Investigating the Impact of Climate and Extreme Weather on Greenhouse Gas Emissions from Irish Soils: An Evaluation of the ECOSSE Model

Soil is a complex material capable of storing and releasing large amounts of carbon, making it an integral part of the global carbon cycle. As soil is included in national inventory assessments where countries quantify their emissions, it is important that estimates of the soil carbon flux are as accurate as possible. Measuring the flux of carbon from every surface on the planet is unrealistic, therefore modelling emissions is the next best method of emissions assessment. Among modelling approaches, process-based modelling allows for the highest specificity of input data. Process-based models split soil carbon into pools with different decay rates influenced by environmental (biotic and abiotic) factors and management. The Estimation of Carbon in Organic Soils – Sequestration and Emissions (ECOSSE) model has previously been recommended as the optimum process-based model for simulating greenhouse gas (GHG) emissions from Irish soils. Results from the site-specific evaluation of the ECOSSE model for an Irish arable site showed a temporal offset between measurements and model outputs which could not be explained by model parameters, motivating further in-depth analysis of the model. This subsequent analysis indicated that the rate modifiers, which control the release of carbon from different pools in the soil, were functioning as intended, but highlighted issues with the model simulation of soil water. The modelled soil dried out fully when in reality it was above field capacity, attributable to erroneous simulation of evaporation on this well-drained sandy soil. In parallel, a spatial process-based model (GlobalECOSSE) was employed to derive national emissions estimates for multiple GHGs and CO2 equivalents for designated agricultural land uses on the island of Ireland. The justification for using the spatial process-based model was to provide an assessment of a model that had fewer input requirements and operated on a monthly timescale to simulate areal emissions. Re