather and climate extremes and different management methods. The model comparison revealed that uncertainties in model parameterization can significantly affect simulation outcome, highlighting the need for further model development, calibration, and evaluation.
Jönsson et al. (Fri,) studied this question.