Improving the temporal variability of agricultural greenhouse gas emissions for Germany

Greenhouse gas (GHG) emissions, particularly carbon dioxide (CO₂) and methane (CH₄) from human activities, are the primary drivers of global warming. Additionally, methane contributes to ozone formation and therefore contributes to air pollution, posing risk to human health. Agriculture is a significant contributor to the global GHG emissions, with methane primarily emitted through enteric fermentation in livestock and manure management practices, while carbon dioxide largely arises from the use of machinery in various land management operations. Hence, to better understand and represent the intra - annual variability of GHG emissions within the agricultural sector, it is crucial to obtain spatial and temporal information about all contributing activities.

Within the ARTEMIS project we are further developing and refining a dynamic emission model to capture the spatio-temporal variability of anthropogenic GHG and air pollutant emissions in Germany and its surroundings. Inside the emission model the spatial allocator estimates the total yearly emissions with the gridded GHG emission inventory of TNO - CAMS for Europe and UBA - GRETA for Germany.

To account for temporal variability, different agricultural emission activities are parameterized individually. The temporal emission distribution for machinery use during land management operations gets estimated by deriving the emission timings from phenology observation data from the German Weather Service as well as using remote sensed phenology data from the COPERNICUS project. Additionally we incorporate an agricultural timer (Ge et al. 2020, 2022) developed to estimate the start of the growing season, which allows us to derive key dates such as sowing and manure application. The temporal variability of methane emissions from enteric fermentation are parameterized using literature-based emission factors linked to livestock feed intake and animal population data from national statistical agencies.

These emission datasets were integrated into a LOTOS-EUROS model simulation to demonstrate their added value. The comparison using the new dynamic emission model indicated an improved representation of intra-annual GHG concentration variability. Furthermore also the depiction of the diurnal concentration cycle showed a better alignment with measured concentrations. Additionally, evaluation against ICOS tall tower measurements revealed improvements in correlation (up to 0.06) and reductions in root mean squared error (up to 15%) between modeled and observed concentrations at nearly all stations. These findings highlight the importance of disentangling the agricultural GHG emissions into seperate subsectors, enabling a more accurate depiction of temporal variability in anthropogenic emissions. We conclude that further improving the spatio-temporal emission information should be extended on other sectors such as the industry and energy, the road traffic or the landfills as well.