Large differences in variability between land surface models (LPJ-GUESS and ORCHIDEE) and inventory estimates: N₂O emissions and emission factors in Italy

In recent years, emission factors published by the Intergovernmental Panel on Climate Change (IPCC) are widely used by national inventory compilers as a Tier 1 methodology for estimating N₂O emissions at national levels. Whilst emission factors are straightforward to implement and offer several practical advantages, conventional emission factors tend not to account for the impacts of interannual climatic variability or spatial heterogeneity.

To investigate the added value of using spatially and temporally explicit processes included in land surface models, we assess the spatial and temporal variability of N₂O emissions associated with land management and extreme weather events over Italy for the 2010-2021 period. We compare N₂O emissions and N₂O emission factors estimated from two process-based land surface models, LPJ-GUESS and ORCHIDEE, against those estimated using national inventory data and published emission factors from IPCC guidelines.

Inventory-derived emissions do not show a trend over the study period and have limited interannual variation. In contrast, both models show a positive trend in emissions over the study period with interannual variability that extends well beyond the variability suggested by the inventory. We investigate the variability in emissions simulated by both models and assess whether this is indicative of a sensitivity to climate that is largely muted in IPCC based emission factors.

Both models show higher emissions from croplands than grasslands (total and per square meter) but higher emission factors from grasslands than croplands, indicating that the addition of (organic) fertilisers to pastures is more likely to be emitted as N₂O emissions than the same fertiliser added to croplands. We discuss key structural difference in how the models treat grasslands and pastures and how these discrepancies underscore the simplifications present in land surface model representations of these systems, especially regarding grazing, harvests, and manure management.

The substantial interannual variability in emission factors produced by both models exceed those estimated by inventory estimates and indicated by IPCC emission factors. These temporal patterns highlight the potential relevance of considering climate anomalies when using emission-factor methodologies, particularly with the increasing occurrence of extreme climate events.